Review articles

Population structure and evolutionary dynamics of pathogenic bacteria John Maynard Smith,1* Edward J. Feil,2 and Noel H. Smith3 Summary Evidence concerning the significance of recombination within natural bacterial populations has historically come from two main sources: multilocus enzyme electrophoresis (MLEE) and nucleotide sequence data. Here we discuss evidence from a third method, multilocus sequence typing (MLST), which is a development of MLEE based on nucleotide sequencing that combines the advantages of both approaches. MLST has confirmed both the existence of clones and the high rates of recombination for several bacterial pathogens. The data are consistent with ``epidemic'' population structures, where clones are superimposed upon a backdrop of frequent recombination, thus, in the short term, resisting the homogenising effect of recombination. The nature of the selective advantage of clones, however, and how this advantage relates to virulence are unclear. The current evidence also has broader implications concerning bacterial species definition, the management of antibiotic-resistant bacteria and the assessment of the dangers of releasing genetically modified organisms into the environment. BioEssays 22:1115±1122, 2000. ß 2000 John Wiley & Sons, Inc. Introduction Until recently, bacteria were low on the list of organisms that evolutionary biologists considered: no wings or beaks, no sexual selection or social behaviour, and little phylogeny or fossil record. Two things have changed all that. The first is the evolution of antibiotic resistance; economically this is the most important evolutionary change that we have been able to study as it happened. The second is the advent of molecular data on a vast scale. The molecular data have become available in three overlapping waves. (1) In multilocus enzyme electrophoresis


School of Biological Sciences, University of Sussex, UK. 2 Wellcome Trust Centre for the Epidemiology of Infectious Disease, University of Oxford, UK. 3 35 Newton Street, Canterbury, UK. *Correspondence to: Professor John Maynard Smith, School of Biological Sciences, University of Sussex, Falmer, Brighton, BN1 9QG, UK. Abbreviations: MLEE, multilocus enzyme electrophoresis; MLST, multilocus sequence typing; ET, electrophoretic type; ST, sequence type

BioEssays 22:1115±1122, ß 2000 John Wiley & Sons, Inc.

(MLEE), variation within housekeeping gene loci is detected indirectly through differences in the electrophoretic mobility of their products. (2) The generation of nucleotide sequence data, initially of samples of up to 50 related isolates, sequenced with a degree of accuracy, made a study of population variation possible. (3) In ``multilocus sequence typing'' (MLST), typically, seven housekeeping gene fragments of approximately 500 nucleotides have been sequenced from several hundred isolates. The questions that can be answered differ for the various methods. In this essay, we attempt a brief historical review. We will illustrate our account primarily with data from the Neisseria, a genus that includes the agents responsible for gonorrhoea and bacterial meningitis. A similar picture could be given for other pathogens, for example Streptococcus. We suspect, however, that a very different picture will emerge when comparable data become available for free-living bacteria such as Bacillus. A central question that can be addressed concerns the role of genetic recombination in bacterial populations. Although the occurrence of recombination in bacteria was demonstrated experimentally over 50 years ago,(1) there is still uncertainty about its role in natural populations. Horizontal gene transfer is of obvious practical relevance to the epidemiology of pathogenic bacteria, and will become increasingly important as genetically engineered bacteria are more widely used. A second question that is beginning to be addressed is the relationship between pathogenic strains of bacteria and the often much more abundant populations of related bacteria present in humans that cause no symptoms. Genetic comparisons of these two populations will help to explain the origins of new pathogenic strains and to identify the genetic differences involved. Electrophoresis and the clonal paradigm Caugant et al.(2) published electrophoretic data for 15 loci from 688 isolates of Neisseria meningitidis. They found an average of seven distinguishable alleles per locus, and 331 distinguishable electrophoretic types (ETs). Despite this variability, 61 ETs were found more than once, of which 19 were found in more than one continent and up to 15 years apart. This reoccurrence of the same ETs despite great variability was found in a number of studies involving several species. It confirmed the view that, despite the occasional transfer within genes

