Co-cladogenesis spanning three phyla: leafhoppers (Insecta: Hemiptera: Cicadellidae) and their dual bacterial symbionts

Molecular Ecology (2006)

15

, 4175–4191 doi: 10.1111/j.1365-294X.2006.03071.x

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

Blackwell Publishing Ltd

Co-cladogenesis spanning three phyla: leafhoppers (Insecta: Hemiptera: Cicadellidae) and their dual bacterial symbionts

DANIELA M. TAKIYA,

*

PHAT L . TRAN,

†

CHRISTOPHER H. DIETRICH

*

and NANCY A. MORAN

†

*

Center for Biodiversity, Illinois Natural History Survey, 1816 S. Oak Street, Champaign, IL 61820, USA. E-mail: [email protected],

†

Department of Ecology and Evolutionary Biology, University of Arizona, Biosciences West, Room 310, 1041 East Lowell Street, Tucson, AZ 85721-0088, USA

Abstract

Endosymbioses are a major form of biological complexity affecting the ecological and evo-lutionary diversification of many eukaryotic groups. These associations are exemplified bynutritional symbioses of insects for which phylogenetic studies have demonstrated numerouscases of long-term codiversification between a bacterial and a host lineage. Some insects,including most leafhoppers (Insecta: Hemiptera: Cicadellidae), have more than one bacterialsymbiont within specialized host cells, raising questions regarding the patterns of codiver-sification of these multiple partners and the evolutionary persistence of complex symbioticsystems. Previous studies reported the presence of two dominant symbiont types in amember of the leafhopper subfamily Cicadellinae (sharpshooters). In this study, 16S rRNAsequences were obtained and used to examine the occurrence and evolutionary relation-ships of the two dominant symbiont types across 29 leafhopper species. Candidatus

Sulciamuelleri

(

Bacteroidetes

) was detected in all leafhopper species examined, a finding that isconsistent with a previous report of its ancient association with the Auchenorrhyncha (agrouping that includes leafhoppers, treehoppers, cicadas, planthoppers, and spittlebugs).

Baumannia cicadellinicola

(

Proteobacteria

), previously known from only five sharpshooterspecies, was found only in the sharpshooter tribes Cicadellini and Proconiini, as well as inthe subfamily Phereurhininae. Mitochondrial and nuclear gene sequences were obtainedand used to reconstruct host phylogenies. Analyses of host and symbiont data sets supporta congruent evolutionary history between sharpshooters,

Sulcia

and

Baumannia

and thusprovide the first strong evidence for long-term co-inheritance of multiple symbionts duringthe diversification of a eukaryotic host.

Sulcia

shows a fivefold lower rate of 16S rDNAsequence divergence than does

Baumannia

for the same host pairs. The term ‘coprimary’symbiont is proposed for such cases.

Keywords

:

Baumannia

, Cospeciation, mutualism, phylogeny,

Sulcia

Received 8 April 2006; revision accepted 19 June 2006

Introduction

One of the most prominent aspects of biological complexityis the occurrence of intimate symbiosis, linking the ecologyand evolution of phylogenetically distant lineages. Recentstudies have shown that such symbioses can persist forperiods of hundreds of millions of years and that they canhave major impact on the evolution of interacting partners.

Major instances of such symbioses occur in insects, manyof which rely on mutualistic associations with bacteriato supplement amino acid or vitamin deficiencies inspecialized, nutritionally unbalanced diets (Moran 1998).Such symbionts typically live within host cells calledbacteriocytes that form aggregates called bacteriomes andare transmitted by infection of eggs within the mother(Buchner 1965; Nault & Rodriguez 1985). So far, molecularstudies of insect symbioses have been focused on cases inwhich only a single bacterial lineage is universally presentin any given insect group, restricted to bacteriocytes, and

Correspondence: Daniela M. Takiya, Fax: 217 2440729; E-mail:[email protected]

4176

D . M . T A K I Y A

E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

cospeciating with their hosts. Obligate symbionts, livingwithin bacteriomes with a universal distribution in a cladeof hosts, are referred to as primary symbionts (Baumann2005). Both genomic and experimental studies indicate thatprimary symbionts studied to date have a major role inprovisioning nutrients to their hosts (Shigenobu

et al

. 2000;Akman

et al.

2002; Baumann 2005). This nutritional role hasbeen elucidated in the primary symbiont of aphids,

Buchneraaphidicola,

which provides its hosts with essential aminoacids (Douglas 1998; Baumann

et al

. 1999; Moran

et al

. 2003a;Baumann 2005). Although primary symbionts providebenefits to hosts, they also present constraints on host evolu-tion, especially because they typically undergo degenerativeevolution involving irreversible loss of genes and regulatorycapacities (Tamas

et al

. 2002; Moran

et al

. 2003a).Multiple symbiont types often co-occur in one insect

host (summarized in Buchner 1965; Fukatsu & Nikoh 2000;Baumann 2005). In some cases, insects with a primary sym-biont harbour additional symbionts with less restricteddistributions among host tissues; these are called second-ary or guest symbionts, and typically are not required forhost development or reproduction (Buchner 1965; Moran1998; Baumann 2005). Secondary symbionts tend to bephylogenetically diverse associates with relatively shorthistories in host lineages. Thus, closely related host speciesoften differ in the presence of particular secondary symbi-onts, which can vary in presence among members of thesame host species. A large number of phylogenetic studiesindicate that secondary symbionts do not show phyloge-netic congruence with hosts over long periods (e.g. Thao

et al

.2000 for psyllid hosts; Thao & Baumann 2004b for whiteflyhosts; review in Baumann 2005), although they may persistwithin a closely related cluster of species such as membersof a genus of mealybugs (Pseudococcidae) (Thao

et al

.2002) or a subgenus of aphids (Sandström

et al

. 2001). Ingeneral, primary symbionts are also distinguished by theirunusually large cell size and shape, whereas secondarysymbionts have more typical cell size and shape (overviewin Baumann 2005). Finally, recent studies have revealedmajor genomic differences between the two kinds of sym-bionts, with primary ones having near-minimal genomesthat do not take up new genes (Tamas

et al

. 2002; Degnan

et al

. 2005) and secondary ones featuring larger genomesthat are dynamic and capable of gene acquisition andrecombination (Moran

et al

. 2005b; Toh

et al

. 2006).In numerous insects, more than one symbiont type

appears to be an obligate associate, restricted to bacteriomesand universally present within a species. Among the mostprominent examples are the complex assemblages ofsymbiotic microbes that occur in most members of thehemipteran suborder Auchenorrhyncha, a diverse groupthat comprises approximately 45 000 species of leafhoppers,treehoppers, cicadas, planthoppers, and spittlebugs(Buchner 1965). The most evident ecological explanation

for the high frequency of symbionts in Auchenorrhyncha isthat they feed on plant fluids, which are notoriously unbal-anced nutritionally with low nitrogen and essential aminoacid content (Sandström & Moran 1999). Leafhoppers(Cicadellidae) include more than 20 000 described species(Dietrich

et al

. 2001) and some of the most important vectorsof plant diseases (Purcell & Hopkins 1996; Purcell

et al

. 1999;Hendson

et al

. 2001; Redak

et al

. 2004). While members ofmost leafhopper lineages feed preferentially on phloemsap, those from the subfamily Cicadellinae (commonlycalled sharpshooters) specialize on the very dilute sap ofxylem. Xylem sap is among the most nutritionally limiteddiets used by any animals, with a severe scarcity of organiccarbon and nitrogen concentrations that are 10 times lessthan those of phloem sap (Andersen

et al

. 1989; Redak

et al

.2004).

According to earlier microscopy-based studies of symbio-sis, performed before the availability of molecular methods,auchenorrhynchan species may harbour up to six morpho-logically distinct symbiont types, but the most commoncondition involves the presence of two obligate bacteriome-associated symbionts and from zero to two secondarysymbionts. Sharpshooters exhibit the latter condition (Buchner1965; Chang & Musgrave 1972; Kaiser 1980). Sharpshooterbacteriomes are paired structures located near the anteriorend of the abdomen (Buchner 1965; Kaiser 1980; Moran

et al

.2003b). As is the case for all obligately symbiotic bacteria, theidentities and evolutionary relationships of organisms inthese bacteriomes could not be ascertained before the adventof PCR and DNA sequencing techniques. Recently, Moran

et al

. (2003b, 2005) linked these two morphologically dis-tinct organisms to respective 16S rDNA sequences, en-abling their placement on the bacterial phylogenetic treeand correlating sequences with cells observed in light andelectron microscopy. Candidatus

Baumannia cicadellinicola

(hereafter called ‘

Baumannia

’), an irregularly spherical bac-terium approximately 1.5–2.0

µ

m in diameter, occurs in thebacteriomes of five sharpshooter species studied previously(Moran

et al

. 2003b). Although these represented a verylimited sample of species, phylogenetic analyses indicatedthat they formed a strongly supported clade within thegamma-3 Proteobacteria and that their relationships wereconsistent with those of the insect hosts. However, thisevidence for codiversification is inconclusive because sofew species were included, representing only two genera.More intensive studies of

Homalodisca vitripennis

(formerly

H. coagulata

and also called the glassy-winged sharpshooter)revealed that a second organism, named Candidatus

Sulciamuelleri

(hereafter called ‘

Sulcia

’) in the phylum

Bacteroidetes

,is present in all sampled populations (Moran

et al

. 2003b;Moran

et al

. 2005a).

Sulcia

is widespread in Auchenorrhynchaand is documented in hosts from Fulgoroidea, Cicadoidea,and Cercopoidea, in addition to Membracoidea (which in-cludes leafhoppers). Furthermore, phylogenetic relationships

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S

4177

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

among

Sulcia

strains from diverse auchenorrhynchan hostsare congruent with current estimates of host phylogenies,suggesting that this symbiont descended from an ancestorthat infected an ancient ancestor of Auchenorrhyncha,over 260 million years ago (Moran

et al

. 2005a).

