Horizontal Gene Transfer in Osmotrophs: Playing with Public Goods

ORE Open Research Exeter

TITLE Horizontal gene transfer in osmotrophs: playing with public goods.

AUTHORS Richards, Thomas A; Talbot, Nicholas J.

JOURNAL Nat Rev Microbiol

DEPOSITED IN ORE 19 November 2014

This version available at

http://hdl.handle.net/10871/15898

Open Research Exeter makes this work available in accordance with publisher policies.

A NOTE ON VERSIONS

The version presented here may diﬀer from the published version. If citing, you are advised to consult the published version for pagination, volume/issue and date of publication PERSPECTIVES

however, not unique to fungi. Many bacteria, OPINION for instance, feed in an analogous man ner, and other eukaryotic groups, such as Horizontal gene transfer in hyphochytriomycetes (FIG. 1c) and oomycetes (FIG. 1d) (sometimes collectively termed the pseudofungi26), also feed osmotrophically osmotrophs: playing with public and adopt filamentous growth habits, allow ing invasive growth in heterogeneous sub goods strates. Importantly, these eukaryotes also lost the ability to carry out phagotrophy and Thomas A. Richards and Nicholas J. Talbot became obligately osmotrophic26,27. Osmotrophy has a number of distinct Abstract | Osmotrophic microorganisms, such as fungi and oomycetes, feed advantages as a feeding strategy. External by secreting depolymerizing enzymes to process complex food sources in the digestion of large and complex polymers extracellular environment, and taking up the resulting simple sugars, micronutrients allows greater control over substances that and amino acids. As a consequence of this lifestyle, osmotrophs engage in the are allowed to enter a cell (FIG. 2a), thus acquisition and protection of public goods. In this Opinion article, we propose that minimizing potential routes of infection and intake of harmful substances. Furthermore, horizontal gene transfer (HGT) has played a key part in shaping both the repertoire osmotrophy allows greater fidelity in nutri of proteins required for osmotrophy and the nature of public goods interactions in ent acquisition, such that the repertoire of which eukaryotic microorganisms engage. digestive enzymes and uptake transporters expressed by an osmotrophic species can Osmotrophic microorganisms feed by secret the fossil record, fungi colonized the ter be altered to match a particular need, like ing extracellular depolymerizing enzymes restrial environment more than 400 million the colonization of a new substrate28–30. into the environment to degrade complex years ago, in close association with early However, osmotrophy also carries some polymers, such as cellulose, lignin and land plants6–9. Extant fungi occupy diverse disadvantages or risks, such as the utilization proteins, and by transporting the resulting ecosystems and undertake a wide variety of secreted enzymes by competitors within a simple, monomeric sugars and amino acids of interactions, ranging from highly mutu community and the loss of derived nutrient into their own cells. Fungi, oomycetes and alistic symbioses to devastating diseases of sources to neighbouring microorganisms many bacteria feed by osmotrophy, and in both animals and plants10. Fungi are diverse (FIG. 2b,c) or by diffusion30–32. An important using this trophic mechanism, these organ in terms of species complexity11–17, gene consequence of osmotrophy is therefore isms have become the principal degraders repertoire and biochemical capabilities18–24, that microorganisms which use this trophic of biomass in most terrestrial ecosystems, as and form a range of cell types, including mechanism must engage in producing, well as important pathogens of plants and unicellular yeasts, motile flagellated zoo protecting and acquiring public goods33,34. animals. One of the most important conse spores and polarized multicellular hyphae. Thus, osmotrophy-associated genes pre quences of osmotrophy is the participation They also form specialized feeding struc dominantly encode secreted depolymerizing of microorganisms in both competitive tures, such as appressoria (FIG. 1a), haustoria enzymes and cognate transporter proteins and cooperative public goods interactions, and rhizoid structures (FIG. 1b), which can go (FIG. 2), as well as genes associated with toxin because the food sources of these organ on to form complex tissues within multi production or detoxification. Toxin produc isms reside outside the cells even as these cellular fruiting bodies12. It is clear that tion can be used by osmotrophs to exclude foods are being processed, and are therefore early in the diversification of fungi, the competitors, thereby protecting public goods available to others. In this Opinion article, ability to carry out phagocytosis — a mode from being used by others (for examples, see we focus primarily on microbial eukaryotes of feeding that generates private goods — REFS 35,36). The acquisition or reconfigura and propose that genes encoding proteins was lost, and the overwhelming majority tion of these traits is likely to be favourable, associated with osmotrophy have been of extant fungi are instead dependent on as both allow microorganisms to colonize important in the evolutionary ecology and osmotrophy13, a process that generates both new environments and/or make use of addi public goods interactions of these micro public and club goods. tional metabolites, but also equip microor organisms. We also present evidence that Osmotrophy has proved to be a success ganisms with new capabilities with which to horizontal gene transfer (HGT) has had a ful feeding strategy for fungi in particular, engage in social competition (for example, major role in reconfiguring osmotrophic and has allowed them to colonize diverse stealing bacterial siderophores30). functions in fungi and oomycetes. These heterogeneous terrestrial environments results are part of a growing body of evi where nutrients are plentiful but largely Osmotrophs and public goods interactions dence which suggests that diverse evolution inaccessible to most competitors because As discussed above, the production of ary mechanisms operating on genomes have they take the form of complex biological osmotrophic phenotypes represents a cost shaped social interactions across a range of molecules such as cellulose and lignin (for to the individual, because a fraction of the microorganisms1–5. example, within leaf litter or soil) or the cel protein produced or food digested might be lulose- and protein-rich tissues of plant and lost to competitors32,34. As a result, selection Osmotrophic eukaryotic microorganisms animal hosts, respectively. As a consequence, will tend to favour individuals that stop or The largest group of osmotrophic eukary fungi have evolved to become important minimize their production of public goods otic microorganisms, in terms of biodiver decomposers of biomass in most terres but still make use of public goods manu sity, is the kingdom Fungi. According to trial ecosystems25. Obligate osmotrophy is, factured by others. Cooperative behaviour