BioEssays 22.12


Review articles where recombinants are likely to confer a selective advantage (e.g. flagellin genes in Salmonella(3)), bacterial populations were predominantly clonal.(4,5) The first reasons for doubting this conclusion came from sequence data, described in the next section. A re-analysis of the electrophoretic data(6) was stimulated by a relatively small dataset for Neisseria gonorrhoeae,(7) of nine loci from 327 isolates, collected worldwide. These data showed random assortment of alleles, or linkage equlibrium, where the presence of an allele at one locus is independent of the presence or absence of alleles at other loci. For example, the commonest ET occurred 35 times, but this was not evidence for clonality since the number expected by chance (calculated by multiplying the frequencies of each of the alleles of this ET within the dataset) was 32.2. Although linkage equilibrium is difficult to explain other than by assuming high rates of recombination, the reverse is not true; linkage disequilibrium (the non-random association of alleles) may be observed in a sample of strains even if the population from which the sample was drawn is recombining at a high frequency. One possibility, notably absent in the gonococcus, is geographical structuring within the population. Linkage between alleles may result from the combination of two or more geographically isolated subpopulations, between which gene flow is limited, within a single sample, even if recombination on a geographically localised scale, or within each subpopulation, is frequent. Souza et al.(8) reported evidence of geographical structuring within the nitrogen-fixing bacterium Rhizobium leguminosarum biovar phaseoli which is associated with wild and cultivated beans. By comparing the extents of linkage between alleles from samples over a spatial scale that ranged from the Western Hemisphere down to the level of an individual host plant, they noted that the degree of linkage (and allelic diversity) increased with increasing spatial scale. From this they inferred low rates of migration and significant geographical structuring. In contrast, in other species such as E. coli, rates of migration are very high, and a large proportion of the global diversity can be detected locally.(9,10) In other non-pathogenic species, such as Bacillus subtilis, efficient dispersal mechanisms such as spore formation ensure high rates of migration and help to maintain a single common gene pool. Geographical structuring also appears to be limited in N. gonorrhoeae and other pathogenic species such as N. meningitidis and S. pneumoniae. Large nucleotide datasets for these latter two species, which are discussed later, appear to be highly inclusive; most allelic exchanges involve imported alleles also present in unrelated strains in the dataset. As these data have been drawn from global samples, this is good evidence for high rates of migration. For example, an allele first noted in a Dutch strain may be found to have replaced by recombination an allele in a strain from, say, the Gambia. In the case of most pathogenic species, the lack of geographic structuring and the


BioEssays 22.12

efficient dispersal of the bacteria probably owes more to the high rate of migration of the human population than to longrange bacterial transmission between hosts. Although there is good evidence for high rates of recombination in many bacteria, the rates of recombination probably vary widely between different bacterial populations. A useful statistical test to provide a benchmark for degree of clonality within a population is the ``index of association'',(11) a multilocus measure of linkage disequilibrium; a value of the index significantly greater than zero indicates that recombination has taken place. However, the index is less effective in comparing the frequency of recombination in different bacteria, because its expected value for a given frequency of recombination depends on the number of loci analysed. A measure that avoids this drawback has been suggested by Burt.(12) A second difficulty lies in distinguishing the causes of disequilibrium. It may be that recombination is rare or absent or, as discussed above, that there is geographical structuring within the population. Alternatively, it may be that a population has rather frequent recombination, but that occasionally a genotype arises that is strongly favoured by selection. It will then increase rapidly in frequency, generating disequilibrium until recombination has had time to randomise the genetic background. Such ``epidemic structure was suggested by the electrophoretic data for N. meningitidis, and can be demonstrated more convincingly using MLST data as described below (Fig. 1). Gene sequences Evidence for the role of recombination in bacterial evolution came first from sequence data. In bacteria, the ``sexual'' process leads to the replacement of a region of the chromosome by homologous recombination from a donor cell, either by transduction (essentially accidental), conjugation (mediated by a plasmid) or transformation (an apparently adaptive mechanism present only in some bacteria). If the parental molecules are sufficiently different, such events can be detected from sequence data. If two sequences are compared, regions of close similarity are interspersed with divergent regions. Such mosaic structure was first observed, and interpreted as arising from recombination, in E. coli.(13,14) The phenomenon is most clearly seen, however, in the evolution of antibiotic resistance, because the events are recent and so have not been obscured by subsequent mutation, and because donor and recipient may differ by over 20% of nucleotides. Penicillin kills bacteria by inactivating penicillin-binding proteins (PBPs). These are enzymes needed to synthesise the cell wall. Many bacteria become resistant by acquiring an enzyme that breaks down penicillin. This enzyme is carried on a plasmid, and has been transferred to many taxonomically diverse bacteria. In contrast, Streptococcus and Neisseria, two species that are competent for transformation, have evolved resistance by altering their PBPs

Review articles

Figure 1. The ``epidemic'' bacterial population structure. Such a population is composed of two parts. Firstly, the background population is composed of a large number of relatively rare and unrelated genotypes (small circles) that are recombining at a high frequency. The relationships between these genotypes is most accurately represented as a network, rather than a bifurcating tree, as recombination has overwhelmed the phylogenetic signal. Superimposed upon this background is a limited number of very frequent genotypes, or clusters of closely related genotypes, illustrated as cones. These are clonal complexes, and typically emerge from a single, highly adaptive, ancestral genotype (the large circles). These clones rise to an observable frequency within the population and persist at least for decades. The diversification of these clones is predominantly driven by recombination but also by mutation (indicated by the arrows).

