An In erna tonal Journal of Biological and Biomedical Research

ro INTERNATIONAL MEETING ON <.

OLECULAR EPIDEMIOLOGY ANDEV UTTIONARY GENETlCS

INFECTIOUS DISEAS. MEEGID-3

An In erna tonal Journal ofBiological and Biomedical Research

MEMORIAS DO INSTITUTO OSWALDO CRUZ

EDITORA

IE]I';;;;;=- ==- -- -- -= ::"ocr~uz

Programa de Apoio a Publica'loes Cientificas

SeT/PR @ CNPq IEJ FINEP

re:, 1998 Fundacao Oswaldo Cruz - FiocruzMinisterio da Saude

Mern, Inst. Oswaldo Cruz is covered by

Biological AbstractsCurrent Contents - Life SciencesEntomology AbstractsExcerpta MedicaHelminthological AbstractsIndex MedicusIndex Medicus Latino-Americano

Lit. Latino-Americana - LILA CSMicrobiology AbstractsProtozoological AbstractsReview ofApplied EntomologyScience Citation IndexTropical Diseases BulletinZoological Abstract

Articles from this journal are available through

• the authors• the Genuine Articlef", 350 I Market Street, Philadelphia, PA 19104, USA• the Biosis Document Express, 41575 Joy Road, Canton, MI 48187-2062, Fax: 313-459.8990

E-mail: [email protected]• SourceOne Uncover, Fax: 303-758.5946 or internet: [email protected]

On line access

The address of the home page of the Mem6rias is:http.z/www.pobox.com/vmemoriaswhere the full text of articles can be accessed.

Editorial Office

Administrative EditorsConsulantAdministrative Office

Desktop PublishingOn line version:

Luciane CB SoaresMarly de C Wi11coxWanilda Blandy CorreaGeiza Helena R NevesPaulo Marcio MoreiraLeonardo M Rocha

Subscription information

The subscription price for one volume of six issues including postage is:• Institutional - Brazil R$ 70,00, South America

US$ 80.00, rest of world US$ 120.00.• Personal - Brazil R$ 35,00, South America US$

45.00, rest of the world US$ 60.00.It is possible for Institutions which sponsor scientific journals to exchange issues with the Mem6riasin lieu of payment.

All manuscripts and correspondence concerning editorial matters, subscriptions, and changes ofaddress should be sent to

Mem6rias do Instituto Oswaldo Cruz, Av.Brasil4365, 21045-900 Rio de Janeiro, RJ, Br~;il .~Phone: +55-021-598.4335 Fax: +55-21-280.5048 E.mail: [email protected]

~----------------------------_. -_._-_.

MEMORIAS DO

INSTITUTO

OSWALDO

CRUZ

ISSN-0074-0276

President ofFIOCRUZ: Eloi de Souza GarciaDirector of the lnstituto Oswaldo Cruz: José Rodrigues CouraEditor: Hooman Momen

Editorial Board (Rio de Janeiro)

Ana Carolina VicenteCarlos André SallesCarlos M MorelClaude PirmezClaudio J StruchinerClaudio T Daniel RibeiroDelir Corrêa GomesEloi S GarciaGabriel Grimaldi Jr

Herman G SchatzmayrJosé Rodrigues CouraMaria de Fatima CruzMaria Nazareth L MeirellesMarli Maria LimaRenato S Balào CordeiroRicardo L de OliveiraYara M Traub-Cseko

International Advisory Board

Alberto CC Frasch (Buenos Aires)Alain de Chambrier (Geneva)Allan A Yousten (Blacksburg)Allen W Cheever (Bethesda)Ana Flisser (Mexico)Anthony J Shelley (London)Antoniana U Krettli (Belo Horizonte)Bianca Zingales (Sào Paulo)Brian WJ Mahy (Atlanta)Daniel Colley (Atlanta)Daniel Camus (Villeneuve d'Asc)David H Molyneux (Liverpool)Diane McMahon-Pratt (New Haven)Eric Milstrey (Fort Polk)Franklin A Neva (Bethesda)Gerusa Dreyer (Recife)Humberto Guerra (Lima)JD Thomas (Brighton)Jeffrey Shaw (Brasilia)John R David (Boston)

Jorge Kalil (Sào Paulo)José Luis Ramirez Ochoa (Caracas)KP Chang (Chicago)Luis Travassos (Sào Paulo)M Goreti Rosa-Freitas Sibajev (Boa Vista)Philippe Desjeux (Geneva)R Killick-Kendrick (Ascot)Ralph Harbach (London)Ralph Lainson (Belém)Ruth S Nussenzweig (New York)Rodrigo Zeled6n (San José)Socrates Herrera (Cali)Stephen M Beverley (St Louis)Terry Klein (Washington)Thomas C Cheng (Charleston)Toby Barrett (Manaus)Victor Nussenzweig (New York)Wallace Peters (St. Albans)Walter A Boeger (Curitiba)Zigman Brener (Belo Horizonte)

Publication date

Vol. 93 (5) 10 September 1998

3rd International Meeting on Molecular EpidemlOlogy and EvolutionaryGenetlcs of Infectious Disease,.

.~;J~;~~.

~ ::;" ~~~~~;,;:

. 3rd INTERNATIONAL MEETlNG ON •'t10LECULAR EPiDEMIOlOGY MJD',

E\IOlU rnONARY GENETlCS' Of ...INfEcnOUS DISEASES ";;' '

MEEGID 3 -

:~\ ~~: ,,":"t:f'~'~: .~.. "'t:..:~; "':,.\::j>4.~ f' T~

(1998 Fundaçào Oswaldo Cruz - Flocruz. Ail nghts reserved. ISSN 0074-027611 998/ $4.00

This journal and the indlvldual contributions contamed in it are protected by the copynght of FlOcruz. and the following termsand conditions apply to their use:

PhotocopyingSmgle photocopies of smgle articles may be made for personal use as allowed by national copyright laws. Permission of thepubh,her and payment of a fee IS reqUired for ail other photocopying. including multiple or systematic copying. copying foradvertismg or promotlOnal purposes. resale, and ail forms of documcnt dehvery. Special pennissions are available for educatlonal institutions that wish to make photocopies for non-profit educatIOnal classroom use. In the USA, users may clear pennisslOns and make paymentthrought the Copynght Clearance Center fnc. (CCC account number 1000011958),222 RosewoodDrive. Danvers MA 01923, USA. Tel. (508) 750-8400; Fax: (508) 750-4744. In other countnes where a local copyrightclearance centre eXlsts, please contact it for information on required permission and payments.

Derivative WorksSubscnbers may reproduce tables of contents or prepare lists of articles including abstracts for internaI circulation within theirmslItutions. Permission is required for resale or dlstnbution outside the mstltution. PermiSSIOn IS required for ail other derivatlve works. includmg compilations and translations. Permission IS reqUired to store e1ectronically any matenal contained in thisjournaL including any article or part of an article. Except as outlmed above. no part of this publication may be reproduced,stored in a retneval system or transmitted m any form or by any means. e1etronic. mechanical, photocopying, recordmg orotherwise. without prior written permiSSIOn.

FICHA CATALOGRAFICA

Mem6nas do InslItuto Oswaldo Cruz. v. 1. 1909- Rio de Janeiro: Fiocruz: Secretana das Mem6rias, 1909 - v.: Il.; 27cm.

Irregular: 1909-1976.Interromplda: 1976-1980.Tnmestral. 1981-1994.Bllnestral a partir do Vol. 90. 1995.ISSN 0074-0276.

1

1. Medicma experimental - Pen6dicos - Brasi!. 2. Pesquisa cientifica - Peri6dicos - Brasil.

!. Fundaçào Oswaldo Cruz. II. Instituto Oswaldo Cruz. Secretaria das Mem6rias.COD - 001.4305

-619.05

COU - 61.001.6(81 )(05)- 001.891/.2(81)(05)

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 565, Sep./Oct.1998 565

3rd International Meeting on Molecular Epidemiology andEvolutionary Genetics of Infectious Diseases

7-10 June 1998 - Rio de Janeiro, RJ

FOREWORD

The Third International Workshop on Molecular Epidemiology and Evolutionary Genetics oflnfectious Diseases was held at the Hotel Gloria in Rio de Janeiro, Brazil, from June 7 to 10, 1998. The titleof this third meeting was broadened to coyer infectious diseases so as to include both vector and hostaspects as weil as pathogenic micro-organisms.

The Il plenary lectures and 14 round-tables presented during this workshop covered a wide varietyof diseases from a number of different perspectives. The abstracts received from over 20 countries andsix continents attested to the popularity and widespread appeal of these meetings. Brazil was an appropriate setting for this meeting as most of the infectious diseases discussed during this workshop areeither emerging, re-emerging or endemic in this country. These international meetings started from anidea shared between Michel Tibayrenc and Altaf LaI. The first meeting was held in June 1996 in AtlantaGA, USA and the second in Montpellier in May 1997. These two meetings were co-sponsored byORSTOM (the National French Agency for scientific research in developing countries), CNRS (theNational French Agency for basic research) and the Centers for Disease Control and Prevention (CDC).For this third meeting the Oswaldo Cruz Institute of the Oswaldo Cruz Foundation joined the originalsponsors. The Oswaldo Cruz Institute was founded in Rio de Janeiro in the beginning ofthis century andhas a distinguished record of achievements in the field of research and control of infectious diseases.Since these meetings were founded the importance of the molecular epidemiological and evolutionarygenetic approach to infectious diseases has been increasingly demonstrated in the identification andcontrol ofmany outbreaks. Several practical examples of the use of this approach were given in the talksduring the meeting. The full program and abstracts of ail the presentations (plenary lectures, roundtables and posters) are available at the web-site for the event http://www.dbbm.fiocruz.br/www-mem/meeting. In addition the speakers of the oral presentations were invited to submit manuscripts to beconsidered for publication in the Memorias. In order for the manuscripts to be published shortly after themeeting a deadline was imposed for the submission of the manuscripts. Due to the short time availablemany speakers were unable to make submissions, however those who sent manuscripts and which wereapproved for publication are included in this issue of the journal.

We would like to thank the following organizations for their financial support ofthis meeting: CNPq(The Brazilian National Research Councii), FAPERJ (The State of Rio de Janeiro Research Council),CAPES (The Brazilian Agency for post-graduate studies), FNS (The Brazilian National Health Foundation), INTERACTIVA Biotechnologie Gmbh and Sigma Chemical Co. (Brazil). We would also like toacknowledge the support of the Brazilian societies of Mycology, Virology and Microbiology.

From the many comments received both during and after the workshop it can be concluded that themeeting was very successful, both in terms of the high quality of the presentations and in the opportunities provided by the intervals and social program for contacts and interactions among the participants.The National press also took great interest in the workshop and articles appeared in newspapers andmagazines, before, during and after the meeting as weil as material, on television news and radio.

The success ofthis meeting bodes weil for the next workshop which is planned for Dakar, Senegal inJune 1999. Further information about this meeting can be obtained from Dr Michel Tibayrenc (fax: +334-67416299) or from the organizers be10w.

The organizers

Hooman Momen([email protected])

AltarA Lai([email protected])

Michel TIbayrenc([email protected])

3rd INTERNATIONAL f'lIEETING ON MOLECULAREPIDEMIOLOGY AND EVOLUTIONARY GENETICS

OF INFECTIOUS DISEASES

f'lIEEGID-3

HOTEL GLORIARIO DE JANEIRO, JUNE 7-10, 1998

tj Mlnisterio da sauce

Fundacao Oswaldo Cruzlnsntuto Oswaldo Cruz

coc(enters for DiseaseControl and Peevenncn

~ C"Ct\TR[r",'..TlQ"Ii",1

DE L-. r,ElHEJKHE"UH..TlHOU~

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 567-576, 5ep./Oct. 1998 567

Evolutionary Control of Infectious Disease: Prospects forVectorborne and Waterborne Pathogens

Paul W Ewald/+, Jeremy B Sussman, Matthew T Distler, Camila Libel, WahidP Chammas, Victor J Dirita*, Carlos André Salles**, Ana Carolina Vicente***,

Ingrid Heitmann****, Felipe Cabello*****

Department of Biology, Amherst Collegc, Amherst. MA 01002-5000, USA *Laboratory of Animal Mcdicinc,UniversIty Michigan School of Mcdicll1c, Ann Arbor, MI 48109, USA **Dcpartamento dc Bioquimica e

Biologla Molecular ***Oepartamento dc Genétlca, Instituto Oswaldo Cruz, Av. Brasi14365, 21045-900 Rio dcJaneiro, RJ, Brasil ****Sub Departamcnto de Microbiologia Clinica, Instituto de Salud Publica, Maraton 1000,

Nunoa Santiago, Chilc *****Dcpartment of Immunology, New York MedIcal College, Vahalla, NY, USA

EvolutionaJY theOl'Y ma)' contribute to practical solutionsfor control ofdisease by identlji.'ing intervelltions that may cm/se pathogens to evolve to reduced virulence. Theory predicts, for example, thatpathogens transmitted by water or arthropod vectors shOl/ld evolve to relatively Illgh levels ofvirulencebecal/se Sl/ch pathogens can gain the evolutionary bene/its ofrelative~l' high levels ofhost exploitationwhile paying tittle pricefrom host i/lness. The entrance ()( Vibrio cholerae into South America in 1991has generated a natural experiment that al/olVs testing olthis idea by determining whether geographicand temporal variations in toxigenicity correspond to variation in the potential fàr waterborne transmission. Pretiminary studies show such correspondences: toxigenicity is negatively associated withaccess to uncontaminated water in Brazil; and in Chile, where the potential for waterborne transmission is particularly low, toxigenicity of .l'trains declined between 1991 and 1998. /n theOl)' vectorproofing ofhouses should be simitarly associated with benignity o(vectorbOlïle pathogens, such as theagents ofdengue, malaria, and Chagas , disease. These preliminary studies draw attention to the needfor definitive prospective experiments to determine whether interventions such as provisioning ofuncontaminated water and vector-proofing ofhOllses callse evolutionar,v reductions in virulence.

Key words: infectious diseases - control - pathogens - waterborne transmission

AN EVOLUTIONARY APPROACH TO VIRULENCE

The ongoing synthesis of epidemiology, molecular biology, and evolutionary biology promises to improve our understanding of the temporaland geographic variation in pathogens and the diseases they cause. From a practical viewpoint thisimproved understanding may prove usefui in identifying new possibilities for the control and prevention of infectious disease. One aspect of theinfectious process that seems particularly amenableto this control is virulence, which is defined hereas the level of haml to the host. Although viru-

This study was supported by two grants from LeonardX Bosack and Bette M Kruger Charitable Foundation(PWE), an Amherst College Faculty Research Award(PWE), a Hughes student fellowship (CL), and theWebster Fund of the Biology Departmcnt at AmherstCollege.+Corresponding author. Fax: +413-542.7955Received 15 June 1998Accepted 30 July 1998

lence depends on the interplay between pathogenand host characteristics, it is useful to consider theinherent virulence of a pathogen as the pathogen'scontribution to this harmfulness. In practice thiscontribution is not separable from the host in whichthe harmfulness is assessed, yet conceptually reference to the inherent virulence of pathogens inthe context of the spectrum of infectious agents.The smallpox viruses are inherently more hannfulthan rhinoviruses even though some of the mildestsmal1pox virus infections may be no more severethan the most severe rhinovirus infections.

Evolutionary considerations emphasize that theinherent virulence of pathogens should depend ona tradeoffbetween fitness benefits and fitness coststhat are associated with particular levels of virulence. The fitness benefits are accrued throughincreased replication of the genetic instructions forthe characteristic. Costs are typically accruedthrough reductions in the transmission of the genetic instructions, for example, due to negativeeffects of hast illness on pathogen transmission.Evolutionary theory generally does not propose thatvirulence per se is beneficial. Rather, the logic

568 Evolutionary Control of Infectious Disease • Paul W Ewald et al.

assumes that disease organisms may benefit byexploiting their hosts. Such exploitation al10ws adisease organism to secure resources that it can useto reproduce, and thereby contribute more copiesof the instructions for that exploitation into futuregenerations. These fitness benefits ofexploitationare weighed against the costs. The illness causedby intense levels ofexploitation may make the hostimmobile, host mobility may be necessary for transmission to new hosts (as is the case, for example,with the common cold virus). In this case pathogen variants that exploit hosts so intensely that theycause host immobility may get more resources inthe short run, but lose in the slightly longer runbecause of reductions in transmission. Pathogensthat do not rely on host mobility for transmissionpaya relatively low price iftheir exploitation immobilizes the host. According to the tradeoff reasoning presented above, pathogens in such categories should be particularly virulent. One of thesecategories involves waterbome transmission.

WATERBORNE TRANSMISSION

Waterbome transmission al10ws diarrhealpathogens to be transported from immobilized infected hosts to uninfected hosts. Where water supplies are not protected, a person with incapacitating diarrheal illness will release the diarrheal pathogens into clothes, bed sheets, or containers for collecting excreta. These items then tend to be removed by attendants and washed in bodies ofwater such as canals or rivers, which may be used assources of drinking water or may flow into supplies of drinking water. Either way, the cycle iscompleted when susceptible individuals drink thecontaminated water. In this situation, hightlyexploitative (and hence highly virulent) pathogen

variants should be favored by natural selectionbecause the benefits of intense exploitation aregreat and the costs of exploitation are small. Thebenefits are great because large numbers ofsusceptibles can be infected by the increasedJ1umbers of propagules in the water. The costs are lowbecause the incapacitating iIIness associated withthis propagule production should have relativelylittle negative effect on the waterbome transmission of the propagules-rather than relying on themobility of the infected individuals to enact transmission, the pathogens are using the mobility ofthe attendants and the water.

This hypothesized effect of waterbome transmission has been tested by determining whetherthe lethality ofbacterial agents ofhuman diarrheais positively correlated with the degree to whichthey are waterbome (Ewald 1991). Fig. 1 showsthat this correlation exists. Variation in the virulence ofhuman diarrheal diseases can thus be explained in an evolutionary sense by variation inthe degree to which different diarrheal pathogensare waterbome. This association offers sorne insight into the variation in virulence that occursamong diarrheal bacteria, but perhaps more importantly it suggests a new means for lessening thedamage associated with diarrheal diseases. By reducing the potential for waterbome transmissionwe may be able to force diarrheal pathogens toevolve reduced virulence.

Whether this possibility is feasible depends onthe validity of applying the trend apparent acrossthe broad spectrum ofdiarrheal pathogens depictedin Fig. 1 to particular pathogens. Would a particular kind of pathogen evolve reduced virulence inresponse to a reduced potential for waterbornetransmission? If so, what time period would be

16 •14~~12c0

1'5 10~

8.~ •~6 •If)..c 4roQ)

0 2 •- -, .o ----_ -- --- ----, --- -- ---,----o ~ ~ 00 W

Outbreaks involving waterborne transmission (%)

Fig. 1: waterborne transmission and mortallty of diarrheal bacteria of humans. Pathogens ordered from most to least waterborneare classical v/hria cha/erae, Sh/gel/a dysenteriae type 1. Sa/monel/a typhl, el tor V. cha/erac. Shigel/aj/exneri, Shigel/a sonnei.entcrotoxigenic E5cherichia coli, Campy/obacterjejum, and nontyphoid Sa/monel/a (for other details see Ewald 1991).

Mem Insl Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep'/Oct. 1998 569

required? If this period were a few years, thenefforts to improve access to clean water supplieswould have evolutionary effects over an intervalthat is comparable to the intervals envisioned forconventional, nonevolutionary interventions.

V cholerae is an excellent study subject for thiskind of analysis particularly because its virulenceis largely attributable to its toxin production. Theinherent virulence ofparticular strains therefore canbe assessed by quantifying levels oftoxin production in vitro. Toxin production generates an effiuxof fluid into the smal1 intestine, which appears toprovides two benefits to V cholerae: (1) it flushesout competitors throughout the intestinal tract, allowing V cholerae to pass down and out of thetract intact, and (2) it creates a fluid stool that probably facilitates transmission by contamination ofthe external environment and dissemination inwater supplies. V cholerae can persist in the intestine during this tumult because it can swim andadhere to the intestinal lining. The costs of toxinproduction include (1) the metabolic costs ofproducing the toxin and (2) the negative effect oftoxinon host mobility and the probability ofhost death.Death from cholera results primarily from the dehydration which in turn results trom the loss offluid due to the toxin.

The cholera epidemic that has been unfoldingin South America during the 1990s offers a naturaI experiment with which to assess the generaltheory. The first reported cases occurred in Peruat the beginning of 1991. The interval since thenthus al10ws an assessment of whether any evolutionary effects ofwaterborne transmission can occur over a time interval comparable to the intervalnecessary for other categories of interventions suchas vaccination or hygienic improvements to reducethe frequency of infection.

Within two years of the first reports of cholerafrom Peru the descendants of the Peruvian Vcholerae had spread from this epicenter throughout most countries of South and Central America(Tauxe et al. 1995). This spread set up a temporaland geographic pattern of infection that may allow detailed testing of the proposed evolutionaryassociation between waterborne transmission andtoxigenicity of V cholerae.

We first focused on Brazil because water quality varies throughout Brazil, and the 8razilian Ministry of Health provides summaries of the proportion of the population with access to potable water. Moreover the large size of Brazil offers thepotential for V cholerae to evolve in different directions within the country. The first reported caseof cholera in Brazil was in April 1991, about 2.5months after the first reported case in Peru (Tauxeet al. 1995).

Although this analysis is still in progress, theresults are consistent with an influence of waterquality on virulence. If the mean for each state isused as a separate data point, there is a statisticallysignificant negative association between access topotable water and V cholerae toxigenicity (onetailed p<0.05, Spearrnan rs= -0.62). These dataare, however, preliminary in several respects: (1)additional strains need to be obtained to make theaccuracy of each data point more comparable.Some data points are based on multiple isolatesothers are based on only one isolate; (2) eachstrain was considered to be an independent datapoint in the statistical test; however, the degree towhich the different data points are independent isunknown. Use of molecular phylogenies shouldal10w the generation of tests that use independentpair-wise comparisons (Harvey & Pagel 1991).This kind of comparison should be feasible eventual1y, but will probably need to be unusual1y extensive because nucleotide sequencing and pulsedfield gel e1ectrophoretic studies to date have detected almost no variation among the pandemic eltor strains (Salles & Momen 1991, Karaolis et al.1995); (3) changes in toxigenicity need to betracked to deterrnine whether harrnful strains thatenter areas with relatively pure water evolve reduced virulence over time.

Although the data from Brazil suggest that Vcholerae has evolved toward a lower level ofvirulence, they do not indicate how mild it could eventually become in response to cycling in areas withuncontaminated drinking water. To provide suchan indication, Fig. 2 also plots the rate of toxinproduction ofstrains isolated from Texas and Louisiana in coastal areas of the Gulf of Mexico whereV cholerae has been endemic. Zymodeme analysis indicates that these US strains cluster with theel tor strains of V cholerae (rather than strains ofthe c1assical biotype) but are only distantly relatedto these "mainstream" el tor strains (Salles &Momen 1991). They therefore appear to have beenpresent in the US for decades, perhaps being theremnant of a global outbreak of cholera that occurred many decades ago. Their low toxigenicityprovides an indication ofhow low V cholerae toxigenicity could become in an area with uncontaminated ofdrinking water. Accordingly, although thefrequency of seropositivity to V cholerae in localpopulations in this coastal area of the Gulf ofMexico can be substantial (M.M. Levine, personalcommunication), cases of cholera there are rare.Only about 50 cases have been reported from thisregion from 1965 through 1991 (Weber et al. 1994).

We are currently evaluating whether toxigenicity declines over time in regions with a low potential for waterborne transmission after V cholerae


600

Rondoma

_500 Clox400Ë

~300 .;:;.Tl'1:200Cl)Cl'x~100 -

o ,

ftAlagoas

Amazonas

..Espmto Santo

(8RIo de JaneIro

RIo Grande do Sul.

•Mmas GermsChlle.

LoulSIana xas

o 20 40 60Access of population to potable water (%)

80 100

FIg 2. toxigenicities of el tor i'ihrio cl1o/eme from Brazil, Chile, and the United States. Toxigenicities werc assayed usingstandard ELISA techmques and AKI growth conditIOns. Names of states are given next to the data pOlllt that corresponds to thegeometric mean toxigemclty of the stralll(s) isolated from the state. Numbers next to each data pomt refer to the number ofdifferent strams tested. About 20 separate measurements of toxigenicity were made for each strain. The geometnc mean tmmproduction was calculated for each stram. When more than one strain was obtained for a state, the geometnc mean of the strainmeans was calculated. Although the V. cl1o/eme strams from the United States are only distantly related to those circulating inBraztl, the US strams are presented to provide a sense of how bemgn strams of V. cho/erae might become if they are exposed for along period oftlme to dean water supplies. The Chtlean strains are presented to illustrate how the mean generated from a collection of data associated with a reduction in toxigemcity through time corresponds with the overall geographic trend (see Fig. 2 andtext).

enters such areas from regions with a higher potential for waterbome transmission. Our mostcomplete data set in this regard, although still preliminary, cornes from Chile. Chile is a particularly important country for evaluation of this hypothesis because it has one of the lowest potentials for waterborne transmission among LatinAmerican countries for which el tor V. choleraeinfections have become endemic. There is goodaccess to uncontaminated drinking water and asteep elevational gradient that would limit cyclesof waterbome transmission. V. cholerae enteredChile from Pem at the onset of the pandemie, thefirst case being reported from Chile about tenweeks after the first case was reported in Peru(Tauxe et al. 1995). Water supplies have been morecontaminated in Peru than in Chile since the onsetof the South American epidemic.

We have obtained and tested eight Chileanstrains isolated from c1inical cases over a time spanthat ranged from 1991 (the first year of the SouthAmerican epidemic) to the beginning of 1998.The toxigenicity of the tested Chilean strainsdropped significantly as a function oftime (Fig. 3,one-tailed p<0.02, Spearman rs=-0.81).

The geometric mean toxin production of theChilean strains presented in Fig. 3 is also given inFig. 2 to allow an assessment ofthe degree to whichtheir toxigenicity conforms to that of the Brazilianand North American strains (for ail data in Fig. 2;

Spearman, p<O.OOI, rs=-0.75; N=II; the Louisiana and Texas values were treated as a single datapoint as a conservative measure).

The data from the most recent Chilean isolatesare particularly interesting in light of the toxigenicities of the strains isolated in recent years, whichare nearly as low as those isolated from the Gulfcoast of the US (compare the values for the Chilean isolates of 1998 with those for the Texas andLouisiana). The drop in toxigenicity in Chile corresponds to a very low number of cholera cases.In 1994, for example, when nearly 50,000 caseswere reported in Brazil and nearly 25,000 in Peru,the number of reported cases in Chile dropped toone (Tauxe et al. 1995). These figures coupIed withthe similar differences between the toxigenicitiesof the US Gulfstrains and the most recent Chileanstrains further supports the idea that the evolutionary management of virulence is feasible for V.cholerae if the potential for waterbome transmission can be sufficiently reduced.

At least four explanations exist for the evolution of reduced toxigenicity of V. cholerae in response to reduction in waterbome transmission: (1)the reduction in toxigenicity could result from theincreased costs and decreased benefits of toxinproduction as outlined above; (2) the reductioncould result from a variation on this theme, in whichthe growth of V. cholerae in marine environmentsdisfavors toxigenicity, much in the same way the

Mem /nst Oswa/do Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 1998 571

culturing of parasites outside of hosts causes evolutionary attenuation when genes for virulence nolonger provide a fitness benefit to the organism;(3) the decline in Chile could be interpreted as aresult of the duration oftime that the outbreak hadbeen cycling. Theory suggests that as an outbreakbecomes endemic, pathogens might evolve reducedvirulence (Lenski & May 1994). To evaluate thishypothesis analogous data are needed from "control" countries invaded by V. cholerae at the sametime, but for which water quality has remained low.If the reduction in toxigenicity of V. cholerae inChile is attributable at least in part to its low potential for waterbome transmission, this reductionshould be stronger than that found in such controlcountries. We have not yet obtained such a dataset, but this comparison is feasible because strainsof V. cholerae have been isolated in various countries throughout the pandemic; (4) the decrease inwaterborne transmission might favor decreasedvirulence by reducing the genetic heterogeneity ofthc population of pathogens within a host. Although this hypothesis is probably generally applicable across a broad range ofdisease organisms,it does not appear to be particularly applicable toV. cholerae because its pathogenicity does not involve direct use of host resources, but instead involves the secretion of a product that benefits ailof the other V. cholerae in the intestinal lumen (seethe description of toxin action presented above).

Additional studies are needed to assess thesefour alternatives. It should be noted however, thatfrom the practical perspective ofevolutionary control of disease virulence, the precise mechanism isnot so critical as recognition of the association.

800

That is, whatcver combination of these explanations is correct, virulence of V. cholerae would stillevolve toward lower levels in esponse to investments that reduce waterborne transmission.

The comments about phylogenetically pairedcomparisions mentioned in the context of Fig. 2also apply to the data in Fig. 3, and neither data setcontroJs for several other variables. Seing basedon strains that have been isolated and archived, thecomparisons do not control, for example, for thesource of material. The source of ail or virtuallyail ofthe strains was clinical material, but the sourcewas often not recorded explicitly in the archivedinformation. Nor was the gathering ofstrains regimented so as to eliminate gathering biases. Strainsisolated at the onset of an outbreak might be disproportionately gathered from severe infections,because severe infections would attract the attention of investigators, who would then develop approaches during the outbreak that would generatesamples that were more representative of the existing sample. Although this kind ofsampling biasmight have contributed to the trend presented inFig. 3, particularly with regard to the high valuefor the 1991 isolate, a sampling bias seems inadequate as an explanation of the overall trend, whichresults from the extremely low levels oftoxin production of the strains isolated during the last fewyears. The toxin production ofthese strains is oneto two orders of magnitude below that oftypical eltor strains. Any biases associated with identification of the early cases in the Chilean epidemicshould not have created the uniformly low Jevelsof toxin production that were associated with thestrains during the latter halfofthe epidemic; more-

~700oox 600

~500Clc';400 ..!2g300 -'C

e0.200c'x~ 100

o1990

•

1991

•1992

• •1993 1994 1995 1996

Year of Isolation

1997 1998 1999 2000

FIg. 3: toxigeniclties ofel tor Vihrio cho/erae Isolated from Chile from the beginning of the South American outbreak through thebeginning of 1998. Each data pOint corresponds to a different Isolate. Figures for access to water supplies are from ministries ofhealth statistics for 1996 (Water access percentages from Louisiana and Texas are artificially separated to allow visuahzation ofdata points.). Other detalis are as described in Fig. 1.


over, the statistical test used is sensitive relativerather than the absolute amounts oftoxin production. If the 1991 strain from Chi le had, for example, been only one-fifth of its measured value(and substantially less than a "typical" el tor strain),the statistical significance would have remainedunchanged.

VECTORBORNE TRANSMISSION

Evolutionary theory identi fies vectorbornetransmission as a second factor favoring evolutiontoward relatively high levels of virulence. Ifa disease organism is transmitted by a biting arthropodvector such as a mosquito or reduviid bug, then itcan still be transmitted even if a person is entirelyimmobilized with illness because such bitingarthropods come to feed at immobile people. Infact, experimental studies indicate that mosquitoesare better able to bite a laboratory animal when itis sick with a vectorborne disease such as malariathan when it is healthy, and reduviid bugs (whichare vectors for Chagas' disease) typically feed onsleeping individuals. As a consequence, naturalselection should favor relatively high levels ofhostexploitation by vectorborne pathogens, and weshould therefore see a particularly high virulenceamong vectorborne diseases.

The mortality associated with untreated infections is highly variable among both vectorborneand directly transmitted pathogens, but it is greaterfor vectorborne pathogens than for directly transmitted pathogens (Ewald 1983, 1994). Just as reduction of waterborne transmission should favorevolutionary decreases in virulence, reduction inthe potential for arthropodborne transmission fromimmobilized humans should favor decreases invirulence. This effect can occur through twomechanisms. One mechanism is the direct analogof the argument for waterborne transmission,namely that reduction of transmission from immobilized humans causes a greater reliance on human mobility for transmission. Much as provisioning of uncontaminated drinking water is anintervention that should cause evolutionary reductions in the virulence of diarrheal pathogens, mosquito-proofing of houses is an intervention thatshould cause an evolutionary reduction in the virulence of vectorborne pathogens such as the agentsof malaria or dengue. Ifa person ill with malariaor dengue stays in bed in a vector-proofhouse (orhospital), then the transmission of any pathogensin that person will be blocked during that period.To the extent that those variants tend to be inherently more virulent than variants that allow infectious people to be feeling weil enough to movearound outside oftheir homes, the composition ofthe pathogen population will shift toward a greater

representation of the milder variants. That is, thepathogen population will have evolved towardmildness. This prediction has not yet been tested,but the information available in the literature, bothsupports the key steps in logic and suggests thatthe next stage of large-scale testing is warrantedand would be beneficial even if the hypothesis isincorrect.

First, illness tend to be associated with infectiousness. For vectorborne viral diseases, such asdengue, the evidence is straight-forward: viremiaoccurs during the symptomatic period (e.g., Vaughnet al. 1977). In parasites with more complicatedlife histories, such as plasmodia, the evidence ismore complex because the critical variable is thetiming of infectious life history stages (i.e., the gametocytes) is the critical variable. In this case theevidence still supports the idea that much ofthe transmissibility will be associated with the period ofreduced host mobility (e.g., see Ewald 1994).

Geographic variation indicates that parasiteshave the potential to cause largely miId infectionswhere opportunities for vectorborne transmissionare limited. P vivax strains, for example, tend tobe more mi Id in geographic areas associated withlow and sporadic mosquito transmission (Ewald1994). The variation in P vivax 's distribution appears to be largely a result of differences in theparasite's tendency to generate dormant restingstages (i.e., "hypnozoites").

P falciparum infections are often similarly miIdwhere the potential for vectorborne transmissionis low, for example, in low transmission areas inthe Sudan and Columbia (Elhassan et al 1995,Gonzalez et al. 1997). This tendency also occursin Mali and more generally along the northern edgeof Pfalciparum's range in subsaharan Africa (D.S. Peterson, pers. comm.), where the parasite's distribution may be limited by the restricted abundanceofmosquitoes. The relative importance ofhost andparasite characteristics in determining the mildnessofPfalciparum infections has not been determinedin any ofthese areas, however. If the mildness ofsuch Pfalciparum infections results at least in partfrom the mildness of the P. falciparum variants,evolution toward reduced virulence would seemparticularly feasible. With pre-existing mildstrains, detectable evolutionary shifts toward mildness could occur relatively quickly if mosquitoproofing programs were enacted at the edges of P.falcipanim sdistribution. Ifthese programs provedsuccessful the interventions could progress towardthe center of the ranges, because the mild strainsthat would be needed to replace the more severestrains would already be present in the P.falciparum gene pool. Although such a progression might facilitate a rapid evolutionary shift to-

Mem /nst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), 5ep./Oct. 1998 573

ward benignity, it may not be necessary, as variations in pathogen virulence appear to be presenteven in areas with intense transmission (e.g., Kunet al. 1988).

Influences of exposure to infection on host resistance is a potential confounding variable in anyefforts to control malaria through reduction in frequencies of transmission. One hypothesis attracting recent attention proposes that reductions inentol1lological inoculation rates (EIRs) will havelittle etTect on overall mortality and morbidity inareas with moderate to high bite frequencies, wherethe benefits of reduced EIRs might be otTset byreductions in acquired resistance (Snow & Marsh1995). With regard to evolutionary etTects, thisconcern is applicable primarily to areas with moderate EIR. In areas with low EIR, mosquito-proofing should lower frequencies of infections to thepoint of eradication (Watson 1949). In areas withhigh EIRs, one would expect that mosquito-proofing would cause an evolutionary shi ft toward benignity with relatively little effect on frequency ofinfection, and hence with little effect on benefitsof acquired immunity. If the evolutionary hypothesis is incorrect, great epidemiological benefits canbe expected at least in areas with low EIRs; suchnonevolutionary benefits at higher EIRs are uncertain. If the evolutionary hypothesis is correctthis benefit at low EIRs will be supplemented withreduced virulence of infections across the spectrurnofEIRs.

As is the case with waterborne transmission,vectorproofing of houses can be expected to provide evolutionary reductions in virulence across aspectrum ofvectorbome diseases. DitTerent strainsofdengue, for example, vary in virulence, with themore virulent strains being more productive in cellculture (Morens et al. 1991). Vector-proofing ofhouses against dengue's vector, Aedes aegypti.should similarly favor the milder less exploitativevariants, driving the dengue population to a morebenign state. When more than one vectorbornedisease is occurring in an area, the overall cost effectiveness may increase in proportion to the number of diseases, because the same interventionshould have similar evolutionary effects for each.

The next stage oftesting of these ideas will befeasible only if those who control the sources offunds consider the effort worthwhile. The chancesof such a positive assessment would be improvedif vector-proofing of houses could be shown tohave traditional nonevolutionary epidemiologicalbenefits (i.e., reduction in the frequency of infection) in addition to the hypothesized evolutionaryepidemiological benefits (i.e., reduction in theharmfulness of the causative organisms). Theavailable evidence indicates that traditional ben-

efits do occur. The etTectiveness ofmosquito-proofhousing against transmission of dengue, for example, is suggested by the resistance to invasionwhen such housing is generally present. Over thepast two decades thousands of cases of dengue fever have occurred on the Mexican side of the US;Mexico border along the Gulfof Mexico. Denguehas been introduced repeatedly into Texas therebut has failed to spread in spite of the ubiquitouspresence ofAedes vectors. For every reported caseacquired on the Texas side of the border there areabout 1000 reported cases on the Mexican side(CDC 1996). The pervasiveness ofmosquito-proofon the Texas side appears to be responsible for thisditTerence. Similarly, malaria has been introducedon numerous occasions in recent years to areas inthe U.S. where it had previously been endemic.Appropriate vectors are abundant, yet little secondary transmission occurs; when it does, it has beenself-limited and localized (Wyler 1993, Dawsonet al. 1997; for an analogous example involvingsevere diarrheal disease, see Weissman 1974 ).

The most thorough experimental test of the effectiveness of mosquito-proof housing on malariatransmission was conducted from )939 through the1940s in a large section of northern Alabama, bythe Tennessee Valley Authority (TVA), which wasoverseeing the construction dams in the area(Watson 1949). The TVA was concemed aboutmalaria because the construction of dams in theregion had previously contributed to the malariaproblem there (Ackerman 1956, Derryberry 1956).

During the 1930s about half of the people inthe area tested positive. In 1939, the TVA began acampaign to mosquito-proof ail houses in the areaand accomplished this goal within seven years.They divided the area into II zones and completedthe mosquito-proofing of each zone at differenttimes. The results of their study show that mosquito-proofing virtually eradicated malaria from thearea, with the decline occurring earlier in thosezones in which mosquito-proofing was completedearlier (Fig. 4). No other intervention was enactedprior to the decline (Watson 1949).

These results do not represent a test of the ideathat malaria pathogens evolve to lower levels ofvirulence in response to mosquito-proofing ofhouses. The results do, however, demonstrate several important points.

First, the results show that Plasmodium populations are influenced by mosquito-proofing. Ifthe population as a whole declines so strongly inresponse to screening, it seems probable that certain variants within the population will be moresubstantially reduced by screening than others,leading to an evolutionary change in the Plasmodium gene pool.


,

- ------------:

Fig. 4: seroposltlvlty of blood samples for Plasmodium presented as a function of year dUTIng the mosquito-prooting programcalTied out in Alabama by the Tennessee Valley Authority. Each row corresponds to one of the Il geographlc zones that comprisedthe study. The asterisk designates the year in which mosquito-prooting was completed for ail houses in the zone. See text for otherdetails (data from Watson 1949).

Second, the results demonstrate nonevolutionary benefits necessary to justify the large-scaleevolutionary experiment that would be needed toassess virulence management through mosquitoproofing. To justify the experiments from bothethical and economic perspectives, new areas forexperimentation could be selected on the basis ofhaving a slightly more difficult control problemthan those for which nonevolutionary success hasbeen demonstrated (e.g., a slightly higher prevalence of infection than occurred in northern Alabama just prior to the mosquito-proofing).

Third, the results show that the experiment isfeasible logistically and financially even with thelimitations of 1940s technology. The costs ofmosquito-proofing (in 1944 dollars) was about $100per house for the area with the poorest quality ofhousing; the costs of maintaining the mosquitoproofing was about $12 per house per year (Watson1949). Modern technology has generated materiaIs that are more effective, more durable, easier toapply and maintain, and more pleasant to live withthan those used in the TVA study. Costs shouldtherefore not be as greatly increased as wouId beindicated by a simple adjustment of the TVA costsfor inflation. The actual costs may be influencedup or down depending on the details of a particular area such as the quality of existing houses, thedegree to which materials couId be generated 10cally and the costs of local labor.

Finally, the results of the TVA study demonstrate that mosquito-proofing worked even thoughthis geographic area can be stiflingly hot and humid during the malaria season. Skeptics couId haveargued that people would not stay inside ofhouses

sufficiently under such conditions for the antimalarial effects of mosquito-proof housing to work.Or, skeptics could have argued that people woulddeliberately destroy screens to increase air-flowthrough houses, but such vandalism was rare inthe Alabama study (Watson 1949).

These ideas should be generally applicableacross the spectrum of vectorborne diseases, although the particular details of the application willdepend on the details of the vectorborne disease.Chagas disease offers an informative illustrationofone variation on the theme. The agent ofChagasdisease, Tr)panosoma cruzi, is transmitted by reduviid bugs that bite sleeping individuals. It is therefore transmitted largely while people are immobilized in their houses. The frequencies of infectionshould therefore be reduced by vector-proofing ofhouses.

The details of T end transmission indicatethat this intervention could reduce the virulence ofT enlZÎ through two evolutionary processes. Thefirst process is analogous to that proposed abovefor malaria and dengue. To the extent that transmission does sometimes occur from mobile hostsoutside ofhouses, virulence could be reduced.

The second evolutionary process concerns theeffects ofalternative vertebrate hosts on virulencein humans. The extent ofhuman-bug-human transmission varies substantially geographically; substantial human-bug-human transmission occurringthroughout most of T erllzi ·s range but is virtuallyifnot entirely absent in the US. Theory and comparative data indicate that vectorborne pathogensshould tend to be relatively mi Id in humans whenthey rarely cycle in humans (Ewald 1983). About

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep'/Oct. 1998 575

5-25% ofnonhuman vertebrate hosts (racoons andoppossums) in the southern US are infected withT. cru:::i (Burkholder et al. 1980, Karsten et al.1992, Pung et al. 1995), and a comparison ofsuchstrains with strains from humans in Brazil that theyare genetically distinct (Clark & Pung 1994). Inthe US humans rarely acquire T. cru::i via vectors,and appear to be dead-end hosts, probably becauseof the vector proof housing and low vector densities (Burkholder et al. 1980, Kirchhoff 1993,Barrett et al. 1997). In accordance with theoryabout the evolution of virulence, such infectionsappear ta be particularly mild in humans, so muchso that only three cases of acute Chagas' diseasefrom bug bites had been reported in the US as of1993 (Woody et al. 1961 a, b, Kirchhoff 1993).

This situation is of importance to evolutionarycontrol of T. cru:i in countries with endemic Chagasdisease because by making houses vector-proof,the importance ofhuman-bug-human cycling relative to enzootie cycling should become greatly reduced, thus causing the evolution ofincreased specialization of T. cruzi on nonhuman vertebrates,reduced specialization on humans, and consequently, reduced virulence in humans. As in thecase ofmalaria, the presence ofbenign strains couIdbe beneficial through protection against severestrains like a free live vaccine, because benignstrains of T. cruzi can protect against highly virulent clones (Lauria Pires & Teixeira 1997).

THE MERGING OF EPIDEMIOLOGY WITHEVOLUTIONARY BIOLOGY

The ideas presented above illustrate how theevolutionary considerations of virulence is bringing the health sciences is adding a new dimensionto the ideas of the early epidemiologists. AfterEvandro Chagas deciphered the mode of transmission of T. cruzi, he stressed the importance ofbreaking the domestic cycle of transmissionthrough the vector-proofing of houses. Decadeslater, in the 1940s, the architects of the mosquitoproofing campaign in Alabama stressed the samepoint for the control of malaria and demonstratedits utility (Watson 1949). But just as the results ofthe Alabama study were becoming available, DDTwas introduced and successfully used to controlmalaria in the Mediterranean and South Asia(Harrison 1978). Also at that time the powerfulquinine derivatives against malaria were being discovered in response to the eut-off of natural quinine to the Allied powers during World War II. Theevidence of the epidemiological value of vectorproofing houses as a control measure against malaria was set aside and largely forgotten in favor ofthese lwo more attractive options. A half-century

of expcrience has demonstrated how these two alternatives are incapable of the sort of global eradication that was envisaged at mid-century and haveleft researchers narrowing their hopes on vaccinesas their remaining option for eradication. Butbroadly effective vaccines have proved elusive;moreover, the evolutionary versatility of plasmodia casts doubt on the long-term success of vaccination - the generation of effective vaccines maybe a less fOlmidable challenge than maintainingthe cfficacy ofvaccines after they are put into use.The present therefore seems an opportune time toinvestigate the possibility of using the evolutionary versatility of plasmodia to our advantage, togenerate milder variants.

Like Chagas, by quantifying the frequencies ofcholera in areas of London one-and-one-half centuries ago, John Snow demonstrated that the frequencies ofcholera were associated with contamination of water supplies (Snow 1855). By integrating evolutionary insights with this kind of cpidemiological insight we can add a second dimension to studies of cholera, namely that the harmfulness of pathogens (and hence the harmfulnessper infection) is also associated with contamination of water supplies.

Evolutionary considerations strengthen arguments for improving housing and water quality intwo ways. First, evolutionary considerations reveal weak spots in programs based on insecticidesand antibiotics: the target organisms evolve resistance. Second, evolutionary considerations suggest a previously unrecognized evolutionary benefit of such improvements: the target pathogensshould evolve reduced virulence. To evaluate theval idity ofsuch evolutionary benefits the suggestedinterventions (making water supplies pure andhouses vector-proof) need to be enacted and studied prospectively in human populations. The molecular and genetic tools are already available orcould be readily developed for target pathogensin each category. Molecular detenninants ofvirulenee are needed to determine whether evolutionary changes in virulence occur. Molecular phylogenies are needed to categorize pathogens according to their epidemiological history and to structure statistical tests.

Our current state of knowledge already seemssufficient to justify such investments in these interventions ethically and economically. The experimental tracking ofepidemiological changes inthe frequencies of virulent and mild genotypesshould provide conclusive answers to these evolutionary questions while simultaneously providingthe epidemiological benefits envisioned by Snowand Chagas.


REFERENCES

Ackerman EA 1956. TVA in its larger setting, p. 244256. In RC Martin. TVA. The First Twenty Years. AStafIReport, University of Alabama Press & University of Tennessee Press.

Barrett VJ. Leiby DA, Odom JL, Otani MM, Rowe JD.Roote JT, Cox KF, Brown KR, Hoiles JA. SaezAlquezar A, Turrens JF 1997. Negligible prevalenceof antibodies against lIypanosoma end amongblood donors in the southeastern United States. AmJ Clin Patholl08: 499-503.

Burkholder JE, Allison TC, Kelly VP 1980. lI}panosomacru::i (Chagas) (Protozoa: Kinetoplastida) in invertebrate, reservoir, and human hosts of the lower RioGrande Valley of Texas. J Parasitol66: 305-311.

CDC - Centers for Disease Control 1996. Dengue feverat the U.S.-Mexico border, 1995-1996. Morb MortWeeklr Report 45: 841-844.

Clark CG', Pung OJ 1994. Host specificity of ribosomalDNA variation in sylvatic lIypanosoma end fromNorth America. Mol Biochem Parasitol66: 175-179.

Dawson M, Johnson PT, Feldman L, Glover R, KoehlerJ, Blake P, Toomey KE 1997. Probable locally acquired mosquito-transmitted Plasmodium vivax infection - Georgia, 1996 (Reprinted from MMWR,vol 46, p. 264-267, 1997). JAMA 277: 1191-1193.

Derryberry OM 1956. Health, p. 193-205. In RC Martin, TVA. The First Twenty Years. A Staff Report,University of Alabama Press & University of Tennessee Press.

Elhassan lM, Hviid L, Jakobsen PH, Giha H, Satti GMH,Arnot DE, Jensen 18, Theander TG 1995. High proportion ofsubclinical Plasmodiumfàlciparum infections in an area of seasonal and unstable malaria inSudan. Am J lI'Op Med Hyg 53: 78-83.

Ewald PW 1983. Host-parasite relations, vectors, andthe evolution ofdisease severity. Ann Rev Ecol Syst14: 465-85.

Ewald PW 1991. Waterborne transmission and the evolution of virulence among gastrointestinal bacteria.Epidemiollnfèct 106: 83-119.

Ewald PW 1994. Evolution ofInfèctious Disease, Oxford University Press, New York.

Gonzalez JM, OIano V, Vergara J, Arevalo Herrera M,Carrasquilla G, Herrera S, Lopez JA 1997. Unstable,low-Ievel transmission ofmalaria on the ColombianPacific Coast. Ann lI'Op Med Parasitol91: 349-358.

Harrison G 1978. Mosquitoes. Malaria & Man: A HistOIJ' of the Hostilities Since 1880, EP Dutton, NewYork.

Harvey PH, Pagel MD 1991. The Comparatil'e Methodin Evolutionary Biology, Oxford University Press,Oxford.

Karaolis DKR, Lan R, Reeves PR 1995. The sixth andseventh cholera pandemics are due to independentclones separately derived from environmental.nontoxigenic, non-O 1 Vibrio cholerae. J Baeteriol177: 3191-3198.

Karsten V, Davis C, Kuhn R 1992. lIJpanosoma el1lziin wild raccoons and opossums in North-Carolina. JParasitol 78: 547-549.

Kirchhoff LV 1993. Current concepts - American trypa-

nosomiasis (Chagas' disease) - a tropical disease nowin thc United States. New Engl J Med 329: 639-644.

Kun JFJ, SchmidtOtt RJ, Lehman LG, Lell B, LucknerD, Greve B, Matousek P, Kremsner PG 1998. Merozoite surface antigen 1 and 2 genotypes and resetting of Plasmodium fàlciparum in severe and mildmalaria in Lambarene, Gabon. lI'ans R Soc lI'Op MedHyg 92: 110-114.

Lauria Pires L, Teixeira ARL 1997. Protective etTect ofexposure to non-virulent lIypanosoma end cloneson the course of subsequent infections with highlyvirulent clones in mice. J Comp Pathol 117: 119-126.

Lenski RI, May RM 1994. The evolution ofvirulence inparasites and pathogens: reconciliation bctwcen twocompeting hypotheses. J Theor Biol 169: 253-265.

Morens DM, Marchette NJ, Chu MC, Halstead SB 1991.Growth of dengue type-2 virus isolates in humanperipherai blood leukocytes correlates with severeand mild dengue disease. Am J lI'op Med Hyg 45:644-51.

Pung OJ, Banks CW, Jones DN, Krissinger MW 1995.lIypanosoma cru=i in wild raccoons, opossums, andtriatomine bugs in southeast Georgia, U.S.A. JParasitol 81: 324-326.

Salles CA, Momen H 1991. Identification of Vibriocholerae by enzyme electrophoresis. Trans R SoclI'op Med Hyg 85: 544-547.

Snow J 1855 (1996 reprint). On the Mode ofCommunication ofCholera, 2nd ed., London, Churchill.

Snow RW, Marsh K 1995. Will reducing Plasmodiumfalciparum transmission alter malaria mortalityamong African children? Parasitol Today 1/: 188190.

Tauxe RV, Mintz ED, Quick RE 1995. Epidemic choiera in the new world: translating field epidemiologyin now prevention strategies. Emerg Infect Dis 1:141-146.

Vaughn DW, Green S. Kalayanarooj S, Innis BL,Nimmannitya S, Suntayakorn S, Rothman AL, EnnisFA, Nisalak A 1997. Dengue in the early febrilephase: viremia and antibody responses. J fJ!fèct Dis176: 322-330.

Watson RB 1949. Location and mosquito-proofing ofdwel1ings, p. 1184-1202. In MF Boyd, Malariology.A Comprehensive Survey o.f'AlIAspects ofThis Groupof Diseases from a Global Standpoint. Saunders,Philadelphia.

Weber JT, Levine WC, Hopkins DP, Tauxe RV 1994.Cholera in the United States, 1965-1991. Risks athome and abroad. Arch Intern Med 154: 551-556.

Weissman 18, Murton KI, Lewis JN, Friedemann CHT,Gangarosa EJ 1974. Impact in the U.S. of the Shigadysentery pandemic ofCentral America and Mexico:A review of surveilIance data through 1972. J InfèctDis 129: 218-23.

Woody NC, Woody HB 1961 a. American trypanosomiasis 1. Clinical and epidemiological background ofChagas' disease in the U.S. J Pediat 58: 568-580.

Woody NC, DeDranous N, Woody HB 1961b. American trypanosomiasis II. Current serologic studies inChagas' disease. J Pediat 58: 738-745.

Wyler DJ 1993. Malaria: overview and update. Clin Infect Dis 16: 449-458.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 577·580, Sep./Oct. 1998 577

Integrated Genetic Epidemiology of Infectious Diseases:The Chagas Model

Michel Tibayrenc

Centre dEtudcs sur le Polyrnorphismc des Microorganisrncs (CEPM), UMR CNRS/ORSTOM 9926. ORSTOM,BP 5045. 34032 Monptcllier Cedcx 0 I, France

Genetic typing a/pathogenic agents and ofvectors has known impressive developments in the last 10years. thanks to the progresses ofmolecular biology. and to the contribution of the concepts ofevolutional)' genetics. Moreover; we know more and more on the genetic susceptibility 0/man to infectiousdiseases. I propose here to settle a new, synthetic field of research, which 1 call 'integrated geneticepidemiology ofinfectious diseases' (IGEID). 1aim at evaluating, by an evolutionary genetic approach.the respective impact, on the transmission and pathogenicity of infectious diseases. of the host 5. thepathogen 5 and the vector 5genetic diversity. and their possible interactions (eo-evolution phenomena).Chagas' disease constitutes afine model to develop the IGEID methodology, by both field and experimental studies.

Key words: eo-evolution - genetic typing - evolutionary genetics - Trypanosoma cruzi

Genetic studies dealing with infectious agents,vectors and hosts (for example: genetic susceptibility of man to infectious diseases) have developed until now separately, in a compartmentalizedmanner. Nevertheless, in an evolutionary point ofview, the three actors of infectious disease transmission (the pathogen, the host, and in the case ofvector-borne diseases, the vector) have evolved together, and should be considered as the three linkedcomponents of a unique phenomenon of co-evolution. When the host evolves (for example, develops specific immune defenses to escape fromthe damage caused by the pathogen), it shapes inreturn the evolution and the genetic diversity ofthe pathogen. It is therefore distressing to analyzeseparately these three components. I have proposed(Tibayrenc I998a, b) to settle a new, synthetic fieldof research, the 'integrated genetic epidemiologyof infectious diseases' (IGEID), that wiIl take intoaccount simultaneously the impact, on the transmission and pathogenicity of infectious diseases,of the host's, the pathogen's and the vector's genetic diversity, as well as the interactions (phenomena of eo-evolution) of these three parameters. IwiIl advocate here that Chagas' disease constitutesa fine model for throwing the first bases of thisambitious approach.

Fax: +33-4-6741.6299.E-mail: [email protected] 15 June 1998Accepted 30 July 1998

WHAT ABOUT THE GENETIC DIVERSITY OFTRYPANOSOMA CRUZJ?

1f we consider the putative impact of the host's,the pathogen's and the vectors' genetic diversityon the transmission and pathogenicity of Chagas'disease, there is little doubt that the best knownelement is T. cruzi genetic variability. Many studies have been published on this theme, and it ispossible that T. cruzi is one ofthe pathogenic agentswhich evolutionary genetics is the best explored.Main results can be briefly summarized as foIlows:T. cruzi natural populations show considerablegenetic polymorphism, as revealed by isoenzymeelectrophoresis (Miles et al. 1978), kDNA RFLPanalysis (Morel et al. 1980) and RAPD (Tibayrencet al. 1993). The most parsimonious hypothesis toaccount for this huge genetic polymorphism is thatit is the result of long-term clonal evolution withpossible occasional bouts of genetic exchange(Tibayrenc et a1. 1986, Tibayrenc & Ayala 1988).Recently. these suspected recombinant genotypeshave been more precisely characterized as stablehybrid lines, that would propagate clonally afterthe hybridization event (Bogliolo et al. 1996,Carrasco et al. 1996, Brisse et al. 1998). Amongthe natural clones of T. cruzi, some are widespreadand more frequently sampled. They have beengiven the name of 'major clones' (Tibayrenc &Ayala 1988), since it can be suspected that theirepidemiological and pathogenic relevance is considerable. It is most probable that these 'clonalgenotypes' identified by a limited set of geneticmarkers do not correspond to real clones, but rather,tho families ofclosely related clones. We have pro-

578 Integrated Genetic Epidemiology of Chagas • Michel Tibayrenc

posed (Tibayrenc & Ayala 1991) the term of'clonet' to refer to sets of stocks that appear identical for a given set of genetic markers in a clonaispecies. T. eruzi clonets are distributed into twomain phylogenetic lineages within each of whichgenetic diversity remains considerable (Tibayrenc1995, Souto et al. 1996). The second main phylogenetic lineage of T. eruzi appears as structuredinto five lesser subdivisions (Brisse et al. 1998),of which some correspond to either hybrid lines orto formerly identified 'major clones' (Tibayrenc& Ayala 1988) or both. By comparison with otherpathogens, T. eruzi population structure can bedefined as follows: it is a clonai species (Tibayrencet al. 1986) that is structured into durable geneticsubdivisions ('discrete typing units' or DTUs;Tibayrenc 1998a, b). The whole species T. eruzi isa DTU, as weil as its main and lesser genetic subdivisions. Ali these DTUs can be characterized byspecific genetic markers or 'tags' (Tibayrenc1998a, b). To some extent, T. eruzi DTUs and tagscan be equated respectively to monophyletic lineages (clades) and synapomorphic characters, although a strict cladistic approach is difficult here,due to the existence of occasional hybridizationevents. Still the fact remains that T. eruzi overallintraspecific phylogeny appears as robust, considering the strong agreement between the speciesphylogenies generated by independent sets of genetic markers: isoenzymes and RAPDs (Tibayrencet al. 1993), and microsatellites (Macedo & Pena,pers. comm.). This striking concordance betweenthree different kinds ofgenetic markers is clear evidence that the strong genetic distances recordedwithin T. cruzi are due to a real evolutionary divergence rather than to individual genetic diversity within a hypothetical, recent ancestral sexualspecies, as formerly envisaged (Tibayrenc et al.1984). It is reasonable to expect that the evolutionary divergence accumulated between T. eruziclonallineages involves also those genes that govem relevant medical properties such as virulenceor resistance to drugs. A possible link between T.eruzi genetic variability and Chagas'disease clinical diversity has been suspected by Miles et al.( 1981). Montanat et al. (1996) have recently corroborated this hypothesis. Long-term experimentsperformed in our laboratory show a clear correlation between evolutionary divergence among T.eruzi clonallineages and amount ofdifferences forrelevant biological properties such as pathogenicity in mice, in vitro drug sensitivity or culturegrowth speed (Laurent et al. 1997, Pinto et al. 1998,Revollo et al. 1998, De Lana et al. 1998). Certainexperiments suggest an interaction between clonaigenotypes in artificial mixtures (De Lana et al.

1998). Macedo and Pena (1998) have recently proposed a 'clonal-hystotropic model', which statesthat T. cruzi clonai genotypes infecting the samehost have each a specific tropism for given organs.These proposaIs as weil as our results dealing withinteractions of clonai genotypes lead to considerthat the idea: 'one strain, one pathology' is possibly too simplistic. Still the fact remains that convergent lines ofresults suggest a profound impactof the phylogenetic diversity of T. enlZÎ naturalclones on their relevant biomedical properties.

For studies dealing with the integrated geneticepidemiology, T. eruzi eonstitutes an ideal mode!,for it is clearly subdivided into clear-cut discreteentities: upper and lesser DTUs, and at a lower levelofphylogenetic divergence, the natural clones. TheRAPD technique is an abundant source of markers for designing probes and PCR diagnoses specific of either DTUs or natural clones. These specific molecular tools can be conveniently used inthe context of integrated genetic epidemiology ofChagas' disease.

THE VECTOR

Although less known than the parasite's geneticdiversity, triatomine bugs have been the materialfor various evolutionary genetic analyses. Thesestudies were based mainly on multilocus enzymeelectrophoresis, and have focused either on theintraspecific level (population genetics analysis;Tibayrenc et al. 1981a, b, Dujardin & Tibayrenc1985, Dujardin et al. 1998) or on between-speciescomparisons (phylogenetic analysis; Pereira et al.1996, Solano et al. 1996). These data provide afine starting basis to include the study of the vector in the integrated genetic epidemiology ofChagas'disease. Nevertheless, it will be necessaryto complement isoenzyme typing with more modern molecular tools such as RAPDs ormicrosatellites, in order to increase the resolutionpower oftriatomine bug genetic characterization.

THE HOST

From the genetic point ofview, ofthe three linksof Chagas transmission chain, man is the lessknown. As a matter of fact, contrary to other parasitic diseases such as malaria or schistosomiasis(Abel & Dessein 1997), nothing is known aboutpossible human genetic susceptibility toChagas'disease and its different clinical fonns.Now the genetic variability of the human specieshas been widely explored (HLA and microsatellitetyping, gene mapping), which should make easierto explore the parameter of host genetic susceptibility in the specific case ofChagas' disease.


INTEGRATED GENETIC EPIDEMIOLOGY OFCHAGAS' DISEASE: EXPERIMENTAL APPROACH

Chagas' disease constitutes a very fine modelfor experimental studies, since it is possible to establish a complete artificial cycle in the laboratory.The parasite is relatively easy to culture, underepimastigote, trypomastigote and amastigoteforms. Rearing the vector is easy too, includingthrough artificial feeding devices, which makes iteasier to monitor the experimental parameters(Pinto et al. 1998). Lastly, many mammiferousmodels (mainly mice) can be used as vertebratehosts. The principle of an experimental approachofintegrated genetic epidemiology is to have onlyone parameter vary at the same time, while the twoother ones are kept as constant as possible. Forexample, if the impact on Chagas' disease of T.end is to be explored (either with pure clonai genotypes or artificial mixtures of genotypes), homogenous triatomine bug and mice strains will be used.When the influence of the vector is explored (bothat the level of subspecific and interspecific variability), this will be done, in a given experiment,with only one T. erllzi clonai genotype and with aunique mouse strain. Lastly, when the host is considered, various populations of a given strain andvarious strains (males and females) will be usedwith the same T. eruzi clonai genotype and the sametriatomine bug strain. Apart from the empiricalobservation of the respective impact of the host's,the vector's and the pathogen 's genetic diversityon Chagas transmission and pathogenicity, it willbe possible to identify the genes that are impliedin the infectious process, and to analyze gene regulation phenomena through the analysis ofmRNAswith the RNA AP-PCR technique (Welsh et al.1992). For example, it will be possible to analysegene expression ofgiven T. erllzi clonai genotypes(amastigote, epimastigote and trypomastigoteforms) before and after passage through given vector and host populations, or before and after infection ofcell cultures, or to compare infected vs noninfected cardiac or digestive cells of dissectedmice. Again in these RNA AP-PCR analyses, onlyone parameter will be allowed to vary at a giventime, while the other ones are kept as constant aspossible.

FIELD STUDIES

The experimental step is the easiest one to master, and is indispensable. Nevertheless, it definitelyhas to be completed with a more ambitious approach, which is field studies. This involves thejoint analysis of man, triatomine bug and parasitepopulations. When man is considered, the nowwell-codified screening with microsatellite mark-

ers will have to be used. This makes it possible,through the study offamilies and control, Chagasfree, populations, to look for possible associationsbetween given parts of the human genome andsusceptibility to Chagas' disease and its variousclinical fonns, through a statistical analysis oflinkage disequilibrium. In the same time, isolation ofT. crll=ï stocks from the same populations of patients gives the opportunity to explore possibleassociations between T. end clonaI genotypes andclinical forms ofChagas' disease. Lastly, the jointanalysis of the genetic variability oftriatomine bugpopulations and ofthe T. erllzi stocks isolated fromthem could make it possible to increase by far thelevel ofresolution ofgenetic epidemiological tracking. As a matter of fact, genetic evolution of thevector and of the parasite do not have the samepatterns and the same speed, although they arelinked. They can give therefore non-redundant,complementary indications on the spread ofChagas'disease epidemics.

CONCLUSION; PERSPECTIVES

IGElD is a very ambitious endeavor, that willneed the joint efforts of many different teams having complementary expertises. These variouscompetences are difficult to find in only one country. For these reasons, it is a typical field of research that should be launched in the recently-proposed project of 'European Centre for Control ofInfectious Diseases' (ECCID; Tibayrenc 1997a, b).The Chagas model gives the opportunity to launchthe first bases of this approach, with the advantages ofeasy-to-master experimental protocols, andaboundant amount of knowldege on the geneticdiversity of the pathogen, and, to a lesser extent,of the vector. The general methodologies developed for Chagas'disease will be applicable to alarge extent to other infectious models, especiallyto the ones that involve related parasites (Leishmania and African trypanosomes). The comparative approach advocated for in the case of evolutionary genetics of pathogens (Tibayrenc 1995,1996) should be retained for IGEID. lndeed, onlya comparative IGEID approach will permit to drawthe generallaws that govern pathogen/host/vectorcoevolution, and in the same time, to enlighten thespecificities of each model.

REFERENCES

Abel L, Dessein Al 1997. The impact ofhost geneticson susceptibility to human infectious diseases. Cl/,.,.Op 1mmullo1 9: 509-516.

Bogliolo AR, Lauriapires L, Gibson WC 1996. Polymorphisms in TTypanosvma cru::i: evidence of genetic recombination. Acta TT'op 61: 31-40.

Brisse S, Barnabé C, Tibayrcnc M 1998. TTJ'Panosoma

580 Integrated Genetlc Epidemiology of Chagas • Michel Tibayrenc

cmzi: how many relevant phylogenctic subdivisionsarc thcre? Parasitol Today 14: 178-179.

Carrasco HJ, Frame lA, Valente AS, Miles MA 1996.Genetic exchange as a possible source of genomicdiversity in sylvatic populations of Trypanosomacruzi. Am J Trop Med Hvg 54: 418-424.

De Lana M. Pinto A da S, Barnabé C, Quesney V, NoélS, Tibayrenc M 1998. n:l'Panosoma cnô: comparedvectorial transmissibility of3 major clonal genotypesby n-iatoma in{estans. ExpParasitoI, in press.

DUJardinJP, TibayrencM 1985. Etude de II enzymes etdonnées de génétique formelle pour 19 lociisoenzymatiques chez n'iatoma in{estans (Hemiptera: Rediviidae). AIlIl Soc Belge Méd n·op 65: 271280.

Dujardm JP, Schofield CJ. Tibayrenc M 1998. Population structure of Andean Triatoma il~/estans:

allozymc frequencles and their epidemiological relevance. Med Vet Entomoi 12: 20-29.

Laurent JP. Barnabé C. Quesney V, Noél S, TibayrencM 1997. Impact of clonai evolutlOn on the biological diversity of n-ypanosoma crllzi. Parasitology114: 213-218.

