Antibiotic resistance in bacteriaN. Symonds the bacterial genome. It is now known that this mutation is in a gene which has something to do with the structure of the ribosomes, and

Postgraduate Medical Journal (April 1972) 48, 216-221.

Antibiotic resistance in bacteria

N. SYMONDSB.A., BSc., Ph.D

Professor of Microbial Genetics, School of Biological Sciences, University of Sussex

To a microbial geneticist two aspects of the problemsarising in the field of bacterial drug resistance are ofspecial interest. The first is that it is one of the fewareas where there is a direct contact between medicalpractice and basic research in molecular genetics:and the second is that it is a subject in which evolu-tion can be observed in one's own time.

Before discussing aspects of bacterial drugresistance from the medical point of view, it isnecessary first of all to set the stage, so to speak, andto talk about certain facets of what has been found inthe last decade concerning sexual processes ofbacteria. In order to get a feel for this it is importantto have some idea of what the relative sizes are ofbacteria and of the mammalian cells, and also tohave a vague notion of some of the fundamentaldifferences between mammalian cells and bacteria.Probably the most useful index to bear in mind whenone is comparing these different organisms is theamount of genetic information which is carried intheir nucleic acid. Roughly speaking, a typicalbacterium contains sufficient DNA to specify 104genes; that is, it can code for about 10,000 differentenzymes or proteins. On the other hand, a mam-malian cell, again speaking very roughly, has about100 times as much nucleic acid; so it can code forsomething like a million different proteins or en-zymes. The other type of organism that we will bediscussing is the bacteriophage, the viruses thatgrow in bacteria. These are about 1% the size ofbacteria, again speaking very approximately; sothey contain about 100 genes. That is all we have toremember as far as size is concerned. The otheraspect that we can just touch on is the fundamentaldifferences between bacterial and mammalian cells.Obviously a good antibiotic has to be one which willkill a bacterium but yet have very little effect onmammalian cells themselves. Two components ofcells which differ markedly in the mammalian andbacterial cases are on the one hand the wall thatprotects the cell, and on the other the ribosomeswhich are the site for the synthesis of protein. Thesetend to be markedly different in the two cases andfor the most part the antibiotics which are medicallysuccessful are those which will attack either the cellwall or the protein synthesizing apparatus of thebacteria without interacting with these componentsin mammalian cells.

Turning to the types of sexual process which haveevolved over the years in bacteria, these are quitedifferent in kind from those which one is accustomedto think about in considering the genetics of higherorganisms. The essential difference is that bacteriaare haploid while the higher organisms are mostlydiploid. In other words bacteria contain only onecopy of genetic material within each cell whilemammalian cells contain two. This distinction has aprofound effect on the mechanisms of the geneticprocesses in the two cases. Allied to this distinctionis another which relates to the concept of sex in thetwo cases. In higher organisms one is accustomed tothink of sex as a process in which a complete set ofmale genes is transferred into a female cell andsubsequently, after a certain amount of juggling, anew organism arises with characteristics acquiredfrom each parent. With bacteria it is the exceptionthat complete genomes are transferred during sexualprocesses. Normally only a small part of the geneticmaterial in the donor (or male) cell is transferred intoa recipient (or female) cell. And the fascination ofbacterial genetics is the variety of methods by whichthis partial transfer is achieved. As far as the dis-cussion today is concerned we have to only considertwo of these, termed transduction and conjugation.

Transduction is a sexual process in which genesare transferred via the intermediary of a phageparticle. Normally when a phage particle attacks asensitive bacterium it acts as a parasite. It uses thebacterium as a factory for producing the proteinsand nucleic acids which are necessary to make newphage particles, but it diverts the direction of thesesyntheses into making phage DNA and phage coats,and these subsequently get put together to form in-fectious phage particles. The host cell is killed andfrom infection with one phage there is derived after afairly short period of J hr or so a hundredormorenewviral particles. Now sometimes, as the phage grows,instead of wrapping up in a phage head the piece ofphageDNA which as we said is about 100 genes long,it sometimes picks up a piece of the bacterial DNAwhich has by chance been broken down to about thesame size; so that rarely (and this happens only inabout one case in 10,000) a phage comes out whichhas a normal phage head but in which the DNA isnot phage DNA but is a small segment of the bac-terial genome. This fragment of bacterial DNA can