by transformation so that penicillin no longer binds to them. Figure 2 shows part of the history of resistance in Neisseria. Before 1950, both the gonococcus and meningococcus were sensitive to penicillin. They have become resistant by acquiring short DNA regions from commensal Neisseria species,

which, presumably for accidental reasons, were resistant before 1950. The analysis of mosaic structure, however, is possible only if recombination events have not been frequent enough to scramble the different gene sequences within the population, and have occurred between very different strains. In the case of the PBP loci, because recombinants conferring resistance to a cell will be selectively favoured, they can be detected even at a very low frequency. This is less likely to be true of recombinants in housekeeping genes, such as those encoding essential metabolic enzymes, where the variation is more likely to be selectively neutral. Analysis of these genes therefore provides a more realistic impression of the effect of recombination on the genome as a whole. If recombination within a population is sufficiently frequent, it will break up any mosaic structure, and there will be no localised runs of diverged sequence. One method for detecting frequent within-population recombination is to construct a phylogenetic tree for a set of sequences, ideally of housekeeping genes, and to count the number of apparent ``homoplasies''Ði.e., sites that have changed more than once. This can be compared with the estimated number of homoplasies expected if descent is truly clonal, or if recombination has been so frequent as to generate linkage equilibrium. The observed number lies between these extremes, and provides a comparative measure of recombination rate.(15) It has been found that bacteria vary over the whole range, from the strictly clonal (e.g. Borellia, the causative agent of lime disease(16)) to the effectively panmictic (e.g. Helicobacter, the cause of stomach ulcers(17)), with most species lying somewhere in between. In these ``weakly clonal'' species, recombination will dominate the long-term evolution of the population, but will not prevent the transient emergence of widespread clones.

Figure 2. Mosaic structure of the penA gene in Neisseria. A: penicillin-sensitive N. meningitidis; B: penicillin-resistant commensal, N. flavescens; C: susceptible commensal, N. lactamica; D-H: resistant (recombinant) strains of N. meningitidis; I: resistant (recombinant) strain of the commensal species, N. lactamica. Unshaded regions are similar to susceptible strains of N. meningtidis. Shaded regions, indicating the proposed origin of different block, differ from susceptible N. meningitidis by between 10% (most of N. lactamica; light-shaded) and 23% (N. flavescens; darkshaded) of nucleotides.

BioEssays 22.12


Review articles More light is shed on the structure of an evolving population if sequences are available for several gene loci. Dykhuizen and Green(18) pointed out that phylogenetic trees derived from different genes should be congruent if reproduction is strictly clonal, but different if there is recombination. They found that, in Borellia, the trees for different genes were almost identical. It is comforting that the same conclusion, of clonality, is indicated both by the homoplasy test and the comparison of phylogenetic trees. The history of disease outbreaks A combination of gene sequencing and MLEE has made it possible to follow the history of disease outbreaks (reviewed by Caugant(19) and Achtman(20)). Disease-causing meningococci fall into three main serogroups, A, B and C, as defined by the antigenicity of their polysaccharide capsules. These serogroups are characterised by differing patterns of disease. Serogroup A strains are responsible for major epidemics, particularly in China and Africa, but have been absent from Europe and North America since the second world war. Serogroups B and C meningococci cause endemic and hyperendemic disease in Europe, but particular hypervirulent clones are long lasting and are globally distributed. At least 10 other serogroups have been identified, but these are believed to be mostly harmless. Electrophoretic studies have revealed a highly variable population, most of whose member genotypes rarely cause disease, although they may occasionally do so, perhaps because their host is immuno-compromised or is infected by a second microorganism. In contrast, maybe 80% of disease world-wide is caused by no more than 10 discrete groups of closely related bacteria, or clonal complexes, belonging to serogroups A, B and C. These particularly virulent strains spread rapidly, possibly because of their improved ability to infect new hosts. However, we might expect that these virulent clones will become genetically more variable through recombination, and may become less virulent over time because of immunological change in the host population. We will illustrate this picture with two examples. In the 1980s, a pandemic caused by a highly virulent clonal complex, serogroup A subgroup III, began in China and Nepal. In 1987, 7000 cases caused by this subgroup occurred among pilgrims to the annual Haj in Mecca. Returning pilgrims carried the infection to many parts of the world. Epidemics have subsequently occurred in a number of African countries, most recently in West Africa in 1997; subgroup III meningococci had not been reported from any African country prior to 1988. Although transmission from the Far East, via Mecca, to Africa seems certain, the genetic constitution of the bacteria differs. Sequence studies reveal at least three independent differences between Chinese and African isolates, involving two separate horizontal transfers. Isolates from the Mecca outbreak show that most resembled the later African type,


