* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Patchy distribution of flexible genetic elements in bacterial
Transposable element wikipedia , lookup
Cre-Lox recombination wikipedia , lookup
Quantitative trait locus wikipedia , lookup
Therapeutic gene modulation wikipedia , lookup
Human genome wikipedia , lookup
Gene expression programming wikipedia , lookup
Heritability of IQ wikipedia , lookup
Genetic drift wikipedia , lookup
Metagenomics wikipedia , lookup
Polymorphism (biology) wikipedia , lookup
Minimal genome wikipedia , lookup
Non-coding DNA wikipedia , lookup
Vectors in gene therapy wikipedia , lookup
Koinophilia wikipedia , lookup
Public health genomics wikipedia , lookup
Designer baby wikipedia , lookup
Helitron (biology) wikipedia , lookup
Extrachromosomal DNA wikipedia , lookup
Genome editing wikipedia , lookup
Genomic library wikipedia , lookup
Genome (book) wikipedia , lookup
Pathogenomics wikipedia , lookup
Artificial gene synthesis wikipedia , lookup
Genetic engineering wikipedia , lookup
Human genetic variation wikipedia , lookup
Genome evolution wikipedia , lookup
Site-specific recombinase technology wikipedia , lookup
Population genetics wikipedia , lookup
No-SCAR (Scarless Cas9 Assisted Recombineering) Genome Editing wikipedia , lookup
MINIREVIEW Patchy distribution of £exible genetic elements in bacterial populations mediates robustness to environmental uncertainty Holger Heuer1, Zaid Abdo2 & Kornelia Smalla1 1 Federal Research Centre for Cultivated Plants, Julius Kühn-Institute (JKI), Braunschweig, Germany; and 2Departments of Mathematics and Statistics and Initiative of Bioinformatics and Evolutionary Studies, University of Idaho, Moscow, ID, USA Correspondence: Holger Heuer, JKI, Institute for Epidemiology and Pathogen Diagnostics, Messeweg 11-12, D-38104 Braunschweig, Germany. Tel.: 149 531 299 3831; fax: 149 531 299 3006; e-mail: [email protected] Received 13 December 2007; revised 23 May 2008; accepted 23 May 2008. First published online 8 July 2008. DOI:10.1111/j.1574-6941.2008.00539.x Editor: Jim Prosser Keywords flexible genome; bacterial population heterogeneity; broad host-range plasmids; evolvability; robustness; mobile genetic elements. Abstract The generation and maintenance of genetic variation seems to be a general ecological strategy of bacterial populations. Thereby they gain robustness to irregular environmental change, which is primarily the result of the dynamic evolution of biotic interactions. A benefit of maintaining population heterogeneity is that only a fraction of the population has to bear the cost of not (yet) beneficial deviation. On evolutionary time frames, an added value of the underlying mechanisms is evolvability, i.e. the heritable ability of an evolutionary lineage to generate and maintain genetic variants that are potentially adaptive in the course of evolution. Horizontal gene transfer is an important mechanism that can lead to differences between individuals within bacterial populations. Broad host-range plasmids foster this heterogeneity because they are typically present in only a fraction of the population and provide individual cells with genetic modules newly acquired from other populations or species. We postulate that the benefit of robustness on population level could balance the cost of transfer and replication functions that plasmids impose on their hosts. Consequently, mechanisms that make a subpopulation conducive to specific conjugative plasmids may have evolved, which could explain the persistence of even cryptic plasmids that do not encode any traits. Introduction One of the key findings of comparative genome analyses was the large difference between closely related strains. With each sequenced genome of eight strains of Streptococcus agalactiae, or seven strains of Escherichia coli, a significant number of new genes were found (Abby & Daubin, 2007). These results suggest that the pan genome, comprised of the different genes of all strains of a species, is typically several times larger than the core genome, which is shared by the majority of conspecific strains. One part of the core genome evolves by optimization of the intracellular networks for maintenance and replication (Fig. 1). The other part of the core genome, which forms the basis of the ecotype, is shaped by long-term correlations of environmental conditions in the primary habitats of the species (Fig. 1). The counterpart of the core genome is called the flexible genome (Hacker & Carniel, 2001). It is determined by the high plasticity and modularity of its genetic elements and a high rate of gene FEMS Microbiol Ecol 65 (2008) 361–371 acquisition and loss (Fig. 1). Comparative genomics of freeliving bacteria indicated that horizontal gene transfer is the primary source of intraspecies genomic diversity (Lerat et al., 2005). It seems that by transformation, transduction or conjugation most bacteria constantly acquire genes from a large available gene pool. However, the majority of the acquired genes do not persist within lineages over evolutionary times. Free-living bacteria in particular have a high incidence of horizontal gene transfer but relatively few mobile genetic elements in their genomes because, due to their large effective population sizes, selection efficiently removes acquired genes (Ochman & Davalos, 2006). Over evolutionary times, foreign DNA is more likely to be deleterious than to increase the fitness of the recipient and consequently become fixed in the lineage (Kurland, 2005; Hao & Golding, 2006). Over ecologically relevant time periods, the contribution of evolutionarily successful gene acquisition by the core genome to bacterial population fitness seems to be small. For E. coli K-12, it was estimated 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c 362 Fig. 1. Components of the population metagenome that differ in flexibility as a result of variation in continuity of the shaping selective forces. that only one transfer event in 400 000 years gets fixed in the lineage (Lawrence & Ochman, 1998). Nevertheless, mobile genetic elements, the agents of horizontal gene transfer, are abundant in bacterial populations of many habitats, which points to a major ecological function. Horizontally transferred genes can persist for a long time in a fraction of a bacterial population even if they are slightly deleterious (Novozhilov et al., 2005). Thereby, mobile genetic elements create standing genetic variation, by which adaptation to a rapid environmental change or habitat expansion can more likely be achieved rather than by new mutations (Jain et al., 2003; Hermisson & Pennings, 2005). In this minireview, we discuss whether the primary benefit of bacterial populations to maintain mobile genetic elements is to gain robustness to uncertain environmental change, and whether a secondary benefit is evolvability, i.e. the heritable ability of an evolutionary lineage to generate and maintain genetic variants that are potentially adaptive in the course of evolution. We summarize evidence for the generation of population heterogeneity by mobile genetic elements as a basis for robustness in fluctuating environments. We place special emphasis on conjugative broad host-range plasmids, as they are major vehicles of frequent interspecific gene transfer, which may be the basis for their evolutionary success. Broad host-range conjugative plasmids Plasmids are typically composed of (1) conserved backbone modules coding for