Macedo AM, Pena SDJ 1998. Genetic variability of nypanosoma end: implicatIOns for the pathogenesisof Chagas' disease. Parasitai Today 14: 119-124.

Miles MA, Povoa M, Prata A, Cedillos RA, De SouzaAA. Macedo V 1981. Do radically dissimilar n-ypallosoma eru::i strains (zymodemes) cause Venezuelan and Brazilian forms ofChagas' disease? Lancet 8234: 1336-1340.

Miles MA. Souza A, Povoa M, Shaw JJ, Lainson R,Toyé P.I 1978. Isozymic heterogeneity of Trypanosoma cru::i in the tirst autochtonous patients withChagas' disease in Amazonian Brazil. Natllre 272:819-821.

Montanat EE, De Luca GM, Gallerano RH, Sosa R,Blanco A 1996. Characterization of nypanosomaeruzi populations by zymodemes: correlation withclinical pictures. Am J n·op Med H)'g 55: 625-628.

Morel CM. Chiari E, Plessmann Camargo E, Mattei DM,Romanha AJ, Simpson L 1980. Strains and clonesof n)'Panosoma cl'llzi can be characterized by pattern of restriction endonuclease products of kinetopIast DNA minicircles. Proc Nati Acad Sei USA 77:6810-6814.

Pereira J. Dujardin JP, Salvatella R, Tibayrenc M 1996.Enzymatic variability and phylogenetic relatednessamong Triatoma in{estans, T. pIatcl/sis. T. delponteiand T. ruhrovana. Heredity 77: 47-54.

Pinto A da S, de Lana M, Bastrenta B, Barnabé C,Quesney V, Noël S, Tibayrenc M 1998. Comparedvectorial transmissibility of pure and mixed clonaigenotypes of TI:vpanosoma crZl::i in Triatomain{estans. Parasitol Res 84: 348-353.

Revollo S. Oury B, Laurent JP, Barnabé C. Quesney V,Carrière V, Noël S. Tibayrenc M 1998. Try'Panosoma

crZlzi: impact of clonai evolution of the parasite onits biological and medical properties. Exp Parasitoi89: 30-39.

Solano P, Dujardin JP, Schofield CJ, Romanan C,Tibayrenc M 1996. Isoenzymes as a tool forRhodnills species identification. Res Rel' Parasitol56: 41-47.

Souto RP, Fernandes 0, Macedo AM, Campbell DA,Zingales B 1996. DNA markers define two majorphylogenetic lineages of n:vpallosoma crllzi. MolBiochem Parasitol83: 141-152.

Tibayrenc M 1995. population genetics ofparasitic protozoa and other microorganisms. Adv Parasitol 36:47-115.

Tibayrenc M 1997a. European Centres for Disease Control. Natllre (correspondence) 389: 433-434.

Tibayrenc M 1997b. Microbes sans frontières and theEuropean CDe. Parasitoi Today 13: 454.

Tibayrenc M 1998a. Genetic epidemiology of parasiticprotozoa and other infectious agents: the need foran integrated approach. ln! J Parasitol 28: 85-104.

Tibayrenc M 1998b. Beyond strain typing and molecular epidemiology: integrated genetic epidemiologyofinfectious diseases. Parasitoi Today 14: 323-329.

Tibayrenc M, Ayala FJ 1988. Isozyme variability of Try'panosoma cruzi, the agent of Chagas' disease: genetical, taxonomical and epidemiological significanee. Evolution 42: 277-292.

Tibayrenc M, Ayala FJ 1991. Towards a population genetics ofmicroorganisms: the clona] theory of parasi tic protozoa. Parasitoi Toda)' 7: 228-232.

Tibayrenc M, Echalar L, Carlier y 1981 a. ComparaisonIsoenzymatique de deux populations boliviennes(altitude et plaine) de n'iatoma iI!festans (HemipteraReduviidae). Cah ORSTOM sér Ent méd Parasitol19: 125-127.

Tibayrenc M, Echalar L, Carlier y 1981 b. Données degénétique formelle pour six loci isoenzymatiqueschez Triatoma infestans (Hemiptera, Reduviidae).Cah ORSTOM sér Ent méd Parasitoi 19: 121-123.

Tibayrenc M, Neubauer K. Barnabé C, Guerrini F,Sarkeski D, Ayala FJ 1993. Genetic characterization of six parasitic protozoa: parity of randomprimer DNA typing and multilocus isoenzyme electrophoresis. Proe Natl Acad Sei USA 90: 1335-1339.

Tibayrenc M, Solignac M, Cariou ML, Le Ray D.Desjeux P 1984. Les souches isoenzymatiques den)'panosoma cnd: origine récente ou ancienne,homogène ou hétérogène? CR Acad Sei Paris 299:195-198.

Tibayrenc M, Ward P, Moya A, Ayala FJ 1986. Naturalpopulations of Trypanosoma cru::i, the agent ofChagas' disease, have a complex multiclonal structure. Pme Nat Acad Sei USA 83: 115-119.

Welsh J, Chada K, Dalal SS, Cheng R, Ralph D,Mcclelland M 1992. Arbitrarily Primed PCR Fingerprinting of RNA. NlIcl Ac Res 20: 4965-4970.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 581-585, Sep'/Oct. 1998 581

Molecular Epidemiologie Typing Systems of 8acterialPathogens: Current Issues and Perpectives

Marc JStruelens

Service de Microbiologie, Hôpital Erasme and Unité d'Epidémiologie des Maladies Infectieuses, Ecole de SantéPublique, Université Libre de Bruxelles 808, Route de Lennik 1070 Bruxelles, Belgium

The epidemiologic typing of bacterial pathogens can be applied to answer a number of dilJerentquestions: in case ofoutbreak, what is the extent and mode of transmission ofepidemic clone(s }? Incase oflong-term surveillance, what is the prevalence over time and the geographic spread ofepidemicand endemic clones in the population? A number oimolecular typing methods can be used to classifj;bacteria based on genomic diversity into groups ofclosely-related isolates (presumed to arise [l'Dm acommon ancestor in the same chain of transmission) and divergent, epidemiologically-unrelated isolates (arising/inm independent sources ofinfection). Ribotyping. IS-RFLPfingerprinting. macrorestrictionanalysis ofchromosomal DNA and PCR:fingelprinting using arbitra!)! sequence or repeat elementprimersare useful methods for outbreak investigations and regional surveillance. Lihrmy typing systems hasedon multi/ocus sequence-based analysis and strain-specific probe hybridization schemes are in developmentlor the international surveillance ofmajor pathogens like Mycobacterium tuberculosis. Accurateepidemiological interpretation ofdata obtained with molecular typing systems still requires additionalresearch on the evolution rate ofpolymorphie loci in bacterial pathogens.

Key words: bacterial typing - DNA polymorphism - pulsed-ficld gel electrophoresis - ribotyping - RFLP typing PCR fingerprinting - randomly amplified polymorphie DNA (RAPD) - cross-infection - epidemiology

surveillance

WHY DO WE NEED EPIDEMIOLOGIC TYPING?

Epidemiologie typing systems can be used foroutbreak investigations, to confirm and delineatethe patterns of transmission of one or more epidemie clone(s), to test hypotheses about the sourcesand vehicles of transmission ofthese clones and tomonitor the reservoirs of epidemic organisms.Typing also contributes to epidemiologic surveillance and evaluation ofcontrol measures, by documenting the prevalence over time and circulationof epidemic clones in infected populations.Clearly, different requirements will be needed forthese distinct applications (Maslow & Mulligan1996, Struelens et al. 1996).

The basic premise of epidemiologic typing isthat isolates of an infectious agent that are part ofthe same chain of transmission are clonally related,that is the progeny of the same ancestor cell. Extensive genomic and phenotype diversity existswithin populations of microbial pathogens of thesame species. This diversity reflects the evolutionary divergence arising from mutations and gene

Fax: +32.555.64.59. E-mail: [email protected] 15 June 1998Accepted 30 July 1998

flux. Clonally related isolates exhibit significantlymore similar characters than unrelated isolates.These distinctive characters, called epidemiological markers, are scored by typing systems whichare designed to optimize discrimination betweenepidemiologically related and unrelated isolates ofthe pathogen ofinterest (Maslow & Mulligan 1996,Struelens et al. 1996). The threshold of markersimilarity used for definition ofa clone need to beadjusted to the species studied, the typing systemused, the environmental selective pressure and thetime and space scale of the study (Tibayrenc 1995,Struelens et al. 1996). Mutation rate and gene fluxvary between species, pathovars and environments.In vivo micro-evolution ofmost pathogens remainspoorly understood. Subclonal evolution and emergence of variants that occur in individual hosts orduring prolonged transmission can be recognizedby several high resolution molecular typing systems, like, for instance, macrorestriction analysisby pulsed-field gel electrophoresis (Struelens et al.1993, 1996).

CURRENT TECHNOLOGIES: HOW WELL DO THEYFULFILL OUR NEEDS?

ln recent years, the deve10pment and extensiveuse of high resolution molecular typing systemsbased on direct analysis ofgenomic polymorphismhave greatly improved the understanding of the

582 Molecular Typing Systems· Marc J Struelens

epidemiology of infectious diseases (Maslow &Mulligan 1996, Struelens et al. 1996). However,the rapid diversification and incomplete comparative evaluation of these methods leave the microbiologist and the epidemiologist faced with a number of questions dealing with selection of the appropriate typing system(s) for solving a particularproblem, as weil as a lack of consensus about interpretation and communication of results.

Several criteria are proposed for evaJuating theperformance of typing systems (Maslow &Mulligan 1996, Struelens et al. 1996). These criteria include: ~vpeahility. reproducihility. stahility.diseriminatory powel: and epidemiologic concordance. Typeability refers to the proportion of isolates that can be scored in the typing system andassigned a type, ideally all isolates. Reproducibili~r refers to the ability of the typing system to assign the same type on repeat testing of the samestrain. Stability is the biological feature ofclonallyderived isolates to express constant markers overtime and generations. The stability ofmarkers maybe acceptable even in the presence of variation,provided that the typing system enables recognition of clonaI relatedness and does not lead tomisclassification ofsubclonal variants as epidemio10gically unrelated. Diseriminatory power is a keycharacteristic oftyping systems, because it conditions the probability that isolates sharing identicalor closely-related types are truly clonaI and part ofthe same chain of transmission. Discriminatingpower can be ca1culated based on Simpson's index of diversity. Ideally, the index, based on testing a large number ofepidemiologically unrelatedisolates, should equal 1. In other words, each independent isolate should be sufficiently differentto be assigned to a distinct clone. In practice, atyping system, or combination of systems, displaying a discrimination index greater than 0.95 is acceptable. This level ofdiscrimination correspondsto a 5% probability of erroneously assigning independent isolates to the same clone. Epidemiologieconcordance is the capacity of a typing system tocorrectly classify into the same clone ail epidemiologically related isolates from a well-describedoutbreak. Additional comparative studies areneeded to establish the relative value of systemscurrently used for typing microbial pathogens.Moreover, there are important variations in theperformance of a given method depending on thespecies and on modifications of the procedure asapplied by different investigators.

In addition to its intrinsic perfonnance whenapplied to a particular microbial pathogen, a typing system should have practical advantages. Versatility, or the ability to type any pathogen, givenminor modifications of the method, is an impor-

tant advantage for the study of nosocomial infections. Other practical aspects of typing systemsinclude ease of performance and ease ofresult interpretation, as weil as cost and availability ofreagents and equipment. Moreover, results shouldbe obtained rapidly enough to be useful in makingdecisions about management of an outbreak. Infection control problems which require rapid typing data include confirmation that an outbreak isoccurring and identification of carriers of the epidemic clone to implement isolation precautions ordecolonization therapy. Because there is no optimal typing system that meets ail the above requirements, it is as a rule neccssary to use a combination of systems. Rapid screening systems can beused initially for preliminary assessment ofclonality. Confirmation can be obtained subsequently, ifrequired, by using more reliable but lessefficient typing systems. Recent reviews have proposed "optimal" first pass and alternate methodsas weil guidelines for interpreting differences fora typing number of bacteria when faced with theneed to investigate outbreaks (Maslow & Mulligan1996, Struelens et al. 1996, Tenover et al. 1997).

Methods that index chromosomal DNA polymorphism are the best options for comparative typing of most bacteria, especially nosocomial pathogens (Tenover et al. 1997). Good resolution of genomic restriction fragment length polymorphisms(RFLP) analysis is obtained by: (i) transfer ofrestriction fragments onto membranes, followed bySouthem-blot hybridization with DNA probes, and/or (ii) use of endonucleases that have infrequent« 30) recognition sites in the chromosome, followed by separation ofthese macrorestriction fragments by pulsed-field gel electrophoresis (PFGE).Different types ofnucleic acid probes are used fortyping: (i) genes encoding metabolic, virulence orresistance functions; (ii) multicopy elements, inc1uding insertion sequences and transposons, and(Ïii) rRNA orrDNA sequences (ribotyping). Southern blot analysis ofgene polymorphism was foundmoderately discriminating but highly reproducibleand stable. Examples include the mee detenninantfor discriminating strains of methicillin-resistantStaphyloeoeeus aureus and the exotoxin A probefor typing Pseudomonas aeruginosa strains fromcystic fibrosis patients. fS:/Îngerprinting, or Southem blot analyses by using insertion sequences asprobes, provides a very reproducible and highlydiscriminating typing tool. Discrimination is relatedto the presences of multiple copies of these elements at diverse locations in the chromosome.Careful selection and optimization of probe sequence, restriction endonucleases, electrophoresisand hybridization conditions need to be developedfor each species or pathovar to be typed. These

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 1998 583

techniques are not rapid and required specializedreagents and expertise. International standardization of technique, reagents, type strains and nomenclature was established by public health reference laboratories for IS6110 RFLP-fingerprintingof Mycobacterium tuberculosis, which integratesstandard computer analysis ofpatterns and a common database, and is now widely applied for largescale surveillance of tuberculosis (Bauer et al.1998).

Ribotyping is the most versatile and the mostwidely used strategy of Southern blot analysis ofbacterial genome polymorphism. The evolutionary conservation of ribosomal RNA makes it applicable as a universal bacterial probe. Many important pathogens, including Enterobacteriaceae.Listeria. Pseudomonas sp. and staphylococci havemore than five ribosomal operons and thus produce ribotype patterns of5 to 15 bands. Ribotypingis a robust method that exhibits excellent reproducibility and stability, both in vitro and in vivoduring the course of outbreaks. It is commerciallyavailable in a fully automated and well-standardized format. However, its discriminatory power isonly moderate, at a level equal or inferior to thatof multilocus enzyme e1ectrophoresis. This is related to the fact that ribosomal operons coyer lessthan 0.1 % ofchromosomallength and tend to cluster in one particular region of the genome. Discrimination of ribotyping depends on speciesand on choice and number of restriction endonucleases used. No consensus has been achieved onoptimal procedure and no general mies are available for interpretation of technically prob1ematicresults, like weakly hybridizing fragments.

Macrorestriction analysis resolved by pulsedfield gel electrophoresis has recently emerged as agold standard for genome fingerprinting of microbial pathogens (Maslow & Muligan 1996, Tenoveret al. 1997). Careful selection of low-frequencyc1eaving enzymes enables cutting the whole bacterial chromosome ofany species into 1ess than 30fragments, typically lOto 700 kb in size. Periodicchange in the orientation of electric field duringagarose electrophoresis, or pulsed field gel electrophoresis, allows separation and size determination of these macrorestriction fragments. Withminor modifications in the selection of enzymesand "pulsing protocols", PFGE can be applied toany bacterium or yeast. Although direct probingof recognition sequences by rare cutters detectsvariation in less than 0.01% of the chromosome,large size rearrangements, like sequence duplication, deletion, or insertion, will be readily detectedas a shift in fragment size and/or number. In comparison with other typing methods, PFGE hasshown equal or greater discriminatory power

(Maslow & Muligan 1996). PFGE requires two tofour days before results are available and spccialized equipment that is more expensive than thoserequired for PCR or Southern hybridization. Nevertheless, because ofits superiorversatility, reproducibility and resolution, genome macrorestrictionanalysis is currently a method ofchoice for typinga majority ofnosocomial pathogens and sorne community-acquired pathogens (Tenover et al. 1997).The sensitivity of PFGE to detect genomic rearrangements makes appropriate interpretation ofminor pattern differences a key to its correct application to outbreak investigations and surveillancestudies. Interlaboratory standardization has not yetreached a sufficient leveJ to allow the use of common type nomenclature or direct DNA pattern exchange.

In recent years, a number ofPCR-based strategies have been developed for strain discriminationof microbial pathogens. In PCR-gene RFLP typing, a target sequence, 1 to 2 kb long and known toshow polymorphism among strains of the speciesof interest, is amplified at high stringency. Theamplified product is cut with restriction endonuc1eases and isolates are compared by RFLP pattern. The PCR-serotyping method takes advantagesof the conserved sequences at each end ofproteinantigens genes, like flagellin and outer membraneproteins ofGram-negative pathogens, for amplification of allehc variant sequences encoding thecentral, antigenically variable portion ofthese proteins. The polymorphic alleles can be determinedby amplicon characterization with suitable restriction endonucleases (PCR-RFLP serotyping(Harrington et al. 1997) or conformational analysis (e.g., single strand conformation ana1ysis, orPCR-SSCP serotyping). The advantages ofthesemethods over conventional serotyping include theunlimited availability of specific reagents, use ofuniversal techniques and typeabi1ity of variantstrains with cryptic antigens.

Although it is a rapid, simple and reproducibletechnique, PCR-RFLP typing has shown so far onlymoderate discrimination. Moreover, it can be biased either by mosaicism due to horizontal transfer, e.g., flagellin gene in Campylobacter jejuni(Harrington et al. 1997) or confounded byhypermutation rate at so--called contingency locithat undergo rapid rearrangements in response toenvironmental changes (e.g., protein A gene polymorphism in S. aureus (van Belkum et al. 1996).

As the logical next step, nucleotide sequencing of PCR-amplified genes is the most sensitiveand accurate means of indexing localized DNApolymorphism for strain typing. However, the timerequired and cost of the procedure are currentlylimiting the use of this method which has been

584 Molecular Typing Systems· Marc J Struelens

applied to type viruses such as hepatitis virusesand HIV, but also bacteria such as Streptococclispyogenes (Perea Mej ia et al. 1997). With the rapidprogress ofautomated, high troughput methods likeDNA chip technology (Chee et al. 1996), it is likelythat PCR resequencing will be increasingly usedfor epidemiologic typing of viruses, bacteria andother pathogens in the years to come.

Arhitrarily-primed PCR (AP-PCR) typing, andsimilar methods like RAPD (random amplifiedpolymorphie DNA) and DAF (DNA amplificationfingerprinting), are based on low-stringency PCRamplification by using a single, lOto 20-mer primerof arbitrary sequence. In the early cycles of thePCR reaction, the primer anneals to multiple sequences with partial homology, and fragments ofDNA Iying within less than a few kb between annealing sites on opposite DNA strands are amplified. After additional cycles, a strain-specific array ofamplified DNA segments ofvarious sizes isobtained. This simple and rapid technique has beensuccessfully applied to genotypic strain delineation and genetic population analysis of a broadrange of microbial pathogens, including bacteria,fungi and protozoans. Ali isolates are typeable andno prior knowledge oftarget genome sequences isnecessary. Discrimination is good and correlatesweil with other genotyping techniques. The discriminatory power is variable according to number and sequence ofarbitrary primers and amplification conditions. In spite of its attractive efficiency, AP-PCR typing suffers from problems inreproducibility and from the lack ofconsensus rulesfor interpretation of pattern differences (Maslow& Mulligan 1996, Struelens et al. 1996). A number oftechnical factors need to be strictly standardized for optimal reproducibility (Grundmann et al.1997). Progress toward enhanced resolution andreproducibility of analysis of PCR products isachieved by incorporating fluorescent primers inthe reaction and performing computer-analysis ofamplimer patterns by an automated laser fluorescence detection system. In general, differences inprotocols, equipment, or even the batch ofreagentused result in different AP-PCR patterns, but theoverall clustering and grouping of isolates intoidentical, similar, or divergent patterns is reproducible. This makes the method adequate for rapidcomparative typing but less suitable for library typing in surveillance programs.

Repetitive element PCR (rep-PCR) typing consists of PCR amplification of spacer fragments lying between repeat motifs of the genome by use oftwo outwardly-directed primers at high stringency.Short, repetitive elements which have been successfully used as targets for rep-PCR typing include the repetitive extragenic palindromes (REPs),

the enterobacterial repetitive intergenic consensus(ERIC) sequences, insertion sequences and otherspecies-specific repeat elements (Maslow &Mulligan 1996, Deplano et al. 1997). These repPCR strategies produce fewer amplified DNA fragments than AP-PCR, but can nevertheless providegood discriminatory power. Their major advantageis a better reproducibility as compared with APPCR analysis, which may enable their standardization for use as library typing systems.

Another set of innovative PCR-based strategy,which also appears to offer high resolution andgood reproducibility, are the amplified fragmentlength polymorphism (AFLP) method (Vos et al.1995) and inji-eqllent restriction site amplification(fRS-PCR) (Mazurek et al. 1996). In these methods, a restriction-ligation step produces restrictedgenomic DNA fragments tagged with speciallydesigned adapters. A set of different primerscomplementary to these adapters and adjacentnucleotides are then used to PCR amplify variousparts of the tagged restriction fragments, therebyselectively highlighting a subset ofrestriction fragments. More studies are needed to determine thestability ofthese markers over time, establish criteria for interpretation of pattern differences andevaluate inter-Iaboratory reproducibility.

Finally, specialized genotyping schemes usereverse dot hlot or line blot binary hybridizationpatterns of crude genomic DNA or amplified regions thereof with immobilized, clone-specificDNA probes. This method has been developed fortyping ofS. aureus (van Leeuwen et al. 1996) andM. tuberculosis (Kamerbeek et al. 1997). Theselibrary probe genotyping systems provide unambiguous, numeric clonai signatures that should bereproducible between laboratories. Inclusion ofadditional polymorphic sequences should increasethe discrimination to the level needed for surveillance of major pathogens. The power ofthese genotypic hybridization schemes could be much enhanced by the use ofhigh density DNA probe assays, as this technology currently allows parallelanalysis of 104 target sequences within a few hours.

HOW TO INTERPRETE DIFFERENCES OBSERVEDBETWEEN GENOTYPES?

We use molecular typing systems in epidemiologic studies to determine if isolates are clonallyrelated and thus belong to the same chain of transmission. When a set of isolates show identicalDNA banding patterns, this clue to clonality is proportional to the number oftyping systems used andtheir discriminatory power. A problem arises whenpatterns are similar but not identical. What levelof pattern similarity can be used to define clonallyor epidemiologically related organisms? This level

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), 5ep.lOct. 1998 585

needs to be adjusted to the resolving power of thesystem used, the genomic plasticity of the organism under study and the time scale of the investigation. Suggested rules for interpretation of differences in PFGE patterns, as applied to outbreakinvestigations (Struelens et al. 1996, Tenover etal. 1997), relate the graduai increase in the number of restriction fragment mismatches with increasing number of genetic differences and withdecreasing probability of epidemiologic relatedness. Calculation of restriction/hybridization pattern similarity coefficients and graphical displayof pattern relatedness as dendrograms is also usefuI for interpretation, particularly for large scalestudies (Struelens et al. 1996). Altough this quantitative analysis has been criticized as invalid forphylogenetic inferences, because DNA restrictionfragment pattern variation is not due to independent events, it is supported by population analysis, e.g. of P aeruginosa. Additional populationgenetic and micro-evolution studies are needed toprovide a better understanding of the nature andfrequency of molecular events giving rise to genomic polymorphisms exploited empirically byepidemiologists for strain typing (Tibayrenc 1995,Struelens et al. 1996)

Molecular typing systems are undergoing rapidtechnical improvements. Advances in the understanding of biological basis of microbialbiodiversity at subspecies levels will improve theconceptual framework required for proper epidemiologic interpretation of typing results. Widerapplication of these systems should shed light tothe epidemiology of hospital and community-acquired infections and, therefore, allow for moreeffective control and prevention strategies.

REFERENCES

Bauer J, Yang Z, Poulsen S, Andersen AB 1998. Results from 5 years of nationwide DNA fingerprinting ofMycohacterium tuberculosis complex isolatesin a country with a low incidence ofM. tuberculosisinfection. J Clin Microbiol 36: 305-308.

Chee M, Yang R, Hubbell E, et al. 1996. Accessing genetic information with high-density DNA arrays.SCIence 274: 610-614.

Deplano A, Vaneechoutte M, Verschraegen G, StruelensMJ 1997. Typing ofStaphylocoeeus aureus and Staphylococcus epidermidis strains by PCR analysis ofinter-IS256 spacer length polymorphisms. J ClinMicrobiol 35: 2580-2587.

Grundmann HJ, Towner KJ, Dijkshoom L, et al. 1997.Multicenter study using standardized protocols andreagents for evaluation of reproductibility of PCRbased fingerprinting of Acinetobacter spp. J ClinMicrobiol 35: 3071-3077.

Harrington CS, Thomson-Carter FM, Carter PE 1997.Evidence for recombination in the fJagellin locus of

Campylohaeter jejuni : implications for the fJagellin gene typing scheme. J Clin Microhiol 35: 23862392.

Kamerbeek J, Schould L, Kolk A, et al. 1997. Simultaneous detection and strain differentiation of Mycohaeterillm tubereulosis for diagnosis and epidemiology. J Clin Micmbiol 35: 907-914.

Maslow J, Mulligan ME 1996. Epidemiologie typingsystems. Infect Control Ho.\p Epidemiol 17: 595604.

Mazurek GH, Reddy V, Marston BJ, Haas WH, CrawfordJT 1996. DNA fingerprinting by infrequent-restriction-site amplification. J Clin Microbiol 34: 23862390.

Perea Mejia LM, Stockbauer KE, Pan X, Cravioto A,Musser JM 1997. Characterization ofgroup A Streptoeoeells strains recovered from Mexican childrenwith pharyngitis by automated DNA sequencing ofvirulence-re1ated genes : unexpected1y large variation in the gene (sic) encoding a complement-inhibiting protein. J Clin Microbiol 35: 3220-3024.

Struelens MJ, and the Members of the European StudyGroup on Epidemiologica1 Markers (ESGEM), ofthe European Society forClinical Microbiology andInfectious Diseases (ESCMID) 1996. Consensusguidelines for appropriate use and evaluation ofmicrobial epidemiologic typing systems. ClinMicrobiollnfect 2: 2-11.

Struelens MJ, Schwam V, Deplano A, Baran D 1993.Genome macrorestriction analysis of diversity andvariability of Pseudomonas aeruginosa strains infecting cystic fibrosis patients. J Clin Microbiol 31:2320-2326.

Tenover FC, Arbeit RD, Goering RV, the MolecularTyping Working Group ofthe Society for HealthcareEpidemiology of America 1997. How to select andinterpret molecular typing methods for epidemiological studies of bacterial infections: a review forhealthcare epidemiologists. Infect Control HospEpidemiol18: 426-439.

Tibayrenc M 1995. Population genetics ofparasitic protozoa and other microorganisms. Advances Parasitol136: 47-115.

van Belkum A, van Leeuwen W, Kaufmann ME, et al.1998. Assessment of resolution and intercenter reproducibility of results of genotyping methicillinresistant Staphyloeoccus aureus by pulsed-field gelelectrophoresis ofSmaI macrorestriction fragments:a multicenter study. J Clin Microbiol36: 1653-1659.

van Belkum A, Riewerts Eriksen N, et al. 1996. Arevariable repeats in the spa gene suitable targets forepidemiological studies ofmethicillin-resistant Staphylococcus aureus strains? Eur J Clin MicrobiolInfect Dis 15: 768-769.

van Leeuwen W, Sijmons M, Sluijs J, Verbrugh H, vanBelkum A 1996. On the nature and use ofrandomlyamplified DNA from Staphylococcus aUl"eus. J ClinMierobiol 34: 2770-2777.

Vos P, Hogers R, Bleeker M, et al. 1995. AFLP: a newtechnique for DNA fingerprinting. Nucleie AcidsRes 23: 4407-4014.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 587-588, Sep.lOct. 1998 587

Vancomycin-resistant Enterococci in IntensiveCare Hospital Settings

Daren J Austin/+, Marc JM Bonten*

Wellcome Trust Centre for the Epidemiology of Infectious Diseases, University of Oxford, South Parks Road,Oxford OXI 3PS, UK *Department oflntemal Medicine, University Hospital Utrecht, PO Box 85000, 3508, GA

Utrecht, Netherlands

Vancomycin-resistant enterocucci (VRE) have recently emerged as a nosocomial pathugen andpresentan increasing threat to the treatment ofseverely il! patients in intensive-care hospital settings. We outtine results of a study of the epidemiology o( VRE transmission in fCUs and define a reproductivenumber Ro; the number ofsecondary colonization cases induced by a single VRE-coloni;:ed patient in aVRE-free lCU. (or VRE transmission. For VRE to become endemic requires RO> l. We estimate that inthe absence ofinfection control measures Ro lies in the range 3-4 in defined ICU settings. Once infectioncontrol measures are included Ro=O.6, suggesting that admission of VRE-colonized patients can stabitize endemic VRE.

Key words: vancomycin resistance - transmission dynamics - epidemiology - infection control

Increasing antibiotic resistance in common bacterial pathogens presents a growing threat worldwide. The emergence of vancomycin-resistant enterococci (VRE) as a nosocomial pathogen is a striking example of this new danger to vulnerable patients. In both the United States and the UnitedKingdon, the frequency with which isolates havebeen recovered has increased dramatically duringthe past seven years (CDC 1995, PHLS 1996). Forpatients infected with VRE treatment options areoften limited and control ofoutbreaks relies heavilyon conventional infection control procedures(Husani & Raad 1997).

Molecular epidemiological studies ofearly endemic infections suggested that single clones wereprimarily responsible (Boyce et al. 1995). Morerecently, many outbreaks appear to involve morethan one clone, indicating reintroduction (Morriset al. 1995, Siaughter et al. 1996, Bonten et al.1996). Since enterococci form part of our naturalflora, it was believed that new hospital outbreakswere a result of endogenous sources. However recent studies have revealed that transmission ofVREvia the hands of transiently colonized heath-careworkers (HCWs) is a very important determinantof spread and persistence (Bonten et al. 1996).

DJA thanks the WeIlcomc Trust for continued support.MJMB thanks Merck Sharpe & Dohme and the Infectious Diseases Society of the Netherlands and Flandersfor grant support.+Corresponding author. Fax: +44-1865-281245. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

ln a previous study conducted at the CookCounty Hospital (CCH) Chicago, IL, that measuredthe relative efficacies of various barrier infectioncontrol precautions (Slaughter et al. 1996) rectalcultures were taken daily and environmental cultures monthly (Slaughter et al. 1996). Using pulsedfield gel electrophoresis, a total of 19 strain typesof VRE were identified. In a follow-up study itwas subsequently confirmed that patients ratherthan the environment provide the major reservoirofVRE (Bonten et al. 1996).

The transmission dynamics ofVRE in an ICUsetting can be represented by a set of coupled differential equations with framework summarized byFig. If we view patients as definitive hosts andHCWs as vectors for transmission, then the structure of the model reduces to that of the RossMacdonald equations for malaria transmission(Anderson & May 1991). A central concept in infectious disease transmission is the reproductivenumber Ro;the number ofsecondary cases ofVREcolonization generated by a single primary case ina VRE-free ICU. IfRo> 1 an outbreak ofVRE willpersist and become endemic with prevalence 1-1/Ro' If Ro<1 the outbreak will fade to extinction.For indirect VRE transmission via HCWs Ro isdefined as

RO= m bp bs ,.2Dp D~

where m is the staff-patient ratio, bp and bs are therespective probabilities of transmission fromHCW-patient and vice-versa, r is the staff-patientcontact rate (patient contacts per unit time) andparameters Ds and Dp represent the average duration VRE remains transmissible on the hands ofHCWs (typically one hour) and from patients (typi-

588 Vancomycin-resistant Enterococci • DJ Austin, MJM Bonten

cally the duration oftheir stay in the ICU i.e., days).The contact rate appears as a squared quantity reflecting the patient-HCW-patient nature of transmission.

, '--__----l,, ,, ,

;<,, ,, "

1~~:'I: :1_V_~_c_;_:_e_1Model of indirect patient-health care workers (HCW)-patientvancomycin-resistant enterococcl (VRE) transmission. Patientsare admitted at a rate A per day with a fraction lj> already colonized. Dashed arrows indicate indirect transmission. Colonizedpatients remain in the ICU for duration D and HCWs can trans-

. PmIt VRE for duratlOn D,.

Infection control measures will influence Ro indifferent ways. Barrier precautions such as handwashing reduce the probability ofHCWs transmitting VRE, once colonized, by a factor (l-p) wherep is the observed compliance with hand-washingmeasures. Cohorting HCWs will reduce the effective staff-patient ratio m by a factor (l-q) where qis the proportion of staff cohorted to a single patient. Increasing the numbers of HCWs will increase m but may reduce the patient contact rate rand hence Ro'

Estimates of compliance with barrier precautions have been reported as less than 50%(Doebbbelling et al. 1992), suggesting that VREwill not controlled ifRo>2. Cohorting ofstaffmembers via one-to-one nursing can give very highcohorting levels, perhaps 80% with a correspondingly higher eradication Ro<5. Indeed outbreakshave been brought under control using just suchmethods (see e.g., Haley et al. 1995).

During the course of a 133-day study of endemic VRE at CCH a mean compliance of 51 %was observed and the level ofcohorting ofHCWswas estimated to be 80%. The mean endemic prevalence ofVRE was found to be 36% (95% CI 3-68)and 15% of patients were already colonized onadmission (Slaughter et al. 1996). Our analysesindicate that the effective reproductive number (inc1uding infection control measures) Ro(p,q)=0.6,corresponding to a true reproductive numberRo=3.11 in the absence of infection control measures. We deduce that since Ro(p,q)<I, infectioncontrol would ordinarily control VRE. Howeverthe admission of colonized patients continued tostabilize endemic VRE. The observed reduction inVRE transmission is considerable. In the absence

of infection control, the predicted endemic prevalence ofVRE is 75% compared with an observedmean of36%.

The use of molecular epidemiology has demonstrated that the primary determinant of endemicVRE is indirect patient-HCW-patient transmission,rather than environmental or endogenous sources.Using a precise mathematical framework enablescareful analysis of the transmission dynamics ofVRE and allows for quantitative measurements ofboth transmission and, more importantly, intervention can be made. As treatment options becomemore limited, c1inicians will become ever morereliant on conventional infection control procedures. The quantitative measurements outlined canbe used to assist in enabling better management oflimited resources to combat the threat of VRE inICU hospital settings.

REFERENCES

Anderson RM, May RA 1991. Infectious Diseases ofHl/mans: Dynamics and Control, Oxford UniversityPress, 755 pp.

Bonten MJM, Hayden MK, Nathan J, van Voorhis J,Matushek M, Siaughter S, Rice T, Weinstein RA1996. Epidemiology of colonization of patients andenvironment with vancomycin-resistant enterococci.Lancet 348: 1615-1619.

Boyce JM, Mennal LA, Zervos MJ, Rice LB, PotterBynoe G, Giorgio C, Medeiros AA 1995. Controlling vancomycin-resistant enterococci. Infect Control Hosp Epidemiol 16: 634-637.

CDC 1995. Recommendations for preventing the spreadofvancomycin-resistance. MMER 44 (RRI2): 1-13.

Doebbelling BN, Stanley GL, Sheetz CT, Pfaller MA,Houston AK, Annis L, Li N, Wenzel RP 1992. Comparison of efficacy of alternative hand-washingagents in reducing nosocomial infections in intensive-care units. NEJM 327: 88-93.

Haley RW, Cushion NB, Tenover FC, Bannennan TL,Dryer D, Ross J, Sanchez PJ, Siegel JD 1995. Eradication of endemic methicillin-resistant Staphylococcl/ aureus from a neonatal intensive-care unit.J b!fect Dis 171: 614-624.

Husani R, Raad l 1997. Treatment and prevention ofvancomycin-resistant enterococcus. Cl/rI' Opin Intensive Care 10: 431-434.

Morris Jr JG, Shay DK, Hebden JN 1995. Enterococciresistant to multiple antimicrobial agents, includingvancomycin: establishment of endemicity in a University Medical Center. Ann Internai Med 123: 250259.

PHLS 1996. Vancomycin-resistant enterococci in hospitals in the United Kingdom. CDR Weekly 6: 1.

Siaughter S, Hayden MK, Nathan C, Hu TC, Rice Tl,van-Voortis J, Matusheka M, Franklin C, WeinsteinRA 1996. A comparison of the universal use ofgloves and gowns with that of glove use alone onthe acquisition ofvancomycin-resistant enterococciin a medical intensive-care unit. Ann Internai Med125: 448-456.

Mem fnsl Oswafdo Cruz, Rio de Janeiro, Vol. 93(5): 589-594, Sep./Oct. 1998 589

Molecular Genetic Analysis of Multi-drug Resistance inIndian Isolates of Mycobacterium tuberculosis

Noman Siddiqi/++, Md. Shamim/++, NKJain*, Ashok Rattan**, Amol Amin,VM Katoch***, SK Sharma****, Seyed E Hasnain/+

National Institute of Immunology, New Delhi, 110067, India *New Delhi T.B. Centre. New Delhi, India **Department of Microbiology, A.I.I.M.S., New Delhi, India ***Central .Talma Institute of Leprosy, Agra, India

****Department of Medicine, A.I.I.M.S., New Delhi, India

A total of 116 isolatesfrom patients attending the out-patient department at the Alllndia Institute ofMedical Sciences, New Delhi and the New Delhi Tuberculosis Centre. New Delhi. India were collected.They were analyzedfor resistance to drugs prescribed in the treatmentfor tuberculosis. The drug resistance was initially determined by microbiological techniques. The Bactec 460TB system was employedto determine the type and level ofresistance in each isolate. The isolates werefurther characterized atmolecular leve!. The multi-drug loci corresponding to rpo ~, gyr A, kat G were studiedfor mutation(s)by the polymerase chain reaction-single strand conformational polymorphism (PCR-SSCP) technique.The SSCP positive samples \Vere sequenced to characterize the mutations in rpo ~, and gyr A loci.While previously reported mutations in the gyr A and rpo ~ loci were found to he present, several novelmutations were also scored in the rpo ~ locus.

Interesting~y. analysis ofthe gyr A locus showed the presence ofpoint mutation(s) that could not hedetected by PCR-SSCP Furthermore, r(fampicin resistance was found to be an important marker forchecking multi-drug resistance (MDR) in clinical isolates of Mycobacterium tuberculosis. This is the.first report on molecular genetic analysis ofMDR tuberculosisfrom India, and highlights the increasingincidence ofMDR in the Indian isolates of M. tuberculosis.

Key words: clinical isolates - gyrA gene - multi-drug resistance - Mycobaclerium tuberculosis - rpo f3 genepolymerase chain reaction-single strand conformational polymorphism

Until recently, the cornrnon beliefheld that tuberculosis (TB) no longer posed a major threat topublic health, at least in developed countries. However due to various reasons there is an increasingincidence ofTB leading to high morbidity and mortality rates (Bloom & Murray 1992). Furthermore,the association of TB with the AIDS pandemicleading to increase in fatality rates and emergenceofmuiti-drug resistant (MDR) strains ofMycobacterium tuberculosis is a cause of grave concemworldwide (Iseman 1994).

Resistance to rifampicin, isoniazid andfluoroquinolones have been well studied and characterized at the molecular level (Honore & Cole

This project was supported by a grant from the Department of Biotechnology, Ministry of Science and Technology, lndia.+Corresponding author. Fax: +9111-616.2125. E-mail:[email protected]!.in++Both authors contributed equally.Received 15 June 1998Accepted 30 July 1998

1993, Heym et al. 1993, Miller 1994, Takiffet al.1994). Rifampicin resistance arises due to mutations in rpo ~ gene encoding the DNA-dependentRNA polymerase. The primary target of rifampicin is the ~-subunit of RNA polymerase. The association of the RNA polymerase ~ (rpo ~) subunit gene with resistance to rifampicin has beendocumented previously and subsequent reportsfrom various groups have confirmed this association in clinical isolates of M. tuberculosis (Kapuret al. 1994, Williams et al. 1994, Musser 1995,Hasnain et al. 1998). Most of the mutations havebeen mapped to the 27 codons located at the center of rpo ~ gene that is known to bear mutationsthat confer rifampicin resistance in Escherichiacoli. Many of the reported mutations are missensewith a few cases of insertions and deletions also(Telenti et al. 1993a, b). Resistance to rifampicinis a relatively rare event and leads to selection ofmutants that are already resistant to other components of short-course chemotherapy. Therefore,rifampicin resistance is often regarded as an excellent surrogate marker for MDR-TB (CDC 1993).

Isoniazid acts as the prodrug which is convertedto an active form (isonicotinic acid or aldehyde-

590 MDR-M.tb in India • Noman Siddiqi et al.

bearing groups or free radicals) by the kat G-encoded catalase-peroxidase enzyme in M. tuberculosis. Isoniazid resistance is due to conversion ofArg463 to Leu in the kat G prote in (Heym et al.1993, 1994). The second mechanism, conferringlow level resistance, is mutation in the inh A genewhich encodes fatty acid synthase. This enzymerequires NAOH as a cofactor; the mutant enzymehas been shown to have a lower affinity for NADHand cannot be saturated at NADH concentrationsexisting within M. tuberculosis (1ohnsson &Schultz 1994).

The target of fluoroquinolones action is theDNA gyrase, an ATP-dependent type II DNAtopoisomerase that catalyses the negative supercoiling of DNA. This enzyme is made up of fourunits (a2132), which are encoded by the gyrA andgvrB genes respectively. Fluoroquinolones bindto the gyrase and inhibit the supercoiling ofDNA.The gyrA and gyrB genes of M. tuberculosis havebeen c10ned and mutations in the quinolone-binding site have been mapped (Takiff et al. 1994).

The present study represents the first report ofmolecular genetic analysis of c1inical isolates ofMOR M. tuberculosis from India. While association ofrifampicin resistance with MOR is evident,we also demonstrate the utility ofpolymerase chainreaction-single strand conformational polymorphism (PCR-SSCP) for rapidly scoring mutationswithin the rpo 13 locus. Sequence analysis of therpo 13 and gyrA loci shows that the more commonmissense mutations are also prevalent in the Indian isolates. This study reaffirms the growing incidence of MOR-TB in India.

MATERIALS AND METRODS

Clinical isolates of M. tuberculosis were procured from TB patients attending the out-patientdepartments at the Ali India Institute of MedicalSciences and the New Delhi Tuberculosis Center,New Delhi, India. Susceptibility testing of ail theM. tuberculosis isolates was done by Bactec 460TBsystem. Minimum inhibitory concentration (MIC)was defined as the lowest drug concentration thatinhibited bacterial growth by at least 99%.

DNA was extracted from c1inical isolates grownon LJ slants. The colonies were scraped, suspendedin TE and subjected to freeze thawing (30 min at 70°C to 100°C for 10 min). This was followed bytreatment with lysozyme (40 ~g/ml), SOS (0.5%)and proteinase K (50 ~g/ml) at 37C for 2 hr. Theprotein and other contaminants were removed byCTAB precipitation. DNA was finally precipitatedwith 0.6 volumes of isopropanol. The precipitatewas washed twice with 70% ethanol, air dried andre-dissolved in water.

PCR was performed using 150 pmoles ofeachprimer with 1 U of Taq polymerase (BangaloreGenei, India), 200 ~moles of each dNTPs and 1.5mM magnesium chloride. Mutations in the rpo 13gene, conferring rifampicin resistance were detected with previously reported forward (5 TACGGT CGG CGA GCT GAT CC 3) and reverse (5TAC GGC GTT TCG ATG ATG AAC 3) primers.250 ~g of template was ampl ified in 30 cycles inPerkin Elmer Cetus thermal cycler using the following conditions: 94°C_I min; 56°C-1 min; noc2min. The 350 bp rpo 13 amplicon was checked byelectrophoresis on a 1.5% agarose gel.

Similar conditions were used for amplificationof the kat G genes mutational hot spot region using forward (5 GCC CGA GCA ACA CCC 3)and reverse (5 ATG TCC CGC GTC AGG 3) primers.

Sixteen isolates resistant to ofloxacin werechecked for mutations in gyr A. For gyr A amplification the primers used were 5 CAG CTA CATCGA CTA TGC GA 3 and 5 GGG CTT CGG TGTACC TCA T 3. The PCR amplification conditionswere similar to that of rpo 13 gene amplificationexcept that the annealing temperature was 45°C.The amplicons were purified using Qiaquick PCRpurification kit (Qiagen, USA). Samples were initially analyzed for SSCP. Briefly, the samples wereheat denatured and electrophoresed on a composite gel (0.25% agarose, 5% acrylamide and 5%glycerol). The gel was then silver stained (BioRadSilver stain Kit) and the DNA bands visualized. Achange in the banding pattern as compared to thewild type H37Rv strain was taken as indicative ofmutation(s).

Cycle sequencing was performed following themanufacturers protocol using the cyclist ?(u Exokit (Stratagene, USA).The forward primers of rpo13 and gyr A were used to sequence the SSCP positive samples. The 81 bp and 30 bp sequence corresponding to the hot spot region of the rpo 13 andgyr A respectively, were read and compared withthe respective sequences ofstandard H37Rv strain.

RESULTS

The MOR data for the 116 c1inical isolates isshown in Table 1. A majority of the strains (69%)turned out to be resistant to at least one drug. It isimportant to mention at the outset that this studyhas a sampling bias in terms of the drug resistancecases and does not represent the observed frequency ofthe occurence ofdrug resistance (-13%)in Indian clinical isolates of M. tuberculosis(Ramalingaswami 1998). Between various drugsused in short term chemotherapy for tuberculosis,the isoniazid resistant was most common (56%),

Mem /nst Oswa/do Cruz, Rio de Janeiro, Vol. 93(5), 5ep./Oct. 1998 591

TABLE! A 2 3 4 5 6 7 8Summary of mulli drug resitance in Mycobaclerium

lubercu/osis isolates from Indian patients

Drugs No. of isolates MRD Status Samples

Isoniazid 65 Idrug 15Rifampicin 62 2 drugs 25Ethambutol 27 3 drugs 20Streptomycin 25 4 drugs 5Ofloxacin 16 5 drugs 13None 38

Total No. of samples: 116 B 1 2 3 4 5 6 7 8

closely followed by rifampicin (53%). The number of isolates resistance to the rest of the drugswas lesser (about 25% for ethambutol and streptomycin) and was least in case offluoroquinolones.This data corelates weil with the treatrnent regimefollowed in the TB clinics in India wherefluoroquinolones represent the last line of drugs(Pande 1998). We observed that 58 isolates out of62 positive for rifampicin resistance, were also resistant to atleast one other drug. Therefore, a majority (93%) of the rifampicin resistant strainsshowed an association with resistance to otherdrugs thereby supporting earlier observations(Kapur et al. 1994, Williams et al. 1994, Musser1995, Hasnain et al. 1998) on rifampicin as a surrogate marker for multidrug resistance.

The principle of PCR-SSCP is based on the factthat the two denatured strands ofDNA (in this casePCR-amplified) adopt stable intramolecular conformations which may differ from the wild typeupon mutation. This causes a change in the electrophoretic mobility of the strands. We utilizedSSCP to conduct a primary screening of the rpo ~,

kat Gand gyr A amplicons for the presence ofmutations. The results ofSSCP analysis reveal thatwhile most of the rifampicin resistant strains didexhibit the expected mobility shifts correlating withpoint mutations, a very large percentage ofisolatesresistant to fluoroquinolone and carrying point mutations (revealed upon sequencing), however didnot display altered electrophoretic mobility. Theresults of SSCP analysis of rpo ~ amplicons forfew isolates are summarized in Table II. Sorne typical SSCP gel electrophoresis patterns corresponding to the rpo ~ (Fig. la), kat G (Fig. lb) or gyr Aamplicons (Fig. Ic) are represented.

The precise mutations within the drug resistantloci were identified by direct sequencing of theamplified regions. Sequence analysis (Fig. 2) ofthe rpo ~ gene hotspot for the rifampicin resistantisolates revealed the presence ofmany of the common mutations reported earlier (Kapur et al. 1994)in addition to several novel mutations (data not

Fig. ta: PCR-SSCP analysis of representative rpo f3amplicons.gel picture of few rpo f3 amplicons. Arrow denotesthe shift in bands due to the conformational polymorphism as aresult of point mutation. Lanes 3 is the control lane; Fig lb:typical PCR-SSCP pattern displayed by kat G amplicons.Arrowmark denotes the DNA mobility shift due to the conformational polymorphism as a resull of point mutation within thekat G locus; Fig. 1c: typical PCR-SSCP gel picture of few mutant gyr A amplicons. Arrow mark denotes the DNA mobilityshift due to the conformational polymorphism as a resull ofpoint mutation within the gyrA locus. Lanes 210 8 are ampliconsfrom quinolone resistant strains while lane 1 has the wild typeH37Rv strain.

shown). Mutations in the gyr A gene, however wereonly of the types already reported earlier (Takiffetal. 1994 ). The S95T and A90Y mutations werethe two more common mutations within the Indianisolates ofquinolone resistant M. tuberculosis. Theamino acid changes caused by the correspondingpoint mutations in the DNA within the gyr A lociare shown in Fig. 3.

592 MDR-M.tb in India • Noman Siddiqi et al.

TABLE II

Susceptibility to rifampicin and polymerase chainreaction-single strand conforrnational polymophism

(PCR-SSCP) results. '1' mark denotes that theseisolates were resistant but were not positive in SSCP

analysis

DISCUSSION

In this study we present molecular geneticanalysis of rifampicin, isoniazid and quinoloneresistance in Indian clinical isolates of MDR M.tuberculosis. Point mutations within l'pO ~ , kat Gand gyr A genes respectively lead to amino acidpolymorphism in the target prote in of the drug resulting in drug resistance (Rattan et al. 1998). MDRdoes not appear to arise due to the acquisition of atransposable element or a plasmid carrying drugresistant marker, but is perhaps a reflection ofstepwise acquisition ofnew mutations in the genesfor different drug targets. Alterations in the chromosomal genes are random but get selected due topoor compliance or prescription. Inadequate prescription ofchemotherapy, poor compliance of thedrug regime and in recent times infection with HIVhave caused an increase in the selection of MDRstrains ofM. tuberculosis (Lederberg 1998). A lackof monitoring programs and poor fol1ow up of thepatients health has caused an increase in relapsecases for TB. In most cases, the secondary infection is by drug resistant mycobacterium. What isof graver concem is the increasing incidence ofprimary infection by MDR- Mycobacterium tuberculosis (Hasnain et al. unpub. data). Patients infected with a rifampicin resistant strain of M. tuberculosis general1y have a poor prognosis, particularly because rifampicin resistance is often associated with resistance to other frontiine drugs.

We found a majority of the rifampicin resistantIndian isolates to be resistant to at least one another anti-tubercular drug, supporting the idea ofusing rifampicin resistance as a surrogate markerfor MDR TB. PCR-SSCP analysis while offeringa rapid method for detecting MDR particularly forrifampicin resistance however, has limitations asit failed to detect mutations within the gyr A locus.These results representing a first report on themolecular genetic analysis ofMDR in clinical isolates ofM. tuberculosis from India have importantbearing on the management and control ofthis reemerging infectious disease.

Isolate no.anaIysis

C46C49C56C64C68C7JC74C78C80C93C97AIA2A3A4A5A6A7A8A9BIB2B3B4B5B6B7FIF2F3F4F5F6F7F8F9DI020304060709DIO011012018VIV2V3V4V5V6

Rifampicinsensitivity (> Il!g/mi)

SensitiveResistantResistantResistantSensitiveResistantSensitiveResistantSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveSensitiveResistantSensitiveResistantResistantResistantResistantResistantResistantResistantResistantResistantSensitiveResistantResistantResistantResistantResistantResistantResistantResistantResistantSensitiveResistantResistantResistantResistantResistantResistant

SSCP

NegativePositivePositivePositiveNegativeNegative?NegativePositiveNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativeNegativePositiveNegativePositivePositivePositivePositivePositivePositivePositivePositivePositiveNegativeNegative?Negative?Negative?Negative?PositivePositivePositivePositivePositiveNegativePositivePositivePositivePositivePositivePositive

V7V8V9VIOVIIVI2Vi3Vl4VI5VI6VI7VI8

SensitiveResistantResistantResistantResistantSensitiveResistantResistantResistantResistantResistantResistant

NegativePositivePositivePositivePositiveNegativePositivePositivePositivePositivePositivePositive

Mem Ins! Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 1998 593

CG C.\ T TTG

TTOGA

TG C

CTG ACiC CAA TTC .\ 1(, G,\C ('AC; AAC AAC CCG ('TG TCC, (;(iCi TI Ci ACC CAC AAG CGC CGA l'TG TCG GC('. ( 1(i

511 L S Q F .~f D Q N N P L (i H K R R L A 5)~

P K Y Y l- l'

R L V D W

E LRQ

Fig. 2: common muwtlons in the 'po f3 gene. Tor ranel shows the nueleollde changes that arc sn11llar to rreviollsly rerortedmutatIons. The bottom ranel deplcts the corresronding amino aCld changcs within the RNA polymerase f3 subumt.

1: c G ~

P,ï • • • • 96

2564 CAC GGC GAC GCG TCG ATC TAC GAC AGC CTG 2594

HGDAS IYD SL1 1 1 1v P G T

FIg 3: mlilatlOns wlthin the KI'f A genc. The top panel showsIhe nllcleotide change while the bollom ranel depicts the corresponding amino aCld changes. The common mutations 111 theIndian isolales are depicted in bold leuers.

ACKNOWLEDGMENTS

To Mr Sunder Singh Bisht for his help in sequencing and SSCP.

REFERENCES

Bloom BR. Murray IL 1992. Tuberculosis: commentary on a reemergent killer. Science 257: 1055-1064.

CDC - Centers for Disease Control 1993. Initial therapyfor tuberculosis in the era of multi-drug resistance.recommendations of the advisory council for theclimination oftuberculosis. MMWR 42: RR-7.

Hasnain SE. Amin A. Siddiqi N, Shamim M, Jain NK,Rattan A, Katoch VM, Sharrna SK 1998. Moleculargenetics of multiple drug resistance (MOR) in Mycohacterillln luherClllosis, p. 35-40. In RLSinghal& OP Sood (cds), Drug Resislance: Mechanism andManagemenl, Proceedings of the Fourth AnnuaJRanbaxy Science Foundation SymposiumCommunicore Publishers, New Delhi. India

Heym B, Alzari PM, Honore N, Cole ST 1994. Missense mutations in the eatalase-peroxidase gene,katG. arc assoeiated with isoniazid resistance inMycohaclerium tllherculosl:Ç. Mol Mlcrohiol15: 235245.

Heym B. Zhang Y, Poulet S, Young D. Cole ST 1993.Charaeterization of the katG gene eneoding a eatalase-peroxidase required for the isoniazid suseeptibility of Mycobactertllm tllherclllosis. J Bacteriol175: 4255-4259.

Honore N, Cole ST 1993. The moleeular basis ofrifampin resistanee in Mycohacleriu/ll leprae.Antimicroh Agents Chemother 37: 414-418.

Iseman MD 1994. Evolution of drug-resistant tubereulosis: A tale oftwo species. Proc Natl Acad SCI USA91: 2428-2429.

Johnsson K. Schultz PG 1994. Mechanistic studics ofthe oxidation of isoniazid by the eatalasc peroxidasefrom Mycohacter/llm tuherclllosis. J Am ('hem Soc1/6: 7425-7426.

Kapur V, Ling Ling LI, lordaneseue S, Krieswirth BN,Musser lM 1994. Charaeterization by automatedDNA sequencing of mutations in the gene (lpO ~)

encoding the RNA polymerase ~ subunit in rifampinMycohacterillm tllhelt'1I1osis strains from New Yorkand Texas. J Clin Microbiol 32: 1095-1098.

Lederberg l 1998. The future of infectious disease, p.514.ln RLSinghal & OP Sood (eds), DJ1Ig Resistance:Mechanism and Management, Proceedings of theFourth Annual Ranbaxy Science Foundation Symposium Communicore Publishers, New Delhi, lndia.

Miller LP, Crawford ]T, Shmnick TM 1994. The l'pO ~

gene of Mycohacterillm tuherclllosis. AntimicrohAgents Chemother 38: 805-811.

Musser JM 1995. Antimicrobial agent resistance in mycobacteria: molecular genetic insights. ClinMiClvhiol Rel' 8: 496-514.

Pande JN 1998, MuItidrug resistant tuberculosis: Currcnt concepts and future directions for management,p. 87-90. In RLSinghal & OP Sood (eds), DmgReSistance: Mechanism and Management. Procecdings of the Fourth Annual Ranbaxy Science Foundation Symposium CommunIcore Publishers, NewDelhi, India.

Ramalingaswami V 1998. Opening address. p. 1-3. InRLSinghal & OP Sood (eds), Drug Resistance:Mechanisl11 and Management. Proceedings of theFourth Annual Ranbaxy Science Foundation Symposium Communicore Publishers, New Delhi, lndia.

Rattan A, Kalia A, Ahmad N 1998. Multidrug resistanttuberculosis: Molecular perspectives. Emerg II/fecDis 4: 195-209.

TakifT HE, Salazar L, Guen'ero C, Philipp W, HuangWM, Kreiswirth B, Cole ST, Jacobs Jr WR, TeJentiA 1994. Cloning and nucleotide sequence of theMrcohacterillm tuherculosis g)'rA and grrB gencs,and characterization of quinolone resistance mutations. Alltimicroh Agents Chemother 38: 773-780.

594 MDR-M.tb in India • Noman 5iddiqi et al.

Telenti A, Imboden P, Marchesi F, Lowrie D, Cole S,Bodmer T 1993a. Detection of rifampicin-resistance mutations in M. tuberculosis. Lancet 341: 647650.

Telenti A, Imboden P, Marchesi F, Schmidheini T,Bodmer T 1993b. Direct automated detection ofrifampicin-resistant Mycobacterium tlIherculosis by

polymerase chain reaction and single-strand conformation polymorphism analysis. AntimicrohAgents Chernother 37: 2054-2058.

Williams DL, Wanguespack C, Eisenach KD, Bates JH,Crawford JT 1994. Characterization of rifampin resistance in pathogenic mycobacteria. AntirnicrobAgents Chernother 38: 2380-2386.

Mem /nst Oswaldo Cruz, Rio de Janeiro, Vol. 93(S): 595-599, Sep./Oct. 1998 595

Molecular Basis of Ribotype Variation in the SeventhPandemie Clone and its 0139 Variant of Vibrio cholerae

Ruiting Lan, Peter R Reevesj+

Department of Microbiology, University of Sydney, NSW 2006, Sidney, AustraiJa

Ribotyping has been widely lIsed ta c!wractcrise the se1'cnth pandemie clone ine/lldillg SOllth Americali alld 0139 J'ariclllts whieh appeared in 1991 and 1992 respectiveZr. To re1'eal the moleclliar ha.lis ofribOf.}pe variation wc anaZ1'sed the rrn operon,l and their /lanking regioll.l. A1/ bllt olle 1'Uriation d('tcctedby BglI, tire most discriminatOl:v en::yme, \Vas fè>lInd to he dlle ta changes lI'itlrin the rrn operolls, resnltingfi'om recomlJlllation hetwccn operons. The recombinants are deteeted beeal/se of the presence otaBglI site in tire 16S gene in three orthe nine rrn operons and/or changes o/ll1fe/gellie spaeer types orwhich/our variants were idelltified. As theji-eqllency ofrrn recombination is high, ribot1plllg hecomes aless IIsefitl tool {or emllitiona/y stlldies and long rerm monitorillg of tire patlrogenie clones of Vibriocholerae as 1'{//ïatioll could IIndergo precise re1'ersion by the same reeombinatlOn e1'ent.

Key words: ribotyping - /'l'JI reeombination - seventh pandemie - 0139 - ~'ilJlïo cho/erae

Vibrio cholerae is the agent responsible forcholera which was tirst described in 1854. However, the natural habitat of V. cholerae is the aquaticenviromnent. Environmental V. cholerae are diverse and most are nontoxigenic. There are morethan 1900 antigens identitied (Yamai et al. 1997).The best known forms are 01 and recently 0139.Both cause cholera, currently mostly in developing countries. Seven pandemies of cholera arerecognised since 1817. The seventh started in 1961and continues to the present day.

The seventh pandemie clone has been studiedby various molecular methods. Ribotyping, a fonnof restriction fragment length polymorphism analysis using l'RNA genes, was shown to be very discriminatory in revealing variation. l'RNA sequencesare highly conserved and the genes (rl'l1) are presentas multiple copies in the genome of many bacteria. The typing therefore provides infonnation onseveral flanking regions simultaneously, Therehave been quite a few studies on the epidemiologyand molecular typing of V. cholerae. Koblavi et al.(1990) were the tirst to employ ribotyping to tingerprint V. cholerae strains and Popovic et al.(1993) proposed a standardised scheme for typingV. clrolera using Bg/I restriction enzyme to allowpublic laboratories to follow the movement and

This work was supported by a grant l'rom the NationalHealth and Medical Researeh Couneil of Australia.+Corresponding author. Fax: +612 9351.4571. E-maIl:[email protected] ved 15 June 1998Accepted 30 Ju1y 1998

identify the ongll1S of V. cholerae strains.Ribotyping has since been widely uscd tocharacterise the sixth pandemie clone (Faruque etal. 1993) and the seventh pandemie clone (Karaoliset al. 1994, Faruque et al. 1995), the South American and 0139 variants (Wachsmuth et al. 1993,Popovic et al. 1995, Da1sgaard et al. 1997) andother 01 outbreaks (Coelho et al. 1995). Ribotypevariation has been found in the South Americanisolates in the four years of its spread from 1991(Dalsgaard et al. 1997) and in the 0139 variantwhich appeared in 1992 (Popovic et al. 1995,Faruque et al. 1997).

As the rrn operons are conserved, variationdetected in ribotyping has been generallyassumedto be duc to variation in flanking regions. Wc studied the changes behind ribotype variation to helpus understand the evolution of pandemie clones(Lan & Reeves 1998).

RIBOTYPE VARIATION IN THE SEVEl'iTH PANDEMICCLONE

The seventh pandemie clone is an very interesting clone to study because accurate dates ofdevelopment are known. In our ribotyping studywith a total of47 strains isolated from 1961 (Indonesia and Hong Kong) to 1993 including Africanand Asian isolates (Karaolis et al. 1994). BglI detects most of the polymorphisms. The Bg/Iribotypes are summarised in Table 1. There are IIribotypes, Ribotype G was present at the start ofthe pandemie and 14 type G strains are ail fromAsia. In the tirst 10 years of spread there was nodetectable variation. In fact after 1966 there was alull period with relatively little cholera. RibotypeH appeared in 1970 in Asia and spread to Africa;

596 Molecular Basics of Ribotype Variation in V. cho/erae • R Lan, PR Reeves

5ehematie representation ofBglI ribotype data of V/brIO cho/cracseventh pandemie iso1ates to indieate variation in specifie operons among nbotypes G, l, J, M and Q and the 0139 variantR. The operons are indieated to the left for eaeh band representing the 5' 165 gene. Alternative bands of the same operonare indieated by a dotted line for operons B, C, and G. The sizeofeaeh band IS indieated on the left. After Karaolis et al. (1994)and Lan and Reeves (1998).

tandem operon. However, we could not reconci1eour data with their conclusion of eight independent rm loci.

IDENTIFICATION OF ALTERNATIVE FRAGMENTSFOR EACH RIBOTYPE CHANGES BY OPERON SPECIFIC PROBING

Due to the presence of a BgII site at the end ofthe 23S genes, changes in the 23S proximal DNAare Ilot detected by ribotyping and ail Bg/l variation involves the 16S genes or their flanking regions. PCR walking was used to obtain sequenceimmediately upstream of the 16S genes. We thenmade operon specifie probes and probed BglI digests of different ribotypes to analyse the variation. The probing experiments are summarised inthe Fig. with schematic representation ofeach BgIIchange.

For operons B, C, and G, alternative bands wereidentified through probing. For operon B, the alternative bands are 7.2 kb, 12.3 kb and 19.1 kb.We found that these alternative bands result fromloss of the BglI site in the 16S gene.The 7.2 kbband has only 5' end of the 16S gene because ofthe Bg/l site in the 16S gene. We sequenced theregion around the BglI site in M803, M811, andM662 of operon 8. A single base difference fromC to T at base 849 determines the presence or absence of the elevenmer BgII recognition sequence.However the change to 19.1 kb in M662 from 7.2kb in M803 or 12.3 kb in M811 also involves

- --- - -- - ----- --- -. --'- --------- - -----

123

Operon A 107

OperonB"2Operon C -.... 69

Operon D _6'Operon E - ~~ =Operon H:;:" ,Operonl /"Operon F : ~

'"

R

,'=

,,,,

,,

====

iG

191

Rlbotype

"Operon G _ 2'\

Type stram

TABLE 1

Summary of Egil ribotypes detected in the seventhpandemie clone (Karaolis et al. 1994)

rrn OPERON FEATURES IN THE EARLY SEVENTHPANDEMIC ISOLATE M8D3

Majumder et al. (1996) mapped seven operonsto the genome of V cholerae strain 5698. However, ribotyping in V cholerae detects 10 or morebands in Bgll digests in the seventh pandemie isolates (Karaolis et al. 1994) and it was not clearhow to assemble seven operons from the patterns.We know that in general a rm operon has threegenes in the order 16S-23S-5S and there is a uniqueI-Cell 1 site in the 23S gene (Liu & Sanderson1995). We used 2 DNA probes, one specifie to the5' region of the 16S gene and the other specifie tothe 3' region of the 23S gene, and an oligo probefrom 16S 5' end to probe Southern blots of BgIIdigests. From the probings we concluded that thereare nine operons in the strain of M803, an earlyseventh pandemie isolate. Three operons have aBgII in their 16S gene. We also used the 23S probeto probe strain M803 DNA digested with l-Cell 1and BglI, which showed only a strong 1 kb bandrather nine bands as expected and thereby discovered that there is a Bg/l site near the end of the 23Sgene. However, the number of operons is differentfrom the seven operons detected by Majumder etal. (1996). Later we discovered that two of thenine operons are in tandem with another operon.Thus there are only seven loci on the chromosome.A recent study using pulsed-field gel electrophoresis of l-ceu 1 digest from the Majumder group(Nandi et al. 1997) found that there are nine fragments, one of which is 6 kb, the equivalent of a

G 14 1961-1991 Asian region onlyH 22 1970-1993 Asia and Africa1 1 1971 BurmaJ 2 1971,1974 ChadK 1 1972 SenegalL 1 1978 MalaysiaM 1 1988 ZaireN 2 1989, 1992 HK & Indonesia0 1 1990 Malawip 1 1991 IndiaQ 1 1993 lndonesia

Ribotype No. Year of Locationisolates isolation of isolation

Africa being free from cholera before 1970. Theresurgence in 1970 seems to be associated with asubstantial increase in genetic variation. There areother minor ribotypes with one or two iso1ates fromAsia and Africa.


change in the flanking region of the operon. Wehave yet to walk to that Bgl1 site. For operon Coperon-specifc probing showed that the 2.4 kb bandin M825 is a replacement of the 6.9 kb fragmentwhich is due to the creation of a Bg/l site in the16S gene. For operon G the alternative band forthe 2.3 kb band is similar in size to the 6.9 kb operon C fragment. Apparently this resulted in thenewly created band being masked. This change isdue to the loss of Bg/I site in the 16S gene.