BioEssays 22.12

but that some ancestral types were present, and also at least one intermediate type, carrying only one of the imported regions. Turning to serogroup B meningococci, many recent cases of meningitis caused by this serogroup have been associated with a particular electrophoretic type, ET5. This MLEE genotype was first noted in Norway in 1969 and caused a hyperendemic wave in that country in 1975. ET5 has since spread to much of Western Europe, Cuba, North and South America, South Africa, Morocco and Israel. The earliest isolates were very uniform: almost all had exactly the ET5 genotype, but by the 1990s one third of the isolates, although clearly closely related, differ at one or more loci from this ancestral ET5 genotype. There have also been serological changes, and changes in proteins likely to provoke immunological attack (including a porin, porA; a transferrin-binding protein, tbpB; and pilQ, a gene associated with pilus biogenesis). Thus, an originally uniform strain has become increasingly variable, both in housekeeping genes and, in particular, at immunologically relevant loci. Many puzzles remain. We do not know the nature of the genetic changes responsible for increased virulence. The repeated recovery of strains with identical, or very closely related, genotypes from cases of disease (such as the ET5 cluster) suggests that these changes are associated with increased fitness or, more specifically, transmissibility. In the absence of a selective advantage, it is difficult to explain how these widespread clones could emerge and persist over a matter of decades before being broken down by recombination. However, the cause of the association between virulence and fitness is unclear. Most meningococci are found in the nasopharynx, where they are transmissible but not disease causing. Once in the bloodstream or cerebrospinal fluid, they are potentially lethal, but unlikely to infect a new host. Presumably, virulence genes are pleiotropic, increasing both transmissibility and the ability to cause invasive disease. A serious difficulty arises because most of our information concerns bacteria recovered from patients; we know far less of the carriage population, harmlessly inhabiting people's throats. It is as if we tried to understand human evolution by analysing the genetic constitution of the inhabitants of death row in an American penitentiary: an atypical sample of the population, and one with little reproductive future. Fortunately, this is already being addressed as data are being collected on carriage populations of Neisseria and other bacteria. The species problem If, as was long supposed, bacteria reproduce clonally, there would be little reason to expect isolates to fall neatly into species. Such a system might work for Salmonella, which appears to be genuinely clonal and to inhabit different host species, although the difficulties inherent in microbial species definition are well illustrated even for this genus. A little over

Review articles 10 years ago, the nomenclature of the Salmonellae was rationalised, leading to a reduction in the number of recognised ``species'' from over two thousand to just twoÐS. enterica and S. bongori.(21,22) The new nomenclature is cumbersome, S. thyphimurium becomes S. enterica subsp. enterica serovar Typhimurium. For human Neisseria species, which are highly recombinogenic and live commensally in the nasopharynx, classification is even more difficult. These bacteria have been classified according to the Linnaean binomial system: for example, we have Neisseria meningitidis, N. mucosa, N. lactamica and so on. The discovery that genetic exchange is widespread might at first suggest that isolates could be grouped into ``biological species'', freely exchanging genes within species but genetically isolated from other species. Because there is genetic exchange between the most distantly related Neisseria, this hypothesis is not valid. It would be uninformative, however, to place all the Neisseria into one species, because there is substantial genetic structure within the taxon. A study of the genetic and phenotypic variation in a taxon such as Neisseria should be compulsory for all philosophers who believe in the existence of natural kinds, for all cladists who believe in the universal validity of phyolgenetic classification, and for all pheneticists, whatever they believe. In the end, we are forced to adopt a pragmatic approach, and view the Neisseria genus as a kind of commonwealth of phenetic and genetic clusters, each in turn partially characteristic, but also sharing some common identity with other clusters through horizontal gene transfer. Phenetic classification, based mainly on the biochemical properties, has placed the Neisseria into varying numbers of species. The genus is characterised by the colonisation of mucous membranes, primarily of mammals. The primary habitat of those species that colonise humans is the nasopharynx, the one notable exception being N. gonorrhoeae, which normally resides in the urinogenital tract. Barret and Sneath(23) concluded that few of the human commensal species can be well defined phenotypically. In an attempt to bring order out of chaos, Smith et al.(24) sequenced three housekeeping genes and the short subunit of ribosomal RNA, from 30 isolates belonging to seven commensal species and the two pathogens. They were able to place the isolates into five ``groups'': for each of the four genes, the average diversity within a group was smaller than between-group diversity, and bootstrap values were significant. This might suggest that five ``species'' have been identified, but sadly this is not so, for several reasons. First, the classification bears little similarity to a grouping suggested by Barret & Sneath on phenetic grounds. Second, there is genetic structure within groups; sometimes a set of isolates similar at one locus are very different at another. Third, and most important, if phylogenetic trees are constructed, a different tree is obtained for each gene.

What sense is to be made of these observations? Given that, with the exception of the gonococcus, the various strains have frequent opportunities for recombination, recombination between genetically similar strains is frequent, and, as shown for the pbp gene, recombination occurs occasionally between more distant strains, the picture is perhaps not surprising. The different phylogenetic trees for different genes, together with the low within-group diversity, implies that, occasionally, a whole gene is transferred from one group to another. It also implies that, after transfer, the locus is homogenised within the group, by selective spread of the new gene and by withingroup recombination. Because it is both reassuring and useful to be able to name isolates, there is a reluctance among bacteriologists to accept that there are no such entities as ``species'' in bacteria. A comparison with higher organisms may help. Obligatorily clonal organisms, for example, apomictic plants such as Alchemilla (Lady's mantle) and Taraxacum (Dandelions), belong to an indefinite number of clones, which cannot usefully be grouped into species. There is, therefore, no reason to expect bacteria that rarely or never recombine to fall into discrete kinds, unless, like Salmonella, they are parasitic in different hosts, in which case it may be possible to distinguish types living in different host species. Most higher organisms, however, are sexual. The members of a sexual species breed with other members of the same species, but not with others: this fact is both the reason why sexual organisms fall into distinct groups, and a criterion that can be used in classifying them. It is a criterion that obviously cannot be used for populations living in different places; it is a matter of taste whether geographically isolated populations with recognisable differences are regarded as species or varieties. Sometimes, particularly in plants, the criterion of interbreeding breaks down even for sympatric populations. In bacteria, it hardly seems applicable at all. Thus instead of falling into groups that either interbreed freely, or not at all, the frequency of recombination seems to fall off continuously with genetic distance. There is no discontinuity that can be used in species identification. Multilocus sequence typing In Neisseria(25) and Streptococcus(26) sequence data are now available for 450 bp fragments of seven housekeeping genes, from several hundred isolates. Comparative data will soon be available for Staphylococcus aureus, Haemophilus influenzae, Campylobacter jejuni and S. pyrogenes (http:// These data were collected primarily for the identification and surveillance of the ``types'' of bacteria responsible for specific disease outbreaks or with the ability to survive particular antibiotics. The purpose of the original meningococcal MLST dataset was essentially to validate the technique;(25) therefore all 106 strains used in this study had been previously typed by MLEE.