copyright. on June 9, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.48.558.216 on 1 April 1972. D

ownloaded from

Antibiotic resistance in bacteria

be any part of the bacterial genome. Amongst thelive phage that arise from infecting bacteria withphage there are therefore a few phage particles whichare not infectious, but contain a piece of bacterialDNA which incorporates something of the order of100 bacterial genes. Let us suppose, just as anexample, that we infect a bacterial population inwhich the bacteria are able to synthesize the aminoacid tryptophan. If we grow phage particles on thisstrain then we will pick up amongst the live phage theodd phage particle which contains the bacterialgenes which are responsible for the synthesis oftryptophan. If we now isolate the phage particlesvery carefully, get rid of all the host bacteria, anduse these phage particles to infect another bacterialculture which is unable to synthesize tryptophan,then in the odd cell which receives this defectivephage particle the recipient cell will now have theopportunity to acquire the ability to synthesizetryptophan by incorporating into its genome themissing genes which have been introduced into it viathe phage particle.

Transduction is a fairly specific process because,for the most part, phage particles are only able toattack one particular kind of bacterium, due to thespecificity of the adsorption step which is necessaryif the phage particle is ever going to be able toinject its DNA. For a long time I think it wasconsidered that transduction was a kind of labora-tory trick indulged in by the bacteria, because itwas not obvious that there was any real evolutionaryor medical significance in this type of genetic transfer.However, as we will see later, it does turn out that itis of profound importance in certain cases.The other type of genetic mechanism that goes on

in bacteria which we will have quite a lot to talkabout later is conjugation. This is a much moreanthropomorphic type of sexual process in whichthere is direct contact between two bacterial strains,and in which the genome of one strain, the donor, isactually transferred via a bridge between the twocells, into the second recipient strain. It turns outthat bacterial strains can be divided into two classestermed males and females. If you mix two malestrains together you get no transfer; if you mixtwo females together you get no transfer. It is onlywhen you mix a male and a female strain togetherthat you get genetic transfer and it is always in thedirection which we can arbitrarily call from male tofemale. It is appropriate that the person to first putforward the model by which we now interpret thisconjugation mechanism was Dr Hayes, who at thetime was a bacteriologist at Hammersmith Hospital.He postulated that the difference between the maleand the female strains arose because the male con-tained an extra piece of DNA, a sex factor, and thisextra piece of DNA was something quite distinct

from the normal bacterial genome. It was a piece ofDNA which now we think of as being about 100genes long (again about the size of the genes in atypical phage particle), and which is able to replicatein phase with the bacterial replication cycle. At everydivision, as well as the bacterial genome dividing, thesex factor also divides. It is quite easy to changemales into females, and vice versa. If a femaleacquires a sex factor it automatically becomes amale; if a male loses a sex factor it then becomes afemale. One property which is conferred on a malestrain by the genes in the sex factor is an alterationto the cell surface. It seems that there is a particularprotein structure, something like a sexual organ,which protrudes from the surface of the male strainand when the males get mixed with females it is thisorgan which joins up with the female and establishesa bridge through which nucleic acid can be trans-ferred. Normally the nucleic acid which is transferredis the sex factor itself, so the first effect of the matingis to turn females into males. The female acquires anew sex factor but the male does not become femalebecause an extra sex factor is synthesized in the malestrain, and it retains one and transfers a newlyformed one into the female. Sometimes as well as theDNA carried in the sex factor, bacterial genes arealso transferred. Then one gets an actual transfer ofthe genetic material from the genome of the maleinto the female. However, the type of transfer whichis most important from the point of view of bacterialdrug resistance is the transfer of the sex factor itself,and this takes place with a very high efficiency.The sex factor called F that was first identified byHayes is just one of a large family of what are nowcalled plasmids. These are extra chromosomal piecesofDNA usually about 100 genes long. Some of themare sex factors analogous to F, while some of themare not sex factors; but they all carry in them genesthat can confer new properties on the host bacteria.For instance some of them confer the property ofcolicin resistance, some confer new surface proper-ties on the bacteria and, as we will see later, some ofthem are involved in the acquisition of bacterial drugresistance. The characteristic properties of a plasmidare twofold. They are an extra-chromosomal pieceofDNA in the bacteria, and they are able to controltheir own replication.We come finally to the topic of this paper which