Sulcia

and

Baumannia

are the dominant microbes resid-ing in the bacteriomes, based on an intensive study on

H. vitripennis

; large clone libraries were obtained and fewbacterial sequences were obtained that could not bedefinitively assigned to these organisms (Wu

et al

. 2006).Furthermore, these two organisms possess complementarysets of pathways for the biosynthesis of nutrients neededby their insect hosts, with

Baumannia

able to provision alarge set of cofactors (including most B-vitamins) and

Sulcia

encoding genes for the biosynthesis of essential aminoacids (Wu

et al

. 2006). This result along with fluorescent

insitu

hybridization studies of

Baumannia

and

Sulcia

(Moran

et al

. 2005a; Wu

et al

. 2006) verified earlier conclusions thattwo bacteriome-associated symbionts are consistently presentin sharpshooters (Buchner 1965; Chang & Musgrave 1972;Kaiser 1980). Although numerous symbioses involving aninsect host group and a primary symbiont have been ana-lysed, including symbioses of aphids (Moran

et al

. 1993),psyllids (Thao

et al

. 2001), mealybugs (Thao

et al

. 2002),whiteflies (Thao & Baumann 2004a), cockroaches (Lo

et al

. 2003), carpenter ants (Sauer

et al

. 2000), and tsetse flies(Aksoy

et al

. 1997), the sharpshooters are the first insect cladefor which two apparently widespread bacteriome-restrictedsymbionts have been characterized using molecular data.

In addition to

Baumannia

and

Sulcia

,

Wolbachia pipientis

was the most common organism represented in the clonelibraries of

H. vitripennis

, although much less abundantthan the other two bacteria and sometimes absent fromsome host individuals (Moran

et al

. 2003b; Wu

et al

. 2006).

Wolbachia

are found in all major insect orders (Werren

et al

.1995a) and, similarly to

Baumannia

and

Sulcia

, are verticallytransmitted through the egg cytoplasm. On the other hand,

Wolbachia

undergo extensive horizontal transmission betweeninsect taxa (Werren

et al

. 1995b), leading to completelyincongruent histories with their hosts independent of geneor taxonomic-level studied (Moran & Baumann 1994; VanMeer

et al

. 1999; Shoemaker

et al

. 2002; Kikuchi & Fukatsu2003).

In this paper, we address the question of the distributionof

Baumannia

,

Sulcia

, and

Wolbachia

among sharpshootersand related leafhoppers. Our central question is the extentto which these specialized associations are stable, long-termassociations that have persisted throughout the diversifi-cation of the group and that therefore might impose majorconstraints on host evolution. Thus, we focus on the evalu-ation of whether the two dominant bacteria have under-gone long-term codiversification with their hosts. Prior toour study, sequences of sharpshooter symbionts were fewand insufficient for rigorous testing of the extent of codi-

versification of these two symbionts with their hosts. Here,24 additional leafhopper species, spanning six tribes withemphasis on sharpshooters, were characterized for both

Baumannia

and

Sulcia

. Phylogenetic analyses were performedbased on symbiont rDNA gene sequences and on fournuclear and mitochondrial insect gene sequences to gener-ate independent hypotheses regarding the evolutionaryhistories of insects and symbionts. These data sets andphylogenies were then compared to assess the extent ofsupport for a history of codiversification between hostspecies and their two microbial associates.

Materials and methods

Taxon sampling

Included in the analyses were 29 species belonging to fourleafhopper subfamilies, with emphasis on the Cicadellinaeand related leafhopper lineages (Dietrich 1999; Dietrich

et al

. 2001). Table 1 lists the species names and collectiondata. Leafhopper higher-level classification follows Young(1968, 1977), Oman

et al

. (1990), Hamilton & Zack (1999), andDietrich (2004). Specific names used herein for the glassy-winged sharpshooter and the smoketree sharpshooter(

Homalodisca liturata

) follow Takiya

et al

. (2006) and Burks& Redak (2003), respectively. Samples from Moran

et al

.(2003b) were used for

Homalodisca vitripennis

(=

H. coagulata

),

H. liturata

(=

H. lacerta

),

Graphocephala hieroglyphica

,

Graphocephala aurora

, and

Graphocephala cythura

. Sequencesof host and symbiont genes were obtained for the sameleafhopper species with the exception of the three previouslystudied species of

Graphocephala

, which were not analysedfor host genes. In most cases, host and symbiont sequenceswere obtained from specimens collected on the same dayand locality (Table 1), including from the same insectspecimen in eight of the species studied.

Symbiont 16S rDNA sequences were newly determinedin this study, except for five previously deposited sequencesfor

Baumannia

(AF465793–7) and the single

Sulcia

(AY147399).Sequences of 16S rDNA for symbiont outgroups wereobtained from GenBank. Outgroups were chosen to re-present several of the closest

blast

hits for each of the sym-biont groups, based on

blastn

searches of GenBank. Theseincluded both free-living bacteria and some other insectendosymbionts, belonging to the same bacterial division.Insect symbionts included were those associated with aphids,mealybugs, psyllids, and ants for the phylogenetic analysesof

Baumannia

which falls in the gamma-Proteobacteria(NC_002528, AF476100, AF476106, AF263556, AJ250715,AF476102, AF263560, AF366378, and AE005632), and thosewith ladybird beetles, termites, and cockroaches for theanalyses of

Sulcia

which falls in the

Bacteroidetes

(M58789,M93152, AB071953, Z35665, Z35666, AF363713, Y13889,and AJ009687).

4178

D. M

. TA

KIY

A

ET

AL

.

© 2006 T

he Authors

Journal compilation ©

2006 Blackwell Publishing Ltd

Table 1

Material examined with accession numbers for insect and endosymbionts Candidatus

Baumannia cicadellinicola

and Candidatus

Sulcia muelleri

genes. Symbiont DNA wasextracted either from specimens from the same collection event as the specimen from which host DNA was extracted (‘same collection’), which may include the same specimen, or fromspecimens from a different locality as indicated. Numbers in parentheses following symbiont locality are the number of host individuals that were tested positive for the presence of

Wolbachia

using

wsp

and

ftsZ

diagnostic primers/total number of individuals tested. The total number of individuals tested is the same as the number of individuals from whichcomparable sequences of 16S for

Baumannia

and

Sulcia

were recovered. Dash (—) indicates that no

Wolbachia

test was performed

Taxon

Locality GenBank Acession

Host Symbiont Host: COI, COII, 16S, H3

Baumannia

: 16S Sulcia: 16S

Coelidiinae: TeruliiniJikradia olitoria (Say) USA: Illinois same collection (0/1) —, —, AY869828, AY869758 — AY676913

Evacanthinae: PagaroniiniPagaronia tredecimpunctata Ball USA: California same collection (0/2) AY869733, AY869785, AY869827, AY869755 — AY676911

Phereurhininae: PhereurhininiClydacha catapulta Kramer Peru: Huánuco same collection (2§/3) AY869743, AY869795, AY869804, AY869769 AY676878 AY676898

Cicadellinae: BathysmatophoriniHylaius oregonensis (Baker) USA: Oregon same collection (0/2) AY869729, AY869784, AY869825, AY869753 — AY676905

Cicadellinae: CicadelliniCicadella viridis (Linnaeus) Kyrgyzstan: Dzhalal-Abad Germany: Brandenburg (0/2) AY869735, AY869786, AY869826, AY869760 — AY676915Graphocephala aurora (Baker) — USA: Arizona (–/1) — AF465797* †Graphocephala coccinea (Forster) USA: Illinois USA: Illinois (1§/4) AY869730, AY869789, AY869807, AY869763 AY676891 AY676916Graphocephala cythura (Baker) — USA: Arizona (0/1) — AF465795* AY676919Graphocephala hieroglyphica (Say) — USA: Arizona (–/1) — AF465796* †Helochara communis Fitch USA: Arizona same collection (0/10) —, AY869783, A2Y869819, AY869752 AY676877 AY676897Pamplona spatulata Young Peru: Pasco same collection (2§/2) AY869744, AY869779, AY869821, AY869770 AY676887 AY676908Paromenia isabellina (Fowler) Costa Rica: San José same collection (0/3‡) AY869734, AY869782, AY869822, AY869759 AY676896 AY676914

Cicadellinae: ProconiiniAcrogonia virescens (Metcalf) Peru: Junín same collection (2§/5) AY869746, AY869797, AY869809, AY869772 AY676883 AY676903Cuerna costalis (Fabricius) USA: Florida USA: Illinois (0/7) AY869739, AY869791, AY869808, AY869765 AY676895 AY676918Cuerna gladiola Oman & Beamer USA: California same collection (2§/2) AY869741, AY869793, AY869818, AY869767 AY676892 AY676917Cuerna sayi (Nielson) USA: Maryland USA: Wisconsin (0/5) AY869751, AY869802, AY869823, AY869777 AY676894 AY676921Cuerna striata (Walker) USA: Michigan USA: Wisconsin (0/5) AY869750, AY869801, AY869824, AY869776 AY676893 AY676922Cyrtodisca major (Signoret) Mexico: Jalisco Mexico: Jalisco (0/3) AY869749, AY869800, AY869810, AY869775 AY676886 AY676907Diestostemma excisum Schmidt Peru: Pasco same collection (0/3) —, AY869778, AY869817, AY869756 AY676889 AY676910Diestostemma stesilea Distant Peru: Pasco same collection (0/1) AY869732, AY869780, AY869813, AY869754 AY676884 AY676904Homalodisca vitripennis (Germar) USA: Florida USA: California (1§/1) AY869740, AY869792, AY869803, AY869766 AF465793* AY147399*Homalodisca elongata Ball USA: Arizona USA: Arizona (2§/2) AY869747, AY869798, AY869806, AY869773 AY676881 AY676901Homalodisca liturata Ball USA: California USA: Arizona (0/1) AY869738, AY869790, AY869820, AY869764 AF465794* AY676920Homoscarta irregularis (Signoret) Peru: Pasco Peru: Huánuco (0/2) AY869745, AY869796, AY869814, AY869771 AY676882 AY676902Oncometopia orbona (Fabricius) USA: Illinois USA: Illinois (0/2) AY869736, AY869787, AY869811, AY869761 AY676879 AY676899Paraulacizes irrorata (Fabricius) USA: Illinois USA: Illinois (2§/3) AY869737, AY869788, AY869815, AY869762 AY676880 AY676900Phera obtusifrons Fowler Mexico: Puebla same collection (0/1) AY869748, AY869799, AY869805, AY869774 AY676888 AY676909Proconosama alalia (Distant) Peru: Pasco same collection (0/1) AY869742, AY869794, AY869812, AY869768 AY676885 AY676906Proconosama columbica (Signoret) Peru: Junín same collection (3§/4) AY869731, AY869781, AY869816, AY869757 AY676890 AY676912

*Sequences from Moran et al. (2003); all other sequences were obtained in current study. †Lack of DNA for diagnostic PCR. ‡Baumannia and Sulcia sequences recovered from only a single specimen. §Accession numbers for Wolbachia wsp are DQ450148–DQ450164.