BioEssays 22.12


Review articles These strains were chosen to represent all of the major disease-causing clonal complexes within the meningococcal population, and very few of the isolates had been recovered from asymptomatic carriage. The total number of meningococcal strains typed using MLST now exceeds 1000. As discussed earlier, clonal complexes within freely recombining populations may arise through the acquisition of an adaptive mutation and a subsequent rise in frequency within the population. Alternatively, there is a more mundane explanation. Since only a small proportion of the meningococcal population is liable, for genetic reasons, to cause disease, it is also possible that the selective advantage of strains belonging to specific disease-causing clonal complexes may be largely limited to their ability to be represented in strain collections. This alternative cannot be discounted without more data from the carried population. Despite these difficulties, the presence of clonal complexes, somewhat ironically, can be used to confirm high rates of recombination. Ten meningococcal clonal complexes are resolved using MLST, showing a high level of congruence to the groups resolved previously using MLEE. Clonal complexes are composed of two parts, a ``consensus group'', i.e., a group of bacteria that are identical at all seven loci, and ``single locus variants'' (SLVs) which are identical to a consensus group at six loci but which differ at the seventh. Since it is highly probable that each SLV has arisen from a member of a consensus group by a single event, mutation or recombination, they provide direct evidence concerning the relative contributions of these two processes to clonal diversification. This then facilitates meaningful comparisons between species, and between subpopulations of the same species. The following conclusions can be drawn from an analysis of the meningococcal and pneumococcal datasets.(27,28) (i) New alleles are generated approximately five times more frequently by recombination than by mutation in the meningococcus (and approximately ten times more frequently by recombination than by mutation in the pneumococcus). (ii) Over 50% of the alleles arising through recombination in the meningococcal dataset are present elsewhere in the dataset; for the pneumococcal dataset this figure is over 80%. This suggests the datasets represent most of the common alleles in the wild. (iii) Recombinational exchanges, on average, involve the exchange of alleles that are approximately 4% diverged in the meningococcus and 1% diverged in the pneumococcus. The higher degree of divergence in the meningococcal population is almost certainly a consequence of the import into the meningococcal population of alleles from commensal Neisseria species.


BioEssays 22.12

(iv) In most cases, recombination replaces the whole gene fragment. This enables us to estimate the average size of the inserted piece as 5±10 kb in both the meningococcus and the pneumococcus. (v) At some loci, a comparison of the different alleles shows clear mosaic structure, suggesting the recent introduction of genetically distant DNA from other species (this is observed more commonly in the meningococcus than in the pneumococcus). Other loci, although genetically diverse, show no obvious mosaic structure, indicating that within-population recombination has randomised the sequences. The conclusion is that the meningococcal population is highly diverse, with frequent within-population recombination, and with its diversity being maintained by the occasional introduction of DNA from other species. It is not clear whether the situation within the meningococcal population is stable in the long term, or whether it is a temporary consequence of the recent increase in human numbers, geographic range and migration. Does the present diversity of the Neisseria resident in humans, both within and between the named ``species'', reflect the relatively recent invasion of humans by other Neisseria species? If so, why do we not observe a similar effect in the pneumococcus, which also cohabits the nasopharynx with numerous commensal streptococcal species? The answers to these questions are simply not known. The MLST data raise another question. Are the ``consensus groups'' merely the result of sampling bias, the stochastic increase in numbers of particular haplotypes, or do they require a selective explanation? In other words, is the existence of consensus groups evidence of an epidemic structure? To answer this question, it is helpful to know how old these groups are; i.e., when was the last common ancestor of the group. Obviously, the more recent the common ancestor, the more rapidly the group must have grown, and the more likely it is that selection has been responsible. The existence of penicillin resistance enables us to answer this question with some confidence. Prior to the clinical use of penicillin, all pneumococci were sensitive (penS), yet all members of some consensus groups are resistant (penR). The common ancestor of a penR group cannot be more ancient than 50 years. Calculation shows that penR groups could have reached their present numbers without selection in so short a time only if the effective population size of S. pneumoniae was absurdly small. Almost certainly, these groups have increased by selective advantage conferred by resistance to penicillin. What of the penS groups; could they be much older? The evidence is against this. If this were true, we would expect them to be surrounded by a cloud of isolates differing from them at one or two of the gene loci. In fact, there is little difference between the numbers of such neighbours of penS