is bacterial resistance to antibiotics. If we are carry-ing out experiments in the laboratory and have astrain which is sensitive to streptomycin, then it isa fairly easy process to isolate from this strain amutant, which usually arises at a low frequency,which has become resistant to streptomycin. If oneexamines the genetic characteristics of the strepto-mycin-resistant mutant, it is found that an alterationhas occurred in a gene which in this case is part of

217

copyright. on June 9, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.48.558.216 on 1 April 1972. D

ownloaded from

N. Symonds

the bacterial genome. It is now known that thismutation is in a gene which has something to dowith the structure of the ribosomes, and strepto-mycin is one of those types of antibiotics which wetalked about earlier which kill the bacteria by inter-fering with the protein synthetic machinery of thebacterial cell and specifically the structure of theribosomes. It is not just streptomycin which has achromosomal location of this type, but a lot of theearly genetic work which used drug resistance asmarkers also managed to locate resistance to variousother antibiotics on the bacterial genome. However,in the field, rather than in the laboratory, it turnsout that genes conferring bacterial resistance to anti-biotics are, for the most part, not located on thebacterial genome at all but are extrachromosomaland located on plasmids.

I would like to discuss here two different cases inwhich this has now been substantiated and a greatdeal of evidence has accumulated over the past 10years. The first is drug resistance in Staphylococci.The great drug against Staphylococci is penicillin.The first signs of the occurrence of penicillin-resistant strains occurred in the late 1940s but theywere quite rare. In the 1950s however penicillin-resistance in Staphylococci became much morecommon and by the mid-1950s about 50Yoof theStaphylococci that were isolated in hospitals hadbecome resistant to penicillin. About that time otherdrugs were introduced to deal with Staphylococci. Inparticular streptomycin, tetracycline, chloram-phenicol, kanamycin and neomycin were used. Whatwas later found was that many of the Staphylococcithat were isolated in hospitals had become resistantto these drugs also and many strains were multiply-resistant to anything of up to four or five of thesedifferent antibiotics. In the 1960s with the introduc-tion of methicillin and cloxacillin the incidence ofthe multiples declined, but none-the-less penicillin-resistance stayed high, and even at present some-thing of the order of 50%o of Staphylococci isolatedin hospitals are resistant to penicillin.As far as the genetic analysis of staphylococcal-

resistance is concerned it has now been establishedthat this resistance is transferred from one cell toanother via transduction and that the resistance isconferred by a plasmid (which is not a sex-factor)but which contains the genes which confer on thecells resistance to various antibiotics. It turns outthat in Staphylococci there are a large number ofphage particles which are capable of acting as inter-mediaries in transduction. The plasmid, as we said,is a separate piece ofDNA which contains about 100genes, and this is about the size of DNA which cancomfortably be wrapped up inside a phage head.So the transduction process with these staphylo-coccal phages is one in which a complete plasmid is

able to be wrapped up inside a phage, transferredintact from one cell to the other, and finally conferson the recipient cell the ability to withstand theantibiotics. Although transduction is a relatively in-efficient process, none-the-less with the terrificselective advantage that the resistant strains haveover the sensitive ones as far as antibiotics are con-cerned, it has been instrumental in causing thisdramatic degree of multiple antibiotic resistance inStaphylococci which has now spread virtuallythroughout the world.The other mechanism for the transfer of antibiotic