replication, maintenance and transfer functions, and (2) variable accessory modules. By definition, they do not carry genes that are essential for growth of their host, except in special environmental situations. Most evidence points to plasmids being unable to survive as purely parasitic elements or by only temporarily providing beneficial traits to the cell (Bergstrom et al., 2000). Even if a 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c H. Heuer et al. plasmid carries beneficial accessory modules, these could easily be integrated into the chromosome to minimize the cost associated with plasmid backbone modules (Thomas, 2004). In the absence of sexual reproduction, a genetic element that lowers the fitness of its host and initially resides in only a fraction of the population will inevitably be lost by natural selection (Johnson & Brookfield, 2002). Plasmidbearing cells will be outcompeted by plasmid-free relatives if the cost of carrying the plasmid backbone is not balanced by a benefit for the cell. This problem is especially evident for broad host-range plasmids, because their genetic load may not be adaptive in the ecological niche of a new host after transfer, even if it was at some point beneficial for the previous one. Some of their accessory genes may be temporarily beneficial for many hosts in special environments. However, evidence from plasmid genome comparisons found that broad host-range IncP-1 plasmids even without any accessory genes existed in microbial communities over evolutionary times and occasionally acquired accessory elements at different insertion sites (Heuer et al., 2004). The recently sequenced IncP-1 plasmids pA1 and pBP136 are thought to resemble such a cryptic plasmid without accessory genes (Harada et al., 2006; Kamachi et al., 2006). However, each carries two ORFs of unknown function in addition to the backbone. The effects of these accessory genes and of the plasmid backbone on fitness of the host have not yet been investigated. The long-term existence of cryptic plasmids could be explained if bacterial populations actively maintain a type of plasmid backbone, regardless of accessory elements, thus taking advantage of their capacity to capture genetic modules, which could increase phenotypic diversity, and thereby the chance to respond to uncertain environmental change or to grasp an opportunity for transient niche expansion. Heterogeneous distribution of plasmids in bacterial populations The incidence of plasmids in bacterial isolates from diverse habitats, or in collections of strains, range from nearly 100% to o 10% (Sobecky, 1999; Cook et al., 2001; Malik et al., 2002; Thomas, 2004). In a survey of 521 diazotrophic isolates from four salt marsh grass rhizoplanes, the proportion of plasmid-containing strains was between 12% and 30% (Beeson et al., 2002). The analysed strains belonged to more than six genera, and 71 different plasmids, which ranged in size from 2 to 4 100 kbp, were detected in 134 of the strains. These data suggest that plasmids are abundant in many populations, but cannot determine estimates on the plasmid distribution within populations. An examination of 74 strains of the fish pathogen Photobacterium damselae ssp. piscicida isolated from amberjacks (Seriola spp.) in Japan found that all strains had several plasmids (Kijima-Tanaka FEMS Microbiol Ecol 65 (2008) 361–371 363 Population-level robustness through genome flexibility et al., 2007). Even though this locally emerging population showed only subtle genomic differences among strains, the distribution of plasmids was heterogeneous. There was no analysis of plasmid types and no discussion of the potential involvement of these plasmids in pathogenicity that could explain the high incidence in this population. Data on the frequency of a particular plasmid type, as defined by its backbone, and to what extent it confers phenotypic diversity by varying accessory elements in an environmental population of bacteria are scarce. A glimpse into the natural heterogeneity of soil populations due to conjugative plasmids was given by a recent study by Kuhn et al. (2008). The incidence and diversity of pSmeSM11a-like plasmids was analysed in isolates of Sinorhizobium meliloti from two field sites, representing two populations. Five of 21 strains isolated from one field, and four of 16 strains from the other field contained plasmids with pSmeSM11a-like backbone genes, while the majority of isolates from both populations did not carry this type of plasmid. The pSmeSM11a-like plasmids differed in size and accessory gene content. Thus, the isolates represent two natural field populations that are considerably heterogeneous with respect to their plasmids. Diversification of an introduced population by acquisition of naturally occurring conjugative plasmids was shown in a field experiment, using a genetically marked Pseudomonad as a recipient that colonized sugar beet plants (Lilley & Bailey, 1997b). At least four different large mercury-resistance plasmids were captured at frequencies between 10 6 and 1 for a few samples. One of the captured plasmids, the 330-kbp mercury-resistance plasmid pQBR103, was further studied in a greenhouse experiment. It conferred a transient fitness advantage to the plasmid-bearing cells compared with the plasmid-free subpopulation of Pseudomonas fluorescens SBW25 in the phytosphere of mature sugar beet plants (Lilley & Bailey, 1997a). The detoxification of mercury encoded by this plasmid could result in the coexistence of plasmid-free and plasmid-bearing cells at a certain range of mercury concentrations (Ellis et al., 2007). The higher the frequency of plasmid-bearing cells, the more the neighbouring plasmid-free cells benefit from detoxification without paying the cost of plasmid carriage. However, this kind of stabilizing frequency-dependent selection (Levin, 1988) cannot be a general explanation for the observed plasmid incidences in bacterial populations, but only in those special cases where a plasmid-encoded gene product alters the external environment in a way that plasmid-free cells benefit. Such gene products are occasionally encoded by accessory elements of plasmids, but occur chromosomally encoded as well. In conjugation experiments with close contact of plasmid-containing donor cells and plasmid-free recipient cells, only a fraction of cells receive a plasmid even under optimized conditions, i.e. in filter matings on nutrient media. In the natural environment, much lower transfer FEMS Microbiol Ecol 65 (2008) 361–371 rates are expected. In the rhizosphere of pea, which was shown to be a hot spot of plasmid transfer, the IncP-1 plasmid pKJK5 could transfer only to a tiny fraction of an introduced recipient strain or indigenous bacteria even if the potential recipients were heavily surrounded by plasmidcontaining cells (Mølbak et al., 2003, 2007). In contrast to this and many other studies, in a recent mating experiment the plasmid pKJK5 could completely invade plasmid-free populations of an E. coli and a Kluyvera sp. strain in bacterial mats without providing any obvious selective advantage (Bahl et al., 2007). This exceptional result could be due to the special conditions of the bacterial mat and also to the recipient strains applied. Often environmental strains are surprisingly reluctant to receive an IncP-1 plasmid in biparental mating (Heuer et al., 2007a), while conjugative transfer from a donor to cells of a bacterial