Review articles and penR groups. So it seems likely that penS groups have also increased by selection. The following question is more difficult to answer. Does the high proportion of isolates in a sample of pathogens that fall into consensus groups imply that virulence is associated with some factor that increases fitness, or is it merely a consequence of the fact that relatively few genotypes are potentially virulent, and that a sample of pathogens contains only these genotypes. Although N. meningitidis, S. pneumoniae and Staphylococcus aureus are considered pathogenic species, typically they all exist as harmless commensals. In the UK, some 10% of the human population harbour meningococci asymptomatically and it seems clear that virulence itself confers no advantage to the meningococcus. In fact, the opposite is true; a bacterium in the cerebrospinal fluid has no chance of infecting another host. But it is possible that virulence is a pleiotropic effect of other selectively advantageous changes, and that these changes are responsible for the selective spread of virulent clones. Given data only for virulent isolates, it is hard to decide. We also need to know something about the carriage population from which they evolved. A recent study of Staphylococcus aureus by Dr Nick Day and his colleagues illustrates what may follow such an approach. They used MLST to characterise bacteria isolated from patients acquiring the disease in the community, and also strains isolated from the nasopharynx of healthy individuals in the same area, in Oxfordshire, between 1997 and 1998. The picture that is emerging from an analysis of these data is that occasionally a new genotype arises, by mutation or recombination, with relatively high transmissibility and pathogenicity. Because of the high transmissibility, such a genotype gives rise to a successful clone. Most of the pathogens sampled belonged to one of these recent clones. However, even in the early days of such a clone, it is likely that most such infections do not lead to disease. As time passes there is gradual loss of virulence and transmissibility, perhaps because of immunity within the host population, and the descendents of the original genotype become more variable as a result of recombination and mutation, finally merging into the background population. Such a picture of occasional ``selective bursts'' is of course not a new one. It is probably true of many bacterial populations, and corresponds to the ``epidemic structure'' originally suggested for the meningococcus to account for the combination of linkage disequilibrium with a high rate of recombination (Fig. 2). What is new is that the picture is being confirmed by a simultaneous study of carriage and pathogenic bacteria from the same region. Conclusions We are approaching the time when we will be able to give a general picture of the history of human diseases and their impact on human history. How ancient are such diseases as gonorrhoeae and meningitis? Where did they come from? To

what extent did their spread depend on an increase in human numbers or mobility? What has been the role of horizontal gene transfer? Such historical understanding will have obvious practical relevance today, in understanding the origin and spread of new pathogenic strains. This understanding will be hastened as data become available, not only for pathogenic strains, but also for carriage populations, which will allow an insight into the comparative population structures of virulent and asymptomatically carried strains as well as the genetic basis of virulence. Microbial population biology is of direct relevance to a number of increasingly urgent problems. Most notable of these are the management of pathogenic bacteria resistant to several antibiotics and an assessment of the dangers from releasing genetically modified organisms into the environment. Both problems concern the inherent difficulties in predicting the likelihood and outcome of horizontal genetic exchange in the wild. The available evidence for the pathogenic bacteria indicates that frequent recombination in the wild may be the norm rather than the exception, and the rapid spread of antibiotic resistance through diverse microbial genera vividly illustrates the caution required in assuming that genes introduced into bacteria will remain within the confines of the cell wall. Rapid advances in sequencing technology have begun to answer some of the fundamental questions concerning microbial population structures, the genetic basis of virulence, and the significance of recombination. MLST datasets in particular allow a quantitative estimate of the impact of recombination on clonal diversification. As more laboratories adopt the technique for routine typing, so the current datasets will grow, allowing a constant refinement of the estimates. As datasets become available for other species, it will be possible to draw meaningful comparisons between them. Only then will we be able to tease apart the biological or ecological factors that determine recombination rates in different microbial species, and begin to tackle the question that should be the Holy Grail for any microbial population biologistÐwhy do bacteria recombine? References

1. Ledeerberg J, Tatum EL. Gene recombination in Escherichia coli. Nature 1946;158:558. 2. Caugant DA, Mocca LF, Frasch CE, Froholm LO, Zollinger WD, Selander RK. Genetic structure of Neisseria meningitidis populations in relation to serogroup, serotype, and outer membrane protein pattern. J Bacteriol 1987;169:2781±2792. 3. Smith NH, Beltran P, Selander RK. Recombination of Salmonella phase 1 flagellin genes generates new serovars. J Bacteriol 1990;172:2209± 2216. 4. Selander RK, Levin BR. Genetic diversity and structure in Escherichia coli populations. Science 1980;210:545±547. 5. érskov F, érskov I. Summary of a workshop on the clone concept in the epidemiology, taxonomy, and evolution of the enterobacteriaceae and other bacteria. J Infect Dis 1983;346±357. 6. Maynard Smith J, Smith NH, O'Rourke M, Spratt BG. How clonal are bacteria? Proc Natl Acad Sci USA 1993;90:4384±4388.