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4179

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

DNA preparation

Insect specimens were collected in the field directly into95–100% ethanol and stored at −20 °C until processed. Foramplification of host genes, genomic DNA was extractedfrom a single hind leg and associated muscles using amodified ethanol precipitation/resuspension protocol(Bender et al. 1983) or the DNeasy tissue kit (QIAGEN). Allleafhopper vouchers were dried, pinned, and deposited atthe Illinois Natural History Survey (Champaign, Illinois).

For amplification of bacterial genes, bacteriomes wereisolated in the laboratory by immersing the freshly col-lected etherized insect in 0.85% saline solution under a dis-secting microscope, slitting the cuticle and teasing out thestructure using insect pins. Bacteriomes were separatelyplaced into 95% ethanol and subjected to DNA extractionusing the DNeasy tissue kit (QIAGEN). An individualbacteriome was placed in a 1.5 mL microfuge tube, frozenby immersion in liquid nitrogen and crushed by grindingwith a disposable pestle. Following addition of 180 µL ofthe tissue homogenization buffer from the DNeasy kit, thetissue was homogenized further by grinding. Then 20 µLof Proteinase K were added, followed by vortexing andincubation at 65 °C for 30 min. To remove RNA, 4 µL ofRNase A (100 mg/mL) were added, followed by 200 µL ofthe A1 buffer from the DNeasy kit, to enhance precipitation.

This mixture was incubated at 70 °C for 10 min, followedby purification steps as specified in the kit. Each extractionwas performed on one bacteriome from a single individual,with 1 to 10 individuals extracted for each species (seeTable 1).

PCR and sequencing of host genes

Modified primers based on those published by Simon et al.(1994) were used to amplify parts of the host mitochondrialgenes cytochrome oxidase subunit I (COI), cytochromeoxidase subunit II (COII), and 16S rDNA (see Table 2).Nuclear histone H3 sequences were amplified using theprimers HexAF and HexAR (Ogden & Whiting 2003; seeTable 2). Host templates and controls were amplified withTaq DNA polymerase (Promega) added at 80 °C and afterinitial denaturation at 94 °C for 3 min, followed by 30cycles of 94 °C for 1 min, 50 °C for 1 min, 72 °C for 2 min;and a final extension of 72 °C for 7 min. Double-strandedpolymerase chain reaction (PCR) amplification productswere checked for yield and specificity on 1% agaroseelectrophoresis gels stained with ethidium bromide underUV light. Amplicons were purified using QIAquick PCRpurification kit (QIAGEN) and both strands sequencedusing ABI PRISM BigDye terminator kit version 3 (PEApplied Biosystems). Sequencing products were run on an

Table 2 Oligonucleotide primer sequences used in polymerase chain and sequencing reactions for the following loci: COI and COII,cytochrome oxidase I and II; H3, histone H3; 16S, large subunit of ribosomal RNA, wsp, and ftsZ. Primers indicated with asterisk (*) wereonly used in sequencing reactions

Primer Organism Locus Sequence 5′→3′

C1-J-2195 Leafhopper COI TTGAT TTTTT GGTCA YCCWG AAGTTL2-N-3014 Leafhopper COI TTCAT TGCAC TAATC TGCCA TACTATL2-J-3037 Leafhopper COII TAGTA TGGCA GATTA GTGCA ATGAAC2-N-3661 Leafhopper COII CCRCA AATTT CWGAR CATTG ACCAHexAF Leafhopper H3 ATGGC TCGTA CCAAG CAGAC GGCHexAR Leafhopper H3 ATATC CTTGG GCATG ATGGT GACLR-J-12887 Leafhopper 16S CCGGT YTGAA CTCAR ATCALR-N-13398 Leafhopper 16S CRMCT GTTTA WCAAA AACAT10F Baumannia 16S AGTTT GATCA TGGCT CAGAT TG35R Baumannia 16S CCTTC ATCGC CTCTG ACTGC320R* Baumannia 16S ACCAG CTAGA GATCG TTGC650R* Baumannia 16S CACCG GTACA TATGA AATTC T1128R* Baumannia 16S GGGAC TTAAC CCAAC TTTCA C1507R Baumannia 16S TACCT TGTTA CGACT TCACC CCAG10FF Sulcia 16S AGTTT GATCA TGGCT CAGGA TAA270F* Sulcia 16S TTAGT TGGTA AGGTA ATGGC700R* Sulcia 16S ACATT CCAGC TACTC CAAAC T1370R Sulcia 16S CGTAT TCACC GGATC ATGGC559F* Baumannia and Sulcia 16S CGTGC CAGCA GCCGC GGTAA TACWspF Wolbachia wsp TGGTC CAATA AGTGA TGAAG AAACT AGCTAWspR Wolbachia wsp AAAAA TTAAA CGCTA CTCCA GCTTC TGCACFtsZF1 Wolbachia ftsZ GTTGT CGCAA ATACC GATGCFtsZR1 Wolbachia ftsZ CTTAA GTAAG CTGGT ATATC

4180 D . M . T A K I Y A E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

ABI 3730 capillary sequencer or purified using Sephadexcolumns and run on an ABI 377 lane automated sequencerat the Biotechnology Center of the University of Illinois atUrbana–Champaign.

PCR and sequencing of symbiont genes

For all species except those of the sharpshooter genusCuerna, PCR for Baumannia was conducted using theeubacterial primers 10F and 1507R, which together amplifymost of the 16S rDNA sequence (∼1500 bases) for mostProteobacteria. For Cuerna species, PCR was conductedusing 10F and primer 35R, which together amplify most ofthe 16S, a small part of the 23S rDNA, plus the intergenicspacer. These reactions consistently yielded a single product,corresponding to the Baumannia gene. Products were purified(PCR purification kit, QIAGEN) and submitted for DNAsequencing using the PCR primers, as well as 320R, 559F,650R, and 1128R (Table 2). All intergenic spacer sequencewas removed prior to analyses, using Escherichia coli as areference. As previously noted (Moran et al. 2003b), the 10Fprimer does not amplify Sulcia 16S rDNA sequences, dueto mispairing at the 3′ end. To obtain a portion of the Sulciasequence, primer 10FF was used instead, which binds inthe same position as 10F, but has a change in the 3′ endcompatible with sequences available in GenBank for mostmembers of the Bacteroidetes phylum. Approximately 1350 bpof 16S rDNA of Sulcia was amplified using primer 10FFplus 1370R, binding near the 3′ end of the 16S rRNA.Amplifications and sequencing protocols for rRNA genesof both Sulcia and Baumannia followed Moran et al. (2003b).Sequences were obtained for both Sulcia and Baumanniawith 1 to 10 individuals of each host species. Becausemultiple sequences from the same host species revealedalmost no sequence differences, we obtained polishedsequences for only one individual per species, with at leasttwo reads in each direction. Other individuals weresequenced using only the sequencing primers, giving justone read over most of the sequence.

Wolbachia screenings were performed for 27 of the 29species included in our analysis (DNA samples were notsufficient for screening two of the species), using the PCRconditions listed above. The same individuals screened forSulcia and Baumannia were also screened for presence ofWolbachia using Wolbachia-specific primers for the geneswsp (Zhou et al. 1998) and ftsZ (Werren et al. 1995b). Reac-tions were performed using Taq DNA Polymerase (Promega).A ‘touchdown’ PCR cycle was used, with denaturationstep (94 °C for 2 min); followed by 10 cycles of 94 °C for45 s, 65 °C for 45 s, 72 °C for 1 min; and 28 cycles of 94 °Cfor 45 s, 55 °C for 45 s, 72 °C for 1 min, and a final extensionof 72 °C for 5 min. Diagnostic PCRs consistently yielded asingle product at ∼0.6 kb and ∼1 kb for Wolbachia wsp andWolbachia ftsZ genes, respectively. All wsp sequences and

some ftsZ products were sequenced, using the PCR primersas sequencing primers (Table 1).

Sequencing products were run on an ABI 377 sequencerat the University of Arizona Genomic Analysis and Tech-nology Center.

Alignments

GenBank accession numbers for the sequences generatedare listed in Table 1. Correction of chromatogram sequences,reconciliation of complementary strands, and alignment ofprotein coding host genes across species were facilitatedby sequencher 4.1.2 (Gene Codes Corp.). Alignmentsfor the host ribosomal 16S rRNA sequences were madeusing clustal x 1.81 (Thompson et al. 1997) with gapopening:extension costs 50 : 1 and International Union ofBiochemistry (IUB) DNA weight matrix. Alignments of thehost genes resulted in 783 bp of COI, 591 bp of COII, 328 bpof H3, and 481 bp of 16S rDNA (total of 2183 bp, 804parsimony-informative). Including outgroup taxa, alignmentsfor the 16S rDNA of Baumannia was 1495 bp (352 parsimony-informative) and of Sulcia was 1489 bp (372 parsimony-informative) in length.

Although sequences for 16S rDNA and histone H3 forthe coelidiine Jikradia olitoria were sequenced and depositedin GenBank, this species was not included in the hostphylogenetic analyses. Preliminary results for this specieswere highly inconsistent with the currently accepted clas-sification and previous phylogenetic analyses (Oman et al.1990; Dietrich 1999; Dietrich et al. 2001), probably due tothe unavailability of COI and COII sequences for thisspecies, which accounted for 63% of character data in thecombined host data set.

Phylogenetic analyses

All analyses were run using paup* 4.0b10 (Swofford 1998)unless otherwise stated. Combination of molecular datasets for the host analysis was supported by the lack ofincongruence among matrices as suggested by theincongruence length difference test (ILD; Mickevich &Farris 1981; Farris et al. 1995) run with 1000 replicates andheuristic tree searches of 10 random addition replicatesand tree-bisection–reconnection (TBR) branch-swapping.Constant and uninformative characters were excluded priorto the test (Cunningham 1997). Pairs of partitions testedwere COI vs. COII (P = 0.15); mitochondrial proteinencoding (COI and COII) vs. ribosomal (16S) (P = 0.27);and mitochondrial (COI, COII, and 16S) vs. nuclear (H3)(P = 0.60). All gaps were treated as missing data.

Models of molecular evolution for use in maximum-likelihood analyses were estimated by a hierarchical likeli-hood-ratio test (Felsenstein 1981; Huelsenbeck & Crandall1997) using modeltest 3.06 (Posada & Crandall 1998).