community, e.g. from the rhizosphere or activated sludge, resulted in diverse transconjugants (Pukall et al., 1996; DeGelder et al., 2005). Selected recipient strains like Pseudomonas putida KT2442 consistently show high frequencies of plasmid capture. Pseudomonas putida KT2442 was originally isolated as the plasmid-bearing strain mt-2, which was then cured of the TOL plasmid but may have retained or even increased its conduciveness to plasmids through a long history of cultivation and mutation (Regenhardt et al., 2002). Recent evidence suggests the conduciveness for a plasmid is strain specific rather than species specific in P. putida (Heuer et al., 2007a), which was subsequently confirmed also for other species (DeGelder et al., 2007). In P. putida strain H2, IncP-1 plasmids were very unstable and conferred a high fitness cost. However, analysis of plasmid stability in eight lineages derived from single cells revealed high instability of the plasmid in only six of the lineages in batch culture (Heuer et al., 2007a). This population heterogeneity was still significant after 1000 generations of adaptive evolution, while plasmid stability and cost ameliorated. Large differences in the expression profiles of the plasmid between a conducive E. coli host and the unfavourable P. putida host H2 suggested considerable host–plasmid interactions, but could not give a mechanistic explanation for the differences in cost and stability of the plasmid. A specific mechanism of interaction between plasmid and chromosome was found in strain P. putida KT2440, which is known to be conducive for plasmids. Transcriptome analysis revealed that the chromosomal gene parI is positively regulated by the plasmid-encoded ParA and by ParI itself (Miyakoshi et al., 2007). This might support plasmid partitioning and maintenance in this strain. It remains to be investigated at which scale a population is heterogeneous with respect to conduciveness for plasmids. Seemingly, strains of a global population that define a particular species vary in that respect. But is this also true for strains of a population that occupy a single and contiguous spatial area, thereby providing a mechanism for 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c 364 sympatric diversification? Answering such a question requires a clear definition of a bacterial population, a task complicated by the ecological and genetic interactions between different bacterial strains. Population definitions differ, depending on the scale at which a study of interest is conducted. On the ecological scale, a population is defined to be a group of individuals of the same species within the same habitat at the same time that interact with one another (Lowe et al., 2004; Waples & Gaggiotti, 2006). On the evolutionary scale, populations are generally defined to be all individuals connected by gene flow (Hedrick, 1999; Lowe et al., 2004; Waples & Gaggiotti, 2006). These two definitions are coined, for the most part, for the study of sexually reproductive populations where a clear separation can be identified between ecological and evolutionary time scales. Bacterial populations do not adhere to such time-scale separation. Moreover, the asexual nature of bacteria further complicates the task of identifying an appropriate definition due to the level and mechanisms of gene flow that need to be considered before identifying an appropriate population concept. When plasmids are considered as an independent entity acting as a parasite, or mutualist, in association with a bacterial strain or species, they can be thought to form populations separate from their bacterial host populations (Bergstrom et al., 2000). In this case, the host bacterial population can be defined based on common ancestry, where strains close enough in their phylogenetic history are considered to be part of the same population. A challenging task is defining the meaning of ‘close enough’, which is usually based on an arbitrary cut-off imposed on the genetic distance between taxonomic units (Meyer & Paulay, 2005; Abdo & Golding, 2007). We favour a definition where a population includes strains that have similar genomic structure of the core genome as defined in Fig. 1, and that considers plasmids as a mechanism of gene flow between these strains. This definition restricts the boundaries of the population to include only those that share an evolutionary history yet allows for gene flow through plasmid conjugation and other mechanisms. Plasmids can be thought of as part of the overall gene content that contributes to the diversity of a population. Such structure facilitates the application of phylogenetic and population genetics modelling concepts on a higher evolutionary scale, i.e. the scale of genes rather than nucleotides (Hao & Golding, 2006), and facilitates the study and comparison of the diversity of microbes with respect to their plasmid content and other flexible genome differences. Population genetics modelling concepts allow for studying the effects of migration, selection, and spatial and temporal structure as part of the genome evolutionary process. To eliminate the use of arbitrary cut-off limits in identifying strain memberships to populations, we suggest the use of probabilistic, modelbased classification methods that provide a measure of 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c H. Heuer et al. confidence in the population membership of a strain (Abdo & Golding, 2007). Such methods quantify the uncertainty in grouping different strains to form a population based on their evolutionary history and provide objective decision rules for identifying population memberships of the different strains. Environmental data on incidence and diversity of plasmids within a population could be integrated into population genetics models to improve our understanding of the contribution of plasmids to standing genetic variation in stable or frequently perturbed environments. Coupling the use of these models with probability theory allows for evaluating the statistical significance of the effect of population heterogeneity, due to broad host-range plasmids, and facilitates testing the hypothesis that such heterogeneity contributes to the robustness of bacterial populations under environmental uncertainty. Probability theory forms the base for statistical inference by providing the link between observed environmental data and population genetics models that can take into account population heterogeneity, population and plasmid evolution, and plasmid proliferation under relevant selection pressures. This can overcome limitations of existing studies on the incidence of plasmids, which are difficult to interpret because the host population was not well defined with respect to the phylogenetic relationship of the strains, the spatial area occupied, the likelihood of ecological and genetic interaction between strains, and migration between populations. Mechanisms potentially leading to heterogeneity in bacterial populations with respect to conduciveness to horizontal gene transfer The generation and maintenance of genetic variation seems to be a general ecological strategy of bacterial populations. Various mechanisms have been described to foster intrapopulation heterogeneity. Mutation rates can be genetically altered in a subpopulation of mutator cells. This can be regulated in response to environmental conditions, or focused on specific loci in the genome to minimize