We found that operons H and 1are tandem operons and absence ofa band means deletion of theoperon. We did Southern probing ofl-ceu 1digeststo show that H and 1are immediately downstreamofother rrn operons. For tandem operons I-Ceu 1digestion, which cuts uniquely in 23S gene, wouldrelease a fragment from the I-Ceu 1site of the upstream operon to the I-Ceu 1site ofthe downstreamoperon.

SIZE VARIATION IN INTERGENIC SPACERS

The bands of 4.0,4.1 and 4.3 kb in the Fig. arefragments having part of 16S gene and the wholeof23S gene and variation in size ofthese bands isdue to size differences of the intergenic spacer. Weanalysed the intergenic spacer of the three operonsB, e and G. The spacer region was amplified bynested peRo Three types of spacers were initiallyfound: small (S) 411 bp, large (L) 692 bp, andmedium (Ml) 488 bp. We later identified anotherspacer variant (M2) 587 bp.

REASSORTMENT OF SPACER TYPES AND rrn RECOMBINATION

The nature of variation detected by BglIribotyping is summarised in Table Il, includingeleven ribotypes of the seventh pandemic and its0139 variant. It is evident that there was extensivereassortment of spacers. For operon B, there arefour forms: S, BglI-; L, BglI+; L, BglI-; and M2,BglI- generated from the S, BglI+ form of the earlystrains. Both operons e and G have two additionalforms. So there are in total eight types of changesto ITI1 operons which we attribute to homologousrecombination between operons. We don't knowwhy the most frequent changes are in operon B.The other frequent change is deletion of operon 1.Deletions might be due to single or multiple events.The majority of the ribotypes seems to have arisenfrom ribotype G by a single recombination eventthough M, N and Q require two or more recombinational changes.

RIBOTYPE VARIATION IN THE 0139 VARIANT

lt is weil recognised from multilocus enzymeelectrophoresis (MLEE) and other data that the0139 Bengal strain is derived from the seventhpandemic strain. A single ribotype, R, (representedby strain M831) was identified in our previousstudy (Karaolis et al. 1994). The 0 139 clone is verysimilar to the Asia-dominant ribotype G. The onlydifference in the Bg/I digests ofribotypes Rand Gis that in ribotype R the 6.9 kb operon e fragment

TABLE Il

Spacer types and status of BglI site in the 16S RNA gene for operons B, C and Gand status of operons H and 1inthe ribotypes (G-Q) of the seventh pandemie clone and ribotype R of the 0139 variant

Operon B Operon C Operon G Operon Operon

Ribotype Strain Spacer a Bg/I site b Spacer BglI site Spacer Bg/I site H

+

+ S

G M803 S +PresentH M8û7 c

I M811 L

J M812 L

K M813L M82û L

M M825 L

N M799 M20 M826 L

P M654Q M662 M2R M831

Ml

S +

MI + Present

Del

DeldeiDel

Del"

Del DelDel

Del

a: the spacer types are small (S) 431 bp, large (L) 711 bp, medium (M 1) 5û9bp and a variant of size between MI andL (M2) 6û7bp; b: +/- indicates presence or absence ofBgll site at base 838 of the 16S RNA gene; c: status identicalto ribotype G is indicated by a dot; d: deletion of an operon.

598 Molecular Basics of Ribotypc Variation in V. cho/erae • R Lan, PR Reeves

is replaced by a 2.4 kb band due to the gain of aEgIl site in the 16S gene. It is very interesting thatPopovic et al (1995) detected two ribotypes, 3aand Sa, in 0139 isolates. Ribotype 3a is identicalto ribotype R. Ribotype Sa differs from ribotype3a by the absence ofoperon I. Popovic et al (1995)noted that ribotypes 3a and Sa are very similar totheir ribotypes 3 and 5 of seventh pandemic clonerespectively, which implied that ribotype 3a maybe derived from ribotype 3 and Sa derived from 5.Most likely ribotype Sa is derived from ribotype3a through an independent operon 1deletion ratherthan from ribotypc 5. If 3a and Sa were derivedseparately, it would require independent transferof a new 0 antigen gene cluster. This also illustrates the wcakness of ribotypmg for defining relationships between strains.

RIBOTYPE VARIATION IN THE SOUTU AMERICANVARIA]\'T

ln our ribotyping study (Karaolis et al. 1994),only one South American isolate was included whichbelongs to ribotype H. Dalsgaard et al. (1997)analysed 50 South American isolates isolated in Perufrom 1991 to 1995 and found four ribotypes: RI,R2, R3 and R4. The publication of good qualitySouthern blot by Dalsgaard et al. (1997) and use ofprobes identical to our study allow us to interprettheir results in terms ofthe basis ofribotype changes.RI is identical to our ribotype H. R2 differs fromRI by the appearance of a 4.3 kb band and disappearance of a 4.0 kb band representing the changein spacer type from small to large in operon C byrm recombination. R3 differs from RI in the presence of a 5.6kb band which is equivalent to the position of operon I and the pattern is identical to ourribotype G, the frequent ribotype isolated in Asianregion only. It seems that a RI (ribotype H pattern)strain reversed back to ribotype G pattern by recreation of the tandem operon r. presumably also dueto rrn recombination. This is an good example whereidentical ribotypes may not mean genetic similarity.It is possible that the R3 strain is actually a ribotypeG strain separate]y introduced into South Americaand R3 is not derived from RI. The early SouthAmerican isolates differ at the locus ofleucine aminopeptidase in MLEE from other seventh pandemieisolates. Therefore the allele profile could be usedto detennine whether R3 is derived from RI. R4 isvery different from the other three ribotypes. In ouropinion R4 is not developed from RI.

GDIERAL CONCLUSIONS

It has always been assumed that ribotype variation is due to changes outside the rrn operons. Thisstudy shows that ribotyping detects two types ofchanges of very different nature: within operon

changes produced by rm recombination or mutational changes outside the operons. For the seventh pandemie clone the varaition observed is predominantly rrn recombination. The frequency ofrrn recombination is high. Ten new ribotypes werefound in 46 isolates over 33 year span in our previous study (Karaolis et al. 1994). Nine of the tenribotypes were generated by rm recombination.There are new ribotypes found in the South American isolates in its first four years spread (Dalsgaardet al. 1997). Simi larly new ribotypes were foundin the 0139 variant (Popovic et al. 1995, Faruqueet al. 1997).Therefore, the level of paraI lei and reversaI changes will also be high. And similarribotype does not necessarily refiect genetic similarity. Thus ribotyping is not suitable for long tcnnmonitoring of the seventh pandemic clone or anyV. cholerae clones. Other species need to be studied to see whether this is a general phenomenon.

However, some laboratories may wish to continue to use EgIl ribotyping for typing seventh pandemic isolates as an accessary tool as it has beenwell established technically. In such case we recommend using a fragment from the 16S gene encompassing the Egil site as a probe, eg the 1 kbfragment from base 21 to base 1097 of the 16S genein this study, which will produce bands of unifonnintensity and identical patterns to probing using amixture of 16S and 235 rRNA or rDNA (Lan &Reeves 1998). Using a 16S and 235 rRNA as probethe hybridisation signal for those fragments havingonly part of the 165 gene is much weaker than forother fragments which may lead to misidentificationofpattems. The use of probes including 5S gene orany fianking sequences in addition to 16S and 235genes is not recommended as it produces ribotypingpatterns not comparable to those of other laboratories. Separation of the four bands between 6.3 kb to5.8 kb, and the three bands between 4.0 kb to 4.3kb is usually poor. Reference ribotypes representing each band variation should be included for comparison to detennine the present or absent ofa bandto increase accuracy.

REFERENCES

Coelho A. Andrade JRC, Vincente ACP, Salles CA. 1995.New variant of Vibrio cholcrac 0 1 from elinical isolates ln Amazonia. J Clhl Microhiol 33: 114-118.

Oalsgaard A, Skov MN, Seriehantalcrgs O. EehevernaP, Meza R, Taylor ON 1997. Mo1eeu1ar evolution ofVi/m'o cholcrac 01 strains isolated in LIma, Peru,from 1991 to 1995.JClinMic:robio/35: 1151-1156.

Faruque SM, Ahmed KM, Siddique AK, Zaman K, AlimARMA, Albert MJ 1997. Moleeular analysls oftoxigenie Vi/mo cholerae 0139 bengal strains isolatedin Bangladesh between 1993 and 1996: evidenee foremergcnce of a new clone of the bengal Vibrios. JClin Microhiol 35: 2299-2306.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), 5ep./Oct. 1998 599

Faruque SM, Alim ARMA, Rahman MM, Siddique AK,Sack RB, Albert MJ 1993. Clonai relationshipsamong Classical Vibrio cholerae 0 1strains isolatedbetween 1961 and 1992 in Bangladesh. J ClinMicrobiol 31: 2513-2516.

Faruque SM, Roy SK, Alim ARMA, Siddique AK,Albert MJ 1995. Molecular epidemiology of toxigenic Vibrio cholerae in Bangladesh studies by numerical analysis of rRNA gene restriction patterns.J Clin Microbiol 33: 2833-2838.

Karaolis DKR, Lan R, Reeves PR 1994. Molecular evolution of the 7th pandemic clone of Vibrio choleraeand its relationship to other pandemic and epidemicV. cholerae isolates. J Bacteriol 176: 6199-6206.

Koblavi S, Grimont F, Grimont PAO 1990. Clonai diversity of Vibrio cllOlerae 01 evidenced by rRNAgene restriction patterns. Res Microbiol 141: 645657.

Lan R, Reeves PR 1998. Recombination between rRNAoperons created most of the ribotype variation observed in the seventh pandemic clone of Vibriocholerae. Microbiology 144: 1213-1221.

Liu S-L, Sanderson KE 1995. Rearrangements in thegenome ofthe bacterium Salmonella typhi. Proc NatlAcad Sei USA 92: 1018-1022.

Majumder R, Sengupta S, Khetawat G, Bhadra R, K,Roychoudhury S, Das J 1996. Physical map of the

genome of Vibrio cholerae 569B and localization ofgenetic markers. J Bacteriol 178: 1105-1112.

Nandi S, Khetawat G, Sengupta S, Majumder R, Kar S,Bhadra RK, Roychoudhury S, Das J 1997. Rearrangements in the genome of Vibrio cholerae strainsbelonging to different serovars and biovars.int J SystBacteriol 47: 858-862.

Popovic T, Bopp CA, Olsvik 0, Wachsmuth K 1993.Epidemiologic application ofa standardized ribotypescheme for V. cholerae 01. J Clin Microbiol 31:2474-2482.

Popovic T, Fields PI, Olsvik 0, Wells JG, Evins GM,Cameron ON, Farrner III JJ, Bopp CA, WachsmuthK, Sack RB, Albert MJ, Nair GB, Shimada T, FeeleyJe 1995. Molecular subtyping of toxigenic Vibriocholerae 0139 causing epidemic cholera in Indiaand Bangladesh, 1992-1993. J infect Dis 17/: 122127.

Wachsmuth TK, Evins GM, Fields Pl, Olsvik 0, POPOVICT, Bopp CA, Wells JG, Carrillo C, Blake PA 1993.The molecular epidemiology of cholera in LatinAmerica. J infèct Dis 167: 621-626.

Yamai S, Okitsu T, Shimada T, Katsube Y 1997. Distribution ofserogroups of Vibrio cholerae non-O 1 non0139 with specific reference to their ability to produce cholera toxin, and addition ofnovel serogroups.J Japan Assoc infect Dis 7i: 1037-1045.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 601-607, 5ep./Ocl. 1998 601

The Amazonia Variant of Vibrio cho/erae: MolecularIdentification and Study of Virulence Genes

MAS Baptista, JRC Andrade*, ACP Vicente* *, CA Salles***, A Coelho/+

Dcpartamento de Genética- 1. Biologia, Universidade Federal do Rio de Janeiro, I1ha do Fundào, Cx Postal680 Il,21944-970 Rio de Janeiro, RJ, Brasi1 *Serviço de Microbiologia e Imunologla, Universidade do Estado

do Rio de JaneIro **Departamcnto de Genética, Instituto Oswaldo Cruz ***Departamenlo de Bioquimica eBiologia Molecular, Instituto Oswaldo Cruz, Rio de Janeiro. RJ. Brasi1

The pathogenic 01 Amazollia variallt of Vibrio cholerae has been shawn previous(v ta have a cytotoxinacting on cultured Vero and Y-l cells, and ta lack important vimlellce factors such as the cholera toxin(eoelho et al. 1995a). This study extends the mo/ecular allalysis ofthe Amazonia strains. detecting tht' presence ofthe toxR gene, with a very) simitar sequence to that of'the El Tor and classical biotypes. The outermembrane proteins are analyzed. detecting a variation among the group ofAmazonia straills. with threedifferent patternsfound. As a by-product ofthis work a polymerase chain reaction fragment was sequenced.reading part of the sequence of the LOIl protease ~f the Amazonia strains. This gelle was not previous(vdescribed in V. cholerae, but its sequence is present in the TIGR database specificfor this species.

Key words: Vihrio cho/erae - Amazonia - foxR - outer membrane proteins - protease - Lon

The Amazonia variant of Vibrio cholerae wasisolated from a group of clinical 01 isolates obtained from the northwest of Brazil in 1991-1992(Coelho et al. 1995a). The Latin American epidemie was caused by a strain of the El Tor biotype. Il spread out from the Pacifie coast of LatinAmerica inwards, mainly following the Amazonriver basin, reaching the northeast of Brazil andcoming down up to the Rio de Janeiro region. Acollection ofstrains from the beginning of the epidemie in Brazil was analyzed by the random amplified polymorphie DNA (RAPD) discriminativetechnique (Coelho et al. 1995b), and a surprisingresult was obtained. There was a group of strainswith a different fingerprint pattern from the epidemie El Tor strains. Ali ofthese strains presentedthe same pattern, showing that they represented adistinct group.

These 14 strains were analyzed by various techniques. Biochemically they are undistinguishablefrom other V cholerae strains. Ali ofthese strainswere Ogawa, in contrast to the majority of strainscollected at this time that were Inaba. However

This work was supported by a combination of grantsfrom the following institutions: Pronex (CNPq), CAPES,CNPq, Univer~idade Federal do Rio de Janeiro,Universidade do Estado do Rio de Janeiro, FUJB (Brazil). PAPES (Fiocruz).+Corresponding author. Fax:+55-21 -280-8043. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

other El Tor strains from the same time and areawere also Ogawa. The isozyme method (Salles &Momen 1991) was used on these strains, and theywere c1assified into a new group. The same thinghappened with ribotypes (Popovic et al. 1993), andthey formed a new group.

The Amazonia strains were tested for the presence of the ctx gene (Kaper & Levine 1981, Salleset aL 1993, Kaper et al. 1994), encoding the choiera toxin, and the result was negative. The presence of other virulence genes was tested by polymerase chain reaction (PCR), and neither the ST(thermo-stable toxin) (Ogawa et al. 1990, Vicenteet al. 1997a) nor the zot (zonula occludens) toxin(Baudry et al. 1992) were found. The tcpA gene,coding for the colonization pilus (Taylor et al. 1987,Rhine & Taylor 1994, Manning 1997, Vicente etal. 1997b), was not found by PCR or Southernhybridization.

When tested on rabbit ligated ileal loop, thestrain did not produce an accumulation of liquid,but did show a destruction of the intestinalephitelium, a heavy mucus production, with a largenumber oferythrocytes and epithelial cells embedded in il. In in vitro studies on cultured Vero cells,the production of a cytotoxin was detected, leading to morphological alterations of the cells, theirdetachment from the plastic and death.

fn this paper the analysis of virulence genes ofthe Amazonia strain is extended, mainly with thestudy of the regulatory gene toxR (Peterson &Mekalanos 1988, DiRita & Mekalanos 1991,DiRita et al. 1991). toxR is considered a main regulatory gene, responsible for the recognition of en-

602 Virulence Genes of the Amazonia V. cho/erae • MAS Baptista et al.

vironmental stimuli for expression of a number ofgenes collectively denominated the ToxR virulenceregulon (Skorupski & Taylor 1997, Champion etal. 1997). The major outer membrane protein of V.cho/erae, OmpU, is directly regulated by the ToxRprotein (Miller & Mekalanos 1988, Sperandio etal. 1995, Chakrabarti et al. 1996, Crawford et al.1998). The outer membrane proteins of theAmazonia strains are also analyzed here.

MATERIALS AND METHODS

Bacteria/ strains - Fourteen V. cho/eraeAmazonia strains were previously described(Coelho et al. 1995a). A further group ofsix strainswas obtained from Cholera Reference Center(Fiocruz, Brazil). A streptomycin resistant derivative of one of the original strains, 4010 was usedfor ail the experiments described in this paper.Control classical and El Tor strains were 0395 andE7946 respectively.

DNA preparation, PCR reaction conditions andprodllct ana/ysis - Bacteria were grown ovemightin alkaline peptone water (1 ml) and DNA was extracted (Silhavy et al. 1984). The program used forPCR consisted of35 cycles, at 94°C for 1min, 55°Cfor 1min 30 sec and noc for 1min 30 sec. The reactions included 1 III of each primer (500ng/IlI),100ng ofDNA, dNTP's 50 IlM each, 5 III reactionbuffer (1.5mM MgCI2 final concentration), 0.5 IIITaq polymerase (2.5U) (Pharmacia) and distilledwater to a total volume of 50 Ill. An Ml Researchthermocycler (Watertown, Mass.) was used for thetemperature cycling. Primers used for the toxR fragment of 560bp were: OL.I: 5' TCGGAITAGGACACAACTC and 0L.2: 5' CTGCGAGGGGAAGTAAGAC. DNA was analyzed on 1.4% agarosegels in TBE IX, prepared according to Sambrooket al. (1989), and running at 100 Volts for approximately 2 hr 30 min, until the bromophenolbluereached the end of the gel.

Southern transfer and hybridization - Southem transfers to nitrocellulose were done according to Sambrook et al. 1989. The hybridizationsolution was 50% formamide, 6X SSC, 0.7% SDS.DNA (200ng) was labeled with the Random Primerkit (Life Technologies) employing uP32 dCTP.

P/asmid preparations, DNA restriction and Iigation - Plasmid preparations employed Qiagen p100 columns according to instructions ofthe manufacturers. Restrictions were done as described bythe enzyme manufacturers (Life Technologies). Afive to one proportion of insert fragment was usedin the cloning experiments. Electroporation wasdone into the Escherichia coli strain DH5u.

Outer membrane proteins preparation andprotein ana~vsis on SDS-po/yaCl}'/amide gels - Outermembrane proteins were prepared from 1ml ofcells.

Bacteria were spun down and treated for 10 minwith 0.06M Tris HCI (pH 8.0)/ 0.2M sucrose, 0.2mMEDTA and 0.04 mglmllysozyme (total volume 500Ill). 10 III of 1 mg/ml DNAse were added, and then500 III ofTriton extraction buffer [(2%Triton X-I 00,10mM MgCI2, 50mM Tris-Hel (pH 8.0)] wereadded. Outer membrane fragments were spun downand washed with water for three times. Proteins wereresuspended in SOS-PAGE sample buffer, boiledfor 5 min, and loaded on 12% SDS-polyacrylamidegels with 5% stacking gels. Electrophoresis wascaITied out at constant CUITent (35mA), and the gelswere stained with 0.25% Coomassie blue, anddestained with methanollacetic acid.

DNA sequencing and ana~vsis - DNA sequencing was done employing the Thermo-sequenase kit(Amersham) and uP33 labeled dideoxy nucleotides.Plasmid DNA or PCR amplicons were sequenced.Specifie bands on agarose gels were eut, and theDNA purified with the Sephaglass kit (Pharmacia).300 ng ofDNA and 0.5 III ofthe ddNTPs were usedfor sequencing. Standard 60 cm 6% polyacrylamidebisacrylamide gels were used, with a glycerol tolerant buffer provided with the kit. The gels were fixedand dried and Hyperfilm was exposed for the visualization of the bands.

Databank searches with the sequences weremade against the specifie V. cho/erae TIGR database at the Institute for Genomic Research(www.ncbi.nlm.gov/cgi-bin/BLAST/nph-tigrbl)and against the non-redundant combined databasethrough the Blast Search (www.ncbi.nlm.nih.gov/cgi-bin/BLAST/nph-newblast?lform=O). Individual sequences were retrieved from the Genbank(www 2. ncbi. n1m. n ih. gOY /gen ban k/ q uery_form.html).

RESULTS

Presence o/the toxR gene - Primers OL.I and0L.2 were used in a test to check whether the toxRgene was present. Fig. lA shows the presence ofastrong band with an expected size of 560 bp, forvarious Amazonia strains. The gel was transfeITedand hybridized to a probe prepared from a classical strain, by amplification with the same primers.Hybridization was positive, as shown in Fig. 1B.

C10ning ofa fragment of the toxR gene - The560bp PCR fragment of the Amazonia toxR wascloned into the pBluescript SK vector, using Smal,and producing the plasmid pMB560R. This fragment was then transfeITed to a second plasmid,pCVD442 (Oonnenberg & Kaper 1991), a suicideplasmid that does not replicate in V. cho/erae. XbaIand Sa/I were used in the second cloning, and theplasmids obtained were denominated pCVD560R(Fig. 2). A restriction analysis of two such clones(Fig. 3A), and hybridization to the Amazonia toxR

Mem /nsl Oswa/do Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 1998 603

(Fig. 38) were done, in a structural analysis of theclones.

Partial sequencing ofthe IOxR gene - The loxRAmazonia 56übp fragment was sequenced, usingboth a universal primer for the pM856üR or internaltuxR primers. A high similarity of the sequence(98.6%) was found to that oftheclassical biotypes(Fig. 4A). A translation of this sequence shows a98.4% identity to the c1assicai and El Tor ToxRprotein (Fig. 48). An arginine (R) for threonine125(T) substitution in particular could cause a difference in the secondary structure of the protein.

A2 M 3 4 5 6 7

A 2 3 4 B 2 3 4

Fig. 1A: IOrR PCR amplification l'rom Am~70nia strains. torRprimers were OL.I and OL.2, which delimit a fmgment of560bp.Lanes 1 and 2 eontain products l'rom Ihe Escherichia enti DH5a(negative conlrol) and the classical srrain 0395, respectively.Lanes 3 through 7. PCR products l'rom the Amazonia strains. 3.4010; 4, 3729: 5, 3506; 6, 3439 and 7, L-34. M indicmes thesize marker. 1 kb ladder (Life Technologies). Fig. 1B: Soulhemhybridizalion of the gel on A with Ihe 560bp torR fragmentl'rom the 0395 classical strain.

Fig. 3A: PstJ restriction analysis of pCVD442 and two di fferenl pCVD560R clones. Lane 1, 1kb ladder (Life Technologies);lane 2, pCVD442; lane 3, pCVD560R c1.\; lane 4, pCVD560Rc1.2. Fig. 3B: hybridization of the gel on A to a 40 10 Amazonia560bp torR probe.

6 72 M 3 4 5B

Tox R 885

/

......-.

bla

Fig. 2: cloning of Ihe 560bp PCR fragment of the Amazonia strain 401 OSm' inlo the Smal sile of pBluescripl SK, producingplasmid pMB560R and further cloning, with Xbal and Sali, into pCVD442, to yield plasmids pCVD560R.

604 Virulence Genes of the Amazonia V. cho/erae • MAS Baptista et al.

A ToxR lmS

4 564

ATGTTCGGAT TAGGACACAA CTCAAAAGAG ATATCGATGA GTCATATTGG TACTAAATTC Clas.*.*******. ***--*-**- -********* ********** ******--** Amaz.

ATTCTTGCTG AAAAATTTAC CTTCGATCCC CTAAGCAATA CTCTGATTGA CAAAGAAGAT Clas.............. ...................... .... .............................. ........................................ ................ *----- ******** •• Amaz.

AGTGAAGAGA TCATTCGATT AGGCAGCAAC GAAAGCCGAA TTCTTTGGCT GCTGGCCCAA Clas.********** *********- ---*---**- *****T**** --*-***-*- *-**---*-- Amaz.

CGTCCAAACG AGGTAATTTC TCGCAATGAT TTGCATGACT TTGTTTGGCG AGAGCAAGGT Clas.*-******** ****G***** **--*-*-*- -***-*-*-* --****-*** *--**----* Amaz.

TTTGAAGTCG ATGATTCCAG CTTAACCCAA GCCATTTCGA CTCTGCGCAA AATGCTCAAA Clas.******-*-* *-*-*-*--* *---*-*--- --*-*---** ---******* ****--*--* Amaz.

GATTCGACAA AGTCCCCACA ATACGTCAAA ACGGTTCCGA AGCGCGGTTA CCAATTGATC clas.

--*--*-**- ---*--**** ---**----- ---*--*--- *A-*--**-* *-*-**-**- Amaz.

GCCCGAGTGG AAACGGTTGA AGAAGAGATG GCTCGCGAAA ACGAAGCTGC TCATGACATC Clas.

*-----**-- ***G****** *-------** -*-*----** *---*-**-* --*-*---** Amaz.

TCTCAGCCAG AATCTGTCAA TGAATACGCA GAATCAAGCA GTGTGCCTTC ATCAGCCACT Clas.*****A**G* ******---- ---------- **G******* *-----*--- -*-*-***** Amaz.

GTAGTGAACA CACCGCAGCC AGCCAATGTC GTGGCGAATA AATCGGCTCC AAACTTGGGG Clas.• **** ........ ......... * ...... ••••••• *-* -*._*_ ............. ....**** .... *** .... ** ........ *** ........ * Amaz .

AATCGACTGT TTATTCTGAT AGCGGTCTTA CTTCCCCTCG CAGTATTACT GCTCACTAAC Clas.*********C ********** ********** ********** *** Amaz.

CCAAGCCAAT CCAGCTTTAA ACCCCTAACG GTTGTCGATG GCGTAGCCGT CAATATGCCG Clas.AATAACCACC CTGATCTTTC AAATTGGCTA CCGTCAATCG AACTGTGCGT TAAAAAATACAATGAAAAAC ATACTGGTGG ACTCAAGCCG ATAGAAGTGA TTGCCACTGG TGGACAAAATAACCAGTTAA CGCTGAATTA CATTCACAGC CCTGAAGTTT CAGGGGAAAA CATAACCTTACGCATCGTTG CTAACCCTAA CGATGCCATC AAAGTGTGTG AGTAG

B MFGLGHNSKE ISMSHIGTKF lLAEKFTFDP LSNTLIDKED Classical********** ********** ********** ********** El Tor

******** ********** ********** ********** Amazonia

SEEIIRLGSN ESRILWLLAQ RPNEVISRND LHDFVWREQG Classical**** .... ***** ********** ********** ********** El Tor********** ********** ********** ********** Amazonia

FEVDDSSLTQ AISTLRKMLK DSTKSPQYVK TVPKRGYQLI Classical********** ********** ********** ********** El Tor********** ********** ********** ********** Amazonia

ARVETVEEEM ARENEAAHDI SQPESVNEYA ESSSVPSSAT Classlcal********** ***5****** ********** ********** El Tor****R***** ***5****** ********** ********** Amazonla

VVNTPQPANV VANKSAPNLG NRLFILIAVL LPLAVLLLTN Classical********** *T******** ***L****** ****** .... *** El Tor********** *A******** ***L****** Arnazonla

PSQSSFKPLT VVDGVAVNMP NNHPDLSNWL PSIELCVKKY Classlcal***T****** ********** ********** ********** El Tor

NEKHTGGLKP IEVIATGGQN NQLTLNYIHS PEVSGENITL Classical********** ********** ********** ********** El Tor

RIVANPNDAI KVCE Classical********** El Tor

Fig. 4A: DNA sequence of the lo,R fragment from the Amazonia strain 4010 compared to the sequence of the classical 5698strain. Numbers correspond to the number ofnucleotldes. * are used to mark the same nucleotide as m the previous Ime. FIg. 48:ammoacld comparison between a translation of the DNA sequence of the Amazonia strain 4010. compared to the sequence of theToxR protein of c1assical strain 5698 and El Tor stram E7946. * are used to mark the same ammoacid as in the previous line.

Mem Insl Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep.lOct. 1998 605

Fig. 5: SDS- polyacrylamide gel electrophoresis of outer membrane proteins. Lane 1: c1assical strain 0395; 2-6, Amazoniavariant strains (2, L-34; 3, 3218; 4. 3439; 5,4010; 6, 3506); 7,El Tor (607); 8 and 9, Amazonia variant strains (8, 3501; 9,4132); 10, protein molecular weight markers.

The major Ol/ter membrane pr01eins ofAmazonia strains - The outer membrane proteinsof the Amazonia strains were analyzed on polyacry lamide gels (Fig. 5). Three types of patternswere obtained. Some strains present one major protein with a size of 38kDa, as the 38kDa OmpUprote in of the El Tor and classical strains. Otherstrains present one smaller protein of 35kDa. A

2 3 4 5 6 7 8 9 10

third group presents both ofthese proteins. Different media were lIsed for growth, to check whetherthis major proteins would vary according to gro\vthconditions (data not shown), but it seems that thepattern is strain determined.

Presence of the Ion gene - In the process oflooking for virulence related genes, we obtained aspurious 800bp PCR arnplicon that was partiallysequenced. This sequence corresponds to the Lonprotcase of various bacteria. This is one of the mainproteases. responsible for the degradation ofmistolded proteins (Gottesman 1996). rn Fig. 6 wecompare the sequence of the Amazonia strain putative Lon protein to a sequence found in the TIGRdatabank for V cholerae and to the other bacteriaand archae.

DISCUSSION

The pathogenic 01 Amazonia Vibrio choleraemay be considered a new human pathogen. It wasdiscovered with the use of discriminative molecular techniques (Coelho et al. 1995b, 1997). Thisstrain belongs to the same species as the epidemicstrains but seems to cause diarrhea by a differentmechanism. The Jack of the cholera toxin, and thepresence of a cytotoxin leading to intestinal tissuedamage in rabbits strongly suggest a different roll tefor the disease.

The loxR gene was found in the Amazoniastrains, and its sequence is highly homologous tothat of the El Tor and classical biotypes. A toxRhomologous gene has been described in Vibrioparahaemolyticl/s, V/brio fischeri and a Pholobacterium sp., with much lower DNA homology values, ail below 70%.

The presence of the regulatory gene toxR, inthe absence ofvarious genes that it nonnally regulates, leads to the question of its further role. Onepossibility is the regulation of the outer membrane

E. coli EQQIPFSASLTFEQSYSEVDGDSASMAELCALISALADVPVNQSIAlTGSVDQFGRAQPV........................................... .... .. .... ..

V. cholerae

V.chol. Amazonia

Hae. inf.

Arch.fulgidus

TARIPLTTTITFEQSYGGVDGDSASMAEFCAIVSAFSKQPNRQDIAITGSMNHFGESQAl........................................ .. .......................................... .. ..TTITFEQSYGGVDGDSASMA?LCAIVSAFSKQPNRQDIAITGRLR?K

.............. .PSQLPFSASLVFEQSYGEIDGDSASLAIFCVLVSALADLPLPQHIAITGSIDQFGLVHSV

DISNMDVHIQFVGTYEGVEGDSASISIATAVISAIEGIPVDQSVAMTGSLSVKGEVLPV............. . .... . .. . . . .Meth. therruo. TDISNYDIHIQFLQAYDGVEGDSASVSVATAVISALEEIPVDQSVALTGSLSIRGDVLPV

Fig. 6: aminoacid sequence comparison of the Amazonia putative Lon protease and the corresponding protein in otller bac(eria andarchae. Bacteria: Escherichia coli, Vibrio cho/erae. Haemophillls in(lllellzae. Archae: Archaeoglobllsfulgidus and Merhallobacrerill177rhermoallxorrophicum.

606 Virulence Genes of the Amazonla V. cho/erae • MAS Baptista et al.

proteins. It is known that toxR directly regu)atesthe OmpU protein, which has been proposed as anadhesin with a role in virulence (Sperandio et al.1995, 1996). The Amazonia strain, as in the caseof El Tor and c1assical strains, presents major outermembrane proteins, which could have functionalhomologies to OmpU. Further developments in thiswork wiIl include studies on the presence of otherregulatory genes such as genes of the tcp c1uster,Iike the tcpP and tcpH (Carroll et al. 1997, Manning 1997, Hase & Mekalanos 1998) and otherputative virulence genes (Karaolis et al. 1998,Tacket et al. 1998). It will also aim at the construction of toxR mutants to evaluate the role of thisgene in the pathogenicity of the Amazonia strain,including an analysis of its adhesion to cells, cytotoxicity, and behavior in rabbit ileal loops. Thesecomparative studies may shed light not only on thevirulence mechanisms involved in the Amazoniastrain but also on V. cholerae itself.

ACKNOWLEDGMENTS

To the Cholera Rcference Centcr (Fiocruz) that provided a second group of six strains that we typcd as belonging to the Amazoma group

REFERENCES

Baudry B, Fasano A, Ketley J. Kapcr .1 B 1992. Cloningof a gene (=01) encoding a new tOXll1 produced byVi/m'o cho/crae. In{ect Imll1un M): 42R-434.

Carroll PA, Tashima KT, Rogers MB, DiRita VJ,Calderwood SB 1997. Phase variation in tcpH modulatcs expression of the ToxR regulon in Vihriocho/erae. Mo/ Mierohio/25: 1009-1011.

Chakrabarti SR, Chaudhuri K, Sem K, Das J 1996. Porinsof Vi/,rio cho/erae: purification and characterizationof OmpU. J Bacteno/ 178: 524-530.

Champion GA, Neely MN, Brennan MA, Di Rita VJ1997. A branch In the ToxR regulatory cascade ofVihrio cho/erae revealed by charactcrizatlon oftoxTmutant strains. Mo/ MiclVbio/ 23: 323-331.

Coelho A, Andrade JRC, Baptista MAS. Vicente ACP,Fcrraz LR, Salles CA 1997. Genomic fingerprintsof Vihrio cho/erae using arbitrary primer PCR areuseful epidemiological tools. p. 203-211. In GTKeusch. M Kawakami (eds), Cytokines. Choleraand the Gut. lOS Press. Netherlands.

Coelho A, Andrade JRC, Vicente ACP, Salles CA 1995a.A new variant of Vibrio cho/erae 01 l'rom clinicalisolates in Amazonia. J Clin MiclVhio/33: 114II R.

Coelho A, Vicente ACP, Baptista MAS, Momen H, dosSantos FRW, Salles CA 1995b. The distinction ofpathogenic r'ihrio cho/erae groups using arbitrarilyprimed PCR fingerprints. Res Microhio/ 146: 671683.

Crawford lA, Kaper JB, DiRita VJ 1998. Analysis ofToxR-dependent transcription activation of ompU,the gene encoding a major envelope protein in Vibriocho/crae. Mo/ Microhio/ ln press.

DiRita Vl, Mckalanos 11 1991. Periplasmic interactionbetween two membrane regulatory proteIns, ToxRand ToxS, results in signal transduction and transcriptional activation. Ce1/64: 29-37.

DiRita V, Parsot C, lander G, Mekalanos JJ 1991. Regulatory cascade controis virulence In v'i/mo cho/erae.Proc Nat/ Acad SCI USA liS: 5403-5407.

Donnenbcrg MS, Kaper 18 1991. Construction ofan caedcletion mutant ofenteropathogenic Eschcrichia coliby using a positive- selection suicide vector. InfectImmun 59: 4310-4317.

Goltesman S 1996. Proteases and thelr targets in Eschericliia coli. Ann Rcv Gcnct 30: 465-506.

Hase CC, Mekalanos JJ 1998. TcpP is a positive regulator ofvirulence gene cxpression in Vihrio cllOlcrac.Proc Nat/ Acad Sei USA 95: 730-734.

Kaper JB, Fasano A, Trucksis M 1994. Toxins of Vihriocho/erae, p. 145-176. ln IK Wachsmuth. PA Blake.o Olsvik (eds), VihrlO cho/crae and clio/era. 1170

/ecu/ar to g/oha/ perspcctil'es, ASM Press, Washington.

Kaper JB, Lcvine MM 1981. Cloned cholera cnterotoxingenes in study and prevention of cholera. Lancet il:1162-1163.

Karaolis DK, Johnson JA, BaIley CC, Boedeker EC,Kaper JB, Rceves PR 1998. A Vlhno cho/crac pathogel11city island associated with epidemic and pandemic strains. Proc Nat/ Acad Sei USA 95: 31343139.

Manning PA 1997. The tcp gene clustcr of r'ihriocho/crac GCIIC 192: 63-70.

Miller VL, Mekalanos 11 1988. A novel suicide vectorand its use in construction on insertion mutations:osmoregulation ofouter membrane protellls and virulence determinants in Vihrio cho/crae requircs 10xRJ Bactcrio/ 170: 2575-2583.

Ogawa A, Kato J-l, Watanabe H, Nair BG, Takeda T1990. Cloning and nuclcotide sequence of a heatstable enterotoxin gene l'rom ~1'hrio cho/crac non01 isolated l'rom a patient \Vith trave]er's diarrhea.Infcct 11171171111 5il: 3325-3329.

Peterson KM, Mekalanos JJ 1988. Characterization ofthe Vibrio cho/crac ToxR regulon: identification ofnovel genes involved in intestinal colonization. In(cct Immuil 56: 2822-2829.

Popovic T, Bopp CA, Olsvik 0, Wachsmuth K 1993.Epldemiologic application ofa standardized ribotypescheme for Vihrio cho/crac. J C/;'1 Microhiol 31:2474-2482.

Rhine lA, Taylor RK 1994. TcpA pIlin sequences andcolonization requircments for 01 and 0139 Vihriocho/crae. Mo/ MiclVhio/ 13: 1013-1020.

Salles CA, Momen H 1991. Identification of Vihriocho/crae by enzymc electrophoresis. Ti'ans R Socn'Op Meil Hyg il5: 544-547.

Salles CA, Momen H, Vicentc ACP, Coelho A 1993.Vihrio clIO/crac in South America: polymerase chainreaction and zymovar analysis. Ti'ans R Soc Ti'OpMed H.l'g il7: 272.

Sambrook J, Fritsch EF, Maniatis T 1989. Mo/ecu/al'Clonillg. a Lahoratol)' Manua/. 2nd cd. Cold SpringHarbor Lab Press, New York.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 1998 &07

Silhavy TJ, Berman M L. Enquist LW 1984. DNA extraction from bacterial ceIls, p. 137-139. In E.\periments with Gene Fusions, Cold Spring Harbor LabPress, New York, 303 pp.

Skorupski, K, Taylor RK 1997. Control of the ToxRvirulence regulon in Vihrio cho/erae by environmental stimuli. Mo/ Microhio/25: 1003-1009.

Spcrandio V, Bailey C, Giron JA, DiRita VJ, SilveiraWD, Vettore AL, Kaper JB 1996. Cloning and characterization of the gene encoding the OmpU outermembrane protein of Vibrio c11O/erae. Infect Immun64: 5406-5409.

Sperandio V, Giron JA, Silveira WD, Kaper JB 1995.The OmpU outer membrane protein, a potential adherence factor of Vihrio cho/erae. Infeet Immun 63:4433-4438.

Tacket CO, Taylor RK, Losonsky G, Lim Y, Nataro JP,

Kaper JB, Levme MM 199~L Investigation of theroles of toxin- coregulated pili and manose-sensitive hemagglutinin pili in the pathogenesis of Vihriocho/erae 0139 intèction. Infect Immun 66: 692-695.

Taylor RK, Miller VL, Fur10ng DB, Mekalanos JJ 19R7.Use of phoA gene fusions to identify a pilus colontzation factor coordinatcly regulated wlth choleratoxin. Prae Nat/ Acad Sei USA 84: 2833-2837.

Vicente ACP, Coelho A, Salles CA 1997a. DetectIon ofVihrio eho/erae and Vibno nllmie/ls heatstable tOXIngene sequence by PCR. J Med Microbio/ 47: 39R402.

Vicente ACP, Novaes RC, Coelho A, Salles CA 1997b.Molecular characterization of TCP cluster in non01, non-0139 Vibrio cho/erae, p. 243-24R. In GTKeusch, M Kawakami (eds), Cytokll1es, Cholera andthe Gut. lOS Press, Netherlands.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5):609-614, Sep'/Oct. 1998 609

Molecular Epidenliology and Emergence of Rift Valley Fever

AA Sall/+, PMA Zanotto*, PVialat* *, OK Sène, M Bouloy* *

Institut Pasteur de Dakar, BP 220, Dakar, Senegal *Oepartamento de MieroblOlogl3. lnstituto de ClênclasBiomédie3s, Umversidade de S:io Paulo, Cidade Universitaria, 05508-900 Sào Paulo. SP. 8rasIi

**Institut Pasteur, 25, Rue du docteur Roux, 75724 Paris cedex 15, France

Rift Valley {ever (RVF) is a mosquito-borne viral disease l\'hich mani{ested ilselfduring recell! epidemics and revea/ed its significant potential o{emergence. Studies on mo/ecu/ar epidemi%gr undertaken to better understand the fàctors leading ta RVF emergence, have co/!firmed the mode o{circu/ation ofthe virus and highlithted probable risks and obstacles {or prevention and control. As/br sel'era/other vira/ agents, mo/ecu/ar epidemi%gy is becoming a IIse{u/ too/ in the study o{the emergence ofRVF as a serious infecliolls disease.

Key words: arboviruses - Rift Valley fever - molecular epidemiology - emergence - phylogeny

Rift Valley fever (RVF) is an arboviral diseasetransmitted by mosquitoes in Africa. RVF affectsprimarily ruminants causing high mortality in offspring and abortions in pregnants females and occasionally humans, whose infection leads to a clinical picture which ranges from a mild febrile caseto hemorragic fever with complications such ashepatitis, encephalitis and retinitis (Laughlin et al.1979). In 1977, a severe outbreak of RVF occurredin human and livestock populations of Egypt(Meegan 1981). Although RVF was known formore than 40 years at that time, the extensive morbidity and mortality observed in humans appearedas a novelty in the history of this disease, therefore, emphasizing RVF as a serious emerging threatfor humans and animaIs health. Futher large scaleepidemics in Mauritania (Digoutte & Peters 1989),Madagascar (Morvan et al. 1991, 1992a, b), Egypt(Arthur et al. 1993) and very recently in eastemAfrica (Anonymous 1998) confinned the majorimpact of RVF on public health through its continuing emergence. Thus, RVF constitutes an excellent model to overview factors involved in arboviruses emergence because most ofthe conceptsrelative to emerging diseases may be illustratedalong its natural history.

Control of RVF implies the better identification of factors involved in its emergence and itsmaintenance in nature. Tt is also necessary to un-

+Corresponding author. Fax: +221.839.92.10. E-mail:[email protected] ved 15 June 1998Accepted 30 July 1998

derstand the rules and modalities ofcirculation andevolution of RVF virus (RVFV) in Africa. Theselatter objectives have been addressed by studyingthe variability among RVFV isolates by serological (Besselar et al. 1991) and molecular methods(Battles & Dalrymple 1988, Sali et al. 1997a, b).This paper aims the discussion of sorne of the aspects and contributions of molecular epidemiology towards the elucidation of RVF emergencc.

BACKGROUND

Discovery of RVFV and recent major epidemic!epizootics

RVFV was first isolated in 1930 near lakeNaivasha in Kenya by Daubney et al. (1931). Sincethen, the virus has been shown to be widespread insubsaharian Africa and in Egypt (Meegan & Bailey1989). Major epidemic!epizootics occurred inEgypt in 1977 (200,000 humans infcctions and 600deaths) and in 1993, Mauritania in 1987 (200 human deaths), Madagascar in 1991 and in easternAfrica (89,000 infections and more than 500 deathsreported so far) with the last recent outbreak in1997-1998 in Kenya, Tanzania, Soma1ia.

The etiological agent of RVF

RVFV is a member of Bunyaviridae family,Phlebovirus genus (Murphy et al. 1995). Its genome consists in three negative single strandedRNA segments referred as L, M and S respectivelyfor large, medium and small. The L segment codesfor the L protein which is the viral polymerase.The M segment codes for glycoproteins G1 andG2 and two others proteins of78 and 14 K. The Ssegment codes for the nucleoprotein N and the nonstructural NSs protein using an ambisense strategy (Bouloy 1991, Elliott et al. 1991, Giorgi 1996,Schmaljohn 1996).

610 Emergence of RVFV • M Sali et al.

Epidemiology of RVF

The epidemiology ofRVF eonsists in both epizootie and interepizootie eycles (Meegan & Bailey1989). Epizooties of RVF in Afriea oeeured oftenwhen unusually heavy rainfall were observed.During an epizootie, virus eirculates among infected arthropod vectors and mammalian hosts,particularly cattle and sheep, which represent themost signifieant livestock amplifiers of RVFY. Theinter-epizootic survival of RVFV is believed todepend on transovarial transmission of virus infloodwater Aedes mosquitoes (Linthicum et al.1985). Virus can persist in mosquitoes eggs untilthe next period of heavy rainfall when they hatchand yield RVFV infected mosquitoes. Dependingon factors such as availability of sufficient numbers of competent mosquito vectors, presence ofsusceptible vertebrates, appropriate environmental conditions, infected mosquitoes have the potential to infect a relatively small number of vertebrate hosts or to initiate a widespread RVF epizootic.

Control and prevention of RVF

Vaccines have been the principal mean used tocontrol RVE Two types of vaccines have been described for use against RVF: inactivated and liveattenuated.

Formalin-inactivated RVF vaccines have beenused to immunize animais, laboratory workers,veterinarians and other people at high risk of exposure to RVFY. The cost of the vaccine, the requirement for multiple inoculations and the timeinterval required to mount a protective immuneresponse, alliimit its use for veterinary purposes.

Two live attenuated vaccines, the Smithbumvaccine, also referred as Smithbum neurotropicstrain or SNS (Smithbum et al. 1949), and MP 12(Caplen et al. 1985) have been developped. TheSmithbum strain is the only widely available veterinary vaccine but has serious limitations in practical use, because it has been proven to be teratogenic, cause abortions and encephalitis in younglambs.

Possessing attenuation markers in ail three segments, MP 12 has a very 10w probability ofreversion (Saluzzo & Smith 1990, Vialat et al. 1997)and has been inoculated into more than 100 peopleand shown to be safe and immunogenic (Peters1997). MP 12 was also promising in laboratory triais in domestic animais (Morrill et al. 1987, Morril& Mc Clain 1996), but vaccination of pregnantewes revealed that the virus caused teratogenieeffect if inoculated during the first trimester ofpregnancy (Erasmus and Bishop, pers. commun.). Another attenuated virus, clone 13, a naturally attenu-

ated strain, is very promising regarding the preliminary results obtained in terms of immunogenicity and safety (Muller et al. 1995).

MOLECULAR EPIDEMIOLOGY OF RVFV

Investigation on the variation among RVFVisolates using serological tests based on the antigenicity of structural proteins (Saluzzo et al.1989a,b, Besselar et al. 1991) or genetic methodssuch as T I-oligonucleotide fingerprints (Peters &Linticum 1994) and, more recently sequencing,(Battles & Dalrymple 1988) indicated only minorvariations among RVFV natural isolates. To further analyze the genetic diversity ofRVFV (Sali etal. 1997b), we selected a panel of 18 strains (Table1) isolated over sorne 50 years from various hostsand geographical origins and we sequenced directlytheir NSs coding region on the S segment after astep of reverse transcription-polymerase chain reaction amplification (RT/PCR). A 50% majorityrule consensus tree derived from the sequences anaIyzed are presented in Fig. 1. The NSs coding region sequences clustered in three major lineagessupported by high bootstrapping values and byusing different phylogenetie inference procedures(e.g., maximum likelihood, parsimony and di~tanc.e

methods) and correlated with the geographlc ongin of the isolates and are referred as West Africa,East-Central Africa and Egypt. While the WestAfrica group was homogenous with strains fromMauritania, Senegal, Guinea and Burkina Faso, theEast-Central Africa and Egypt ones appeared tobe heterogenous.

As expected, the Egyptian group containsstrains isolated in 1977 and 1993 epidemics, whiehappear in the phylogeny as sister groups suggesting that either the virus remained endemic betweenthe two outbreaks or have been reintroduced in1993 from the same source (probably Sudan) as in1977. To explain the reemergence ofRVF in Egyptafter years ofsilence despite intensive surveillance,Peters (1997) proposed that the virus was reintroduced through an incompletely inactivated RVFveterinary vaccine. Furthermore, Ar MAD 79,which is the tirst isolation of RVFV in Madagascar, clustered in Egypt group and then confinneddata obtained by Morvan et al. (1991) who anaIyzed the antigenic properties of the N protein.

Secondly, the East-Central African group clustered isolates from Uganda, Central African Republic, Mauritania and Senegal. The presence ofAn MAD 91 in that group suggested that this latterstrain was probably introduced in Madagascar fromthe eastem coast of Africa. This latter assumptionalso implies that there was at least two introductions of the virus in Madagascar but also severallineages coexist in East Africa. Moreover, in

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Ocl. 1998 611

TABLE 1

Characteristics of the Rift Valley fever virus isolates analyzed by sequencing

Code Strain Year of isolation Q!igin Source-----------------

SNS Smithbum 1944 Uganda Entebbe strain

Ar UG 55 Lunyo 1955 Uganda MosquitoArRCA 69 Ar B 1976 1969 CAR MosquitoH EGY 77 ZH 548 1977 Egypt HumanMP 12" MPI2 1985 Egypt ZH548 strainAr MAD 79 ArMg RII 1979 Madagascar MosquitoAr SEN 84 Ar 0 38661 1984 Senegal MosquitoAn GUI 84 An K 6087 1984 Guinea BatAr BUF 84 Ar 0 38457 1984 Burkina Faso MosquitoHI MAU 87 H047502 1987 Mauritania HumanH2 MAU 87 H 047311 1987 Mauritania HumanH3 MAU 87 H047408 1987 Mauritania HumanH4 MAU 87 H048255 1987 Mauritania HumanAn MAD 91 An Mg 990 1991 Madagascar BovineAr SEN 93 Ar 0104769 1993 Senegal MosquitoAn SEN 93 An 0106417 1993 SenegaJ ZebuB EGY 93 B EGY 93 1993 Egypt BuffaloH EGY 93 H EGY 93 1993 Egypt Human

a: laboratory-attcnuated strain derived from a wild strain; SNS: Smithbum neurotropic strain; H: human; Ar:arthropode; An: animal; B: buffalo.

~~ L SNS

An MAD 91

Ar UGA 55

100

100

HI MAU 87

Ar SEN 93

Ar CAR 69

Ar BIIF 84

Easl-CentralAfriea

100100

6.92

.792

B EGY 93H EGY93

Ar MAD 79

9.100

Hl MAU 87

H4 MALI 87

HJMAl! 87

An SEN 93

AnGlII84

0.01

WestAfrica

Egypt

Fig. 1: phylogcnetic tree for the NSs gene of several Rift Valley fever virus isolates. Values above branches indicate the level (%)of bootstrap support usmg maxmlUm parslmony after 500 iterations. Values below branches indicate the number of times a givennode was observed on a maJority rule consensus of 50 trees with eqUivalent likelihood (LnL). Branch lengths are shown proportional to the number of substitutions per 100 residues. The rooting shown here was detennined by the inclusion of the SSF NSssequence.

TABLE II

Summary of Rift Valley fever (RVF) emergencefactors described and analyzed by Wilson (1994)

Factors of emergence Examples relative to RVF

Economic developmentl Dams and irrigation,Land use pasturage improvement

and agricultural pratices change, the outbreak inMadagascar in 1991 has been shown to be a veryinstructive example (Peters & Linthicum 1994, Peters 1997).

Although these two key factors were clearlyidentified and characterized, data derived frommolecular epidemiology are needed for a comprehensive view of RVF and its emergence process.Ourwork, although still preliminary allowed to illustrate the contribution of molecular epidemiology for, (i) the understanding oftwo modes of circulation of viral strains and (ii) delineation of genetic aspects of the virus, which may turn out tobecome potential obstacles for the prevention andcontrol of the disease.

Modes of circulation of RVFV

Regarding the molecular epidemiology dataabout RVFV, two modes of circulation may be illustrated (Fig. 2): (i) distant spread from one region to another and (ii) local circulation in an enzootic/endemic area.

Distant spread was illustrated by introductionof RVFV in Egypt (1977) and Madagascar (1979and 1991) probably from eastern or central Africa(see molecular epidemiology). It is interesting toemphasize that in both cases, possibly an antigenically and phylogenetically "new" virus was introduced in an area exempt of RVFV, raising the issue about the raie of herd immunity for both, humans and animais populations, as a factor ofemergence of the virus.

Concerning local circulation in an enzootic/endemic area, Senegal is an instructive example

612 Emergence of RVFV • AA Sali et al.

Uganda, despite II years separating the isolationsof Ar UGA 55 and Entebbe strain, the parentalstrain of Smithbum vaccinal strain, RVFV did notshow much genetic diversity, suggesting a maintenance mechanism through an endemic/enzooticcycle, possibly involving comparatively little virai activity, since increased genetic diversity for agiven mutation rate entails an increase on the effective viral population size. Surprisingly, HIMAU 87 and Ar SEN 93 belonged to the EastCentral Africa group.

The West Africa group appeared to be homogenous and suggested circulation ofsimilar variantsin Senegal, Mauritania, Guinea and Burkina Faso.lt is noteworthy that H2, 3, and 4 MAU 87 clustered near each other and were isolated from fatalcases whereas HI MAU 87 which was isolatedfrom a febrile case clustered together with Ar SEN93 unexpectedly in the East-Central Africa lineage.Moreover, one may deduced, from the strains distribution on Fig. 1, that there are two areas of circulation of RVFV in Senegal: (i) the NorthernSahelian zone where Ar SEN 93 and HI MAU 87were isolated and, (ii) the Sudano-Guinean zonewhere An SEN 93 was isolated.

Groupings ofAr SEN 84 with Egyptian strainson one hand and HIMAU 87 and Ar SEN 93 witheastern and central African strains on the otherhand, were quite unexpected and led us to hypothesize genetic exchange through reassortment toexplain these puzzling clusterings. In order to checkthis hypothesis further sequencing and phylogenelic analysis were undertaken both on Land Msegments. Although, this hypothesis is still underinvestigation, one may obviously speculate by anticipation that such a mechanism in nafura wouldhave important implications on epidemiology andemergence of RVF in Africa (see below).

EMERGENCE OF RVFV

Various factors contributing to the emergenceofinfectious diseases were classified by Lederberget al. (1992) and analyzed from the point of viewofRVF by Wilson (1994) and summarized in TableII. Emergence of RVF was also discussed by Peters (1997) with special reference to Madagascar,distant spread of the virus to Egypt and historicalspeculations. These two papers emphasized themultifactorial aspect of RVF emergence and thecentral raie ofwater and ecological change as factors triggering epidemics. Water is usually involvedeither through dams or irrigation for the sake ofagriculture development, as illustrated by Egyptin 1977 and Mauritania in 1987 or, under excessive rainfall and floodings as observed during the1997-98 outbreak in eastern Africa. Concerningthe impact of ecological changes as deforestation

Human demography andbehavior

International travel andcommerce

Biological adaptationand changc

Climate events

Living with domesticungulates, slaughter of sickanimal, Vaccination ofhealthy animais

Domestic ungulates export,human travel andmigration

Increased viral virulence,improved vectorcompetence, greateranimal susceptibility

Excessive rainfall


Reintroduction in Madagascarin 1990 probably fromeast Afriea

Distant spread

Er)7(\(\tic/endemiccycle

Fig. 2: possible modes of circulation of Rift Valley fever.

because it showed reemergence of the virus froma pool of existing enzootic/endemic strains undera similar process although the ecological contextof transmission is different between the north andthe south of that country. As far as RVF is concemed, Senegal can be divided in two areas (SaIlet al. 1997b) from which the virus have cmergedas demonstrated by isolations in 1993 (Zeller et al.1997): (i) the Sahelian zone, where southernMauritanian and northem Senegalese strains arecircu1ating and, (ii) the Sudano-Guinean zonewhere southem Senegalese strains are in contactwith those from bordering countries.

Prevention and control

In the field ofprevention and control of RVFV,molecular epidemiology studies highlighted a potential major obstacle to the use oflive attenuatedvaccines. Indeed, the possibility of the existenceofreassortment in nature raised by the unexpectedgroupings (Ar SEN 84, Hl MAU 87 and Ar SEN93) would emphasize the risk of generating uncontrolled chimeric viruses.

CONCLUSION

Although molecular epidemiology has beenshown to he informative for a better understanding on different facets ofRVFV emergence, manyquestions such as those relative to the sylvatic cycleofthe virus for instance remain unanswered. Mean-

while, surveillance of RVF and awareness shouldbe improved and reinforced since it is so far theonly conceivable way to prevent RVFV emergencewith its toll ofdeaths, sickness and economic loss.

ACKNOWLEDGEMENTS

To C Mathiot and J Thonnon for critically readingthe manuscript.

REFERENCES

Anonymous 1998. An outbreak of Rift Valley fever. Eastern Africa, 1997-1998. WER 73: 105-112.

Arthur R, EI-Sharkawy MS, Cope SE, Botros BA, OunS, Morrill JC, Shope RE, Hibbs RG, Darwish MA,Imam IZE 1993. Recurrence of Rift Valley fevcr inEgypt. Lancel 342: 1149-50.

Battles JK, Dalrymple JM 1988. Genetic variation amonggeographic isolates of rift valley fever virus. Am JTrop Med Hyg 39: 617-631.

Besselar TG, Blackburn NK, Meenehan GM 1991. Antigenic analysis of Rift Valley fever isolates: monoclonal antibodies distinguish between wild-type andneurotropic virus strains. Res Viro/142: 469-474.

Bouloy M 1991. Bunyaviridae: Genome organizationand replication strategies. Adv Vi,. Res 40: 235-266.

Caplen H, Peters CJ, Bishop DHL 1985. Mutagen-directed attenuation of Rift Valley fever virus as amethod for vaccine development. J Gen Viro/ 66:2271-2277.

Daubney R. Hudson JR, Garnham PC 1931. Enzootiehepatitis or Rift Valley fever. An undescribed virusdisease of sheep, cattle and man from East Africa. JPalho/ Baclerio/ 34: 545-79.

614 Emergence of RVFV • M Sali et al.

Digoutte JP, Peters CJ 1989. General aspects of the 1987Rift Valley fever epidemic in Mauritania. Res Vîrol140: 27-30.

Elliott RM, Schmaljohn CS, Collett MS 1991.Bunyaviridae genome structure and gene expression.Curr Top Mîcro Immunol, Springer- Verlag, Berlin:91-142.

Giorgi C J996. Molecular biology of phleboviruses, p.105-128. In RM Eliott, The Bunyavîrîdae, PlenumPress, New York.

Laughlin LW, Meegan JM, Strausbaugh LW, MorensDM, Watten RH 1979. Epidemie Rift Valley feverin Egypt: observations of the spectrum of humanillness. Trans R Soc Trop Med Hyg 73: 630-633.

Lederberg J, Shope RE, Oaks SC 1992. Emergîng Injections, Mîcrohial Threats to Health ill the UnîtedStates, National Academy Press, Washington, DC.

Lmthicum KJ, Davies FG, Kairo A 1985. Rift Valleyfcver virus (family Bunyaviridae, genus Phlehovirus). Isolations from Diptera collected during aninter-epizootic pcriod in Kenya. J /~\'g Camh 95:197-209.

Meegan JM 1981. Rift Valley fever in Egypt. An overview of epizootics in 1977 and 1978, ContrîbEpîdemîol Bîostat 3: 100-113.

Meegan JM, Bailey CJ 1989. Rift Valley fever, p. 5176. In TP Monath, The Arboviruses: Epidemiologyand Ecolog)' IV, CRC Press Inc., Boca Raton, FI.

Morvan J, Saluzzo JF, Fontenille D, Rollin PE,Coulanges P 1991. Rift Valley fever in the east coastof Madagascar. Res Viral /42: 475-482.

Morvan J. Rollin PE, Laventure S, Rakotoarivony l,Roux J 1992a, Rift Valley fever epizootic in thecentral highlands ofMadagascar. Res Vi'rol /43: 407415.

Morvan J, Rollin PE, Roux J 1992b. La fièvre de la Valléedu Rift à Madagascar en 1991. Enquête séroepidémiologique chez les bovins. Rel' Ele Med PayTrap 45: 121-127.

Morrill Je, Mc Clain DJ 1996. Epidemiology and pathogenesis ofRift Valley fever and other phleboviruses.p. 281-294 ln RM Ell iott, The bUllyavîridae, Plenum Press, New York.

Morrill JC, Jennings GB, Caplen H, Turell MJ, JohnsonAJ, Peters CJ 1987. Pathogenicity and immunogenicity of a mutagen-attenuated Rift Valley fever virus immunogen in pregnant ewes. Am J Vet Rel> 48:1042-1047.

Muller R, Saluzzo JF, Lopez N, Dreier T, Turell M, SmithJ, Bouloy M 1995. Characterization of clone 13, anaturally attenuated avirulent isolate of Rift Valleyfever virus, which is altercd in the small segment.Am J Trop Med Hvg 53: 405-411.

Murphy FA, Fauquet CM, Bishop DHL, Ghabrial SA,Jarvis AW, Martelli GP, Mayo MA, Summers MD

1995. Family Bunyaviridae, p. 300-315. ln VîrusTaxonomy. Class(fication and Nomenclature of Viruses. Sixth report of the international committeeon taxonomy ofviruses. Springer-Verlag, Wien, NewYork.

Peters Cl 1997. Emergence of Rift Valley fever, p. 253264. ln JF Saluzzo, B Dodet (eds), Factors in theEmergence ofArbovîl1ls Dîseases, Elsevier, Paris.

Peters CJ, Linthicum KJ 1994. Rift Valley fever, p. 125138. In Beran GW, Steele JH (eds), Hanhhook ofZoonoses, SectÎon B: Viral, 2nd ed., CRC Press, BocaRaton, FI.

Sail AA, Zeller H, Zanotto PM de A, Billecocq A, VialatP, Préhaud C, Yadani F, Bouloy M 1997a. Rift Valley Fever virus: the NSs protein as a target to evaluate genetic variability and its roles in pathogenesis,p. 265-271. In JF Saluzzo, B Dodet (eds), FactorsofEmergence oIA,-hovîrllse,I, Elsevier, Paris.

Sali AA, Zeller H. Thiongane Y, Digoutte JP, Bouloy M1997b. Genetic variability ofthe NSs protein amongRift Valley Fever Virus isolates. J Gen Virol 78:2853-2858.

Saluzzo JF, Smith JF 1990. Use of reassortants virusesto map attenuating and temperature sensitive mutations of the Rift Valley fever virus MP 12 vaccine.Vaccine 8: 369-75.

Saluzzo JF, Anderson GW, Smith JF, Fontenille D,Coulanges P 1989a. Biological and antigenic relationship between Rift Valley fever virus strains isolated in Egypt and Madagascar. Trans R Soc TropMed Hyg 83: 701.

Saluzzo JE Anderson GW, Hodgson LA, Digoutte JP,Smith JF 1n9b. Antigenic and biological properties of Rift Valley fever isolated during the 1987Mauritanian epidemic. Res Vîral 140: 155-64.

Schmaljohn CS 1996. Bunyaviridae: the viruses and theirreplication, p. 1447-1471. ln Fields BN, Knipe DM,Howley PM (eds), Fields Viralogy. 3rd ed., RavenPress, New York.

Smithburn KC 1949. Rift Valley fever: the neurotropicadaptation of the virus and the experimental use ofthis modified virus as a vaccine. BrîtJ Exp Path 30:1-16.

Vialat P, Muller R, Vu TH, Prehaud C, Bouloy M 1997.Mapping of the mutations present in the genome ofRift Valley fever virus attenuated MP 12 strain andtheir putative role in attenuation. Virus Research 52:43-50.

Wilson ML 1994. Ecology and the epiderniology of disease emergence. Ann New York Acad Sei 740: 169180.

Zeller HG, Fontenille D, Traoré-Laminzana M,Thiongane Y, Digoutte JP 1997. Enzootie activityof Rift Valley fevcrvirus in Senegal. Am J Trop MedHyg 56: 265-272.


Molecular Epidemiology of Human Polyomavirus JC in theBiaka Pygmies and Bantu of Central Africa

Sylvester C Chima+, Caroline F Ryschkewitsch, Gerald L Stoner

Neurotoxicology Section, National Institutes ofNeurological Disorders and Stroke, National Institutes ofHealth, Bethesda, MD 20892, USA

Polyomavirus lC (lCV) is ubiquitous in humans and causes a chronic demyelinating disease ofthecentral nervous system, progressive multifocalleukoencephalopathy which is common in AIDS. leV isexcreted in urine of30-70% ofadults worldwide. Based on sequence analysis oflCVcomplete genomesorfragments thereof, lCV can be class!fied into geographically derived genotypes. Types 1 and 2 are ofEuropean and Asian origin respectively while Types 3 and 6 are African in origin. Type 4, a possihlerecombinant ofEuropean and African genotypes (l and 3) is common in the USA. Ta delineate the lCVgenotypes in an aboriginal African population, ralldom urine samples were collectedfrom the BiakaPygmies and Bantufrom the Central African Republic. There were 43 males and 25females aged 4-55years, with an average age of26 years. After PCR ampl!fication oflCV in urine, prodllcts were direct(vcycle sequenced. Five of23 Pygmyadults (22%) andfour of20 Bantu adults (20%) were positivefor lCviruria. DNA sequence analysis revealed lCV Type 3 (two), Type 6 (two) and one Type 1 variant inBiaka Pygmies. Ali the Bantu strains were Type 6. Type 3 and 6 strains oflCV are the predominantstrains in central Africa. The presence ofmultiple subtypes oflCV in Biaka Pygmies may be a resu/t ofextensive interactions ofPygmies with their African tribal neighbors during their itinerant movementsin the equatoria/forest.

Key words: polyomavirus - Je virus - genotypes - Pygmies - Bantu - Africa

The dsDNA polyomavirus JC (JCV) is ubiquitous in humans and bears close sequence homology with other species ofthis genus, BK virus andthe simian virus 40. Sero-epidemiologic studieshave shown that up to 90% of adults are positivefor antibodies to JCV (Walker & Frisque 1986).Infection with JCV is acquired in early childhoodpossibly via the respiratory tract. This is fol1owedby persistent infection of the kidneys from whichJCV is excreted in urine. Studies with polymerasechain reaction (PCR) show that 30-70% of adultsworldwide are positive for JC viruria (Agostini etal. 1996, Sugimoto et al. 1997, Shah et al. 1998).JCV has been established as the causative agent inprogressive multifocal leukoencephalopathy(PML), a fatal demyelinating disease of the centrai nervous system (Zurhein & Chou 1965). PML,previously a rare disorder found in immunocompromised patients with hematologic malignancies,is now prevalent in 5-7% ofAIDS cases in the USAand Europe (Berger & Concha 1995, Martinez etal. 1995), but in only 0.8% of Brazilian AIDS pa-

+Corresponding author. Fax: +301-402-1030.E-mail: [email protected] 15 June 1998Accepted 30 Ju/y 1998

tients (Chimel1i et al. 1992) and 1.5% in West African AIDS cases (Lucas et al. 1993).

The complete genome ofprototype JCV (Mad 1)from the brain ofa patient with PML was sequencedin 1984 (Frisque et al. 1984). The genome consistsof a single molecule of dsDNA, 5.1kb in length,which is transcribed bidirectional1y from the originofDNA replication (ori). It codes for the early region proteins, large T and smal1 t antigens whichregulate transcription ofthe late region proteins VP 13 and agnoprotein. Jev regulatory region can beclassified into two major configurations: an "archetype" which is amplified from urine of normal individuals with JC viruria (Yogo et al. 1990) and a"PML type" when sequenced from the brain of patients with PML. PML-type regulatory regions arederived from the archetypal forrn by unique rearrangements, consisting ofdeletions and duplicationswithin the JCV promoter/enhancer (Ault & Stoner1993, Agostini et al. 1997c).

Based on sequence analysis of JCV completegenomes, as wel1 as segments of the VPI and Tantigen genes, JCV can be classified into severalgeographically based genotypes and subtypes (Ault& Stoner 1992, Agostini et al. 1995, 1997d,Sugimoto et al. 1997). The major genotypes so fardescribed are Type 1, which is of European origin,Type 2, which is Asian, and Types 3 and 6 whichare African in origin (Agostini et al. 1995, 1998).

616 JC Virus Genotypes in African Pygmies and Bantu • Sylvester C Chima et al.

Type 4 which appears to be a recombinant of African and European Types (1 and 3)(Agostini et al.1996), is prevalent within the United States withthe highest frequency in African-Americans. Anew clade of lCV strains, consisting ofthree possible subtypes has been identified in Southeast Asia(Ou et al. 1997) (Chima et al. unpublished data).

Biaka Pygmies (singular 'Aka'), are a groupof aboriginal peoples in central Africa who livepredominantly as hunter-gatherers in the tropicalforest and have a shorter stature when comparedto other Africans. Genetic studies have identifiedPygmies to have distinctive genetic markers whichmay be described as "ultra-African" (CavalliSforza 1986). The Biaka show a Ievel of admixture with other Africans, with a residual incidenceof 18-35% ofancient Pygmy genes (Cavalli-Sforza1986, Cavalli-Sforza et al. 1994). It is estimatedthat the differences between Pygmies and theirclosest African neighbors are great enough to haverequired at least 10-20,000 years of isolation, considering that gene flow between this two groupsoccurs at the rate of only 0.7% per generation(Cavalli-Sforza 1986).

The Biaka Pygmies presented in this study aremembers of the Babenzele clan, the easternmostsubgroup of Aka or "Western" Pygmies, who livein the Dzangha-Sangha dense forest reserve on thebanks of the Sangha river, below 4°N of the equator in Central African Republic (CA.R) (CavalliSforza 1986, Sarno 1995).

The Bantu are African agriculturalists whospeak a group of related languages and occupythe southern third of Africa starting from their putative origin in the Nigeria-Cameroon border in thewest, to the Kenya-coastline in the east and as farsouth as Port Elizabeth in South Africa (Hrbek etal. 1992). Pygmies and their Bantu neighbors havea symbiotic relationship of mutual interdependence(Turnbull 1986, Bahuchet 1993, Sarno 1995). Itis estimated that the Bantu first made contact withPygmies during the Bantu expansion about 2-3,000years ago (Cavalli-Sforza 1986, Hrbek et al. 1992).The Bantu villagers presented in this study live inclose proximity and interact extensively with thePygmies. Indeed, the Biaka and other Pygmy tribesspeak a forrn of Bantu or Nilotic language borrowed from their neighbors having lost their ownlanguage over a long period of contact with otherAfrican tribes. However, ethnologists and linguistscan still recognize cornmon language elementsbetween the Biaka in the west and the most genetically ancient and distant Pygmies (Mbuti), who livein the Ituri forest sorne 800 miles to the east(Bahuchet 1993, Sarno 1995).

It is assumed that lCV, like any good parasite,has co-evolved with its human host. Due to the

stable and distinct lCV genotypes which characterize different populations, urinary lCV has beenshown to be a valuable tool in tracing human migrations (Agostini et al. 1997d, Sugimoto et al.1997). To delineate the lCV genotypes circulating among the aboriginal peoples of central Africa, we undertook a study of the genotype profileof lCV excreted in the urine of the Biaka Pygmiesand their Bantu neighbors with a view to determine whether unique strains of lCV may be circulating within these remote people and to comparethe rates and pattern of lC vilUria with other population groups around the world.

MATE RIALS AND METHODS

Patients and samples - Single urine samples(5-50 ml), were collected from 33 Biaka Pygmiesfrom the Pygmy settlement of Yandoumbe and 28Bantu villagers from Amopolo within the DsanghaSangha dense forest reserve in Bayanga prefectureCA.R. Seven additional urine sampies were alsocollected from two female and five male Bantusliving in the city of Bangui, CA.R. There were 43males and 25 females with an average age of 26years and a range of 4-55 years. Adults 20 yearsand older made up 65% of the sample population.Age detennination of the Pygmy population utiIized educated estimates by an experienced Pygmynurse practitioner. Ali subjects included in thestudy population were healthy volunteers.

DNA extractiol1- Urine samp1es (5-15 ml) werecentrifuged at 4,300 x g for 10 min and cell pelletswere resuspended in phosphate buffered saline(PBS), recentrifuged and the supernatant was discarded. Cells were suspended in 100-200 ,..li digestion buffer containing 0.2 mg/ml ofproteinaseK, 50 mM KCI, 10 mM Tris/HCI (pH 8.3), 2.5 mMMgClz, 10% (wt/vol) gelatin, 0.45% (vol/vol)NP40 and Tween20. After overnight incubationat 55°C in a waterbath, enzyme reactions werestopped by boiling for 10 min. DNA extracts werestored at -70°C until used and 2-10 !lI of the extract was used for subsequent PCR.

PCR - Initial tests for lCV were designed toamplify DNA fragments from the VPI and large Tantigen genes. lCV specific primers for the VPlcoding region were lLP-15 & 16 which arnplify a215-bp fragment frorn this region. This DNA fragment provides up to 15 typing sites for differentiating lCV genotypes and subtypes (JLP-15, nuc1eotides 1710-1734, 5'ACAGTGTGGCCAGAATTCACTACC-3' and lLP-16, nucleotides 1924-1902,5'-TAAAGCCTCCCCCCCAACAGAAA-3'). Asegment of the large T antigen was amplified using the primer pair JTP-5&6 which amplify a 276bp fragment from the T-antigen encoding the zincfinger motif. This region is the site of a mutation

Mem Insl Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep'/Oct. 1998 617

changing a glutamine codon to leucine at aminoacid 301. This point mutation is characteristic ofail African and sorne Asian strains of lCV so farstudied (Agostini et al. 1995, 1997a) (lTP-5 nucleotides, 3621-3642, 5'-CTTTGTTTGGCTGCTACAGTAT-3' and lTP-6 nucleotides, 3896-3877, 5'GCCTTAAGGAGC ATGACTTT-3'). The noncoding regulatory regions and T-antigen intronwere amplified using the primer pairs lRR-25 &28 and lSP-I & 2 respectively. lRR -25 & 28amplify the entire regulatory region (341-bp) including three typing sites to the left of ori for distinguishing Types 1 and 2 strains (lRR-25, nucleotides, 4981-5004 5'-CATGGATTCCTCCCTATTCAGCA-3' and lRR-28, nucleotides, 291-2685' -TCACAGAAGCC TTACGTGACAGC-3 ').Specific mutations at positions 133 and 217 of thearchetypal regulatory region can be used to further characterize African genotypes. Deletion ofcertain pentanucleotide repeats within the regulatory region has been used to subtype lCV strainsin Taiwan (Ou et al. 1997). The lCV specific primers lSP 1&2 amplify a 402-bp fragment from theT-antigen intron which provides additional typingsites for confinning genotype assignments (lSP-Inucleotides, 4390-4412, 5'-ACCAGGATTCCCACTCATCTGT-3' and lSP-2 nucleotides, 47914769, 5'-GTTGCTCA TCAGCCTGATTTTG-3').

Following an initial heating at 94°C for 1.5 min(hot start), the 50-cycle, two-step PCR programinclude 1min for annealing and elongation at 63°C,denaturation at 94°C for 1 min and extension atnoc for 1 min. After a final extension for 10 minreactions were tenninated at 4°C. PCRs were perfonned using UITma DNA polymerase with 3'-5'proofreading activity (Perkin Elmer Cetus) in astandard buffer containing 1.5 mM MgCI2.

Cycle sequencing - Gel-purified PCR productswere sequenced directly using the Exce1 Kit(Epicentre Technologies, Madison, WI) with thesame primers used for DNA amplification endlabeled with 33p-ATP (Amersham, ArlingtonHeights, IL). Initial denaturation at 95°C was followed by 30 cycles of 30 sec at 95°C for denatur-

ation and 1 min at 63°C for annealing and elongation. Products were electrophoresed on a 6% polyacrylamide gel containing 50% urea. Gels werefixed with 12% methanol and 10% acetic acid,transferred to 3MM chromatography paper, driedunder vacuum, then exposed to X-ray film for 1248 hr.

lCV genotypes were identified as previouslydescribed (Ault & Stoner 1992, Agostini et al.1995, 1997b, 1997e, 1998). Sequence relationships were analyzed with GCG programs, Unixversion 8 (Genetics Computer Group, Madison,WI). Primer design was assisted by the OLIGOprogram version 5.0 (NBI, Plymouth, MN).

Reference sequences - The following areGenBank accession numbers for lCV sequencesreferred to in this work: lCV archetypal regulatory region lCV(CY) M35834 (Yogo et al. 1990);lCV coding region lCV(Mad-1 ), 102227 (Frisqueet al. 1984); lCV Type 6 coding and regulatoryregions, AFO 15537 and AFOl5538 (Agostini et al.1998); lCV Type 3 strains #309, U73178, #311,U73501 (Agostini et al. 1997a); lCV strain# 123,subtype lB, AFOI5527 (Agostini et al. 1997b).

RESULTS

The age and gender of the Biaka and Bantuadults tested for lC viruria is given in the Table.Of the 43 adults tested by PCR amplification ofthe VPI coding region, 22% (5 of 23) Pygmiesand 20% (4 of 20) Bantus were shown to excretethe virus in urine. Overall, males had a higher excretion rate than females, seven out of 27 (26%)compared with two out of 16 (13%). None of the24 children and adolescents aged 18 years oryounger included in the sample population werepositive for lC viruria. One of seven samples collected from Bantus in the city of Bangui was positive. This strain, LI 081, was obtained from theurine ofa47-year old Cameroonian ofthe Bemountribe long domiciled in C.A.R.

JCV coding regions - The lCV genotypes excreted by the nine adults were further analyzed bydirect cycle sequencing ofthe lLP-15 & 16 amplified fragments from both directions. Within this

TABLE

Age and gender of Pygmy and Bantu adults screened for Je viruria

Cohort Gender No. adults Age range (years) No. positives % positives

Pygmy M 15 25-55 3 20F 8 30-55 2 25

Total 23 5 22

Bantu M 12 21-55 4 33F 8 22-40 0 0

Total 20 4 20

618 jC Virus Genotypes in African Pygmies and Bantu • Sylvester C Chima et al.

fragment up to 18 typing sites have been identified for differentiating lCV genotypes and subtypes. Fourteen ofthese sites are illustrated in Fig.1. lCV Type 6 can be clearly distinguished fromboth Types 1 and 3 at positions 1790 and 1837.Type 1 strains can be separated from both Types 3and 6 at position Inl, while the two subtypes ofType 1, (1 A and lB) can be differentiated fromeach other at positions 1843 and 1850.

Analysis of the lCV strains from Pygmiesyielded three different types oflCV from five positive samples. These were two Type 3 strains, oneType 1and two Type 6 strains. One of the Type 3strains (L 1059) showed identical sequence in theVP 1fragment to the DNA sequence ofstrain #309previously amplified from the urine of an Africanfrom Mara region in Tanzania (Agostini et al.1995). The other Type 3 strain (L 1066) showedpartial sequence homology with #311 (Type 3B),previously sequenced from an African-American,but differed from this strain at position 1870 wheredeoxyadenosine was inserted in place ofdeoxyguanosine. The latter strain was therefore

tenned a variant of Type 3B pending analysis ofthe complete genome. Strain L 1132, from a BiakaPygmy showed very close sequence homology inthe VP 1 fragment when compared to a Type 1Bstrain, # 123, sequenced from a Caucasian (Agostiniet al. 1997b). However this Aka strain had a distinct point mutation at position 1830, wheredeoxythymidine (T) was replaced by a 'G'. Thismutation caused a change in the codon for amineacid inserted at this position from valine to glycine. This point mutation at position 1830 of Akastrain L 1132 has not been described previously inany Type 1 strains (Agostini et al. 1997b). BothType 6 strains sequenced from Aka were identicalwith the previously reported Type 6 sequence(#601). A total offour lCV strains were sequencedfrom the Bantu. These four strains when analyzedshowed exact sequence homology in the lLP-15and 16 amplified fragments when compared tostrain #60 l, sequenced from the brain ofan African-American patient with PML. The Bantu Type6 strains were also identical to the Aka Type 6(Fig. 1).

Strain

No.

Ethnicity Ol'-MOO\M M '<t li") \Cl00 00 00 00 00.............. ....... .......

o1'00

lCV

Mad1 Caucasian A CGT A T A G T T GAG G Type lA

#123 Caucasian A CGT A T A G T T T G G G Type lB

Ll132 Alea A CGT A T A G ~ T T G G G Type lB

#309 Tanzanian T A A TAC G C T T TAC A Type 3A

Ll059

#311

Ll066

Alea

AfAm

Alea

T A A T

~ :[;J~A C G

:~~C

C

C

T

T

T

T

T

T

T A

T A

T A

1"----'

C !A i Type3A1 11 1

C : G : Type3B1 1

C !A ! Type 3BL .J

Ll069 Alea

Ll076 Aka

Ll044 Bantu

Ll052 Bantu

Ll081 Bantu

Ll138 Bantu

#601 AfAm A A A C A TAC T C T A G G Type 6

A A A C A TAC T C T A G G Type 6






Fig. 1: typing sites within the lLP-15& 16 amplified fragments of the VPI gene. Bantu and Pygmy strams are compared to leVMadl sequence and strains #123 (Type lB) (Agostmi et al. 1997b). #309 (Type 3A) from Tanzania. #311 (Type 3B) and # 601(Type 6) from African-Americans (Agostini et al. 1997a, 1998). L 1132 shows a point mutation at nucleotide 1830. LI 066 showssimllarity with Type 3B nucleotides at positions 1786 and 1804 (solid frame). while it resembles Type 3A at position 1870 (brokenframe). Numbering is based on the sequence of leV Mad 1 (Frisque et al. 1984).


A 276-bp fragment was sequenced from thelarge T antigen of six lCV strains (three Aka andthree Bantu) using the Primer pair lTP- 5 and 6.This T antigen fragment encodes the zinc fingermotif. A specifie point mutation in this fragmentcharacterizes ail African strains of lCV so far described and some Asian strains. This mutation is anon-conservative nucleotide base substitution atposition 3768 from 'T' to 'A' , causing a change inthe amino acid coded from hydrophilic glutamineto hydrophobie leucine (Agostini et al. 1997a). Thesix Bantu and Pygmy strains amplified from theT-antigen zinc finger region showed a mutation atposition 3768 (Fig. 2). Typing sites within thisfragment confirm strain LI 059 as a Type 3 strainand strains LI052, L1069, L1076, LI081 andLI138 as Type 6 strains.

leV noncoding regiolls - Noncoding regu1atory regions of six lCV strains from Bantus andPygmies were sequenced by the primers lRR-25and 28 from both directions. The DNA sequencewas compared to the consensus archetypa1 sequence ofType 1 (Agostini et al. 1996) and a Type3 regulatory region sequence #309 from an Tanzanian (Agostini et al. 1997a). The Aka Type 3 strain(L 1059) showed sequence identity with #309 inc1uding a point mutation at position 133 where 'C'is characteristic ofail Type 3 strains. Four Type 6strains from Bantus and Pygmies, (LI052, LI 069,LI 076, and Ll138) ail showed an archetypa1 configuration without deletions. Strains LI 081 (Type6, Bantu) and LI 059 (Type 3, Aka) both show a

10-bp deletion at nucleotides (51-60),just preceding the first NF1 site (Fig. 3). The deletion at thissite is identical to those observed in strains #307and #309 from Tanzania (Agostini et al. 1995,1997a). Ali the Type 6 strains and the single Type3 strain were characterized by the nucleotide "G"at position 217, however only the Type 3 strainshowed deoxycytosine at position 133 ofthe regulatory region.

A 402-bp fragment was amplified from thenoncoding T-antigen intron using the primers lSP1 and 2. This fragment provides up to 15 additional typing sites for confirmation of lCV typesand subtypes from the coding region sequences.Seven lCV strains were amplified from this fragment in the Pygmy and Bantu cohorts. Cycle sequencing confirmed the previous type assignmentsfrom the VP 1gene. LI 044 (Bantu, Type 6) showedtwo nucleotide mutations at positions 4562 and4648 while LI 059 (Aka, Type 3) showed a singlemutation at position 4435 (Fig. 4). The significanee of these point mutations is unknown sincethe primary function of introns is to be spliced outprior to protein translation.

DISCUSSION

This study delineates the genotype profile oflCV strains circulating among the Biaka Pygmiesand Bantu from Bayanga prefecture ofC.A.R. Thisaboriginal African population excretes lCV in urineat a lower rate (21%) when compared to rates ofexcretion in urban populations in the United States

[- JTP 5&6

JCV 3680 3710 3722 3743 3768 3770 3809 3830 3836 3848

Mad! A T G C T T A T G A

Type 3 A C G C A T G G A T

Type 6 A C A C A T A T G A

Ll052 A C A C A T A T G A

Ll059 A C G C A T G G A T

Ll069 A C A C A T A T G A

LI076 A C A C A T A T G A

Ll08! A C A C A T A T G A

LI 138 A C A C A T A T G A

Fig. 2: typing sites within the JTP-5&6 amplified fragment of large T antigen Including the zinc finger motif. Position 3768(frame) shows site of nucleotide mutation from "T" to "A" in ail AfTican genotypes including Bantu and Pygmy strains whencompaTed to JCV Mad 1.

620 JC Virus Genotypes in African Pygmies and Bantu • Sylvester C Chima et al.

Archetype# 309LI 059Ll052LI 069Ll076Ll081LI 138

Or;1 60GCCTCGGCCTCCTGTATATATAAAAAAAAGGGAAGGTAGGGAGGAGCTGGCTAAAACTGG______________________________________________[~~-l______________________________________________ f- 1

______________________________________________ [~---~l

120Archetype ATGGCTGCCAGCCAAGCATGAGCTCATACCTAGGGAGCCAACCAGCTGACAGCCAGAGGG# 309Ll059Ll052LI 069LI 076Ll081Ll138

Archetype# 309LI 059LI 052LI 069Ll076Ll081LI 138

Archetype# 309Ll059LI 052LI 069LI 076Ll081LI 138

Archetype#309Ll059LI 052LI 069LI 076Ll081Ll138

ln lWAGCCCTGGCTGCATGCCACTGGCAGTTATAGTGAAACCCCTCCCATAGTCCTTAATCACA___________C ____________C ____________A ____________A ____________A ____________A ____________A _

217 240AGTAAACAAAGCACAAGGGGAAGTGGAAAGCAGCCAAGGGAACATGTTTTGCGAGCC__________________________________G ___________________________________G ___________________________________G ___________________________________G ___________________________________G ___________________________________G ___________________________________G _

270AGAGCTGITITGGCTTGTCACCAGCTGGCCATG [q Start agnoprotein]

Fig. 3: regulalory region sequences amplified from Pygmy and Bantu strains is compared to the consensus archetypal regulatoryregion of Type 1 (Agostini et al. 1996) and #309 from Tanzania. Dashed lines denote unifonnity with the consensus archetypalsequence. Solid lines show areas of nucleotide deletion initially observed in strains #307 and #309 (Agostmi et al. 1995. 1997a)and now found in LI059 from a Blaka Pygmy and LI08l from a Bantu. At position 133. "A" IS replaced by "c" in aU Type 3strains. At position 217, both Type 3 and Type 6 strains substitute deoxyguanosine for deoxyadenosme. Numbering IS based onarchetypal numbering of strain CY (Yogo et al. 1990).

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(S}, Sep./Oct. 1998 621

JSP 1&2

JCV 0 '<t V'l 00 N \0 r-- '<t ~N '<t N 00 N :! r--

M M ..., ..., '<t \0 r-- 00 00 N ..., \0 '<t V'l \0'<t '<t ~ ~

'<t~

'<t '<t '<t V'l V'l V'l V'l \0 \0 r-- r--'<t '<t '<t '<t '<t '<t '<t '<t '<t '<t '<t '<t '<t '<t

Mad 1 A A A C T G G A A G G T G G C G C

Type 3 C A A A T G C G T G A C G G C A T

#601 A A A A T A C A T A G C G G T G C

LlO44 A A A A T A G A T G G C IT A IT G C

Ll052 A A A A A A G A T G G C G G T G C

Ll059 C A ~A T G C G T G A C G G C A T



LI081 A A A A A A G A T G G C G G T G C


Fig. 4: the JSP-l &2 amplified fragment of the T antigen intron further confinn genotype assignments from the VPI and large Tantigen genes. Typing in this region is compared to the consensus sequence ofType 3 (Agostini et al. 1997a), stram #601 (Agostmiet al. 1998) and Mad 1. Framed sets denote sites ofspecific point mutations in LI 044 and Ll059 from Biaka Pygmies. Numberingis based on Mad 1 sequence.

(41 %) (Agostini et al. 1996) and Europe (Stoneret al. 1998a). Native American tribes in the UnitedStates and the Pacifie Islands show a rate of JCvirus excretion in urine (65%) (Agostini et al.1997d), which is three times the rate observed inthis African cohort. However the rate ofexcretionamong the Bantu and Pygmies are somewhat closerto a reported incidence rate of 30% in HIV positive patients from the Mara region of northwestTanzania (Agostini et al. 1995). The reasons forthe differences in rates of JCV virus excretion indifferent populations is not yet explained. However, it may be related in part to the difference inage ofvarious sample populations. Studies in Caucasians and African-American cohorts within theUnited States have shown that the rate of JC virusexcretion in urine rises dramatically in the fifthdecade oflife (Agostini et al. 1996), (Chima, unpublished observations). It therefore fol1ows thatsample populations with older age groups are morelikely to yield a higher rate ofJC viruria. The Afriean cohort studied here had only three adults estimated to be aged 50 years or older.

Analysis of the JCV strains from Pygmy urinerevealed four different subtypes from the five positive cases. These were two Type 3 strains (one 3A

and one 3B variant), two Type 6 and one Type 1Bvariant. The Type 3A strain showed close identitywith Type 3 strains previously reported amongNilotic Africans of the Luo tribe from the Mararegion of Tanzania. The Type 3B strain showed asimilar sequence to that recently found in an African-American (strain A 179) (Chima, unpublisheddata). This is a variant ofstrain #311 also found inan African-American with an 'A' to 'G' substitution at position 1870 of the VP 1 gene. The twoType 6 strains were identical to those sequencedfrom the urine of the Bantu in this study.

JCV Type 6 was first sequenced from the brainof an African-American patient with PML(Agostini et al. 1998). This was later identified asa new subtype of JCV when similar strains weresequenced from the urine of Africans from Ghana(Guo et al. 1996). Type 6 strains have also beensequenced from the brains of AIDS patients withPML from the Ivory Coast (Stoner et al. 1998b) aswel1 as the urine of an immunocompetent individual from Sierra Leone (Chima, unpublisheddata). The four JCV strains excreted in the urineof Bantus reported here are Type 6. Of the fourBantu strains, (L 1081 ) showed a 1O-bp deletion inthe regulatory region sequence similar to that found

622 Je Virus Genotypes in African Pygmies and Bantu • Sylvester e ehima et al.

in #309 from Tanzania and LI 059 in Pygmies.However, LI 059 also displays another marker ofType 3 strains, i.e., deoxycytosine at position 133of the archetypal regulatory region. It is more likelytherefore, that these two strains arose independentlyof each other rather than as a result ofviral recombination. We can hypothesize that the two Africangenotypes of lCV (Types 3 and 6) may have coevolved, independently of each other, in their respective African hosts. Ali genotype studies onlCV in Africans so far have shown that both Type3 and 6 strains can be found in West and CentralAfrica (Guo et al. 1996, Sugimoto et al. 1997,Stoner et al. 1998b), while Type 3 is the only genotype so far described from East Africa (Agostini etal. 1995).

Archeological and linguistic data have shownthat the Biaka Pygmies migrated to their presentlocation from a region north of the lturi around thesouthem Sudan, first to northem Zaire and then ina northwest direction to their present location inthe southwest tip ofC.A.R. around the Sangha river(Cavalli-Sforza 1986, Bahuchet 1993). The putative site ofBiaka Pygmy origin around the southem Sudan is closer to the region occupied by previously studied Africans from northwest region ofTanzania. The latter population are in part Niloticsof the Luo tribe (Agostini et al. 1995). This groupexcrete Type 3 lCV strains similar to those foundin Biaka Pygmies. The Bantus on the other handare migratory fanners thought to have come intocontact with the Pygmies about 2000 years agoduring the Bantu expansion from West Africa(Cavalli-Sforza 1986, Hrbek et al. 1992). Archeologists and historians estimate that during the second stream of the Bantu expansion, there was amigration along the banks of the Sangha river intocentral Africa (Hrbek et al. 1992). It is thereforeIikely that Bantu descendants of the first immigrants still occupy the present location and carrylCV strains transmitted from their parents. Due tothe close interaction between the Pygmies and theirBantu or Nilotic neighbors in equatorial Africa, itmay be speculated that Type 6 strains were transmitted to the Biaka during their later interactionswith Bantus while the Type 3 strains were broughtalong during their migration from southem Sudanand East Africa.

A Type 1B variant of lCV was sequenced fromthe urine of a 55 year old female Pygmy. Type 1strains are generaUy characteristic of Europeans.This Aka strain bears a unique mutation at position 1830 not previously reported in Type 1strainsoflCV (Agostini et al. 1997b, Stoner et al. 1998a).The significance of this Type 1 strain is unknownalthough in another study, it has been reported thata pocket ofthe European subtype oflCV was found

in Bangui, C.A.R. (Sugimoto et al. 1997). Analysis of the complete genome of the Aka Type 1Bvariant and identification ofmore lCV strains withsimilar mutations will facilitate characterization ofthis subtype. lt is possible that on analysis of thecomplete genome, this strain may represent aunique subtype oflCV different from Type 1strains

We conclude that human polyomavirus lCV isexcreted in the urine ofBiaka Pygmies and Bantusof central Africa, though at a lower rate than thatobserved in other population groups. This studyconfirms Types 3 and 6 as the predominant genotypes of lCV in central Africa. The finding offourdifferent subtypes of lCV in the urine of BiakaPygmies may be explained by the extensive interactions of Pygmies with their various African tribalneighbors over a long period oftime, as they movedfrom place to place in the equatorial forest.

ACKNOWLEDGMENTS

To Hansjurgen T Agostini for initial studies on African gcnotypes of JC virus. To the entire staff of theWorld Wildlife Fund in Bangui and Bayanga for theirkind hospitality and assistance throughout our stay inthe Central African Republic.

REFERENCES

Agostini HT, Brubaker GR, Shao J, Levin A,Ryskcwitsch CF, Blattner WA, Stoner GL 1995. BKvirus and a new type of JC virus excreted by HIV-Ipositive patients in rural Tanzania. Arch Virol 140:1919-1934.

Agostini HT, Ryschkewitsch CF. Stoner GL [996. Genotype profile ofhumam polyomavirus JC excreted inurine of immunocompetent individuals. J ClinMicrobiol34: 159-164.

Agostini HT, Ryschkewitsch CF, Stoner GL 1998. Thecomplete genome ofJC Virus Type 6 from the brainof a African-American with progressive multifocalleukoencephalopathy (PML). J Hum Viral: in press.

Agostini HT, Ryschkewitsch CF, Brubaker GR, Shao J,Stoner GL 1997a. Five complete genomes of JC virus Type 3 from Africans and African AmericansArch Viral /42: 637-655.

Agostini HT, Ryschkewitsch CF, Singer CF, Stoner GL1997b. JC virus Type 1 has multiple subtypes: threenew complete genomes. J Gen Vùv! 79: 801-805.

Agostini HT, Ryschkewitsch CF, Singer EJ, Stoner GL1997c. JC virus regulatory region rearrangementsand genotypes in progressive multifocal leukoencephalopathy: two independent aspects of virusvariation. J Gen Viml 78: 659-664.

Agostini HT, Ryschkewitsch CF, Yanagihara R, DavisV, Stoner GL J997d. Asian genotypes of Je virus(lCV) in Native Americans and in a Pacifie Islandpopulation: markers ofhuman cvolution and migration, Prac Natl Acad Sei US 94: 14542-14546.

Agostini HT, Shishido Y, Ryschewitsch CF, Stoner GL1997e. JC Virus Type 2: definition ofsubtypes basedon analysis of ten complete genomes. J Gell Viral:in press.

Mem /nst Oswa/do Cruz, Rio de Janeiro, Vol. 93(5), 5ep./Oct. 1998 623

Ault GS, Stoner GL, 1992. Two major types of JC virusdefined in progressive multifocalleukoencephalopathy brain by early and late coding region DNA sequence J Gen Virol 73: 2669-2678.

Ault GS, Stoner GL 1993. Human polyomavirus JCpromoter/enhancer rearrangement patterns from progressive multifocal leukoencephalopathy brain areunique derivatives ofa single archetypal structure. JGen Viro174: 1499-1507.

Bahuehet S 1993. History of the inhabitants of the Centrai African Rain Forest: Perspectives from comparative linguistics, p. 37-54. ln CM Hladik, A Hladik,o Linares, H Pagezy, A Semple & M Hadley (eds),Tropical Forests. People and Food. UNESCO, Paris.

Berger JR, Concha M 1995. Progressive multifocalleukoencephalopathy: the cvolution of a disease onceconsidered rare. J Nel/rovirol 1: 5-18.

Cavalli-Sforza LL 1986. Aji"ican Pygmies, AcademiePress, Orlando.

Cavalli-Sforza LL, Menozzi P, Piazza A 1994. Africa,p. 159-194. ln Cavalli-Sforza LL, Menozzi P, Piazza A (eds), The History and Geography of HI/man Genes. Princeton University Press, Princeton.

Chimelli L, Rosemberg S, Hahn MD, Lopes MBS,Barretto-Netto M 1992. Pathology of the centralnervous system in patients infected with the humanimmunodeficiency virus (HIV): a report of252 autopsy cases from Brazil. Neuropathol Appl Neurohiol18: 478-488.

Frisque RJ, Bream GL, Cannella MT 1984. Humanpolyomavirus JC virus genome. J Viro15/: 458-469.

Guo J, Kitamura T, Ebihara H, Sugimoto C, Kunitake T,Takehisa J, Na YQ, AI-Ahdal MN, Hallin A, KawabeK, Taguchi F, Yogo Y 1996. Geographical distribution of the human polyomavirus JC virus type A andB and isolation of a new type from Ghana. J GenVirol77: 919-927.

Lwango-Lunyiigo S, Vansina J 1992. The Bantu-speaking peoples and their expansion, p. 75-85. In 1Hrbek,General History ofAfrica. Vol. 111, Africa from theSeventh to the Eleventh CentulJ'. UNESCO, Paris.

Lucas SB, Hounnou A, Peacock C, Beaumel A, DjomandG, N'gbichi J-M, Yeboue K, Honde M, DiomandeM, Giordano C, Doorly R, Brattegaard K, KestensL, Smithwick R, Kadio A, Ezani N, Yapi A, De CockKM 1993. The mortality and pathology of HIV infection in a West African city. AlDS 7: 1569-1579.

Martinez AJ, Sell M, Mitrovics T, Stoltenburg-DidingerG, Inglesias-Rojas JR, Giraldo-Velasquez MA,Gosztonyi G, Schneider V, Cervos-Navarro J 1995.The neuropathology and epidemiology of AlOS. ABerlin experience. A review of200 cases. Path ResPract 191: 427-449.

Ou W, Tsai R, Wang M, Fung C, Hseu T, Chang D, 1997.Genomic Cloning and sequence analysis ofTaiwan3 Human Polyomavirus JC Virus. J Formos MedAssoc 96: 511-516.

Sarno L 1995. Bayaka: the ExtraordinGl)' Music of theBahen::ele Pygmies. Ellipsis Arts, New York, 66 pp.

Shah KV, Daniel RW, Strickler HD, Goedert JJ 1998.Investigation ofhuman urine for genomic sequencesof the primate polyomaviruses simian virus 40, BKvirus, and JC virus. J Infect Dis 176: 1618-1621.

Stoner GL, Agostini HT, Ryschkewitsch CF, Komoly S1998a. JC virus excreted by multiple sclerosis patients and paired controls from Hungary. MultipleSclerosis 4: 45-48.

Stoner GL, Agostini HT, Ryschewitsch CF, Mazlo M,Gullotta F, Wamukota W, Lucas S 1998b. Two casesof progressive multifocal leukoencephalopathy(PML) due to JC virus: detection ofJCV Type 3 in aGambian AIDS patient. J Med Microbiol 47: 1-10.

Sugimoto C, Kitamura T, Guo J, AI-Ahdal MN,Schelnukov SN, Otova B, Ondrejka P, Chollet JY,El-Safi S, Ettayebi M, Gresenguet G, Kocagoz T,Chaiyarasamee S, Thant KZ, Thein S, Moe K,Kobayashi N, Taguchi F, Yogo Y 1997. Typing urinary JC virus DNA offers a novel means oftracinghuman migrations. Proc Natl Acad Sci USA 94:9191-9196.

Turnbull CM 1986. Survival factors among Mbuti andother hunters of the Equatorial Rain Forest, p. 103123. In Cavalli-Sforza LL, African Pygmies. Academie Press, Orlando.

Walker DL, Frisque RJ 1986. The biology and molecular biology of JC virus, p. 327-377. In NP SalzmanThe Papovaviridae. Vol. l, The Polyomaviruses. Plenum Press, New York.

Yogo Y, Kitamura T, Sugimoto C, Ueki T, Aso Y, HaraK, Taguchi F 1990. Isolation of a possible archetypai JC virus DNA sequence from non immmunocompromised individuals. J Virol 64: 3139-3143.

Zurhein GM, Chou SM 1965. Particles resemblingpapova viruses in human cerebral demyelinating disease. Science 148: 1477-1479.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 625-626, 5ep.lOct. 1998 625

RESEARCH NOTE

Molecular Epidemiology ofDen-2 Virus in Brazil

MP Miagostovichj+, RMR Nogueira,HG Schatzmayr, RS Lanciotti*

Laborat6rio de Flavivirus, Departamento de Virologia,Instituto Oswaldo Cruz, Av. BrasiJ 4365, 21045-900

Rio de Janeiro, RJ, Brasil *Centers for DiseaseControl and Prevention, CDe, Fort Collins, CO, USA

Key words: dengue virus type 2 - sequencing - Brazil

Dengue (DEN) viruses belong to the familyFlaviviridae, genus Flavivirus, and occur as fourantigenically related, but distinct serotypes designated DEN-l, 2, 3 and 4 (EG Westaway et al. 1985Intervirology 24: 183-192). The viruses are characterized by a single strand of RNA associatedwith a core protein, in a nucleocapside surroundedbya Iipid envelope. The genoma consist of a singleopen reading frame coding for core protein (C),precursor of the membrane protein (prM) and enveIope (E) structural proteins, followed by the nonstructural proteins NS1, NS2a, NS2b, NS3, NS4a,NS4b and NS5 (FZ Heinz & JT Roehring 1990 p.289-305. In MHV Van Regenmorte, AR Neurath(eds), Immunochemistry of Vin/ses, ll. The Basisfor Seradiagnosis and Vaccines, Elsevier).

The genetic variation among DEN viruses hasbeen demonstrated by numerous methods including oligonucleotides fingerprinting, restriction enzymes, primer extension sequencing and nucleotidesequences from different fragments of genoma (RRico-Hesse 1990 Viralogy 174: 479-493, lA Lewiset al. 1993 Virology 197: 216-224, V Deubel et al.1993 Arch Viral 129: 197-210, V Vorndam et al.1994 Arch Viro1136: 191-196, DW Trent et al. 1989Viralogy 172: 523-535, RS Lanciotti et al. 1994 JGen Viral 75: 65-75, E Chungue et al. 1995 J GenVirol 76: 1877-1884, KZ Thant et al. 1995 MicrobiolImmunol 39: 581-590).

Since intra-serotypic antigenic variations (genetic subtypes) could be associated with severedisease it is important to monitor the distributionand eventual introduction of new genotypes of

Financial support: CNPq and Fundaçào Banco do Brasil.+Corresponding author. Fax: + 55-21-270.6397. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

existing serotypes into areas where dengue activity are troublesome (Vorndam et al. 1994 lac. cit.).

ln this report we sequenced the E fragmentfrom geographically and temporally distinct DEN2 viruses isolated in Brazil during 1990-1995, inorder to investigate the genetic subtype distribution of this serotype virus in the country.

DEN-2 viruses analyzed in this study were obtained from the collection of the Laboratory ofFlavivirus, Department of Virology, IOC, Fiocruz.These strains were isolated from sera by inoculation into Aedes albopictus clone C6/36 ceillinc (AIgarashi 1978 J Gen Viral 40: 531-544) and wereidentified by immunofluorescence using type-specific monoclonal antibodies. Virus seeds were amplified once by inoculation into C6/36 (01 Gubleret al. 1984 Am J Trop Med Hyg 33: 158-165).

Viral RNA was extracted from infected C6/36ceUs by using the acid-guanidin isothiocyanateprocedure previously described (RS Lanciotti etal. 1992 J Clin Microbiol 30: 545-551 ). Oligonucleotide primers used in the amplification and sequencing protocols were designed with the aid ofthe oligo program (National Bioscience Inc., Plymouth, MN).

Nucleotides from positions 1685 to 2504 coding for the fragment ofE gene were amplified usingRT-PCR. The RT reaction was performed in 4111 of5X RT reaction buffer (BRL), 4111 ofwater, 2111 ofO.IM DTT, 5111 of251lM dNTP's, 0.2111 ofRnasin(40U/IlI), 2111 of 10 mM downstream primerD2CP2504 (5' GGGGATTCTGGTTGGAACTTATATTGTTCTGTCC 3'),2 III of RNA and 1 III of200 U Superscript RT (Gibco). The reaction wasincubated at 50°C for 10 min then 50 min at 45°C.A PCR amplification was followed by adding 10III of RT reaction to 90 III of PCR reaction mix (74III of water, 9 III of 10X C buffer, 5 III of 25 mMdNTP, 2111 of 10 mM upstream primer D2P 1685(5'CTAGGATCTCAAGAAGGAGC AATGCA 3')and 0.5 III Taq. The DNA molecules were denaturatedat 94°C for 4 min and subjected to a 35 amplificationcycles (94°C for 1 min, 55°C for 1 min, noc for 8min) and to one of72°C for 10 min.

After an eletrophoresis on a 1% agarose gel,the amplified DNA bands were excised and purified by using the Bio 101 Gene-Clean kit. Purified DNAs were then sequenced by using the following primers: D2P1685- 5'CTAGGATCTCAAGAAGGAGCAATGCA 3'; P760- 5' GGATCACAAGAAGGAGCCATGCA 3'; CPI171 - 5'ATGG AGCTTCCTTTCTTCTTGAACCA 3';CPI234 - 5' CCAAAGTCCCAGGCTGTGTCTCCCAGAATGGCCAT 3'. The sequencing reactionwas perfonned by using the Taq DyeDeoxy Terminator Cycle Sequencing Kit (Applied Biosystem,Inc., USA). Cycle sequencing parameters were exactly as described in the manufactures protocol.

626 Molecular Epidemiology of Den-2 Virus· MP Miagostovich et al.

The overlapping nucleic acid sequences obtained from individual sequencing reactions werecombined for analysis and edited using theDNASTAR program (Madison, WI). The DEN-2virus nucleic acid sequences were then aligned witheach other, and with DEN-2 envelope sequencesobtained from GENEBANK, using the multiplesequence alignment algorithm CLUSTAL (0Higgins, Heidelberg, Germany). Phylogenetic treeswere reconstructed from the aligned nucleic acidsequences using algorithms based upon parsimony(program PAUP, 0 Swofford, Champaign, IL).

The comparison of our results with thephylogram generated by the sequencing of the entire E gene (Lewis et al. 19931oc. cit.) showed thatail the isolates belong to subtype III (Figure). Theresults confirmed the asiatic origin of DEN-2 strainsisolated in the State of Rio de Janeiro previouslydemonstrated by Rico-Hesse (1990 loc. cit.) andLewis et al. (1993 loc. cil.). The circulation of thesame genotype in ail areas studied demonstratedthe dispersion of DEN-2 virus from Rio de Janeiroto the other states of the country. The subtype IIIhas been refered to have a greater potential to causesevere disease causing concern in those areas inwhich high rates ofantibody to DEN-l and DEN-4viruses predispose populations to severe disease(Vorndam et al. 19941oc. cil.).

ln Brazil, the increasing incidence of DHF/

DSS was associated with the introduction of theDEN-2 viruses in the states of Rio de Janeiro,Ceani and recently in the State of Rio Grande doNorte (RMR Nogueira et al. 1993 EpidemiollnJect III: 163-170, RV Souza et al. 1995 Mem InstOswaldo Cruz 90: 345-346, PFC Vasconcelos etal. 1995 Rev Inst Med Trop Sào Paulo 37: 253255, SMü Zagne etaI. 1994 Trans R Soc TropMedHyg 88: 677-679) after a period of high DEN-)virus activity. In the states of Bahia and EspiritoSanto, where DEN-2 virus was responsible initiallyfor primary infections, signs and symptoms ofclassic dengue fever were observed (RMR Nogueiraet al. 1995 Rev Inst Med Trop Sào Paulo 37: 507510). In those states a higher percentage ofexhantema and pruritus were observed when compared with signs and symptoms due to DEN-I primary infection during 1986 in Rio de Janeiro.

Recently, R Rico-Hesse et al. (1997 Viralogy230: ]-8) demonstrated the direct association between the introduction ofsoutheast Asian DEN-2viruses severe disease in America and showed acirculation of a new subtype responsible for DHFepidemics in Mexico and Venezuela, in 1995. Thisdata point out the need to continue rnolecular epidemiological studies in dengue endemic areas inorder to monitor the introduction of a new subtypeand the impact of it over the population.

Acknowledgement: to Dr J Chang for supplying theprimers.

129 21

436 12

336

22

1 1911

9 13

49

702

30 22

9 0022

4û 12 11

4

26130

22

118

1567

17 615

527

11 5

2121

2731

9130 3

14

466

New Guinea C-44Sri Lanka -68Sri Lanka -69Philippines -83Taiwan -87Thailand -58Thailand -64Thailand -80Malaysia -86Jamaica -83Brazil-9039122 - RJ 199040247 - RJ 199040678 - RJ 199048576 - CE 199448577 - CE 199448586 - CE 199448643 - BA 199448844 - BA 199449255 - RJ 199549367 - RJ 1995H140700 - AL 1991H140749 - AL 1991Indonesia -76Somalia -84Burkina Faso -83seychelles-77Sri Lanka -82Sri Lanka -85Sri Lanka -89Sri Lanka -90Trinidad -53ndia -57

Puerto Rico -69Tonga -74Dengue 3

Phylogram gene!ated by pa.rsimony analysis ofnucleic acid sequences from the prM/M and E genes of 12 DEN-2 Brazilian strainsand 24 DEN-2 vlruscs obtamcd from Gcncbank.


Antimalarial Drug Resistance: Surveillance and MolecularMethods for National Malaria Control Programmes

Umberto D'Alessandro

Prince Leopold lnstitute of Tropical Medicine, Antwerp, Belgium

National malaria control programmes have the responsibility to develop a policyfor malaria disease management based on a set ofdefined criteria as e'/jicacy, side ef{ects, costs and compliance. Thesewillfluctuate over time and nationalguidelines will require periodic re-assessment andrevisioll. Changinga drug policy is a major undertaking that can take several years before beingfully operational. Thestandard methods on which a decision can be taken are the in vivo and the in vitro tests. The latter allowa quantitative measurement o{the dmg response and the assessment ofseveral drugs at once. Howevel;in terms ofdrug policy change its results might be difficult to interpret although they may be IIsed as anearly warning system{or 2nd or 3rd line dnlgs. The new WHO l4-days in vivo test addresses mainly theproblem oftreatment failure and o{haematological parameters changes in sick children. lt gives valuable information on whether a drug still 'works '. None (J{ these methods are weil suitedfor large-scalestudies. Molecular methods based on detection (J{mutations in parasite mo/ecules targeted by antimalurial drugs could be attractive tools for surveillance. Howevel: their relationship with in vivo testresults needs to be established.

Key words: antimalarial drug resistance - in vivo test - il1 vitlV test - polymerase chain reaction - surveillance

Despite considerable efforts done during thiscentury to eradicate or control it, Plamodiumfalciparum malaria is still the most prevalent andthe most devastating disease in the tropics (WHO1993). In the last decades, its control and treatmenthas been complicated by the emergence of resistance to widely used antimalarial drugs such aschloroquine. Drug resistance of malaria parasiteshas been defined as the ability of a parasite strainto multiply or to survive in the presence of concentrations of a drug that normally destroy parasites of the same species or prevent their multiplication. Its dynamics and occurrence are the resultof several interactions:

Parasites and drugs - Natural populations ofP falciparum are heterogeneous mixtures of individuals with different, genetically determined degrees of drug response. The efficacy of medication will depend on the concentration of the drugin relation to the parasite's sensitivity and the timeover which concentrations above this threshold aremaintained. A small fraction of the original parasite population might always survive to the drugbut it will be eventually removed by the immunesystem. However, the infection will not be c1earedif the surviving fraction is too large due either to

Fax: +32-3-247.6362. E-mail: [email protected] 15 June 1998Accepted 30 July 1998

reduced sensitivity or to subcritical drug concentrations. Selection of resistant strains could occurwhen a particular drug is misused (Wernsdorfer1991). The transmission of such parasites mightalso be enhanced by an increased production ofgametocytes (Robert et al. 1996).

Humans and drugs - Readily absorbed drugswith a long half-life, like mefloquine andsulfadoxine-pyrimethamine (SP), can permit effective single dose treatment of malaria and the following chemoprophylactic period prevents infection for several weeks and may be important inrecovery from anaemia. However, these drugs arelikely to exert undesirable drug pressure for a longtime once their concentrations drop below the critical threshold and may select resistant parasites. Thishas been shown in Kenya where a potent selectivepressure for resistance operates even under conditions of supervised drug administration and optimal dosage (Watkins et al. 1997).

Vector andparasite - Vectors may be more receptive to resistant strains and may produce moreparasites compared to sensitive strains. Enhanceddrug pressure and uninhibited transmission mightproduce a fast selection and spread of resistantparasites (Wernsdorfer 1991).

NATIONAL DRUG POLlCIES

It is the responsibility ofNational Programmesto develop a policy for malaria disease managementas early diagnosis and adequate treatment remainthe basic elements of any malaria control action

628 Drug Resistance Surveillance and Molecular Methods • Umberto D'Alessandro

(WHO 1993). The basis for the national drug policyshould be a set of detined criteria for efficacy, sideeffects, costs and compliance. Rational prescribingin the public sector should be promoted through thedevelopment and introduction of treatment guidelines. These activities should be supported by thesupply system to ensure that the drugs health workers have been trained to prescribe are actually available (Baudon 1995). Many of the factors influencing the national drug policy, such as parasite drugsusceptibility, drug pricing and availability, will fluctuate over time and national guidelines for malariatreatment will require periodic re-assessment andrevision (WHO 1996). This should be done on thebasis ofreliable information that, in case ofparasitedrug susceptibility should be collected by means ofan appropriate surveillance system.

Implementing a change in drug policy is a major undertaking that can take several years beforebeing fully operational. The 'reaction-time' couldbe quite long in countries where the capabilities oftraining and re-training of health workers is limited and the distribution system of drugs is complex, fragmentary and uncontrolled. A study donein Congo in 1993 has shown that 77% of GeneralPractitioners (GPs) did not know the national drugpolicy and that they continued to apply a strategy(weekly malaria chemoprophylaxis in childrenunder 5) abandoned seven years before (Baudon1995). In optimal conditions the reaction time hasbeen estimated to at least two years (Baudon 1995).Considering these difficulties, once the decisionofchanging the first line drug has been taken thereis no go back.

SURVEILLANCE

One of the major questions is whether the firstline drug is still 'working'. Unfortunately, the criteria on which this decision can be taken are notclear (Bloland et al. 1993). The standard methodsto assess the efficacy of a given drug can be divided in two broad groups: in vivo and in vitro tests.

In vivo test - During the in vivo test the recommended dose of an antimalarial drug is administered to infected subjects and the parasite's response in the host is assessed. The test could bedone on symptomatic or asymptomatic people(Wernsdorfer & Payne 1988) with a 7-day followup (Prasad et al. 1990). It assesses only the initialparasitological response and, to a limited extent,clinical response to therapy during the follow-up.However, it does not address the implications andmanifestations ofpersistent parasitaemia occurringafter poor response, for example its impact on othercondition such as anemia and malnutrition (Blolandet al. 1993). A study carried out in Malawi and inKenya among young children with clinical malaria

showed a shortening of the duration of clinicalimprovement and a decreased haematological recovery after therapy with chloroquine as comparedto SP. In addition the health care system is burdened by both the drug cost and the manpower requirements of frequent and repetitive visits fortreatment with a poorly efficacious drug. The new14-days ill vivo test proposed by WHO (WHO1996) tries to address this problem by lookingmainly at treatment failure and change in thehaematological parameters (Hb/PCV at day 0 and14). It is a more clinical test, carried out on sickchildren (fever + parasitaemia) aged 6 months-5years (D'Alessandro et al. 1997). This test doesnot permit a quantitative assessment of the drugsensitivity of individual parasite populations, mayoccasionally be influenced by the abnormal fateof the drug in individual patients and it is influenced by the immunological host response to theparasite. However, it gives an information that iscloser to real-life situation and therefore essentialin deciding drug policy changes.

III vitro test - The in vitro test consists in measuring the inhibition of schizont maturation by increasing doses of a given antimalarial drug. It allows the quantitative measurement of drug response, permits to test several drugs at once andimitate the non-immune state. However, it is generally held that the in vitro tests do not reflect thedegree of in vivo resistance, since the latter is substantially determined by factors related to the host'sresponse (Draper et al. 1988). Furthermore, it isestimated that the technical capacity to conduct thenecessary assays is difficult to develop and maintain in national malaria control programmes. Therefore, in vitro testing cannot substitute in vivo observations of malaria therapy and is inappropriatefor making policy decision on drug use. Nevertheless, in vitro tests may provide an early warning ofimpeding resistance before this becomes clinicallyapparent. The optimal deployment of in vitro testsshould be to define specific issues related to temporal and geographical trends ofparasite's responseto drugs. Such issues include the longitudinal follow-up of parasite drug susceptibility, monitoringthe patterns of parasite cross-resistance to different drugs and the establishment of baseline dataon the susceptibility oflocal parasites to new drugs(WHO 1990).

MOLECULAR METHOOS FOR SURVEILLANCE

Recently, molecular diagnostic methods fordetecting resistant parasites have been proposedfor monitoring the lever and spread of resistance(Plowe et al. 1995). The methods are suited foruse on large numbers of samples in a laboratory ina malaria-endemic country and have major advan-


tages on in vitro tests that require parasite cultivation and take days to perfonn (Plowe et al. 1996).These molecular tools are based on the detectionby PCR of point mutations in the parasite genesresponsible for in vitro resistance. Presently, several mutations associated with resistance to SP havebeen identified while the role of those linked tochloroquine resistance is more controversial. Thelatter has been associated to Pfmdr 1 genepolymorphysm, although multiple mutations indifferent genes are probably required for clinicalresistance (Foote et al. 1990). Wellems et al. (1990)have linked chloroquine resistance to a single genetic locus yet to be identified. The point mutations linked to SP resistance have been observedin the parasite genes encoding for dihydrofolatereductase (DHFR) and dihydropteroate synthetase(DHPS), the targets for pyrimethamine andsulfadoxine respectively (Wang et al. 1995). Aserine in position 108 of the DHFR gene is linkedto in vitro sensitivity to both pyrimethamine andcycloguanil. A mutation to asparigine at position108 seems to be the key mutation for conferring invitro pyrimethamine resistance (de Pecoulas et al.1996), although a genotype without this mutationhas been recently described (Wang et al. 1997).An asparagine to isoleucine change at position 51and a cysteine to arginine at position 59 appear tomodulate higher levels of in vitro pyrimethamineresistance when they occur with the asparagine108 mutation, and an isoleucine to leucine mutation at position 164 in combination with the asparagine-I08 and arginine-59 mutations has beenfound in P. falciparum lines that are highly resistant to both pyrimethamine and cycloguanil (Bascoet al. 1995, Reeder et al. 1996). Point mutations ofthe DHPS gene have been less extensively studied. Thirteen variants over the wild type have beenidentified so far in samples from different countries, the most common alteration being in position 437 (Wang et al. 1997). A certain amount ofcorrelation was found between the prevalence ofknown DHFR and DHPS mutations and increasing levels of in vivo SP resistance in four differentcountries: Mali, Kenya, Malawi, Bolivia (Ploweet al. 1997). Different degrees of in vivo parasitological resistance and clinical failure to SP mightbe due to the progressive accumulation of OHFRand DHPS mutations (Plowe et al. 1997). However, this still needs to be established by carryingout prospective studies on the relationship betweenparasite genotype and clinical outcome in individual infections treated with SP. No relation between clinical SP resistance and mutations in theDHFR gene couId be established in a study carried out in Tanzania. Although ail the isolatesshowed a point mutation in at least one codon of

the OHFR gene, only 43% of the children had detectable parasitaemia seven days after treatment(Jelinek et al. 1997).

The characterisation of the two polymorphiemerozoite surface antigens, MSP 1 and MSP2 hasalso been used to establish whether a parasitaemiaobserved after treatment is caused by a recrudescence of drug-resistant parasites or by a new infection (Babiker et al. 1994, AI-Yaman et al. 1997).This couId be an important infonnation when carrying out in vivo tests, particularly in areas with aconsiderable amount of transmission where, aftera certain time, it is impossible to distinguish between recrudescence and new infection. However,it is unlikely that national malaria controlprogrammes will routinely use such methods.

In vitro test results have been used as the goldenstandard to establish the link between resistance toSP and point mutations in the DHFR and DHPSgenes. Therefore, molecular tests identifying mutant parasites might replace in vitro tests, as theyare easier to perfonn and can be carried out on alarger number of samples. They will complementthe information on treatment failure and might actually identify areas where a given drug is likely tobe less efficacious. However, for antimalarial drugsother than SP, these tests are not available yet anddrug susceptibility ofdifferent parasite populationswill continue to be established by in vitro tests.Decisions on the national drug policy will continueto be based on the results of in vivo tests as thesereflect more closely the therapeutic efficacy of agiven drug.

REFERENCES

AI-Yaman F, Genton B, Reeder JC, Anders RF, AlpersMP 1997. Evidence that recurrent Plasmodiumfalciparum infection is caused by recrudescence ofresistant parasites. Am J Trop Med Hyg 56: 436-439.

Babiker H, Randford-Cartwright L, Sultan A, Satti G,Walliker D 1994. Genctic evidence that RIcholoroquine resistance of Plasmodium falciparumis caused by recrudescence of resistant parasites.Trans R Soc Trop Med Hyg 88: 328-331.

Basco LK, de Pecoulas PE. Wilson CM, Le Bras J,Mazabraud A 1995. Point mutations in thedihydrofolate reductase-thymidylate synthase geneand pyrimethamine and cycloguanil resistance inPlasmodiumfalciparum. Mol Biochem Parasitol69:135-138.

Baudon D 1995. La prise de décision dans la politiqued'utilisation des antipaludiques en réponse à unemodification de l'efficacité de la chloroquine.Réflexions appliquées à l'Afrique. Méd Trop 55:37S-40S.

Bloland PB, Lackritz EM, Kazembe PN, Were JBO,Steketee R, Campbell CC 1993. Beyond chloroquine:implications of drug resistance for cvaluating malaria therapy efficacy and treatment policy in Af-

630 Drug Resistance Surveillance and Molecular Methods • Umberto D'Alessandro

rica. J Infect DIS 167: 932-937.D'Alessandro U, Mutabingwa TK, Ouma JH, Watkins

WM 1997. Report of the workshop on monitoringantimalarial treatment efficacy: an East African regional network. Malaria Consortium, UK.

De Pecoulas PE, Basco LK, Le Bras J, Mazabraud A1996. Association between antifol resistance in vitroand DHFR gene point mutation in Plasmodium.Iàlciparllm isolates. Trans R Soc Trop Med Hyg 90:181-182.

Draper CC, Hills M, Kilimali VAEB, Brubaker G 1988.Seriai studies on the evolution of drug resistance inmalaria in an area ofEast Africa: findings from 1979up to 1986. J Trop Med Hyg 9 f: 265-273.

Foote SJ, Kyle DE, Martin RK, Oduola AM, Forsyth K,Kemp DJ, Cowman AF 1990. Several alleles of themultidrug-resistance gene are closely linked to chloroquine resistance in Plasmodium falciparum. Nature 345: 255-258.

Jelinek T, Ronn AM, Curtis J, Duraisingh MT, LemngeMM, Minha J, Bygbjerg IC, Warhurst DC 1997. Highprevalence of mutations in the dihydrofolate reductase gene ofPlasmodiumfalciparum in isolates fromTanzania without evidence ofan association to clinical sulfadoxine/pyrimethamine resistance. Trop Medfntern Hlth 2: 1075-1079.

Plowe CV, Djimde A, Bouare M, Doumbo 0, WellensTE 1995. Pyrimethamine and proguanil resistanceconferring mutations in Plasmodium falciparumdihydrofolate reductase: polymerase chain reactionmethods for surveillance in Africa. Am J Trop MedHvg 52: 565-568.

Plowe CV Djimde A. Wellems TE, Diop S, Kouriba B,Doumbo OK 1996. Community pyrimethaminesulfadoxine use and prevalence ofresistant Plasmodium falciparum genotypes in Mali: a model fordeterring resistance. Am J Trop Med Hyg 55: 467471.

Plowe CV, Cortese JF, Djimde A, Nwanyanwu OC,Watkins WM, Winstanley PA, Estrada-Franco JG,Mollinedo RE, Avila JC, Cespedes JL, Carter D,Doumbo OK 1997. Mutations in Plasmodiumfalciparum dihydrofolate reductase anddihydropteroate synthase and epidemiological patterns of pyrimethamine-sulfadoxine use and resistance. J Infect Dis f 76: 1590-1596.

Prasad RN, Prasad H, Virk KJ, Sharma VP 1990. Application of a simplified in-vivo test system for determining chloroquine resistance in Plasmodiumfalciparllm. Bulletin WHO 68: 755-758.

Reeder JC, Rieckmann KH, Genton B, Lorry K, WinesB, Cowman AF 1996. Point mutations in the

dihydrofolate reductase and dihydropteroate synthetase genes and in vitro susceptibility to pyrimethamine and cycloguanil of Plasmodiumlàlciparum isolates from Papua New Guinea. Am JTrop Med Hyg 55: 209-213.

Robert V, Molez JF, Trape JF 1996. Short report: gametocytes, chloroquine pressure, and the relative parasite survival advantage of resistant strains offalciparum malaria in West Africa. AmerJ Trop MedHyg 55: 350-351.

Wang P, Brooks DR, Sims PFG, Hyde JE 1995. A mutation-specifie PCR system to detect sequence variation in the dihydropteroate synthetase gene of Plasmodiumfalciparum. Mol Biochem Parasitol 7f: 115125.

Wang P, Lee CS, Bayoumi R, Djimde A, Doumbo 0,Swedberg G, Dao LD, Mshinda H, Tanner M,Watkins WM, Sims PFG, Hyde JE 1997. Resistanceto antifolates in Plasmodium falciparum monitoredby sequence analysis of dihydropteroate synthetaseand dihydrofolate reductase alleles in a large number of field samples ofdiverse origins. Mol BiochemParasitol89: 161-177.

Watkins WM, Mberu EK, Winstanley PA, Plowe CV1997. The efficacy ofantifolate antimalarial combinations in Africa: a predictive model based on pharmacodynamie and pharmacokinetic analyses.ParasitaI Today f 3: 459-464.

Wellems TE, Panton LJ, Gluzman IY, do Rosario VE,Gwadz RW, Walker-Jonah A, Krogstad DJ 1990.Chloroquine resistance not linked to mdr-like genesin a Plasmodium falciparum cross. Nature 345: 253255.

Wernsdorfer WH, Payne D 1988. Drug sensitivity testsin malaria parasites, p. 1765-1794. In WHWernsdorfer, 1McGregor (eds), Malaria - Principlesand Practice olMalariology, Churchill Livingstone,Edinbourg, London, Melbourn, New York.

Wernsdorfer WH 1991. The development and spread ofdrug-resistant malaria. Parasitol Today f f: 297-303.

WHO - World Health Organization 1990. Practical chemotherapy of malaria. Technical Report Series 805.WHO, Geneva.

WHO - World Health Organization 1993. Implementation ofthe global malaria control strategy. Report ofa WHO Study Group on the implementation of theGlobal Plan of Action for Malaria Control 19932000. WHO. Geneva.

WHO - World Health Organization 1996. Assessmentoftherapeutic efficacy of antimalarial drugs for uncomplicatedfalcipal1lm malaria in areas with intensetransmission. WHO/MALl96. 1077.


Allelic Diversity at the Merozoite Surface Protein-1 (MSP-l)Locus in Natural Plasmodium falciparum Populations:

a 8rief Overview

Marcelo U Ferreira/+, Osamu Kaneko*, Masatsugu Kimura**, Qing Liu ***,Fumihiko Kawamoto***, Kazuyuki Tanabe****

Departamento de Parasitologia, lCB, Universidade de Sào Paulo, Av. Prof. Lineu Prestes 1374,05508-900 SâoPaulo, SP, Brasil *Department of Medical Zoology **Laboratory of Biophysics, Osaka City University MedicalSchool, Osaka, Japan ***Department oflntemational Health, Nagoya University School of MedIcine, Nagoya,

Japan ****Laboratory of Biology, Osaka Institute of Technology, Osaka, Japan

The merozoite surface protein-I (MSP-I) locus of Plasmodium falcipamm codes for a major asexualblood-stage antigen currently proposed as a major malaria vaccine candidate. The protein, however,shows extensive polymorphism, which may compromise its use in sub-unit vaccines. Here we comparethe patterns ofallelic diversity at the MSP-I locus in wild isolales from three epidemiologically distinctmalaria-endemic areas: the hypoendemic southwestern Brazilian Amazon (n =54), the mesoendemicsouthern Vietnam (n =238) and the holoendemic northern Tanzania (n =79). Fragments o/the variableblocks 2, 4a, 4b and 6 or 10 ofthis single-copy gene were ampl(fied by the polymerase chain reaction,and 24 MSP-I gene types were defined as unique combinations ofalleUc types in each variable block.Ten d(fferent MSP-I types were identified in Brazil, 23 in Vietnam and 13 in Tanzania. The proportion ofgenetically mixed infections (isolates with parasites carrying more than one MSP-I version) rangedfrom 39% in Brazil to 44% in Vietnam and 60% in Tanzania. The vast majority (90%) of the typedparasite populations /rom Brazil and Tanzania belonged to the same seven most frequent MSP-I genetypes. In contrast, these seven gene types corresponded to only 61% o/the typed parasite populationsfrom Vietnam. Non-random associations were /ound between allelic types in blocks 4a and 6 amongVietnamese isolates, the same pattern being observed in independent studies performed in 1994, 1995and 1996. These results suggest that MSP-1 is under selective pressure in the local parasite population.Nevertheless, the finding that similar MSP-I type frequencies were found in 1994 and 1996 arguesagainst the prominence ofshort-termfrequency-dependent immune selectionl?fMSP-1 polymorphisms.Non-random associations between MSP-I alleUc types, however, were not detected among isolatesfromBrazil and Tanzania. A preUminary analysis of the distribution of MSP-I gene types per host amongisolatesfrom Tanzania, but not among those/rom Brazil and Vietnam, shows significant deviation/romthat expected under the null hypothesis ofindependent distribution l?f parasites carrying difJerent genetypes in the human hosts. Some epidemiological consequences ofthesefindings are discussed.

Key words: Plasmodium falciparum - malaria - allelic diversity - merozoite surface protein-l - populationgenetics - vaccine candidate

Supported by grants from the Ministry of Education,Science, Sports and Culture of Japan, Toyota Foundation, the Program for Malaria Control in the Amazon ofthe National Health Foundation (Brazilian Ministry ofHealth), Fundaçào de Amparo à Pesquisa do Estado deSâo Paulo, Brazil, and the UNDP/World Bank/WorldHealth Organization Special Programme for Researchand Training in Tropical Diseases. MUF was supportedby a research student scholarship from the Ministry ofEducation, Science, Sports and Culture of Japan.+Corresponding author: Fax: +55-11-818.7417. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

The polymorphie merozoite surface protein-l(MSP-l) of Plasmodium falciparum is a majorasexual blood-stage malaria vaccine candidate(Holder 1996). Comparisons of nucleotide sequences led to the identification of seven variableblocks in the gene, which are interspersed with fiveconserved and five semi-conserved blocks (Fig. 1).There are essentially two versions of each block,named after the representative isolates MAD2ü andKI (Tanabe et al. 1987). A major exception to thisdimorphic mIe is the variable block 2, that has athird version originally described in the isolateRD33 (Certa et al. 1987). Most allelic diversity isgenerated by intragenic recombination between

632 Genetic Diversity in Plasmodium falciparum • Marcelo U Ferreira et al.

these representative sequences at the 5' end of thegene, within blocks 3, 4 and 5. Minor differencesalso exist between homologous versions of thesame variable block, and nucleotide substitutions(most ofwhich are dimorphic) occur in semi-conserved and conserved blocks (Tanabe et al. 1987).

Major MSP-/ gene types may be defined asunique combinations of: (a) one of three versionsofblock 2 (MAD20, KI or Rü33), (b) one offourpossible versions ofblock 4, because recombination within this region generates MAD20/K 1 andKIIMAD20 hybrids in addition to the 'pure' allelic types MAD20 and KI (Conway et al. 1991 b,Kaneko et al. 1996), and (c) one of two versions(MAD20 or KI) of the segment between the variable blocks 6 and 16, that comprises about 60% ofthe gene. Recombination events have not been described in this portion of the gene (Tanabe et al.1987, 1989, Peterson et al. 1988, Conway et al.1991b, Jongwutiwes et al. 1991, Kaneko et al.1996, 1997). Therefore, the 24 MSP-/ gene typesshown in Table 1may theoretically be observed innatural parasite populations (Kaneko et al. 1997).

The extent of allelic diversity in different malaria-endemic areas should be evaluated ifthe variable domains of MSP-l, that are highly inununogenic (Holder & Riley 1996), are to be included insubunit malaria vaccines. A novel polymerasechain reaction (PCR)-based strategy was recentlydeveloped to group c1inical isolates ofP.falciparuminto the 24 MSP-l gene types defined in Table 1(Kaneko et al. 1997). This strategy has been successfully applied to type wild isolates from themesoendemic southern Vietnam (Kaneko et al.1997, Ferreira et al. 1998b), the hypoendemic Brazilian Amazon (Ferreira et al. 1998a), and theholoendemic Tanzania (Ferreira et al. 1998c). Inthis communication we analyze available data regarding complete MSP-l typing of isolates fromthese three malaria-endemic areas.


Table Il summarizes basic infonnation regarding typed P fa/ciparum isolates in each malariaendemic area. Genomic DNA was extracted directly from the blood of P. falciparum-infectedpatients, without previous in vitro cultivation ofparasites. Locations of the oligonucleotide primers are shown in Fig. 1. Primer sequences and PCRprotocols are given elsewhere (Kaneko et al. 1997).The basic PCR-based typing procedure developedby Kaneko et al. (1997) may be described as it follows:

First step - Block 2 was typed in three separatereactions with the allelic specific forward primersM2F, K2F and RlF and the common reverse primerC3R.

TABLE 1

Mero::oite surface protein-l (MS?-l) gene typesdefined as unique combinations of allelic types in each

variable block

Variable black

Gene typeU 2 4a 4b 10

1 KI KI KI KI2 MAD20 KI KI KI3 R033 KI KI KI4 KI MAD20 KI KI5 MAD20 MAD20 KI KI6 R033 MAD20 KI KI7 KI KI MAD20 KI8 MAD20 KI MAD20 KI9 Rü33 KI MAD20 KI10 KI MAD20 MAD20 KIII MAD20 MAD20 MAD20 KI12 R033 MAD20 MAD20 KI13 KI KI KI MAD2014 MAD20 KI KI MAD2015 R033 KI KI MAD2016 KI MAD20 KI MAD2017 MAD20 MAD20 KI MAD2018 R033 MAD20 KI MAD2019 KI KI MAD20 MAD2020 MAD20 KI MAD20 MAD2021 R033 KI MAD20 MAD2022 KI MAD20 MAD20 MAD2023 MAD20 MAD20 MAD20 MAD2024 Rü33 MAD20 MAD20 MAD20

a: each gene type is defined as a unique combination ofallelic types detected in the variable blocks 2, 4a (5'segment ofblock 4), 4b (3' segment ofblock 4) and 6-16 of the MS?-l gene. Since there is no recombinationat the central and C-terminal portions of this gene, theallelic type detected in black lOis considered ta be thesame for the variable blacks 6, 8, 14 and 16. Allelic typesare named after the reference isolates MAD20, KI andR033.

Second step - The gene fragments between theconserved block 5 and the variable block 6 wereamplified in two separate reactions with the common forward primer C5F and the type-specific reverse primers M6R or K6R. Altematively, block10 was typed with the semi-conserved forwardprimer C9F and the type-specific reverse primersMI OR and KI OR. Since there is no recombinationbetween blocks 6 and 16, the allelic type found inblocks 6 or lOis the same for variable blocks 8, 14and 16.

Third step - Segments between blocks 2 and 6were amplified in three separate reactions with thetype-specific forward primers M2F, K2F or RlF,and the type-specific reverse primers M6R or K6R.The PCR fragments amplified in this step were usedas template in the next step. As an alternative, this


TABLE II

Recent polymerase chain reaction-based studies involving complete typing of the merozoite surface protein-lgene in natural Plasmodiumfàlciparum populations

Malaria No. of typedArea endemicity isolates Reference

Brazilian AmazonSouthem VietnamSouthem VietnamNorthem Tanzania

LowIntermediateIntennediateHigh

5413610279

Ferreira et al. 1998aKaneko et al. 1997Ferreira et al. 1998bFerreira et al. 1998c

a

b

Block

108

-'1-K8R

64b_CGR __

'1- ~ .-- .--M4R L-----.J 118R [2'2'2J M10R-.C8F _

.-Ki OR-.--K4R

2 3 4a- C3R --~ '1- -.M2FI 1 M4F

-~... C3F --~ -~K2F K4F

~-~R2F

MAD2D-type

K1-type

R033-type

Fig. I-a: structure of the mero=oite surfàce protein-I gene of Pll1snlodlllm fl1Jc,parunI. Conserved, seml-conservcd and variableblocks of the gene are shown as open, hatched and c10sed boxes, respectively. Block numbers arc after Tanabe et al. (1987): b:locations and directions of the oligonucleotide primers used to type blocks 2, 4a, 4b, 6, and 10 are also indicated. Rcdrawn fromKaneko et al. 1997.

template may be prepared with the conserved forward primer C3F and the conserved reverse primerC5R.

Forth step - Block 4 was typed by nested PCRin four separate reactions with the type-specifieforward primers M4F or K4F and type-specifiereverse primers M4R or K4R. As an alternative,the first step may be eliminated, and block 2 maybe typed by detecting allelic-specific fragments inthe second step (Ferreira et al. 1998b).

The detection ofPCR products in the expectedsize ranges after 1.5-2% agarose gel electrophoresis defined the presence of each allelic type inblocks 2, 6 or 10, 4a and 4b. As MS?-] is a singlecopy gene in the haploid genome of blood-stageparasites, we consider that isolates harboring morethan one gene type have mixed infections withgenetically distinct P. falciparom subpopulations.Each subpopulation may be separately typed bythis approach (Kaneko et al. 1997).

RESULTS AND DISCUSSION

Are al! theoretical!y possible MSP-l gene typesfound in natural Plasmodium falciparum populations? - As shown in Fig. 2, ail but one of the 24possible MS?-] gene types were detected in

mesoendemic Vietnam. Only gene type Il wasabsent in that area at both occasions (Kaneko etal. 1997, Ferreira et al. 1998b). This suggests thatalmost aIl possible combinations ofMS?-] allehctypes may be found in parasites that are able toinfect human hosts. In contrast, only 10 and 13MS?-] types were found in hypoendemic Braziland holoendemic Tanzania, respectively. Moreover,essentially the same MS?-] gene types were foundto predominate in both countries, and about 90%of the typed parasite populations belonged to theseven most common gene types, namely the types13, 16, 17, 18,22,23 and 24 as defined in Table 1(Ferre ira et al. 1998c). Nevertheless, these sevengene types were found in only 61% of the typedparasite populations in Vietnam.

Is there any association between the extent ofMSP-l diversity and the intensity ofmalaria transmission? - If we compare the proportions of genetically mixed infections (that is, patients harboring more than one MS?-] gene type) and the average number of MS?-] gene types found per patient, an apparent positive association is found between malaria endemicity and MS?-] diversity(Table III). However, if we compare the numberof different MS?-] gene types found in each en-

634 Genetic Diversity in Plasmodium falclparum • Marcelo U Ferreira et al.

Fig. 2-a: frequency distribution of the merozoite suiface prolein-/ (MSP-I) gene types in 54 Plasmodiumfalciparum isolates collected in July 1995 in the cIty of Porto Velho, State ofRondônta, southwestern Brazilian Amazon (Ferreira et al.1998a): b: frequency distribution of the MSP-/ gene types in79 P falciparum isolates collected between July and September 1996 in the city ofTanga and the nearby village ofPangani,in northem Tanzanta (Ferreira et al. 1998c); c: frequency distribution of the MSP-J gene types in 238 Pfa/cipal1lm isolatescollected between July 1994 and July 1996 trom malaria palients belonging to the ethnic majority Kinh and the minonlyK'ho living in the towns ofBao Loc and Phu Rieng and nearbyareas in southem Vietnam (Kaneko et al. 1997, Ferrelra el al.1998b). The 24 MSP-/ gene types are numbered as in Table 1.

demic area, no such association can be detected.Therefore, despite the fact that most infected hostsin Tanzania carry two or more parasite clones whichmay be ingested by the vector and recombine during meiosis, the resulting repertoire ofMSP-1 variants seems to be relatively restricted in humanhosts, ifcompared with the situation found in Vietnam. Strong selective pressure related to the sequential use of several different antimalarials in afew years, in the context of multi-drug resistance,may have resulted in increased genetic diversityof P falciparum populations in Vietnam.

Are the patterns ~fMSP-I diversity temporal/ystable in a given malaria-endemic area? - Fig. 3compares the distribution of MSP-/ gene types inparasite populations sampled in the same communities in southem Vietnam at intervals of 12 months(Fig. 3a) and 18-24 months (Fig. 3b). There is nosignificant difference when both pairs offrequencydistributions are compared. The stability in the frequencies ofMSP-/ gene types over periods of 1224 months does not imply that long-term changescan be ruled out. Under the present conditions ofmalaria transmission in southem Vietnam, just afew infections with parasites carrying distinct versions of the MSP-I antigen are expected per hostat a one-year or two-year interval. As a consequence, natural acquisition of effective anti-MSP1 immunity may occur at a rather slow rate. Wehave now examined this issue in relation to theBrazilian Amazon by typing MSP- J variable blocksin Pfalciparum isolates collected over a period of12 years (LA Silveira & MU Ferreira, unpublisheddata).

Are there non-random associations betweenMSP-l variable bloc/cs in natural parasite populations? - If intragenic recombination occurs frequently at the MSP- J locus in the absence of major selective constraints, the distribution of MSPJ gene types would be described by a simple probability model analogous to those used in population genetics to estimate expected frequencies ofmultiple-locus genotypes (Tibayrenc 1995). Forinstance, the expected frequency of gene type 1(as defined in Table 1) is given by multiplying the

(c)

(b)

(a)

5

10

25

20

15

o -N~~~~~~~O-N~~~~~~~O_~~----------NNNNN

35

o -N~~~~~~~O-N~~~~~~O-N~~__________C'lNNNN

M5P-l gene type

30

15

10

5

35

25

20

15

30

25

30

20

35

~co

'Z;

(5Co0:

TABLE III

The extent of allelic diversity at the merozoite surface protein-l (MSP-/) locus in natural Plasmodium {alciparumpopulations from areas with different levels ofmalaria endemicity

AreaMalariaendemicity

No. of MSP-/ Proportion (%)gene types of isolates withdetected by PCR > 1 MSP-/ type

Average no. ofMSP-/ typesper patient

BrazilVietnamTanzania

LowIntermediateHigh

10 3923 4413 60

1.421.762.37

c:o

'-2o0..e

Cl...

30

20

10

K' ho people 1994-95

(a)

Kinh people 1994-96


lies (either MAD20 or KI) in blocks 4a and 6 or10 were found more frequently than expected(Kaneko et al. 1997, Ferreira et al. 1998b). Thereasons why similar results were not found inholoendemic Tanzania and hypoendemic Brazilremain to be elucidated.

Non-random associations between allelic typesmay result from: (a) geographic isolation leadingto random genetic drift, (b) limited chances for intragenic recombination during meiosis in the mosquito vector due to the presence of few differentMS?-] versions and the low prevalence ofmixedinfections in human hosts, and (c) biological constrains which bias for particular associations. As

30

25

20

15

10

(b)35

30

25

20

15

(a)

Brazil1995

M5P-1 gene type

10

5

o ':'-~N~",:-:...~",~","",...~"':-:en:-:o=-:-:...~...~ '" "' .... '" en 0 _ N'" ...-- __ NNNNr-I

Fig. 4-a: expected (c1osed bars) and observed (striped bars) frequencies of melV=oite sU/face prolein-/ (MS?-/) gene types mPorto Velho, southwestem Brazilian Amazon (n = 54) (data FromFerreira et al. 1998a); b: expected (c1osed bars) and observed(striped bars) frequencies of MS?-/ gene types in Tanga andPangani, northem Tanzania (Il = 79) (data From Ferreira et al.1998c). Expected frequencies were computed under the nuithypothesis of random associations of allelic types in variableblocks of the gene (see the text for details). There is no significant ditference between expected and observed frequencies byX2 tests of goodness of fit.

Fig. 3-a: frequency distribution of merozoite surface prolein/(MS?-I) gene types in isolates From K'ho people living in hillareas surrounding Bao Loc, southem Vietnam, collected between July-August 1994 (n '" 34) (c1osed bars) and in August1995 (n =28) (striped bars) (redrawn From Kaneko et al. 1997);b: frequency distribution ofMS?-/ gene types in isolates FromKinh people living in the town ofBao Loc, southem Vietnam,collected between July-August 1994 (n = 44) (c1osed bars) andbetween January-July J996 (n = 95) (striped bars) (redrawn rromFerreira et al. 1998b).

observed frequencies of the allelic type KI inblocks 2, 4a, 4b and 6-16 in a given population.Fig. 4 shows expected frequencies ofMS?-] genetypes under the null hypothesis of random association of allelic types (MAD20, KI or RD33) ineach variable block in Brazil and Tanzania. No significant difference between expected and observedfrequencies was detected by the X? test for goodness of fit in both cases (Ferreira et al. 1998a, c).In contrast, significant differences between expected and observed frequencies of MS?-] genetypes were found in two surveys in southern Vietnam (Fig. 5). Non-randomassociations were foundto occur, in both cases, between blocks 4a and 616: MS?-] gene types with concordant allelic fami-

c:o

'-2o0..o0:

Tanzania 1996

30

(b)

20

1.5

10

O_N~.~~~~~O_N~~~~~~~O_N~~--- NNNNN

M5P-1 gene type


Vietnam 1994-95

FIg. 5-a: expected (closed bars) and observed (stnped bars) frequencies of mem=oite surface protein-/ (MS?-/) gene types inIsolates from Bao Loc, Vietnam, coliected between July-August 1994 from both Kinh and K'ho people (n = 108) (datafrom Kaneko et al. 1997); b: expected (closed bars) and observed (striped bars) frequencies of MS?-/ gene types 10 isolates from Bao Loc, Vietnam, coliected between January-July1996 from both K'ho and Kinh people (II = 102) (data fromFerreira et al. 1998b). Expected frequencles were computedunder the null hypothesls ofrandom associations ofalle1ic typesin variable blocks of the gene (see the text for details). ln bothcases slgnificant differences between expected and observedfrequencies were detected by X2 tests of goodness of tit.

discussed elsewhere, the first two possible explanations do not match available data from Vietnam,and speculations regarding the third hypothesis arelimited by the fact that the function of MSP-f remains unknown (Ferreira et al. 1998b).

Are parasite populations carrying d~fferent

MSP-l gene types independently distributed in thehost population? - A basic assumption of recentmathematical models ofmalaria transmission is thatinfections by different 'strains' are independent.This means that, in genetically mixed infections, apatient infected by a parasite carrying a given MSPf gene type (for instance type 1) is as likely to beco-infected with a given second type (for instancetype 2) as someone infected with any other MSP-ftype. This does not take into account the possibilities of: (a) frequent multiple-clone infections byvectors carrying two or more gene types includingrecombinant gene types resulting from the unionoftwo different clones from one previous host (Hill& Babiker 1995), and (b) either facilitation or competition between parasites carrying different versions ofa polymorphic antigen which co-infect thesame host (Gilbert et al. 1998). We applied here asimple statistical analysis to test this assumption.

The expected distribution ofMSP-f gene typesper host under the hypothesis of independent distribution of MSP-f gene types may be describedas the sum ofN independent binomial distributions,where N is the number of different MSP-f genetypes observed in host population. The varianceofthis summed binomial distribution, or expectedvariance a 2, was calculated and compared to theobserved variance s 2 using a X2 test as described(Lotz & Font 1991). The difference between theexpected and observed variances was statisticallysignificant in Tanzania (Table IV), suggesting thatthe MSP-f gene types are not independently distributed in the host population. Therefore, we tested21 possible pairwise associations between genetypes in 2 x 2 contingency tables using either standard X2 or Fisher's exact tests when appropriate,with the significance level adjusted for multiple

Vietnam 1996

M5P-1 gene type

(a)

10

15

2:1

20

.30

~t:0

'-e0Cl..0et .30

25(b)

20

15

10

TABLE IV

Statistical comparison of expected (a2) and observed (s2) variances of the distribution of merozoite surfaceprotein-l (MSP-l) gene types among human hosts living in areas with difTerent levels of malaria endemicity

AreaMalaria Expected Observed X2

endemicity variance (a2) variance (s2) (d. f.)a P

Brazil

Vietnam

Tanzania

Low 1.04 0.32 16.45 (53)

Interrnediate 1.35 1.42 105.65 (101)

High 1.59 2.10 103.08 (78)

> 0.05

> 0.05

< 0.05

a: degrees of freedom.


comparisons with the Bonferroni's correction (Lordet al. 1997). At least 1 pair of MSP-! gene types( 18 and 24) was found to be positively associated(P = 0.0008 by X2 test). With the Bonferroni'scorrection applied to these data, an association isstatistically significant at the 5% level ifP <0.0024.This means that the genetically similar types 18and 24, that differ only in the block 4b allelic type(Table I), tend to co-occur more frequently thanexpected under the null hypotheses that they areindependently transmitted. Nevertheless, departures from the null hypothesis of independent transmission were not detected in areas of lower endemicity such as Brazil and Vietnam (Table IV).

Mathematical models have recently regardedmalaria as a heterogeneous disease caused by several independently transmitted and antigenicallydistinct parasite subpopulations or 'strains' thatdo not interact wi thin the human hosts and are ableto elicit 'strain' -specifie protective irnmunity. Thesemodels estimate the basic reproduction number Roof malaria, defined as the average number of secondary infections generated by one primary infection in a fully susceptible population, as a weightedaverage of Ro values for each 'strain'. This estimate is substantially lower than R0 values obtainedby conventional methods, suggesting that malariaeradication in Africa may be quite feasible (Guptaet al. 1994). Nevertheless, the finding that genetically and antigenically distinct parasite populationsare not independently distributed in the humanhosts in areas ofhigh endemicity, such as northernTanzania (Ferreira et al. 1998c) and the Gambia(Conway et al. 1991a), implies Ro values considerably higher than those provided by the weightedaverage approach (Lord et al. 1997).

In conclusion, this study provides examples ofthe use of simple molecular and statistical approaches to investigate the extent of antigenic diversity in malaria parasites and to test hypothesesregarding the patterns of transmission and interaction ofgenetically distinct parasite subpopulationsin endemic areas.

ACKOWLEDGMENTS

To HV Thien (Lam Dong II Provincial Hospital, BaoLoc, Vietnam), BT Ndawi (Primary Health Care Institute, Iringa, Tanzania), M Zhou and S Isomura (NagoyaUniversity School of Medicine, Nagoya, Japan), AMKatzin and EAS Kimura (Department of Parasitology,University of Sào Paulo, Brazil).

REFERENCES

Certa U, Rotmann D, Matile H, Reber-Liske RA 1987. Anaturally occuring gene encoding the major surfaceantigen precursor pl90 of Plasmodium falciparumlacks tripeptide repeats. EMBO J 6: 4137-4142.

Conway DJ, Greenwood BM, McBride JS 1991a. Theepidemiology of multiple-clone Plasmodiumfalciparum infections in Gambian patients. Parasitology 103: 1-6.

Conway DJ, Rosario V, Oduola AMJ, Salako AL, Greenwood BM, McBride JS 1991 b. Plasmodium{alciparum: intragenic recombination and nonrandom associations between polymorphlc domains ofthe precursor to the major surface antigens. ExpParasitol 73: 469-480.

Ferreira MU, Liu Q, Kaneko 0, Kimura M, Tanabe K,Kimura EAS, Katzin AM, Isomura S, Kawamoto F1998a. Allelic diversity at the mero;:oite slllface protein-I locus of Plasmodium falciparum in clinicalisolates from the southwestern Brazilian Amazon.Am J Trop Med Hyg, in press.

Ferreira MU, Liu Q, Zhou M, Kaneko 0, Kimura M,Thien HV, Isomura S, Tanabe K, Kawamoto F1998b. Stable patterns ofallelic diversity at the merozoite surface protein-I locus of Plasmodiumfalciparum in clinical isolates from southern Vietnam. JEuk Microbiol45: 131-136.

Ferreira MU, Liu Q, Kimura M, Ndawi BT, Tanabe K,Kawamoto F 1998c. Allelic diversity in the merozoite surface protein-I and epidemiology ofmultipleclone Plasmodium falciparum infections in northern Tanzania. J Parasitol, in press.

Gilbert SC, Plebanski M, Gupta S, Morris J, Cox M,Aidoo M, Kwiatkowski D, Greenwood BM, WhittleHC, Hill ASV 1998. Association ofmalaria parasitepopulation structure, HLA, and immunological antagonism. Science 279: 1173-1177.

Gupta S, Trenholme K, Anderson RM, Day KP 1994.Antigenic diversity and transmission dynamics ofPlasmodium falciparum. Science 263: 961-963.

Hill WG, Babiker HA 1995. Estimation of numbers ofmalaria clones in blood samples. Proc R Soc London B 262: 249-257.

Holder AA 1996. Preventing merozoite invasion oferythrocytes, p. 77-104. In SL Hoffman, Malaria Vaccine Development. A Multi-immune Response Approach, ASM Press, Washington D.C.

Holder AA, Riley EM 1996. Human immune responseto MSP-I. Parasitol Today 12: 173-174.

Jongwutiwes S, Tanabe K, Nakazawa S, Uemura H,Kanbara H 1991. Coexistence of gp 195 alleles ofPlasmodiumfalciparum in a small endemic area. AmJ Trop Med Hyg 44: 299-305.

Kaneko 0, Jongwutiwes S, Kimura M, Kanbara H, IshiiA, Tanabe K 1996. Plasmodium falciparum: variation in block 4 of the precursor to the major surfaceproteins (MSPI) in natural populations. ExpParasitol 84: 92-95.

Kaneko 0, Kimura M, Kawamoto F, Ferreira MU,Tanabe K 1997. Plasmodium falciparum: allelicvariation in the merozoite surface protein 1 in wildisolates from southern Vietnam. Exp Parasitol 86:45-57.

Lord CC, Woolhouse MEJ, Barnard BJH 1997. Transmission and distribution of virus serotypes: Africanhorse sickness in zebra. Epidemiol Infect 118: 43-50.


Lotz lM, Font WF 1991. The role of positive and negative interspecific associations in the organization ofcommunities of intestinal helminths ofbats. Parasitology /03:127-138.

Peterson MG, Coppel RL, Moloney MB, Kemp Dl 1988.Third forrn of the precursor to the major surface antigens of Plasmodiumfalciparum. Mol Ce/l Biol 8:2664-2667.

Tanabe K, Murakami K, Doi S 1989. Plasmodium

falciparum: dimorphism of the p 190 alleles. ExpParasitai 68: 470-403.

Tanabe K. Mackay M. Goman M, Scaiffe lG 1987. AIlelic dimorphism in a surface antigen genc of themalaria parasite Plasmodium .làlciparum. J MolBiolol/95: 273-287.

Tibayrenc M 1995. Population genetics ofparasitic protozoa and other microorganisms. Adv Parasitai 36:47-115.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(S}: 639-646, 5ep./Oct. 1998 639

Evaluation of DNA Recombinant Methodologies for theDiagnosis of Plasmodium falciparum and their Comparison

with the Microscopy Assay

L Urdaneta/*, P Guevara/+, JL Ramirez*

Escuela de Malariologia y Saneamiento Ambiental "Dr. Arnoldo Gabaldon", Maracay,Venezuela*Grupo de Genética Molecular, lnstituto de Biologia Experimental, Universidad Central de Venezuela,

Caracas,Venezuela

Since 1984, DNA tests based on the highly repeated subtelomeric sequences of Plasmodium falciparum(rep 20) have been frequently used in malaria diagnosis. Rep 20 is very specific for this parasite, and ismade of21 bp units, organized in repeated blocks with direct and inverted orientation. Based in thispm'ticular organization, we selected a unique consensus oligonucleotide (p{-21) to drive a PCR reaction coupled to hybridization to non-radioactive labeled probes. The pf-21 unique oligo PCR (pf-21-1)assay produced DNA amplificationjingerprints when was applied on purified P. falciparum DNA samples(Brazil and Colombia), as weil as in patient:S blood samplesfrom a large area ofVenezuela. The performance of the Pf-21-1 assay was compared against Giemsa stained thick blood smears from samplescollected at a malaria endemic area ofthe Bolivar State, Venezuela, at the jield station ofMalariologiain Tumeremo. Coupled to non-radioactive hybridization the p{-21-1 pe/formed better tJzan the traditional microscopic method with a 1'=1.7:1. In the case ofmixed infections the r value of P. falciparumdetection increased to 2.5:1. The increased diagnostic sensitivity ofthe test produced with this homologous oligonucleotide could provide an alternative to the epidemiological diagnosis of P. falciparumbeing currently used in Venezuela endemic areas, where low parasitemia levels and asymptomatic malaria arefrequent. In addition, the DNAjingerprint could be tested in molecular population studies.

Key words: Plasmodium falciparum - diagnosis - polymerase chain reaction - malaria

Plasmodiumfalciparum, the agent of the mostlethal form of malaria, causes 1.5 to 2.7 milliondeaths each year, mostly among children. The incidence of malaria in the world is estimated to be300-500 million clinical cases annually (WHO1997). An approximate of 2,300 million peoplelived in areas with malaria risk, distributed in 100endemic countries incIuding Venezuela.

The Venezuela global incidence in 1997 was28,056 malaria cases (Direccion de EndemiasRurales, Venezuelan Malaria Program, pers.commun.). Bolivar State located in the Amazonbasin is responsible for 40-50% of the global inci-

Financial support from Escuela de Malariologia ySaneamiento Ambiental "Dr. Arnoldo Gabaldon",Maracay-Venezuela, la Fundacion para el Oesarrollo dela Ciencia y la Tecnologia de Aragua, Venezuela andGrupo de Genética Molecular del Instituto de BiologiaExperimental, Universidad Central de Venezuela, Venezuela.+Corresponding author: Fax: +58-2-753-5897. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

dence for the whole country. The number of malaria cases has been in a constant rise during theprevious years, due mainly to local economic activities (gold and diamond mines), changes in human migration patterns and a high prevalence ofP. falciparum resistant malaria to chloroquine andpyrimethamine/sulfadoxine drugs.

Studies on epidemiology ofparasitic infections,on measures to control disease and clinical evaluation of treatments, ail require identification of theinfecting species. The identification ofmalaria parasite is usually perforrned with traditional microscopiediagnosis of Giemsa-stained thick blood, which isinexpensive and of easy application in the field.Nevertheless, the sensitivity ofmicroscopic methoddepends on highly trained examiners, is time-consuming when large numbers of samples must beexamined, and is thus not the most appropriatemethod for large-scale epidemiological surveys.

Assay strategies have been proposed that direct/y detect abundant parasite nucleic acid sequences, including repetitive DNA (Franzen etal. 1984, Barker et al. 1986, Oquendo et al. 1986,McLaughlin et al. 1987, Zolg et al. 1987) or ribosomal RNA (rRNA) (Lai et al. 1989, Wathers &McCutchan 1989).

640 DNA Recombinant Methodologies· L Urdaneta et al.

ln the P falciparum genome there are severalrepetitive sequences. One of these consists of 21bp blocks, imperfectly repeated in tandem clustersoriented in opposite directions (rep-20). These rep20 sequences are found in ail chromosomes of theparasite (Oquendo et al. 1986). DNA probes directed to these repeats hybridize with Pfalciparumstrains from South America, Africa, and South andSoutheast Asia (Buesing et al. 1987), and their sensitivity is comparable to that of conventional microscopy (Lanar et al. 1989). The sensitivity ofDNA and RNA approaches have been further increased through the amplification ofthe target DNAusing the polymerase chain reaction (PCR).

Considering the high specificity of the rep-20for Pfalciparum and its genomic organization, ourobjective was directed to design a diagnostic PCRassay coupled to non-radioactive hybridization, forthe detection of P falciparum in human bloodsamples. In the present communication we reportthis diagnostic assay using a single oligo anddigoxigenin labeled probes detected by photoluminescence. Ils application to blood samples obtained and processed in the endemic areas, compared to traditional thick smear technique is discussed. We also report the application of the assayto isolated Pfalciparum DNA to produce amplification patterns differentiating strains of severalgeographical locations and proposing its use as atool in molecular epidemiological studies.


Blood samples collection and treatment - Thesample blood was obtained from patients withmalaria symptoms assisting to the malaria diagnostic post of the Venezuelan Malaria Program inTumeremo, Bolivar State. The samples were collected during two visits of one week each to theendemic area during July 1993 and March 1994.

Duplicate samples of fingerprick blood werecollected from 33 individuals in heparinized capillary tubes; 50 III were deposited in Wathman paper filter, dried at room temperature and stored inindividual and labeled sealed plastic bag, in case itwas necessary to repeat the PCR assay. Another50 III were transferred to a 1.5 ml centrifuge tubein which the Pfalciparum DNA was isolated using the chelex-IOO-iron protocol (Wooden et al.1993).

Microscopie examination - Two thick bloodfilms were prepared for each patient during theblood collection process. The blood smears werestained with Giemsa, and one reading was performed at field site under routine conditions ofwork (100 fields examined under oil immersionoptics before a slide was considered negative), byan expert microscopist at the diagnostic post of the

Malaria Program. Forcomparison purposes, a second microscopic diagnostic was done by one ofus(L.U.); in this case, the sample was considerednegative after 200 microscopic fields were examined. The parasite number was registered with respect to 200 white blood cells.

P falciparum reference strain DNAs - Pfalciparum reference strain DNAs were a gift ofthe fol1owing researchers: (1) one Colombian strainfrom Dr Moises Wasserman (lnstituto Nacional deHigiene de Colombia), (2) 12 Brazilian isolatesfrom Dr Hernan Del Portillo (Universidade de SàoPaulo, Brasil) and (3) 2 cultured strains from DrErlinda Sanchez (Universidad de Carabobo andMalariologia, Venezuela). The parasite genomicDNA was extracted using proteinase K, followedby phenol-chloroform extraction and ethanol precipitation.

Primers selection and PCR assays - Two oligonucleotides were designed: (1) a primer of 16bp, pf-16 (S'-ACT AACTTA GGTCTTA-3'), and(2) a primer of 21bp pf-21 (5'-ATG TTA GTCAAC TTA AGA CCT-3 '), both derived from thepf-21 consensus sequence reported by Oquendo etal. (1986).

The P falciparum assay described by Tirasophon et al. (1991), was included in this study asa reference test. This PCR uses primers KI14-PI(S'-CGC TAC ATA TGC TAG TTG CCA GAC3') and K114-P2 (S'-CGT GTA CCA TAC ATCCTA CCA AC-3') for amplification of a 206 bpfragment and, primer KI-14 (5'-GCT ATA ACCACT ATT GCA ACG-3') for hybridization. Wedesignated the PCR assays 1to III according to theoligonucleotid used: PCR 1: pf-21 primer alone,PCR II: two primers pf-16 and pf-21, and PCR III:reference test using oligos K 114-P 1and KI14-P2.

PCR reaction conditions - Amplification reaction mixtures consisted of 50 mM KCL, 10 mMTris HCL pH 8.8, 1% Triton X-I 00, 3mM dNTPs,and 3 mM MgCI2. Primers pf-21 and pf-16 wereat4llM. Primers K114-Pl and K114-P2 were at 1IlM. 1.25 units of Taq polymerase. The total volume was 25 Ill. The specificity and sensitivity ofthe assays were tested against different Pfalciparum DNA concentrations (200 ng/Ill to 0.02pgllll), including in each case a negative control.Amplification products were analyzed by electrophoresis in 3% agarose gels and detected by staining with ethidium bromide.

PCR 1and PCR II initial denaturation was doneat 94QC for 5 min followed by 35 cycles of 94QCfor 1 min, annealing at 40QC for 1 min, and extension at 72QC for 2 min, and a final elongation at72QC for 5 min.

PCR III initial denaturation at 94QC for 5 min;35 cycles of denaturation at 94QC for 1 min, an-


cBA

tary and degenerate primers directed against rep20. Nevertheless, the use of two primers PCR Ilyielded nonspecific amplification with humanONA (Fig. IC; lane 3) and negative control without ONA (Fig. lC; lane 4). In the amplificationwith a single primer (PCR 1) (Fig. 1A; lanes J, 2),a broad range of reproducible amplified productswas observed, clearly showing a 344 bp band forthis Pfalciparum isolate. No amplification was observed in the negative control (Fig. 1A; lanes 3,4). The PCR III reference system yielded an expected single 206 bp band specific for Pfalciparum(Fig. 1B; lanes l, 2), without amplifying the human ONA (Iane 3).

Considering the specificity and the amplification profile in P falciparum ONA, the PCR 1system was selected for detection of P falciparumfrom field samples.

Sensilivity ofPCR 1and peR 111 systems - Oncethe specificity of these systems was detennined,experiments were done to examine the detectionsensitivity ofboth systems against seriai dilulionsof purified P falciparum DN A. Fig. 2 shows theresults ofamplification products in 3% agarose gelelectrophoresis for PCR 1 (Fig. 2A), PCR III (Fig.2B) and Southern PCR III hybridizations (Fig. 2C).Both systems showed a detection level of 0.2 pgof P falciparum ONA on agarose gels (Figs 2A,2B; lane 7), equivalent to ten parasites. The South-

M123412341234

•

Fig. 1: PCR systems J, Il and III. Electrophoresis in 3% agarose gel of amplified products by the systems. A: PCR 1pf-21;B: PCR III KI 14-PI/KI14-P2: C: PCR II pf-21+pf-16. Lanes- 1: 200 ng of Plasmodiumfa/ciparum purified DNA; 2: 50 ngof P jiJ/ciparum purified DNA; 3: negative control wilh human DNA; 4: negative control wilhoul DNA; M, 123 bp ladder.

344pb~..206pb~

nealing at 602C for 1 min, and extension at 722Cfor 1 min, and a final elongation at 72QC for 5 min.

Blois and DNA hybridizalions - AmplifiedONAs were size fractionated by electrophoresis in3% agarose gel and transferred onto nylon membranes (Hybond N+, Amersham) using theVacuGene apparatus (Phannacia LKB) followingthe manufacturer instructions. Slot blots were perfonned as described in Davis et al. (1986).

The filters were pre-hybridized in 2X SSC, 10XDenhardt's solution (Maniatis et al. 1982) and 1.5%blocking reagent (Boehringer-Mannheim) for 1 hrat 372C. The hybridization with digoxigenin labeledoligonucleotides were incubated for 2 hr at 53 QCfor pf-21, and 502C for K 1-14. Filters were washedin 6X SSC-I % SOS at room temperature for 15min, 2X SSC-O.I % at room temperature for 15 min,and finally 0.2X SSC-O.l % SOS at 372C for 15min. In the case of ONA probes, the filters werepre-hybridized in 6X SSC, 5X Oenhardt's, 0.5%SOS, 5% blocking reagent and 50% fonnamidefor 2 hr at 372C. The hybridizations with digoxigenin labeled amplified products pf-21 andKI14-PIIKI14-P2 were incubated at 372C for 1218 hr. Filters were washed every 15 min in 2X SSC0.1 % SOS at room temperature for 1 hr, 1X SSC0.1 % SOS at 6S2C each 15 min for 1 hr, and 0.1 XSSC-O.I % SOS at 682C twice each 15 min. Themaxim concentration used for both the oligonucleotides and the probes was 20 ng/1l1 of hybridization solution.

DNA probe labeling and detection - The Genius Kit from Boehringer-Mannheim was used forON A labeling and detection by photoluminiscence.The oligonucleotides were labeled incorporatingthe dUTP-OGX using tenninal transferase, following the manufacturer recommendations. The ONAprobes were labeled by the multiprimer system incorporating the dUTP-OGX nucleotide. The photoluminescent substrate used was Lumi-phos 530and the signais were detected using X-rays films.

RESULTS

PCR assays - Two PCR amplification systemswere designed and standardized: PCR (1) pf-21primer and PCR (II) pf-16+pf-21 primers. Both systems with target in the 21 bp repeat (rep 20). Thereference system PCR (III) KI14-PI/KI14-P2primers, was standardized to our conditions as described above. Fig. 1 shows the results of theseamplification systems using ONA from two Pfalciparum strains from Colombia. Both the PCR1and II directed to rep 20, yielded a broad range ofamplified products, with band sizes from 200 bpto 3 Kb (Fig. 1A, C; lanes 1,2), with higher intensity in the PCR II system. Similar results were reported by Barker et al. (1992), using complemen-

642 DNA Recombinant Methodologies' L Urdaneta et al.

em hybridization tests witb digoxigenin labeledprobe, yielded the following results: (a) the PCR [hybridized to the pf-21 primer showed a sensitivity level lower than the one observed in the electrophoresis, suggesting problems with the hybridization conditions (results not shown); (b) the PCRIII hybridization (Fig. 2C) showed an increase inthe sensitivity to 0.02 pg ofDNA, equivalent to oneparasite (Fig. 2C; lane 8).

A

Erfec/s ofan/icoagulants on the PCR assays Previous reports have revealed the inhibitory effect in PCR analysis ofcertain anticoagulants suchas heparin (Barker et al. 1992, Tirasophon et aL1991), citrate and EDTA (Tirasophon et aL 1994).We tested the PCR assay amplifying vertebraterDNA included as a positive control for the PCRreaction (Premoli-de-Percoco et al. 1993), takingblood samples with different anticoagulants (heparin and citrate), including a sample in Wathmanpaper. The resu lts showed the specific vertebrate 'srDNA band of 126 bp in ail the samples, indicating that there is no inhibition by the anticoagulants(results not shown).

PCR detection ofPfa/ciparum DNA in humanblood samp/es - Fig. 3 shows the agarose 2% electrophoresis of PCR III for a total of 32 samples.There is a 206 bp band (positive signal) in Ilsamples (Fig. 3A; lanes 1,2,6, 7, 8, 9, 10, Il; Fig.3B; lanes 2, 10, 14). There was inhibition of theamplification reaction in one sample (Fig. 3B; Jane6) as revealed by the absence of the vertebrate'srDNA amplification band of 126 bp. The samesampi es were examined using the PCR 1 system,with the single pf-21 primer (Fig. 4). Those samples

1

iNciN

B

B 1 2 3 4 5 6 7 8 9 10 M 11 12 l3 14 15 16 17 la 19

l.'ll'h·..h .....

Fig. 2: sensilivityof PCR 1 and PCR Il systems. Electrophoresis in 3% agMosc gel of amplified producls from seriaidilutions of Plasmodium /alciparum pllrified ONA. A: PCR 1system pf-21: 8: PCR III system K114-P I/K 114-P2; C: Southun-blot hybridization ofB against capture K-14 primer labeledwith digoxigenin. The cxposure time was 15 min. Lanes - 1:200 nglpl; 2: 20 ng/~t1; 3: 2ng/pl: 4: 0.2 ngill/; 5: 20 pglltl; 6:2 pgllll: 7: 0.2 pglpl: 8: 0.02 pg/Ill: M - Fig. 2A: ladder 1 Kb;Fig. 2B: ladder J 23 bp.

Fig. 3: PCR on human samples. PCR III system. Eleclrophoresis in 2% agarose gel ofamplified products by PCR III. A: Janes1 to 10 and Il to 15, human sam pics 1 10 15: lane 16. negativccontrol with human ON A; lane 17. negative control no ONA;lane 1R. positive control with purified Plasmodium/alcipal'//IIIDNA. B: lanes 1 to 10 and Il 10 16. human samples 17 to 32:lane 17. negative control with hllman ON A; Jane 18, negativecontrol no DNA; tane J 9, positive control with purified P/alcipuI'um ONA; M , 123 bp ladder.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Ocl. 1998 643

with a band pattern similar to the P .falciparumDNA positive control (Fig. 4A; lane 18), were considered positives. This positive pattern was observed in 24 samples (Fig. 4A; lanes l, 2, 5, 6, 7,8,9.10,12,13,14; Fig.4B; lanes2,4,5, 7,9, la,Il, 12, 13, 14, 15, 16, 17). The results showed thespecific vertebrate's rDNA band of 126 bp in ailthe samples, indicating that there is no inhibitionof the Taq polimerase by the anticoagulants used.

~ • :.. • .:. • 1.._ li.: 6

. . --

:;11··I.~ .1,1,: .•-.- -- - 1

C19 1817 16 15 14 13, 1 1 •, • 1 • 1 1 •12" 11 10 9 8 7 6 5 4 3 2 1

Fig. 4: PCR on human samples, PCR 1system. Electrophoresisin 2% agarose gel of amplifïed products by PCR 1 pf-21. A:lanes 1 to 15, human samples 110 15; lane 16, negative controlwith human DNA: Jane 17, negative control no DNA; lane 18.positive conlrol with purifïed PlasmodiulII fulcipumm DNA.13: lancs 1 to 17 are human samples 16 to 32; lane 18. posilivecontrol with purifïed P (alcipurum DNA; lane 19. negative control with human DNA; lane 20. negative control no DNA; M,1 Kb ladder. C: slot blot hybridization ofamJ1lifïed prodllcts byPCR 1pf-21 on human samples 16 to 32 against the pf-21 primeramplifïed prodllct. labeled with digoxigenin. Exposition time:30 min. Siots 1 \0 17 are human samples 161032; slol 18. positive control with purifïed Pfitlcipumm DNA; slot 19, negativecontrol with human DNA.

Slot-blot hybridization \Vith PCR [pf2l probe- Fig. 4C shows the slot-blot results ofPCR amplifications with PCR l system on human samples. Apositive signal was observed in stots l, 2, 5, 8, 9,la, Il, 12, 14, 15, 16, and 17.

DNA amplificationfingerprint (DAF) of P.falciparum - Six of the 12 Amazonian Brazilianisolates were amplified by PCR l. Fig. 5 shows adifferent and speci fic polymorphic fingerprint foreach isolate, within the range of 200 bp and 3 Kb.This method couId be used as a molecular markerfor genetic diversity studies in population genetic.

~ 1 2 3 4 5

--_344bp-220

Fig. 5: DAF of Plasmodium falcipurum isolates from Braziland Colombia. Electrophoresis in 2% agarose gel of amplifïedproducls by PCR 1pf-21. Brazilian isolates: 1 - 608; 2 -365; 351; 4 - 54 and Colombian isolate: 5; M - PGEM molecularweight marker.

Microscopie diagnostic - Table 1 shows thecomparative results of both microscopic readings(100 and 200 fields with immersion lens) frompatients blood samples. Detection level was clearlyimproved when the number offields was increased.Taking into account this improvement in the microscopic exam sensitivity, we used the 200 fieldsreading to compare the PCR and hybridization systems. The comparative results of the PCR 1, PCRIII and microscopy are presented in Table II. Table[][ jllustrates the results of 16 samples derived fromthe PCR (1) pf-21 test coupled to a hybridizationassay, in comparison with the microscopic diagnosls.


DISCUSSION

The current diagnosis of human malaria isachieved by microscopie examination of Giemsastained blood smears. Although weIl adapted to thefield situation, this methodology is not practical interms oftime and labor involved when large number of samples are required for epidemiologicalstudies. The availability of DNA derived techniques for diagnosis of infectious agents with highspecificity and sensitivity has represented an attractive alternative. Several groups have identifiedspecies specifie repetitive sequences in the genomeof P falciparum (Franzen et al. 1984, Oquendo etal. 1986, Barker et al. 1986, McLaugh1in et al.1987) and derived DNA probes in an attempt toimprove diagnosis. The highly repeated subtelomeric rep-20 sequence has been target ofDNAprobes as weil as PCR diagnostic assays (Barker

TABLE 1

Comparative detection levels of two microscopiereadings. Tumeremo Bolivar State, Venezuela 1993-94

Parasite 100 field examined 200 field examinedspecies Samples No. % Samples No. %

Plasmodiumfalciparum 7 24 II 38P. vivax 7 24 6 21Mixed infection 2 7Negatives 15 52 10 34

Total 29 100 29 100

et al. 1992). This late approach has increased thedetection level but has not significantly surpassedthe microscopy diagnosis probably due to selfcomplementation between the oligonucleotidesused.

Exploiting the genomic organization of the pf21 repeat, we have designed a PCR assay for Pfalciparum based on a single consensus primer thatcoupled to non-radioactive hybridization, can beapp1ied to the epidemiological diagnosis of Pfalciparum with significant improvement over themicroscopie ana1ysis.

ln order to improve the performance of the diagnostic PCR reaction it was necessary to optimizethe protocol for blood sample preparation. The direct treatment of blood with che1ex-100 Fe wasadopted over the proteinase K incubation and samp1epreservation on filter paper. Consistent amplification ofpatient samp1es was obtained independentlyofthe treatment with anticoagulants, as demonstratedby the vertebrate rDNA PCR assay.

We app1ied the PCR 1 system to 32 samp1es ofmalaria patients with moderate and low parasitemia. As a reference PCR, we used the PCR IIIsystem targeted on moderately repetitive sequences(Fucharoen et al. 1988, Tirasophon et al. 1991) thatamplifies an unique band of 206 bp. The sensitivity of these DNA amplification assays was compared to the improved microscopie reading of200fields. This more extended examination of the slideincreases the positive diagnosis of P falciparumfrom 24% in 100 fields examined to 38% (Table

TABLE II

Comparative detection levels between PCR 1, PCR III and microscopy.Tumeremo Bolivar State, Venezuela 1993-94

Parasite PCRI PCR III Microscopyspecies No. samples % No. samples % No. samples %

Plasmodium falcipamm 19 66 Il 38 Il 38P. vivax 6 21Mixed infection 5 17 5 17 2 7Negatives 5 17 13 45 10 34

Total 29 100 29 100 29 100

TABLE III

Comparaiive detection levels between PCR 1 coupled to an hybridization assay and microscopy.Tumeremo Bolivar State, Venezuela 1993-94

Parasite PCRI PCR III Microscopyspecies No. samples % No. samples % No. samplcs %

Plasmodium falciparum 12 75 12 75 7 44P. vivax 2 12Mixed infection 1 6 2 12Negatives 3 19 2 13 7 44

Total 16 100 16 100 16 100


1). Overall, examining the slides at 200 fields favored the diagnosis of mixed infections and thedetection of low parasitemia over the 100 field'sroutine methodology.

Three of the 32 samples showed inhibition ofthe PCR reaction as evidenced by the absence ofthe 126 bp product of the vertebrate rDNA PCRassay, and were not include in the comparison analysis. The summary ofthe results from the 29 patientsanalyzed by microscopic examination and PCR systems 1and III are shown in Table II. The detectionratio ofthe PCR 1system in relation to both the PCRIII system and the microscopic examination was1.7:1. Five of the six diagnosed P. vivax infectionswhere detected positive for P. falciparum by bothPCR 1 and III systems, increasing the detection ofmixed infections more than two fold.

When the PCR 1 analysis was coupled to hybridization (Fig. 4C) and these results were compared with the microscopic exam (Table III), a similar detection ratio of 1.7: 1, with a higher sensitivity for the PCR was observed. The two reportedcases of P. vivax infection were confirmed positive to P. falciparum by hybridization. In generalthe PCR 1system had a higher detection level thanbath the PCR III system and the 200 field microscopic examination. This Increment in the detection ofP.falciparum infections is significant whensub-patent parasitemia and mixed infections are notdetected by the traditional method.

Additionally, the single oligonucleotide amplification produces a DNA fingerprint specific forP. falciparum isolates of Immediate application topopulation genetic studies. It was possible to observe a pattern of reproducible bands, specificallya band of344 bp for the P. falciparum isolate fromColombia (Fig. lA; lane 1,2; Fig. 5; lane 5). Therep-20 repeat is found in the subtelomeric regionsofP.falciparum chromosomes (Triglia et al. 1992).It is in this region where chromosome rearrangements responsible for chromosomic size changesobserved is postulated to occur. The origin ofthesesize changes is not weil understood, but it is attributed to genetic exchange during the different typesof cellular division (meiosis-mitosis). The possibility ofobtaining an amplification pattern oftheseregions specific for each isolate (as the result ofitschromosomic arrangement), wouId probably permit to follow in an epidemiological way the different circulating isolates, and the relationship oftheseDNA amplification fingerprints (DAFs) with otherimportant characters, such as drug resistance andvirulence.

ln the group ofpatients studied, it was possibleto distinguish at least three DAF variation patternsin the isolates circulating in the zone (Fig. 4A; lanes1,2, 8). A totally different pattern was observed in

the positive control DNA from Colombia strain(Fig. 4A; lane 18).

These results constitute an important tool inthe epidemiology of P. falciparum malaria. Thisapproach will allow the analysis of large numberof samples for diagnosis, detecting mixed infections and low parasitemia. Concomitantly, circulating isolates can be identified trough specificDAF for P.falciparum. We propose the combineduse ofthis polymorphism detection analysis withother similar assays (RAPD, isoenzymes, and micro-satellites) that detect population changes, tofurther evaluate its discrimination capabilities inidentifying different isolates circulating in endemic areas.

ACKNOWLEDGMENTS

To the patients that voluntarily accepted to participate in this study. To Dr Alejandro Caraballo and ail theworkers in the malaria diagnostic post of the Venezuelan Malaria Program in Tumeremo, Bolivar State, forhelping ln the sample collection stage. To Dr MarceloMazzarri and Dr Alexis Rodriguez for their helpful support. To Dr Moises Wasserrnan, Dr Heman Del Portilloand Dr Erlinda Sanchez for providing the reference strainDNAs.

REFERENCES

Barker R, Banchongaksom T, Courval J, Suwonkerd W,Rimwungtragoon K, Wirth 0 1992. A simplemethod to detect Plasmodium falciparum directlyfrom blood samples, using the polymerase chain reaction. Am J Trop Med Hyg 46: 416-426.

Barker R, Suebsaeng L, Rooney W, AJecrim G, DouradoH, Wirth D 1986. Specifie DNA probe for the diagnosis of Plasmodium falciparum malaria. Science231: 1434-1436.

Buesing M, Guerry P, Diesanti C 1987. An oligonucleotide probe for detecting P.falciparum: an analysisofclinical specimens from six countries. J Infect Dis155: 1315-1318.

Davis L, Dibner M, Battey J 1986. Basic Methods inMolecula,. Biology. Elsevier, N.Y.

Franzen L, Shabo R, Perlmann H, Wigzell H, Westin G,As/und L, Persson T, Pettersson U 1984. Analysisof c1inical specimens by hybridization with probecontaining repetitive DNA from Plasmodiumfalciparum. A novel approach malaria diagnosis.Lancet 1: 525-527.

Fucharoen S, Tirawanchai N, Wilairat P, Panyim S,Thaithong S 1988. Ditferentiation of Plasmodiumfalciparum clones by means of a repetitive DNAprobe. Trans R Soc Trop Med Hyg 82: 209-211.

LaI A, Changkasin S, Holingdadi M, McCutchan T 1989.Ribosomal RNA-based diagnosis of P. falciparummalaria. Mol Biochem Parasitol 36: 67-72.

Lanar D, McLaughlin G, Wirth D, Barker R, Zolg J,Chulay J 1989. Comparison of thick films, in vitroculture and DNA hybridization probes for detectingP. falciparum malaria. Am J Trop Med Hyg 40: 3-6.


Maniatis T, Fritsch E. Sambrook J 1982. Molecular Cloning: A Laboratory Manual, Cold Spring HarborLaboratory.

McLaughlin G. Breman J. Collins W. Schwartz J,Brandling-Bennet A, Sulzer A, Skinner J. Ruth J.Andrysiak P, Kasaje D, Campbell G 1987. Assessment of a synthetic DNA probe for Plasmodiumfalciparum in African blood specimens. Am J TropMed Hyg 37: 27-36.

Oquendo P, Goman M, Mackay M, Langsley G, WalhkerD, Scaife J 1986. Characterization ofrepetitive DNAsequence from the malaria parasite, Plasmodiumfalciparum. Mol Biochem Parasitol 18: 89-] 0 1.

Premoli-de-Percoco G, Pinto-Cistemas J, Ramirez JL,Galindo 1 1993. Focal epithelial hiperplasia: Humanpapillomavirus induced disease with a genetic predisposition in Venezuela family. Hum Gent 91: 386388.

Tirasophon W, Ponglikitmongkol M, Wilariat P,Boonsaeng N. Panyim S 1991. A novel detection ofa single Plasmodium (alciparum in infected blood.Biochem Biophys Res Comm 175: 179-184.

Tirasophon W, Rajkulchai P, Ponglikitmongkol M,Wilairat P, Boonsaeng V, Panyim S 1994. A highlysensitive, rapid and simple polymerase chain reaction-based method to detect human malaria (Plasmodiumfalciparum and Plasmodium vivax) in bloodsamples. Am J Trop Med Hyg 51: 308-313.

Triglia T. Wellems T, Kemp D 1992. Towards a highresolution map of the Plasmodium jàlciparum genome. Parasitol Today 8: 225-229.

Wathers AP, McCutchan T 1989. Rapid sensitivity diagnosis of malaria based on ribosomal RNA. Lancet 8651: 1343-1346.

WHO - World Health Organization 1997. World malariasituation in 1994. Week~v Epidemiol Record. Geneva72: 36-269.

Wooden J, Kyes S, Sibley C 1993. PCR and strain identification in Plasmodium jalciparum. Parasitol 1'0day 9: 303-305.

Zolg J. Andrade L, Scott E 1987. Detection of Plasl1lodiumfalcipal'llm DNA using repetitive DNA clonesas species specific probes. Mol Biochem ParasitaI22: 145-151.

Mem Inst Oswaldo Cruz, RIO de Janeiro, Vol. 93(5): 647-650, Sep'/Oct. 1998 647

Systematics and Population level Analysis ofAnopheles darlingi

JE Conn

Department of Biology, University of Vermont, Burlington, VT, USA

A new phylogenetic ana(vsis ofthe Nyssorhynchus subgenus (Danorr-Burg and Conn, unpub. data)using six data sets {morphological (alliife stages); .l'canning electron micrugraphs of eggs; nuclearITS2 sequences; mitochondrial COlI, ND2 and ND6 sequences} revealed c!irferent topologies wheneach data set was analyzed separate(v but no heterogeneity between the data sets using the am test.Consequently, the most accurate estimate of the ph.vlogeny was obtained when ail the data were combined. This new phylogeny supports a monophyletic Nyssorhynchus subgenus but both previousZl' recognized sections in the subgenus (Albimanus and Argyritarsis) were demonstrated to be paraphyleticrelative to each other andfour ofthe seven clades included species previousZv placed in both sections.One ofthese clades includes both Anopheles darlingi and An. albimanus, suggesting that the ability tovector malaria effectively may have originated once in this subgenus.

Both a conserved (315 bp) and a variable (425 bp) region of the mitochondrial COI genefrom 15populations of An. darlingifrom Belize, Bolivia. Brazil, French Guiana, Peru and Venezuela were usedto examine the evolutionary history of this species and to test several analytical assumptions. Resultsdemonstrated (l) parsimony analysis is equally informative compared to distance anaZvsis using NJ; (2)clades or clusters are more strangly supported when these mo regions are combined compared to eitherregion separately; (3) evidence (in the form ofremnants ofolder haplotype lineages) for mo colonization events; and (4) significant genetic divergence within the populationfrom Peixoto de Azevedo (Stateof Mato Grosso, Brazil). The oldest lineage includes populations from Peixoto, Boa Vista (State ofRoraima) and DOl/rado (State ofSào Paulo).

Key words: Anopheles - Nyssorhynchus - Anopheles darlingi - phylogeny - combined analysis - parsimony distance analysis

In the neotropics, species in the subgenusNyssorhynchus are responsib1e for many of theestimated 20 million annual cases of malaria(Goriup & Pull 1988). The original phylogenetichypothesis for this important subgenus (Faran1980, Faran & Linthicum 1981, Linthicum 1988)was based on morphological characters. Thesubgeneric treatment ofPeyton et al. (1992) placedNyssorhynchlls species in three purportedly monophyletic sections: Albimanus, Argyritarsis andMyzorhynchella. A recent parsimony analysis ofeight species from the Albimanus section usingpartial sequences of the mtDNA genes ND2 andND6 (Perera I 993) had no nodes in common withthose of Faran (1980). Relationships among species in Nyssorhynchus remained unresolved until

These projects received financial support from theNational Institutes ofHealth (NIH; USA) grant AI 31034to LP Lounibos and NIH grant AI 40116 to JE Conn.Fax:+802-656.29 14. E-mail: [email protected] 15 June 1998Accepted 30 July 1998

the recent analysis of Danoff-Burg and Conn(unpub. data) which forms the basis for the systematic portion of this presentation. A subset oftheir objectives was: (1) an analysis of membersof the Albimanus (15 species) and Argyritarsis (8species) sections using six data sets (morphology,egg ultrastructure, ITS2 region, and mitochondrialgenes COll, ND2 and ND6); (2) an examinationof character congruence between these data setsusing the am and Templeton tests (Larson 1994,Farris et al. I995); and (3) a reevaluation of theearlier classifications of Nyssorhynchus with regard to a new total evidence phylogeny based onparsimony where ail available data were equallyweighted and included.

Anopheles darlingi has historically been considered the most important malaria vector throughout much of South America (Deane et al. 1946).Prior to the analysis of Danoff-Burg and Conn(unpub. data), it was placed in the Argyritarsis section of the Nyssorhynchus subgenus (Linthicum1988). lt is broadly distributed and has been incriminated as an important regional vector ofPlasmodium falciparum, the most dangerous of themalaria parasites (Deane et al. 1946), but it is also

648 Anophe/es darlingi • JE Conn

a competent vector of other malaria species (Kleinet al. 1991). Although the taxonomic status of An.darlingi has recently been reevaluated throughoutits range and is now considered a single species byallozyme, RAPD-PCR, ITS2 and morphologicalanalysis (Manguin et al. unpub. data), questionsremain concerning populations which ditfer in biting times (reviewed in Rosa-Freitas et al. 1992),life history characteristics (Lounibos et al. 1995)and genetic divergence (discussed in the presentwork). These heterogeneous traits (and others) mayinfluence the vector competence ofAn. darlingi indifferent regions in the neotropics. We sequencedtwo regions (one conserved and one variable) ofthe mtDNA gene COI from 15 populations ofAn.darlingi from Belize to Sào Paulo to conduct apopulation level analysis comparing parsimony anddistance methods (Conn & Hennig, unpub. data).Assumptions or hypotheses to be tested at the population level were: (Hl) analysis based on distancemeasurements is more informative than cladisticanalysis; (H2) ditferent mo1ecular models (for distance analysis) provide different estimates of genetic divergence; (H3) the more variable region ofthe mtDNA COI gene gives better resolutionamong haplotype lineages than the more conservedregion; (H4) ana1ysis of combined conserved andvariable regions is more informative at various levels of divergence than analysis of either regionalone.


Systematics - Methods of extraction and amplification of DNA, as weil as methodology forthe phylogenetic analysis of the six data sets, thecongruence tests, and outgroup treatment are foundin Danoff-Burg and Conn (unpub. data).

Population level - DNA was extracted fromindividual wild-caught mosquitoes following theprotocol in Collins et al. (1987). These mosquitoes were collected from BZ (Belize), EJ (El Juval,Trujillo, Venezuela), AY (Puerto Ayacucho,Amazonas, Venezuela), FG (French Guiana), IQ(Iquitos, Peru), GU (Guayaramerin, Bolivia), BV(Boa Vista, Roraima, Brazil), MC (Macapa,Amapa, Brazil), CP (Capanema, Para, Brazil), NS(Tefé, Amazonas, Brazil), IT (Itacoatiara,Amazonas, Brazii), AB (Porto Velho, Rondônia,Brazil), PX (Peixoto de Azevedo, Mato Grosso,Brazil), SP (Araraquara, Sào Paulo, Brazil) and DO(Dourado, Sào Paulo, Brazil) (Fig.). Primers forthe two regions of the COI gene were from Lunt etal. (1996). For the parsimony analysis the following parameters were used with PAUP 3.1.1.(Swotford 1993): unrooted trees, heuristic search,and trees were assessed by bootstrap analysis ( lOO1000 replicates)(Felsenstein 1985). Three distance

CollectIOn localities for sarnples of Anophe/es dar/mgi

models were assessed by the Neighbor-Joining (NJ)distance method: uncorrected p, Tamura-Nei andHKY85 and aH were tested by 1000 bootstrap replications. Details of this analysis are from Connand Hennig (unpub. data).

RESULTS

Systematics - The reanalysis of the 49 morphologica1 characters from Faran (1980), Faran andLinthicum (1981) and Linthicum (1988) produceda largely unresolved strict consensus tree whichwas eight steps shorter than the original (DanoffBurg & Conn, unpub. data). The analysis of thesix different data sets (morphology, egg ultrastructure, ITS2 region, and mitochondrial genes COlI,ND2 and ND6) each produced a distinctivetopology even though the character-based heterogeneity was not significantly different between anytwo matrices. The tree based on the ITS2 analysisalone most accurately reflected the new phy10genetic hypothesis which was produced when ail datawere combined. This new phylogeny supportedmonophyly of the subgenus Nyssorhynchus butboth the Albimanus and Argyritarsis sections wereparaphyletic relative to each other and four cladescontained species previously placed in both sections. Both An. albimanus (previously in theAlbimanus section) and An. darlingi (previouslyin the Argyritarsis section) arose basally in thephylogeny.

Population level - For the 27 individuals sequenced for the conserved region alone (315 bp),there were 20 characters that were parsimony informative. With 1,000 bootstrap replicates, fourclades were supported: BV-NS at 53%; CP 1-CP2DO at 66%; FR-MC at 53%; PXI-PX2 at 100%.


Individuals from the same locality were found inthe same clade 50% ofthe time and little geographicpartitioning was apparent. For the variable regionalone (425 bp), 66 characters were parsimony informative among the 24 individuals sequenced.

Two clades were supported ( 100 bootstrap repIicates): EJ I-GU l-EJ2-DO I-NS4-AB3 at 52% andSP2-PXI-BV3-PX2-D02-NS3-AB4 at 61%. Except for PX and EJ, individuals from the samepopulations were not found in the same clades. Thecombined region (740 bp) had 81 characters thatwere parsimony informative for the 22 individualssequenced. Three clades were supported at 200bootstrap replications: EJI-GU l-EB at 68%; PX1PX2-BV3-D02 at 100% and NS4-AB3 at 82%.The clade composition was similar to that foundin the variable region even though one additionalclade was supported by the analysis of the combined regions. The parsimony trees for each of theCOI regions were tested to determine ifthere wasphylogenetic signal by comparing each ofthe threetree lengths {conserved region (TL = 53), variableregion (TL = 256), combined regions (TL =289)} toa distribution of tree lengths of 1000 randomlygenerated trees. Ali three trees contained significant phylogenetic signal using this test.

A comparison of the parsimony analysis of thecombined regions with each of the three distancemodels demonstrated that the trees from the parsimony and the HKY85 model were identical in cladecomposition (EJ I-GU I-EJ3, PX I-PX2-BV3-D02,NS4-AB3) and differed very slightly in levels ofbootstrap support. For the uncorrected p andTamura-Nei models, three additional populationswere supported (SP2, AB4, and NS3) either as aseparate clade (uncorrected p) or as part ofa largerclade (Tamura-Nei).

Results of the four hypotheses were: (HI) treeswere either more resolved using parsimony (conserved and variable regions analyzed separately)or nearly equally resolved (combined analysis) ascompared with distance analyses; (H2) ofthe threedistance models tested, HKY85, the most parameter-rich model, gave the least amount of resolution (i.e., lowest levels ofbootstrap support); (H3)for conserved vs. variable regions of the COI gene,parsimony analysis resulted in more lineages being supported (four) for the conserved region compared with the variable region (two) but most ofthe same populations were grouped with both regions; while for the distance analyses between regions, each model contained the same number ofIineages; (H4) the combined analysis for both theparsimony and distance was more informative thanfor either region alone.

A graph of the frequencies of pairwise geneticdistances of both regions combined using uncor-

rected p (x-axis) compared with the number ofpairwise comparisons (y-axis) resulted in a bimodal distribution. The genetic distances for the firstpeak ranged from 0.004-0.05. Ali of the pairwisecomparisons ofhigh genetic distances (0.06-0.13;the second peak) included at least one individualfrom PX, BV or DO.

DISCUSSION

Systematics - The parsimony analysis ofbothmolecular and morphological data does not support the earlier phylogenies of Faran (1980), Faranand Lithicum (1981), Linthicum (1988) and Perera(1993). The discordant topologies of the six datasets were probably the result of two factors: (1)homoplasy, and (2) data sets were informative at adifferent taxonomic levels. Perhaps the most significant aspect of this new phylogenetic hypothesis is the basal position of two of the majorneotropical malaria vectors, An. albimanus and An.darlingi, suggesting that the ability to effectivelyvector malaria parasites may have arisen once inthe ancestor of Nyssorhynchus (Danoff-Burg &Conn, unpub. data). This pleisiotypic ability appears to have been retained by many species in thissubgenus which act as important local or regionalvectors when population densities of major vectors are low or when environmental conditions aresignificantly altered (Cruz Marques 1986, P6voaet al. unpub. data). The similarity between the ITS2tree and the total evidence tree suggests that nuclearmarkers are potentially more accurate in reconstructing the true phylogeny at this hierarchicallevel compared with the mitochondrial or morphological markers presented.

Population level - The similar results betweenthe parsimony and distance models were presumably because there was low overall sequence divergence among populations (i.e., not near saturation), there were no secondary hits, and many ofthe mutations were unique to single individuals.

If the genetic distances and Iineage support areaccurate portrayals of the evolutionary history ofAn. darlingi, this suggests that there have been atleast two waves ofcolonization events across SouthAmerica. Alternatively, the haplotype lineage withthe greatest genetic divergence, PX-BV-DO, maybe the remnant ofan older lineage which has goneextinct in other regions ofthe range ofAn. darlingi;this scenario is favoured by the strong support forthe PX-BV-DO lineage and how rarely individuaIs from the same geographic locality are found inthe same clade. This pattern of distribution (multiple divergent haplotypes in the same population)has also been found in Ail. nuneztovari (Conn etal. 1998) and may be more appropriately explainedby local extinction of once widespread lineages

650 Anopheles darlingi • JE Conn

than by heterogeneous effective population sizesor immigration from previously isolated areas(Slatkin 1985).

Information on genetic divergence and heterogeneity among populations ofAn. darlingi may beuseful for existing malaria control strategies in thatlocal solutions will need to be implemented forsuccessful transmission reduction, as proposed byWHO (1992). One ofthe worst recent malaria outbreaks in Brazil has been documented in Peixotode Azevedo (R. Zimmermann, pers. comm.) whereAn. darlingi is considered to be the main vector.

ACKNOWLEDGMENTS

For Jogistic and technological support in the field 1thank collaborators in Bolivia (Unidad Sanitaria,Riberalta), Brazil (Fundaçào Nacional da Saude,Fundaçào Oswaldo Cruz, Institituo Evandro Chagas,Instituto de Pesquisas Cientificas e Tecnol6gicas doEstado do Amapa), France (LIN-ORSTOM), Peru(NAMRID, Iquitos) and Venezuela (Univ. de los Andes,Trujillo; Univ Central, Caracas; Malariologia, Maracay).To RC Wilkerson for additional specimens from Brazil.To WC Kilpatrick for constructive criticism. To lADanoff-Burg and A Hennig for the excellent collaboration.

REFERENCES

Collins FH, Mendez MA, Rasmussen MO, MehaffeyPC, Besansky Nl, Finnerty V 1987. A ribosomalRNA gene probe differentiates member species ofthe Anopheles gambiae complex. Am J Trop MedHyg 37: 37-41.

Conn JE, Mitchell SE, Cockburn AF 1998. Mitochondrial DNA analysis of the neotropical malaria vector Anopheles mmeztovari. Genome, in press.

Cruz Marques A 1986. Um estudo sobre a dispersào decasos de malaria no Brasil. Rev Bras Malar D Trop38: 51-75.