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4181

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

Models chosen were TVM + G + I for the combined hostdata set, GTR + G + I for Baumannia, and TrN + G + I forSulcia. All models assume unequal base frequencies, un-equal rates across sites modelled by a gamma distribution(Yang 1996), and some estimated proportion of invariantsites. DNA substitution matrices assumed by each modelinclude one transition rate in TVM and two different ofsuch rates in TrN and GTR, as well as one transversion ratein TrN and four transversion rates in TVM and GTR. Thesemodel parameters were used when searching heuristicallyfor the most likely trees with 10 random addition replicatesand TBR branch-swapping.

Heuristic parsimony tree searches were performedwith 1000 random addition replicates and TBR branch-swapping without setting a maximum limit on number oftrees saved in memory, except for the Sulcia data set. Due tothe immense numbers of trees that resulted from preliminaryanalyses of Sulcia sequences, it was necessary to set themaximum number of trees (‘maxtrees’) to 100 000 and 100independent replicates of heuristic parsimony searcheswere performed, each with five random addition replicatesand subtree pruning-regrafting (SPR) branch-swapping.

Branch support was assessed by nonparametric characterbootstrapping (Felsenstein 1985), using 100 replicates inlikelihood [starting tree obtained by neighbour-joining,followed by nearest neighbour interchanges (NNI) branch-swapping] and 1000 replicates in parsimony (five randomadditions and SPR branch swapping) analyses. Once more,due to memory and time limitations, ‘maxtrees’ was set to100 000 for the Sulcia data set. Bremer and partitionedBremer decay indices (Bremer 1994) were calculated basedon symbiont and host most likely trees, respectively, withthe aid of treerot 2.0 (Sorenson 1999). These indices areconservative estimates as heuristic parsimony searcheswere conducted with only 20 random replicates and TBRbranch-swapping (‘maxtrees’ set to 10 000 when analysingthe Sulcia data set).

Host–symbiont associations

To assess the evidence for codiversification of the twosymbionts with their hosts, three statistical tests based ondifferent assumptions and algorithms were conducted. Aparsimony-based ILD test was conducted under the nullhypothesis that the host combined data set and each symbiontrDNA data set are congruent, suggesting a history ofcospeciation. Settings for ILD tests were as specified above.The advantage of the ILD test is that it is not biased due touncertainty regarding the topology of individual trees, asare the following two tests. Congruence of host andsymbiont topologies was also assessed with a Shimodaira–Hasegawa likelihood-based test (S–H; Shimodaira &Hasegawa 1999; Goldman et al. 2000) run with 10 000 RELL(re-estimation of likelihoods) bootstrap replicates. Consider-

ing the null hypothesis that the log-likelihood score of agiven host tree calculated using the host data set and modelof evolution is the same as the score calculated using thesymbiont data set and model of evolution (and vice versa),a failure to reject H0 is suggestive of a perfect cospeciationscenario. For these tests, a new data matrix containing onlythe taxa represented in all three data sets (n = 22) wasconstructed, and those taxa not included were prunedfrom most likely and parsimonious trees tested. All mostlikely and parsimonious trees were tested, except inthe case of Sulcia, where only 10 randomly chosen mostparsimonious trees were tested (topologies for host, n = 3;Baumannia, n = 6; Sulcia, n = 13). Finally, an event-basedtree-fitting method, implemented in the program treefitter1.0 (Ronquist 1998, 2002), was used to hypothesize parallelevolutionary events between host and symbiont trees. Thismethod also tests whether cospeciation events hypothesizedare more numerous than expected by chance, by comparinghost trees with 1000 randomly generated symbiont trees.Because treefitter does not allow input of polytomoustrees, the most likely tree for Baumannia, which containedone trifurcation, was resolved into the only two possiblestrictly bifurcating trees. The less resolved tree for Sulciawas resolved randomly into 20 different bifurcatingtopologies. These 22 symbiont topologies were comparedto the most likely host tree and events calculated based onthe following costs: 0 for codivergence and duplication, 1for sorting, and 0–30 for horizontal transfer (= host switches).

Evolutionary rates of bacterial 16S rDNA

To test whether the 16S rDNA of Baumannia and Sulciawere evolving with a constant rate across different host-associated lineages, a likelihood-ratio test was performed(Felsenstein 1981; Huelsenbeck & Crandall 1997). Afterremoving outgroups from the symbiont data sets, evolu-tionary models were chosen using modeltest as describedabove. Maximum-likelihood analyses were then run inpaup*, with one analysis constrained to follow a molecularclock and another unconstrained. The likelihood-ratio teststatistic, which should follow a chi-squared distribution,was calculated as twice the difference between the negativelog likelihoods (–ln Ls) of trees resulting from the un-constrained and constrained analyses, i.e. LRT = 2 [(–ln Lnonclock) – (–ln L clock)]. The null hypothesis that the rateof substitution is homogeneous among all branches in thetopology was tested, with degrees of freedom equal to thenumber of terminal taxa minus 2, i.e. d.f. = 23 for Baumanniaand d.f. = 25 for Sulcia.

Because phylogenetic results were consistent with anorigin of symbionts pre-dating or simultaneous with theorigin of the common ancestor to sharpshooters, we inferredthat the symbionts had likely evolved in sharpshootersover the same time interval. To compare rates of 16S rDNA

4182 D . M . T A K I Y A E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

evolution between Baumannia and Sulcia, nodes withcompatible groupings of taxa in both phylogenies, i.e.potentially cospeciating nodes, were identified. Branchesleading from these nodes to their most recent ancestor ineach phylogeny were recognized as copaths (i.e. homologousevolutionary branches, Page 1996). Copath maximum-likelihood lengths were compared using a Wilcoxon signedrank test (α = 0.05, two-tailed). Failure to reject the nullhypothesis that members of a copath have undergone thesame amount of evolution would imply that the two sym-bionts have the same substitution rate in their 16S rDNA.The slope of a reduced major axis regression (McArdle 1988;Johnson et al. 2003), i.e. the ratio of standard deviations ofcopath members, was calculated as a relative measure ofthe rate difference between the 16S rDNA of the twosymbiont lineages.

Results

Distribution of Baumannia, Sulcia and Wolbachia among hosts

Of the 29 species included in the study, Baumannia waspresent in 25, including all species of the cicadelline tribesProconiini and Cicadellini except Cicadella viridis, andalso in the single species of the subfamily Phereurhininae.Baumannia was not detected in Hylaius oregonensis (Bathy-smatophorini), Pagaronia tredecimpunctata (subfamilyEvacanthinae) or in Jikradia olitoria (subfamily Coelidiinae).To further test for presence of Baumannia in species notinitially yielding a Baumannia product, we used several otherprimer pairs expected to amplify Baumannia 16S rDNA andstill obtained no positive reactions. All negative reactionswere repeated, and positive controls, run concurrently, didyield PCR products. Some species not yielding Baumanniasequences did produce 16S rDNA sequences from unrelatedbacteria (not in the gamma-Proteobacteria), possiblycorresponding to other symbionts or to contaminants thatamplified in the absence of the high copy number of theBaumannia chromosome.

All species considered, representing six tribes, possessedSulcia, based on PCR amplification followed by DNAsequencing, blast searches, and phylogenetic analyses asdescribed below. In every case, the first blastn hit againstthe nonredundant GenBank nucleotide database was thesingle sequence of this symbiont type from the sharpshooterhost H. vitripennis, previously deposited in GenBank (Moranet al. 2003b).

From 1 to 10 individual insects from each host specieswere used for independent determinations of sequencesfrom both symbionts (Table 1). In all cases, the sequenceswere identical or near-identical for different individuals ofthe same species, with no cases of more than three nucleotidedifferences between pairs of either Sulcia or Baumannia

sequences within a species. The earlier study of Baumanniafrom H. vitripennis from Florida and California recoveredidentical 16S rDNA sequences from the two localities(Moran 2003b).

The most common additional sequences obtained corre-sponded to Wolbachia pipientis, a widespread symbiont ofinsects and other arthropods that is known to modifyreproductive biology of some hosts. Using screens basedon diagnostic primers for two protein coding genes (wspand ftsZ), we found Wolbachia in a total of 9 of the 27 speciesscreened. Sequences of wsp were obtained and depositedin GenBank (Table 1). Because Wolbachia was absent frommost species and often not universal in species in which itwas found, it was not considered to be an obligate symbiont.Its distribution showed no evident pattern with respect tohost phylogeny or classification. The sequences indicatedthe presence of several Wolbachia haplotypes among oursamples, and no two were identical. Because recent find-ings indicate that Wolbachia in insects shows high rates ofrecombination among genes and among regions of the wsplocus (Baldo et al. 2006), we did not use these sequences forphylogenetic reconstructions. The few other sequencesobtained from 16S rDNA amplifications were restricted tothe few species not containing Baumannia. These were eithercontaminants or possibly additional obligate or facultativesymbionts and are not reported since they were notconfirmed.

Host trees

The combined host data set yielded a single most likelytree (–ln L = 19 451.13), shown in Fig. 1, and two mostparsimonious trees. Clades that were present in theparsimony trees are indicated in bold, and clade support isshown in Fig. 1. Host trees were largely consistent withgeneric and tribal taxonomic groupings, but some differenceswere observed (discussed below).

Symbiont phylogenies and divergences

Phylogenetic analyses of Baumannia 16S rDNA sequencesyielded a single most likely tree (–ln L = 9175.09) (Fig. 2)and five most parsimonious trees. Sequence divergenceswithin Baumannia ranged from 1% to 19% (Table 3 A). Theanalyses gave strong support for the monophyly of Baumanniawithin the clade of gamma-3 Proteobacteria. Phylogeneticanalyses of the Sulcia 16S rDNA sequences resulted inthree most likely trees: the least resolved of these is shownin Fig. 3. Parsimony analyses gave approximately 200 000most parsimonious trees (length = 1012, CI = 0.73, RI = 0.78).The immense number of parsimonious trees reflects thelow divergence among the sequences from different hostspecies, with sequence divergence ranging from 0% to7% (Table 3B). Nonetheless, some nodes were strongly

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4183

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

supported in both maximum-likelihood and parsimonyanalyses, indicating that some phylogenetic informationwas present in the Sulcia data set. Both likelihood andparsimony analyses gave strong support for the monophylyof Sulcia within the Bacteroidetes.