resistance which I want to discuss and which is moregeneral than that with the Staphylococci (which is atransfer system confined to this particular type ofbacteria because of the specificity of the transducingphages), relates to the type of resistance which iscarried by what is now called a resistance transferfactor, or RTF. This RTF is again a plasmid about100 genes in size which contains amongst its genesthose which confer resistance to various antibioticson the host bacteria. In contrast to the situation withthe Staphylococci these types of plasmids are alsosex factors. The study of these RTFs has taken placemostly in Gram-negative strains. Historically, therealization that in these cases antibiotic resistancewas being transmitted by a plasmid came fromstudies in Japan in the 1950s. After the war in Japanthere were a large number of epidemics of dysenteryand these were treated with the antibiotics whichwere available at the time. It turned out that by themid 1950s the Shigella strains that were being iso-lated in Japan from these dysentery outbreaks hadalready become resistant to the four antibiotics whichwere commonly being used to treat the disease;these were streptomycin, chloramphenicol, tetra-cycline and sulphonamide. This multiple resistanceto these antibiotics grew in later years to includekanamycin, neomycin, erythromycin and ampicillinwhich came into play for treatment of the Gram-negative strains somewhere in the 1960s. At thepresent time, well over 504 of the Shigella andSalmonella strains which are turning up in hospitalsnot only in Japan, but in any country where they arelooked for, are now resistant to a large number ofthis array of antibiotics. I would like to present acertain amount of more specific data with regard tosome of these strains. The first is to do with the out-breaks of Shigella sonnei which have been looked atin London. Amongst these in the late 1950s therewas a small incidence of resistance to sulphonamide;something much less than 10%. In the late 1950sstrains were also turning up which were resistant totetracycline and streptomycin and during the 1960sresistance to ampicillin began to be observed. In1970 when the latest data are available, more than70%0 of the Shigella sonnei strains which are being

218

copyright. on June 9, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.48.558.216 on 1 April 1972. D

ownloaded from

Antibiotic resistance in bacteria 219

isolated in London hospitals are resistant to three ormore of the common antibiotics. The actual patternof the multiple resistance varies widely from oneisolate to the other as can be seen from Fig. 1, andin some cases it can be observed the bacteria havebecome resistant to up to seven different antibiotics.Much the same pattern of resistance can be tracedin two Salmonella strains, Salmonella panama andSalmonella typhimurium which have been looked atin Western Europe in the last decade. Their dramaticincrease in tetracycline resistance is noted in Fig. 2.Another extremely interesting case which has been

studied is that of type 29 of Salmonella typhimurium.It can be seen from Fig. 3 that up till 1963 thisparticular type of Salmonella was very rare both inanimals and in humans, and that the strains that wereisolated were almost entirely sensitive to antibiotics.Around 1963-64 there was an increase both in theincidence of type 29 and also in the fraction of themwhich were becoming resistant to antibiotics, until inthe late 1960s the incidence of this type reachedabout 50°/ of all Salmonella isolates, and almost all

A S T C K Su ST SuA S T K Su STCASTC Su SSuS T C K Su ATASTSu AKA S K Su S SuS T C Su TKS T K Su T

FIG. 1. Patterns of transmissible drug-resistance. A,resistance to ampicillin; S, resistance to streptomycin;T, resistance to tetracyclines; C, resistance to chlor-amphenicol; K, resistance to kanamycin/neomycin;Su, resistance to sulphonamides.

40-

O

30

U)

-C 030U.

U_

0

1959 1960 1961 1962 1963 1964FIG. 2. Incidence of drug-resistance amongst isolates ofS. typhimurium (open columns) and S. panama (diagon-ally hatched columns) in the Netherlands (MANTEN,GUINEE & KAMPELMACHER (1966) Zentralblattfur Bak-teriologie, Infektionskrankheiten und Hygiene, 200, 13).