deleterious effects on fitness (Moxon & Thaler, 1997; Hallet, 2001). Furthermore, isogenic cells exposed to the same environmental conditions can show significant variation in molecular content and phenotypic characteristics. Cell cycle, cell ageing and epigenetic regulation are among the drivers of this heterogeneity (Avery, 2006). The phenomenon of persisters is well studied, i.e. antibiotictolerant subpopulations of otherwise susceptible bacteria that ensure survival of the population in case of antibiotic pressure (Dhar & McKinney, 2007). Differences within a population in conduciveness to horizontal gene transfer can be the result of genetic or epigenetic variation in defence against foreign DNA or in DNA uptake. Mechanisms that confine DNA uptake to a FEMS Microbiol Ecol 65 (2008) 361–371 365 Population-level robustness through genome flexibility subpopulation could minimize the cost for integration of potentially not (yet) beneficial foreign DNA. Evidence for variation in conduciveness to horizontal gene transfer was recently found within a global collection of E. coli isolates. Lineages were identified by multilocus sequence typing, which bore specific pathogen types arising several times independently through horizontal gene transfer, whereas in other lineages few pathogenic strains have arisen (Wirth et al., 2006). It is plausible that the pathogen-bearing lineages evolved a higher or more variable conduciveness to mobile genetic elements, which resulted in episodic selection of strains that transiently expanded their ecological niche by escaping the host immune response. Pathogenicity-related genes were often found on genomic islands, i.e. large chromosomal regions, in a subset of strains of the same species, that contain genes for chromosomal integration and excision and other mobility-related genes (Dobrindt et al., 2004). They were shown to contribute to strain diversity in populations of Mezorhizobium and Pseudomonas aeruginosa (Klockgether et al., 2007; Nandasena et al., 2007). Site-specific excision can compromise the maintenance of genomic islands in individual cells. As shown for uropathogenic E. coli, excision is stimulated by low temperature and high cell density (Middendorf et al., 2004), which could induce patches of cells without genomic islands in environmental biofilms. Also, epigenetic mechanisms can be the basis for heterogeneity in conduciveness to horizontal gene transfer. Noise in gene expression, when subject to positive feedback regulation, can result in bistability, i.e. the expression of two different genetic programmes within a clonal population of bacteria under identical external conditions (Dubnau & Losick, 2006). Bistability was shown, for different species, to direct uptake of foreign DNA to a fraction of the population. Steinmoen et al. (2003) showed that the partitioning of Streptococcus pneumoniae populations into lysing ‘donor’ and DNA-uptaking ‘recipient’ subpopulations is achieved epigenetically by the ComCDE signal transduction pathway. Kaern et al. (2005) found that in populations of Bacillus subtilis, a minority of cells became competent in DNA uptake due to intrinsically random fluctuations in comK mRNA production and degradation. ComK is the master regulator for competence and positively autoregulates comK expression. It was suggested that the cellular processes, which modulate the level of noise in gene expression and thereby the proportion of competent cells in a population, evolved under selection to maintain a well-tolerated rate of diversification (Maamar et al., 2007). Various barriers limit the transfer, uptake and stabilization of foreign DNA in bacteria (Thomas & Nielsen, 2005). These limitations could be relaxed in a fraction of a population, leading to a subpopulation with an increased likelihood of extant DNA acquisition. As an example, bacterial strains with increased potential for genomic integration of foreign DNA FEMS Microbiol Ecol 65 (2008) 361–371 by recombination have been found in natural populations at proportions of 1–20%, which was in particular caused by altered methyl-directed DNA repair (references in Thomas & Nielsen, 2005). Frequencies of DNA uptake by natural transformation can be highly variable within populations of Pseudomonas stutzeri, B. subtilis or Haemophilus influenzae (Lorenz & Sikorski, 2000, and references therein). Only a fraction of the strains analysed could be transformed, and frequencies varied over more than three orders of magnitude. In addition, natural competence for transformation is under physiological and environmental modulation, so that individual cells of a population in a spatially heterogeneous environment will differ in transformation potential. Some of the mechanisms that limit the transfer, uptake and stabilization of foreign DNA may also be the basis for differential conduciveness to plasmids within natural populations of bacteria. Restriction–modification (R–M) systems play an important role in protection against foreign DNA by significantly decreasing the efficiencies of conjugation, transformation and phage infection (Thomas & Nielsen, 2005; Hoskisson & Smith, 2007). R–M systems are ubiquitous in bacteria and archaea, but they can vary among strains of the same species, and are often part of the flexible genome (Naderer et al., 2002; Bishop et al., 2005). Even more, R–M systems were found to be phase variable in a variety of species, including Helicobacter pylori, Mycoplasma pulmonis, Pasteurella hemolytica, Neisseria meningitidis, Streptomyces coelicolor and H. influenzae (Hoskisson & Smith, 2007), i.e. R–M systems can be switched on and off in individual cells of a population and thereby affect plasmid uptake. In populations of H. influenzae, the high-frequency phase variation of the R–M system HindI was shown to be mediated by tetra- and pentanucleotide repeats (Zaleski et al., 2005). A widely distributed and highly flexible defence system against foreign DNA might be associated with clustered regularly interspaced short palindromic repeats (CRISPRs), which were identified in the genomes of a wide range of bacterial and archaeal genera (Godde & Bickerton, 2006). CRISPR loci consist of direct repeats separated by stretches of variable sequences called spacers. Bacteria can become immune against specific foreign genetic elements when they acquire fragments of the foreign DNA into the spacers. This is mediated by the function of CRISPR-associated cas genes (Barrangou et al., 2007), probably via a mechanism based on RNA interference (Makarova et al., 2006). It was shown that variation of the CRISPR region is fast enough to promote individuality in otherwise nearly clonal populations of Leptospirillum sp. (Tyson & Banfield, 2008). This hypervariability most likely will cause microdiversity in DNA acquisition within natural populations of bacteria. Among others, spacer sequences from plasmids were also found (Makarova et al., 2006), but experimental evidence for interference with plasmid stability is still missing. 