Host–symbiont associations

The maximum-likelihood topologies for hosts and symbiontswere slightly different. However, these differences were notsignificant under the criterion of the maximum-likelihood-based S–H test. That is, the Baumannia data set andevolutionary model generated topologies not significantlyless likely than those produced from the host data set underits respective evolutionary model (diff –ln = 40.66–52.07,P = 0.35–0.42) and vice versa (diff –ln = 85.89–98.55, P =0.18–0.23). This result is consistent with the hypothesis ofperfect phylogenetic congruence of Baumannia and hosts.Results of the S–H test involving Sulcia were not self-consistent. Although host and Baumannia best topologies

were not statistically different from Sulcia best topologieswhen optimized with the Sulcia dataset and evolutionarymodel (diff −ln = 44.33–57.51, P = 0.17–0.24 and diff –ln =40.52–45.83, P = 0.23–0.27, respectively), Sulcia best topo-logies were found to be statistically different when optimizedwith host or Baumannia data sets and evolutionary model(diff –ln = 559.07–1053.85, P < 0.05 and diff –ln = 402.63–666.49, P < 0.05, respectively). In the event-based parsimonymethod of tree-fitting, whenever one to three horizontaltransfers (host switching cost = 10–15) were allowed, themaximum number of cospeciation events between hostsand symbionts (12 for both symbionts) was found. Thisnumber was significantly higher than would be expectedto occur by chance (P < 0.05 for all of Baumannia trees andfor most Sulcia trees), supporting a significant history ofcospeciation of each symbiont with hosts. Whenever no(host switching cost = 30) or a high number of horizontaltransfers (7–12, host switching cost = 0–2) were allowed,estimates of codivergence decreased. When host switchingwas not allowed, the null hypothesis of no switching could

100/100

Fig. 1 Single maximum-likelihood phy-logram for leafhopper hosts, based on thecombined gene data set (–ln L = 19 451.13).Thicker branches represent clades alsorecovered by maximum parsimony analyses.Numbers above branches are likelihood/parsimony bootstrap percentages (asterisksrepresent those below 50), and those beloware Bremer decay indices. Node-associatedgraphs represent partitioned Bremer decayindices (y-axis scale ranging from −15 to 30)for the partitions (from left to right) histoneH3, COI, COII, and 16S rDNA. Numberednodes indicated by black circles arecongruent with those indicated in bothsymbiont phylogenies.

4184 D . M . T A K I Y A E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

not be rejected for most Sulcia trees. Also, we note that thehigher number of transfer events is not biologicallyrealistic for Sulcia, which consists of huge cells confined tohost bacteriocytes. The disparate result for Sulcia in the S–H test and the failure to reject the null hypothesis for someof its reconstructions in treefitter is likely to be an artefactof the low phylogenetic information content of the 16SrDNA sequences of Sulcia, which yielded polytomoustopologies and low clade support statistics overall.

Finally, the ILD test did not reject the null hypothesesthat the gene sequences of hosts and each of the two sym-bionts share a common evolutionary history (P = 0.46 forhost and Baumannia and P = 0.96 for host and Sulcia).

In summary, the results of all tests suggest that thediversification of both endosymbionts was largely or entirelydependent on the phylogenetic history of their hostleafhoppers.

Evolutionary rates of bacterial 16S rDNA

For both symbionts, the null hypothesis of rate constancyamong lineages was rejected (P < 0.001). The differencesbetween the –ln Ls of analyses with unconstrained ratesand rates constrained to be constant were 37.35 and 49.65for Baumannia and Sulcia, respectively.

If Baumannia and Sulcia both cospeciated with theirhosts, as supported by our analyses, then comparing cor-responding sequence divergences for symbionts from thesame host pairs will give an estimate of the relative substi-tution rates, since codiversification implies that thesedivergences accumulated over the same time intervals, foreach pair. Inspection of the values in Table 3 suggests thatSulcia sequences evolve more slowly. However, many ofthese pairwise comparisons share overlapping branchesof the tree, or correspond to parts of the tree in which

Fig. 2 Single maximum-likelihood phy-logram for the Baumannia symbionts andrelatives in the gamma-Proteobacteria,based on partial 16S rDNA sequences (–lnL = 9175.09). Thicker branches representclades also recovered by maximum-parsimony analyses. Numbers abovebranches are likelihood/parsimony bootstrappercentages (asterisks represent thosebelow 50), and those below are Bremerdecay indices. Numbered nodes indicatedby black circles are congruent with bothhost and Sulcia phylogenies, while theone indicated by the black diamond iscongruent only with the Sulcia phylogeny.

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4185

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

codiversification was less strongly supported. In order toderive a more accurate estimate of the relative amount ofsequence evolution in Baumannia and Sulcia, substitutionrates were estimated by the lengths of branches in thetopologies generated, and only on those branches leadingto cospeciating nodes that are present in both symbionttrees (copaths). Based on the maximum-likelihood topo-logies, six copaths between Baumannia and Sulcia wereidentified and five of these are congruent with the hosttopology (Figs 1–3). A paired statistical analysis showedthat copath branch lengths from Baumannia are signifi-cantly different from those of Sulcia (W = 10.5, P = 0.03),

which implies different 16S rDNA substitution rates forthose two symbionts. In fact, the 16S rDNA of Baumanniaappears to be evolving 4.88 times faster than that of Sulcia,as suggested by the slope of the major axis regression ana-lysis (SD of Baumannia = 0.0156 and SD of Sulcia = 0.0032).

Discussion

Codiversification in a dual symbiosis

A congruent evolutionary history of both Baumannia andSulcia with their sharpshooter hosts is supported based on

Table 3 Maximum-likelihood pairwise distances among 16S rDNA sequences of selected Baumannia (A) and Sulcia (B) endosymbionts andrelated bacteria. Host names are in parentheses. Values represent number of substitutions per 100 bases based on models of molecularevolution specifiedA: GTR + G + I

B: TrN + G + I

2 3 4

Candidatus Baumannia cicadellinicola

5 6 7 8 9 10 11 12 13 14 15

1. Yersinia enterocolitica 7.64 23.79 18.69 17.76 19.69 23.09 19.05 25.53 21.08 20.53 22.96 22.86 22.21 22.492. Escherichia coli 17.68 19.46 18.72 18.19 24.81 19.82 26.07 22.53 23.46 24.29 23.28 25.03 25.223. Buchnera aphidicola 21.92 24.41 25.91 28.01 24.77 25.98 25.77 28.78 23.31 28.29 22.94 21.684. S-symb (A. lichtensioides) 14.22 15.86 14.23 13.44 13.65 14.30 14.60 14.87 16.02 15.70 13.865. S-symb (A. inermis) 16.86 16.69 13.97 15.72 14.09 15.49 16.60 15.30 18.37 16.716. Blochmannia (C. herculeanus) 17.22 16.14 18.00 17.67 19.39 19.75 17.63 21.17 18.387. (G. coccinea) 9.30 10.15 9.14 11.17 14.33 10.58 14.63 14.658. (P. isabellina) 8.56 6.53 9.38 12.23 8.56 13.62 12.479. (C. catapulta) 6.74 10.20 12.59 10.14 14.78 12.7910. (A. virescens) 7.90 12.17 7.31 13.78 12.7411. (C. costalis) 14.30 4.49 17.14 14.7112. (D. excisum) 16.11 8.36 7.2513. (H. vitripennis) 19.25 16.1914. (H. irregularis) 5.8015. (P. alalia) —

2 3 4

Candidatus Sulcia muelleri

5 6 7 8 9 10 11 12 13 14 15

1. Flexibacter tractuosus 23.27 25.30 27.54 30.24 29.91 34.85 39.04 30.91 30.83 31.01 31.42 29.37 29.32 29.442. Weeksella virosa 25.98 26.44 32.76 29.51 33.40 61.56 29.23 32.35 29.70 31.08 28.14 30.24 29.503. Endoparasite (A. variegata) 13.89 15.59 16.38 18.79 22.38 16.33 16.80 16.53 17.60 16.92 16.31 16.044. Blattabacterium (M. darwiniensis) 18.80 18.99 21.79 29.28 18.55 20.47 18.52 20.28 18.85 19.39 18.925. (J. olitoria) 2.12 4.13 3.13 2.58 2.73 2.58 3.12 3.11 2.21 2.136. (P. tredecimpunctata) 2.35 2.93 1.65 1.83 1.58 1.11 1.94 1.11 1.037. (G. coccinea) 4.87 1.15 1.84 0.98 1.93 0.74 2.28 1.708. (P. isabellina) 0.20 1.42 0.62 3.74 1.30 1.69 2.329. (C. catapulta) 0.40 0.49 1.15 0.83 1.23 1.0610. (A. virescens) 0.08 2.09 0.57 1.26 1.2711. (C. costalis) 1.08 0.66 1.16 1.1612. (D. excisum) 1.43 0.71 0.0813. (H. vitripennis) 1.52 1.5114. (H. irregularis) 0.2315. (P. alalia) –

4186 D . M . T A K I Y A E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

all (Baumannia) or most (Sulcia) statistical tests conductedhere, suggesting a long-term association of these bacteriawith their hosts. Prior to our study, molecular phylogeneticstudies of hosts and symbionts have documented long-term cospeciation of only a single symbiont clade with itshosts. These examples of primary symbionts includeB. aphidicola in aphids (Moran et al. 1993; Baumann 2005),Wigglesworthia in tsetse flies (Chen et al. 1999), Blochmanniain carpenter ants (Sauer et al. 2000), Blattabacterium incockroaches and termites (Bandi et al. 1994; Lo et al. 2003),Carsonella in psyllids (Spaulding & von Dohlen 2001; Thaoet al. 2001), Tremblaya in mealybugs (Thao et al. 2002), andPortiera in whiteflies (Thao & Baumann 2004a). Each ofthese host groups is considered to have a single primarysymbiont representing an ancient infection followed bycodiversification.

We identified Sulcia and Baumannia in all but one of thesharpshooters tested (Baumannia was absent from C. viridis),and both have other features of typical primary symbiontssuch as large irregularly shaped cells, restriction to bacte-

riomes, and small genome sizes (Moran et al. 2003b).Neither Sulcia nor Baumannia can be considered a secondarysymbiont, in view of their ancient associations with hosts,supported by our results, and their apparent nutritionalcontributions to hosts, supported by genomic results (Wuet al. 2006). We suggest the term ‘coprimary’ symbionts forthis and other similar cases, in which two or more symbiontsare obligate and ancient bacteriome-associates of a hostgroup. Baumannia was found only in the sharpshooter tribesCicadellini and Proconiini and the subfamily Phereurhininae;whereas Sulcia was present in all leafhoppers tested(Table 1). Preliminary observations on members of othergroups of Auchenorrhyncha suggest that Baumannia is notfound in groups outside of Cicadellidae, but that Sulcia iswidely distributed among auchenorrhynchan lineages(Moran et al. 2005a).