BioEssays 22.12


Review articles

7. O'Rourke M, Stevens E. Genetic structure of Neisseria gonorrhoeae populations: a non-clonal pathogen. J Gen Microbiol 1993;139:2603± 2611. 8. Souza V, Ngyen TT, Hudsen RR, Pinero D, Lenski RE. Hierarchical analysis of linkage disequilibrium in Rhizobium populations: Evidence for sex? Proc Natl Acad Sci USA 1992;89:8389±8393. 9. Caugant DA, Levin BR, Selander RK. Genetic diversity and temporal variation in the E. coli population of a human host. Genetics 1981;98: 467±490. 10. Whittam TS, Ochman H, Selander RK. Geographic components of linkage disequilibrium in natural populations of Escherichia coli. Mol Biol Evol 1983;1:67±83. 11. Brown AHD, Feldman MW, Nevo E. Multilocus structure of natural populations of Hordeum spontaneum. Genetics 1980;96:523±526. 12. Burt A. Population genetics of human-pathogenic fungi. In Thompson RCA, editor. Molecular Epidemiology of Infectious Diseases. London: Edward Arnold, 2000. 13. Dykhuizen DE, Green L. DNA sequence variation, DNA phylogeny and recombination. Genetics 1986;113:s71. 14. Stoltzfus A, Leslie JF, Milkman R. Molecular evolution of the Escherichia coli chromosome. I. Analysis of structure and natural variation in a previously uncharacterized region between trp and tonB. Genetics 1988; 120:345±358. 15. Maynard Smith J, Smith NH. Detecting recombination from gene trees. Mol Biol Evol 1990;15:590±599. 16. Dykhuizen DE, Polin DS, Dunn JJ, Wilske B, Preac-Musric V, Dattwyler RJ, Luft BJ. Borrelia borgdorferi is clonal: implications for taxonomy and vaccine development. Proc Natl Acad Sci USA 1992;90:10162± 10167. 17. Suerbaum S, Maynard Smith J, Bapumia K, Morelli G, Smith NH, Kuntsmann E, Dyrek I, Achtman M. Free recombination within Helicobacter pylori. Proc Natl Acad Sci USA 1998;95:12619± 12624.


BioEssays 22.12

18. Dykhuizen DE, Green L. Recombination in Escherichia coli and the definition of biological species. J Bacteriol 1991;173:7257±7269. 19. Caugant DA. Population genetics and molecular epidemiology of Neisseria meningitidis. APMIS 1998;106:505±525. 20. Achtman M. Microevolution during epidemic spread of Neisseria meningitidis. Electrophoresis 1998;19:593±596. 21. Le Minor L, Popof MY. Antigenic formulas of the salmonella serovars. WHO Collaborating Centre for Reference and Research on Salmonella. Institut Pasteur, Paris. 1987. 22. Reeves PR, Evins GM, Heiba AA, Plikyatis BD, Farmer JJ. Clonal nature of Salmonella typhi and its genetic relatedness to other salmonellae as shown by multilocus enzyme electrophoresis, and proposal of Salmonella bongori comb. nov. J Clin Microbiol 1989;27:313±320. 23. Barrett SJ, Sneath PHA. A numerical phenotypic taxonomic study of the genus Neisseria. Microbiology 1994;140:2867±2891. 24. Smith NH, Holmes EC, Donovan GM, Carpenter GA, Spratt BG. Networks and groups within the genus Neisseria: analysis of Argf, Reca, Rho and 16s RNA sequences from human Neisseria species. Mol Biol Evol 1999;6:773±783. 25. Maiden MCJ, Bygraves JA, Feil E, Morelli G, Russell J, Urwin R, Zhang Q, Zhou J, Zurth K, Caugant DA, Feavers IM, Achtman M, Spratt BG. Multilocus sequence typing: a portable approach to the identification of clones within populations of pathogenic microorganisms. Proc Natl Acad Sci USA 1998;95:3140±3145. 26. Enright M, Spratt BG. A multilocus sequence typing scheme for Streptococcus pneumoniae: identification of clones associated with serious invasive disease. Microbiology 1998;144:3049±3060. 27. Feil EJ, Maiden MCJ, Achtman M, Spratt BG. The relative contributions of recombination and mutation to the divergence of clones of Neisseria meningitidis. Mol Biol Evol 1999;16:1496±1502. 28. Feil EJ, Maynard Smith J, Enright MC, Spratt BG. Estimating recombinational parameters in Streptococcus pneumoniae from multilocus sequence typing data. Genetics 2000;154:1439±1450.

Population structure and evolutionary dynamics of ...

sequence data, described in the next section. A re-analysis of the electrophoretic data(6) was stimulated by a relatively small dataset for Neisseria gonorrhoeae ...