720 | OCTOBER 2013 | VOLUME 11 www.nature.com/reviews/micro

31 a b and glucose outside the cell , where a suite of hexose transporters are then utilized to recover extracellular sugars29,41,45. This osmo trophic process can result in the loss (into the environment) of up to 99% of the mono saccharides produced by invertase activity31 (but see REF. 46 for a scenario suggesting that this 99% is not all lost). The production of extracellular enzymes such as invertase clearly carries a cost to the organism, and it is therefore likely that selection favours indi viduals which have lost their invertase activity but still survive as cheats in a population of c d invertase secretors (a scenario analogous to that set out in the Black Queen hypothesis1). Alternatively, acquisition of activities such as invertase production might confer on a microbial lineage the ability to colonize new environments and utilize previously untapped food sources. Complex patterns of gain and loss of the genes encoding secreted depolymerizing enzymes and cognate nutri ent uptake transporters are thus likely to occur within osmotrophic lineages. It is Figure 1 | The osmotrophic lifestyle in eukaryotic microorganisms. a | Scanning electron micro- therefore clear to see how a process such as graph (SEM) of Magnaporthe oryzae, showing appressoria, which areNature specialized Reviews structures | Microbiology for the HGT might provide a means for such traits invasion of plant tissue. b | Light micrograph of sporangia and rhizoid structures of Pseudorhizidium to rapidly move within a population of endosporangiatum (isolate JEL 221)106. c | SEM of a fraction of a clonal multicellular assembly (or myce- different species. lium) of the pseudofungus Hyphochytrium catenoides, showing hyphal growth and interconnections. d | SEM of Phytophthora infestans attacking the leaf surface of a host plant. Scale bars represent 10 μm. Osmotrophic and social evolution Part b image courtesy of J. Longcore, University of Maine, USA; part d image courtesy of S. Kamoun, There have been considerable advances The Sainsbury Laboratory, UK. in our understanding of social inter actions in microbial systems (for reviews, see REFS 33,34). These interactions have can therefore give way to the emergence of — those organisms that possess a compat been interpreted mainly using classic cheats within a population34,37. It is generally ible uptake system (a cognate transporter social evolution theory as set out by W. D. assumed that functions are maximized if protein) to take up the nutrient (FIG. 2b,c). Hamilton34,47,48, which posits that individu there is extensive cooperation (sometimes In this way, possession of a cognate trans als gain inclusive fitness through improving referred to as a ‘conspiracy of doves’), but in porter might be restricted to individuals of the reproductive success of closely related some populations, a mixture of cooperators the same species or clonal group, providing individuals of the same species. Hamilton’s and cheats can be optimal38 — although even one mechanism for kin selection. However, rule therefore predicts that higher levels of in these cases, cheats and cooperators are possession of the cognate transporter by an public goods interactions will occur when still in conflict. As a consequence of such organism of the same species that does not individuals are within the same species conflict, it is important that osmotrophs cooperate by helping to produce the public group and closely related. The kin selection acquire nutrients rapidly, thereby making goods, or by a member of a different species hypothesis for public goods interactions public goods inaccessible to others, or that that does not contribute to generation of has received support from experimental the organisms protect released food sources the public goods, would allow such cheats manipulations of populations of the bacte by excluding competitors from the local to proliferate in the population (FIG. 2c). rial pathogen Pseudomonas aeruginosa, for environment. Adaptation to these pressures Acquiring transporters would therefore be instance, demonstrating that higher levels of can take the form of utilization of high- predicted to be a means by which an indi siderophore production (a public goods trait affinity nutrient transporters39–43, modifica vidual could cheat within a population of for iron acquisition) evolve in communities tion of the public goods produced so that osmotrophs. with higher genetic relatedness49. However, it they can be acquired by only a subset of the Similarly, the secretion of enzymes that is worth noting that the spatial distribution of microbial community3,44, or (as mentioned generate public goods also determines these interactions and the relative density of above) exclusion of competitors by toxin which microorganisms provide the ‘work’ microbial communities are important factors production35,36. within a population (FIG. 2b,c). For example, for understanding the fate of both cheats and Although some metabolites generated the secretion of invertase by Saccharomyces cooperators46,49–52. This is because the spatial by extracellular digestion are likely to be cerevisiae is an osmotrophic phenotype and distribution of a species is connected to the universally accessible (glucose or phosphate, a popular model for investigating public likelihood that related individuals will be in for instance), others can be available to only goods interactions31,32. Invertase catalyses close proximity to each other and will there a subset of the population (club goods) the hydrolysis of sucrose to liberate fructose fore benefit from public goods produced by