Our findings suggest that both Sulcia and Baumannia areancient bacteriome-associates that reflect long-term co-diversification of two bacteria clades with sharpshootersand relatives. These findings are in general agreement with

Fig. 3 Selected maximum-likelihoodphylogram (one of three) for the Sulciasymbionts and relatives in the Bacteroidetesphylum of Bacteria, based on partial 16SrDNA sequences (–ln L = 6993.82). Thesmaller cladogram on the left shows thesame topology as the phylogram, but withoutgroups pruned for better visualizationof clades and support. Thicker branchesrepresent clades also recovered by maximum-parsimony analyses. Numbers abovebranches are likelihood/parsimonybootstrap percentages (asterisks representthose below 50), and those below are Bremerdecay indices. Numbered nodes indicatedby black circles are congruent with bothhost and Sulcia phylogenies, while the oneindicated the black diamond is congruentonly with the Baumannia phylogeny.

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4187

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

Buchner’s views: he recognized leafhoppers and theirrelatives as having the most elaborate symbioses of anyinsects and hypothesized long-term co-evolution of two ormore bacterial associates with particular leafhoppersubfamilies or tribes. Thus, in his proposed evolutionaryscenario, based largely on the work of his student, H. J.Müller, one symbiotic association was ancestral to theAuchenorrhyncha, with different host lineages acquiringand losing additional symbionts during the radiation ofthis large clade, resulting in a mosaic of different symbiontcombinations across modern subfamilies and tribes.Buchner hypothesized that the sharpshooters retained theancestral (‘a’) symbiont plus a more recently acquiredsymbiont, designated the ‘t’ symbiont. Under this view,the paired, bilateral bacteriomes are homologous organsretained by most groups of Cicadomorpha (the lineage thatincludes leafhoppers, treehoppers, spittlebugs and cica-das), during their evolution from an ancestor containing anoriginal symbiont type.

Nutritional roles for primary symbionts and implications for host diversification

The evidence for co-inheritance of both Baumannia andSulcia across millions of years strongly suggests that bothsymbionts are essential for the host insect. Other ancientbacteriome-associated symbionts are known to provideneeded nutrients to host insects, a role inferred in part fromgenomic analyses of the symbionts (e.g. Shigenobu et al.2000). The xylem sap diet of sharpshooters lacks manynutrients required by animals, including essential aminoacids and cofactors (Redak et al. 2004). Stable coprimarysymbiosis might result if different symbiont types areproviding different sets of required nutrients. Genomicanalyses of Baumannia and Sulcia of the host H. vitripennisprovide strong evidence that the two symbionts indeedserve complementary nutritional roles in the provisioningof amino acids and cofactors (Wu et al. 2006). Despite itssmall genome size, Baumannia retains pathways for numerouscofactors (vitamins) needed by the host, with a total ofabout 12% of its genome devoted to these processes. Sulciawas partially sequenced and has genes for production ofmost essential amino acids.

The dependence on more than one obligate symbiont ispotentially a constraining factor in the evolution and hostplant range of sharpshooters and other auchenorrhynchans.Genome comparisons for primary symbionts of both aphidsand carpenter ants indicate that these bacteria do notacquire foreign genes and do undergo continued erosion ofthe ancestral genome (Tamas et al. 2002; Degnan et al.2005). For example, Buchnera lineages continue to losegenes, and some of these appear to impose new nutritionalrequirements on hosts, such as a requirement for dietarysources of fixed sulphur, available in only some plants

(Tamas et al. 2002). In H. vitripennis, Baumannia has a verysmall genome (686 kb; Moran et al. 2003b; Wu et al. 2006),and Sulcia also has a genome considerably under 1 mega-base (Wu et al. 2006; P. Tran & N. Moran, unpublisheddata), as do all primary symbionts for which genome sizesare known. Indeed, an apparent distinction between pri-mary and secondary symbionts is the extent to which theirgenomes continue to acquire novel genes and genearrangements, and thus the extent to which they representa source of novelty in host evolution. In both aphids andtsetse flies, recent genomic studies indicate that secondarysymbionts have larger and more dynamic genomes, withphage, repetitive elements and evidence of recent geneacquisition, contrasting with the primary symbionts in thesame hosts, which have near-minimal genome sizes(Moran et al. 2005b; Toh et al. 2005).

Although the Baumannia-Sulcia-sharpshooter symbiosisis the first case in which multiple obligate and apparentlyancient microbial symbionts have been studied usingmolecular data, many more similar cases of complexsymbioses appear to exist, based on microscopy of othersystems (Buchner 1965). Hosts with multiple primary sym-bionts are likely to be highly dependent on the gene sets ofeach endosymbiont, with possible consequences for theirown evolution. Alternatively, stable co-occurrence of morethan one symbiont might enable exchange of metabolitesbetween them that enables even more extensive loss ofgenes in one or both genomes. Further studies of symbiontgenomes and interactions with hosts will be required toassess how co-evolution with multiple symbionts affectshost biology and diversification.

Variable rates of evolution in symbiont sequences

Our estimates of divergences for 16S rDNA sequencesof Baumannia, based on maximum-likelihood models, aresimilar to the distance divergences estimated using aJukes–Cantor model for those five representatives studiedpreviously by Moran et al. (2003b). In their study, themaximum divergence of about 7% was found betweenlineages of Graphocephala (Cicadellini) and Homalodisca(Proconiini). Furthermore, they suggested that Baumanniahas an approximately equal rate of nucleotide basesubstitution for 16S rDNA to Buchnera aphidicola, theprimary endosymbiont of aphids, based on the observationthat Baumannia and Buchnera sequences show very similardivergences to outgroups such as Escherichia coli and Yersiniapestis (Table 3A). Based on previous molecular studies,B. aphidicola shows a per-lineage substitution rate of about1–2% per 50 million years, corresponding to a divergencerate between paired lineages of 2−4% per 50 million years(Moran et al. 1993). Accordingly, Moran et al. (2003b)hypothesized that the Baumannia common ancestor wasassociated with the sharpshooter common ancestor around

4188 D . M . T A K I Y A E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

80–175 million years ago. However, the larger taxon sampleof the present study facilitated our discovery that theminimum and maximum divergences between membersof the most basal clade of Baumannia (associated withhosts of the Proconiini clade which includes Diestostemma)and other Baumannia representatives were 11% and 19%,respectively. If the 16S rDNA of Baumannia is evolving atthe same rate as that of Buchnera, then these observeddivergences would imply that the sharpshooter lineage isbetween 138 and 475 million years old. These dates aresubstantially greater than those suggested by the fossilrecord, suggesting that at least some lineages of Baumanniahave a faster 16S rRNA substitution rate than do Buchnera.Although the earliest fossils of Cicadomorpha are 270million years old, leafhoppers (Cicadellidae) do not appearin the fossil record until 130 million years later (Shcherbakov2002), and fossil leafhoppers definitely assignable to extantsubfamilies do not appear until 50 million years ago (Balticamber, Szwedo 2002). Nevertheless, no true sharpshooteris known from this period (including Ambericarda skalskiiSzwedo & Gebicki 1998; which was incorrectly placed inProconiini, J. Szwedo, personal communication). Only inDominican amber (25–40 million years ago) do true sharp-shooters first appear as fossils (Dietrich & Vega 1995). Thisimplies that at least some lineages of Baumannia evolve ata faster rate than does Buchnera, and/or that the sharpshooterlineage is older than indicated by available fossil evidence.Previous studies have shown that bacterial groups canhave very different rates of substitution in DNA sequences(Ochman et al. 1999).

Indeed, a primary result of the present study is thatBaumannia and Sulcia have dramatically different rates of16S rDNA substitution, with Baumannia evolving approxi-mately five times faster than Sulcia. The basis for the muchslower rate of evolution in Sulcia is not known. Because itis the first bacteriome-associate outside of the Proteobacteriato be studied using molecular sequence data, its slow ratemay reflect basic differences in mutational processesbetween bacterial phyla. We note that Sulcia does showsome other genomic features that are characteristic ofbacteriome-associated symbionts, including AT-biasedgenome composition (about 24% G + C) and small genomesize (Wu et al. 2006). Finally, results of the likelihood-ratiotest suggest that both Baumannia and Sulcia 16S rDNA sub-stitution rates are not constant among the different line-ages, which would introduce errors in age estimates basedon the assumption of a constant molecular clock.

Wolbachia in leafhoppers

Our findings for Wolbachia are consistent with previousstudies showing that leafhoppers are frequent hosts forthis symbiont group and that it has an erratic distributionwith respect to host phylogeny (Mitsuhashi et al. 2002;

Kittayapong et al. 2003). Noncongruent phylogenies aretypically found between Wolbachia and their hemipteran(Kikuchi & Fukatsu 2003) or other insect hosts (Werrenet al. 1995b; Shoemaker et al. 2002; Bordenstein & Rosengaus2005; Kyei-Poku et al. 2005). Results for Wolbachia are astriking constrast to those for Baumannia and Sulcia andweigh against a role for Wolbachia as an obligate mutualistof leafhoppers. Indeed its most well-documented effect oninsects hosts are modifications of the reproductive biology,such as cytoplasmic incompatibility and son-killing (Werrenet al. 1995b). We do not know what phenotypes Wolbachiaconfers upon the leafhopper hosts that we examined.Wolbachia is often variably present within host species.Since most of our study species were represented by one orfew collections, more extensive sampling might revealWolbachia infections in some of the species for which wefound none.

Implications for leafhopper phylogenetics and classification

Results of the present analysis of sharpshooter hosts agreewith recent higher-level phylogenetic studies of leafhopperssuggesting that the subfamily Cicadellinae sensu Omanet al. (1990) is not a monophyletic unit (Dietrich 1999, 2004;Dietrich et al. 2001). These previous studies, based onanalyses of morphology and 28S rDNA sequences, supportYoung 1968, 1977) inclusion of only two tribes, Cicadelliniand Proconiini, in the subfamily, but also suggest thatPhereurhininae is derived from the same lineage. Thisdelimitation of Cicadellinae is consistent not only withour present phylogenetic analyses of host nuclear andmitochondrial, but also of symbiont genes, suggesting thepossibility of using these endosymbiont genes for inferringhost phylogenetic histories. Moreover, Baumannia appearsto be found only in this leafhopper lineage, and its presencecould itself be considered a highly informative characterfor phylogenetic reconstruction.