of them had acquired multiple resistance to anti-biotics. (This particular set of data was that whichled to the Swann report on the use of antibiotics infoodstuffs for animals.) In all these cases one cansee the same trend, this really fantastic and fright-ening increase in the percentage of the bacterialstrains which are resistant to this battery of thecommon antibiotics.Now in all these latter cases that we have been

discussing the resistance to the antibiotics is carriedon a plasmid and this plasmid is transferred fromresistant to sensitive strains via conjugation. Theplasmid in this case is a sex factor. The resistantstrain conjugates with non-resistant strains, andduring conjugation transfers the plasmid and soconfers bacterial resistance to the sensitive strain.This strain in turn will act as a donor to other sensi-tive strains. So one has the situation where there is acontinual increase in the fraction of strains which areresistant to the antibiotics. With the RTFs the situa-tion is muchmore serious than with the Staphylococcibecause with these RTFs there is nowhere near thespecificity that exists in that case due to the limitedhost range of the transducing phage. It turns out thatRTFs can be transferred very widely between quitedifferent types of bacteria and it has now beenestablished that resistance to antibiotics which isconferred by RTFs very similar to those in Shigellaand Salmonella are now common in other Entero-bacteriacae such as Vibrio, Pseudomonas andProteus. Also amongst urinary infections at present20% of isolates are carrying plasmids, as are Gono-cocci, Pneumococci and haemolytic Streptomyces.The danger seems to be that in the normal flora

of the bowel, particularly amongst the E. coil, thereis a very high incidence of bacteria which are carryingRTFs. There can then occur a very efficient transferfrom these resident bacteria to infecting organismswhich may well start as sensitive strains, but will veryquickly due to the efficient transfer of the RTF bealtered to resistant strains inside the body. So muchthen for the medical problem which obviously is onewhich has reached quite staggering proportions overthe last ten years.

In conclusion, I would like to talk very brieflyabout three other aspects of this type of plasmid drugresistance. The first is the mechanism of the resis-tance; just why are strains which carry these plasmidsresistant to so many antibiotics? Secondly, whattheories can one put forward to explain the originof this type of plasmid resistance? Thirdly whatpossibilities are there for treating drug resistance ofthis type? With regard to the mechanism of resistanceit was first thought that all the antibiotic resistancescarried on the plasmids were conferred by geneswhich changed the properties of the cell wall in sucha way that the antibiotic could not enter the cell. It

copyright. on June 9, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.48.558.216 on 1 April 1972. D

ownloaded from

220 N. Symonds

50 -

tE~~~~~F~ASTNKSuFu40

* Sensitive strains

a Lin ~Drug - resistant strainsE . Appearance of respective

(ma an anials

30 resist ance patterns

ASTSuFu20 STSu

E0-0 cc '-c)

0 -- SS._u

0Year:- 1961 1962 1963 1964 1965 1966

Total 4090 3523 3689 3859 5131 3540S. typhimurium(man and animals)FIG. 3. Incidence of type 29 as a percentage of total S. typhimurium from 1961 to66 (from ANDERSON, E.S. et al (1968) British Medical Journal, iii, 333 with kindpermission of the Editor and the author).

was further considered that a single gene could inducea change in the cell wall which would render it im-permeable to a large number of antibiotics. Morerecently, however, this has been found not to be thecase, and as far as the RTFs are concerned it seemsthat for every drug resistance there is a specific geneon the plasmid which confers resistance. What isknown at the present time about mechanisms of thisresistance is indicated in Table 1. Resistance topenicillin for example is conferred by cells generatinga particular type of enzyme, penicillinase, whichactively breaks down the penicillin. This is a veryinteresting system because it contains not one genebut many, and these are under a very strict mecha-nism of control. They are in fact an inducible systemof enzyme production, and so for this the plasmidhas to carry a number of genes which are co-ordinated in a very particular way. In the case ofstreptomycin it turns out that the resistance is quitedifferent from what it is in laboratory cases which we

TABLE 1. Cause of drug resistance in bacteria conferredby RTF.