NATURE REVIEWS | MICROBIOLOGY VOLUME 11 | OCTOBER 2013 | 721

a Figure 2 | Competitors and collaborators in an osmotrophic ecosystem. a | A schematic polarized cell showing adaptation to an osmotrophic function. The close-up shows a view of this polarized cell in the process of degradative enzyme secretion (step 1), breakdown or diges- 3 tion of complex molecules in the extracellular environment (step 2) and selective transport of broken-down metabolites back into the cell (step 3). These are the key functions associated with osmotrophic feeding. b | Two cells collabo- rating, with each cell producing a different set of enzymes that release different club goods from which both cells can profit because both possess cognate transporter proteins. c | A situation in which one cell (right) is doing the work to make public goods available, whereas both cells are 2 1 making use of the public goods because they possess the cognate transporter. The cell on the left could be described as a defector, or even a cheat.

their neighbouring kin. Microorganisms that make clonal patches (for example, fungi that b develop hyphal networks or yeast colonies) produce situations in which the proximity of kin can favour public goods-producing behaviours46,50. However, there is evidence that this is not a simple rule: low levels of dis tribution (high viscosity) can increase local competition between relatives of the same species, meaning that intermediate dispersal (intermediate viscosity) might in many cases produce optimal situations for the evolu tion of public goods traits53–55. Consistent with this idea, artificial alterations in spatial structure among mycorrhizal fungi has been shown to favour cooperative fungal species56. In these cases, kin selection operates within species, family or clonal groups, but the term ‘relatedness’ in these contexts often refers to whether the interacting microor ganisms share compatible alleles for public goods traits, rather than average genome c similarity, although generally the two are likely to be correlated unless HGT or differ ential gene loss1,2 has played a part. We note, however, that the species concept is difficult to apply to many microbial groups, includ ing many fungi, that reproduce by asexual means and for which ‘species groups’ can contain considerable diversity in terms of sequence variation, morphology and genome content57–59. This makes the concept of kin selection equally difficult to apply to these microorganisms, and ideas like spe cies, genetic relatedness and kin selection are therefore fluid in many microbial groups and are dependent on the distribution of certain genes rather than on patterns of vertical or familial ancestry (that is, true relatedness).