Conclusions and general significance of this study

Initial molecular studies on bacterial–eukaryote symbiosesfocused on relatively simple systems, involving a singlebacterial lineage or a single primary symbiont combinedwith opportunistic secondary symbionts (Baumann 2005).However, complex symbioses involving multiple partnersare common in eukaryotes. For example, Buchner (1965)describes numerous systems with several co-inheritedbacterial and/or fungal symbionts in animal hosts. Ourresults support the hypothesis that the sharpshooter dualsymbiosis has been stably inherited during the evolution ofthis insect group and raise the likelihood that bacterialsymbiosis has been a major element governing the ecologicaldiversification of these insects. In the case of the single

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4189

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

primary symbiont Buchnera in aphids, the symbiosis benefitshosts by providing a source of needed nutrients (Shigenobuet al. 2000; Baumann 2005). But, because Buchnera lineagesundergo irreversible gene loss over time, this symbiosiscan also be a constraining factor, and current evidence fromgenomics suggests that gene losses and genome degradationin Buchnera may limit ecological capabilities of aphids,confining them to certain host plant groups (Tamas et al.2002). In the case of multiple bacterial partners, theseconstraints are likely to be more complex, Baumannia andSulcia resemble Buchnera is having reduced genomes andlimited metabolic capabilities (Wu et al. 2006). The evolutionof each symbiont is likely to have had major consequencesfor the other as well as for hosts. For the first time, acombination of phylogenetic and genomic approaches isenabling us to recognize these intertwined histories and tounderstand their consequences for host ecology andevolution.

Acknowledgements

Gathering of material used in this project would not have beenpossible without the help of the following individuals: B. Alvarado(INRENA, Peru), C. Godoy (INBIO, Costa Rica), J. Grados, G. Lamas,C. Peña, and T. Pequeño (Museo de Historia Natural, UMSM,Peru), P. Lozada and G. Solis (SENASA, Peru), S. McKamey(United States National Museum, USA), R. Mizell III (Universityof Florida, USA), G. Moya-Raygoza (Universidad de Guadalajara,Mexico), and R. Rakitov (Illinois Natural History Survey, INHS,USA). K. Johnson (INHS, USA) kindly provided primers for H3gene, and, along with J. Banks (University of Illinois at Urbana-Champaign, USA), suggestions on the analyses of host–symbiontcospeciation. We also thank H. Dunbar, J. Russell, and H. Ochman(University of Arizona, USA) for assistance and comments, P.Baumann (University of California, Davis, USA) for comments,including the suggestion of the term ‘coprimary symbiont’, andBecky Nankivell for helping to prepare the manuscript. Prelimin-ary versions of this manuscript benefited by the useful commentsof R. Rakitov (INHS). D. M. Takiya’s graduate studies werefunded by Conselho Nacional de Desenvolvimento Científico eTecnológico (CNPq, Brazil). This study was supported by NationalScience Foundation (USA) grants DEB-0089671 to C. Dietrich andDEB-0313737 to N. Moran.

ReferencesAkman L, Yamashita A, Watanabe H et al. (2002) Genome

sequence of the endocellular obligate symbiont of tsetse flies,Wigglesworthia glossinidia. Nature Genetics, 32, 402–407.

Aksoy S, Chen X, Hypsa V (1997) Phylogeny and potential trans-mission routes of midgut-associated endosymbionts to tsetse(Diptera: Glossinidae). Insect Molecular Biology, 6, 183–190.

Andersen PC, Brodbeck BV, Mizell RF III (1989) Metabolism ofamino acids, organic acids and sugars extracted from the xylemfluid of four host plants by adult Homalodisca coagulata. EntomologiaExperimentalis et Applicata, 50, 149–159.

Baldo L, Bordenstein S, Wernegreen JJ, Werren JH (2006)Widespread recombination throughout Wolbachia genomes.Molecular Biology and Evolution, 23, 437–449.

Bandi CG, Damiani L, Magrassi A et al. (1994) Flavobacteria asintracellular symbionts of cockroaches. Proceedings of the RoyalSociety of London, Series B, Biological Sciences, 257, 43–48.

Baumann P (2005) Biology of bacteriocyte-associated endosym-bionts of plant sap-sucking insects. Annual Review of Microbiology,59, 155–189.

Baumann L, Baumann P, Moran NA, Sandström J, Thao ML (1999)Genetic characterization of plasmids containing genes encodingenzymes of leucine biosynthesis in endosymbionts (Buchnera) ofaphids. Journal of Molecular Evolution, 48, 77–85.

Bender W, Spierer P, Hogness DS (1983) Chromosomal walkingand jumping to isolate DNA from the ace and rosy loci and thebi thorax complex in Drosophila melanogaster. Journal of MolecularBiology, 168, 17–34.

Bordenstein S, Rosengaus RB (2005) Discovery of a novel Wolbachiasupergroup in Isoptera. Current Microbiology, 51, 393–398.

Bremer K (1994) Branch support and tree stability. Cladistics, 10,295–304.

Buchner P (1965) Endosymbiosis of Animals with Plant Microorganisms.John Wiley, New York.

Burks RE, Redak RA (2003) The identity and reinstatement ofHomalodisca liturata Ball and Phera lacerta Fowler (Hemiptera:Cicadellidae). Proceedings of the Entomological Society of Washington,105, 674–678.

Chang KP, Musgrave AJ (1972) Multiple symbiosis in a leaf-hopper, Helochara communis Fitch (Cicadellidae: Homoptera):envelopes, nucleoids and inclusions of the symbiotes. Journal ofCell Science, 11, 275–293.

Chen X, Li S, Aksoy S (1999) Concordant evolution of a symbiontwith its host insect species: molecular phylogeny of genus Gloss-ina and its bacteriome-associated endosymbiont, Wigglesworthiaglossinidia. Journal of Molecular Evolution, 48, 49–58.

Cunningham CW (1997) Can three incongruence tests predictwhen data should be combined? Molecular Biology and Evolution,14, 733–740.

Degnan PH, Lazarus AB, Wernegreen JJ (2005) Genome sequenceof Blochmannia pennsylvanicus indicates parallel evolutionarytrends among bacterial mutualists of insects. Genome Research,15, 1023–1033.

Dietrich CH (1999) The role of grasslands in the diversificationof leafhoppers (Homoptera: Cicadellidae): a phylogenetic per-spective. Proceedings of the 15th North American Prairie Conference,44–49.

Dietrich CH (2004) Phylogeny of the leafhopper subfamilyEvacanthinae with a review of Neotropical species and noteson related groups (Hemiptera: Membracoidea: Cicadellidae).Systematic Entomology, 29, 455–487.

Dietrich CH, Vega FE (1995) Leafhoppers (Homoptera, Cicadellidae)from Dominican amber. Annals of the Entomological Society ofAmerica, 88, 263–270.

Dietrich CH, Rakitov RA, Holmes JL, Black WC (2001) Phylogenyof the major lineages of Membracoidea (Insecta: Hemiptera:Cicadomorpha) based on 28S rDNA sequences. MolecularPhylogenetics and Evolution, 18, 293–305.

Douglas AE (1998) Nutritional interactions in insect-microbialsymbioses: aphids and their symbiotic bacteria Buchnera. AnnualReview of Entomology, 43, 17–37.

Farris JS, Källersjö M, Kluge AG, Bult C (1995) Constructing a sig-nificance test for incongruence. Systematic Biology, 44, 570–572.

Felsenstein J (1981) Evolutionary trees from DNA sequences: amaximum likelihood approach. Journal of Molecular Evolution,17, 368–376.

4190 D . M . T A K I Y A E T A L .

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

Felsenstein J (1985) Confidence limits of phylogenies: an approachusing the bootstrap. Evolution, 39, 783–791.

Fukatsu T, Nikoh N (2000) Endosymbiotic microbiota of thebamboo pseudococcid Antonina crawii (Insecta: Homoptera).Applied and Environmental Microbiology, 66, 643–650.

Goldman N, Anderson JP, Rodrigo AG (2000) Likelihood-basedtests of topologies in phylogenetics. Systematic Biology, 49, 652–670.

Hamilton KGA, Zack RS (1999) Systematics and range fragmenta-tion of the Nearctic genus Errhomus (Rhynchota: Homoptera:Cicadellidae). Annals of the Entomological Society of America, 92,312–354.

Hendson M, Purcell AH, Chen DQ, Smart C, Guilhabert M,Kirkpatrick B (2001) Genetic diversity of Pierce’s disease strainsand other pathotypes of Xylella fastidiosa. Applied and Environ-mental Microbiology, 67, 895–903.

Huelsenbeck JP, Crandall KP (1997) Phylogeny estimation andhypothesis testing using maximum likelihood. Annual Review ofEcology and Systematics, 28, 437–466.

Johnson KP, Cruickshank RH, Adams RJ, Smith VS, Page RDM,Clayton DH (2003) Dramatically elevated rate of mitochondrialsubstitution in lice (Insecta: Phthiraptera). Molecular Phylogeneticsand Evolution, 26, 231–242.

Kaiser B (1980) Licht und elektronenmikroskopische Untersuc-hung der symbionten von Graphocephala coccinea Forstier [sic!](Homoptera: Jassidae). International Journal of Insect Morphologyand Embryology, 9, 79–88.

Kikuchi Y, Fukatsu T (2003) Diversity of Wolbachia endosymbiontsin Heteropteran bugs. Applied and Environmental Microbiology,69, 6082–6090.

Kittayapong P, Jamnongluck W, Thipaksorn A, Milne JR,Sindhusake C (2003) Wolbachia infection complexity amonginsects in the tropical rice-field community. Molecular Ecology,12, 1049–1060.

Kyei-Poku GK, Colwell DD, Coghlin P, Benkel B, Floate KD (2005)On the ubiquity and phylogeny of Wolbachia in lice. MolecularEcology, 14, 285–294.

Lo N, Bandi C, Watanabe H, Nalepa C, Beninati T (2003) Evidencefor cocladogenesis between diverse dictyopteran lineages andtheir intracellular endosymbionts. Molecular Biology and Evolution,20, 907–913.

McArdle BH (1988) The structural relationship: regression in biology.Canadian Journal of Zoology, 66, 2329–2339.

Mickevich MF, Farris JS (1981) The implications of congruence inMenidia. Systematic Zoology, 30, 351–370.

Mitsuhashi W, Saiki T, Wei W, Kawakita H, Sato M (2002) Twonovel strains of Wolbachia coexisting in both species of mulberryleafhoppers. Insect Molecular Biology, 11, 577–584.