363KB Sizes 0 Downloads 304 Views

Recommend Documents

Population structure and evolutionary dynamics of ...
them to be surrounded by a cloud of isolates differing from them at one or .... tional parameters in Streptococcus pneumoniae from multilocus se- quence typing ...

pdf-4\population-games-and-evolutionary-dynamics-economic ...
economic theorists. (Drew Fudenberg, Professor of Economics, Harvard University). About the Author. William H. Sandholm is Professor of Economics at the University of Wisconsin--Madison. Page 3 of 8. pdf-4\population-games-and-evolutionary-dynamics-e

Evolutionary imitative dynamics with population-varying ...
Mar 22, 2014 - Keywords: Learning; Imitation; Multiple Populations; Aspiration level; Wright manifold. JEL Classification Numbers: ...... MIT Press. GOLMAN, R. (2009): “Essays on Population Learning Dynamics and Boundedly Ra- tional Behavior,” Ph

Evolutionary game dynamics of controlled and ... - Squarespace
Jul 27, 2015 - evolution of automatic and controlled decision-making processes. We introduce a ..... where for the sake of illustration we fix a¼0.15 and vary b and q. ..... tion to, variation caused by the population).33,34 Our basic framework ...

Evolutionary game dynamics of controlled and ... - Squarespace
Jul 27, 2015 - simulations that support this suggestion.26 In these simula- ... pA for automatic agents and pC for controlled agents. ..... 365, 19–30 (2010).

Population Dynamics
and T ∈ R+ is the time period, we define the domain of the solution as. Q = Ω × [0,T]. ..... Our purpose is to find out, for segregated initial data, solutions of (43).

Effects of Population Size on Selection and Scalability in Evolutionary ...
scalability of a conventional multi-objective evolutionary algorithm ap- ... scale up poorly to high dimensional objective spaces [2], particularly dominance-.

Population dynamics and life cycle of the introduced ... - Springer Link
Oct 11, 2008 - Ó Springer Science+Business Media B.V. 2008. Abstract Marine introductions are a ... intensive. Microcosmus squamiger is a widespread.

Population dynamics and life cycle of the introduced ...
Centre d'Estudis Avançats de Blanes (CEAB, CSIC), .... in the centre, through which the dish was fastened to ... contact with ethanol they quickly contracted. We.

Complex networks: Structure and dynamics
Jan 10, 2006 - 255. 6.2. The Internet and the World Wide Web . ... Structure of the Internet . ..... Watts' pioneering book on the subject deals with the structure and the dynamics .... emerging, e.g., in mobile and wireless connected units.

Population genetic structure of two primary parasitoids of Spodoptera ...
subsequent comparison by molecular analysis. ..... meters using the ArcGIS 9.0 software (ESRI 2004). ... analysis was performed in R CRAN 2.6.2 (R Develop-.

Population genetic structure of two primary parasitoids of Spodoptera ...
subsequent comparison by molecular analysis. Genetic ..... (data not shown). Permutation tests ..... DNA haplotypes—application to human mitochondrial DNA.

PDF Online Structure and Dynamics
PDF Online Structure and Dynamics: An Atomic View of Materials. (Oxford Master Series in Physics) - Read Unlimited eBooks and. Audiobooks - By Martin T.

Investigation of structure and dynamics in the sodium metallocenes ...
Solid-state 23Na NMR spectra of two organometallic complexes, cyclopentadienylsodium (CpNa) and the tetrahydrofuran (THF) solvate of CpNa (CpNa·THF), are presented. Analytical simulations of experimental spectra and calculated 23Na electric-field gr

Dividend Dynamics and the Term Structure of Dividend Strips
Dividend Dynamics and the Term. Structure of Dividend Strips. FREDERICO BELO, PIERRE COLLIN-DUFRESNE, and ROBERT S. GOLDSTEIN∗. ABSTRACT. Many leading asset pricing models are specified so that the term structure of dividend volatility is either fl

Structure and Dynamics of Parallel b-Sheets ...
We study three systems (S1, S2, and S3) that differ in the initial staggering ... connecting loop, with V24–N27 being more extended in the S3 system. Additional ...

Dividend Dynamics and the Term Structure of Dividend ...
We thank the editor, Cam Harvey, as well as the associate editor and an ..... the per-year standard deviation of dividend growth across each horizon T for the two.

Data on the distribution, population structure and ...
io) (F : M) and biometric ... into the lagoons and river mouths [7, 9, 11, 12, 17, 13 ... in the Viluni Lagoon. Nr of females. Nr of males. Total n. 3. 4. 7. 17. 9. 2. 14. 17.

Impact of population age structure on Wolbachia ...
models, parameterized using field data, are essential for estimating thresholds (Turelli and ... Uninfected. U Bo. U B1. БББ U BxА1. U Bx. U p0. 0. БББ 0. 0. 0. U p1. БББ 0. 0 ..... Bockarie, M.J., Service, M.W., Barnish, G., Toure, Y.T., 1

Temporal structure of neuronal population oscillations ...
neural network, such as the organization and size of the neural network. To give .... in CA1 is more significant by comparison with the distribution of frequency .... 272–280. [4] G. Lantz, C.M. Michel, M. Seeck, O. Blanke, T. Landis, I. Rosen, Cli

Historical population structure of White Sturgeon ... - Wiley Online Library
Analysis of capture ⁄ recapture and sonic telemetry data identified spatial patterns in habitat use, with individuals showing a preference for one of four high-use ...

Population structure and local selection yield high ...
bits a significantly positive genomewide average for Tajima's D. This indicates allele frequencies .... In this way, migration can gen- .... calls to alternate) from (ii).