Nature Reviews | Microbiology 722 | OCTOBER 2013 | VOLUME 11 www.nature.com/reviews/micro

Cooperation without relatedness In this scenario, the production of public in natural heterogeneous communities66, Public goods interactions are difficult to goods is dependent on a fixed or threshold and such a pattern is indeed consistent with explain in systems for which genetic related number of cooperators. In such social dilem observations from one study, in which it was ness is low, such as communities composed mas, each individual in a community would found that 12% of S. cerevisiae and 10% of of many different species, or situations in rather avoid the cost of volunteering alone Saccharomyces paradoxus yeast cells sampled which there is no kin discrimination. Many and will seek to exploit public goods pro from natural and industrial environments community interactions involving osmo duced by others. However, if a threshold of do not cooperate by secreting invertase, trophic microorganisms such as fungi and participants is not reached, public goods are suggesting that invertase cheats are main pseudofungi fall into this category, as these not produced, and every individual in the tained at a low relative frequency in these organisms are often found in heterogeneous community then pays a cost that is higher heterogeneous populations73. The volunteer’s communities16,17,25 such as those in soil and than that of volunteering (the so‑called dilemma might therefore be significant in leaf litter. ‘volunteer’s dilemma’ (REFS 67,69)). This explaining how osmotrophs, such as fungi Given the high heterogeneity of these key feature of the volunteer’s dilemma cor and pseudofungi, engage in public goods communities, it is also important to consider responds well with osmotrophic functions production in the heterogeneous microbial alternative mechanisms that can support because in many cases a threshold level communities in which they live. cooperative behaviours, such as those that of enzyme secretion (work) is required to In reality, however, natural communities enhance public goods production. These sustainably digest a complex food source, are complex and dynamic, so it is unlikely mechanisms can include reciprocal coop implying that a certain level of participa that one factor alone (be it kin selection, erative behaviour60,61, or punishment of tion or a certain number of participants is a community density, repeated interactions, cheats62,63. As these features require repeated prerequisite. A good example of this would reciprocal cooperation, punishment or interactions or stable microbial popula be the fungal degradation of wood, which the volunteer’s dilemma) is the continu tions, public goods production is therefore, requires threshold levels of heterogeneous ous driver of public goods interactions. It theoretically, sustainable only in stable populations of both lignin-degrading white is therefore difficult to weigh the relative microbial communities composed of closely rot fungi and cellulose-degrading brown rot importance of these factors, and a dynamic related individuals (such as family groups fungi to facilitate breakdown71,72. mixture of multiple factors should be con or individuals of the same species), which The volunteer’s dilemma should therefore sidered as potential drivers of public goods is true for certain microbial ecosystems, maintain cheats at a low relative frequency interactions in osmotrophic communities. like hyphal networks, yeast colonies46,50 and biofilms, but less so for the communi Glossary ties in more heterogeneous environments. Furthermore, repeated interactions with Appressoria Kin selection 60,61 Specialized infection cells that are used by Selection that favours traits because of their beneficial reciprocal cooperative behaviour or plant-pathogenic fungi to penetrate the host plant surface effects on the fitness of relatives. 62,63 punishment (sanction) of cheats are using mechanical force and/or enzymatic action to breach interactions that are generally more relevant the cuticle. Osmotrophic microorganisms for animals, although there are examples of Microorganisms that take up digested or dissolved Cheats nutrients by osmosis, often facilitated by transporter both in non-animal systems. For example, Individuals within a community that do not carry out proteins to allow molecules to cross the cell membrane. manipulations of mycorrhizal symbiosis cooperative behaviours (or that minimize their cooperation) have demonstrated reciprocal cooperative but derive benefit from the work of others. Phagotrophy behaviours as the plant amends the contri A process governed by the cytoskeleton and involving bution of fixed carbon and the arbuscular Club goods membrane and cytoplasmic manipulation to engulf large Public goods that are accessible to select individuals only particles or other cells for nutrition. mycorrhizal fungus amends the production (‘members of the club’) in the community. of phosphate, thereby rewarding partners Private goods that make ‘better’ contributions to mutual Cooperators Biological or chemical resources that are produced by an ism64. Punishment of cheats is evident in Individuals that provide benefit to others. individual and can be used only by that individual. plant–rhizobium mutualisms, in which Haustoria Public goods bacteria that do not produce fixed nitrogen Specialized fungal feeding structures that are commonly Biological or chemical resources that are produced by an for their plant partners have been shown to produced by biotrophic fungi and occupy living plant cells individual in a community and can be used by all other have reduced reproductive success, poten by invagination of the plant plasma membrane. individuals in the community. tially because of reduced oxygen supply 65 Horizontal gene transfer Relatedness from the plant host . In this case, reduced The transfer of genetic material between genomes (for A measure of genetic or genomic similarity. oxygen supply represents a sanction or a example, across species boundaries). Also called lateral punishment. gene transfer. Rhizoid structures However, some theoretical work sug ‘Hair-like’ protruberances of eukaryotic cells that maximize Hyphae the interface between the cell surface and the gests that cooperation can arise when the Cells of a filamentous morphotype, sometimes forming environment. “population density depends on the average branching structures; this morphotype exists for fungi and population payoff” (REF. 66). For example, some other microorganisms. The development of this Siderophores voluntary participation can be an important cellular morphology is governed by the cytoskeleton, with Small-molecule iron-chelating compounds that are factor for driving public goods interactions, growth and trophic activity directed to the hyphal tip. secreted by microorganisms. producing cooperative behaviour without Inclusive fitness Spiteful behaviour spatial population structures and between The result of individual behaviours on the reproductive Behaviour that is costly to both the producer and the groups of different species or subspecies67–70. output of others, weighted by relatedness. recipient.