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c 366 H. Heuer et al. Table 1. Terms and definitions Environmental uncertainty Sympatric diversification Bacterial population Pan/core genome Flexible genome Mobile genetic elements Flexible genetic elements Evolvability Frequency-dependent selection Bistability Phase variation Surface exclusion Entry exclusion Plasmid incompatibility Broad host-range plasmids Irregular external change that challenge the adaptive potential of a population to withstand adverse conditions, or to transiently expand its ecological niche Development of genetically encoded phenotypic variation in a population that occupies a single and contiguous spatial area Bacterial cells of the same species that live in the same habitat patch and therefore have a chance to interact ecologically and genetically The sum of all genes that occur in any/each strain of a species Counterpart of the core genome that varies between strains of a species, determined by its high rate of acquisition, deletion and recombination of genetic elements Plasmids, transposons, IS elements, integrative conjugative elements, gene cassettes, bacteriophages and other genetic units that can relocate Mobile genetic elements and other genetic structures like repetitive sequences that foster genetic variation in a population The heritable ability of an evolutionary lineage to generate and maintain genetic variants, which are potentially adaptive in the course of evolution In stabilizing (negative) frequency-dependent selection, the fitness of a phenotype decreases relative to other phenotypes in a given population as it becomes more common Nonunimodal phenotypic variation due to the expression of two different genetic programmes within a clonal population of bacteria under identical external conditions ON2OFF switching of a phenotype on the DNA level, e.g. by reversible recombinase-mediated inversion of a DNA segment, leading to a heterogeneous bacterial population Inhibition of mating pair formation in conjugative DNA transfer (mediated by the gene traT of IncF plasmids) Inhibition of plasmid transfer despite mating pair formation (e.g. mediated by genes eexAB of IncP-1 plasmids) Inability of related plasmids with interfering partitioning or replication systems to stably coexist in the same host cell Plasmids that transfer by conjugation between many species and that are stably inherited by hosts belonging to distant phylogenetic branches, e.g. from different subclasses of Proteobacteria. The potential host-range of IncP-1 plasmids comprises virtually all Gram-negative bacteria Furthermore, variation in conduciveness to plasmids among strains of the same species may be based on the molecular mechanisms underlying surface exclusion, entry exclusion or plasmid incompatibility when these traits were acquired by the chromosome of some strains of a population (for an explanation see Table 1). As an example, the acquisition of a 16-bp palindrome sequence, which is the binding site of the partitioning protein ParB, was sufficient to induce incompatibility towards a group of plasmids from Alphaproteobacteria represented by the symbiotic plasmid p42d of Rhizobium etli CE3 (Sóberon et al., 2004). Population heterogeneity with respect to mobile genetic elements could also be achieved by mechanisms leading to variable maintenance of acquired DNA. Enterobacteriaceae and probably also other Proteobacteria can selectively silence the expression of foreign DNA by binding of the nucleoidassociated protein H-NS (Navarre et al., 2007). This could compromise the replication and maintenance of plasmids, thus increasing the rate of segregational loss. In contrast, silencing of accessory elements of the plasmid could decrease the metabolic burden for the cell. Individual cells of a population could have lower levels of H-NS, or activate one of the mechanisms of countersilencing, e.g. mediated by the alternative sigma factor RpoS under conditions of stress. However, the effects of silencing on plasmid maintenance have not been investigated yet. 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c We listed various mechanisms that plausibly lead to population heterogeneity through plasmids or other horizontally transferred DNA. However, it needs to be shown which mechanisms are relevant for natural populations of bacteria. In addition, the relative importance of specific mechanisms that cause heterogeneity in environmental populations still needs to be investigated. Horizontal gene transfer clearly contributes to intrapopulation diversity, but the ecological significance is not well documented. Diversity within a population increases robustness Bacteria evolved adaptive responses to regular variations in their habitats. However, fluctuations of the environment are often irregular and stochastic. In these cases, adaptation through gene regulation may not be sufficient because the repertoire of sensing and response mechanisms is limited, or the time needed for sensing and response may be too short for some threats such as antibiotics. In addition, there is a cost associated with maintaining a sensory machinery to detect external change. In contrast, the generation of population diversity, from which variants are selected in response to environmental change, enables adaptation to challenges for which no sensing mechanism has evolved, which are FEMS Microbiol Ecol 65 (2008) 361–371 367 Population-level robustness through genome flexibility precipitated, or which occur so infrequently in the habitat of the population that maintaining a specific sensor does not pay. The generation of population diversity can be favoured by challenging environmental conditions, e.g. if the underlying cellular mechanism is enhanced in response to stress (Aertsen & Michiels, 2005). According to the insurance hypothesis of biodiversity (Yachi & Loreau, 1999), resistance and resilience of ecosystem functioning increase with species richness because a greater number of species can express a greater range of responses to environmental perturbation. Analogously, genetic diversification of a bacterial population could increase the chance that a subpopulation might be prepared for irregular adverse conditions (Boles et al., 2004). This ecological strategy is known as bet-hedging in population genetics. It was predicted by simulations that in fluctuating environments heterogeneous populations can have a larger net growth rate than homogeneous populations (Thattai & van Oudenaarden, 2004; Gander et al., 2007). Population heterogeneity by stochastic switching is favoured as an adaptive strategy when environments change infrequently, while for adaptation in regularly varying environments it is beneficial to maintain an active sensor for detecting a specific environmental change (Kussell & Leibler, 2005). The longer the environment remains constant, the less it pays to have a sensor. The advantage of stochastic switching will be further increased when individuals remember the last few phenotypic switches and the fluctuating environment exhibits longer correlations (Kussell & Leibler, 2005). Such a stochastic switching with memory can be the horizontal acquisition of a mobile genetic element that confers an adaptive trait under infrequently occurring conditions. As an example of stochastic switching with memory, a bacterial subpopulation may switch the phenotype by acquisition of a plasmid. The plasmid has a history of selection in other cells, for which it has occasionally been beneficial, thereby increasing its abundance and chance of horizontal spread. Environmental uncertainty is typically associated with biotic interactions, because these are dynamically evolving, in contrast to abiotic conditions (Smets & Barkay, 2005). An obvious example of environmental uncertainty is the production of new antibiotic compounds by cohabitants or hosts (including humans), which