Antibiotic Mechanism of resistance

Penicillin Production of penicillinaseStreptomycin Phosphorylation or adenylationTetracycline No concentration of drug inside

cellsChloramphenicol AcetylationKana/neomycin Acetylation or phosphorylation

talked about before, where the resistance is conferredby a change in the structure of the ribosomes them-selves. As far as the plasmid resistance to strepto-mycin is concerned, this comes about by addingextra groups to the streptomycin as it comes into thecell, either phosphorylating or adenylating them, andthis has the effect of rendering the streptomycin in-active. Tetracycline is another interesting situation inwhich sensitive bacteria normally concentrate thedrug inside the cells and in order for the tetracyclineto be active one has to build up high local concentra-tions of the drug. In resistant cells this concentrationof the antibiotic does not take place, it seems that thetetracycline enters the cells and comes out againwithout doing any observable damage. Chloram-phenicol, kanamycin and neomycin act in a similarway to streptomycin in that the plasmid causes theaddition of extra groups to the drug. In the case ofchloramphenicol it is acetylation while with kana-mycin and neomycin it can be acetylation or phos-phorylation.As far as the origin of the particular gene is con-

cerned which is eventually picked up by the plasmidsand then confers resistance to each antibiotic themost pleasing theory is that (as nearly all these anti-biotics are in fact naturally occurring substances)situations have arisen through the years in whichresistance to the naturally occurring substances hasevolved. Then by chance these genes have beenpicked up on plasmids and due to the fantastic

copyright. on June 9, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.48.558.216 on 1 April 1972. D

ownloaded from

Antibiotic resistance in bacteria

selection pressure which has applied over the past 20years these genes have spread throughout the world.In this respect there is an intriguing story abouthedgehogs in New Zealand. In a remote part of thecountry in which few people have ever lived andwhere there has been no history of the use of anti-biotics it has been found that hedgehogs for quite along time have been suffering from ringworm, adisease already described in biblical times. Thisringworm, which is a fungal disease, gives off aparticular type of substance which is similar in manyrespects to penicillin. In the sores which are causedby the growth of the ringworm the skin of thehedgehogs has been invaded by bacteria. In orderto survive in the presence of the penicillin-like sub-stance these bacteria have acquired a type ofresistance which is very similar to that now recog-nized in penicillin-resistant strains. It involves theability to make penicillinase and also the rathersubtle control mechanisms which are necessary insetting up an inducible enzyme system. So in thiscase at any rate it looks as if penicillin-resistantgenes had already developed in the prehistory of theantibiotics; and that the problem which still remainsto be solved is how these genes were ever picked upand spread by the RTF. And one can imagine thatsimilar situations have arisen with all the naturallyoccurring antibiotics. The only one which is indeeda purely synthetic drug is sulphonamide where onewould have to make some special kind of argument.As far as possible treatments of this plasmid type

of antibiotic resistance are concerned, there are two

approaches which are being tried. One is to cut downthe use of the antibiotics themselves especially withregard to incidental uses such as for animals, inparticular in their foodstuffs. It is thought that thisprobably will not make such a great deal of differenceas the incidence of the plasmids is already so wide-spread and it is obviously necessary in medical casesto use antibiotics for people who are acutely ill.So the ability to cut down very much on the spreador on the incidence of this type of antibiotic resis-tance is not very great. A better hope as far astreatment is concerned is to try and utilize theknowledge that has been gained in the laboratoryover the last 10 or 15 years with regard to theproperties of the plasmids and how they can beeradicated in certain cases. We do know in the caseof the sex factor F that it is possible to cure malestrains of the sex factor by treating them with certainchemicals which react with DNA, particularly theacridines. Unfortunately, this treatment is atpresent only successful with a limited number ofplasmids, and even then only in laboratory condi-tions; in medical situations they have little effect.In the long term however the most hopeful approachdoes seem to be an attempt to identify chemicalswhich will eradicate plasmids from bacteria whilebeing safe to use in medical situations.

/

AcknowledgmentI would like to thank Dr Naomi Datta for her patience

while introducing me to the essentials of the topic of this talk.

221

copyright. on June 9, 2020 by guest. P

rotected byhttp://pm

j.bmj.com

/P

ostgrad Med J: first published as 10.1136/pgm

j.48.558.216 on 1 April 1972. D

ownloaded from