NATURE REVIEWS | MICROBIOLOGY VOLUME 11 | OCTOBER 2013 | 723

HGT can drive public goods interactions phenotypes are unlikely to be lost, because multiprotein systems. This prediction has HGT has been proposed to provide a means maintenance of the selfish genetic element therefore been referred to as the complexity by which microorganisms can rapidly gain enforces the public goods trait4,79. When hypothesis90. According to this hypothesis, it new biological functions74–77. A recent, but considered together, these studies show how is less likely that HGT of informational genes steadily growing, body of data suggests that HGT can lead to the evolution of coopera will introduce novel biological functions90,91 HGT has had a minor role in terms of total tive phenotypes within microorganisms4, which confer a clear selective advantage gene numbers within microbial eukaryotes, and how HGT can ‘short-circuit’ patterns of to a recipient, and it is also less likely that but a much more important role in terms of kin selection because of the transfer of com an HGT-acquired informational protein phenotypic change in the evolution of osmo patible cooperative traits between different will fit into the functional networks of the trophic microorganisms74–76,78. An interesting species. recipient cell owing to the requirement for hypothesis has also emerged regarding the Following on from these studies, we specific interactions with multiple foreign role of HGT in determining public goods predict that HGT provides one means by proteins90,91. It has recently been shown, interactions in populations of pathogenic which the spread of public goods traits is however, that the key factor determining bacteria5. Using a mathematical model, it facilitated in fungal and oomycete popula the frequency of HGT is not the biological has been predicted that selection can favour tions. However, comparatively little research function of a gene family, but rather the con HGT-mediated reintroduction of public has looked at HGT of genes that encode nectivity of the encoded proteins, in terms of goods-encoding genes into cheats5. transporter proteins from the perspective protein–protein interactions91. Further evidence is provided by a of public goods interactions, for example. We were interested in investigating study that focused on genes encoding HGT of transporter-encoding genes is likely whether genes involved in the production secreted and outer-membrane proteins in to be advantageous for microorganisms that and acquisition of public goods by osmo 21 Escherichia spp., and looked at the role reside in environments where osmotrophy trophic microorganisms also have low levels of HGT in relation to public goods inter is common, because these genes encode the of protein–protein interaction connectiv actions4. These proteins might be significant mechanism by which microorganisms can ity, suggesting that they therefore have a in public goods interactions because they benefit from the availability of previously greater propensity for HGT. We focused on allow different substrates to be utilized, inaccessible public goods. The acquisition of the S. cerevisiae S288c genome, for which act as effectors to subvert host immune new transporter functions by HGT would, a large proportion of gene products have responses and protect the wider microbial for instance, be predicted to confer a strong protein interaction data available from two- community, or contribute to the manufac selective advantage and lead to the evolu hybrid screens92, gene deletion studies93 and ture of biofilms4. Interestingly, genes encod tion of new cheats in a population (FIG. 2c). a collated database of physical and genetic ing only 3% of the secreted proteins and 6% Consistent with this idea, we have identified interactions94,95, and we investigated those of the outer-membrane proteins were repre numerous