resulted in the enrichment of subpopulations that acquired mobile genetic elements with antibiotic resistance genes (Witte, 1998; McManus et al., 2002). Environmental uncertainty is not necessarily associated with adverse conditions. It could also be the opportunity to access previously unavailable resources. Examples of the contribution of mobile genetic elements to niche expansion of subpopulations are the escape of the host immune system through acquisition of genomic islands associated with either pathogenic or symbiotic interactions (Hochhut et al., 2006; Arnold et al., 2007), or the degradaFEMS Microbiol Ecol 65 (2008) 361–371 tion of recalcitrant compounds through pathway assembly (Top & Springael, 2003). Inheritance of, and selection for, evolvability Evolutionary successful bacteria must have a balance between both resisting genetic change and evolvability (Lenski et al., 2006). Evolvability may be defined as the heritable ability of an evolutionary lineage to generate and maintain potentially adaptive variants, upon which selection can act. The same mechanisms underlying evolvability will generate population heterogeneity. But only a subset of the variants potentially enhancing robustness of a population will also be potentially adaptive for the evolutionary lineage, because some environmental changes may be nonrecurring or too infrequent to shape the evolution of the lineage. It is controversial as to whether evolvability can be favoured by natural selection, because any cost associated with evolvability takes effect immediately while the benefit of adaptation is assumed to lie in the future (Sniegowski & Murphy, 2006). With global mutators in mind, which affect the whole genome of a cell (like mutS defective strains), it is hard to believe that the benefit will ever balance the cost, even though hypermutable cells can reach high frequencies when they are facing novel challenges from the host immune system or chemotherapy (LeClerc et al., 1996). However, the cost can be minimized on cell level by directing genomic change locally (e.g. contingency loci or plasmids) and on population level by directing genomic change to a subpopulation (heterogeneity with respect to rates of mutation or horizontal gene transfer). Plasmids may occasionally act like global mutators when transposable elements of the plasmid integrate into unspecific chromosomal sites, but they will mostly confer genetic change without major cost to their hosts. Selectively neutral or even slightly deleterious variants can persist for a long time in a fraction of a bacterial population (Novozhilov et al., 2005), most likely long enough to balance the cost with the benefit of adaptability in an uncertain environment. Indeed, experimental data and computer simulations suggest that evolvability is a heritable and selectable trait, and the selective pressure for evolvability becomes increasingly strong as the environmental conditions become more uncertain (Earl & Deem, 2004; Jones et al., 2007). As a consequence, selective pressure of therapeutics may enhance the evolvability of pathogens. Various mechanisms that support evolvability by affecting the probability, type and location of genetic change have been described (Caporale, 2003). In response to environmental uncertainty, elements of the flexible genome seem to have evolved genetic structures that increase their rate of diversification. The effect of simple sequence repeats that can modulate the local mutation rate by several orders of 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c 368 magnitude has been well studied, generating phenotypic diversity in traits that directly interact with host structures (Moxon et al., 2006). The virulence-associated avrBs3-like genes of plant-pathogenic Xanthomonas and Ralstonia are an example of horizontally transferred genes with a high rate of intra- and intergenic recombination due to a domain of direct repeats, which was recently shown to affect population diversity and host preference of Ralstonia solanacearum (Heuer et al., 2007b). Also transposable elements, which are often located on transferable plasmids, increase genetic variation and may be amenable to indirect selection for variability (Capy et al., 2000; Poussier et al., 2003). Selection for evolvability may partially explain the genome flexibility and the prevalence of mobile genetic elements among bacteria. Concluding remarks A considerable part of the population metagenome is attributable to the flexible genome, which is characterized by its high plasticity, modularity and rate of horizontal transfer (Fig. 1). The major ecological advantage of maintaining genome flexibility through horizontal gene transfer might be that the resulting within-population variation can offer immediate benefit in realistic situations of uncertain environmental change by providing robustness on population level. If environmental changes stabilize to become longer-term correlations of the environment and shape the specific ecological niche of the ecotype, then genome flexibility can enhance evolvability. This eventually results in new ecotype functions. In consequence, parts of the flexible genome become fixed in the core genome (Fig. 1). Thus, evolvability can be seen as an added value of the mechanisms that generate variation. The flexible genome is shaped by irregular or novel conditions, which are primarily the result of the dynamic evolution of biotic interactions (Fig. 1). These interactions include those between microbes and us, having implications for human health, agriculture or recalcitrance of man-made compounds. A principal ecological role of conjugative plasmids is probably their contribution to make bacterial populations robust to such environmental uncertainty. Ecosystem stability is often thought to depend primarily on species richness, because functionally redundant species may differ in susceptibility to a particular perturbation. However, experimental evidence in microbial ecology is surprisingly scarce, and bacterial community fingerprints reflecting ribotype diversity are often hardly affected by environmental changes, e.g. effects of plants on bacterial ribotype composition in soil (Felske & Akkermans, 1998; Heuer et al., 2002; Kielak et al., 2008). Research on within-population diversity in the environment, rather than ribotype diversity, is needed to clarify how the robustness of bacterial populations contributes to ecosystem stability. Understanding the dynamics of flexible genomes and the 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c H. Heuer et al. underlying mechanisms is essential to estimate the stability and adaptive potential of bacterial populations as a basis of strategies to control bacterial pathogens or to manage the sustainability of ecosystems. Acknowledgements H.H. was supported by the DFG, in the frame of the project FOR566 ‘Veterinary Medicines in Soils’. We would like to thank Jane E. Stewart for her constructive comments and for proofreading this manuscript. References Abby S & Daubin V (2007) Comparative genomics and the evolution of prokaryotes. Trends Microbiol 15: 135–141. Abdo Z & Golding GB (2007) A step toward barcoding life: a model-based, decision-theoretic method to assign genes to preexisting species groups. Syst Biol 56: 44–56. Aertsen A & Michiels CW (2005) Diversify or die: generation of diversity in response to stress. Crit Rev Microbiol 31: 69–78. Arnold DL, Jackson RW, Waterfield NR & Mansfield JW (2007) Evolution of microbial virulence: the benefits of stress. Trends Genet 23: 293–300. Avery SV (2006) Microbial cell individuality and the underlying sources of heterogeneity. Nat Rev Microbiol 4: 577–587. Bahl MI, Hansen LH & Sorensen SJ (2007) Impact of conjugal transfer on the stability of IncP-1 plasmid pKJK5 in bacterial populations. FEMS Microbiol Lett 266: 250–256. Barrangou R, Fremaux C, Deveau H, Richards M, Boyaval P, Moineau S, Romero DA & Horvath P (2007) CRISPR provides acquired resistance against viruses in prokaryotes. Science 315: 1709–1712. Beeson KE, Erdner DL, Bagwell CE, Lovell CR & Sobecky PA (2002) Differentiation of plasmids in marine diazotroph assemblages determined by randomly amplified polymorphic DNA analysis. Microbiology 148: 179–189. Bergstrom CT, Lipsitch M & Levin BR (2000) Natural selection, infectious transfer and the existence conditions for bacterial plasmids. Genetics 155: 1505–1519. Bishop AL, Baker S, Jenks S et al. (2005) Analysis of the hypervariable region of the Salmonella enterica genome associated with tRNAleuX. J Bacteriol 187: 2469–2482. Boles BR, Thoendel M & Singh PK (2004) Self-generated diversity produces ‘‘insurance effects’’ in biofilm communities. Proc Natl Acad Sci USA 101: 16630–16635. Caporale LH (2003) Natural selection and the emergence of a mutation phenotype: an update of the evolutionary synthesis considering mechanisms that affect genome variation. Annu Rev Microbiol 57: 467–485. Capy P, Gasperi G, Biemont C & Bazin C (2000) Stress and transposable elements: co-evolution or useful parasites? Heredity 85: 101–106. Cook MA, Osborn AM, Bettandorff J & Sobecky PA (2001) Endogenous isolation of replicon probes for assessing plasmid FEMS Microbiol Ecol 65 (2008) 361–371 369 Population-level robustness through genome flexibility ecology of marine sediment microbial communities. Microbiology 147: 2089–2101. DeGelder L, Vandecasteele FPJ, Brown CJ, Forney LJ & Top EM (2005) Plasmid donor affects host range of promiscuous IncP-1 beta plasmid pB10 in an activated-sludge microbial community. Appl Environ Microbiol 71: 5309–5317. DeGelder L, Ponciano JM, Joyce P & Top EM (2007) Stability of a promiscuous plasmid in different hosts: no guarantee for a long-term relationship. Microbiology 153: 452–463. Dhar N & McKinney JD (2007) Microbial phenotypic heterogeneity and antibiotic tolerance. Curr Opin Microbiol 10: 30–38. Dobrindt U, Hochhut B, Hentschel U & Hacker J (2004) Genomic islands in pathogenic and environmental microorganisms. Nat Rev Microbiol 2: 414–424. Dubnau D & Losick R (2006) Bistability in bacteria. Mol Microbiol 61: 564–572. Earl DJ & Deem MW (2004) Evolvability is a selectable trait. Proc Natl Acad Sci USA 101: 11531–11536. Ellis RJ, Lilley AK, Lacey SJ, Murrell D & Godfray HCJ (2007) Frequency-dependent advantages of plasmid carriage by Pseudomonas in homogeneous and spatially structured environments. Int Soc Microb Ecol J 1: 92–95. Felske A & Akkermans ADL (1998) Spatial homogeneity of abundant bacterial 16S rRNA molecules in grassland soils. Microb Ecol 36: 31–36. Gander MJ, Mazza C & Rummler H (2007) Stochastic gene expression in switching environments. J Math Biol 55: 249–269. Godde JS & Bickerton A (2006) The repetitive DNA elements called CRISPRs and their associated genes: evidence of horizontal transfer among prokaryotes. J Mol Evol 62: 718–729. Hacker J & Carniel E (2001) Ecological fitness, genomic islands and bacterial pathogenicity. A Darwinian view of the evolution of microbes. EMBO Rep 2: 376–381. Hallet B (2001) Playing Dr Jekyll and Mr Hyde: combined mechanisms of phase variation in bacteria. Curr Opin Microbiol 4: 570–581. Hao W & Golding GB (2006) The fate of laterally transferred genes: life in the fast lane to adaptation or death. Genome Res 16: 636–643. Harada KM, Aso Y, Hashimoto W, Mikami B & Murata K (2006) Sequence and analysis of the 46.6-kb plasmid pA1 from Sphingomonas sp. A1 that corresponds to the typical IncP1beta plasmid backbone without any accessory gene. Plasmid 56: 11–23. Hedrick PW (1999) Genetics of Populations. Jones and Bartlett Publishers, Sudbury, MA. Hermisson J & Pennings PS (2005) Soft sweeps: molecular population genetics of adaptation from standing genetic variation. Genetics 169: 2335–2352. Heuer H, Kroppenstedt RM, Lottmann J, Berg G & Smalla K (2002) Effects of T4 lysozyme release from transgenic potato roots on bacterial rhizosphere relative to communities are FEMS Microbiol Ecol 65 (2008) 361–371 negligible natural factors. Appl Environ Microbiol 68: 1325–1335. Heuer H, Szczepanowski R, Schneiker S, Pühler A, Top EM & Schlüter A (2004) The complete sequences of plasmids pB2 and pB3 provide evidence for a recent ancestor of the IncP-1b group without any accessory genes. Microbiology 150: 3591–3599. Heuer H, Fox RE & Top EM (2007a) Frequent conjugative transfer accelerates adaptation of a broad-host-range plasmid to an unfavorable Pseudomonas putida host. FEMS Microbiol Ecol 59: 738–748. Heuer H, Yin YN, Xue QY, Smalla K & Guo JH (2007b) Repeat domain diversity of avrBs3-like genes in Ralstonia solanacearum strains and association with host preferences in the field. Appl Environ Microbiol 73: 4379–4384. Hochhut B, Wilde C, Balling G, Middendorf B, Dobrindt U, Brzuszkiewicz E, Gottschalk G, Carniel E & Hacker J (2006) Role of pathogenicity island-associated integrases in the genome plasticity of uropathogenic Escherichia coli strain 536. Mol Microbiol 61: 584–595. Hoskisson PA & Smith MC (2007) Hypervariation and phase variation in the bacteriophage ‘resistome’. Curr Opin Microbiol 10: 396–400. Jain R, Rivera MC, Moore JE & Lake JA (2003) Horizontal gene transfer accelerates genome innovation and evolution. Mol Biol Evol 20: 1598–1602. Johnson LJ & Brookfield JFY (2002) Evolutionary dynamics of a selfishly spreading gene that stimulates sexual reproduction in a partially sexual population. J Evol Biol 15: 42–48. Jones AG, Arnold SJ & Burger R (2007) The mutation matrix and the evolution of evolvability. Evolution 61: 727–745. Kaern M, Elston TC, Blake WJ & Collins JJ (2005) Stochasticity in gene expression: from theories to phenotypes. Nat Rev Genet 6: 451–464. Kamachi K, Sota M, Tamai Y, Nagata N, Konda T, Inoue T, Top EM & Arakawa Y (2006) Plasmid pBP136 from Bordetella pertussis represents an ancestral form of IncP-1beta plasmids without accessory mobile elements. Microbiology 152: 3477–3484. Kielak A, Pijl AS, van Veen JA & Kowalchuk GA (2008) Differences in vegetation composition and plant species identity lead to only minor changes in soil-borne microbial communities in a former arable field. FEMS Microbiol Ecol 63: 372–382. Kijima-Tanaka M, Kawanishi M, Fukuda Y, Suzuki S & Yagyu K (2007) Molecular diversity of Photobacterium damselae ssp. piscicida from cultured amberjacks (Seriola spp.) in Japan by pulsed-field gel electrophoresis and plasmid profiles. J Appl Microbiol 103: 381–389. Klockgether J, Würdemann D, Reva O, Wiehlmann L & Tümmler B (2007) Diversity of the abundant pKLC102/PAGI-2 family of genomic islands in Pseudomonas aeruginosa. J Bacteriol 189: 2443–2459. Kuhn S, Stiens M, Pühler A & Schlüter A (2008) Prevalence of pSmeSM11a-like plasmids in indigenous Sinorhizobium 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c 370 meliloti strains isolated in the course of a field release experiment with genetically modified S. meliloti strains. FEMS Microbiol Ecol 63: 118–131. Kurland CG (2005) What tangled web: barriers to rampant horizontal gene transfer. BioEssays 27: 741–747. Kussell E & Leibler S (2005) Phenotypic diversity, population growth, and information in fluctuating environments. Science 309: 2075–2078. Lawrence JG & Ochman H (1998) Molecular archaeology of the Escherichia coli genome. Proc Natl Acad Sci USA 95: 9413–9417. LeClerc JE, Li B, Payne WL & Cebula TA (1996) High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274: 1208–1211. Lenski RE, Barrick JE & Ofria C (2006) Balancing robustness and evolvability. PLoS Biol 4: 2190–2192. Lerat E, Daubin V, Ochman H & Moran NA (2005) Evolutionary origins of genomic repertoires in bacteria. PLoS Biol 3: e130. Levin BR (1988) Frequency-dependent selection in bacterial populations. Philos Trans R Soc Lond B Biol Sci 319: 459–472. Lilley AK & Bailey MJ (1997a) Impact of plasmid pQBR103 acquisition and carriage on the phytosphere fitness of Pseudomonas fluorescens SBW25: burden and benefit. Appl Environ Microbiol 63: 1584–1587. Lilley AK & Bailey MJ (1997b) The acquisition of indigenous plasmids by a genetically marked Pseudomonad population colonizing the sugar beet phytosphere is related to local environmental conditions. Appl Environ Microbiol 63: 1577–1583. Lorenz MG & Sikorski J (2000) The potential for intraspecific horizontal gene exchange by natural genetic transformation: sexual isolation among genomovars of Pseudomonas stutzeri. Microbiology 146: 3081–3090. Lowe A, Harris S & Ashton P (2004) Ecological Genetics: Design, Analysis, and Application. Blackwell Publishing, Malden, MA. Maamar H, Raj A & Dubnau D (2007) Noise in gene expression determines cell fate in Bacillus subtilis. Science 317: 526–529. Makarova KS, Grishin NV, Shabalina SA, Wolf YI & Koonin EV (2006) A putative RNA-interference-based immune system in prokaryotes: computational analysis of the predicted enzymatic machinery, functional analogies with eukaryotic RNAi, and hypothetical mechanisms of action. Biol Direct 1: 7. Malik A, Khan IF & Aleem A (2002) Plasmid incidence in bacteria from agricultural and industrial soils. World J Microb Biot 18: 827–833. McManus PS, Stockwell VO, Sundin GW & Jones AL (2002) Antibiotic use in plant agriculture. Annu Rev Phytopathol 40: 443–465. Meyer CP & Paulay G (2005) DNA barcoding: error rates based on comprehensive sampling. PLoS Biol 3: e422. Middendorf B, Hochhut B, Leipold K, Dobrindt U, Blum-Oehler G & Hacker J (2004) Instability of pathogenicity islands in uropathogenic Escherichia coli 536. J Bacteriol 186: 3086–3096. Miyakoshi M, Shintani M, Terabayashi T, Kai S, Yamane H & Nojiri H (2007) Transcriptome analysis of Pseudomonas putida 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c H. Heuer et al. KT2440 harboring the completely sequenced IncP-7 plasmid pCAR1. J Bacteriol 189: 6849–6860. Mølbak L, Licht TR, Kvist T, Kroer N & Andersen SR (2003) Plasmid transfer from Pseudomonas putida to the indigenous bacteria on alfalfa sprouts: characterization, direct quantification, and in situ location of transconjugant cells. Appl Environ Microbiol 69: 5536–5542. Mølbak L, Molin S & Kroer N (2007) Root growth and exudate production define the frequency of horizontal plasmid transfer in the rhizosphere. FEMS Microbiol Ecol 59: 167–176. Moxon ER & Thaler DS (1997) Microbial genetics. The tinkerer’s evolving tool-box. Nature 387: 659, 661–659, 662. Moxon R, Bayliss C & Hood D (2006) Bacterial contingency loci: the role of simple sequence DNA repeats in bacterial adaptation. Annu Rev Genet 40: 307–333. Naderer M, Brust JR, Knowle D & Blumenthal RM (2002) Mobility of a restriction-modification system revealed by its genetic contexts in three hosts. J Bacteriol 184: 2411–2419. Nandasena KG, O’Hara GW, Tiwari RP, Sezmis E & Howieson JG (2007) In situ lateral transfer of symbiosis islands results in rapid evolution of diverse competitive strains of mesorhizobia suboptimal in symbiotic nitrogen fixation on the pasture legume Biserrula pelecinus L. Environ Microbiol 9: 2496–2511. Navarre WW, McClelland M, Libby SJ & Fang FC (2007) Silencing of xenogeneic DNA by H-NS-facilitation of lateral gene transfer in bacteria by a defense system that recognizes foreign DNA. Genes Dev 21: 1456–1471. Novozhilov AS, Karev GP & Koonin EV (2005) Mathematical modeling of evolution of horizontally transferred genes. Mol Biol Evol 22: 1721–1732. Ochman H & Davalos LM (2006) The nature and dynamics of bacterial genomes. Science 311: 1730–1733. Poussier S, Thoquet P, Trigalet-Demery D, Barthet S, Meyer D, Arlat M & Trigalet A (2003) Host plant-dependent phenotypic reversion of Ralstonia solanacearum from non-pathogenic to pathogenic forms via alterations in the phcA gene. Mol Microbiol 49: 991–1003. Pukall R, Tschäpe H & Smalla K (1996) Monitoring the spread of broad host and narrow host range plasmids in soil microcosms. FEMS Microbiol Ecol 20: 53–66. Regenhardt D, Heuer H, Heim S, Fernandez DU, Strömpl C, Moore ER & Timmis KN (2002) Pedigree and taxonomic credentials of Pseudomonas putida strain KT2440. Environ Microbiol 4: 912–915. Smets BF & Barkay T (2005) Horizontal gene transfer: perspectives at a crossroads of scientific disciplines. Nat Rev Microbiol 3: 675–678. Sniegowski PD & Murphy HA (2006) Evolvability. Curr Biol 16: R831–R834. Sobecky PA (1999) Plasmid ecology of marine sediment microbial communities. Hydrobiologia 401: 9–18. Sóberon N, Venkova-Canova T, Ramı́rez-Romero MA, TéllezSosa J & Cevallos MA (2004) Incompatibility and the partitioning site of the repABC basic replicon of the symbiotic plasmid from Rhizobium etli. Plasmid 51: 203–216. FEMS Microbiol Ecol 65 (2008) 361–371 371 Population-level robustness through genome flexibility Steinmoen H, Teigen A & Håvarstein LS (2003) Competenceinduced cells of Streptococcus pneumoniae lyse competencedeficient cells of the same strain during cocultivation. J Bacteriol 185: 7176–7183. Thattai M & van Oudenaarden A (2004) Stochastic gene expression in fluctuating environments. Genetics 167: 523–530. Thomas CM (2004) Evolution and population genetics of bacterial plasmids. Plasmid Biology (Funnel BE & Phillips GJ, eds), pp. 509–528. ASM Press, Washington, DC. Thomas CM & Nielsen KM (2005) Mechanisms of, and barriers to, horizontal gene transfer between bacteria. Nat Rev Microbiol 3: 711–721. Top EM & Springael D (2003) The role of mobile genetic elements in bacterial adaptation to xenobiotic organic compounds. Curr Opin Biotechnol 14: 262–269. Tyson GW & Banfield JF (2008) Rapidly evolving CRISPRs implicated in acquired resistance of microorganisms to viruses. Environ Microbiol 10: 200–207. FEMS Microbiol Ecol 65 (2008) 361–371 Waples RS & Gaggiotti O (2006) What is a population? An empirical evaluation of some genetic methods for identifying the number of gene pools and their degree of connectivity. Mol Ecol 15: 1419–1439. Wirth T, Falush D, Lan R et al. (2006) Sex and virulence in Escherichia coli: an evolutionary perspective. Mol Microbiol 60: 1136–1151. Witte W (1998) Medical consequences of antibiotic use in agriculture. Science 279: 996–997. Yachi S & Loreau M (1999) Biodiversity and ecosystem productivity in a fluctuating environment: the insurance hypothesis. Proc Natl Acad Sci USA 96: 1463–1468. Zaleski P, Wojciechowski M & Piekarowicz A (2005) The role of Dam methylation in phase variation of Haemophilus influenzae genes involved in defence against phage infection. Microbiology 151: 3361–3369. 2008 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved c