FEMS Microbiology Ecology 42 (2002) 177^185
www.fems-microbiology.org
MiniReview
Ecological and molecular maintenance strategies of mobile genetic elements S.L. Turner a
a;
, M.J. Bailey a , A.K. Lilley a , C.M. Thomas
b
Molecular Microbial Ecology Laboratory, Institute of Virology and Environmental Microbial Ecology, Natural Environment Research Council, Centre for Ecology and Hydrology ^ Oxford, Mans¢eld Road, Oxford OX1 3SR, UK b School of Biosciences, University of Birmingham, Edgbaston, Birmingham B15 2TT, UK Received 5 March 2002; received in revised form 10 July 2002; accepted 10 July 2002 First published online 13 September 2002
Abstract This review considers the influence of selection pressure, fitness and population structures on the evolution of mobile genetic elements (including plasmids, phage, pathogenicity islands, transposons and insertion sequences) that constitute the horizontal gene pool of bacteria. These are considered at different scales using examples from in vitro evolutionary studies of Escherichia coli and associated bacteriophage, detailed molecular analyses of the broad host-range IncP-1 plasmids, population surveys of pseudomonad plasmids and genomic comparisons of members of the Rhizobiaceae. All biological systems show genetic redundancy (the existence of allelic variation) at some population level, i.e. within a cell, a clone, population or community. We consider the level(s) at which redundancy is expressed and how this will affect and has influenced the evolution of mobile genetic elements. 3 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. Keywords : Plasmid; Horizontal gene pool; Evolution; Ecology
1. Introduction In Darwin’s concept of the ‘survival of the ¢ttest’ the unit of selection was the individual. A modern extension of Darwin’s theory, as expounded most eloquently by Dawkins [1] is the concept of the sel¢sh gene, where the unit of selection is the gene and the raison d’e“tre of genetic components (sequences) is simply to ensure their own multiplication. Discovering how genes and other genetic elements a¡ect host ¢tness is a multidisciplinary project that requires us to draw on ecology, population genetics and molecular biology. The aim of this review is to summarise and apply advances in these areas of biology to investigate the persistence and in£uence of mobile genetic elements (MGEs) and the horizontal gene pool (HGP) within bacterial populations. What follows has been tempered by the knowledge that it is dangerous to generalise and that, as with most biological de¢nitions, generalisations serve to demarcate distinct character sets within a continuum.
* Corresponding author. Tel. : +44 (1865) 281630; Fax :+44 (1865) 281696. E-mail address :
[email protected] (S.L. Turner).
MGEs are, by their very nature itinerants, pieces of DNA with no ¢xed abode, which are unable to replicate outside of a suitable host bacterium. MGEs have evolved a variety of mechanisms to enhance their chances of persistence. As the current state of knowledge of the extent of bacterial diversity and ecology remains limited, with only a small proportion of the true diversity identi¢ed, it is likely that we know even less about the ecology and diversity of their MGEs. This idea, that MGEs are adapted to speci¢c niches (host ranges) and are under-sampled, is supported by the observations that many environmental plasmids did not hybridise with any of the known plasmid replication probes [2,3,] described by Couturier et al. [4]. On the other hand, the ability of MGEs to spread between strains, species, genera and habitats could suggest a more uniform, global distribution of MGEs in which they compete for niche (bacterial host) occupancy. This would predict identi¢cation of relatives of MGEs that we already know, in previously unexplored organisms and environments, e.g. the broad host range IncP-1 plasmids have been shown to replicate in diverse bacteria and have been detected in many environments [5]. However, sequencing of new MGEs has so far borne out the prediction that their maintenance and transfer functions often
0168-6496 / 02 / $22.00 3 2002 Federation of European Microbiological Societies. Published by Elsevier Science B.V. All rights reserved. PII : S 0 1 6 8 - 6 4 9 6 ( 0 2 ) 0 0 3 7 2 - 0 Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
178
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
consist of mosaics some components of which at least will be related to previously known components. For example, among rhizobial plasmids the gene essential for replication initiation, repC [6] is unique to this family and closely related K-proteobacterial plasmids, whereas the ¢rst two genes of the minimal replicon, repAB, belong to a large family of chromosomal, plasmid and phage partitioning systems. Consequently generalisations from known systems should be expected to apply to some extent to novel systems. Using examples of characterised MGEs that adopt di¡erent survival strategies we aim to explore the selective pressures and mechanisms driving the evolution of these elements and ask what are the implications of their behaviour for the ecology of bacteria at the individual, microcolony, population and broader community scale. To date comparisons of the available sequence information for microbial genomes reinforce the importance of the HGP in bacterial evolution [7]. Analyses indicate that as much as 24% of the Escherichia coli genome has been acquired due to the transfer and recombination of genetic information [8]. This is a substantial amount for an organism considered to exhibit a predominantly clonal population structure [9]. The signi¢cance of the role of the HGP in genome evolution is no longer in question; however, there are no clear population genetic models for persistence of the HGP. Improved population genetic models are available which seem to be applicable to molecular evolution studies [7,8]. However, in diverse structured communities, mass action models are not accurate for describing gene £ow. Therefore, in these environments it is essential to determine what selection pressures exist, the temporal and spatial distribution of such pressures, how they act across a range of ecological scales and how these factors in£uence the genetic structure of the HGP. Ecological scale is directly relevant in more complex (natural) environments where selection is dispersed, unde¢ned and occurs at several di¡erent levels (individual, colony/microcolony or at the population level). Three recent articles have in£uenced the structure of this review [10^12]. The ¢rst two discuss the concept of molecular ‘warfare’ among bacteria using a three component model comprising two resistant genotypes, toxin plus antidote producing and antidote-only producing, and sensitive bacteria expressing neither gene. Toxins are e¡ectively bactericides, often small, di¡usable compounds, which kill cells lacking the resistance (antidote) gene/protein. These authors consider ecological models that take into account conditions and interactions at both the cellular and population scale. In the model there is an assumed cost associated with both toxin and antidote production. In homogeneous environments toxin producers can kill non-resistant bacteria once their population and hence toxin levels reach a critical density. However, until this critical density is reached the antidote-only producers and sensitive bacteria do better than the producers because
they do not carry the additional metabolic cost of toxin production. Within the non-producing populations the sensitive bacteria have a lower metabolic burden and so can out-compete antidote-only producers. Overall these interactions generate a dynamic system. In structured and dispersed systems, which predominate in nature, the model favours the evolution and maintenance of all three genotypes and is predicted to lead to increased diversity. The population genetic models of Krakauer and Plotkin [12] further illustrate how the genetic pro¢les of bacterial populations are in a state of continual £ux, driven in part by the accumulation of random mutations. These mutations may become ¢xed or lost depending on the population size and relative distribution, and the prevailing selection pressure(s). Basically, the stronger the selection pressure (steeper ¢tness landscape) the narrower the genetic window for survival and hence the more homogeneous the resultant population that emerges as a result of selection against even mildly deleterious mutations (Fig. 1). By contrast the weaker the level of selection (£atter ¢tness landscape) the greater the potential for accumulation of mutations that have no or negligible in£uence on individual ¢tness. This second scenario ¢ts the narrow de¢nition of redundancy used by Krakauer and Plotkin [12]. Table 1 lists some of the mechanisms proposed by these authors [12] that contribute to redundancy and antiredundancy processes in populations. They argue that mutations that have undetectable or minimal impact on the phenotype o¡er minimal selective advantage and are rarely ¢xed to dominate the population. These, ‘£atter’, landscapes are relatively tolerant of the accumulation of redundancy resulting in populations comprising many closely related genotypes that show minimal phenotypic variance under the prevailing conditions. Cells in steeper ¢tness landscapes should pay a high price for the accumulation of redundancy and are therefore counter-selected. This correlates with sensitivity and even hypersensitivity to mutation and leads to populations exhibiting antiredundancy. In such circumstances mutations will a¡ect phenotype and have a detectable impact on host ¢tness : either bene¢cial, leading to ¢xation of the mutation, or deleterious, leading to counter-selection and eventual extinction. Predictably the shape of the selection landscape is highly variable in space and time. It will also vary from one genetic location to another. What we observe results from the balance between these processes. Whether we can deduce anything about the processes that dominate under di¡erent environmental conditions is considered below. If the population genetics of MGEs are to be considered from the standpoint of redundancy/antiredundancy, the more traditional distinctions between the di¡erent classes of genetic element, such as the mode of replication, stability functions, etc. used by a genetic element are relevant. Whilst we are not interested in the precise mechanisms, we are interested in the consequences of the mechanism, e.g. high or low cellular density (copy number), frequency of
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
Fig. 1. These ¢tness landscapes [7] depict the e¡ects of the extent of genetic variation (in the x-y plane) on the relative ¢tness of bacteria. In steeper landscapes (A) the window of genetic variation which permits persistence is narrow, while in £atter landscapes (B) the window is considerably broader and therefore more permissive of genetic redundancy. In natural environments, bacteria are often spatially distributed (C) and will interact with bacteria having di¡erent ¢tness landscapes. These interactions and other spatially and temporally distributed variables can modify the selective pressures and thereby also modify the steepness and shape of ¢tness landscapes.
horizontal transfer, etc. This is a reasonable standpoint, as there are no clear boundaries between the di¡erent types of MGEs, plasmids, phage, transposons (Tn), insertion (IS) elements, integrons, etc. (for review see [13]). Toussaint and Merlin [13] identify clear mechanistic overlaps between processes, such as replication, transposition and transfer, that might be considered typical of di¡erent classes of MGEs. These authors also underline the principle that MGEs are modular, often comprising distinct functional building blocks that can be shared among different classes of MGE, which combine to generate composite MGEs that are mosaics of smaller building blocks. One common phenotype used to discriminate between MGEs is their realised cellular copy number, e.g. some plasmids, IS elements or Tn’s are high copy number (ColE1 and IS1 in E. coli), whereas others persist at low (or unitary) copy number per chromosome (lysogenic P1 and Tn7). Similarly some elements exhibit signi¢cant levels of horizontal spread both within (Tn’s and IS elements) and between (lytic phage and conjugative plasmids) cells, whereas other MGEs adopt a strategy of predominantly vertical inheritance. For example, lysogenic phage and transfer de¢cient plasmids persist within a clonal lineage replicating and segregating along with the host chromosome. These elements can, however, be mobilised by other self-transferring elements, either directly if they possess a suitable transfer origin, or indirectly following a recombination event between the non-mobile and a mobile element. We have chosen four model systems that display redundancy at di¡erent genetic scales. The ¢rst concerns in vitro studies of E. coli-phage interactions and serves to represent the underlying principles of how population size and
179
selection in£uence molecular evolution. The second concerns the broad host range IncP-1 plasmids that have evolved to colonise highly diverse niches (host cell types). The third derives from ecological surveys of pseudomonad plasmids clearly demonstrating temporal and spatial variation in selection pressures and the impact of HGT. Finally we consider the genome sequence data that is available for four rhizobial species. This provides a unique host-plasmid data set including seven plasmid-derived replicons, all of which are larger than 200 kbp. Table 2, presents a summary of whether these di¡erent example MGEs, exhibit redundancy or antiredundancy at di¡erent ecological scales for three general functional modules, replication, transfer and traits carried. By treating MGEs in this way we hope to get a general picture of MGE population structures at di¡erent ecological scales and how this equates with likely evolutionary pressures upon them and their host bacterial populations.
2. In vitro evolutionary/adaptation studies of E. coli-phage populations Following in the footsteps of Lenski [14], many in vitro experiments have been undertaken using E. coli and associated phage and plasmids to investigate some of the fundamental principles governing molecular evolution and adaptation to environmental perturbations. These systems usually involve relatively homogeneous (e¡ectively isogenic) starting populations, high growth rates and strong selection pressures. Although, these systems may not be directly relevant to dispersed and heterogeneous natural populations, they provide detailed information relevant to the molecular principles that underlie molecular adaptive processes. For example, studies of the adaptation of phage phiX-174 to elevated growth temperatures and a novel host demonstrate that mutations occur randomly, facilitating adaptation to the same selective pressure via several di¡erent mutational routes [15]. These studies emphasise the apparent random nature of some evolutionary processes and con¢rm that some genotypes may be transient and lost during adaptation [16]. The fact that distinct populations pass through di¡erent adaptive pathways in response to the same selection process [15] is indicative of the plasticity conferred by genetic redundancy. In all popTable 1 A summary of terms responsible for creating redundancy and antiredundancy at the cellular level relevant to prokaryotic MGEs Redundancy
Antiredundancy
Gene duplication Polyploidy/high copy number No/few genes in coregulatory networks Parallel pathways Bulk transmission
Overlapping reading frames Haploidy/low copy number Multiple genes respond to same regulator Serial/unique pathways Bottlenecks in transmission
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
180
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
Table 2 Summary of MGEs discussed, their survival strategies and the likely consequence, genetic redundancy (R) or antiredundancy (AR) at di¡erent ecological scales MGE system MGE module
Lytic phage in E. coli and rolling circle plasmids in Gram-positive bacteria
IncP1 plasmids
Rhizobial pseudomonad environmental plasmids
Replication
High copy number, intracellular R. Allows the evolution of novel replication incompatibility types, R
Transfer
High levels horizontal spread and population level R
Low copy number, intracellular AR. No or little variation among isolates from distinct environments, indicative population level AR? High levels of HG transfer both within and between species, allowing rapid spread of adaptive genotypes and population level R
Traits carried
Self-propagation of phage or few, highly adaptive/sel¢sh traits, e.g. antibiotic resistance/host killing, AR
Low copy number, AR. However, in rhizobia many homologous, compatible sequences within individual cells. Indicative of past cellular or population level, R? Minority of plasmids are conjugative at low frequencies, but population studies indicate these are su⁄cient to in£uence population dynamics and structures, R? Very few highly conserved traits shared between replicons, AR. But many functionally related traits, e.g. ABC transporters on the same replicon, R?
Only strongly selected traits that di¡er between environments, AR
ulations, mutations arise by chance giving rise to a mixed population (exhibiting redundancy) under mildly selective conditions, large populations should contain numerically more mutations. As seen in the adaptation of phiX-174 to elevated temperature [15], once strong selection is applied, advantageous mutations out-compete wild-type and maladapted strains. These then become dominant and, due to the selection bottleneck, generate a homogeneous population exhibiting reduced levels of redundancy. The frequency and strength of selective sweeps and population structures will result in see-sawing between adapted, homogeneous populations (exhibiting antiredundancy) and heterogeneous populations (exhibiting redundancy). To some extent this restates the process by which plasmids can be maintained in a population via cycles of ‘sweeping innovations’ described by Bergstrom et al. [17], or coadaptation to niche succession by suitable hosts [18]. In both cases combinations of plasmid and host strain arise in circumstances where the association enjoys elevated ¢tness. In structured environments the selective sweeps are likely to be highly localised, probably at the cell/microcolony level. It seems likely that what we observe following population level surveys is an average of disparate but localised selection pressures producing a highly complex pattern that is very di⁄cult to interpret. Messenger et al. [19] also used an in vitro molecular evolution approach to force another E. coli phage, f1, to persist in environments favouring either vertical or horizontal spread for several generations. Phage f1 is ¢lamentous and as such does not kill its host and so can be both vertically and horizontally transmitted within populations. This was achieved by manipulating the frequencies that phage populations were able to transfer horizontally or constrained to solely vertical transmission (eight times longer in one population) in the experiments. The experiments demonstrated rapid adaptation to the prevailing conditions, with viruses exposed to the higher rates of horizontal (infectious) transmission evolving higher titres of virus production. These experiments demonstrate the
general principle of a ‘trade-o¡’ between virulence and reproductive capacity in parasites. Furthermore, this study clearly indicates that the population structure of the host in£uences the adaptation/evolution of the MGE. These types of experiment indicate the adaptive bene¢ts that genetic redundancy can confer on a population. The high intracellular titre of phage means that rare, spontaneous mutant genotypes can persist through molecular cross-feeding. This gives rise to a mixed population of phage with genetic plasticity that may enable the population to respond rapidly to an imposed selection. How do such experiments relate to other systems ? In Gram-positive bacteria, for example, many small plasmids exhibit high copy number and are maintained by rolling circle replication that is regulated by antisense RNA. Given the complementary nature of antisense RNA and its target, encoded by the same DNA segment, the evolutionary landscape may be as follows. Changes that decrease the activity of the antisense RNA will be selected against due to increased genetic load from higher copy number. Changes that increase antisense RNA activity will be lost due to increased segregation rate arising from lower copy number. Changes that are neutral with respect to antisense RNA e⁄ciency but change the speci¢city allowing the new variant not to compete with its parent will persist. This redundancy may increase metabolic load in the short term but will have a bene¢t by allowing di¡erent traits maintained by compatible plasmids to co-exist. Thus genetic redundancy in these populations of plasmids appears to have given rise to many closely related plasmid incompatibility groups that enable di¡erent replicons to stably co-exist within the same bacterial population [20]. Similarly phages have been described that are self-replication de¢cient but which have duplicated origins to increase their chances of replication by coinfective, fully functional phage and maintain themselves in populations. This reliance on the products (active proteins) of other MGE has been taken to an extreme in some cases, e.g. the satellite bacteriophage P4 is dependent on phage P2 to provide
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
lytic functions in E. coli [21]. The deletion processes that result in the need for dependence of one MGE on another can contribute to evolution of MGEs. For example, high copy number mutant IncP-1 plasmids arise spontaneously by illegitimate recombination events and are deleterious to the host. However, these can bene¢t the host by decreasing the copy number of the larger deleterious (because of higher phenotypic or replicative load) wild-type plasmid form, thus ameliorating the harmful phenotype created by the elevated copy number [22].
3. In vitro molecular genetics and population studies of IncP-1 plasmids Many molecular evolution studies involving MGEs have used a high copy number MGE in the devised experimental system. However, not all MGEs fall into this category. In E. coli plasmids such as F and lysogenic phage such as P1/7 are low copy number. Low copy number in£icts an additional challenge on MGEs ^ that of ensuring partitioning of at least one MGE copy to each daughter cell at cell division. Segregation by random processes ¢ts Eq. 1 : P0 ¼ 213n
ð1Þ
which describes the probability of a plasmid-free daughter (P0 ) arising at each cell division for a plasmid of copy number = n. One evolutionary solution to this is to become part of the chromosome by integration (lambda, Tn’s, conjugative Tn’s etc). The other is to actively partition copies to each daughter cell. Mechanisms that enhance replicon maintenance (replication, active partitioning and transfer) are referred to as backbone functions. In low copy number plasmids it may be an advantage for these traits to be co-ordinately regulated. Co-ordinated regulation is an antiredundant process (see Table 1). IncP-1 group plasmids are among some of the best-studied low copy number plasmids at the molecular level [23]. The core backbone maintenance functions are e¡ectively one regulon that is co-ordinately regulated via the multiple regulators ^ KorA, KorB, TrbA and KorC. While the multiplicity of regulators may appear to be redundant, these regulators interact co-operatively and this leads to tighter regulation over small changes in repressor protein concentration [24,25]. This not only creates a sensitive response to changing conditions but also ensures steeper selection curves on the evolutionary landscape ^ that is, once such complexity has arisen, it is unlikely to be lost. This tight regulation of both copy number and backbone functions indicates antiredundancy at the cellular level. Antiredundancy also appears at the wider, population level. In contrast to the small, high copy number plasmids referred to above, drift appears to have happened in the IncP-1 group without creating new incompatibility groups. More than 30% sequence divergence has occurred in intergenic spacer sequences or other regions where changes do not alter
181
function, but key sequences for replication, maintenance and regulation have been highly selected and have remained largely unchanged [26]. IncP-1 plasmids also prevent themselves spreading to bacteria already carrying an IncP-1 plasmid ^ by surface exclusion mechanisms [23]. This may o¡set the load imposed by constantly maintaining a state of transfer pro¢ciency that, even if the transfer genes are tightly regulated, constitutes a burden. This burden can be metabolic or potentially lethal by creating sensitivity to phages that rely on the IncP-1 pilus for attachment and cell infection. All IncP-1 plasmids isolated and studied to date carry one or more known or probable phenotypic determinants ^ none of which appear to be cryptic [27]. Interestingly, many of the selectable phenotypes carried by IncP-1 plasmids are encoded by genes that appear not to be regulated ^ perhaps a higher level of expression is tolerated for genes for advantageous traits. Also, there appear to be regions where insertions are tolerated, so-called ‘redundant’ regions. In the case of the IncP-1L plasmids this redundancy of insertion sites seems to have been extended as a theme ^ the sites where insertions occur contain repeated copies of a palindromic sequence that may attract insertions away from core or backbone functions [26]. It is appropriate to speculate further that those MGEs that may have become ¢xed in a population, or have become highly adapted to a speci¢c niche, continue to be selectively advantageous to their host. Under such circumstances the presence of an adapted plasmid would prevent establishment of a modi¢ed plasmid carrying additional, but non-advantageous phenotypes ful¢lling the antiredundancy processes due to their cost of carriage. This in e¡ect results in a counterselective bottleneck within populations. The integrity of the MGE can only be conserved at the functional level and de¢ned by the essential backbone. This may be illustrated by the apparent broad-host-range (BHR) of IncP-1 elements. For example, does their ability to operate in a range of genomic backgrounds preclude host-specialisation and co-adaptation? Could this indicate that the BHR nature and high transfer rates of IncP-1s are the fundamental explanations for their persistence in the environment? This is an interesting concept as the cost of the BHR phenotype must be a very steep ¢tness landscape (anti-redundant) which maintains backbone homogeneity (Fig. 1). Any apparent homogeneity observed in the structure of the backbone probably results from the selection for maintenance of each function coupled to periodic selection for the (adaptive or accessory) phenotypic determinants. IncP-1 plasmids appear to have contributed to the evolution of bacterial groups [28], through an essentially redundant mechanism, by providing a vehicle on which parts of catabolic pathways can be carried in the same cells as compatible plasmids carrying the whole pathway [29]. Such a scenario allows gene duplication and the consequent evolution of novel functions. Essentially the ability of plasmids to co-exist is a redundant strategy that
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
182
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
promotes evolution and the multiplication of both partners.
4. How host and MGEs contribute to ¢tness and transient redundancy A decade-long study of plasmids in phytosphere pseudomonads at a rural site has identi¢ed three dominant and persistent plasmid types (Groups I, III and IV) and several more-rare types. These plasmids are most often over 200 kb in size and range up to 400 kb with rearrangements and size variation common within groups [30]. Field studies have found that these three plasmid types are carried by a large number of overlapping pseudomonad populations distributed on both the leaves and roots of a variety of crop, ¢eld margin and grassland plants [30]. Field studies have also found that these plasmids are actively transferring in situ between natural populations of phytosphere pseudomonads and have positive or negative e¡ects on bacterial colonisation of plants according to the time in the season and development of the plant [31,32]. Phytosphere bacteria do not confront the same steep ¢tness landscape encountered by gut or pathogenic bacteria where factors such as continuous growth and antigenic phase variation may be essential for survival. However, the phytosphere environment requires adaptation to changing conditions in which no strain remains persistently dominant and many strains exhibit temporal cycling or £uctuations. The genes for traits associated with periodic exploitation of £uctuating conditions are diverse. While the core of these genes is chromosomally (and relatively stably) located [33], many genes of periodic bene¢t are located in the HGP and on plasmids such as those of Groups I, III and IV. For pseudomonads growing on plants, plasmids in these groups have been observed to have conferred strong positive ¢tness bene¢ts on their hosts in the more mature stages of plant growth [30,34]. This type of resource has been interpreted as a shared pool of adaptive genes, increasing diversity and releasing bacteria from the burden of many genes of only periodic value. However, it is possible that many components of the HGP will persist because they can self-transfer and evade selective pressures that would counter-select less mobile loci. A key question with these plasmids is whether or not they (or the genes they carry) experience a steeper ¢tness landscape than chromosomal loci or whether they face a less harsh (less selective), less antiredundancy landscape and therefore more relaxed conditions for their persistence. Certainly sequencing studies of these phytosphere plasmids (unpublished) con¢rm the more general observation that genes partition between the chromosomal and mobile pools and that these two pools are distinct. Most of the ORFs identi¢ed on one of these plasmids (pQBR103), for example, ¢nd no homologues in the genome sequences of pseudomonads or other bacteria [30].
5. In vivo and genomic studies of intracellular and population level MGE redundancy in rhizobia Rhizobia are facultative nitrogen-¢xing symbionts and tumour-inducing pathogens of plants, belonging to the genera Rhizobium, Sinorhizobium, Mesorhizobium, Bradyrhizobium, Azorhizobium and Agrobacterium. Their agronomic importance is re£ected in the large number of population surveys that have assessed their diversity, distribution and host ranges [35^37]. Both the symbiotic genes and the tumorigenic genes are carried on MGE’s, either plasmids or ‘symbiotic’ islands [35]. In addition to these population studies the genomes of three rhizobia have been determined : Mesorhizobium loti MAFF303066 (NC_002678), Sinorhizobium meliloti 1021 (NC_003047) and Agrobacterium tumefaciens C58 (NC_003305). Several whole plasmids from these bacteria have also been sequenced separately (NC_000914, NC_002147, NC_002377 and NC_002575). In addition, the genome of a very closely related animal pathogen, that of Brucella melitensis, is also available (NC_003317). Together these provide us with a powerful and unique data set with which to approach the question of what factors in£uence the make-up and maintenance of the HGP in rhizobia. Rhizobia have a very di¡erent genomic arrangement compared with E. coli. In general their genomes comprise one chromosome and one or more additional low (unitary) copy number replicons, all of which appear to be derived from an ancestral repABC plasmid replicon [38]. The status, whether considered a chromosome or a plasmid, of the repABC replicons is host-dependent. The secondary chromosome of B. melitensis, the linear replicon of A. tumefaciens and the, non-symbiotic plasmid (pSymA) of S. meliloti all carry essential, house-keeping genes and cannot be cured from the host. Thus these former MGEs appear to have been co-opted to become indispensable ‘secondary’ chromosomes by their host bacteria. However, many rhizobium strains carry two or more repABC replicons that are compatible [6]. Multiple, compatible plasmids in a cell, as already discussed in the context of IncP-1 plasmids, constitutes a potential for genetic redundancy and the evolution of novel pathways. Among the repABC plasmids of rhizobia some carry genes of agronomic interest that are intermittently selected, i.e. those required for the pathogenicity of agrobacteria (pTi) or the legume symbiosis of rhizobia (pSym) that have been the focus of population studies. The large proportions of rhizobia persisting in the soil, lacking pTi or pSym, respectively, e.g. [36], con¢rm the facultative nature, and hence intermittent selection for these traits. However, the known functions only represent a small proportion of the total genetic material not carried on the chromosome. Sinorhizobial plasmids appear to carry a large diversity of genes and functions: 54% of the genes of pNGR234a, a 536-kbp pSym of a closely related Sinorhizobium strain, do
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
not have orthologues (genes that have evolved from a shared common ancestor) in the sequenced S. meliloti genome [39]. The absence of many gene orthologues might indicate that antiredundancy processes are operating, because there is little evidence for gene duplication between genomes. However, Galibert et al. [39] state that 42% of the genes within the sequenced S. meliloti genome are paralogues, i.e. have arisen by gene duplication (although many of these duplications were predicted to be ancient and likely code for di¡erent phenotypes). A substantial number of the plasmid-borne sequences are members of the ABC transporter family. Thus while there is little evidence of duplicate genes per se there is a substantial degree of functional redundancy among the gene families associated with plasmids and this might indicate some level of functional redundancy both within and between replicons. Despite their large size and levels of functional redundancy of the types of gene carried it is apparent that the plasmids of rhizobia are not simply ‘depositories’ or ‘genetic collect-alls’ that acquire unlimited numbers of intermittently selected advantageous traits. In the majority of population studies rhizobia have been isolated from plant nodules: strains carrying more than one symbiosis plasmid have not been isolated. In lab experiments passage of a strain containing two pSyms resulted in loss of the ineffective genes, i.e. nodulation appears to provide a selective bottleneck in rhizobial populations [40]. The selective strength of symbiotic interactions was emphasised in this study ^ one pSym-strain combination that was highly unstable under non-selective laboratory culture was maintained through, and could be isolated from nodules during in planta selection. Population and lab-based studies clearly indicate that plasmid exchanges and genomic rearrangements occur in rhizobia, e.g. identical symbiotic genes have been found associated with di¡erent plasmid replication genes [40]. Recent studies by Mavingui et al. [41] suggest a mechanism for genomic re-assortment mediated via repetitive sequences such as IS elements in Sinorhizobium sp. NGR234. Although the wild-type genomic arrangement of NGR234, comprising one chromosome associated with two plasmids (pNGR234a, 0.54 Mbp and pNGR234b, s 2 Mbp), dominates in laboratory cultures the authors were able to isolate strains with di¡erent genomic organisations. These included strains with only one free plasmid, the other having recombined into the chromosome or with one large replicon due to recombination between the two plasmids. Fitness measurements indicate that these rearrangements have minimal a¡ect on phenotype or ¢tness in the laboratory, although the dominance of the wild-type organisation might suggest that this arrangement is optimal. The authors also identi¢ed some genomes of reduced size, indicating that the rearrangements might provide a stochastic process that could lead to the elimination of genetic redundancy by randomly identifying the best combination of traits associated with minimal burden. If much of the incoming DNA is redun-
183
dant (functionally redundant, i.e. not selected for, or genetically redundant, i.e. gene duplications), then most of it can be lost. The distribution of homologous transfer sequences among Rhizobium leguminosarum ¢eld isolates [42] might indicate that genetic redundancy is counter-selected in this way. Population surveys indicate that the plasmid component of rhizobia is highly mobile, this might suggest that many plasmids would, at least, contain an origin of transfer (oriT) to enable their mobilisation. However, when 12 strains of R. leguminosarum carrying a total of 64 plasmids were screened for oriT sequences, only 17 plasmids hybridised with three homologous but diverged (V40%), cross-hybridising probes, i.e. on average each strain contained ¢ve or six plasmids and of these only one or two carried transfer genes. Transfer operons are large and the host could incur a substantial genetic and metabolic burden by their carriage and expression. Thus removal of duplicate copies on compatible replicons, i.e. loss of the redundant copies, might bene¢t the host by removing these genetic and metabolic burdens. One other feature of the rhizobial genomes is that plasmids and chromosomes (whether real chromosomes or coopted plasmids) have distinct G+C% contents : plasmids sensu stricto always exhibit lower G+C% than do the chromosomes [39,43^45]. This phenomenon is not limited to one genus, the plasmids and phage isolated from Pseudomonas spp. have noticeably lower G+C% content than do their host chromosomes (our unpublished results). Also, in parallel with the rhizobial genomes there are distinct regions around the Pseudomonas aeruginosa chromosome that show low G+C%, these all coincide with regions that may at one time have been part of the HGP, e.g. phage-like sequences and the O-antigen region. Explanations for anomalous G+C content have included the idea that the anomalous genes have been acquired from a distantly related organism that itself has a distinct G+C content. However, this explanation seems somewhat incongruous with the known ecology of both rhizobial plasmids (repABC replicons have so far only been identi¢ed in rhizobia and closely a⁄liated K-proteobacteria) and the pseudomonad plasmids and phages (many of which show restricted host range). The reasons for lower G+C content among genes in the HGP are unclear but it is possible that mobile genes are under di¡erent selection pressures than are chromosomal genes. Establishing whether or not this is the case and how the di¡erences are maintained is a major challenge to understanding bacterial genome organisation.
6. Conclusions/broader implications In this review we have tried to ¢t our knowledge of MGEs into an ecological/evolutionary framework which uses redundancy/antiredundancy as a basic way of classifying genetic characteristics in the context of selective
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
184
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
landscapes of di¡ering severity. By considering intensively studied systems we have been able to show that mutational changes in speci¢c aspects of MGEs or types of MGE can have di¡erent consequences depending on the system and the context which allows generalisations (rules) to be formulated which provide a logic for future classi¢cation. These rules will be useful not just for classi¢cation but also for prediction of the likely population structures which have shaped the evolution of a given MGE and hence how it might best be exploited or eliminated. With respect to environments where plasmids have been studied there seems to be general agreement that enteric bacteria in an animal gut, with constant £ux and relatively rapid growth rate, face extreme pressure and thus a steep selective landscape. This appears to contrast with soil or rhizosphere bacteria that are more often subjected to longer term spatially attenuated selection pressures and weak counter-selection. Dykhausen [46], for example, has proposed that the high diversity observed in soil samples results from speciation (accumulation of diversity) combined with relatively low rates of extinction (counter-selection). This contrast is re£ected in the redundancy/antiredundancy of plasmid types commonly observed in these two gut and soil regimes. For example, plasmids of enteric bacteria carrying antibiotic resistance tend to be smaller, trait-focussed, diverse, transfer pro¢cient often with a high transfer rate in contrast to Sym (symbiosis trait), Tol (toluene degradation) or pseudomonad associated Mer (mercury resistance) plasmids. These types of plasmids are larger, generally less transfer pro¢cient, carry diverse traits, and appear to be a re-assembly of mosaics in response to environmental factors such as host-plant range or xenobiotics [35]. In addition, in the gut the components of the HGP appear to be less stable. A dominant MGE trait in a population can be displaced once selection is removed because of competition between MGE-free and MGE-carrying hosts (e.g. R plasmids). Such patterns imply antiredundant selection processes. Plasmids that are able to spread to many species may balance the cost of continuous, albeit controlled expression of transfer functions by their ability to confer adaptive traits on those bacteria that will become their host. The relatively large HGP of rhizobia, represented by the plasmid component of their genomes, might indicate that his pool (rather than individual plasmids) is more stable. These plasmids may even have redundancy-tolerant mechanisms that facilitate HGP acquisition of key genes for survival and would promote the persistence of redundant DNA. Elements such as Tn’s, IS elements, pathogenicity islands etc. are likely to have redundancy pressures similar to the host but mitigated by the capacity to independently replicate. One can thus start to construct a comprehensive theory relating MGE properties to their ecological context and their potential for promoting adaptation and diversity. By choosing archetypes of di¡erent molecular survival strategies and then determining how they perform in di¡erent eco-
logical contexts using modern techniques of quantitative PCR (to detect presence and copy number), in situ PCR and hybridisation to detect the host species, and array technology to assess genetic response to host and environmental variation, it should be possible to build on this foundation a solid understanding of the HGP.
Note added in proof Suggested reading: Kerr et al. (2002) [Kerr, B., Riley, M.A., Feldman, M.W. and Bohannan, B.J.M. (2002) Local dispersal promotes biodiversity in a real-life game of rock-paper-scissors. Nature 418, 171^174]. These authors provide further experimental evidence supporting the rock, scissor, paper model of Czaran et al. [10].
References [1] Dawkins, R.(1976) The Sel¢sh Gene. Oxford University Press, Oxford. [2] Kobayashi, N. and Bailey, M.J. (1994) Plasmids isolated from the sugar beet phyllosphere show little or no homology to molecular probes currently available for plasmid typing. Microbiology 140, 289^296. [3] Beeson, K.E., Erdner, D.L., Bagwell, C.E., Lovell, C.R. and Sobecky, P.A. (2002) Di¡erentiation of plasmids in marine diazotroph assemblages determined by randomly ampli¢ed polymorphic DNA analysis. Microbiology 148, 179^189. [4] Couturier, M., Bergquist, P.L., Bex, F. and Maas, W.K. (1988) Identi¢cation and classi¢cation of bacterial plasmids. Microbiol. Rev. 52, 375^395. [5] Wilkins, B.M.(1995) Gene transfer by bacterial conjugation : diversity of systems and functional specializations. pp. 59^88. In: Population Genetics of Bacteria, 1st edn. (Baumberg, S., Young, J.P.W., Wellington, E.M.H. and Saunders, J.R., Eds.). Cambridge University Press, Cambridge. [6] Palmer, K.M., Turner, S.L. and Young, J.P.W. (2000) Sequence diversity of the plasmid replication gene repC in the Rhizobiaceae. Plasmid 44, 209^219. [7] Ochman, H., Lawrence, J.G. and Groisman, E.A. (2000) Lateral transfer and the nature of bacterial innovation. Nature 405, 299^304. [8] Lawrence, J.G. and Ochman, H. (2002) Reconciling the many faces of lateral gene transfer. Trends Microbiol. 10, 1^4. [9] Maynard Smith, J., Smith, N.H., O’Rourke, M. and Spratt, B.G. (1993) How clonal are bacteria ? Proc. Natl. Acad. Sci. USA 90, 4384^4388. [10] Cza¤ra¤n, T.L., Hoekstra, R.F. and Pagie, L. (2002) Chemical warfare between microbes promotes biodiversity. Proc. Natl. Acad. Sci USA 99, 786^790. [11] Lenski, R.E. and Riley, M.A. (2002) Chemical warfare from an ecological perspective. Proc. Natl. Acad. Sci. USA 99, 556^558. [12] Krakauer, D.C. and Plotkin, J.B. (2002) Redundancy, anitredundancy, and the robustness of genomes. Proc. Natl. Acad. Sci. USA 99, 1405^1409. [13] Toussaint, A. and Merlin, C. (2002) Mobile elements as a combination of functional modules. Plasmid 47, 26^35. [14] Papadopoulos, D., Schneider, D., Eiss, J.M., Arber, W., Lenski, R.E. and Blot, M. (1999) Genomic evolution during a 10,000-generation experiment with bacteria. Proc. Natl. Acad. Sci. USA 96, 3807^ 3812. [15] Wichman, H.A., Badgett, M.R., Scott, L.A., Bouliann, C.M. and
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
S.L. Turner et al. / FEMS Microbiology Ecology 42 (2002) 177^185
[16] [17]
[18]
[19]
[20]
[21] [22]
[23]
[24]
[25]
[26]
[27]
[28] [29]
[30]
[31]
[32]
[33] [34]
Bull, J.J. (1999) Di¡erent trajectories of parallel evolution during viral adaptation. Science 285, 422^424. Kichler Holder, K. and Bull, J.J. (2001) Pro¢les of adaptation in two similar viruses. Genetics 159, 1393^1404. Bergstrom, C.T., Lipsitch, M. and Levin, B.R. (2000) Natural selection, infectious transfer and the existence of conditions for bacterial plasmids. Genetics 155, 1505^1519. Lilley, A.K., Young, J.P. and Bailey, M.J. (2000) Bacterial population genetics: do plasmids maintain diversity and adaptation? In: The Horizontal Gene Pool: Bacterial Plasmids and Gene Spread (Thomas, C.M., Ed.), pp. 287^300. Harwood Academic Publishers. Messenger, S.L., Moulieux, I.J. and Bull, J.J. (2001) Virulence evolution in a virus obeys a trade-o¡. Proc. R. Soc. Lond. B 266, 397^ 404. Spiers, A.J. and Bergquist, P.L. (1992) Expression and regulation of the RepA protein of the repFIB replicon from plasmid P307. J. Bacteriol. 174, 7533^7541. Briani, F., Deho', G., Forti, F. and Ghisotti, D. (2001) The plasmid status of satellite bacteriophage P4. Plasmid 45, 1^17. Thomas, C.M. (1983) Instability of a high copy number mutant of a miniplasmid derived from broad host range IncP plasmid RK2. Plasmid 10, 184^195. Pansegrau, W., Lanka, E., Barth, P.T., Figurski, D.H., Guiney, D.G., Haas, D., Helinski, D.R., Schwab, H., Stanisch, V. and Thomas, C.M. (1994) Complete nucleotide sequence for Birmingham IncPK plasmids. Compilation and comparative analysis. J. Mol. Biol. 239, 623^663. Kostelidou, K., Jones, A.C. and Thomas, C.M. (1999) Conserved C-terminal region of global repressor KorA of broad host range plasmid RK2 is required for co-operativity between KorA and a second global regulator, KorB. J. Mol. Biol. 289, 211^221. Zatyka, M., Bingle, L., Jones, A.C. and Thomas, C.M. (2001) Cooperativity between KorB and TrbA repressors of broad host range plasmid RK2. J. Bacteriol. 183, 1022^1031. Thorsted, P.B., Macartney, D., Akhtar, P., Haines, T., Ali, N., Davidson, P., Sta¡ord, C., Pocklington, M.J., Wilkins, B.M., Pansegrau, W., Lanka, E. and Thomas, C.M. (1998) Complete sequence of the IncPL plasmid R751: Implications for evolution and organisation of the IncP backbone. J. Mol. Biol. 282, 969^990. Smith, C.A. and Thomas, C.M. (1989) Relationships and evolution of IncP plasmids. In: Promiscuous Plasmids of Gram-negative Bacteria (Thomas, C.M., Ed.), pp. 57^78. Academic Press, London. Thomas, C.M. (2000) Paradigms of plasmid organization. Mol. Microbiol. 37, 485^491. Top, E.M., Moe«nne-Loccoz, Y., Pembroke, T. and Thomas, C.M. (2000) Phenotypic traits conferred by plasmids. In: The Horizontal Gene Pool : Bacterial Plasmids and Gene Spread (Thomas, C.M., Ed.), pp. 249^285. Harwood Academic Publishers. Bailey, M.J., Rainey, P.B., Zhang, X-X. and Lilley, A.K. (2002) Population dynamics, gene transfer and gene expression in plasmids, the role of the horizontal gene pool in local adaptation at the plant surface. In: Microbiology of Aerial Plant Surfaces (Lindow, S.E. and Elliott, V., Eds.), pp. 171^189. American Phytopathological Society Press. Lilley, A.K. and Bailey, M.J. (1997) The acquisition of indigenous plasmids by a genetically marked pseudomonad population colonising the phytosphere of sugar beet is related to local environmental conditions. Appl. Environ. Microbiol. 63, 1577^1583. Lilley, A.K. and Bailey, M.J. (1997) Impact of pQBR103 acquisition and carriage on the phytosphere ¢tness of Pseudomonas £uorescens SBW25: burden and bene¢t. Appl. Environ. Microbiol. 63, 1584^ 1587. Rainey, P.B. (1999) Adaptation of Pseudomonas £uorescens to the plant rhizosphere. Environ. Microbiol. 1, 243^257. Lilley, A.K., Fry, J.C., Day, M.J. and Bailey, M.J. (1994) In situ
[35]
[36]
[37] [38]
[39]
[40]
[41]
[42]
[43]
[44]
[45]
[46]
185
transfer of an exogenously isolated plasmid between Pseudomonas spp. in sugar beet rhizosphere. Microbiol. 140, 27^33. Sullivan, J.T., Patrick, H.N., Lowther, W.L., Scott, D.B. and Ronson, C.W. (1995) Nodulating strains of Rhizobium loti arise through chromosomal symbiotic gene transfer in the environment. Proc. Natl. Acad. Sci. USA 92, 8985^8989. Segovia, L., Pinero, D., Palacios, R. and Marinez-Romero, E. (1991) Genetic structure of a soil population of nonsymbiotic Rhizobium leguminosarum. Appl. Env. Microbiol. 57, 426^433. Young, J.P.W. (1996) Phylogeny and taxonomy of rhizobia. Plant and Soil 186, 45^52. Moreno, E. and Moriyo¤n, I. (2002) Brucella melitensis : A nasty bug with hidden credentials for virulence. Proc. Natl. Acad. Sci. USA 99, 1^3. Galibert, F., Finan, T.M., Long, S.R., Puehler, A., Abola, P., Ampe, F., Barloy, H.F., Barnett, M.J., Becker, A., Boistard, P., Bothe, G., Boutry, M., Bowser, L., Buhrmester, J., Cadieu, E., Capela, D., Chain, P., Cowie, A., Davis, R.W., Dreano, S., Federspiel, N.A., Fisher, R.F., Gloux, S., Godrie, T., Go¡eau, A., Golding, B., Gouzy, J., Gurjal, M., Hernandez, L.I., Hong, A., Huizar, L., Hyman, R.W., Jones, T., Kahn, D., Kahn, M.L., Kalman, S., Keating, D.H., Kiss, E., Komp, C., Lelaure, V., Masuy, D., Palm, C., Peck, M.C., Pohl, T.M., Portetelle, D., Purnelle, B., Ramsperger, U., Surzycki, R., Thebault, P., Vandenbol, M., Vorhoelter, F.J., Weidner, S., Wells, D.H., Wong, K., Yeh, K.C. and Batut, J. (2001) The composite genome of the legume symbiont Sinorhizobium meliloti. Science 293, 668^672. Rigottier-Gois, L., Turner, S.L., Young, J.P.W. and Armarger, N. (1998) Distribution of repC plasmid-replication sequences among ¢eld isolates of Rhizobium leguminosarum bv. viciae. Microbiology 144, 771^780. Mavingui, P., Flores, M., Guo, X., Da¤vila, G., Perret, X., Broughton, W.J. and Palacios, R. (2002) Dynamics of genome architecture in Rhizobium sp. strain NGR234. J. Bacteriol. 184, 171^176. Turner, S.L., Knight, K.A.L. and Young, J.P.W. (2002) Identi¢cation and analysis of rhizobial plasmid origins of transfer. FEMS Microbiol. Lett., submitted. Kaneko, T., Nakamura, Y., Sato, S., Asamizu, E., Kato, T., Sasamoto, S., Watanabe, A., Idesawa, K., Ishikawa, A., Kawashima, K., Kimura, T., Kishida, Y., Kiyokawa, C., Kohara, M., Matsumoto, M., Matsuno, A., Mochizuki, Y., Nakayama, Nakazaki, N., Shimpo, S., Sugimoto, M., Takeuchi, C., Yamada, M. and Tabata, S. (2000) Complete genome structure of the nitrogen-¢xing symbiotic bacterium Mesorhizobium loti. DNA Res. 7, 331^338. Wood, D.W., Setubal, J.C., Kaul, R., Monks, D.E., Kitajima, J.P., Okura, V.K., Zhou, Y., Chen, L., Wood, G.E., Almeida, Jr.N.F., Woo, L., Chen, Y., Paulsen, I.T., Eisen, J.A., Karp, P.D., Bovee Sr., D.., Chapman, P., Clendenning, J., Deatherage, G., Gillet, W., Grant, C., Kutyavin, T., Levy, R., Li, M.J., McClelland, E., Palmieri, A., Raymond, C., Rouse, G., Saenphimmachak, C., Wu, Z., Romero, P., Gordon, D., Zhang, S., Yoo, H., Tao, Y., Biddle, P., Jung, M., Krespan, W., Perry, M., Gordon-Kamm, B., Liao, L., Kim, S., Hendrick, C., Zhao, Z.Y., Dolan, M., Chumley, F., Tingey, S.V., Tomb, J.F., Gordon, M.P., Olson, M.V. and Nester, E.W. (2001) The genome of the natural genetic engineer Agrobacterium tumefaciens C58. Science 294, 2317^2323. Goodner, B., Hinkle, G., Gattung, S., Miller, N., Blanchard, M., Qurollo, B., Goldman, B.S., Cao, Y., Askenazi, M., Halling, C., Mullin, L., Houmiel, K., Gordon, J., Vaudin, M., Iartchouk, O., Epp, A., Liu, F., Wollam, C., Allinger, M., Doughty, D., Scott, C., Lappas, C., Markelz, B., Flanagan, C., Crowell, C., Gurson, J., Lomo, C., Sear, C., Strub, G., Cielo, C. and Slater, S. (2001) Genome sequence of the plant pathogen and biotechnology agent Agrobacterium tumefaciens C58. Science 294, 2323^2328. Dykhuizen, D.E. (1998) Santa Rosalia revisited : Why are there so many species of bacteria ? Antonie-van-Leeuwenhoek 73, 25^33.
Downloaded from https://academic.oup.com/femsec/article-abstract/42/2/177/595055 by guest on 22 January 2018
FEMSEC 1426 5-11-02