published examples of trans genes (n = 5,084) that have been classified sented in the core Escherichia spp. genome porter-encoding genes that seem to have as operational or informational, or that are (defined in REF. 4 as gene families conserved been transferred between microbial species previously identified HGT candidates76. We in all 21 bacterial genomes analysed). By by HGT39,40,80–85. Key examples include the also classified osmotrophy-associated genes contrast, genes encoding inner-membrane, transfer of genes encoding high-affinity encoding transporter proteins and secreted periplasmic space and cytoplasmic proteins nitrate transporters40,86, ammonium trans enzymes (FIG. 3). We are aware that analysis are represented within the core genome porters87 and fructose transporters88 into of such a data set does not represent a com at much higher proportions (~24%; and between fungi, and the transfer of genes prehensive analysis of protein interactions, P < 0.0001). This result is consistent with encoding sugar and purine or pyrimidine but the yeast protein–protein interaction genes encoding extracellular proteins being transporters from fungi to oomycetes80,89. studies represent the most developed data frequently lost and acquired by HGT, there sets currently available. Interestingly, these fore driving the potential spread of public HGT has shaped public goods interactions data suggest that the connectivity is com goods traits within the wider population4,5. There are two important factors that can paratively low for proteins encoded by the However, such traits are also likely to be influence the frequency of successfully previously identified HGT candidates76 and under differing patterns of selection because transferring a gene function by HGT. The for osmotrophy-associated gene products, of their importance for host interactions, first factor is whether the gene can be tran consistent with the hypothesis that osmotro resulting in a range of distinct selection pres scribed and translated in a foreign cell, and phy-associated genes are likely to be success sures that is likely to drive high rates of gene whether the protein product can fit into ful HGT candidates for trait acquisition. variation, loss, acquisition and replacement. the functional network of that cell. The Next, we collated all published examples In this example, additional biological second factor is whether the gene acquisi of fungal76,87 and oomycete89,96 genes that mechanisms are also important for ensur tion then provides a selective advantage were acquired by HGT and encode secreted ing transfer and maintenance of public goods that leads to maintenance (fixation) of the depolymerizing enzymes (Supplementary traits. This study demonstrates that the gene within the recipient species. The first information S1 (table)) and transporter mobile genetic elements responsible for consideration suggests that HGT will occur proteins (Supplementary information S2 the transfer of genes associated with public at a higher frequency for ‘operational’ genes, (table)). These lists are derived from a larger goods traits can enforce cooperative traits by which encode housekeeping functions or collated list of previously published and creating ‘cellular addiction’ processes such as individual enzymes, than for ‘informational’ revalidated HGT studies in fungi (a total of toxin–antitoxin and restriction–modification genes, which are involved in transcription, 323 genes76,87), recent studies focusing on systems4. In this way, cheats within a popu translation and signal transduction. This is HGT of fructose transporter genes88 and lation can be converted into faithful pub because informational genes encode proteins ammonium transporter genes87 into fungi lic goods producers4;the newly acquired that are typically members of large, complex (ten and two transfers, respectively), and