Moran NA (1998) Bacteriocyte-associated symbionts of insects.Bioscience, 48, 295–304.

Moran NA, Baumann P (1994) Phylogenetics of cytoplasmicallyinherited microorganisms of arthropods. Trends in Ecology &Evolution, 9, 15–20.

Moran NA, Munson MA, Baumann P, Ishikawa H (1993) Amolecular clock in endosymbiotic bacteria is calibrated usingthe insect hosts. Proceedings of the Royal Society of London. SeriesB, Biological Sciences, 253, 167–171.

Moran NA, Plague GR, Sandström J, Wilcox JL (2003a) A genomicperspective on nutrient provisioning by bacterial symbionts ofinsects. Proceedings of the National Academy of Sciences, USA, 100,14543–14548.

Moran NA, Dale C, Dunbar H, Smith WA, Ochman H (2003b)Intracellular symbionts of sharpshooters (Insecta: Hemiptera:

Cicadellinae) form a distinct clade with a small genome.Environmental Microbiology, 5, 116–126.

Moran NA, Tran P, Gerardo NM (2005a) Symbiosis and insectdiversification: an ancient symbiont of sap-feeding insects fromthe bacterial phylum Bacteroidetes. Applied and EnvironmentalMicrobiology, 71, 8802–8810.

Moran NA, Degnan PH, Santos SR, Dunbar HE, Ochman H(2005b) The players in a mutualistic symbiosis: Insects, bacteria,viruses and virulence genes. Proceedings of the National Academyof Sciences, USA, 102, 16919–16926.

Nault LR, Rodriguez JG (1985) The Leafhoppers and Planthoppers.Wiley, New York.

Ochman H, Elwyn S, Moran NA (1999) Calibrating bacterialevolution. Proceedings of the National Academy of Sciences, USA,26, 12638–12643.

Ogden TH, Whiting MF (2003) The problem with ‘the Paleopteraproblem’: sense and sensitivity. Cladistics, 19, 432–442.

Oman PW, Knight WJ, Nielson MW (1990) Leafhoppers (Cicadellidae):A Bibliography, Generic Check-List and Index to the World Litera-ture, 1956–1985. CAB International Institute of Entomology,Wallingford, UK.

Page RDM (1996) Temporal congruence revisited: comparison ofmitochondrial DNA sequence divergence in cospeciating pocketgophers and their chewing lice. Systematic Biology, 45, 151–167.

Posada D, Crandall KA (1998) Modeltest: testing the model ofDNA substitution. Bioinformatics, 14, 817–818.

Purcell AH, Hopkins DL (1996) Fastidious xylem-limited bacterialplant pathogens. Annual Review of Phytopathology, 34, 131–151.

Purcell AH, Saunders SR, Hendson M, Grebus ME, Henry MJ(1999) Causal role of Xylella fastidiosa in oleander leaf scorchdisease. Phytopathology, 89, 53–58.

Redak RA, Purcell AH, Lopes JR et al. (2004) The biology of xylemfluid-feeding insect vectors of Xylella fastidiosa and their relation todisease epidemiology. Annual Review of Entomology, 49, 243–270.

Ronquist F (1998) Three-dimensional cost matrix optimizationand maximum cospeciation. Cladistics, 14, 167–172.

Ronquist F (2002) Parsimony analysis of coevolving species asso-ciations. In: Tangled Trees: Phylogeny, Cospeciation and Coevolution(ed. Page RDM), pp. 22–64. University of Chicago Press, Chicago.

Sandström J, Moran N (1999) How nutritionally imbalanced isphloem sap for aphids? Entomologia Experimentalis et Applicata,91, 203–210.

Sandström JP, Russell JA, White JP, Moran NA (2001) Independentorigins and horizontal transfer of bacterial symbionts of aphids.Molecular Ecology, 10, 217–228.

Sauer C, Stackebrandt E, Gadau J, Hölldobler B, Gross R (2000)Systematic relationships and cospeciation of bacterial endosym-bionts and their carpenter ant host species: proposal of the newtaxon Candidatus Blochmannia General nov. International Journalof Systematic and Evolutionary Microbiology, 50, 1877–1886.

Shcherbakov D (2002) The 270 million year history of Auchenor-rhyncha. Denisia, 4, 29–36.

Shigenobu S, Watanabe H, Hattori M, Sakaki Y, Ishikawa H (2000)Genome sequence of the endocellular bacterial symbiont ofaphids Buchnera sp. APS. Nature, 407, 81–86.

Shimodaira H, Hasegawa M (1999) Multiple comparison oflog-likelihoods with applications to phylogenetic inferences.Molecular Biology and Evolution, 16, 1114–1116.

Shoemaker DD, Machado CA, Molbo D et al. (2002) The distribu-tion of Wolbachia in fig wasps: correlations with host phylogeny,ecology and population structure. Proceedings of the Royal Societyof London. Series B, Biological Sciences, 269, 2257–2267.

C O D I V E R S I F I C A T I O N O F S H A R P S H O O T E R E N D O S Y M B I O N T S 4191

© 2006 The AuthorsJournal compilation © 2006 Blackwell Publishing Ltd

Simon C, Frati F, Beckenbach A et al. (1994) Evolution, weightingand phylogenetic utility of mitochondrial gene sequences and acompilation of conserved polymerase chain reaction primers.Annals of the Entomological Society of America, 87, 651–701.

Sorenson MD (1999) Treerot, version 2. Boston University, Boston,Massachusetts.

Spaulding AW, von Dohlen CD (2001) Psyllid endosymbiontsexhibit patterns of co-speciation with hosts and destabilizingsubstitutions in ribosomal RNA. Insect Molecular Biology, 10,57–67.

Swofford DL (1998) PAUP*, Phylogenetic Analysis Using Parsimony(*and Other Methods), version 4. Sinauer Associates, Sunderland,Massachusetts.

Szwedo J (2002) Amber and amber inclusions of planthoppers,leafhoppers and their relatives (Hemiptera, Archeorrhyncha etClypaeorrhyncha). Denisia, 4, 37–56.

Szwedo J, Gebicki C (1998) Ambericarda skalskii gen. n. et sp. n.from Baltic amber (Homoptera: Cicadellidae). Estudios DelMuseo de Ciencias Naturales de Alava, 14, 201–206.

Takiya DM, McKamey SH, Cavichioli RR (2006) Validity ofHomalodisca and of. H. vitripennis as the Name for the Glassy-Winged Sharpshooter (Hemiptera: Cicadellidae: Cicadellinae).Annals of the Entomological Society of America, 99, 648–655.

Tamas I, Klasson L, Canback B et al. (2002) 50 million years ofgenomic stasis in endosymbiotic bacteria. Science, 296, 2376–2379.

Thao ML, Baumann P (2004a) Evolutionary relationships of pri-mary prokaryotic endosymbionts of whiteflies and their hosts.Applied and Environmental Microbiology, 70, 3401–3406.

Thao ML, Baumann P (2004b) Evidence for multiple acquisition ofArsenophonus by whitefly species (Sternorrhyncha: Aleyrodidae).Current Microbiology, 48, 140–144.

Thao ML, Clark M, Baumann L, Brennan EB, Moran NA, BaumannP (2000) Secondary symbionts of psyllids have been acquiredmultiple times. Current Microbiology, 41, 300–304.

Thao ML, Clark MA, Burckhardt DH, Moran NA, Baumann P(2001) Phylogenetic analysis of vertically transmitted psyllidendosymbionts (Candidatus Carsonella ruddii) based on atpAGDand rpoC: comparisons with 16S−23S rDNA-derived phylogeny.Current Microbiology, 42, 419–421.

Thao ML, Gullan PJ, Baumann P (2002) Secondary (gamma-Proteobacteria) endosymbionts infect the primary (beta-Proteobacteria) endosymbionts of mealybugs multiple timesand coevolve with their hosts. Applied and Environmental Micro-biology, 68, 3190–3197.

Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG(1997) The ClustalX windows interface: flexible strategies formultiple sequence alignment aided by quality analysis tools.Nucleic Acids Research, 24, 4876–4882.

Toh H, Weiss BL, Perkin SA et al. (2006) Massive genome erosionand functional adaptations provide insights into the symbioticlifestyle of Sodalis glossinidius in the tsetse host. Genome Research,16, 149–156.

Van Meer MMM, Witteveldt J, Stouthamer R (1999) Phylogeny ofthe arthropod endosymbiont Wolbachia based on the wsp gene.Insect Molecular Biology, 8, 399–408.

Werren JH, Windsor D, Guo LR (1995a) Distribution of Wolbachiaamong neotropical arthropods. Proceedings of the Royal Society ofLondon, Series B, Biological Sciences, 262, 197–204.

Werren JH, Zhang W, Guo LR (1995b) Evolution and phylogeny ofWolbachia: reproductive parasites of arthropods. Proceedings ofthe Royal Society of London, Series B, Biological Sciences, 261, 55–63.

Wu D, Daugherty SC, Van Aken SE et al. (2006) Metabolic comple-mentarity and genomics of the dual symbiosis of sharpshooters.Public Library of Science — Biology, 4, e188.

Yang Z (1996) Among-site rate variation and its impact on phylo-genetic analyses. Trends in Ecology & Evolution, 11, 367–372.

Young DA (1968) Taxonomic study of the Cicadellinae (Homoptera:Cicadellidae), Part 1, Proconiini. Bulletin of the United StatesNational Museum, 261, 1–287.

Young DA (1977) Taxonomic study of the Cicadellinae (Homo-ptera: Cicadellidae), Part 2, New World Cicadellini and the genusCicadella. Technical Bulletin of the North Carolina AgriculturalExperiment Station, 239, 1–1135.

Zhou W, Rousset F, O’Neill S (1998) Phylogeny and PCR-basedclassification of Wolbachia strains using wsp gene sequences. Pro-ceedings of the Royal Society of London, Series B, Biological Sciences,265, 509–515.

This work was a collaborative study between the laboratories ofDr Nancy Moran at the University of Arizona (Tucson) and DrChris Dietrich at the Illinois Natural History Survey (Champaign).The data presented are part of research conducted for Phat Tran’sBS honors thesis and Daniela M. Takiya’s PhD dissertation. NancyMoran is broadly interested in the evolution and ecology of insectsand their microbial symbionts. Chris Dietrich focuses on thesystematics and evolution of leafhoppers and related groups ofinsects (Hemiptera: Membracoidea).