Deane LM, Causey OR, Deane MP 1946. Studies onBrazilian anophelines from the Northeast and Amazan regions. 1. An illustrated key by adult femalecharacteristics for the identification of thirty-fivespecies of Anophelini, with notes on the malariavectors (Diptera, Culicidae). Am J Hyg Mono Series18: 1-18.

Faran ME 1980. Mosquito studies (Diptera, Culicidae)XXXIV. A revision ofthe Albimanus section of thesubgenus Nyssorhynchus ofAnopheles. Contr AmerEnt Inst 15: 1-215.

Faran ME, Linthicum lU 1981. A handbook ofAmazonian species ofAnopheles(Nyssorh)'nchus) (Diptera:Culicidae). Mosq Syst 8: 1-107.

Farris lS, Kallersjo M, Kluge AG, Bult C 1995. Constructing a significance test for incongruence. SystBiol 44: 570-572.

Felsenstein 1 1985. Confidence limits on phylogenies:an approach using the bootstrap. Evolution 39: 783791.

Goriup S, Pull JH 1988. Field research in the context ofmalaria control, p. 1741-1764. In WH Wernsdorfer,1 McGregor (eds), Malaria: Principles and Practice ofMalariology, Vol. 2, Churchill and Livingstone, New York.

Klein TA, Lima 1BP, Tada MS, Miller R 1991. Comparative susceptibility of anophelines in Rondônia,Brazil to infection by Plasmodium vivax. Am J TropMed Hyg 45: 463-470.

Larson A 1994. The comparison of morphological andmolecular data in phylogenetic systematics, p. 371390. In G Schierwater, B Streit, GP Wagner, RDeSalle (eds), Molecular Ecology and Evolution:Approaches and Applications, Birkhauser, VerlagBasel/Switzerland.

Linthicum Kl 1988. A revision of the Argyritarsis section of the subgenus Nyssorhynchus of Anopheles.Mosq Syst 20: 99-271.

Lounibos LP, Nishimura N, Conn J, Lourenço-deOliveira R 1995. Life history correlates ofadult sizein the malaria vector Anopheles darlingi. Mem InstOswaldo Cruz 90: 769-774.

Lunt OH, Zhang DX, Szymura lM, Hewitt GM 1996.The insect cytochrome oxidase 1gene: evolutionarypatterns and conserved primers for phylogeneticstudies. Insect Mol Biol 5: 153-165.

Perera OP 1993. Phylogenetic Analysis ofTwo Mitochondrial Genes from Several Species of the SubgenusNyssorhynchus (Culicidae: Anopheles) and the Development ofSpecies-specific DNA Probes for TheirIdentification, PhD Thesis, University of Florida,Gainesville, 110 pp.

Peyton EL, Wilkerson RC, Harbach RE 1992. Comparative analysis of the subgenera Kerteszia andNyssorhynchus of Anopheles (Diptera: Culicidae).Mosq Syst 24: 51-69.

Rosa-Freitas MG, Broomfield G, Priestman A, MilliganJJM, Momen H, Molyneux OH 1992. Cuticular hydrocarbons, isoenzymes, and behavior ofthree populations of Anopheles darlingi from Brazil. J AmerMosq Control Assoc 8: 357-366.

Slatkin M 1985. Gene flow in natural populations. AnnRev Ecol Syst 16: 393-430.

Swofford DL 1993. PAUP: phylogenetic analysis usingparsimony. Version 3.1.1. Illinois Natural HistorySurvey, Champaign.

WHO - World Health Organization 1992. World malariasituation in 1990. Bull WHO 70: 801-807.


Anopheline Species Complexes in Brazil. CurrentKnowledge of Those Related to Malaria Transmission

Maria Goreti Rosa-Freitas/+, Ricardo lourenço-de-Oliveira*, Carlos José deCarvalho-Pinto*/**, Carmen Flores-Mendoza*, Teresa Fernandes

Silva-do-Nascimento*

Laboratorio de Sistematica Bioquimica, Departamento de Bioquimica e Biologia Molecular *Laboratorio deTransmissào de Hematozoarios, Departamento de Entomologia, Instituto Oswaldo Cruz, Av. Brasil 4365,

21045-900 Rio de Janeiro, RJ, Brasil **Departamento de Microbiologia e Parasitologia,Universidade Federal de Santa Catarina, Florianopolis, SC, Brasil

A summary ofthe prohlems related to the systematics ofprimary and seconda')) Brazilian anophelinesvectors ofmalaria is presented.

Key words: Anopheles systematics - species complexes - malaria vectors - Nyssorhynchl/s - Kerteszia

Many neotropical anopheline species are eithercandidates or fonned by complex of cryptic species. The taxonomie elucidation ofthese complexesreflects on the epidemiology of malaria transmission and ultimate1y to the control.

In Brazil, there are 54 species belonging to fivesubgenera of Anopheles Meigen (Nyssorhynchus,Kerteszia, Stethomyia, Lophopodomyia, Anopheles). Anopheline species reported as human malaria vectors in the country belong to the subgenera Nyssorhynchus and Kerteszia (Deane 1986,Consoli & Lourenço-de-Oliveira 1994).

In the subgenus Nyssorhynchus, the speciesfound harboring human plasmodia include Anopheles darlingi Root 1926, An. aquasalis Curry 1932,An. alhitarsis sensu lato Lynch-ArribaIzaga 1878(including An. deaneorum Rosa-Freitas 1989), An.oswaldoi Peryassu 1922, An. nuneztovari Gabaldon1940 and An. triannulatus (Neiva & Pinto 1922).In the subgenus Kerteszia natura1 infections werereported for Ail. cruzii Dyar & Knab 1908, An.heUator Dyar & Knab 1906 and Ail. homunculusKomp 1937. It is our opinion that other speciesreported naturally infected do not play a raIe inmalaria maintenance as they are exophilic, zoophilie, oflow density and their distribution and frequency do not coincide with that of malaria. Except for An. darlingi, the natural history of the species listed above points out for zoophilie and/orexophilic behavior in sorne areas, in such a fash-

+Corresponding author.E-mail: [email protected] 15 June 1998Accepted 30 July 1998

ion that their raie in malaria transmission is doubted(Deane 1986). Are these characteristics an indication that these species are indeed complexes?

To decide whether a given species is high polymorphie or a complex of closely related species,integrated approach studies on distinct populations,including on that of the type-Iocalities and wheremorphological/behavioral/molecular difTerenceshave been reported, are mandatory. Most of theBrazilian anopheline species has been taxonomically investigated by morphology, behavior andmolecular tools such as isoenzymes and DNAanalyses (mitochondrial and ribosomal DNA restriction analysis, random amplification and sequencing of specifie regions) as summarized onTable.

More than 99% of the malaria cases reportedin Brazil occur in the Amazon in which transmission is due to Nyssorhynchus species only.

An. darlingi is the most important Brazilianmalaria vector (Shannon 1933, Rachou 1958). Thespecies is the most anthropophilic and endophilicamong the Amazonian anophelines. It is frequentlyfound infected and its distribution and density areclearly related to malaria transmission. Eventhough many populations of the species have beenlate1y reported as biting outdoors, An. darlingi continues successfully transmitting malaria both indoors and at the close vicinity of the houses(Lourenço-de-Oliveira 1995). Isoenzymatic, behavioral and mitochondrial DNA studies on eitherBrazilian (Rosa-Freitas et al. 1992, Freitas-Sibajevet al. 1995) or other Latin-American (Manguin etal. 1998) populations, showed that An. darlingi isa monotypic species.

An. aquasalis is the lowland coastal vector inBrazil. Chromosoma1 banding pattern and mtDNA

TABLE <:1'U1N

Summary of differences in behavior, morphology, isoenzyme, mtDNA, rDNA, RAPD and cytogenetic data reported in the literature for populations of neotropicalanopheline species related to malaria transmission ;..

:>

Species Behavior Morphology Isoenzyme mtDNA Cytogenetics References Conclusion0

rDNA RAPD -0:r-

(lTS2) <Ilin

Anopheles darlingi Consoli & Lourenço-de-Oliveira 1994 Monotypic '"'" - - - ND ND '" V1

(Peak and place (Except from Freitas-Sibajev et al. 1995 ""C<Il

"ofbiting) Belize) Harbach et al. 1993 (ij'<ft

Kreutzer et al. 1972 nManguin et al. 1998 0

3Rosa-Freitas et al. 1992 ""C

iDAn. aquasalis ND Conn et al. 1993a High

x

'" '" - '" - - .(Host and place (Egg) Cova-Garcia et al. 1977 polymorphie 3::ofbiting) Flores-Mendonza 1994 (VI) !!;

0;'Monaca-Perez & Conn 1991 Cl

Complex0

An. albitarsis '" - '" '" ND '" '" Kreutzer et al. 1976~.(Host and place (Except An. Narang et al. 1993 (4 species);>0

ofbiting) deaneorum) Rosa-Freitas et al. 1990 0<ft

Wilkerson et al. 1995 <li

~

An. oswaldoi VI ND ND ND Causey et al. 1946 Complex:~

'" '" '" ;::;:<li

(Host and place (Male Consoli & Lourenço-de-Oliveira 1994 (at least 2 <ft

ofbiting) genitalia) Flores-Mendoza pers. comm. forms - VI) ~<li

Klein & Lima 1990 -

Marrelli et al. 1998

An. nuneztovari '" '" '" '" '" ND '" Delgado & Rubio-Palis 1992 Possiblya(Host and place (Egg, male Fritz et al. 1994 complexofbiting) genitalia Hribar 1994, 1995

and female) Linley et al. 1996

An .triannulatus - '" '" '" ND '" ND Silva-do-Nascimento 1995 Complex:(Male genitalia, Silva-do-Nascimento pers. corn. (at least 2larva and egg) species - VI)

An. cruzii '" '" ND ND '" ND '" Deane et al. 1971 High(Acrodendrophily) (Larva) Malafronte et al. 1997 polymorphie

Ramirez et al. 1989 (VI)Zavortink 1973

ND: non determined; VI: under investigation.


restriction profiles of specimens from Venezuelaand Brazil were identical (Moncada-Pérez & Conn1991, Conn et al. 1993a). Isoenzymes from threepopulations of Venezuela and Surinam (Steiner etal. 1981) and two from Brazil (Flores-Mendoza1994) with behavioral differences also revealedonly intraspecific variation. Egg morphology ofAn.aquasalis varies intraspecifically (Maldonado et al.1997). In fact, variation was seen in a single female oviposition (Flores-Mendoza 1994). Resultsofmitochondrial DNA and egg morphology analyses however, suggest that there might be an interspecific division inAn. aquasalis populations northand south of the Amazon River delta (Conn et al.1993a, Linley et al. 1993).

An. albitarsis is a complex formed by, at least,four sibling species: An. marajoara Galvào &Damasceno 1942, An.albitarsis sensu strictoLynch-ArribaIzaga 1878, An. deaneorum and afourth form to be formally described (Rosa-Freitaset al. 1990, Wilkerson et al. 1995). Since An.deaneorum is the only morphologically distinguishable member of the albitarsis complex, therole of each member in malaria transmission hasnot been determined yet. The incrimination ofothermembers of the complex, except An. deaneorum(Klein et al. 1991 a, b), were based solely on theirpresumed geographical distribution.

An. oswaldoi has been regarded as a potentialmalaria vector in sorne localities of the Amazon(Arruda et al. 1986, Oliveira-Ferreira et al. 1990,Branquinho et al. 1996) although sorne authorsbelieve that most populations of this species aremuch more related to the natural environment andprefer to feed on animais than on man indoors(Deane et al. 1948, Consoli & Lourenço-deOliveira 1994, Lourenço-de-Oliveira & Luz 1996).The taxonomic status of the species is under investigation (PhD work of CFM). Preliminary results from morphological analyses demonstrate thatat least two forms are present under An. oswaldoi:An. oswaldoi sensu stricto and An. konderi Galvào& Damasceno 1942, distinguished pratically onlyby the shape of the apex ofaedeagus (Causey et al.1946, Lounibos et al. 1997).

Morphology, behavior, cytogenetics, isoenzymes and mtDNA studies favor the existence ofat least two cryptic species in An. nuneztovari:one in Venezuela and Colombia northwest ofOrinoco and another in the Amazon (Conn et al.1993b, Fritz et al. 1994, Linley et al. 1996). Thespecies is considered a primary malaria vector inVenezuela and Colombia (Gabaldon 1969,Gabaldon et al. 1975). In Brazil however, the species is not related to malaria transmission, althoughnatural infection by Plasmodium viven: has beendetected in areas where darlingi was the primary

vector (Arruda et al. 1986).An. triannulatus is constituted by at least three

sibling forms. These forms can be differentiatemorphologically (egg, larva and male genitalia) andisoenzymatically (Silva-do-Nascimento 1995). Thetypical triannulatus is the most known and largelydistributed form. The other two forms seem to berestricted mostly to central Brazil and are not related to malaria transmission.

The mosquitoes ofthe subgenus Kerteszia sharethe common characteristic of using bromeliads asbreeding places. An exception is An. (Ker.)bambusicolus Komp 1937 that also breeds in bamboo.

An. (Kersteszia) cruzii and An. bel/ator wereprimary vectors of the malaria once endemic insoutheastern/southern Brazil (Rachou 1958). An.cruzii is currently involved in the maintenance ofthe oligosymptomatic malaria occurring in the valleys of the Atlantic Coastal Rain Forest in bothRio de Janeiro and Sào Paulo states (Carvalho etal. 1988, Azevedo 1997, Branquinho et al. 1997).Larval differences were observed in An. cruziipopulations from Rio de Janeiro and Santa Catarina(Zavortink 1973). Besides, chromosomal bandingpattern differences were also found among severalAn. cruzii populations (Ramirez 1989, Dessen pers.comm.). An. homunculus is a morphologicallyclose related species and there is the possibility ofbeing a sibling species in the cruzii complex (PhDwork of CJCP). The remaining Kerteszia speciesdo not seem to be important in malaria transmission in Brazil.

In summary, An. darlingi is a monotypic species. An. aquasalis and An. nuneztovari are possibly complexes. An. albitarsis. An. triannulatus andAn. oswaldoi are complexes ofspecies. Anophelinespecies of the subgenus Kersteszia are still underinvestigation (Table).

The refinement of the taxonomic tools and theaddition of other populations are likely to lead tonew insights into the knowledge and understanding of the neotropical species complexes.

REFERENCES

Arruda ME, Carvalho MB, Nussenzweig RS, MaracicM, Ferreira AW, Cochrane AH 1986. Potential vectors of malaria and their different susceptibility toPlasmodiumfalciparum and P. vivax in northem Brazil identified by immunoassay. Am J Trop Med Hyg35: 873-881.

Azevedo AL 1997. Aspectos da Epidemiologia daMalaria e da Biologia de Anopheles (Kerteszia)cruzii Dyar & Knab em Va/es Montanhosos doSistema de Mata Atlântica, MSc Thesis, InstitutoOswaldo Cruz, Fiocruz, Rio de Janeiro, 94 pp.

Branquinho MS, Araujo MS, Natal D, Marrelli MT,Rocha RM, Taveira FA, Kloetzel JK 1996. Anoph-

654 Anophe/es Species Complex • Maria Goreti Rosa-Freitas et al.

eles oswaldoi a potential malaria vector in Acre,Brazil. Trans R Soc Trop Med Hyg 90: 233.

Branquinho MS, Marrelli MT, Curado l, Natal D, BarataJM, Tubaki R, Carreri-Bruno GC, de Menezes RT,Kloetzel JK 1997. Infecçào de Anopheles (Kerteszia)cruzii por Plasmodium vivax e Plasmodium vivaxvariante VK247 nos municipios de Sào Vicente eJuquitiba, Sào Paulo. Rev Panam Salud Pûblica 2:189-193.

Carvalho ME, Glasser CM, Ciaravolo RMC, Etze! A,Santos LA, Ferreira CS 1988. Sorologia de malariavivax no foco Aldeia dos Indios, municipio dePeruibe, Sào Paulo, de 1984 a 1986. Cad Saûde Pûbl4: 276-292.

Causey OR, Deane LM, Deane MP 1946. Studies onBrazilian anophelines from the northeast and Amazon regions. Il. An illustrated key by male genitaliacharacteristics for the identification of thirty-fourspecies ofanophelini, with a note on dissection technique. Am J Hyg Mono Series 18: 21-31.

Conn JE, Cockburn AF, Mitchell SE 1993a. Populationdifferentiation of the malaria vector Anophelesaquasalis using mitochondrial DNA. J Hered 84:248-253.

Conn JE, Rangel-Puertas Y, Seawright J 1993b. A newcytotype of Anopheles nuneztovari from WesternVenezuela and Colombia. J Am Mosq Control Assoc9: 294-301.

Consoli R, Lourenço-de-Oliveira R 1994. PrincipaisMosquitos de fmportância Sanitaria no Brasil, Ed.Fiocruz. 225 pp.

Cova-Garcia P, Pulido J, Amarista RJ 1977. Anopheles(Nyssorhynchus) deltaorinoquensis n. sp. (Diptera:Culicidae) de Venezuela. Bol Dir malariol San Amb17: 150-160.

Deane LM 1986. Malaria vectors in Brazil. Mem fnstOwaldo Cruz 81(SuppI.II): 5-14.

Deane LM, Causey OR, Deane MP 1948. Notas sobre abiologia do anofelinos das regiôes nordestina eamazônica do Brasil. Rev Sery Esp Saûde Publ 1:827-965.

Deane LM, Deane MP, Ferreira-Neto JA, Almeida FB1971. On the transmission of simian malaria in Brazil. Rev fnst Med Trop Sào Paulo 13: 311-319.

Delgado N, Rubio-Paris Y 1992. Morphometric characterization of the malaria vector Anophelesnuneztovari (Diptera: Culicidae) from Western Venezuela. Mosq Syst 24: 241.

Flores-Mendoza C 1994. Um Estudo sobre AlgunsAspectos Comportamentais, MOIfologicos eBioquimicos de Anopheles aquasalis Curry 1932,MSc Thesis, Instituto Oswaldo Cruz, Rio de Janeiro,172 pp.

Freitas-Sibajev MGR, Conn J, Mitchell S, Cockburn AF,Seawright JA, Momen H 1995. Mitochondrial DNAand morphological analyses of Anopheles darlingipopulations from Brazil. Mosq Syst 27: 79-85.

Fritz GN, Conn J, Cockburn A, Seawright J 1994. Sequence analysis ofthe ribosomal DNA internai transcribed spacer 2 from populations of Anophelesnuneztovari. Mol Biol Evo111: 406-416.

Gabaldon A 1969. Global malaria erradication: changes

of strategy and future outlook. Am J Trop Med Hyg18: 641-656.

Gabaldon A, Martin GG, Sifontes R 1975. Necesidadesen el campo de la investigacion dei ProgramaNacional de Eradicacion y Control de la Malaria deVenezuela. Bol Dir Malariol San Amb 15: 263-285.

Harbach RE, Roberts DR, Manguin S 1993. Variationin the hindtarsal markings of Anopheles darlingi(Diptera: Culicidae) in Belize. Mosq System 25: 192197.

Hribar LJ 1994. Geographie variation of male genitaliaofAnopheles nuneztovari (Diptera: Culicidae). MosqSyst 26: 132-144.

Hribar LJ 1995. Costal wing spot variation within andamong progeny of single female Anophelesnuneztovari (Diptera: Culicidae). Mosq Syst 27: 115.

Klein TA, Lima lBP 1990. Seazonal distribution andbiting patterns of Anopheles mosquitoes in CostaMarques, Rondônia, Brazil. J Am Mosq ControlAssoc 6: 700-707.

Klein TA, Lima lBP, Tada MS 1991 a. Comparative susceptibility ofanopheline mosquitoes to Plasmodiumfalciparum in Rondônia, Brazil. Am J Trop Med Hyg44: 598-603.

Klein TA, Lima lBP, Tada MS, Miller R 1991 b. Comparative susceptibility of anopheline mosquitoes inRondônia, Brazil, to infection by P vivax. Am J TropMed Hyg 45: 463-470.

Kreutzer RD, Kitzmiller lB, Ferreira E 1972. Inversionpolymorphism in the salivary gland chromosomesof Anopheles darlingi. Mosq News 32: 555-556.

Kreutzer RD, Kitzmiller lB, Rabbani MG 1976. Cytogenetically distinguishable populations of the mosquitoAnopheles albitarsis. Acta Amazonica 6: 473481.

Linley JR, Lounibos LP, Conn J 1993. A descriptionand morphometric analysis of the eggs of four SouthAmerican populations ofAnopheles aquasalis. MosqSyst 25: 198-214.

Linley JR, Lounibos LP, Conn J, Duzak D, NishimuraN 1996. A description and morphometric comparison of eggs from eigth geographic populations ofthe South American malaria vector Anopheles(Nyssorhynchus) nuneztovari. J Am Mosq ControlAssoc 12: 275-292.

Lounibos LP, Duzak D, Linley JR 1997. Comparativeegg morphology of six species of the AlbimanusSection of Anopheles (Nyssorhynchus) (Diptera:Culicidae). J Med Entomol 34: 136-155.

Lourenço-de-Oliveira R 1995. Quai a importância dahematofagia extradomiciliar do Anopheles darlingina Amazônia? Rev Patol Trop 23 (supp1.): 100-101.

Lourenço-de-Oliveira R., Luz SLB 1996. Simian malaria at two sites in the Brazilian Amazon. Il. Vertical distribution and frequency ofanopheline speciesinside and outside the forest. Mem fnst OswaldoCruz 91: 687-694.

Malafronte RS, Marrelli MT, Carreri-Bruno, GC,Urbinatti, PR, Marinotti 0 1997. Polymorphism inthe second internai transcribed spacer (ITS2) ofAnopheles (Kerteszia) cruzi (Diptera: Culicidae)

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), 5ep./Oct. 1998 G55

from the State of Sào Paulo, Brazil. Mem /nstOswaldo Cruz 92 (Suppl. 1): 306.

Maldonado V, Finol HJ, Navarro JC 1997. Anophelesaquasalis eggs from two Vcnezuelan localities comparcd by scanning electron microscopy. Mem /l1stOswaldo Cruz 92: 487-491.

Manguin S, Wilkerson RC, Conn JE, RublO-Palis Y,Danoff-Burg JA, Roberts DR 1998. Population structure of the primary malaria vector in South America,Anopheles darlingi using isozyme, RAPD,ITS2 andmorphological markers. Am J Trop Med Hyg, inpress.

Marrelli MT, Malafronte RS, Flores-Mendoza C,Lourenço-de-Oliveira R, Kloetzel JK, Marinotti 01998. Polymorphism in the second internai transcribed spacer (lTS2) of ribosomal DNA amongspecimens ofAllopheles oswaldoi. /nseet Mol Biol.in press.

Moncada-Pérez A, Conn J 1991. A polytene chromosome study of four populations of Anophelesaquasalis from Venezuela. Genome 35: 327-331.

Narang SK, Klein TA, Perera OP, Lima 18, Tang AT1993. Genetic evidence for the existence of crypticspecies in the Anopheles albitarsis complex in Brazil: allozymes and mitochondrial DNA restrictionfragment length polymorphisms. Bioehem Genet 3/:97-112.

Oliveira-Ferreira J, Lourenço-de-Oliveira R, Teva A,Deane LM, Daniel-Ribeiro CT 1990. Natural malaria infections in anophelines in Rondonia State,Brazilian Amazon. Am J Trop Med Hyg 43: 6-10.

Rachou RG 1958. Anofelinos do Brasil. Comportamentodas espécies vetoras de malaria. Rev Bras MalariolDoenças Trop /0: 145-181.

Ramirez CL 1989. Estudo Cromossâmico em uma

Populaçào de Anopheles (Kerteszia) cruzii Dyar &Knah /909. MSc Thesis, Instituto de Biociências,Universidade de Sào Paulo, Sào Paulo, 96 pp.

Rosa-Frcitas MG, Broomfield G, Priestmann A, MilliganP, Momen H, Molyneux DH 1992. Studics on cuticular components, isoenzymes and behaviour of 3populations of Allopheles darlingi from Brazil. JAmer Mosq COlltrol Assac 8: 357-366.

Rosa-Freitas MG, Deane L, Momen H 1990. A morphological, behavioural and isoenzymatic study ln

Anopheles albitarsis from 10 populations. Mem /nstOswaldo Cruz 85: 275-289.

Shannon RC 1933. Anophclines of the Amazon Valley.Proe Entamai Soc Washington 35: 117-143.

Silva-do-Nascimento T 1995. Estudo Taxonâmico eNotas sobre a Biologia de Anopheles trianl1ulatusde 8 Localidades. MSc Thesis, Instituto OswaldoCruz, Fiocruz, Rio de Janeiro, 88 pp.

Steiner WWM, Kitzmiller 18, Osterbur, DL 1981. Onthe genetic identity and evolution of the malaria vectors Allopheles aquasalis and Anopheles emilianus.ln R Pal, 18 Kitzmiller & T Kanada (eds), Cytogenetics and Genetics ofVeetors. Proceedings ofa symposium of the XVI International Congress of Entomology, Amsterdam, Elsevier Biomedical,Amsterdam.

Wilkerson RC, Parsons Tl, Klein TA, Gaffigan TV,Bergo E, Consolim J 1995. Diagnosis by randomamplified polymorphic DNA polymerase chain reaction of four cryptic species related to Anophelesalbitarsis from Paraguay, Argentina and Brazil. JMed Entamai 32: 697-704.

Zavortink Tl 1973. Mosquito studies XXIX. A reviewofthe subgenus Kerteszia ofAnopheles. Contr AmerEntomol/nst 9: 1-54.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 657-661, Sep./Ocl. 1998 657

Implications of a Neotropical Origin of the GenusLeishmaniaHarry Noyes

Liverpool School of Tropical Medicine, Pembrokc Place, Liverpool, L3 5QA, UK

The hypothesis of a Neotropical origin of the LeishmaniaJEndotrypanum clade is reviewed. Theposition ofthe L. (Sauroleishmania) external to the subgenus L. (Leishmania) is not consistent with theNeotropical origin of the latter subgenus. It is suggested that this may be a consequence ofa fasterevolutionary rate in the L. (Sauroleishmania). The implicationsfor the classsification ofthe phlebotominesandflies ofthe hypothesisfor a Neotropical origin ofthe Leishmania is also considered. The classification ofGalati (1995) is proposed to be most consistent with the hypothesis ofa Neotropical origin oftheLeishmania. whilst class!fications which place the New and Old World species in separate taxa areinconsistent with this hypothesis.

Key words: Endotrypanum - Leishmania (Sauroleishmania) - Leishmania hertigi - Phytomonas - porcupines sloths - Phlebotominae - biogeography

In recent years DNA sequence based phylogenies have transformed our understanding of theevolutionary relationships amongst a wide rangeof protozoa. The growing number of taxa examined and the increasing range of molecules usednow makes it possible to consider how host parasite systems have coevolved with more confidence.

The family Trypanosomatidae consists ofninegenera of parasitic flagellated protozoa. Five ofthese genera are parasites of arthropods only andare transmitted contaminatively (Wallace 1979).Three genera have digenetic lifecycles in vertebrates and invertebrates and one genus thePhytomonas is parasitic in plants and insects. Phylogenies of the ribosomal RNA genes ofmembersof the Trypanosomatidae indicate that vertebrateparasitism has arisen on at least two separate occasions within this family, once in the Trypanosomaand once in the Leishmania/Endotrypanum clade(Fernandes et al. 1993, Hol1ar & Maslov 1997,Lukes et al. 1997, Noyes 1998). The increasingnumber of taxa used in these phylogenies makes itpossible to calibrate these phylogenies against specific events in the evolution of the hosts of theseparasites and hence to consider how host parasitesystems may have coevolved through time.

The genus Trypanosoma is a cosmopolitanparasite ofalmost all classes ofvertebrates and mayhave very ancient origins in the Palaeozoic. The

Fax: +151-708-8733. E-mail: [email protected] 15 June 1998Accepted 30 July 1998

Trypanosoma that are infective to humans may beof much more recent origin in the continents inwhich they are now found (Stevens et al. 1998).The genus Phytomonas is parasitic in floweringplants to which it is transmitted by a range ofHemiptera and Diptera. Since the first flowering plantsappeared in the fossil record in the early Cretaceous (130 million years ago - MYA) and began todominate the terrestrial flora during the second halfof the Cretaceous it is possible that the genusPhytomonas made the transition from monogeneticparasites ofplant feeding insects to digenetic parasites of plants and insects during the Cretaceous.The remaining two genera of digenetic parasitesthe Endotrypanum and the Leishmania are the twomost closely related genera in the rRNA phylogenies of the Trypanosomatidae and their commonancestor may have made a transition to digeneticparasitism around the time of the mammalian radiation in the late Cretaceous or early Cenozoic(Fernandes et al. 1993).

Leishmania and Endotrypanum are both transmitted by phlebotomine sandflies but the genusEndotrypanum only infects sloths in the Neotropicswhilst the genus Leishmania infects at least nineorders ofmammals and reptiles and is found in thetropics and subtropics worldwide. Recent phylogenies of the Leishmania/Endotrypanum clade haveshown that the parasites which are endemic in theNew World are closer to the root ofthis clade thanthe Old World parasites (Fig. 1). Consequently ithas been proposed that this clade made the switchfrom monogenetic parasites of phlebotominesandflies to digenetic parasites of sandflies andvertebrates in the Neotropics (Croan et al. 1997,Noyes et al. 1997). This switch is believed to have

658 Implications of Neotropical Origin of Leishmania • Harry Noeys

occurred at the time of the mammalian radiation atthe end ofthe Cretaceous when the Neotropics werebecoming isolated from the rest of the world (Fig.2). At this time only four groups of mammals areknown to have been established in the Neotropics:the monotremes, the marsupials, the notoungulatesand the xenarthrans (Fig. 1). The monotremes andnotoungulates became extinct in the Neotropicsduring the Cenozoic leaving the marsupials andxenarthrans as the only indigenous mammals tosurvive in the Neotropics to the present day(Patterson & Pascual 1972).

The Xenarthra consists ofthe sloths, the armadillos and the anteaters. The two genera ofmodemsloths Bradypus and Choloepus belong to two different families which diverged at the end of theCretaceous over 70 MYA (H6ss et al. 1996). Themodem genera are the remnants of a much morediverse group of terrestrial, arboreal and evenaquatic sloths that flourished during most of theCenozoic (Patterson & Pascual 1972, Engelmann1985, de Muizon & McDonald 1995). Since thetwo modem genera of sloths are hosts to both genera within the LeishmanialEndotrypanllm clade itis possible that sloths were the first vertebrate hostsof the common ancestor ofthese parasites. Monogenetic trypanosomatid parasites of sandflies arenot uncommon, although they are rarely described.These trypanosomatids may have become preadapted to development in the blood ofvertebratesby frequent exposure to blood in the gut of theirsandfly hosts. The relatively low body temperature of modern sloths, and presumably extinct

sloths as weil, may have facilitated the switch fromsandfly to vertebrate hosts. Altematively it has beensuggested that since a number of genera ofmonogenetic and digenetic trypanosomatids can flourish in the anal scent glands ofmarsupials, that thisorgan provided a transitional stage in the acquisition ofvertebrate parasitism and that consequentlymarsupials were the primary hosts of the Leishmania/Endotlypanum clade (Deane & Jansen 1988,Jansen et al. 1988). However the sandfly gut is anenvironment in which the parasite is more likelyto be routinely exposed to vertebrate blood overmany generations, than in the marsupial anal glandwhere the parasites have a lower chance ofonwardtransmission. It may be difficult to resolve thisquestion until more is known of the lifecycle andhost specificity of the monoxenoustrypanosomatids of sandflies.

The phylogenies of the LeishmanialEndotrypanum clade show that the L. hertigi complex is more closely related to the genusEndotrypanum than to the Leishmania (Fig 1.)(Croan et al. 1997, Noyes et al. 1997). Since parasites ofthe L. hertigi complex have only been foundin Neotropical porcupines (Rodentia: Hystrichomorpha: Erethizontidae) it is possible that theseparasites did not diverge from the Endotrypanumline until the hystricomorph rodents arrived in theNew World during the Eocene. This provides acalibration point for dating the remaining nodes inthe EndotrypanumlLeishmania tree (Fig.I). Thissuggests that the Endotrypanum and Leishmaniagenera diverged during the Palaeocene 53-65MYA,

~PRr.CA\1DRIAi\

N,

Ut é:,

~N Utw -0- -.1 w w N .."

.." .." 0,0'$ Zln' >: dl ~ 9.~ 3!Z'" 212- 1~i 11 ia8 S. :::~o a. -<:::8 ",,0 n_.~

.gs &lo ~ 1=- ln' ~_·0 .... - .~ B-&l.a N - ::0

~~ICln<j .aa.

if

CENOZOIC

MIOCENEPIlD

L.(S.) _

----....,L.(L.) ------L.(V.) -----------'L.h.

E.sp.

Fig. l' geological periods plotted on a logarithmic scale showing the approximate dates of the appearance of mammal and Insectgroups important for the understanding of the evolution of the Endollypanllln/Leishmania clade. Numbers indlcate millions ofyears before the present. Dates of divergence of mammal groups are from Kumar and Hedges (1998). A phylogeny of theEndollypallum/ Leishmania clade IS shown based on the phylogeny ofRFLPs of the small subunit ribosomal RNA gene of Noyeset al. (1997). The phylogeny IS cahbrated against the divergence of the L. herligi and Elldoll~'Panlimclades which IS assumed tohave occurred no earher than the late Eocene at the time the Hystricomorph rodents first appeared in the Neotropical fossil records.L. (S) L. (Sauroielshmania); L (L.). L (Leishmania): L. (v.): L. (Viannia): L. h.: L herllgi, E. sp.. ElldollJpanUm sp.

Mem /nsl Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), 5ep.lOct. 1998 659

the Neotropical subgenus L. (Viannia) branchedoff during the early Miocene and the L. (Leishmania) and L. (Sauroleishmania) diverged during thesecond hal f of the Miocene. The tree used to estimate these dates was prepared from RFLPs of thesmalJ subllnit rRNA gene (Noyes et al. 1997). Further trees inclllding more taxa and data from moregenes should increase the reliabi lity of these estimates.

Since the subgenus L. (Leishmania) is foundin both the Old and New Worlds a member ofthissubgenus may have migrated to the Old Worldacross the Bering straits region before this regionbecame too cool for the sandfly vectors in the lateMiocene (Wolfe 1994). The subgenus L.(Sauroleishmania) may then have diverged fromthe L. (Leishmania) in the Old World as a consequence of its adaption to reptiles. Although thishypothesis for the origin of the L. (Sauroleishmania) in the Old World requires the minimum number of migrations and extinctions it is not consistent with the phylogenies. If the L. (Sauroleishmania) had evolved in the 01d World from L. (Leishmania) parasites that had migrated from the NewWorld then the RNA and DNA polymerase phylogeny of Croan et al. (1997) wouJd be expectedto show the L. (Sauroleishmania) branching offbetween the L. (L.) mexicana complex which isrestricted to the New World and the L. (L.) major,L. (L.) tropica and L. (L.) donovani complexeswhich are restricted to the Old World. Instead itshows that L. (Sauroleishmania) branched off before any of the L. (Leishmania) subgenus. It is possible that either the L. (Sauroleishmania) firstevolved in the New World and then also migratedto the Old World independently of the L. (Leish-

mania) before becoming extinct in the New World,or that the cornmon ancestor of both sllbgeneramigrated to the Old World and that the ancestorsof L. mexicana migrated back after the L.(Sauroleishmania) had diverged. However it secmsmore likely that the position of the L. (Sauroleishmania) extemal to ail the L. (Leishmania) in theRNA and DNA polymerase phylogeny is an artefact of a faster evolutionary rate in the L.(Sauroleishmania). The long length of the branchleading to the L. (Sauroleishmania) is suggestiveof a faster evolutionary rate a possibility that wasalso indicated by a rate test (Croan et al. 1997).Faster evolving groups are known to be pulled towards the outgroup, a phenomenon known as longbranch attraction, which could have generated theobserved phylogeny (Felsenstein 1988) Tt is conceivable that the changes that were necessary forthe L. (Sauroleishmania) to adapt from mammalian hosts to reptile hosts may have forced a temporarily accelerated rate of evolution. It may bepossible to test this hypothesis more rigorously bythe inclusion of additional taxa in this phylogeny.

Tt is still not known whether the L. (Sauroleishmania) are transmitted by the bite of the sandfly orby contamination when the reptile eats the sandfly(Telford 1995). Since it now appears that the L.(Sauroleishmania) have evolved fTom parasites thatare transmitted by the bite of the fly this is perhapsthe most likely method for transmission of L.(Sauroleishmania) to reptiles as weil.

Most Leishmania parasites are more restrictedin their range ofsandfly vectors than in their rangeof mammalian hosts, implying a much closer coevolutionary relationship with the sandfly than themammal. However the proposed Neotropical ori-

Fig. 2: a map of the world in the laIe Cretaceous (65-1 00 million years ago - M YA) showing the isolation of the Neolropics from thePalaearctic. The Isthmus of Panama did not reconnect the IWO continents until 5 MYA in the Pliocenc. The Beringia region whichconnected the Palaearctic to the Nearctic for most of the Cenozoic (0-65MYA) is shown. The shaded areas indicate shallow seasthat covered large parts of the Palaearctic and Nearctic during the Mesozoic and Cenozoic (after Cox 1973).

660 Implications of Neotropical Origin of Leishmania • Harry Noeys

gin of the Leishmania is not consistent with theexisting classification of the sandflies. Nevertheless relationships between the subgenera and species complexes of sandflies are still controversialand the existing nomenclature may not reflect thetrue relationships within this group (Lane 1993).If the hypothesis for a Neotropical origin is correct then it will be possible to make some predictions for the classification ofthe sandflies that couldbe tested by molecular methods. In the Old Worldmammalian Leishmania are transmitted bysandflies of the genus Phlebotomus and lizard parasites are transmitted by sandflies of the genusSergentomyia. In the New World Leishmania andEndotrypanum are transmitted by sandflies of thegenus Lutzomyia (Fig. 3). If Leishmania migratedacross the Bering region during the Miocene theremust have been a resident population of sandflyvectors throughout this area which may have leftdescendants in both the Old and New Worlds. Consequently the modem sandfly vectors in both theOld and New World may be more closely relatedto each other than they are to sympatric non-vector genera.

One recent phylogeny of the sandflies doessuggest that this is the case. Galati (1995) placesthe Old World genus Sergentomyia in a newsubtribe, the Sergentomyiina, with sorne reptile

biting species that are at present in the New Worldgenus Lutzomyia. In th is c lass ification theSergentomyiina is clustered in a group of NewWorld subtribes which suggests that theSergentomyiina may also have evolved in the NewWorld and that the modem Sergentomyia are descendants of Sergentomyiina that migrated fromthe New World to the 01d World. This impl ies thatsandflies could have crossed through Beringia atsome time. The Sergentomyiina are primarily reptile biters, but members of the Lutzomyia vexatorseries, which Galati places within this subtribe,have been implicated as the vectors ofL. mexicanain Texas (Kerr et al. 1995). Consequently it is possible that ancestors of the widespread Lu. vexatorseries or of a closely related group may have carried the mammalian Leishmania to the Old World.No molecular phylogenies of the Phlebotominaehave been published, but they will provide a valuable test of the hypothesis for a Neotropical originof the Leishmania.

REFERENCES

Cox CB 1974. Vertebrate palaeodistributional patternsand continental drift. J Biogeog 1: 75-94.

Croan DG, Morrison DA, Ellis JT 1997. Evolution ofthe genus Leishmania revealed by comparison ofDNA and RNA polymerase gene sequences. MolBiochem Parasitol89: 149-159.

L (LeishmBniB)L. (VlBnniB)L. hertigiEndotrYPBnum

Fig. 3: the modem distribution ofsandfly vector genera, after Lewis (1974) and Leishmania. The proposed roule by which Leishmania migrated from the Neotropics to the Old World is indicated by an arrow. The distribution of sandflies is limited by summertemperatures which must remain above 20°C for 50 days a year (Lewis 1982). Areas with a mean temperature of less than 20°C inthe hottest month (June for the Northem hemisphere and January for the Southern Hemisphere) are shaded. These areas are notnorrnally suitabJe for Phlebotominae. ln the Northem Hemisphere the 20°C isotherrn is at about 45°N, for the Beringia region to besuitable for sandflies this isothenn must have been approximately 15° further north.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep.IOct. 1998 661

de Muizon, C, McDonald, HG 1995. An aquatic slothfrom the Pliocene ofPeru. Nature 375: 224-227

Deane MP, Jansen AM 1988. From a mono to a digenetic life-cycle: how was the jump for flagellates ofthe family trypanosomatidae. Mem Inst OswaldoCruz 83: 273-275.

Engelmann GF 1985. The phylogeny of the Xenarthra,p. 51-64. In GG Montgomery, The Evolution andEcology of Armadillos, Sloths and Vermilinguas.Smithsonian Institution Press, Washington.

Felsenstein J 1988. Phylogenies from molecular sequences: inference and reliability. Annu Rev Genet22: 521-565.

Fernandes AP, Nelson K, Beverley SM 1993. Evolutionof nuclear ribosomal RNAs in kinetoplastid protozoa - perspectives on the age and origins ofparasitism. Proc Nat! Acad Sei USA 90: 11608-11612.

Galati BEA 1995. Phylogenetic systematics ofPhlebotominae (Diptera, Psychodidae) with emphasis on American groups. Boletin de la Direceion deMalariologia y Saneamiento Amhiental 35 (Suppl.1): 133-142.

Hôss M, Dilling A, Currant A, Paabo S 1996. Molecular phylogeny of the extinct ground sloth Mylodondarwinii. Proc Nat! Acad Sei USA 93: 181-185.

Hollar L, Maslov DA 1997. A phylogenetic view on thegenus Phytomonas. Mol Biochem Parasitol 89: 295299

Jansen AM, Carreira JC, Deane MP 1988. Infection ofamammal by monogenetic insect trypanosomatids(Kinetoplastida, Trypanosomatidae). Mem InstOswaldo Cruz 83: 271-272.

Kerr SF, McHugh CP, Dronen JrNO 1995. Leishmaniasis in Texas: prevalence and seasonal transmissionof Leishmania mexicana in Neotoma micropus. AmJ Trop Med Hyg 53: 73-77.

Kumar S, Hedges SB 1998. A molecular timescale forvertebrate evolution. Nature 392: 917-920

Lane RP 1993. Sandflies (Phlebotominae), p. 78-119.In RP Lane & RW Crosskey (eds), Medical/nsectsand Arachnids, Chapman & Hall, London.

Lewis DJ 1974. The biology of Phlebotomidae in relation to leishmaniasis. Ann Rev Entomol 19: 363-384.

Lewis DJ 1982. A taxonomic review of the genus Phlebotomus (Diptera: Psychodidae). Bull Br Mus NatHist(Ent) 45: 121-209.

Lukes J, Jirku M, Dolezel D, Kral'ova I. Hollar L,Maslov D 1997. Analysis of ribosomal RNA genessuggests that trypanosomes are monophyletic. J MolEvo144: 521-527.

Noyes HA 1998. Can Trypanosoma trees be trusted?Parasitol Today /4: 49-50.

Noyes HA, Arana BA, Chance ML, Maingon R 1997.The Leishmania hertigi (Kinetoplastida;Trypanosomatidae) complex and the !izard Leishmania: their classification and evidence for aneotropical origin of the Leishmania-Endotrypanumclade. J Eukaryot Microbiol 44: 511-517.

Patterson B, Pascual R 1972. The fossil mammal fauna ofSouth America, p. 247-310. In A Keast F Erk & BGlass (eds), Evolution. Mammals and Southem Continents, State University ofNew York Press, Albany.

Stevens JR, Noyes HA, Dover GA, Gibson WC 1998.The ancient and divergent origins of the humanpathogenic trypanosomes, Trypanosoma brucei andT. cruzi. Parasitology, in press.

Telford Jr SR 1995. The kinetoplastid hemoflagellatesof reptiles, p. 161-223. In JP Kreier, Parasitic Protozoa, 2nd ed., Vol. 10, Academic Press, San Diego.

Wallace FG 1979. Biology of the Kinetoplastida ofarthropods, p. 213-240. In WHR Lumsden & DAEvans (eds), Biology of the Kinetoplastida, Vol. 2,Academic Press, London.

Wolfe JA 1994. An analysis of Neogene climates inBeringia. Palaeogeogr Palaeoclimatol Palaeoecol108: 207-216.


Genetic Diversity in Natural Populations of New WorldLeishmania

Elisa Cupolilloj+, Hooman Momen*, Gabriel Grimaldi Jr

Laborat6rio de Leishmaniose, Departamento de Imunologia *Laboratorio de Sistematica Bioquimica,Departamento de Bioquimica e Biologia Molecular, Instituto Oswaldo Cruz, Av. Brasil 4365, 21045-900

Rio de Janeiro, RJ, Brasil

Gur results have shown the wide diversity o[parasites within New World Leishmania. Biochemicaland molecular characterization ofspecies within the gel1lls has revealed that much of the populationheterogeneity has a genetic basis. The source ofgenetic diversity among Leishmania appears to arisefi"Om predominantly asexual, clonai reproduction, although occasional bouts ofsexual reproduction cannot be ruled out. Genetic variation is extensive with some clones widefv distributed and others seemingly unique and localized to a particular endemicfocus. Epidemiological studies of leishmaniasis hasbeen directed to the ecology and dynamics oftransmission of Leishmania species/variants, particularlyin localized areas. Future research using molecular techniques should aim to ident!D' andfollow Leishmania types in nature and correlale genetic typing with important clinical characteristics such as vin/lence, pathogenicit)', drug resistance and antigenic variation. The epidemiological significance ofsuchvariationnot only has important implications/or the control ofthe leishmaniases, but would also help toelucidate the evolutionary biology ofthe causative agents.

Key words: New World Leishmania - leishmaniasis - epidemiology - molecular characterization

Genetic variation in medical1y important protozoan parasites and the nature of the reproductivestrategies which predispose to such variation arecurrently the subject of much interest and controversy (Dye et al. 1990, Tibayrenc et al. 1990, 1991,Tibayrenc & Ayala 1991, Hurst et al. 1992, Sibley& Boothroyd 1992). This genetic heterogeneityproduces different phenotypes which can be associated with a diversity ofclinically important manifestations. At least 13 distinct Leishmania speciesare widespread in the New World and recognizedas causing human illness in the Americas. Each ofthese parasites has a unique zoonotic life cycle,with different sand fly vectors and vertebrate reservoirs (Grimaldi et al. 1989).

Taxonomy ofNew World Leishmania - Taxonomic studies ofleishmanial isolates from the NewWorld indicate tremendous diversity within thisgenus. A number ofnew Leishmania species fromsylvan areas ofthe Neotropics have been describedrecently. Sorne ofthese parasites are associated withdisease in humans, others seem to be restricted tolower orders of mammals, such as rodents andedentates (Lainson & Shaw 1987). Since the origi-

+Corresponding author. Fax: +55-21-280.1589. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

nal description of these parasites, the number ofnamed species has continually increased and several taxa or classification schemes have been proposed (Gardner 1977, Lainson & Shaw 1979,Rioux et al. 1990), including a subdivision of thegenus Leishmania Ross, 1903 into two subgenera,the Viannia and Leishmania, according to the development of the parasite in the gut of sand fly(Lainson & Shaw 1987).

Except for minor differences in size, ail species of Leishmania are morphologically similar.The initial criteria for identification and classification of these parasites were based on extrinsiccharacteristic, such as c1inical manifestations, geographic and epidemiological features, and a variety of other biologic criteria. However, the variation produced by these criteria lead to the development of biochemical, immunological and molecular methods to provide more precise taxonomicmarkers based on intrinsic characteristics of theparasites themselves. Among the techniques currently in use are isoenzyme electrophoresis, species-specific monoclonal antibodies or DNA probeand analysis of restriction fragment length polymorphism (RFLP) using different DNA sequencesas targets (Macedo et al. 1992, Guizani et al. 1994,Mendonza-Leon et al. 1995).

Multi/ocus enzyme electrophoresis (MLEE) The electrophoretic analysis of isoenzymes hasbeen the most wide1y used method for characterizing Leishmania (WHO 1990). Isoenzyme electro-

664 Genetic Diversity in Leishmania • Elisa Cupolillo et al.

phoresis has the ability to examine a very largesample of structural genes, providing genetic evidence to distinguish polymorphism within speciesfrom differences between species as well as information on the reproductive biology of a given organism (Richardson et al. 1986). The techniqueinvolves separating isoenzymes by gel electrophoresis and subsequent visualization of specifieenzymes using appropriate staining reactions. Isolates with identical banding patterns (alleles) areusually referred to as zymodemes (Godfrey 1979).Important taxonomie information can be obtainedby numerical analysis ofthe electrophoretic bands,which may vary according to the different allelesor genotypic frequencies ofioci that are present indistinct parasite strains (Avise 1975).

A large sample of New World Leishmania hasbeen analyzed in our laboratories by MLEE. TheLeishmania strains analyzed until now weregrouped in 68 zymodemes (Cupolillo et al. 1994,1997). Numerical analysis, using phenetic and phylogenetic methods, has demonstrated that the proposed classification ofLeishmania in two subgenera, Leishmania and Viannia (Lainson & Shaw1987), may represent a valid scheme. The parasites were clustered into five phenetic complexes:L. braziliensis. L. nai/fi, L. guyanensis/L.panamensis/L. shawi, L. mexicana, L. major. AliL. chagasi parasites formed a unique zymodemecloser to L. major than to the other Leishmaniaspecies. Within the L. gllyanensis/L. panamensis/L. shawi complex, we found sorne named species

to be as similar as variant strains within each ofthese taxa, which showed that these parasites areclosely related as a group. The L. braziliensis andL. naiffi group showed the highest population heterogeneity, presenting 15 and II zymodemes, respectively. Sorne species, like L. lainsoni. L.eqllatorensis and L. colombiensis were shown tobe very distinct from the other species, but relatedamong themselves (Cupolillo et al. 1994, 1997)(Fig. 1).

fntergenic region typing - DNA analysis provides a means of examining expressed and nonexpressed sequences ofan organism and is not subject to environmental influences. RFLP analysisdetects genetic differences by comparing size variation in DNA banding patterns after restriction endonuclease analysis. The technique has been applied to Leishmania focusing basically on the restriction patterns ofthe minicircle kDNA molecule(Lopes et al. 1984, Pacheco et al. 1986). A PCRbased method has been applied in our \aboratoryto study genetic diversity among parasites, withparticular reference to the intraspecific variabilitythat occurs in natural populations ofa given Leishmania species. In this methodology we amplify theinternai transcribed spacers (lTS) of the rRNA geneby PCR, followed by the digestion ofthe PCR product with several restriction enzymes. The transcribed noncoding regions of rRNA genes (ITS)show extensive variability. Unlike the non-transcribed spacers, the ITS are relatively small(approx. 1kb in Leishmania) and flanked by highly

01 02 03 04 OS 06 07 08 09

Jaccard's coefficient

L. mexicana complexand others L.(Leishmama)

L. chagasi

L. major corn plex

L. naiffi corn plex

L. hraziliensis cornplex

L. gllyanemis corn plex

L. lainso/llL. colomhiensis

L. eqlla/oY(!/lSIS

Leishmaniasubgenus

Vianniasubgenus

Fig. 1: dendrogram showing the level of similarity and the diversity in each phenetic complex/species.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 19_98 665

conserved segments to which PCR primers can bedesigned. This approach can potentially be appliedin many evolutionary situations and amongst abroad range of target loci, we refered to it asintergenic region typing, or IRT (Cupolillo et al.1995). We analyzed many Leishmania isolates representing the Viannia subgenus by this method andoverall, the ITS clustering showed good agreementwith previous organismal or isoenzymatic groupings (Cupolillo et al. 1994). In concordance withthe MLEE data L. braziliensis and L. na(ffi population showed a high level ofpolyrnorphisrn. Ifoneaccepts current species assignments, one interpretation is that both L. naiffi and L. braziliensis arehighly polymorphie, possibly due to their evolutionary antiquity. This view was supported by clineof evolutionary distances within L. naijfi, withoutclearly predominant zymodeme or ITS subgroups(Cupolillo et al. 1994, 1995). In contrast, the levelof polymorphism was less evident in the group L.guyanensis/L. panamensis/L. shawi, reinforcing theidea that these species are very closely related(Lainson & Shaw, 1987, Thomaz-Soccol et al.1993, Cupolillo et al. 1994). However, a small polymorphism was observed at the intra-specific levelfor L. guyanensis and L. shawi, in contrast to theresults obtained by MLEE (Fig. 2). This resultpoints the IRT as a useful method to study geneticvariability amongst intra-species Leishmania isolates from an endemic foci.

Genetic variability - The genetic variabilitywas analyzed in the Viannia subgenus by population-genetic parameters (Cupolillo et al. 1997). Thelevel of polymorphism was 100%, presenting amedia of six alleles/Ioci. The relative level of genetic variability observed among the parasites indicates that they represent a heterogeneous group

L. guyanemis

L. panamensisL. shawi

L. braziliensis

L. colombiensis

L. equatorensis

L. lainsoni

L. naiffi

110 rrs Distance

Fig. 2: internaI transcribed spacers (/TS) relationship and diversity in Leishmania (Vtannia) species.

oforganisrns. The level ofheterozygosity observed(HLobs) among the zymodemes was 0.12, whereasthe level ofheterozygosity expected (HLexp) wouldbe 0.64. Ifa population is in Hardy-Weinberg equilibriurn, both the HLobs and HLexp wouId be similar. The relatively high values of HLexp and thestriking differences found between HLobs andHLexp ail pointed to the existence ofa clonaI structure in natural Leishmallia populations (Selander& Levin 1980, Tibayrenc & Ayala 1988), reinforced by a strong linkage disequilibrium observed,indicating that asexual reproduction in Leishmania is far more common than sexual. Although it isevident that Leishmania has a clonai populationstructure, it does not exclude the possibility ofsexual recombination. Our analysis of the MLEEdata showed a large number ofrecurrent mutationsin the Viannia parasites, which makes it reasonable to attribute sorne variation to recombination.Moreover, rnany authors have reported evidenceof hybrid formation in Leishmania (Evans et al.1987, Kelly et al. 1991, Darce et al. 1991, BonfanteGarrido et al. 1992, Dujardin et al. 1993, Belli etal. 1994, Noyes et al. 1996, Bafiuls et al. 1997,Delgado et al. 1997), reinforcing the idea thatsexual reproduction may occur in Leishmania, butat a level as yet undefined. It is important to emphasize that rare or occasional bouts of sexual recombination in a normally asexual organism canhave a profound effect on the extent of genetic diversity (Cibulskis 1988).

Transmission cycle - A large number of Leishmania isolates have been characterized geneticallyand considerable variability detected. However, theepidemiolgy of leishmaniasis, in particular thetransmission dynamics of the causative agents, isnot completely understood. The transmission pattern ofNew World leishmaniasis involves two distinct cycles: sylvatic and urban. To understand better the role of animal reservoirs and hosts (vertebrate and invertebrate) in these cycles it is important to associated the variability in the parasitepopulations and the ecology of such endemic areas (Fig. 3).

Characterization by MLEE of L. chagasi isolates obtained from a variety of sources (humans,animaIs and sandflies) have indicated a low levelof genetic variation (Momen & Grimaldi 1989),while L. braziliensis isolates were highly polymorphie and L. naijfi showed intra-specific distancescomparable to the largest obtained within al!Viannia (Cupolillo et al. 1994, 1995, 1997). Interestingly, the same phlebotomine and mammalianspecies serve as vector and reservoirs ofL. chagasithroughout its geographic range; with other parasites, such as L. braziliensis, several different sandfly and animal species are involved in distinct eco-


SYLVATIC CYCLE

Wild reservoirsLeishmania varA, B ...Z

LU/::'Oll1yia var l

1Lu/zomyia var l...

Lu/zomyia var III

Scavenger ammals

~Dogs/Humans ~

Leishmania varA, varB'---------'

LUlzomyia var II Lu/zomyia var IIIy ~

DogsLeishmania var B

HumansLeishmania varA

LU/zomyia var II?

Lutzomyia var IILu/zomyia var III

Domestic cycle Peridomestic cycle

URBAN CYCLE

Fig. 3: hypothetical transmissIons patterns of Lelshmania.

logic and geographic regions. Our analyses revealed that the higher molecular diversity foundin natural populations of a given Leishmania species is related with the higher number of sand flyvector(s) and/or animal reservoir(s) involved in thetransmission cycle ofthe parasites (a co-evolutionary phenomenom?). In contrast, the L. braziliensispopulation circulating in the Brazilian Atlanticcoast showed low levels of heterogeneity, and hasLutzomyia intermedia as the principal suspectedvector. There is no apparent relationship betweenpopulation heterogeneity in Leishmania and thecapacity of the parasites to infect their hosts. Accordingly to Mayr (1973), the degree of geneticvariability would be comparatively low in thoseparasites that were naturally selected in single hosts.Leishmania parasites may infect several species ofhost, both vertebrate and invertebrate (Lainson &Shaw 1987, Grimaldi & Tesh 1993). Ifspecificityin the parasite-host relationship is important forLeishmania speciation, this process may also beinvolved in the genetic diversity found among theseorganisms.

Final comments - The strategies for the prevention and control ofleishmaniases are basicallythe interruption of the transmission cycle (vectorand/or reservoir control, personal protection, surveillance, treatment) and vaccination. The controlof these diseases is impeded by the lack of vacci-

nation or efficient treatment as weil as by the sympatric distribution of Leishmania species, thezoonotic nature of the infection and the diversityof the transmission cycles. Although it is not yetevident that the molecular heterogeneity presentin Leishmania is reflected in properties such asvirulence, insect and vertebrate host specificity,geographic range, and drug sensitivity, this polymorphism could have profound consequences forthe etiology and treatment of leishmaniases.

REFERENCES

Avise Je 1975. Systematic value ofe1ectrophoretic data.Syst Zoo123: 465-48I.

Banuls A-L, Guerrini F, Le Pont F, Barrera C, Espinel l,Guderian R, Echeverria R, Tibayrenc M 1997. Evidence for hybridization by multilocus enzyme electrophoresis and random amplified polymorphie DNAbetween Leishmania braziliensis and Leishmaniapanamensislguyanensis in Ecuador. JEuk Microbiol44: 408-41 J.

Belli AA, Miles MA, Kelly JM 1994. A putative Leishmallia panamensislLeishmania braziliensis hybridis a causative agent of human cutaneous leishmaniasis in Nicaragua. Parasitology 109: 435-442.

Bonfante-Garrido R, Meléndez E, Barroeta S, Mejia deAlejos MA, Momen H, Cupolillo E, McMahon-PrattD, Grimaldi Jr G 1992. Cutaneous leishmanisis inwestern Venezuela caused by infection with Leishmania venezue!ensis and Leishmania braziliensisvariants. l'rails R Soc Trop Med Hyg 86: 141-148.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), Sep./Oct. 199B 667

Clbulskis RE 1988. Origms and organization of geneticdiversity In natural populations of Trypanosomahrucel. Paras/tology 96: 303-322.

Cupolillo E, Grimaldi Jr G, Momen H 1994. A generalc1assllication of New World Leishmania using numerical zymotaxonomy. Am J Trop Med Hyg 50:296-311.

Cupolillo E, Grimaldi Jr G, Momen H 1997. Geneticdiversity among Leishmania (Viannia) parasites. AnnTlVp Med ParasitaI 91: 617-626.

Cupolillo E, Grimaldi Jr G, Momen H, Beverley SM 1995.Intergenic region typing ({RT): a rapid molecular approach to the characterization and evolution ofLeishmania. Mol Biochem Parasitol73: 145-155.

Darce MJ, Moran X, Palacios A, Belli F, Gomez-UrcuyoF, Zamora D, Valle S, Gantier Je. Momen H,Grimaldi Jr G 1991. Etiology of human cutaneousLeishmaniasis in Nicaragua. Trans R Soc Trop MedHvg 85: 58-59.

Delgado 0, Cupolil1o E. Bonfante-Garrido R, Silva S.Belfort E, Grimaldi Jr G, Momen H 1997. Cutaneous leishmaniasis in Venezuela caused by infectionwith a new hybrid between Leishmania (Viannia)braziliensis and L. (V.) guyanensis. Mem InstOSlValdo Cruz 92: 581-582.

Dujardin Je. Llanos-Cuentas A, Caceres A, Arana M,Dujardin JP, Guerrini F. Gomez J, Arroyo J, DeDoncker S, Jacquet D. Hamers R, Guerra H, Le RayD, Arevalo J 1993. Molecular karyotype variationin Leishmallia (Viannia) perl/viana: indication ofgeographical populations in Pern distributed alonga north-south cline. Ann Soc Belge Med Trop 87:335-347.

Dye C, Davis CR, Lines JO 1990. When are parasitesclonai? Nature 348: 120.

Evans DA, Kennedy WPR, Elbihari S, Chapman C,Smith V, Peters W 1987. Hybrid formation in thegenus Leishmania? Parassitologia 29: 165-173.

Gardner PJ 1977. Taxonomy of the genus Leishmania:a review of nomenclature and classification. TropDis Bul!etin 74: 1069-1088.

Godfrey OC 1979. The zymodemes of trypanosomes.In Problems in the identification ofparasites and theirvectors. Symp Brit Soc Parasitol 17: 31-53.

Grimaldi Jr G, Tesh RB 1993. Leishmaniases ofthe NewWorld: current concepts and implications for futureresearch. Clin Microbiol Rel' 6: 230-250.

Grimaldi Jr G, Tesh RB. McMahon-Pratt 0 1989. A review of the geographic distribution and epidemiology of Leishmaniasis in the New World. Am J TropMed Hyg 41: 687-725.

Guizani l, Van Eys GJJM, Ismail RB, Dellagi K 1994.Use of recombinant DNA probes for species identification ofOld World Leishmania isolates. Am J IJvpMed Hyg 50: 632-640.

Hurst LD, Hamilton WD, Ladle RJ 1992. Covert sex.Tree 7: 144-145.

Kelli JM, Law JM, Chapman CJ, Van Eys GJJM, EvansDA J 991. Evidence of genetic recombination inLeishmania. Mol Biochem Parasitol 46: 253-264.

Lainson R, Shaw JJ 1979. The role of animais in theepidemiology of the South American Leishmania-

sis, p. 1-116. In HR Lumsden, DA Evans (eds). Biology orthe Kinetoplastida, vol. 2, Academic Press,London.

Lainson R, Shaw JJ 1987. Evolution, c1assilication andgeographical distribution, p. 1-120. In W Peters, RKillick-Kendrick, (cds), The Leishmaniasis, ln Biology and Epidemiology, Vol. 1. Academic Press.London.

Lopes UG, Momen H, Grimaldi Jr G. Marzochi Me.Pacheco RS. Morel CM 1984. Schizodeme andzymodeme characterization of Leishmania in theinvestigation offoci ofviseeral and cutaneous Leishmaniasis. J Parasitol 70: 89-98.

Macedo AM, Melo MN, Gomes RF, Pena SDJ 1992.DNA fingerprints: a tool for identification and determination of the relationship between species andstrains of Leishmania. Mol Biochem Parasitol 53:63-70.

Mayr E 1973. The recent historiography ofgenetics (Essay review). J Hist Biol 6: 125-154.

Mendoza-Leon A, Havercroft JC. Barker OC 1995. TheRFLP analysis of the ~-tubulin gene region in NewWorld Leishmania. Parasitology 111: 1-9.

Momen H, Grimaldi Jr G 1989. Enzyme electrophoreticevidence for the importation of L. in.lantum into theNew World, p. 911-916. In DT Hart, Leishmaniasis. Plenum, New York.

Noyes HA, Belli AA, Maingon R 1996. Appraisal ofvarious random amplified polymorphic DNA-polymerase chain reaction primers for Leishmania identification. Am J Trop Med Hyg 55: 98-105.

Pacheco RS, Lopes UG, Morel CM, Grimaldi GJr,Momen H 1986. Schizodeme analysis of Leishmania isolates and comparison with sorne phenotypictechniques, p. 57-65. In J-A Rioux, Leishmania. Taxonomy and Phylogeny, IMEEE, Montpellier.

Richardson BJ, Baverstock PR, Adams M 1986.Al!ozyme Electrophoresis, Academic Press. Sydney,403 pp.

Rioux J-A, Lanotte G, Serres E, Pratlong F, Bastien P,Perieres J 1990. Taxonomy of Leishmania usingisoenzymes suggestions for a new classification. AnnParasitol Hum Comp 65: 111-125.

Selander RK, Levin BR 1980. Genetic diversity andstructure in Escherichia coli populations. Science210: 545-547.

Sibley LD, Boothroyd JC 1992. Virulent strains of Toxoplasma gondii comprise a single clonai lineage.Nature 359: 82-85.

Thomaz-Soccol V, Lanotte G, Rioux J-A. Pratlong F.Martini-Dumas A, Serres E 1993. Monophyleticorigin of the genus Leishmania Ross, 1903. AnnParasitol Hum Comp 68: 107-108.

Tibayrenc M, Ayala FJ 1988. Isoenzyme variability inTiypanosoma cruzi, the agent of Chagas's disease:genetical, taxonomical, and epidemiological significanee. Evolution 42: 277-292.

Tibayrenc M, Ayala FJ 1991. Towards a population genetics ofmicro-organisms: the clonai theory ofparasitic protozoa. Parasitol Today 7: 228-232.

Tibayrenc M, Kjelberg F, Ayala FJ 1990. A clonai theoryof parasitic protozoa: the populations structures of


Entamoeba, Giardia, Leishmania. Naegleria, Plasmodium, Trichomonas and Trypanosoma and theirmedical and taxonomical consequences. Pmc Nat!Acad Sei USA 87: 2414-2418.

Tibayrenc M, Kjelberg F, Aranaud J, Oury B, Brenière

F, Dardé ML, Ayala FJ 1991. Are eukaryotic microorganism clonaI or sexual? A population geneticsvantage. Proc Natl Acad Sei USA 88: 5129-5133.

WHO-World Health Organization 1990. Control of theLeishmaniasis, WHO Tech Rep Ser 793.


The Evolution of Trypanosomes Infecting Humansand Primates

Jamie Stevens/+, Harry Noyes*, Wendy Gibson

School of Biological Sciences, University of Bristol, Bristol BS8 1UG, UK ·Liverpool School of TropicalMedicine, Pembroke Place, Liverpool L3 5QA, UK

Based on phylogenetic analysis of 18S rRNA sequences and clade tamn composition, this paperadopts a biogeographical approach to understanding the evolutionary relationships of the human andprimate infective trypanosomes, Trypanosoma cruzi, T. brucei, T. rangel i and T. cyclops. Results indicate that these parasites have divergent origins and fundamentally difJerent patterns ofevolution. T.cruzi is placed in a clade with T. rangeli and trypanosomes specifie to bats and a kangaroo. The predominant/y South American and Australian origins ofparasites within this clade suggest an ancientsouthern super-continent origin for ancestral T. cruzL possibly in marsupials. T. brucei clusters exclusively with mammalian, salivarian trypanosomes ofAfrican origin, suggesting an evolutionary historycO'!fined to Africa, white T. cyclops. from an Asian primate appears to have evolved separately and isplaced in a clade with T. (Megatrypanum) species. Relating clade taxon composition to palaeogeographicevidence. the divergence of T. brucei and T. cruzi can be dated to the mid-Cretaceous. around 100million years bf«ore present. following the separation ofAfrica. South America and Euramerica. Suchan estimate ofdivergence time is considerably more recent than those ofmost previous studies based onmolecular clock methods. Perhaps significant/y. Salivarian trypanosomes appear, (rom these data. to beevolving several times faster than Schizotrypanum species. a factor which may have contributed toprevious anomalous estimates ofdivergence times.