724 | OCTOBER 2013 | VOLUME 11 www.nature.com/reviews/micro

but has no information to control which Total genome* (n = 5,084) organisms benefit from those goods — an Operational interaction can be sustained if there is a genes (n = 3,745) cost to a secondary agent entering the envi Informational ronment69. If the rewards and the costs to genes (n = 336) each contributor are optimized appropri Unclassiﬁed ately, only preferred agents can enter such genes (n = 1,002) interactions, so that potential mutualistic Osmotrophic transporter genes (n = 91) symbionts or kin are favoured, depending Secreted public goods on their own capabilities. The principal genes (n = 12) contributor can therefore control which Known horizontally organisms can engage in an interaction or transfered genes (n = 19) an environment69. 0 5 10 15 20 25 30 35 40 45 50 The use of anti-competitor strategies, Interactions of encoded proteins such as the secretion of toxins, is a policing Figure 3 | Comparison of protein–protein interaction conmplexity between horizontally action that seems to be common among 35,36 transferred genes and osmotrophic genes in the SaccharomycesNature Reviews cerevisiae | Microbiology S288c microorganisms and is also an example genome. Horizontal gene transfer (HGT) events moving genes into the yeast genome have occurred of a public goods interaction because it is for genes encoding proteins with low levels of connectivity (that is, protein–protein interactions) (blue). a costly behaviour that harms competi Genes classified as functioning in osmotrophic phenotypes, encoding either secreted proteins that tors37. Such ‘spiteful behaviour’ results in the function in public goods production or transporter proteins that function in the uptake of public goods exclusion of competitors, but only if the (both red), also have a low interaction complexity. This is consistent with such gene classes undergoing spiteful organisms have the capability to HGT. For comparison, we also detail the interaction complexity of the yeast genome as a whole, using resist their own spiteful behaviour (for 90 the available interaction data. Following the classification used in the original complexity hypothesis , example, detoxification or tolerance to we also show the interaction complexity for proteins encoded by operational90, informational90 and toxins)35,98. Interestingly, in microorgan unclassified genes. The data are derived from REF. 95. To show the range of complexity, the median and quartile ranges are shown for each category. isms, multiple social traits can be expressed simultaneously99,100 so that public goods production is often combined with spiteful actions such as toxin production35. Many individual studies of HGT into oomycetes Many of the HGT acquisitions listed in fungal and bacterial species, for example, (a further 35 genes89,96), giving 370 genes in Supplementary information S1,S2 (tables) produce a large diversity of toxins and an total. Genes encoding 28 secreted enzymes are therefore likely to be retained by selec array of secondary metabolites18,101, many of and 24 metabolite uptake transporters were tion primarily because they equip the which have no currently known function. identified among the HGT candidates, recipient species with new metabolic capa Toxins can be used to attack host organisms providing evidence that HGT has played a bilities. For example, HGT acquisitions that (by pathogens), to resist predation or to part in the evolution of osmotrophy and of enable oomycetes to breakdown cutin96 or control microorganisms that are competitors public goods interactions in these groups of other structural components of plant cell in the environment101,102. Toxin production microorganisms. It is not possible to state walls89 might have provided new mecha can therefore be seen as a key strategy in with certainty that HGT has made a signifi nisms for these microorganisms to invade the protection of public goods by allowing cant contribution to the total repertoire of their plant hosts. However, these enzymes mutualistic symbionts or kin to prolifer osmotrophy-related gene functions, without also release nutrients at the leaf surface as ate and excluding cheats or competitors functional analyses of the complete pro public goods, thereby affecting microbial in an ecosystem where public goods are teome and more comprehensive evolution interactions within the phylloplane com available101. ary analyses. Such analyses would be further munity. Therefore, although selection for a On the basis of this idea, we investigated complicated because sometimes the relation horizontally transferred gene might not be whether HGT has played a part in recon ship between a single gene and a new pheno driven by public goods interactions initially, figuring the fungal repertoire of toxins and type is not completely straightforward. For these genes nevertheless influence such detoxifying secondary metabolites. There are example, acquisition of a metabolite uptake interactions. 27 published examples of phylogenetic data transporter can compensate for previous loss demonstrating HGT events for the transfer of an entire biosynthetic pathway, as repeat Does HGT shape additional relevant traits? of toxin and detoxifying genes into fungal edly seen for folate biosynthesis and salvage A key prerequisite for any organism genomes76 (Supplementary information S3 (for instance, see REF. 97). However, puta involved in public goods interactions is the (table)) and one additional example identi tive functional annotation of the genes we capacity to colonize the environment in fied using genome content comparisons103 identified (Supplementary information S1,S2 which particular public goods are located. (making 28 in total). Interestingly, of the (tables)) demonstrates that HGT candidates Hamilton suggested that evolution will ten examples of published HGT events include genes which would significantly act to drive diversification so that strate between fungi76,103, four of the events trans expand the nutrient resources available to gies evolve to exclude cheats and favour ferred whole toxin–detoxification gene osmotrophic microorganisms, suggesting kin47,48. Social theory therefore predicts clusters104,105. Furthermore, because many of that the acquisition of these genes would be that when social interactions over public these 28 HGT events involve gene clusters, selectively advantageous, as predicted by the goods are asymmetrical — a situation in these events led to the transfer of a total models illustrated in FIG. 2b,c. which one player produces public goods of 69 individual genes from the above-