Key words: Trypanosoma brucei - Trypanosoma cruzi - evolution - phylogenetics - small subunit ribosomal RNA- biogeography

The evolutionary relationships ofhuman infective trypanosomes have long been debated (Baker1963, Hoare 1972, Vickerman 1994). However, itis only recently, with advances in molecular andphylogenetic methods, that significant progresstowards resolving trypanosome evolutionary history is being made (e.g. Lake et al. 1988, Fernandeset al. 1993, Wiemer et al. 1995, Maslov et al. 1996,Lukes et al. 1997, Haag et al. 1998, Stevens et al.1998).

Prior to the advent of mo1ecular techniques,great importance was attached to the mode oftransmission as a means of understanding the evo1utionary history of trypanosomes (Baker 1963,Hoare 1972). Most vertebrate trypanosomes aretransmitted from host to host by bloodsucking arthropod or leech vectors. The trypanosomes aretaken up by the vector with a bloodmeal, and usu-

This work was supported by the Wellcome Trust (GrantNo.04713l/Z/96/Z).+Corresponding author. Fax: +44-117-925.7374. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

ally undergo one or more cycles of developmentand multiplication in the alimentary tract of theinvertebrate, before infective forms are transmitted to a new vertebrate host via saliva, contamination with faeces or ingestion of the whole vector.In this respect the human pathogenic trypanosomesdiffer markedly. T brucei, the causative agent ofAfrican human sleeping sickness, together with arange of related species of veterinary importance(T. congolense, T. simiae and T vivax), is transmitted by tsetse flies (genus Glossina) by the sa1ivarianroute. T cruzi, which causes Chagas disease inLatin America, develops in the hindgut oftriatomine bugs; infective forms are excreted in thefaeces and infect a new host by contamination ofwounds or mucous membranes - the stercorarianroute (Hoare 1972). The classification and transmission characteristics ofanother, apparently nonpathogenic species from Latin America, T. rangeli,remain much in debate.

In addition to transmission characteristics, T.brucei and T cruzi also differ in their mode of infection: T brucei resides in the bloodstream andevades the host immune response by antigenicvariation, while T cruzi is an intracellular parasiteand multiplies in tissue pseudocysts, with a tran-

670 The Evolution of Pathogenic Trypanosomes· Jamie Stevens et al.

sient bloodstream phase in the host. ft has thusbeen obvious from even the earliest parasitological studies that T bnlcei and T cruzi are very different organisms, but, just how different? Such aquestion has an important bearing on how far results relating to the biochemistry or metabolism ofone pathogenic species can be extrapolated to theother, for example, in tenns of new chemotherapeutie approaches.

To quantify the evolutionary distance betweenthe two species we have used the divergence ofthe sma11 subunit ribosomal RNA (ssu rRNA) geneto date the evolutionary split (Stevens et al. 1998).Similar molecular phylogenetic studies have previously relied heavily on "molecular c1ocks", calibrated by a variety of methods (e.g. Lake et al.1988, Fernandes et al. 1993, Haag et al. 1998).However, given the almost constant debate surrounding the accuracy of such c10cks (Sibley &Ahlquist 1984, Wilson et al. 1987), we have basedthe date of divergence of T brucei and T cruzi onbiogeographical and clade taxon composition(Nelson & Rosen 1981, Meyers & Gi11er 1988).We believe that this approach is more likely to yielda phylogenetic interpretation with biological relevance, which will contribute to an understandingof the evolution of the genus Tlypanosoma. Fina11y, the addition of more taxa, inc1uding a primate trypanosome from south-east Asia, T cyclops,has a110wed us to explore the robustness of thephylogeny and our biogeographica11y based evolutionary hypotheses.


Choice ofphylogenetic marker - The 18S ssurRNA gene was chosen as a suitable phylogeneticmarker. It is conserved throughout the eukaryotes,while the range ofconserved and variable regionsa110w diverse rates of genetic evolution to be studied, making it ideal for elucidating both higherevolutionary relationships and those betweenclosely related species (Sogin et al. 1986). Its highcopy number also facilitates ease of PCR amplification. The ssu rRNA gene has become the markerof choice for evolutionary analyses of thekinetoplastid protozoa (e.g. Fernandes et al. 1993,Maslov et al. 1994, 1996, Marché et al. 1995, Lukeset al. 1997, Haag et al. 1998).

Trypanosomes - Summary details ofa11 taxa aregiven in the Table. T. cyclops was isolated fromMacaca spp. in Peninsular Malaysia (Weinman1972). This uniquely pigmented trypanosome (described as containing large granules of pigmentderived from haemoglobin) could not readily beplaced in any existing subgenus. The vector isunknown, but transmission by reduviid bugs wasruled out (Weinman 1972).

Ribosomal RNA sequences - The ssu rRNA sequence of T cyclops was sequenced, as describedby Stevens et al. (1998). Briefly, the gene wasamplified by PCR from trypanosome templateDNA as a fragment of- 2 kb using conserved primers (Maslov et al. 1996). The products of 8-10separate PCR reactions were then purified andpooled, prior to automated sequencing in both directions at approximately 300 base pair intervalsusing 12 additional internaI primers (Maslov et al.1996) on a Perkin-Elmer ABI 377 automated sequencer. A consensus sequence was assembledfor each trypanosome strain from the internaIprimer sequences using AutoAssembler v.2.0 (Applied Biosystems, Perkin-Elmer).

Thirty-two Tlypanosoma sp. sequences wereinc1uded from Stevens et al. (1998; Table), togetherwith 15 Trypanosoma sp. sequences from theEMBLIGenBank databases. The suitability offreeliving bodonid taxa as outgroups for phylogeneticanalysis of trypanosomatids has been establishedbya number of studies using a range of ribosomaland protein coding genes (see Stevens et al. 1998).ln this study, Trypanosoma species were comparedwith a range often outgroup taxa (Bodo caudatusX53910; Trypanoplasma borreli L14840; Crithidiaspp. X03450, L29264; Leishmania spp. X53912,X07773, X53913, X53915; and Phytomonas spp.L35076, L35077).

Alignments - The T cyclops sequence was incorporated into the alignment of Stevens et al.(1998). In this alignment, ail sequences werealigned primarily to eight Trypanosoma sequencesdownloaded from the rRNA database maintainedat the University of Antwerp (Neefs et al. 1990);the alignment ofthese eight template sequences isbased on their secondary structure. Sub-sectionsof the alignment, between 'anchor' regions ofhighhomology were then sub-aligned using the programClustal V (Higgins et al. 1992), before final adjustments were made by eye. Hypervariable sites,where nuc1eotide changes were saturated, and regions where it was not possible to produce a singlereliable alignment a.cross ail 58 taxa were excludedfrom the analysis. Fo11owing this, a number ofseparate alignments, representing more or lessstringent subsets of a 'standard' alignment, wereexplored (Stevens et al. 1998) and used as the basis for the phylogenetic analysis presented in thispaper. Certain sites which were 10ca11y infonnative between closely related taxa introduced'noise', resulting in a loss of definition (reducedbootstrap support) at higher phylogenetic levels;such sites were exc1uded from the final analysis(Fig.) and the alignment used inc1uded 1801 nuc1eotide positions (available on request from JRS).

Phylogenetic analyses - Bootstrapped maxi-

Mem /nst Oswa/do Cruz, Rio de Janeiro, Vol. 93(5), 5ep.lOct. 1998 671

Location

Czech RepublicNigeriaUgandaEnglandKenyaCameroonKenyaBrazilChileBraziIBrazilMalaysiaEnglandBelgiumChinaBrazilThe GambiaThe GambiaEnglandAfricaEnglandFranceVenezuelaCanadaThe GambiaGermanyScotlandSenegalEnglandEnglandAustraliaAustraliaGermany

Fringilla coe/ehsHomo sapiensHomo sapiensNoemacheilus harhatll/usCapra sp.Capra sp.Capra sp.Homo sapiensTriatoma in{estansHomo sapiensPhyl/ostomum disc%rMacaca sp.Pipistrel/us pipistrel/usPipistre//us pipistrel/usEquus cahal/usH. hydrochaerisG.m.suhmorsitansG.p.gamhiensisRattus sp.Bufo regu/arisMicrotis agrestisMe/es me/esCanis sp.Rana catesheianaG.m.suhmorsitansBos taurusBos taurusVaranus exanthematicusPipistrel/us pipistrel/usPiscico/a geometraMacropus giganteusVombatlls IIrsinusCel"Vus dama

BirdHumanHumanFreshwater fishDomestic goatDomestic goatDomestic goatHumanTriatomine bugHumanBatMacaqueBatBatHorseCapybaraTsetse flyTsetse flyRatToadVoleBadgerDogFrogTsetse flyCattleCattleLizardBatLeechKangarooWombatDeer

TABLE

Summary details of Trypanosoma spp. analysedU

----------HostSpecies

T. avillmT. hrllcei gamhlenseT hrucel rhodesienseT. cohitisr congo/ense (kilifi)T. congo/ense (forestT. congo/ense (savannah)r cruzi (Z I)T. cnd (Z II)T. eruzi (Z III)T. cnd marinkel/eiT. cyclopsT. dionisiir dionisiir equiperdumr evansiT. godjreyiT. grayiT. /ewisiT. megar mierotiT. pestanaiT. range/ir rotatoriumT. simiaer theileriT. theilerir varaniT. vespertilionisT. sp.T. sp.T. sp.T. sp.

a: details of the 33 Trypanosoma sp. isolates sequenced as part ofthis study. See Stevens et al. (1998) for additionalisolation details; see Weinman (1972) for full details of T. cyclops. Information on the 15 additionalll)panosomasp. ssu rRNA sequences added from EMBLIGenBank can be obtained from database: T. boissoni U39580; T. carassiiLl4841; T. rotatorium U39583; T. trig/ae U39584; T. brucei hrucei M 12676; T. congo/ense (kilifi-type) U22317; T.congo/ense (forest-type) U22319; T. congo/ense (savannah-type) U22315; T. congo/ense (tsavo-type) U22318; T.simiae U22320; T. vivax U223 16; T. cruzi X53917; T. cruz; M31432; T. avium U39578; T. sce/opori U67182.

mum parsimony analysis of 58 kinetoplastid 18Sssu rRNA sequences (Table) was performed with100 replicates (Fig.); again, analyses were repeatedwith a number of more and less stringent alignments (see Stevens et al. 1998). The number oftaxa necessitated the use ofa heuristic search strategy to find the most parsimonious trees. The default options of PAUP were used: TBR branchswapping, zero length branches collapsed and 10random addition sequences (bootstrap analysesused simple addition).

Maximum-likelihood analysis was also performed; however, due to computational constraints,a reduced data set of36 taxa was employed. Taxa

were selected for inclusion in the maximum likelihood analysis so as to maximise the degree of sequence variation analysed, highly homologous sequences being excluded; starting trees were derivedby both parsimony and neighbour-joining. Transition/transversion ratios were estimated from thedata in preliminary runs and then set for full analyses. A1l analyses were performed using test version 4.0d63 of PAUp·, written by David LSwofford.

RESULTS

Phylogenetic aflalysis. The phylogram (Fig.)classifies the genus Trypanosoma into three majorclades, including the brucei-clade and the cruzi-

672 The Evolution of Pathogenic Trypanosomes' Jamie Stevens et al.

CIl

"'"'0~2

:..0

l

98

T. b brucei"T. b. rhodesienseT. b. gamb/ense "T. evansi"T. equiperdum

T. congo/ense (savannah A)T. congo/ense (savannah a)"

T. congo/ense (killfi K45.1)"T. congo/ense (kilifi WG5)

T. congo/ense (forast IL3900)"T. congo/ense (forest CAM 22b)

T. congo/ense (tsavo)T. godfreyi

T. simiae (KETRI 2431/1)"T. sim/ae (KEN 2)

'-------- T. vivax

7

T.cobitisT.carassii"

T. sp. (K&A laech)T. boissoni"

T. trig/aaT. rotatorium (B2-1)"T. rotatoflum (B2-1I)

T. maga"

cruz~clada

T. cruzi (Xl0)T. cruzi (VINCH)"T. cruzi (Can1Il)"

10 T. cruzi (M31432)"T. cruzi (No.3)T. c. marinkellei

T. dioniSiilPJ)T. dionisii P3)"

97 T. rangeliT. vesperlilionis

T. sp. (H25 kangaroo)T./awisi

T. microtiT. avium (LSHTM 144B)T. avium (A1412)"

T. grayiT. varan!

T. sce/oporiT. thei/eri (TREU 124)T. theileri (K127)"T. sp. (030 daer)"

T. cyc/ops7

10

T. sp. (H26 wombat)T. pestanai

Leishmania amazonansisLeishmania major"

9 Leishmania donovani "Le/shmania guyanensisCrithidia oncope/ti "Crithidia fascicu/ata

Phytomonas sp.Phytomonas sp:

i----Trypanop/asma borreli'--------Bodo caudatus

Phylogram constructed by bootstrapped (100 replicates) maximum parsimony analysis of 58 kinetoplastid 185 ssu rRNA sequences and rooted on the free-living kinetoplastid, Bodo cal/da/us. The tree is derived from the 28 most parsimonious trees oflength = 1135 (RI =0.8177, CI =0.5570), based an alignment of 1801 nucleotide sites. Bootstrap values for ail major nodes aregiven and ail branches receiving bootstrap support values >50% are shown; relationships faihng to achieve this level of support areshown as polytomies (i.e. branch points at which three or more branches arise from the ancestralline). Certain clades, referred toin the tex t, are defined by dashed brackets. The 22 taxa not rncluded in the maximum-likelihood analyses are marked '. Details ofail taxa are given in the Table.

clade [the brucei and cruzi clades (and ail speciestherein) are defined in the Fig. The terms 'bruceiclade' and 'cruzi-clade' are used throughout thisstudy to refer to the clades containing (a) ail mammalian Salivarian (Hoare 1972) trypanosomes(brucei-clade) and (b) trypanosomes in the subgenus Schizotrypanum, plus T. rallge/i and an as yetunidentified species of trypanosome from a kangaroo (cruzi-clade)], and six minor clades, whichtogether form a nine-way polytomy. Importantly,

the phylogram is proven to be robust and its structure is largely the same as that presented by(Stevens et al. 1998). However, the addition oftwo Phytomonas sp. to the analysis significantlyreduces phylogenetic definition at the upper levelof the Trypanosoma, such that the clade containing trypanosomes from aquatic and related hostsno longer diverges earlier than other Trypanosoma,and forms part ofa nine-way polytomy within thegenus. Such a result serves to underline the im-


portant influence that the choice of outgroup taxamay exert on phylogenetic analyses.

T. cyclops, isolated from a Malaysian primate,is classified with T. theileri and another weil characterized Megallypanum species from a deer (Fig.;100% bootstrap support), rather than in either thebrucei-clade or the cruzi-clade, which contain ailother human/primate infective trypanosomes[Johnson (1933) reported T. lewisi in the blood ofa child, also in Malaysia]. Such a result does notsupport a close phylogenetic relationship with either African or South American human/primateinfective trypanosomes, whilst its apparent lack ofovert Megatrypanum characteristics (Weinman1972) calls into question the taxonomic basis ofthe subgenus Megallypanum.

The results of the parsimony analyses wereagain strongly supported by maximum-likelihoodanalysis. The positions and branching order ofailmajor clades were identical between methods (irrespective of starting tree), and only minor variations in the positions ofcertain terminal taxa wereapparent. Indeed, the main phylogenetic relationships revealed in the tree are as described byStevens et al. (1998) and are largely robust to theaddition ofthe primate trypanosome and additionaloutgroup species.

Briefly, the phylogenetic analysis confirms themonophyly of the genus Trypanosoma with bootstrap support of63% (Fig.). The human pathogenictrypanosomes, T. brucei and T. eruzi, are placed inseparate clades, each receiving high bootstrap support of>97%. The brucei-clade, contains ail species ofmammalian Salivarian trypanosomes (Hoare1972). Except for T. evansi and T. equiperdum,these trypanosomes are ail of African origin andtransmitted by tsetse flies [analysis of kinetoplast(mitochondrial) DNA (Borst et al. 1987) and isoenzymes (Lun et al. 1992, Gibson et al. 1983) pointsto T. evansi and T. equiperdum being comparativelyrecent mutants of T. brucei, which have been ableto spread outside Africa because they no longerrely on tsetse transmission; the particularities ofthese two species are therefore irrelevant to themore ancient evolution of the clade]. The host exclusivity ofthis clade suggests a distinct evolutionary history initially confined to Africa. Trypanosomes of African origin from other vertebrates arecompletely unrelated (e.g. T. grayi, T. varani fromAfrican reptiles; T. mega from an African toad).A similar result is reported by Haag et al. (1998).

The cruzi-clade contains al! subgenus Schizot'ypanum species - T. cruzi isolates from humans,sylvatic and domestic mammals, including bats andopossums, together with trypanosomes specific toOld and New World bats, T. range/i and an as yetunidentified trypanosome species from an Austra-

lian kangaroo. The origins ofparasites within thisclade thus lie largely in South America and Australia; the only trypanosomes from this clade representing the Old World are those infecting bats.

The taxonomic and evolutionary status of human infective T. range/i (generally classified assubgenus Herpetosoma) remains controversial(D'Alessandro & Saravia 1992, Stevens & Gibson1998). In the current study T. rangeli, albeit onlya single isolate (RGB - Basel), is classified firmlyin a clade with a range of Schizotrypanllm species(bootstrap 97%); the classification of this isolateas T. rangeli is supported by preliminary resultsfrom analysis of the miniexon which indicate it tobe of the correct size and sequence according toMurthy et al. (1992). T. range/i and T. cruzi alsocluster together (bootstrap >90%) and separatefrom Salivarian trypanosomes in phylogeneticanalyses of miniexon sequences (Stevens &Gibson, unpublished data).

Phylogenetic resolution - Despite support forat least nine distinct clades within the Trypanosoma, it is not possible on the basis ofthese rRNAdata to determine the exact order in which theseclades diverged and a nine-way polytomy withinthe genus remains unresolved (Fig.). This may bedue to limitations on the resolution of the ssu geneover this time scale, to a possibly explosive divergence of trypanosome species over a very shortperiod sorne time around 100 million years beforepresent (mybp) or, as seems probable, to a combination of both these factors.

Neverthe1ess, the inclusion of a large and varied range of taxa (Swofford et al. 1996) has enabled elucidation of the complex relationships ofthe human infective trypanosomes and, while saturation ofsorne variable regions within the ssu genemay preclude the accurate determination ofbranching order at more ancient levels, considerable support for the 'correctness' of the phylogenetic re1ationships represented in the tree is provided by thelogical placement ofthe outgroup trypanosomatids,Leishmania spp., Phytomonas spp. and Crithidiaspp., which are weil separated from the Trypanosoma (Fig.).

Comparative rates ofsequence evalution - Thephylogenetic analysis provides evidence of verydifferent rates of evolution within the clades containing T. brucei and T. cruzi. Comparison ofdeeper branch lengths suggests a difference in intra-clade evolution rate of approximately 8-fold.The exact extent to which the rapid evolution ofcertain lineages within the Salivarian clade mayhave distorted the topology of the tree (and henceestimates ofevolutionary rates) is unknown. Nevertheless, the tree appears sufficiently robust tohave avoided the Salivaria being drawn towards


outgroup taxa by long-branch attraction(Felsenstein 1978, Hendy & Penny 1989), a problem encountered in many previous studies (e.g.Fernandes et al. 1993, Maslov et al. 1994, 1996).

DISCUSSION

Phylogenetic analysis ofvariation in ssu rRNAgenes of the 48 trypanosome specimens places thetwo human pathogens, T brucei and T cruzi, unequivocally in two distinct clades. The Asian primate trypanosome, T cyclops, is grouped in anapparently unrelated clade with trypanosomes fromcattle and deer.

The time ofdivergence of T brucei and T cruzican be estimated by a range ofmethods including:sequence divergence analysis (the molecular clockapproach), by reference to host phylogenies andby consideration of palaeo- and biogeographicaldata.

The concept ofa molecular clock was first proposed by Zuckerkandl and Pauling (1965). Sincethen the exact nature of the workings of such'clocks' have remained under almost constant debate (Sibley & Ahlquist 1984, Wilson et al. 1987).It is apparent that, ifthey do exist, they are at bestonly stochastically constant (Fitch 1976), and thatdifferent types of DNA sequence undoubtedlyevolve at significantly different rates. Nevertheless, within given taxonomic groups and definedcategories ofgenetic marker, the concept ofa molecular clock can provide a useful tool for understanding phylogenetic relationships.

In the study oftrypanosome evolution molecular clocks have been used by a number of authorsto attempt to date the divergence events betweenimportant taxonomic groups. In a recent and comprehensive study, Haag et al. (1998) used an estimate of0.85% substitution per 100 million years,derived from rRNA based studies ofApicomplexa(Escalante & Ayala 1995), to date the divergenceof Salivarian trypanosomes from other trypanosomes at about 300 mybp. Such a result places thedivergence of T brllcei and the Salivaria in the lateCarboniferous, at a time when the very first reptiles had just appeared in a world dominated byamphibians, and long before the appearance ofeventhe most primitive mammals. Perhaps significantly,the continents with which several major extantgroups of trypanosomes (i.e. Africa: Salivaria;South America: Schizotrypanum) are now generally associated, had not at that time even begun toseparate, but were grouped together in the solidsouthern land mass known as Gondwana (Cox &Moore 1993, Smith et al. 1994).

A second method for estimating times of divergence in parasite phylogenies is based on congruence of host and parasite phylogenies, a much

debated concept. Using this approach, parasitetrees can be calibrated by reference to time pointswithin host phylogenies, which have been constructed on the basis of independent evidence, e.g.fossils. Such an approach was also used by Haaget al. (1998), who used the divergence offish fromhigher vertebrates (400 mybp) and the divergenceofbirds from rodents (220 mybp), to estimate thesplit ofSalivarian trypanosomes from other trypanosomes at 260 and 500 mybp, respectively.Again, even the most recent of these estimatesplaces the divergence of the Salivaria somewhatunrealistically in the mid-Permian.

A third approach to calibrating organismal phylogenies is by reference to known biogeographical events - vicariance biogeography (Nelson &Rosen 1981) - and several sequence divergencebased studies of trypanosomatids have drawn onthis technique, for example, to date the divergenceofLeishmania and Trypanosoma (Lake et al. 1988)and to date the split between üld and New WorldLeishmania (Nelson et al. 1990, Fernandes et al.1993). Using this approach we previously obtaineda mid-Cretaceous date for the divergence of Tbrucei and T cruzi (Stevens et al. 1998). In summary, we suggested that the exclusively Africanmammalian tsetse-transmitted taxon complementof the brucei-clade (excepting T evansi and Tequiperdum - see above), points to an origin inAfrica. The first time at which Africa became isolated was around 100 mybp, in the mid-Cretaceous,when it finally separated from South America andEuramerica (Parrish 1993, Smith et al. 1994). Atthis time, the first mammals were present, but hadnot yet begun major diversification and it is easyto envisage subsequent coevolution of this cladewith ancient African hosts.

The cruzi-clade would thus have a southernsuper-continent (South AmericanJAntarctica/Australia) origin, an interpretation which makes senseof the inclusion of the Australian marsupial trypanosome in the clade. Indeed, the early evolutionof this clade may have been associated with thedominant marsupial fauna ofthe region. The opossum, Didelphis sp., a not so distant relative of theAustralian kangaroos (Flannery 1989), is a particularly important natural reservoir of T cruzi in SouthAmerica and can maintain a patent parasitaemiathroughout its life, with no apparent clinical symptoms (Deane et al. 1986). The only trypanosomesfrom this clade found in the üld World are thoseinfecting bats. The biological similarity of T cruziand certain species of bat trypanosomes has beenrecognised for sorne time - all are classified in thesubgenus Schizotrypanum. The present day distribution can be explained by the ability of bats todisperse over long distances, particularly across


waterbarriers and, while T (Schizotlypanum) species have been isolated from European bats on anumber of occasions (Baker 1974, Baker &Thompson 1971, Baker & Mewis 1987), reportsof T crllzi-like trypanosomes from other hosts inthe Old World are insubstantial (Hoare 1972).

In the current study, T rangeli was placedfinnly within the cruzi-c1ade. However, the taxonomic position ofthis human infective species haslong been disputed. While it appears morphologically and behaviourally similar to subgenusHerpetosoma trypanosomes (Hoare 1972), it istransmitted by both salivarian and stercorarianroutes (D'Alessandro & Saravia 1992), while Anez(1982) even separated it into a new subgenus,Tejeraia. A limited study based on B-tubulin genesequences (Amorim et al. 1993) suggested T.rangeli to be more c10sely related to T brucei thanto T cruzi. While this result cannot be disputed, itis widely recognised (Swofford et al. 1996) that studies including limited numbers of taxa spanning disparate levels ofrelatedness are highly prone to artifactual effects. Certainly, the close relationship between T rangeli and T. cruzi evident from ssu rRNAanalysis has also been confinned by comparison ofminiexon sequences (Stevens & Gibson, unpublished data). The miniexon sequence also confinnsthat the T rangeli isolate used in the current study isa bona fide T rangeli (Murthy et al. 1992).

The classification of T cyclops with otherwiseapparently unrelated T (Megatrypanum) speciesand apart from other human/primate trypanosomes,suggests that its ability to infect primates hasevolved independently (presumably in Asia) fromspecies in either of the two clades containing human infective trypanosomes.

From the separate evolutionary histories of Tbrucei and T crllzi constructed from the phylogenetic evidence, we can deduce that their pathogenicity to humans developed on very different timescales. In Africa, T brucei would have effectivelyco-evolved with hominids, since the first hominidsevolved 5-15 mybp, the genus Homo 3 mybp(Johanson & Taieb 1993) and Homo sapiens notearlier than 300 000 years bp, presumably in continuous contact with both trypanosomes and tsetseflies. In contrast, human contact with T. crllziwould not have occurred prior to human migration into the Americas, which is generally datedno earlier than 30 - 40 000 years bp. Moreover,there is no evidence for contact earlier than 3000years bp when the first pennanent settlements weremade by previously nomadic cultures (Rothhammer et al. 1985). Humans, Iike other primateswould have become infected as a simple additionto the already extensive host ranges of T cruzi andT rangeli (Hoare 1972).

Finally, to what extent their ditTerent evolutionary histories have affected intra-clade evolutionrates is unknown, however, it appears that hruceiclade species are diverging at a rate up to eighttimes faster than that observed amongst cruzi-cladespecies. Moreover, such differences in evolutionary rates between trypanosome clades are in keeping with resuIts from previous studies (Maslov etal. 1996) and are confinned by maximum-likelihood based rate analyses (Stevens, Rambaut &Gibson, unpublished data), which indicate the differences in rates of evolution between theSalivarian clade and the Schizotlypanllm clade tobe significant. Reasons for these rate differencesremain to be explored.

ACKNOWLEDGEMENTS

To Chris Schofield for vaJuable discussions on theevolutionary relationships of r cruzi.

REFERENCES

Amorim MI, Momen H. Traub-Cseko YM 1993. Trypanosoma rangeli: sequence analysis of f3-tubulingene suggests c10ser relationship to Tlypanosomabrucei than to nypanosoma cruzi. Acta Trop 53:99-105.

Aiiez N 1982. Studies on Trypanosoma rangeli Tejera,1920. IV. A reconsideration of its systematic position. Mem fnst Oswaldo Cruz 77: 405-415.

Baker JR 1963. Speculations on the evolution of thefamily Trypanosomatidae Doflein, 1901. ExpParasitai J3: 219-233.

Baker JR 1974. Protozoan parasites of the blood ofBritish wild birds and mammals. J Zool 172: 169-190.

Baker JR, Mewis GE 1987. Tlypanosoma (Schizotrypanum) sp. inde!. from a Maltese bat. Acta Trop 44:99-100.

Baker JR, Thompson GB 1971. Two species of Trypanosoma from British bats. Trans R Soc n'Op MedHyg 65: 427.

Borst P, Fase-Fowler F, Gibson WC 1987. KinetoplastDNA of Trypanosoma evansi. Mol BiochemParasitai 23: 31-38.

Cox CB, Moore PD 1993. Biogeography: An Ecological and Evolutionary Approach. 5th ed. Blackwell,Oxford, 326 pp.

D'Alessandro A, Saravia NG 1992. Trypanosomarangeli, p. 1-54. In JP Kreier, JR Baker (eds), Parasi/ic Protozoa. 2nd ed, vol. 2, Academie Press, SanDiego.

Deane MP, Lenzi HL, Jansen AM 1986. Double development cycle ofTlypano.wma cruzi in the opossum.Parasitol Today 2: 146-147.

Escalante AA, Ayala FJ 1995. Evolutionary origin ofPlasmodium and other Apicomplexa based on rRNAgenes. Proc Natl Acad Sei USA 92: 5793-5797.

Felsenstein J 1978. Cases in which parsimony or compatibility will be positively misleading. Syst Zool27: 401-410.

Fernandes AP, Nelson K, Beverley SM 1993. Evolutionof nuclear ribosomal RNAs in kinetoplastid proto-


zoa: Perspectives on the age and origins ofparasitism. Proc Natl Acad Sci USA 90: 11608-11612.

Fitch WM 1976. Molecular evolutionary docks, p. 160178. ln FJ Ayala, Molecular Evolution, Sinauer,Sunderland, MA.

F1annery TF 1989. Phylogeny of the Macropodoidea; astudy in convergence, p. 1-46. In G Grigg, P Jarman,1 Hume (eds), Kangaroos. Wallabies and Rat-kangaroos, vol. 1, Surrey Beatty & Sons, ChippingNorton, NSW.

Gibson WC, Wilson AJ, Moloo SK 1983. Characterisation of Trypanosoma (Tlypanozoon) evansifrom camels in Kenya using isoenzyme electrophoresis. Res Vet Sei 34: 114-118.

Haag J, O'Huigin C, Overath P 1998. The molecularphylogeny of trypanosomes: evidence for an earlydivergence ofthe Salivaria. Mol Biochem Parasitol91: 37-49.

Hendy MD, Penny D 1989. A framework for the quantitative study of evolutionary trees. Syst Zool 38:297-309.

Higgins DG, Bleasby AJ, Fuchs R 1992. Clustal V: improved software for multiple sequence alignment.Comp Applns Biosei 8: 189-191.

Hoare CA 1972. The Trypanosomes of Mammals,Blackwell, Oxford, 749 pp.

Johanson DC, Taieb M 1976. Plio-Pleistocene hominiddiscoveries in Hadar, Ethiopia. Nature 260: 293-297.

Johnson PD 1933. A case of infection by Tl)panosomalewisi in a child. Trans R Soc Trop Med Hyg 26:467.

Lake JA, De La Cruz VF, Ferreira PCG, Morel C,Simpson L. 1988. Evolution of parasitism:Kinetoplastid protozoan history reconstructed frommitochondria1 rRNA gene sequences. Proc Natl AcadSci USA 85: 4779-4783.

Lukes J, Jirku M, Dolezel D, Kral'ova 1, Hollar L,Maslov DA 1997. Analysis ofribosomal RNA genessuggests that trypanosomes are monophyletic. J MolEvo144: 521-527.

Lun Z-R, Allingham R, Brun R, Lanham S 1992. Theisoenzyme characteristics of Trypanosoma evansiand Tlypanosoma equiperdum isolated from domestic stocks in China. Ann Trop Med Parasitol 86:333-340.

Marché S, Roth C, Philippe H, Dollet M, Baltz T 1995.Characterization and detection of plant trypanosomatids by sequence analysis of the small subunit ribosomal RNA gene. Mol Biochem Parasitol71: 15-26.

Maslov DA, Avila HA, Lake JA, Simpson L 1994. Evolution of RNA editing in kinetoplastid protozoa.Nature 368: 345-348.

Maslov DA, Lukes J, Jirku M, Simpson L 1996. Phylogeny of trypanosomes as inferred from the smalland large subunit rRNAs: implications for the evolution ofparasitism in the trypanosomatid protozoa.Mol Biochem Parasitol 75: 197-205.

Meyers AA, Giller PS 1988. Analytical Biogeography,Chapman & Hall, London, 578 pp.

Murthy VK, Dibbern KM, Campbell DA 1992. PCRamplification ofmini-exon genes differentiates Try-

panosoma cruzi from T. rangeli. Mol CellularProbes 6: 237-243.

Neefs J-M, Van De Peer Y, Hendriks L, De Wachter R1990. Compilation ofsmall ribosomal subunit RNAsequences. Nucleic Acids Res 18: 2237-2243.

Nelson G, Rosen DE 1981. Vtcariance Biogeography:A Critique, Columbia University Press, New York,500 pp.

Nelson K, Alonso G, Langer PJ, Beverley SM 1990.Sequence ofthe dihydrofolate reductase-thymidylatesynthase (DHFR-TS) gene of Leishmaniaamazonensis. Nucleic Acids Res 18: 2819.

Parrish JT 1993. The palaeogeography of the openingSouth Atlantic, p. 8-27. In W George, R Lavocat(eds), The Africa-South America Connection, Oxford Monographs on Biogeography No. 7, ClarendonPress, Oxford.

Rothhammer F, Allison MI, Nunez L, Standen V, ArriazaB 1985. Chagas disease in pre-Colombian SouthAmerica. AmerJ Physical Anthropology 68: 495-498.

Sib1ey CG, Ahlquist JE 1984. The phylogeny of thehominoid primates, as indicated by DNA-DNA hybridization. J Mol Evo120: 22-15.

Smith AG, Smith DG, Funnell BM 1994. Atlas ofMesozoic and Cenozoic Coastlines, Cambridge University, Cambridge, 99 pp.

Sogin ML, Elwood HJ, Gunderson JH 1986. Evolutionary diversity of eukaryotic small-subunit rRNAgenes. Proc Natl Acad Sci USA 83: 1383-1387.

Stevens JR, Gibson WC 1998. The evo1ution ofSalivarian trypanosomes. Second World Meeting onSalivarian Trypanosomes, Tryplink internet conference, EMBRAPAI Centro de Pesquisa Agropecuariado Pantanal, Brazil.

Stevens JR, Noyes HA, Dover GA, Gibson WC 1998.The ancient and divergent origins of the humanpathogenic trypanosomes, Trypanosoma hrucei andT. cruzi. Parasitology (in press).

Swofford DL, Olsen GJ, Waddell PJ, Hillis DM 1996.Phylogenetic inference, p. 407-514. In DM Hillis,C Moritz, BK Mable (eds), Molecular Systematics,Sinauer Associates, Sunderland, MA.

Vickerman K 1994. The evolutionary expansion of thetrypanosomatid flagellates.lnt J Parasitol24: 13171331.

Weinman D 1972. Trypanosoma cyclops n. sp.: a pigmented trypanosome from the Malaysian primatesMacaca nemestrina and M. ira. Trans R Soc TropMed Hyg 66: 628-636.

Wiemer EAC, Hannaert V, Van Den Ijssel PRLA, VanRoy J, Opperdoes FR, Michels PAM 1995. Molecular analysis ofglyceraldehyde-3-phosphate dehydrogenase in Trypanosoma horelli: An evolutionaryscenario of subcellular compartmentation inKinetoplastida. J Mol Evol 40: 443-454.

Wilson AC, Ochman H, Prager EM 1987. Molecu1artime scale for evolution. Trends Genet 3: 241-247.

Zuckerkandl E, Pauling L 1965. Evolutionary divergenceand convergence in proteins, p. 97-166. In V Bryson,HJ Vogel (eds), Evolving Genes and Proteins, Academie Press, New York.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 677-683, Sep./Oct. 1998 677

Genetic Data Showing Evolutionary Links betweenLeishmania and Endotrypanum

Elisa Cupolillo/+, Luiza OR Pereira, Octavio Fernandes*, Marcos P Catanho*,Julio C Pereira**, Enrique Medina-Acosta**, Gabriel Grimaldi Jr

Laboratorio de Leishmaniose, Departamento de Imunologia *Departamento de Medicina Tropical, InstitutoOswaldo Cruz, Av. Brasi14365, 21045-900 Rio de Janeiro, RJ, Brasil **Laboratorio de Biotecnologia,

Universidade Estadual do Norte Fluminense, 28015-620 Campos, RJ, Brasil

Striking similarities at the morphological. molecular and biological levels exist between manytI}'panosomatids isolatedfrom sylvatic insects and/or vertebrate reservoir hosts that make the identification ofmedically important parasites demanding. Some molecular data have pointed to the relationship between some Leishmania species and Endotrypanum, which has an important epidemiologicalsignificance and can be helpful to understand the evolution of those parasites. In this study, we havedemonstrated a close genetic relationship between Endotrypanum and two new leishmanial species, L.(V.) colombiensis and L. (V.) equatorensis. We have used (a) numerical zymotaxonomy and (b) thevariability ofthe internaI transeribed spacers ofthe rRNA genes to examine relationships in this group.The evolutionary trees obtained revealed high genetie similarity between L. (V.) colombiensis, L. (V.)equatorensis and Endotrypanum, forming a tight cluster ofparasites. Based on further results of (e)minicircle kDNA heterogeneity analysis and (d) measurement of the sialidase aetivity these parasiteswere also grouped together.

Key words: Leishmania colombiensis - Leishmania equatorensis - Endotrypanum - multilocus enzyme electrophoresis - molecular characterization - numerical analysis - sialidase activity - kDNA

Parasitic protozoa of the genus Leishmania(Kinetoplastida: Trypanosomatidae) are biologically diverse group ofmicroorganisms. Taxonomiestudies ofleishmanial isolates from the New Worldindicate tremendous diversity within this genus(Cupolillo et al. 1995). A number of new Leishmania species from sylvan areas ofthe Neotropicsare associated with disease in humans; others appear to be restricted to lower orders of mammals,such as rodents and edentates (Grimaldi et al.1989).

Sloths are reservoir hosts ofat least five namedLeishmania species of the subgenus Vzannia [1.guyanensis Floch, 1954; 1. panamensis Lainson& Shaw 1972; 1. shawi Lainson et al. 1989; 1.eolombiensis Kreutzer et al. 1991 and 1.equatorensis Grimaldi et al. 1992], responsible for

This work was supported by grants of the InstitutoOswaldo Cruz-Fiocruz, FAPERJ and CNPq to EC, OFand GG Jr. Research work at the Laboratorio deBiotecnologia-UENF is supported by grants fromFENORTE, FINEP, FAPERJ and WHOfTDR to EM-A.+Corresponding author. Fax: +55-21 -280.1589. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

human cutaneous and/or mucosal leishmaniasis(Grimaldi & Tesh 1993). Infections with other biologically distinct groups oftrypanosomatid protozoa, such as Endotrypanum and Trypanosoma arealso found in sloths (Deane 1961, Pipkin 1968,Travi et al. 1989, Shaw 1992).

In nature, ail Leishmania spp. are transmittedby the bite of infected phlebotomine sand flies(Diptera:Psychodidae). However, many flagellatesother than Leishmania commonly are found insand flies in Neotropical forests. Arias et al. (1985)identified E. schaudinni and other Endotrypanumsp. infections in sand flies and sloths captured inthe Amazon Region of Brazil. Results of kinetopIast DNA probe identifications ofpromastigotespresent in sand flies captured near Manaus, Brazil also demonstrated Endotrypanum infections inLu. shannoni, as weil as in Lu. umbratilis and Lu.anduzei (Rogers et al. 1988). Further evidence forthe development of Endotrypanum inphlebotomines was obtained by feeding severallaboratory-reared sand fly species on infectedsloths (Christensen and Herrer 1976, 1979, Shaw1981 ).

Endotrypanum spp. are digenetic trypanosomatids in that they are intraerythrocytic parasites ofsloths and are transmitted by phlebotominesand flies (Shaw 1992). Endotrypanum sharesmany other charecteristics with Leishmania. Cul-

678 Evolutionary Links between Leishmania and Endotrypanum • Elisa Cupolillo et al.

tured-derived promastigotes of parasites in bothgenera are morphologically similar. Studies employing monoclonal antibodies for the analysis ofthe genus Endotrypanum have shown antigenicsimilarites between these parasites and sorne Leishmania species (Franco et al. 1997). Furthemore,molecular trees clustered the sandfly-borne digenetic parasites Leishmania and Endotrypanum together, sharing a common ancestor and representing a relatively recent lineage from theTrypanosomatidae family (Fernandes et al. 1993).Results of hybridization using kDNA probes(Pacheco et al. 1990) support the view thatEndotrypanum and the peripylarian leishmanialparasites ofthe subgenus Viannia Lainson & Shaw1987 are phylogenetically close (Shaw 1992). Inaddition, phylogenetic studies have demonstratedthat the most divergent Leishmania species are L.(L.) hertigi and L. (L.) herreri, c1aimed to be closerto Endotr)panum than to the other Leishmania(Croan & Ellis 1996, Noyes et al. 1996, 1997,Croan et al. 1997).

In this study, we have shown evolutionary linksbetween Endotlypanum and sorne leishmanialparasites based on their molecular genetics, ascharacterized using a broad assemblage of methodologies. The data presented here demonstratethat E. schaudinni, L. (V) colombiensis and L. (V).equatorensis form a tight phylogenetic cluster, anevolutionary linked group that should be exploredto understand the origin(s) of neotropical pathogenic Leishmania.


Parasites - Leishmania and Endotrypanum(Table 1) were cultured in Scheneider's Drosophilamedium (Gibco, Grand Island, NY) supplementedwith 10% heat-inactivated FBS (Biolab, Rio deJaneiro, Brazil) at 24°C. In the preparation ofsamples, the parasite (promastigotes in the latephase of growth cultures) were harvested by centrifugation (3,800x g for 15 min at 4°C) and washedtwice in saline pH 8.0, containing the appropriatebuffer.

BiochemicallMolecular characterization - Theprocedures used for characterizing the parasites(multilocus enzyme electrophoresis - MLEE, measurement of the sialidase activity, PCR amplification and restriction enzyme digestion of the parasite ITSrRNA, cloning and sequencing ofthe conserved region of the minicircle kDNA molecules)have been described in detail in previous publications (Cupolillo et al. 1994, 1995, Medina-Acostaet al. 1994, Fernandes et al. 1996). Sialidase activity was measured using a single-cell HITACHI F4500 spectrofluorometer (350 nm excitation and460 nm emission wavelengths). The sequencing

was performed in automatic sequencing(AbiPrisma, Applied Biosystem).

Numerical analysis - The MLEE data was anaIyzed by phenetic methods using the NTSYS software program (version 1.7, exeter software). Principal coordinate analysis was performed based onEuclidian distance between the samples. The similarity level between the Leishmania species andEndotrypanum was calculated using the Jaccard'scoefficient. The kDNA sequences of the parasiteswere analyzed using the MEGA program (Kumaret al. 1993). The number of differences betweenthe sequences were calculated and a similarity treeconstructed by the Neighbor-Joining method. Bootstrap analysis was based on 500 replicates.


Leishmania and Endotrypanum are very closeprotozoan parasites (Fernandes et al. 1993) commonly found in the same vertebrate and insecthosts. Recents studies have showing the relationship between Endotrypanum and sorne New WorldLeishmania species, mainly those from L. (L.)hertigi and L. (L.) helTeri complex (Croan & Ellis1996, Noyes et al. 1996, 1997, Croan et al. 1997).Moreover, DNA analysis of phylogenetically informative RNA polymerase II gene of L. (V)equatorensis and Endotrypanum demonstrated sequence similarities among these parasites (JJ Shaw,pers. commun.). Similarly, it appears that a closeantigenic links may exist between L. (V)colombiensis, L. (V) equatorensis and Endotrypanum (Franco et al. 1997, Grimaldi et al. 1992).

Leishmania (V) colombiensis was found infecting humans, sloths (Choloepus hoffmanni),sandflies (Lu. hartmani and Lu. gomezi), and dogsin Colombia, Panama, and Venezuela (Kreutzer etal. 1991, Delgado et al. 1993, unpublished data).L. (V) equatorensis is an enigmatic parasite, whichwas isolated from the viscera of a sloth (c.hoffmanni) and a squirrel (Sciurus granatensis),captured in humid tropical forest on the PacifieCoast of Ecuador. Data based on biological andmolecular criteria, as well as numericalzymotaxonomy analysis indicated that both theseparasites are clearly distinguishable from all otherknown speeies, but clustered within the L. (V)braziliensis complex (Kreutzer et al. 1991,Grimaldi et al. 1992). Multilocus enzyme electrophoresis data and the restriction fragments of theinternaI transcribed spaeers of the rRNA gene(Cupolillo et al. 1995, 1997) have indicated a closerelationship between L. (V) equatorensis and L.(V) colombiensis, as previously demonstrated(Kreutzer et al. 1991, Grimaldi et al. 1992). In order to better understand their taxonomie positionin the genus, especially in relation to the discrimi-


TABLE 1

Origin and Identification of Leishmania and Endotr)panum strains used in this study

Stock number DeslgnationU Species Geographie origin

L565 MHOM/BRl75/M4147 L. guyanensis Brazil, ParaL566 MHOM/BRlOO/M2903 L. braziliensis Brazil, ParaL575 IFLA/BRJ67/PH8 L. amazonensis Brazil, ParaL579 MHOM/BRl74/PP75 L. chagasi Brazil, BahiaL888 MCHO/EC/82/Lspla L. eqllatorensis Ecuador, GuayasL889 MSCIIEC/82/Lsp2 L. equatorensis Ecuador, GuayasLI023 MHOM/BRJ811M6426 L. lainsoni Brazil, ParaLI 245 IGOM/PA/85/E582.34 L. colombiens!s Panama, ColonLI246 IPAN/PA/85/E696.26 L. colombiensis Panama, ColonLI 247 IGOM/PA/85/E582.36 L. colombiensis Panama, ColonLI545 MHOM/BRJ84/M8408 L. shaw! Brazil, ParaLI365 MDAS/BRJ79/M5533 L. naiffi Brazil, ParaE14 MCHO/BRJ80/M6159b E. schaudinni Brazil, Para

a: host [M=Mammalia: CHO=Choloeplis sp. (ac. hoffmanni, bc. didactylus), DAS=Dasypus novemcinctus,HOM=Homo sapiens, SCI=Scilirus granatensis; I=Insecta: FLA=Lutzomyia flaviscultelata, GOM= Lu. gomezi,PAN=Lu. panamensis]/country of origin/year of isolation/original code.

nation of Leishmania from Endotrypanum andevolutive studies we decided to analyze the genetic similarity among these parasites, using several biochemical and molecular methods. This information will help define the fundamental mechanisms involved in species identification and taxonomic divergence among these microorganism.

The sialidase (EC 3.2.1.18) activity alone hasbeen shown to be a good marker to discriminatebetween morphologically indistinguishable flagellates isolated from human, insects and sylvaticvertebrate reservoir hosts, such as Leishmania andEndotrypanum (Medina-Acosta et al. 1994). Thegeneral concensus is that Endotrypanum referencestocks express clear-cut varying levels of sialidaseactivities whereas the Leishmania reference stocksdo not. In this study, we measured the sialidaseactivity for several neotropical Leishmania speciesand for reference strain E. schaudinni. As expected,Endotrypanum exhibited high levels of sialidaseactivity, whilst the taxonomically unquestionableLeishmania stocks (i.e., L. chagasi) were negativefor this activity. However, high levels of sialidaseactivity were consistently obtained from both cellIysates and culture supernatants of L. (V)colombiensis and L. (V) equatorensis, levels comparable with those obtained for E. schaudinni (thiswork) and those of Trypanosoma rangeli and Trypanosoma leeuwenhoeki (Medina-Acosta et al.1994).

Further, MLEE analyses demonstrated that L.(V) colombiensis and L. (V) equatorensis sharealleles with Endotrypanum for sorne loci, such asG6PDH and IDHNAD, that were previously ad-

mitted as monomorphic for the latter genus and asdiscriminative characters between Leishmania andEndotrypanum (Franco et al. 1996). Moreover, forthe malic enzyme were found two distinct loci(ME 1 and ME2) for L. (V) equatorensis and L.(V) colombiensis, as described for Endotrypanumbut in contrast to other leishmanial parasites(Cupolillo et al. 1994, Franco et al. 1996). According to the phenetic analyses, the resuits showed ahigh level of similarity between the two Leishmania species, as weil as a close relationship betweenthis group and Endol1ypanum (Fig. 1, Table II).The later parasite is genetically closest to L. (V)colombiensis rather than to L. (V) equatorensis(Table II). In addition, the clusters L. braziliensis/L. naiffi andL. guyanensis/L. shawi were observed,as already demonstrated (Cupolillo et al. 1994,1997) and L. lainsoni made a link between L. (V)colombiensis/L. (V) equatorensislE. schaudinniand the Leishmania (Viannia) species.

The Neighbor-Joining tree constructed basedon kDNA sequence data using the number of differences between Leishmania and Endotrypanumshows similar clustering of MLEE for L. (V)equatorensis/L. (V) colombiensislE. schaudinni(Fig. 2). The position ofL. lainsoni was maintained,forming a link between the group L. (V)equatorensis/L. (V) colombiensislE. schaudinniand other Leishmania species. Leishmania (V)lainsoni represents a very divergent monophyleticViannia species, which was clustered as an independent complex (Thomaz-Soccol et al. 1993,Cupolillo et al. 1994, Fernandes et al. 1995, Ereshet al. 1995).


L shawi

/L guyal1ensis

L braziliel1sis

\L naiffi

LeqUalOI~

L colombiensis

V'''''"'': E. schoudinni

Fig. (: principal coordinate analysis of the multilocus enzyme eleClTophoresis data. The three principal coordinales were claculatcdby Euclidian distance and plotted in 3D scale (the three principal coordinates represent 67.51% of the total variance). A minimumspanning tree was superimposed on the ordinations.

TABLE II

Similarity level among Leishmania and Endotrypanum species caJculated by the Jaccard's coefficient

2 3 4 5 6 7 8

1.1. braziliensis2.1. guyanensis 0.213.1. lainsoni 0.22 0.164. 1. equatorensis 0.13 0.08 0.115. 1. colombiensis 0.16 0.07 0.14 0.406. E. schaudinni 0.10 0.13 008 0.18 0.237.1. shawi 0.24 0.59 0.13 0.07 0.05 0.078. 1. naiffi 0.47 0.15 0.26 0.12 0.18 0.12 0.20

Endorrypanum schaudinii

1. equatorensis

1. colombiensis

L. lainsolli

98

97 ......-----

85 ,-------"-------"'-1

0.01

L. braziliensis

L. shawi

L.chagasi

L. amazonensis

Vianniasubgenus

)Leishmaniasubgenus

Fig. 2: phenetic analyze of sequences (83bp) of conserved region of kDNA minicircle. The similarities were evaluated by thenumber of differences among the sequences and the similarity tree constructed by the Neighbor-joining method. ltalic numbersrepresent bootslTap values based on 500 replicates.


The intemal transcribed spacers of the rRNAgene were amplified by PCR and the product digested with several restriction enzymes (Cupolilloel al. J995). The RFLP profiles show a close butnot identical pattern between L. (V) colomhiensisand L. (V) equatorensis. However, through thismethod Endotrypanum can be easily discriminatedfrom the former parasites and other Leishmaniaspecies by most of the restriction enzyme profiles(Fig. 3).

The genetic similarity between Endotrypanumand New World Leishmania was also demonstratedby sequencing comparisons of the small subunitof ribosomal RNA and RNA Polymerase II genes(Croan & Ellis 1996, Noyes et al. 1996, 1997,Croan et al. 1997). The results show that L. (L.)herreri (Zeledon et al. 1975), a sloth parasite, iscloser to Endotrypanum than to other Leishmaniaspecies. Leishmania (L.) hertigi/L. (L.) deanei(Herrer 1971, Lainson & Shaw 1977), which wereisolates from rodents, are also genetically closestto the Endolfypanum/L. (L.) herreri group (Croanet al. 1997, Noyes et al. 1997). Sorne authors suggest that L. (L.) herreri is a misclassified parasiteand therefore probably represents Endotrypanum(Croan & Ellis 1996). Although L. (L.) hertigi andL. (L.) deanei are still enigmatic parasites (Lainson1997) there are evidences supporting their classification as Leishmania. An interesting aspect is thatthese Leishmania species and Endotrypanum arebiologically distinct parasites and do not share the

same hosts.In contrast ta L. (V) colombiensis, which has

been isolated From humans (Kreutzer et al. 1991,Delgado et al. 1993), the public health importanceof L. (V) equatorensis remains to be determined.To date, it has only been isolated from arborealmammals; no human infections with the parasitehave been identified. Likewise, the sandfly vector(s) are unknown. However, the biologicalbehaviour of L. (V) equatorensis is indistinguishable from other members of the L. (V) braziliensiscomplex, based on its virulence and developmentin laboratory animais. Inoculation of culturedpromastigotes into the nose of hamster(Mesocricetus auratus) produced local swellingwithout metaslasis; appearance of the lesions took1-3 months, depending on the size of the inoculum (Grimaldi et al. 1992). Moreover, the restriction profile of the intemal transcribed spacers ofthe rRNA gene showed a close pattern between L.(V) equatorensis and L. (V) colombiensis, but distinct from Endotrypanum, supporting the taxonomic status of the former parasite, and that thetwo Leishmania species represent a link betweenEndotrypanum and Leishmania.

Comparative studies will be needed to addressthe antiquity of this evolutionary link group and,in p3liicular, whether or not it represents a branchpoint on the origin of neotropica lleishmanias. Il isworth noting that sloths, which have always beenrestricted to the American continent, are cons id-

A

on ~00'" ...:::: ~ ~ ~LU ....J -J ...J

B

Fig. 3: restriction enzyme profile of the internaI transcribed spacers of the rRNA genes for Leishmania species and Endotrypanumschaudinni. A. BstUl; B. Taq l.


ered to have evolved trom the basic Xenarthranarmadillo-like stock sorne 60 million years agaduring the Palaeocene period. These early mammals separated between the two and the three-toedgroups of extant sloths later during the Mioceneperiod. With this in mind, we feel that theneotropical leishmanias may weil have evolvedfrom a primitive endotrypanumal Miocene parasite line of South American sloths.

REFERENCES

Arias JR, Miles MA, Naiff RD, Povoa MM, de FreitasRA, Biancardini CB, Castellon EG 1985. Flagellateinfections ofBrazilian sand flies (Diptera: Psychodidae): isolation in vitro and biochemical identification of Endotrypanum and Leishmania. Am J TropMed Hyg 34: 1096-1108.

Christensen H, Herrer A 1976. Neotropical sandflies(Diptera:Psychodidae), invertebrate hosts ofEndotlypanum schaudil1l1i (Kinetoplastida). J MedEntomol 13: 299-303.

Christensen H, Herrer A 1979. Susceptability ofsandflies(Diptera:Psychodidae) to trypanosomatidae fromtwo-toed sloths (Edentata: Bradypodidae). J MedEntomol 16: 424-427.

Croan D, Ellis J 1996. Phylogenetic relationships betwecn Leishmania, Viannia and Sauroleishmania inferred from comparison ofa variable domain withinthe RNA polymerase 1 1 largest subunit gene. MolBiochem Parasitol 79: 97-102.

Croan DG, Morrison DA, Ellis JT 1997. Evolution ofthe genus Leishmania revealed by comparison ofDNA and RNA polymerase gene sequences. MolBiochem Parasitol 89: 149-159.

Cupolillo E, Grimaldi Jr G, Momen H 1994. A generalclassification of New World Leishmania using numerical zymotaxonomy. Am J Trop Med Hyg 50:296-311.

Cupolillo E, Grimaldi Jr G, Momen H 1997. Geneticdiversity amongLeishmania (Viannia) parasites.AnnTrop Med Parasitol 91: 617-626.

Cupolillo E, Grimaldi Jr G, Momen H, Beverley SM1995. Intergenic region typing (IRT): a rapid molecular approach to the characterization and evolution of Leishmania. Mol Biochem Parasitol 73:145-155.

Deane LM 1961. Tripanosomatideos de mamiferos daregiào Amazônica. 1. Alguns flagelados encontradosno sangue de mamiferos silvestres do Estado do Para.Rev 1nst Med Trop Sào Paulo 3: 15-28.

Dclgado 0, Castes M, White Jr AC, Kreutzer RD 1993.Leishmania colombiensis in Venezuela. Am J TropMed Hyg 48: 145-147.

Eresh S, de Brujin MHL, Mendoza-Leon lA, Barker DC1995. Leishmania (Viannia) lainsoni occupies aunique niche within the subgenus Viannia. Trans RSoc Trop Med Hyg 89: 231-236.

Fernandes AP, Nelson K, Beverley SM 1993. Evolutionof nuclear ribosomal RNAs in kinetoplastid protozoa: Perspectives on the age and origins ofparasitism. Proc Natl Acad Sei USA 90: 11608-11612.

Fernandes 0, Bozza M, Pascale JM, Miranda AB, LopesUG, Degrave WM 1996. An oligonucleotide probederived from kDNA minirepeats is specific for Leishmania (Viannia). Mem 1nst Oswlado Cruz 91: 279284.

Fernandes 0, Pacheco RS, Momen H, Degrave W,Campbell DA 1995. Leishmania lainsoni: a pecuIiar Viannia species. Ann Trop Med Hyg 89: 81-82.

Floch H 1954. Leishmania tropica gllyanensis n.sp. agentde la leishmaniose tegumentarie de Guyanes et del'Amerique Centrale. Arch 111st Pasteur de LaGuvane Française et du Teritoire de L 'lnni 15: 328.

Franco"AMR, Machado GMC, NaiffRD, Moreira CFS,McMahon-Pratt D, Grimaldi Jr G 1997. Characterization of Endotrypanum parasites using specificmonoclonal antibodies. Mem 1nst Oswaldo Cruz 92:63-68.

Franco AMR, Momen H, Naiffi RD, Moreira CFS,Deane MP, Grimaldi Jr G 1996. Enzyme polymorphism in Endotrypanllm and numerical analysis ofisoenzyme data. Parasitology 113: 39-48.

Grimaldi Jr G, Tesh RB 1993. Leishmaniases ofthe NewWorld: current concepts and implications for futureresearch. Clin Microbiol Rev 6: 230-250.

Grimaldi Jr G, Kreutzer RD, Hashigushi Y, Gomez EA,Mimory T, Tesh RB 1992. Description of Leishmania eqllatorensis sp.n. (Kinetoplastida: Trypanosomatidae), a new parasite infecting arboreal mammals in Ecuador. Mem 1nst Oswaldo Cruz 87: 221228.

Grimaldi Jr G, Tesh RB, McMahon-Pratt D 1989. A review of the geographic distribution and epidemiology of Leishmaniasis in the New World. Am J TropMed Hyg 41: 687-725.

Herrer A 1971. Leishmania hertigi sp. n. from the tropical porcupine, Coendou rothschi/di Thomas. JParasitol 57: 626-629.

Kreutzer RD, Corredor A, Grimaldi lr G, Grogl M,Rowton ED, Young DG, Morales A, McMahon-PrattD, Guzman H, Tesh RB 1991. Characterization ofLeishmania colombiensis sp.n. (Kinetoplastida:Trypanosomatidae), a new parasite infecting humans,animais, and phlebotomine sand flies in Colombiaand Panama. Am J Trop Med Hyg 44: 662-675.

Kumar S, Tamura K, Nei M 1993. MEGA: Molecularevolutionary analysis, version 1.01. The Pennsylvania State University, Philadelphia, PA.

Lainson R 1997. On Leishmania enriettii and other enigmatic Leishmania species of the neotropics. Mem1nst Oswaldo Cruz 92: 377-387.

Lainson R, Shaw JJ 1972. Leishmaniasis in the NewWorld: taxonomie problems. Brit Med Bull 28: 4448.

Lainson R, Shaw 11 1977. Leishmania of neotropicalporcupines: Leishmania hertigi deanei nov. subsp.Acta Amazonica 7: 51-57.

Lainson R, Braga RR, De Souza AAA, Povoa MM,Ishikawa EAY, Silveira FT 1989. Leishmania(Viannia) shawi sp. n., a parasite of monkeys, slothsand procyonids in Amazonian Brazil. Ann ParasitolHum Comp 64: 200-207.

Medina-Acosta E, Franco AMR, lansen AM, Sampol


M. Nevés N, Pontes-de-Carvalho L, Grimaldi Jr G,Nussenzweig V 1994. Trans-sialidase and sialidaseactivities discriminate between morphologically indistinguishable trypanosomatids. Eur J Biochem225: 333-339.

Noyes HA, Arana BA, Chance ML, Maingon R 1997.The Leishmania hertigi (Kinetoplastida; Trypanosomatidae) complex and the lizard Leishmania:their classification and evidence for a neotropicalorigin ofthe Leishmania-Endo/lypanllm clade. JEukMicrobiol44: 511-517.

Noyes HA, Camps AP, Chance ML 1996. Leishmaniahen'eri (Kinetoplastida; Trypanosomatidae) is morec10sely related to Endotlypanllm (Kinetoplastida;Trypanosomatidae) than to Leishmania. MolBiochem Parasitol 80: 119-123.

Pacheco RS, Thomaz N, Momen H 1990. kDNA crosshybridization between Endotrypanum andperipylarian Leishmania. Trans R Soc Trop Med Hyg84: 531.

Pipkin AC 1968. Domicillary reduviid bug and the epidemiology ofChagas'disease in Panama (Hemiptera:Reduviidae: Triatominae).JMed Entomol5: 107-124.

Rogers WO, Bumheim PF, Wirth DF 1988. Detection

of Leishmania within sandflies by kinetoplast DNAhybridization. Am J Trop Med Hyg 39: 434-439.

Shaw JJ 1981. The behaviour of Endotrypanumschaudinni (Kinetoplastida: Trypanosomatidae) inthree species oflaboratory-bred neotropical sandflies(Diptera:Psychodidae) and its influence of the classification of the genus Leishmania, p. 232-241. InEU Canning, Parasitological TOpICS, Allen,Lawrence, KS.

Shaw JJ 1992. EndO/lypanllm, a unique intraerythrocyticflagellate of New World tree sloths. An evolutionary link or an evolutionary backwater? Ciên Cult44: 107-116.

Thomaz-Soccol V, Lanotte G, Rioux J-A, Prationg F,Martini-Dumas A, Serres E 1993. Monophyleticorigin of the genus Leishmania Ross, 1903. AnnParasitol Hum Comp 68: 107-108.

Travi BL, Zea A, D'Alessandro A 1989. Tlypanosoma(Herpetosoma) leeuwenhoeki in Choloepushoffmanni and Didelphis marsupialis of the pacifiecoast of Colombia. J Parasitol 75: 218-224.

Zeledon R, Ponce C, De Ponce E 1975. The isolation ofLeishmania braziliensis from sloths in Costa Rica.Am J Trop Hyg 24: 706-707.


A Study of Cryptosporidium parvum Genotypes andPopulation Structure

G Widmer/+, l Tchack, FSpano*, S Tzipori

Tufts University School ofVeterinary Medicine, Division ofinfectious Diseases, North Grafton, Massachusetts,USA *Istituto di Parassitologia, Università di Roma "La Sapienza", Rome, Italy

Genetic evidencefor the occurrence oftwo Cryptosporidium parvum subgroups is presented. Thisevidence is based on restriction fragment length polymorphism analysis ofseveral independent loci.Sequence analysis of the fi -tubulin intron revealed additional po/ymorphism. The stabi/ity ofthe genetic profiles fo/lowing passage ofe. parvum iso/ates between d~fJerent hosts was investigated.

Key words: Cryptosporidium parvum - restriction fragment length polymorphism - tubulin

Cryptosporidium parvum is an enteric protozoan parasite which commonly infects immunosuppressed individuals. Ruminants, in particularcalves, are important reservoirs. Recent genotypicanalyses of C. parvum from human cases ofcryptosporidiosis have identified two groups ofgenotypically distinct parasites. One ofthese genotypes, designated genotype C, infects animais andhumans, whereas the other, known as genotype H,is only found in humans. Differences in infectivity between H and C isolates were found in animalmodels. These observations have led to the hypothesis that C. parvum is transmitted via differenttransmission routes, each transmitting parasites ofone genotype. An alternative view is that bothgenotypes circulate among different host species,and that genotypically different populations canarise from mixed infections through selection indifferent host environments.

MATE RIALS AND METHODS

DNA purification - PCR amplification was performed either on DNA isolated directly from stoolor extracted from purified oocysts. For stool DNAextraction, 100 to 200 ~I of stool was incubatedovernight in 0.2% SDS and 200 ~g/ml proteinaseK, extracted with phenollchloroform and ethanolprecipitation. Alternatively, oocysts were purifiedfrom staal and DNA recovered by proteinaseK/SDS treatment.

Restriction fragment /ength po/ymorphism Multilocus RFLP was performed using four unlinked RFLP markers; polyT (Carraway et al.1997), COWP (Spano et al. 1997), TRAP-C 1

+Corresponding author. Fax: +508-839.7977. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

(Spano et al. 1998) and RNR (Widmer et al. 1998).A sequence-specific PCR assay aimed at the ribosomal internai transcribed spacer 1 (Carrawayetal. 1996) was also used.

Isopycnic {ractionation ofoocysts- Semi-purified oocysts were sedimented on a 15-30% (w/v)Nycodenz (Sigma) for 1 hr at 55,000xg. Fractionsofapproximately 1 ml were recovered and oocystsconcentrated by centrifugation.


In order to investigate the epidemiology of C.parvum, we have developed PCR and PCR-RFLPmarkers. Several coding and a non-coding regionwere examined for sequence polymorphism. Using a combination ofpolymorphic markers developed in our laboratories, C. parvum isolates originating from different host species and differentgeographicallocations were subject to a multilocusgenotypic analysis. Isolates were found to segregate into H (41 %), C (52%) and mixed (7%) typeisolates. Significantly, in a sample of 29 isolatesno recombinants were identified, suggesting reproductive isolation between H and C parasites.

RFLP and sequence analysis of a non-codingregion (the p-tubulin intron) identified a high degree ofpolymorphism (Fig.). A multiple sequencealignment of cloned PCR products spanning thep-tubulin intron and part of exon 2, revealed fourgroups ofsequences and additional polymorphismwithin groups. Sequences indicative ofinterallelicrecombination were found in two isolates.

The population structure of isolates seriallytransmitted through caIves or passaged from caIvesto mice, human to mice or calves to humans wasexamined. Several infections showing changes inRFLP profiles following seriai transmission wereobserved. Using isopycnic fractionation ofoocysts,it was possible to separate, in the calf-propagatedisolate GCH l, two populations ofoocysts bearingdifferent genotypes.

686 A Study of C. parvum Genotypes and Populations Structure· G Widmer et al.

~CDCDoN

300-

200--

100--

Restriction site polymorphism in the beta-tubulin gene ofCryplosporidium parvum. PCR produets amplified from theinlTon and adjacent exon 2 were digesled wilh restriction enzyme Tsp5091. Three restriction profiles were detected amongthese samples; one in the bovine isolate GCH 1and TAMU, onein human isolale 06761 and one in human isolates 2066K, 2458Land 0541 L.

These observations indicate that the epidemiology of C. parvum is complex and that individualhosts can excrete heterogeneous populations ofoocysts. The significance of these findings forhuman cryptosporidiosis has not been elucidated.Ofprimary interest is the question whether isolatesof genotype H and C differ in virulence and susceptibility to drug treatment.

REFERENCES

Carraway M, Tzipori S, Widmer G 1996. Identificationof genetic heterogeneity in the Cryptosporidiumparvum ribosomal repeat. Appl Environ Microbiol62: 712-716.

Carraway M, Tzipori S, Widmer G 1997. New RFLPmarker in Cryptosporidium parvum identifies mixedparasite populations and genotypic instability in response to host change. Infect Immun 65: 3958-3960.

Spano F, Putignani L, McLaughlin J, Casemore OP,Crisanti A 1997. PCR-RFLP analysis of theCryptosporidium oocyst wall protein (COWP) genediscriminates between C. wrairi and C. parvum, andbetween C. parvum isolates of human and animalorigin. FEMS Microbiol Letters 152: 209-217.

Spano F, Putignani L, Naitza S, Puri C, Wright S, CrisantiA 1998. Molecular cloning and expression analysisof a Cryplosporidium parvum gene encoding a newmember of the thrombosponding family. MolecBiochem Parasitol 92: 147-162.

Widmer G, Tzipori S, Fichtenbaum CJ, Griffiths JK1998. Genotypic and phenotypic characterization ofCryptosporidium parvum isolates from people withAlOS. J Infect Dis: in press.


Species and Strain-specific Typing of CryptosporidiumParasites in Clinical a.nd Environmental Samples

Lihua Xiao/+, Irshad Sulaiman, Ronald Fayer*, Altaf A Lai

Division of Parasitic Diseases, National Center for Infectious Diseases, Centers for Disease Control and Prevention, US Department of Health and Human Services, Atlanta, GA 30341 *Parasite Immunobiology Laboratory,

Agriculture Research Service, U.S. Department of Agriculture, Beltsville, MD 20705, USA

Cryptosporidiosis has recent/y attracted attention as an emerging waterborne andfoodborne disease as weil as an opportunistic infection in HIV infected individua/s. The /ack ofgenetic information.howevel; has resu/ted in confusion in the taxonomy of Cryptosporidium parasites and in the deve/opment ofmo/ecu/ar too/sfor the identification and typing ofoocysts in environmenta/ samp/es. Phy/ogenetic ana/ysis of the small subunit ribosoma/ RNA (SSU rRNA) gene has shown that the genusCryptosporidium comprises severa/ distinct species. Qur data show the presence ofat /eastfour species:e. parvum, e. muris, e. baileyi and e. serpentis (e. meleagridis, e. nasorum and e. felis were notstudied). Within each species. there is some sequence variation. Thus, various genotypes (geno(vpe l,genorype 2, guinea pig genotype, monkey genorype and koala genotype, etc.) ofe. parvum differfromeach other in six regions of the SSU rRNA gene. Information on po/ymorphism in Cryptosporidiumparasites has been used in the deve/opment ofspecies and strain-specific diagnostic too/s. Use ofthesetools in the characterization ofoocysts in various samp/es indicates that e. parvum genotype 1 is thestrain responsib/efor most human Cryptosporidium infections. In contrast, genorype 2 is probab~v oneof the major sources for environmenta/ contamination, and has been found in most oysters examinedfrom Chesapeake Bay that may serve as biologic monitors ofestuarine waters.

Key words: Cryptosporidium - phylogeny - genotype - ribosoma1 RNA

Cryptosporidiosis is a coccidian infection ofhumans, domestic animaIs and other vertebrates.In young farm animaIs, especially preweaned dairycalves, it causes a severe enteritis resulting in significant morbidity, mortality and economic loss.In humans, it results in an acute infection of thedigestive system in immunocompetent individuais, and chronic, life-threatening disease inimmunocompromised patients. Several transmission routes, including person-to-person, contamination of water or food, and zoonotic infection,are possible. The specifie source ofOyptosporidium oocysts involved in infection orcontamination is frequently unknown, largely dueto a lack of detailed epidemiologic investigationand strain-typing tools. The latter results from a

This work was supported in part by inter-agency agreements (#DW7593 7730-01-0 and DW7593784-0 1-0)from CDC and EPA, and Emerging Infectious Diseasesand Opportunistic Infectious Diseases funds from CDC,USA.+Corresponding author. Fax: +770-488-4454.Received 15 June 1998Accepted 30 July 1998

CUITent paucity of molecular characterization andlack of acceptance of the taxonomy of Cryptosporidium species and genotypes.

CRYPTOSPORlDlUM SPECIES

Since the discovery of Oyptosporidium murisand C. parvum in rodents, over 20 Cryptosporidiumspecies have been described in various animal hosts(0'Donoghue 1995). Species were named basedon the historical be1ief that Cl}ptosporidium spp.are coccidian parasites, and therefore share thestrict host specificity demonstarted by many othercoccidian parasites. Studies conducted in late 1970sand early 1980s, however, indicated that sorne isolates of O}ptosporidium were infectious for several animal species. Thus, one group of investigators suggested that ail Cryptosoridium parasiteswere the same species, C. muris (Tzipori et al.1980). Others demonstrated that host specificitywas present among isolates from different classesofvertebrates (O'Donoghue 1995). Based on theseobservations, Levine (1984, 1986) classified theparasites from mammals, birds, reptiles and fishas C. muris, C. me/eagridis, C. serpentis, and C.nasorum, respectively. Subsequent studies demonstrated that C. parvum from mammals and C.bai/eyi from birds were biologically and morphologically different from C. muris and C. meleagridis

688 CryptosporidlUm Species and Strain • Lihua Xiao et al.

(Upton & Current 1985, Current et al. 1986). Thus,e. parvum, e. muris, e. baileyi, e. meleagridis,e. serpel1tis and e. nasorul11 were considered validCryptosporidium species (O'Donoghue 1995).More recently, based on published reports of hostspecificity, Fayer et al. (1997) added e. relis fromcats and e. wrairi from guinea pigs to 'the list ofvalid species, whereas Tzipori and Griffiths(1998)suggested that current evidence does notsupport the concept that there is more than onespecies of Ctyptosporidium parasites.

The lack of genetic information and the presence of erroneous sequences in a few publishedstudies have added to the present state oftaxonomieconfusion. Cai et al. (1992) compared the smallsubunit (SSU) ribosomal RNA (rRNA) gene, andshowed a greater than 99% identity between onee. parvum and one e. muris isolate. Alignment ofsequences (accession numbers X64430 to 64343)from that study with sequences from us and othersindicates that ail four sequences from Cai et al.(1992) are the e. muris type. Minor sequence errors (one insertion and 12 deletions ofnucleotides)were found in the SSU rRNA sequence (L25642)of another published study (Kilani & Wenman1994). These sequences and five other sequencesdeposited in the GenBank were used recently byTzipori and Griffiths (1998) in a phylogeneticanalysis of Cryptosporidium parasites. Based onthis analysis, they concluded that the observed inter-species and intra-species variation did not favor the designation of separate Ct}'Ptosporidiumspecies, and therefore aU CJyptosporidium oocysts,including those from lower vertebrates, should beconsidered hazardous to humans.

We have recently sequenced the SSU rRNAgenes from various iso\ates ofe. parvum, e. muris,e. baileyi and e. serpentis, and used these sequences in a phylogenetic analysis (Xiao et al.unpub. data). Results of the analysis indicate thatCJ)'Ptosporidium parasites are a multi-species complex containing at least four species: e. parvum,e. baileyi, e. muris and e. selpentis (e. felis, e.nasorum and e. meleagridis were not studied). Theevolutionary distance between theCJyptosporidium guinea pig isolate and e. parvumis too small to warrant a separate species designation.

CRYPTOSPORIDIUM PARVUM GENOTYPES

Results of various studies indicate that there isvariation within the species e. parvum. Two dimensional gel electrophoresis has revealed minordifferences between human and bovine e. parvumisolates (Mead et al. 1990), which has been confirmed by immunoblot (Nichols et al. 1991, Ninaet al. 1992), isozyme (Ogunkolade et al. 1993,

Awad-EI-Kariem et al. 1995), and restriction fragment length polymorphism (RFLP) analysis(Ortega et al. 1991). More recently, random amplified polymorphie DNA (RAPD) markers haverevealed two distinct groups of human e. parvumisolates, one containing most human isolates andthe other containing sorne human isolates and ailanimal isolates (Morgan et al. 1995), indicating thepossibility of zoonotic infection. Similar resultshave been obtained by sequence data or PCR-RFLPanalysis of a repetitive sequence (Bonnin et al.1996), bifunctional dihydrofolate reductasethymidylate synthase (DHFR) (Vasquez et al 1996),rRNA repeats (Carraway et al. 1996),polythreonine motifs (Carraway et al. 1997), 00

cyst wall protein (COWP) gene (Spano et al. 1997),and thrombospondin anonymolls protein-2 (TRAPC2) gene (Peng et al. 1997 Sulaiman et al. unpub.data). It remains unclear, however, whether thesame two genotypes are present in ail these polymorphie loci. Results of our multi-Iocus analysissuggest that indeed the same genotypes are linkedacross ail polymorphie genes (SSU rRNA, TRAPCI, TRAP-C2, CPI5, and B-tubulin intron) examined (Xiao et al. unpub. data).

Our phylogenetic analyses of the SSU rRNAgene have revealed diversities in e. parvum notpreviously observed (Table 1). Human e. parvumisolates differ from bovine isolates in four regionsof the SSU rRNA gene. Likewise, theCryptosporidium isolate from guinea pigs (e.wrairi) also differs from the bovine isolates in fourregions, two of which are the same polymorphieregions between the human and bovine genotypes,thus representing a third genotype of e. parvum.Partial sequences obtained from a monkey by usand from a koala by Morgan et al.( 1997) indicatethe presence oftwo additional genotypes. The difference between the human and bovine genotypesin nucleotides 689-699 has also been observed recently by Morgan et al. (1997). We, however, haveobserved that sorne human isolates have the sequence TTTTTT instead of TTTTTTTTTTT.Based on a partial SSU rRNA gene sequence, another group also identified a new e. parvum genotype (Carraway et al. 1994, 1996). The new genotype sequence (ICP), however, is identical to thee. muris bovine isolate (Xiao et al. unpub. data).

CRYPTOSPORIDIUM GENOTYPES IN CLiNICALSAMPLES

Results of the molecular characterization havebeen used by us in the development of moleculardiagnostic tools. A PCR-RFLP technique based onthe polymorphism in the TRAP-C2 gene was developed and used in the analysis ofhuman clinicalsamples from various outbreak and non-outbreakcases (Sulaiman et al., unpub. data). Results ofour


TABLE 1

Differences among genotypes of Cryptosporidillm parvlIm in the SSU rRNA gene

Location of mutations in the SSU rRNA gene{/

Genotype 129-135 179-]84 262-267 639-656 689-699 795-800

TTTTACT AAACTC AATTAA AAAATATTTTGATGAATA TTTTTTTTTTT TTTTTTorTTTTTT

2 TTT-ACT AAACTC ATTAAA AAAATATTTTGATGAATA TATATTTT TTTCTT

wralfl TTT-ACT AGGCCC ATAAAT ATAATATTTTGAA-AATA TATATTTTT TTTCTT

Monkey unknown unknown AATTAA AATATATTTTGATGAATA TTTTTTTTT TTTTTT

Koalah unknown unknown unknown ATTATACTTTTTAAGGTG TATTTTTTT unknown

a: nuc1eotide positions in the aligned sequences of ail Cryptosporidlum species. Actual positions in individualsequences may vary slightly due to the introduction of gaps in the aligned sequences (1757 bp); h: based on thesequence by Morgan et al. (1997).

studies and those by others (Table Il) indicate thatanthroponotic organisms account for the majorityof the cases and person-to-person transmission islikely to be an important transmission route ofcryptosporidiosis in non-outbreak cases. This isevident from the large number of genotype 1parasites in sporadic cases and HIV patients (Sulaimanet al., unpub. data). This is in agreement with sornerecent observations by others (Table II). Even inoutbreak cases, many cryptosporidiosis outbreaksare caused by anthroponotic (genotype 1) parasites (such as the waterbome outbreaks in Milwaukee in 1993, Nevada in 1994, and Florida in 1995;the Atlanta day care outbreak in 1995, and theWashington outbreak in 1997). It is possible thatgenotype 2 parasites largely cause human infection through contamination ofwater or food or direct contact with infected animais, especia11y inrural areas. Examples are the Maine apple cideroutbreak in 1993, the British Columbia waterbomeoutbreak in 1996, and the Pennsylvania multi-family outbreak in 1997. The reason for the high percentage of genotype 2 in AIDS patients (6/13 patients) in France (Bonnin et al. 1996) is not clear.Taken together, there are two distinct populationsof C. parvum parasites, one cycling only in humans and one cycling predominantly in animais.The latter can cause human infections.

CRYPTOSPORlDlUM PARASITES IN ENVIRONMENTALSAMPLES

One difficu Ity facing the investigation ofwaterborne outbreaks ofcryptosporidiosis is the lack ofa sensitive, specific diagnostic tool. Most of thecurrent PCR diagnostic and genotyping tools aredesigned for analysis of clinical samples. Becausethey cannot differentiate Oyptosporidium speciesand have low sensitivities, they have limitations inthe analysis of water samples. Two PCR-RFLP

techniques based on the SSU rRNA gene haveclaimed to differentiate C. parvum from otherOyptosporidium parasites (Awad-El-Kariem et al.1994, Leng et al. 1996). One technique (Leng etal. 1996) used conserved sequences for primers andtherefore amplify the SSU rRNA gene of a11 eukaryotic organisms. The other technique (AwadEI-Kariem et al. 1994) used erroneous sequenceby Cai et al. (1992) as primers, reducing the efficiency of amplification and making interpretationof the data difficult. Nor have the presentgenotyping techniques been subjected to cross-species testing, making interpretation of results fromenvironmental samples that could contain nonparvum Cryptosporidium virtually impossible.

Based on sequence information on the SSUrRNA gene, we have developed a PCR-RFLP technique for both species identification andgenotyping ofOyptosporidium parasites. Becausethe technique employs nested PCR and targets themulti-copied rRNA gene, it has sufficient sensitivity for use in environmental sampIes. We haveused this technique in the analysis ofOyptosporidium oocysts recovered from the gillwashings and hemolymph of oysters (Crassostreavirginica) collected from the Chesapeake Bay. Weare interested in oysters because they are filter feeders that concentrate and accumu lateCryptosporidium oocysts they have removed fromsurface waters. The use ofoysters enables investigators to avoid the poor recovery rate often associated with filtering hundreds of liters ofwater todetermine the presence or absence ofCryptosporidium oocysts. Before applying ourtechnique Cryptosporidium oocysts were morpho10gical1y identified in oysters, but the species ofmost of the oocysts was unconfirmed (Fayer et al.1998).

690 CryptospondlUm Species and Strain • Lihua Xiao et al.

TABLE Il

Prevalence of genotype 2 in human clinical samples reported in various studies

Location Sample source # of genotypes Technique Referencesamples 1/2 used

England & Guinea Bissau Sporadic cases t 1 10/1 Isozyme Awad-EI-Kariem et al. 1995

Western & South Australia Sporadic cases 14 12/2 RAPD Morgan et al. 1995

USA Sporadic cases 3 2/1 ITSI and SSU Carraway et al. 1996rRNA repeat

Northeast France HIV+ patients 13 617 PCR-RFLP of Bonnin et al. 1996repetitive DNA

UK Sporadic cases 7 5/2 PCR-RFLP Spano et al. 1997of oocyst wallprotein

Western Australia Sporadic cases 32 28/4 PCR ofRAPD Morgan et al., 1997fragment

USA & Canada Outbreaks & 16 13/3 TRAP-C2 Peng et al. 1997sporadic cases sequencing

USA, Canada, India Outbreaks & 50 42/8 PCR-RFLP Sulaiman et al. unpub. data& Guatemala sporadic cases ofTRAP-C2

Preliminary analysis of 65 pooled oystersamples using the SSU rRNA-based PCR-RFLPtechnique has shown the presence ofCryptosporidium oocysts in 26 samples. Twentyfour ofthese positive samples were C. parvum, andeach of the others was C. baileyi and C. selpentis.The majority of Oyptosporidium oocysts were ofgenotype 2 (22 samples), indicating animais maybethe most likely the source ofmost O}ptosporidiumoocyst contamination in the Chesapeake Bay.Even though this is a highly populated area, onlytwo samples had genotype 1 sequences. Theseresults demonstrate that oysters can serve as a biologie monitor for Oyptosporidium oocyst contamination in waters. Because raw oysters are often consumed by humans, Oyptosporidium oocysts inoysters also pose a potential health concern. Otherfilter-feeders such as freshwater clams and marinemussels have also been shown to accumulateO:vptosporidium oocysts (Graczyk et al. 1998,Chalmers et al. 1997). They may serve as similarbiologie monitors for Cryptosporidium oocyst contamination.

CONCLUSIONS

Although the traditional classification of species based on the vertebrate classes of their hostsis largely accurate, it has greatly underestimatedthe diversity various Cryptosporidium isolates. Thishas presented problems in the identification ofparasites in environmental samples. Molecular tech-

niques are now available to identify species ofCryptosporidium and to differentiate known genotypes of C. parvum, and should be very useful inthe investigation of clinical outbreaks ofcryptosporidiosis. The perfonnance ofthese techniques in the analysis of environmental samples,however, has yet to be thoroughly demonstrated.Because of the nature of environmental samples,Ctyptosporidium isolates from various hosts mustbe more extensively characterized before enoughdata have been acquired and interpreted to instillfull confidence in the method.

ACKNOWLEDGMENTS

To lascIf Umor for technical assistance.

REFERENCES

Awad-EI-Kariem FA, Robinson HA, Dyson DA, EvansD, Wright S, Fox MT, McDonald V 1995. DifTerentiation between human and animal strains ofOyptosporidillnl parVlIm using isoenzyme typing.Parasitology 110: 129-132.

Awad-EI-Kariem FA, Warhurst DC, McDonald V 1994.Detection of Cryptosporidillm oocysts using a sytembased on PCR and endonuclease restriction. Parasitology 109: 19-22.

Bonnin A. Fourrnaux MN. Dubremetz JF, Nelson RG,Gobet P, Harly G, Buisson M, Puygauthier-ToubasD, Gabriel-Pospisil F, Naciri M, Camerlynch P 1996.Genotyping human and bovine isolates ofOyptosporidillm parvum by polymerase chain reaction-restriction fragment length polymorphismanalysis of a repetitive DNA sequence. FEMS

Mem Insl Oswaldo Cruz, Rio de Janeiro, Vol. 93(5), 5ep.lOct. 1998 691

Microhiol Left 137: 207-211.Cai J, Collins MD, McDonald V, Thompson DE 1992.

PCR c10ning and nucleotide sequence determinationof the 18S rRNA genes and internaI transcribedspacer 1of the protozoan parasites O)'PtosporidiumparVllm and O~\'Ptosporidium muris. BiochemBlOphys Acta 1131: 317-320.

Carraway M, Widmer G, Tzipori S 1994. Genetic markers differentiate C. parvum isolates. J EukalJ'otMicrobiol41: 26S.

Carraway M, Tzipori S, Widmer G 1996. Identificationof genetic heterogeneity in the O:vptosporidillmparvum ribosomal repeat. Appl Env Microhiol 62:712-716.

Carraway M, Tzipon S, Widmer G 1997. A new restriction fragment length polymorphism fromCrJ'Ptosporidium parvum identifies genetically heterogeneous parasite populations and genotypicchanges following transmission from bovine to human hosts.lnfect Immun 65: 3958-3960.

Chalmers RM, Sturdee AP, Mellors P, Nicholson V,Lawlor F, Kenny F, Timpson P 1997.Oyptosporidium parvum in environmental samplesin the Sligo area, Republic of Ireland: a preliminaryreport. Left Appl Microhiol25: 380-384.

Current WL, Upton SJ, Haynes TB 1986. The life cycleof Cryptosporidium haileyi n. sp. (Apicomplexa,Cryptosporidiidae) infecting chickens. J Protozool33: 289-296.

Fayer R, Graczyk TK, Lewis EJ, Trout JM, Farley CA1998. Survival of infectious CI~vptosporidillm

pan'ulll oocysts in seawater and eastern oysters(Crassostrea virginica) in the Chesapeake bay. ApplEnviron MiCl'ohiol 64: 1070-1074.

Fayer R, Speer CA, Dubey JP 1997. The general biology of c,:vptosporidium, p. 1-4 J. In R Fayer,Oyptosporidium and Cryptosporidiosis, CRC Press,Boca Raton, Florida.

Graczyk TK, Fayer R, Cranfield MR, Conn DB J998.Recovery of waterhorne Cryptosporidium parvumoocysts by freshwater benthic clams (Corbiculajluminea). Appl Envimn Microbiol64: 427-430.

Kilani RT, Wenman WM 1994. Geographical variationin 18S rDNA gene sequence of Cryptosporidiumparvum. Int J Parasitol 24: 303-306.

Leng X, Mosier DA, Oberst RD 1996. DifferentiationofClyptosporidium parvum, C. muris, and C. baileviby PCR-RFLP analysis of the 18s rRNA gene. VetParasitol62: 1-7.

Levine ND 1984. Taxonomy and review of the coccid-ian genus c,yptosporidium (Protozoa,Apicomplexa). J Pmtozool31: 94-98.

Levine ND 1986. The taxonomy of Sarcocystis (Protozoa, Apicomplexa) species. J Parasitol72: 372-382.

Mead JR, Humphreys RC, Sammons DW, Sterling C1990. Isolation ofisolate-specific sporozoite proteinsof Cryptosporidium parvum by two-dimensional gel

electrophoresls. Infect 11I1111un 58: 2071-2075.Morgan UM, Constantine Cc. Forbes DA, Thompson

RCA 1997. Differentiation between human and animai isolates ofClyptosporidium lJaI1'um using rDNAscquencing and direct PCR analysis. J Parasitol 83:825-R30.

Morgan UM, Constantine Cc. O'Donoghue P, MeloniBP, O'Brien PA, Thompson RCA 1995. Molecularcharacterization of Cryptospondiulll isolates fromhumans and other animais uSll1g random amplifiedpolymorphic DNA analysis. Am J Trop Mec! f~l'g

52: 559-564.Nichols GL, McLauchlin J, Samuel 0 1991. A technique

for typing ClJ'Ptosporidilllll isolates. J Protozool38:237S-240S.

Nina JMS, McDonald V, Dyson DA, Catchpole J, UniS, Iseki M, Chiodini PL, McAdam KPWJ 1992.Analysis ofoocyst wall and sporozoite antigens fromthree Cry'Ptosporidiuln species. Infect Immun 60:1509-1513.

O'Donoghue PJ 1995. Cryptosporidilllll andcryptosporidiosis in man and animals.lntJ Parasitol25: 139-195.

Ogunkolade BW, Robinson HA, McDonald V, WebsterK, Evans DA 1993. Isoenzyme variation within thegenus ClJlptosporidium. Parasitol Res 79: 385-388.

Ortega YR, Sheehy RR, Cama VA, Oishi KK, SterlingCR 1991. Restriction fragment length polymorphismanalysis of Cryptosporidillm parvum isolates ofbovine and human origin. J Protozool 38: 540-541.

Pcng MP, Xiao L, Freeman AR, Arrowood MJ,Escalante A, Weltman AC, Ong C, Mac KenzieWR, Lai AA, Beard CB 1997. Genetic polymorphism among CI}'Ptosporidium parl'um isolatessupporting two distinct transmission cycle. sEmerInfect Dis 3: 1-9.

Spano F, Putignani L, McLauchlin J, Casemore OP,Crisanti A 1997. PCR-RFLP analysis of theClyptosporidium oocyst wall protein (COWP) genedicriminates between C. \Vrairi and C. parvum, andbetween C. parvllm isolates of humand animal origin. FEMS Microbiol Left 150: 209-217.

Tzipori S, Angus KW, Campbell 1, Gray EW 1980.Clyptosporidium: evidence for a single-species genus. Infect Immun 30: 884-886.

Tzipori S, Griffiths JK 1998. Natural history and biology of Cryptosporidium parl'um. Adv Parasitol40:5-36.

Upton SJ, Current WL 1985. The species ofCryptosporidium (Apicomplexa: Cryptosporidiidae)infecting mammals. J Parasitol 7/: 625-9.

VasquezJR, Gooze L, Kim K, Gut J, Petersen C, NelsonRG 1996. Potential anti folate resistance deterrninantsand genotypic variation in the bi functionaldihydrofolate reductase-thymidylate synthase genefrom human and bovine isolates ofCryptosporidiumparvum. Mol Biochem Parasitol79: 153-165.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 693-694, 5ep./Oq. 1998 693

RESEARCH NOTE

Human Genetic Bi-allelicSequences (HGBASE), aDatabase of 1ntra-genic

Polymorphisms

Chandra Sarkar/+ , Flavio R Ortigâo,Ulf Gyllensten*/**,

Anthony JBrookes**

Interactiva Biotechnologie GmbH, D-89077 Ulm,Gennany *Swedish Genome Research Center

**Department of Genetics and Pathology, BiomedicalCenter, Uppsala, Sweden

Key words: single nucleotide polymorphisms polymorphisms - intra-genic polymorphisms

databases - bioinfonnatics

The Human Genome Project is providing awealth of information about the human gene repertoire, and promises to furnish a complete genomesequence (and thereby a complete gene catalog)by the year 2005. This enormous output of data isbeginning to be complemented by large scale studies designed to uncover normally occurring variations within human gene sequences. Much ofthisvariability is very subtle, often comprises singlenucleotide polymorphisms (SNPs) which are ideally compatible with a number of large scale detection procedures. SNPs will be the basis of future highly dense polymorphic marker maps, andthose related to known genes can be exploited ingenetic association studies aimed at defining thegenetic basis ofail manner ofcomplex phenotypes,not least disorders such as mental illness, diabetes,cardiovascular disease and cancer. Ali indicationsare that 100,000-200,000 human genome SNPs willbe identified within the next two years.

In light of the above developments, a databaseof gene based polymorphisms is obviously required. To fulfill this need we have constructed andrecently released at http://hgbase.interactiva.de theHGBASE (human genic bi-allelic sequences) da-

+Corresponding author. Fax: +49-731-93579291. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

tabase of intra-genic sequence polymorphism.HGBASE is the result of a joint venture betweenUppsala University Medical Genetics Department,the Swedish Genome Research Centre, andInteractiva Biotechnologie GmbH. Its primary purpose is to facilitate genotype-phenotype association studies based upon the rapidly growing number ofknown, gene related, single nucleotide polymorphisms (SNPs) and other intra-genic sequencevariations. Furthermore, HGBASE will help towards the production of a dense SNP map of thehuman genome, which itselfwill be a valuable research too!.

HGBASE is not designed to include gene 'mutations', but instead is a catalog of intra-genic (promoter to transcription end point) sequence variants found in 'normal' individuals. Although thedistinction between 'mutation' and 'variation' canbe somewhat blurred, the general idea is that thecontent of HGBASE concerns frequently occurring 'normal polymorphisms', whether or not theyare suspected to increase the risk of developing aparticular phenotype. This is in contrast to 'mutant sequences' which are known to cause geneticdisease. Despite its name, HGBASE contains aIltypes of intra-genic variation and is not limited tobi-allelic polymorphisms (though these do represent most ofthe database content). Both functionalpolymorphisms (e.g. promoter and non-silentcodon changes) and non-functional polymorphisms(e.g. intron sequence differences) are included. Thisis for two reasons. Firstly, it is often difficult to becertain about the functional consequence ofa variation. Secondly, regardless offunctional relevance,any intra-genic polymorphism can usually be employed as an effective surrogate marker for an unknown functional variant in an association study,due to close proximity and linkage disequilibrium.

Gene polymorphisms may be retrieved fromHGBASE by using the database search facilitiesto query either by a text string or by a DNA sequence. Data submission to HGBASE is madesimple by provision of a series of Web page datasubmission forms. Ali submitted data is made available to any other public database that wishes todownload it, and continuai efforts are made to access new relevant data from other databases andliterature publications. The exponential growth inpolymorphism discovery requires that scientistsmake every effort to submit their data to theHGBASE database to ensure it remains up to date.HGBASE does not claim any rights to publiclyavailable or submitted data, instead this remainsthe property of the original submitter. Depositionofdata into HGBASE requires only the allelic DNAsequences, the allele frequencies, the host genename, and the intra-genic domain. Additional com-

694 HGBASE, a Database of Intra-genic Polymorphisms • Chandra Sarkar et al.

ments, such as assay conditions, can be suppliedthough are not required. The submitted data is presented in HGBASE along with the submitters nameand contact details to aid discussion and questioning. Database curators will subsequently enhancethe submitted data by adding links to other databases, and by adding information concerning gene

function, gene location, gene expression pattern,disease associations, and suggested assay formats.This 'added value' data is accessible to users following a simple registration procedure that is freeto academia but for which a charge is made to industry to cover the costs of collecting and maintaining the additional data.


Selection, Recombination and History in a ParasiticFlatworm (Echinococcus) Inferred from Nucleotide

Sequences

KL Haag, AM Araujo, B Gottstein*, A Zaha**

Departamento de Genética, Universidade Federal do Rio Grande do Sul, Caixa Postal 15053,91501-970 PortoAlegre, RS, Brasil "Institute of Parasitology, University of Berne, Laengass Strasse 122, PO Box 8466, Berne,CH-300 1. Switzerland .... Departamento de Biotecnologia, Universidade Federal do Rio Grande do Sul, Caixa

Postal 15005, 91501-970 Porto Alegre, RS, Brasil

Three species offlatworms from the gemls Echinococcus (E. granulosus, E. rnultilocularis and E.vogeli) and four strains of E. granulosus (cattle, horse, pig and sheep strains) were analysed by thePCR-SSCP methodfollowed by sequencing, using as tm'gets two non-coding and two coding (one nuclearand one mitochondrial) genomic regions. The sequencing data was used to evaluate hypothesis aboutthe parasite breeding system and the causes ofgenetic divers{fication. The calculated recombinationparameters suggested that cross:(ertilisation was rare in the history ofthe group. However, the relativerates ofsubstitution in the coding sequences showed that positive selection (instead q(purifying selection) drove the evolution q(an elastase and neutrophi/ chemotaxis inhibitor gene (AgBII). The phylogenetic analyses revealed several ambiguities, indicating that the taxonomie status q( the E. granulosushorse strain should be revised.

Key words: Echinococcus - parasites - recombination - SSCP - sequencing - phylogeny

Several new insights about the evolution ofhelminth parasites came out during the last years.Echinococcus, a parasite that causes one ofthe mostimportant and widespread zoonoses, the hydatiddisease, is included in this group. The small flatworm uses herbivores as intermediate hosts and

This work was supported by the Swiss National ScienceFoundation (project nO 31-45575.95), PADCTICNPq(Proc. 620081195-3), EEC (DG XII CI 10284-0), the"Jubilaumsstiftung der Schweitzerischen Lebensversicherungs- und Rentenanstalt für Volksgesundheitund Medizinische Forschung" and the "Sandoz-Stiftungzur Fôrderung der medizinisch-biologischenWissenschaften".This paper reports on research conducted by Karen LuisaHaag as part of her PhD thesis on strain characterisation,genetic variability and breeding systems of Echinococeus. It is a result of a collaborative work between theCentro de Biotecnologia (Universidade Federal do RioGrande do Sul, Brazil) and the Institute ofParasitology(University ofBerne, Switzerland). Arnaldo Zaha worksprimarily with gene organisation and control in E.granu/osus, Aldo Mellender de Araujo works with evolutionary ecology on a variety oforganisms, but mainlyinsects, and Bruno Gottstein is dealing with molecularaspects of host-parasite interactions in E. mufti/ocu/aris.+Corresponding author. Fax: +55-51-319.2011. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

carnivores as final hosts. The adult is hermaphrodite and the larval stage (metacestode) is amplified by asexual reproduction.

Four species within the genus are recognised:E. vogeli and E. oligarthrus, which occur in theneotropical region, E. multi/ocularis, that has anholartic geographic range and E. granulosus, thatis world-wide distributed. Due to a low intermediate host specificity, E. granulosus has been subdivided in several strains, according to the host species used, or to the geographic range of the biological cycle. Sorne of the evolutionary questionsconcerning Echinococcus are: (1) is the adultmainly self- or cross-fertilising? (2) how do thestrains within a species differentiate? (3) what isthe true taxonomic status of these strains?

The first question relates to the second one:depending on the breeding system, only one oftwomodes of strain differentiation can occur. If individual parasites would be mainly selfers (Smyth& Smyth 1964), purifying (negative) selectionwouId quickly eliminate the non-adaptive mutations, due to increased homozygosis. In addition,selfing would lead to a high rate of linkage disequilibrium within parasite populations. In this situation, the genome would be selected as a whole,and not in pieces ofrecombining DNA. If, on theother hand, populations wouId undergo outcrossing (Rausch 1967, 1985), free recombination wouldallow genes to be selected as individual units, and

696 Nucleotide Sequence Variation in Echinococcus • KL Haag et al.

each genomic sequence would be able to respondsingularly to the positive and/or negative selectionimposed by the host.

Il has also been argued (Thompson et al. 1995,Lymbery & Thompson 1996) that the degree ofgenetic differentiation ofsorne strains is larger thanexpected for conspecific groups. Furthennore, ifEehilloeoeeus is an obligatory selfer, the biological species concept cannot be used to solve theproblem (Lymbery 1992, Lymbery & Thompson1996). In the present study we used the nucleotidesequencing oftwo coding and two non-coding regions of Eehinoeoeeus genome to try to elucidatesorne of the questions above. Ifparasite populations would have undergone outcrossing duringtheir evolutionary history, we would expect to findrecombination among sequences. Additionally, byassessing relative rates of substitution in codingand non-coding regions, it would be possible toevaluate the occurrence of positive and/or negative selection. Finally, genetic distances estimatedfrom those sequences could help to decide whetheror not some of the E. granulosus strains should beregarded as different species.


Moleeular analyses - Thirty three E.multi/oeularis isolates from different continents(Asia, Europe and North America), hosts (foxes,humans and rodents) and life cycle stages, as weilas 110 E. grallulosus metacestode isolates fromdifferent geographic regions (Australia, Europe andSouthern Brazil) and strains (bovine, equine, ovineand swine) and one E. vogeli isolate were used forgenomic DNA extraction and further analyses.DNA extraction was done by standard procedures(McManus & Simpson 1985).

For each isolate, four different targets wereamplified by PCR, using primers specific for Eehinoeoeeus DNA (see procedures in Haag et al.1997). Two of them were partial intron sequencesfrom an actin gene (ActII - 266 bp) and from anhomeobox containing gene (Hbx2 - 331 bp). Theother two were coding regions: a partial sequenceof a neotrophil chemotaxis inhibitor nuclear gene(AgB/l - 101 bp) and another partial sequence ofthe mitochondrial NADH dehydrogenase 1 gene(NDl - 141 bp).

The nucleotide variation within the PCR products obtained for the four targets was screened bythe PCR-SSCP method (see procedures in Haag etal. 1997). Subsequently, two isolates from eachSSCP pattern (except in the case ofE. vogeli) werechosen for direct fluorescence sequencing. For this,the single stranded DNA bands were cut out fromthe fresh silver-stained SSCP gels, washed andeluted. One ml ofthe eluted single strands was used

for re-amplification with the corresponding primers. These re-amplification products and their respective primers were used for sequencing.

Statistie andphylogenetie analyses - Sequenceswere aligned by eye (Fig. 1) and the molecular diversity parameters, recombination rates and relative rates ofsynonymous/non-synonymous substitutions (Ka/Ks) were estimated using DnaSP version 2.0 (Rozas & Rozas 1997). The recombination parameter (C) is calculated based on the average number ofnucleotide differences between pairsof sequences (Hudson 1987) and a minimum number of recombination events in the history of thesample (RM) is obtained using a four-gamete test(Hudson & Kaplan 1985).

The genetic distances as weil as the neighbourjoining (NJ) trees were estimated with MEGA version 1.0 (Kumar et al. 1993). The parsimony treeswere constructed using DNA Penny in Phylip version 3.5c (Felsenstein 1993). For the NJ phylogenetic analysis we used a gamma distance (Kimura2-parameter model) with gamma parameter a=1.In the parsimony analysis we made a branch-andbound search to find ail most parsimonious trees.Both kinds oftrees were constructed using E. vogelias outgroup.

RESULTS

The degree of al1ele polymorphism foundwithin E. muftiloeularis and within strains of E.granulosus was low, as shown in our previous studies (Haag et al. 1997, 1998). Indeed, only one transversion and a single base deletion in the Hbx2 intron occurred among isolates of E. multi/oeularis(Haag et al. 1997). Within the cattle, horse, pigand sheep strains of E. granulosus no al1ele polymorphism was found in the four coding and noncoding loci analysed in the present study.

For this reason, further analyses were doneconsidering the most common variant of E.multi/oeularis, the sequences of the four E.granulosus strains and those obtained for the E.vogeli isolate {GenBank assession numbers are:AF003748, AF003749, AF003750, AF024661 andAF024662 (Act II); X66818, AF003976,AF003977, AF024663 and AF024664 (Hbx 2);Z2648 1, Z26482, Z26483, Z26336 and AF024665(AgB/l); U65748 [NDI - authors did not provideinfonnation about variant sequences published byBowles and McManus (1993)]}. The moleculardiversity parameters estimated from this data setare shown in Table 1. The most variable locus wasthe mitochondrial ND1. Surprisingly, one of theintrons (Hbx2) was shown to be very conservedamong the referred strains and species, and theAgB/l nuclear coding region had as much variability as the Act II intron.


A 10 30 50

sheep

cattle

pig

horse

multiloc

vogeli

TCGTCCAAGACATCAGGTTAGTTGGATAGGTAGGCAGTGTTTCAGCCGCACCGGAACTGG

•.••••••••••••••••••..•...••••••••••..••••••• T ••••••••.• G •••

•........•.••••••••......•.•••••••..•..•••••• T •....•.••• G •••

....••••••••..........• A ••• G •..•.••• A ••••••..•••••••••.....•

70 90 110

sheep

cattle

pig

horse

multiloc

vogeli

TACCAACTAGTGGACCAATTTTCTCAAATAAGAGACAGAAATGGTTTGCTTTCATGCACT

........... C C A CA C .

........... C C A CA C .

.T C CTG .. T CA C .

. T C T CTG .. T A CA C .

.T C C T CA C .

130 150 170

sheep

cattle

pig

horse

multiloc

vogeli

AAATGTATGGTGAAGAAGTCGGCTTTTCATCTAACTAGATAGGCATGATTAGTGTGGAGA

•••••••••.•••.•.••••.•.....•.••••.••.•• G••••••••.•.•••••••••

•••••••......•••••..•• T ...•..••...•••••••••••...•.•.••••••..

•••.•.•••.....•.••.•.......•.••.. G...••••••.• A ••.•.•••.•..•.

••••••••.••••.••••...•••••••.•..••.••••••••••. A ••••••••..•••

••.• A •••.•.••.••..•••••••••.••.•••••••.•.•.••• A ••••....•••••

190 210 230

sheep

cattle

pig

horse

multi10c

vogeli

sheep

cattle

pig

horse

multi10c

vogeli

TCAAGTGCTCTCTTGTAGAGTCGCCATCTGAGGGCAGTCTTTCTATTTTCGCCCTGTGAC

....................... T .•. T ........................•.•.....

........................... T GG.C .

......••••••.•..•A ••••••••. T ••••••..•••••••....•••••..••••••

................................................. T .. GGG .....

250

AACGTACCTATTCCGAAATAATCTTT

•...•.•••••••••..•••••• A •.

.•.•..••••••••..•.••••• A •.

.•.••••••••••...••.•••• A ••

•••••••••••••••• G •••••••••

Fig. I-A: nucleotide sequence ahgnments of the Actll intron for the Echinococcus granulosus sheep, caule, pig and horse straÎns asweil as for E. mulliloclIlaris and E. vogeli.


B 10 30 50

sheep

cattle!pig

horse

rnultiloc

vogeli

CGTCTTAGAAGAGCGATTTGATCGACAAAAGTACCTCAGCAGTGCTGAACGCGCCGAGAT

...••••..••••••••.•••••..••.•...•••. A .•.. T ...•.....•••.•• C ••

70 90 110

sheep

cattle!pig

horse

rnultiloc

vogeli

GTCACGAGACCTGGGGCTCTCTGAAACCCAGGTATGTCACAGCCGATGTCATTAAACATG

•••••••......•.....•...•..••••••.........•.••. C ....•••...•..

.•..••••.•.•...•.••.••••••••••••••••••.•.••••• C ••.•••••.••••

•.......••..•...•.•••••••••••..••••••••••••••• C •••••••••••••

..... A.A C .

130 150 170

sheep

cattle!pig

horse

rnultiloc

vogeli

GGAAGGGGTGAGAGTAGTTGGAGCGTCACGAAGTGCCAAATTGGGCGCTTGTCAAGCTGC

•....••...•.•••• T ••.•••••••••••••••.•••.••.••.••••••••••••••

.................. CA .

190 210 230

sheep

cattle!pig

horse

rnultiloc

vogeli

GCCTTTATAACTGTTGAGTGCATCATCACCCATAAAAAATTGGGAGAGAGGGGGGCGGGA

..•.•••..••.•••.••••••.•...••••.•••.••.••.••• T ••••.••.•.••..

250 270 290

sheep

cattle!pig

horse

rnultiloc

vogeli

GCGGGTCAAAAGGGTCATCACGGCTCATGCATTAGTAAGATCGTAAAAGGCATGCCTCTA

.•••••.••.•.••..•••• T ••••.•••••••••••••••.•••...•••.••.•.••.

................... GGT A C .

310 330

sheep

cattle!pig

horse

rnultiloc

vogeli

ATTATGACCCCCACCACTAGGTGAAAATATG

••.•...•. T ••••••.••••••.......•

Fig. 1-8: nucieollde sequence alignments of the Hbx2 intron for the Eehinococeus granu/osus sheep, cattle, pig and horse strainsas weil as for E. mu/ti/oeu/aris and E. voge/i.


c 10 30 50

sheep

cattle

pig

horse

multiloc

vogeli

CTGGTTGGGGTGGTTACAACAATTATTCATTTTTAAGGTCGGTTCGATGTGCTTTTGGAT

.A A T .. T.G T 00 ••••••••• 0.0 •••••

0 •••• 0 •• o' .A T.G T 0 ••••• 0 ••••••••••••••

.G ••••.••.. A •••• T .. T .. o •• C •...•..• G...•• A •••••..• o ••••••••••

•.......••• A .•...•• T .• A Go ....••• T •• T ....•••.•....•... G.

....... A A T .. C C.. G AA G G.

70 90 110

sheep

cattle

pig

horse

multi10c

vogeli

sheep

cattle

pig

horse

mu1tiloc

vogeli

CTGTTAGGTTTGAGGCTTGTTTTATGTGTGTGGTGATTTTTTGTGCTTTGTGTAGTTGTA

.....................•....•....... T ..•............... T .

.................................. T C A.. CT G

............................ C .....•.................. T C.

............. A A T 0 •••• TAC .

..................................A C CT .

130

GGTATAATTTAATTGATTTTT

.......... GG .

•••••••••• G ••••••••••

•.•••••••• GG •••...•••

Fig. I-C: nucleotide sequence alignments of the mitochondrial NO 1 for the Echinococcus granu/osus sheep, catlle, pig and horsestrains as weil as for E. mu/tzlocu/aris and E. vogelt.

D10 30 50

sheep

cattle/pig

horse

mu1ti1oc

vogeli

AGTGGTTGACCTCTTAAAGGAACTGGAAGAAGTGTTCCAGTTGTTGAGGAAGAAGCTACG

•••••••••••••••.•••••••••••••••••••.•• G •••••••••••••••••••••

.T A A A .

.......... A.G .

70 90

sheep

cattle/pig

horse

mu1ti1oc

voge1i

CATGGCACTCAGGTCCCACCTCAGAGGGTTGATTGCTGAAGG

· .C T.A A G .

· .C ••••••.. A .•••..•.•••..• A •... G .••••••.••

• .C .•••••.. A •.•...••••••.• A G .•••••..••

· .C T.A A.AA G .

Fig. 1-0: nucleotide sequence alignments of the AgB/l for the Echinococcus granu/oS/ls sheep, cattle, plg and horse strains as weilas for E mufti/oeu/aris and E. voge/i.


TABLE 1

Nucleotide diversity (1t), theta (El), average number ofnucleotide differences (k), number ofpolymorphic sites (S)and total number of sites (T) of the four non-coding (Act Il and Hbx 2) and coding (AgB/I and NOl) sequences

analysed in this study

PARAMETERa Act II Hbx 2 AgBfl NOl

1t 0.0524 0.0204 0.0559 0.0964(O.OOOI)b (0.0001 ) (0.0001) (0.0002)

El 0.0576 0.0233 0.0618 0.0963(0.0008) (0.0002) (0.0011) (0.0023)

k 13.93 6.70 5.70 13.60S 35 16 13 31T 266 329 102 141SfT 0.1316 0.0486 0.1274 0.2198

a: Nei 1987; b: Numbers in parentheses are standard deviations.

Sheep Catt1e Pig Horse EM EV

TABLE II

Relative rates of non-synonymous and synonymous(KafKs) substitutions within NOl (above diagonal)

and AgB1 (below diagonal) coding sequences amongthe Echinococclis granliioslis strains, E. mliitiloclilaris

(EM) and E. volgeli (EV)

The recombination parameter (C=4Nc, wherec is the recombination rate) among the nuclear sequences was equal to 34.2 (per gene) and 0.0518(between adjacent sites). The minimum number ofrecombination events occurring in the history ofthat sample of sequences was estimated do beRm=2. Additionally, the relative rates of synonymous and non-synonymous substitutions calculatedfor the two coding regions showed that, comparedto the mitochondrial ND l, the rates of non-synonymous substitutions within AgB/I were veryhigh (Table II).

As the results of the NJ and parsimony analyses were very similar, we decided to concentrateon the later. A phylogeny obtained by analysingail loci together is shown in Fig. 2. The topologyof that tree is in accordance with others, obtainedusing a larger number of OTUs and other helminths as outgroups (Lymbery 1995). However, thephylogenies constructed for each sequence separately were not congruent. First, most sequencesdid not provide a single most parsimonious tree:the Hbx2 intron resulted in 15, NDI in 2 and AgB 1in three equally parsimonious topologies. Second,

Sheep 0.20 0.17 0.09Cattle 1.22 0.07 0.06Pig 1.22 0.00 0.05Horse '" 0.31 0.31EM 0.88 0.18 0.18 0.45EV 1.48 2.09 2.09 0.90

'" indeterminacy

0.13 0.100.08 0.070.08 0.090.07 0.04

0.140.61

r---- cattle

plg

sheep

L- horse

multilocularis

vogeli

Fig. 2: maximum parslmony phylogenetic lree of Eehinoeoeeus strains and species obtained using the four coding and noncodmg sequences. The tree requires 113 steps (for details, seeMaterials and Methods).

ambiguities were found regarding the position ofthe horse strain: in sorne instances it is groupedtogether with the E. granulosus strains, and in others it splits before.

A striking result obtained by the genetic distance calculations (Table III) was the high similarity between the caule and the pig strains. As expected, E. vogeli is the most distant group in relation to ail other analysed OTUs. The distance values among the other E. granulosus strains and between each strain and E. multi/oeularis were quitesimilar.

DISCUSSION

Previous studies (Lymbery et al. 1997) concluded that cross-fertilisation occurs within E.granulosus populations. However, there were alsogood evidences that outcrossing is not the predominant mating system, since most loci analysedshowed monomorphism within strains or largedeficiencies ofheterozygotes (Lymbery & Thompson 1988, Lymbery et al. 1990, 1997). The resultsobtained in the present study support those previ-


TABLE III

Jukes-Cantor genetic distances (above diagonal) and their standard deviations (below diagonal) among theEchinococclIs granllloslIs strains, E. mlllti/oclIlaris (EM) and E. volgeli (EV), based on the nuleotide sequences

of the four coding and non-coding loci

Sheep CaUle Pig Horse EM EV

Sheep 0.0329 0.0379 0.0392 0.0455 0.0700

CaUle 0.0064 0.0145 0.0317 0.0442 0.0674

Pig 0.0069 0.0042 0.0405 0.0493 0.0700

Horse 0.0070 0.0062 0.0071 0.0392 0.0622

EM 0.0075 0.0074 0.0079 0.0070 0.0635

EV 0.0095 0.0093 0.0095 0.0089 0.0090

ous findings, suggesting that recombination withinthe nuclear sequences occurred at least twice during the evolution of the genus. Although the coding and non-coding regions tested here were short,the lack of phylogenetic congruence among thetrees constructed for each locus separately couldalso be due to recombination.

Another explanation for those incongruencesis that selection acted independently on each sequence, but this argument could be used only forthe coding regions. Indeed, we showed that positive selection did act during the evolution of theAgBIl gene: most nucIeotide replacements foundby pairwise comparisons of the sequences werenon-synonymous, and the relative rates of non-silentlsilent substitutions (Ka/Ks) were greater thanone in six out of fifteen comparisons.

Selection was also used to explain the high frequency ofheterozygotes found for variant regulatory sequences in populations of E. granulosusfrom the sheep strain. Taken together, ail thosefindings indicate that Echinococcus is not an evolutionary dead-end, unable to adapt quickly enoughto changing environmental conditions. Nevertheless, it seems that a balance between cross and selffertilisation was the best solution found by the parasite to keep evolving. It seems that the recombination rates cannot be neither too high, breaking downcoadapted gene complexes, nor too low, hinderingadaptive changes.

Moreover, the estimated phylogenetic distancesand the trees of Echinococcus species and strainsare in agreement with those reported by Lymbery(1995). The results show that the phylogeneticposition of the E. granulosus horse strain is ambiguous. For this reason, we agree with the proposai of a taxonomic revision of the genus, basednot only on a molecular phylogenetic approachincluding a larger number of OTUs, but also onother comparative biological data.

REFERENCES

Bowles J, McManus DP 1993. NADH dehydrogenase 1gene sequences compared for species and strains ofthe genus EchinococclIs. 1nternlJ Parasitol23: 969972.

Felsenstein J 1993. Phylip (Phylogeny Inference Package) version 3.5c. Distributed by the author. Department ofGenetics, University ofWashington, Seattle.

Haag KL, Zaha A, Araujo AM, GoUstein B 1997. Reduced genetic variability in coding and non-codingregions of EchinococclIs mlliti/oclIlaris genome.Parasitology 115: 521-530.

Haag KL, Araujo AM, Gottstein B, Siles-Lucas M,Thompson RCA, Zaha A 1998. Breeding system inEchinococcus granuloslls (Cestoda; Taeniidae);selfing or outcrossing? Parasitology (in press).

Hudson RR 1987. Estimating the recombination parameter of a finite population model without selection.Gen Res 50: 245-250.

Hudson RR, Kaplan NL 1985. Statistical properties ofthe number ofrecombination events in the history ofa sample ofDNA sequences. Genetics 1JI: 147-164.

Kumar S, Tamura K, Nei M 1993. Mega: MolecularEvollltionary Genetics Analysis. version 1.0. ThePensylvania State University, University Park, PA16802.

Lymbery AJ 1992. Interbreeding, monophyly and thegenetic yardstick: species concepts in parasites.Parasitol Today 8: 208-211.

Lymbery Al 1995. Genetic diversity, genetic differentiation and speciation in the genus EchinococcllsRudolphi 1801, p. 51-88. In RCA Thompson, AlLymbery (eds), Echinococclls and Hydatid Disease,Cab International, Wallingford.

Lyrnbery Al, Thompson RCA 1988. Electrophoreticanalysis of genetic variation in Echinococcusgranuloslls from domestic hosts in Australia. 1nternJ Parasitol 18: 803-811.

Lymbery Al, Thompson RCA 1996. Species of Echinococcus: pattern and process. Parasitol Today 12:486-491.

Lyrnbery Al, Constantine CC, Thompson RCA 1997.Self-fertilization without genornic or population


structuring in a parasitic tapeworrn. Evo/ution 5/:289-294.

Lymbery Al, Thompson RCA, Hobbs RP 1990. Geneticdiversity and genetic differentiation in Echinococcus granu/osus (Batsch, 1786) from domestic andsy1vatic hosts on the mainland of Austra1ia. Parasit%gy 101: 283-289.

McManus OP, Simpson AlG 1985. Identification of theEchinococcus (hydatid disease) organisms usingcloned ONA markers. Mo/ Biochem Parasito/ 17:171-178.

Nei M 1987. Mo1ecu/ar Evo/utionary Genetics. Columbia University Press, New York, 512 pp.

Rausch RL 1967. A consideration of intraspecific categories in the genus Echinococcus Rudolphi.180 1

(Cestoda: Taeniidae). J Parasito/53: 484-491.Rausch RL 1985. Parasitology: retrospect and prospect.

J Parasito/ 71: 139-151.Rozas l. Rozas R 1997. OnaSP version 2.0: a novel

software package for extensive molecular population genetics analysis. Computation App/icated toBiosciences 13: 307-311.

Smyth JO, Smyth MM 1964. Natural and experimentalhosts of Echinococcus granu/osus and E.mu/ti/ocu/aris, with comments on the genetics ofspeciation in the genus Echinococcus. Parasit%gy54: 493-514.

Thompson RCA, Lymbery Al, Constantine CC 1995.Variation in Echinococcus: towards a taxonomie revision of the genus. Advanc Parasito/ 35: 146-176.

Mem Inst Oswaldo Cruz, Rio de Janeiro, Vol. 93(5): 703-704, Sep./Oct. 1998 703

RESEARCH NOTE

Surveillance UsingMolecular Tools: Examples

from Brazil

Ana Carolina Paulo Vicente/+,Hooman Momen*

Departamento de Genética *Departamento deBioquimica e Biologia Molecular, Instituto OswaldoCruz, Av. Brasi14365, 21045-900 Rio de Janeiro, RJ,

Brasil

Key words: surveillance - HIV2 - Vibrio - cholera Leishmania - Escherichia coli

Brazil presents particular problems for surveillance of infectious diseases. These include its continental size, uneven distribution ofresources, difficulty of communication and access in sorne ofthe more remote areas, as weIl as large areas covered by tropical rain-forests. Surveillance for infectious diseases in Brazil has traditionally beencarried out in a passive manner by government authorities or as individual initiatives. Most effort hasbeen directed in the collecting and tabulating ofdata on notifiable infections. A limited amount oflaboratory support has been available for the isolation and identification of the etiological agents.More recently molecular methods have been introduced in the analysis ofthese data.

Here we present sorne practical examples ofthe use of different molecular tools for diagnosisand in the analysis of infectious diseases and inepidemiological monitoring of outbreaks. The examples are taken from work carried out in our Institute.

HIV-2, the second AIDS-causing virus, wasoriginally identified and found to be quite commonin West Afiica. With a more restricted geographicspread than HIV-1, this virus has also been isolatedin countries with socioeconomic links to West Africa (R Marlink 1996 AfDS 10:689-699). Sorne earlyreports analyzing the seroprevalence of HIV-1 andHIV-2 were contradictory about the presence of

+Corresponding author. Fax: +55-21-260.4282. E-mail:[email protected] 15 June 1998Accepted 30 July 1998

HIV-2 in Brazil (L Oyafuso et al. 1989 New Engl JMed 320: 953-958, RM Hendry et al. J Acq lmmD~rSynd 4: 623-627). At that time they concludedthat there was sorne cross-reactivity between HIV2 and HIV-1 which resulted in misinterpretation.Using polymerase chain reaction (PCR) and specific internai probes to HIV-2, Pieniazek et al. (1991AIDS 5: 1293-1299), identified mixed HIV-I/HIV2 infections in Rio de Janeiro, Brazil. In order tovalidate the World Health Organization strategy forHIV testing, sera from 9,885 blood donors from SàoPaulo were screened by HIV enzyme-linkedimmunosorbent assays (ELISA) and Western blotand the results did not support the evidence ofHIV2 circulation in Brazil (MB Carvalho et al. 1996AfDS JO: 1135-1140).

We have applied molecular tools in surveillancefor the detection of HIV-2 in HIV-1 positivesamples (possible dual infections) as weil as insamples with undetermined Western blots. Morethan 200 samples from different parts of the country were screened for the presence of HIV-2 provirai DNA using nested PCR targeting the long terminai repeat (LTR), protease and gag regions. Inthree sampies only PCR products corresponding tothe LTR region were amplified. These products weresequenced and the nucleotide sequence was different from that of HIV-2 LTR. They matched withhuman genome sequence, probably a rare aIleIepresent in few people. At present we have failed todetect and confirm the circulation ofHIV-2 in Brazil. We have shown that the use of LTR diagnosticprimers to HIV-2 has to be carefully analyzed.

Vibrio choleraeoccurs naturally in aquatic systems where it may constitute part of the normalmicroflora of zooplankton and larger animais. V.cholerae is a heterogeneous species with more than140 serotypes, only a few ofwhich are associatedwith biotypes causing human cholera and epidemics. The ongoing cholera pandemic (7th) is causedby the El Tor biotype, serotype 01. In 1991 choiera re-emerged in Brazil after being absent for acentury, the previous pandemic involved V.cholerae classical biotype. The present situation isdifferent in that not only is there a new V. choleraebiotype, but also there is now detailed knowledgeabout the bacterial virulence factors determiningthis disease and the molecular tools available forcharacterization of the isolates. In 1993 a new V.cholerae strain was identified in the State ofAmazonas during surveillance using AP-PCR formolecular characterization ofcholera vibrios. TheV. cholerae amazonia variant is of the 01 serotype; it has distinct multilocus enzyme electrophoresis and AP-PCR profiles from other pathogenic 01 V. cholerae. About 50 isolates have beenmade from cases of diarrhea in the upper Amazon

704 Surveillance Using Molecular Toois • ACP Vicent, H Momen

(Solimôes) River. The microbe apparently doesnot harbor any of the weil known virulence associated genes (e.g. the toxin gene casette and themajor colonization factor, TCP); however sorneisolates present a cytotoxic etfect for Y-1 cells (ACoelho et al. 1995 J Clin Microbiol 33: 114-118).

Since 1997, in Amazonas, ail cholera notification is based on clinical and epidemiological diagnosis. In trying to identify cholera vibrio in apparent outbreaks ofcholera occuring in Sào Paulo deOlivença, Jurua and Envira villages, we appliedPCR - target specific to genes associated with V.cholerae El Tor (cholera toxin/CT and toxin coregulated pilus/TCP) and V. cholerae amazonia(regulatory gene/toxR). The results were negativebut using PCR - target specific to genes associatedwith Escherichia coli enterotoxigenic (termo-Iabel toxin / LT and tenno-stable toxin / ST) (NGTomieporth et al. 1995 J Clin Microbiol33: 13711374) we were able to identify and characterizethis bacteria and thus able to demonstrate that theseacute diarrhea outbreaks, clinically very close tocholera symptoms, were not associated with anyV. cholerae.

In Leishmania, a numerical zymotaxonomicstudy of New World Leishmania was carried out

(E Cupolillo et al. 1994 Am J Trop Med Hyg 50:296-311). The analysis involved the use of phenetic, cladistic and ordination techniques on enzyme electrophoresis data from more than 250 isolates ofLeishmania. This study together with laterwork has revealed a rich diversity among isolatesfrom the New World at both organismal and molecular levels. This diversity has provided numerous opportunities to probe questions concemingparasite evolution and biology, as weil as their rolein human disease. In many localities, more thanone Leishmania species co-exists with overlappinganimal hosts and vectors, as weil as other pathogens. In collaboration with a number of ditferentresearch groups we have studied aspects of theepidemiology ofleishrnaniasis in various countriesof Latin America and in different regions of Brazil, in addition we have been interested in determining the autochtonous origin of certain Leishmania species found in the New World (H Momenet al. 1993 Biol Res 26: 249-255).

In most of these examples molecular identification of the etiological agents was followed bygenetic analysis. The results were then forwardedto the relevant control agencies, usually the FNS(Fundaçào Nacional de Saude).

F

INSTRUCTIONS TO AUTHORS

The Memorias do Instituto Oswaldo Cruz is published by the Oswaldo Cruz Institute and welcomesoriginal contributions from research scientists throughout the world. The journal publishes original research in the fields of parasitology (protozoology, helminthology, entomology and malacology), microbiology (virology, bacteriology and mycology), tropical medicine (pathology, epidemiology and clinicalstudies) as well as basic studies in biochemistry, immunology, molecular and cell biology. physiology,and genetics related to these fields. Short communications in the form of Research Notes are also considered.

Manuscripts will be refereed critically by at least two reviewers; acceptance will be based on scientific content and presentation ofthe material.

Manuscript and figures should be sent in quadruplicate to the Editorial Office. The manuscriptshould be produced using a standard word processing sofware compabtible with DOS or Windows andshould be types (font size, 12) double-spaced, and should be arranged in the following order: runningtitle, title, authors' names, department or laboratory where the work was done name and address of theinstitution, summary, key words, introduction, materials and methods, results, discussion, acknowledgments (if any), and references. Sponsorships should be mentioned as a footnote on the first page.

Alternatively, manuscripts may be submitted following the "Uniform Requirements for Manuscriptssubmitted to Biomedical Journals" produced by the International Committee ofMedical Journal Editors,also known as the Vancouver Style (Annals ofInternal Medicine 1997; 126: 36-47).

For complete instructions, the authors should consult the home-page of the Memorias (http://www.pobox.com/r-rnemorias), or the first issue of each volume, or contact the Memorias Office byphone (55-021-598.4335), fax (55-21-280.5048), or E-mail ([email protected]) for further information.

ERRATA

VoI. 93(1) Jan.lFeb. 1998

p, 109- second paragraph- should read

R. Killick-Kendrick et al. (1994 Ann. Trap. Med. Parasitol88: 183-196) demonstrated the importanceof observation of the base of the spermathecal ducts for the identification of Kenyan Phlebotomus(Larroussius) spp. and described a technique, referred briefly by Leger et al. (1983 Ann Paras HumComp 58: 611-623), for the dissection of the female abdomen. With the insects in Berlese's fluid, theyseparated the terminal part, using entomological pins (size 00) attached to small wooden sticks, andcovered it with a coverslip.

MEMORIAS DO INSTITUTO OSWAlDO CRUZ

Foreword - Hooman Momen, Alia A l.al , Miche!77oo.l·Il!}1C" SoS

Trypa nosornatids

Implicationsof'a I eotropical Originofthe GenusLeish-mania - Harry o)'~ b57

Genetic Diversity in I latural Population of ewW, ridl.eishmania - Elisa Cupolillo, Hooman Momen. GabrielGrimaldiJr 6b .1

The Evolution of Trypanosornes Infecting Humanand Primates - Jamie te ens, Harry Noy s, lVe/ll~v

Gibson 66~

Genetic Oat Showing Evolutionary Links betweenLeishmania and EndOll") 1111/1U/Il - Elisa Cupolillo.LutzaOR Pereira. Octavio Femandes, Marcos P Catanho.Julio C Pereira. Enrique Medina-Aco: tu. GabrielGrimald! Jr 6 77

Allelic Diversity Ilt the Merozoite millet' Protein-l(M P-I) Locus m atural Plasmodium fa / tparumPopulations: Bricf O erview - Marcelo •Ferreira,Osam u Kaneko, MU.Hl/.Wgu Ktmura, Qing Li u,Fumihiko Kawamoto, Kazuyuki Tanabe D31

E aluation of DI 'A Rccombinant Methodologie orthe Diagnosis of Plasmodium falciparum nd the ir

ompari 'on with the lieroscopy Assay - L Urdaneta,P Guevara, JL Ramire: 039

Syst mal ic and Population Level Analy is ofAnopheles darling! - JE Conn 647

Anopheline peel ': Complexes in Brazil. CUlTCnlKnowledge of'Those Related to Malaria Transmission. Maria Goreti Rosa-Freitas, Ricanlo Lourenco-deOliveira, Car/os Jose de Carvalho-Pinto. CarmenFlores-Mendoza . Teresa Fernandes ilva- do-

uscimento &51

Others

A Stud of Cryptosporidium pllrl'lll1l GenotypesandPopulation tructure - G Widmer; L Ichack. F pano,S Tzipori 68 5

peciesand Strain- pecificTyping ofCry pto: poridiumParasites in Clinical and ' nvironmental SarnplLihua Xiao. lrshad Sulaiman, Ronald Foyer; Alta ALal 687

Human leneti Bi-allclicSequ (HGBA E),a Database of lntra-genic Polyrnorphisms - ChandraSarkar;FIOI'io R Orrigtio. UlfGyllensten. Anthony J Brook .(Re rearch lore) 693

election. Recombination and History in a ParasiticFkllworm (EchinocOL'C:IL~) Inferred from uclcolid e-qu nccs - KL Haag. AM ArUlijo. B GO/lSlcil/. I1ZaIUl .. 695

Surveillance U ing Molecular Tool : xampl . fromBrnzil - Alia Carolina Palllo Vi ell/c. }fo<1mwl Mom 11

(Rc earch 'ote) 703

615

609

Vira l Diseases

M lIe ular pidcmiology and Emergen c of Rift Valley Fever -.4. 'all, PMAZanotto. P Vialm. OKSene,AIBouloy .

Mol cular Epid miologyof Human Polyoma irus Jin the Biaka Pygmies and Bantu of Central Africa -vlvester C Chima, Caroline F Rvschkewitsch, Gerald

i Stoner : .

Vancomycin-resistant Enterococci in Intensive CareHospital euings - DarenJ Austin.•HareJ IH Bonten : 587

Mol cular Gen tic Analy is of Multi-drug Resistancein Indian Isolate of Mvcobacterium tuberculosis Noman Siddiqi. Md. S/,,;mim. NK Jail/•Ashok Rattan..-111101Amin, VM Katoch, KSharma, eyed E Hasnain 589

MolecularBasis of Ribotypc Variation in the eventhPandemic loneand its0 139 Variant of Vibrio cholerae- Ruiting UJIl, Peter R Reeves S9S

The Arnazonia Variant of Vibrio cholerae: lolecularIdentification and rudy of Virulence Gene - MASBaptista. JRC Andrade, ACP Vlcente. CA Salles. .-1Coelho 60 1

Jrd INTERN TIO . ·AI. ~1EETI . · • 0 . ' • IOU:El'1D MIOLO{.Y \NDEVOL nONARYGOJ. I. 'FE '10 \)J E,\ E - M E GID-J

Molecular Epidemiology of Dcn-ZVirus In Brazil MP Miago: tovich. RAIR Nog/leira. HG SchalZmayr.RS Landolli (Resea rch I ote) 625

Malaria

ntimnlarial Drug Resistance: Surveillanceand Molecular Methods for Illional Malllria ControlProgrammes - Umberto D'Alessandro 62 7

Plenary ecture

, olutionnry ontrol oflnfectious Disease: Pro pe Ifor Vcctorborne and Watcrborne Pathogens - Paul WEwaltl.. .lerem ' B Sussman. Matthew T Distler; CamilaLibel. WahitJPChammas, VictorJ Dirita, CarlosAndn:Salles, Ana 'arolinu I'h m/e, lngrid Heitmann, Felipc

'abello 567

Integrated ienetic Epidemiology of lnfcctiou: Dis-eases: The hngas Model - Michel Tlbayrenc S77

Molecular Epidemiologic Typing y terns of Bacte-rial Pathogens: urrent I. ues and Perpectivcs - MarcJ Struelens 5a1

Bacterial Dlseases

An In erna tonal Journal of Biological and Biomedical Research

Documents