NATURE REVIEWS | MICROBIOLOGY VOLUME 11 | OCTOBER 2013 | 725

mentioned list of 323 published gene growing body of work1–5 which links mecha 23. Bushley, K. E. & Turgeon, B. G. Phylogenomics reveals subfamilies of fungal nonribosomal peptide synthetases 76 transfers into fungal genomes . nisms that drive genome variation with the and their evolutionary relationships. BMC Evol. Biol. Arguably, the best example of an HGT evolution of social traits in microorganisms1. 10, 26 (2010). 24. Amselem, J. et al. Genomic analysis of the of this kind is the transfer of the 23-gene Thomas A. Richards and Nicholas J. Talbot are at the necrotrophic fungal pathogens Sclerotinia sclerotiorum sterigmatocystin cluster (which is important School of Biosciences, University of Exeter, and Botrytis cinerea. PLoS Genet. 7, e1002230 (2011). in the manufacture of the toxin sterigmato Geoffrey Pope Building, Exeter EX4 4QD, UK. 25. de Boer, W., Folman, L. B., Summerbell, R. C. & Boddy, L. Living in a fungal world: impact of fungi on cystin, related to the carcinogenic aflatoxins) Thomas A. Richards is also at the soil bacterial niche development. FEMS Microbiol. from the Aspergillus clade to Podospora Department of Zoology, The Natural History Museum, Rev. 29, 795–811 (2005). anserina105, a fungus that often shares the Cromwell Road, London SW7 5BD, UK. 26. Cavalier-Smith, T. in Evolutionary Biology of the Fungi (British Mycological Society Symposia) 339–353 (eds same environment as Aspergillus spp. This Correspondence to N.J.T. Rayer, A. D. M., Brasier, C. M. & Moore, D.,1987). acquisition might have allowed P. anserina to e-mail: [email protected] 27. Cavalier-Smith, T. & Chao, E. E. Phylogeny and megasystematics of phagotrophic heterokonts gain an advantage in excluding other com doi:10.1038/nrmicro3108 (kingdom Chromista). J. Mol. Evol. 62, 388–420 petitors from the environment and allowed Published online 10 September 2013 (2006). 105 28. Bermudez Moretti, M., Perullini, A. M., Batlle, A. & its cohabitation with Aspergillus spp. . HGT- 1. Morris, J. J., Lenski, R. E. & Zinser, E. R. The Black Correa Garcia, S. Expression of the UGA4 gene mediated acquisitions of genes involved in Queen Hypothesis: evolution of dependencies through encoding the δ-aminolevulinic and γ-aminobutyric adaptive gene loss. mBio 3, e00036‑12 (2012). acids permease in Saccharomyces cerevisiae is detoxification metabolic pathways might also 2. Cordero, O. X., Ventouras, L. A., DeLong, E. F. & controlled by amino acid-sensing systems. Arch. aid cheats to subvert kin selection mecha Polz, M. F. Public good dynamics drive evolution of Microbiol. 184, 137–140 (2005). iron acquisition strategies in natural bacterioplankton 29. Boles, E. & Hollenberg, C. P. The molecular genetics of nisms and gain entry to an osmotrophic populations. Proc. Natl Acad. Sci. USA 109, hexose transport in yeasts. FEMS Microbiol. Rev. 21, environment where public goods are avail 20059–20064 (2012). 85–111 (1997). 3. Lee, W., van Baalen, M. & Jansen, V. A. An 30. Heymann, P., Ernst, J. F. & Winkelmann, G. able. In addition, the requirement for fungi to evolutionary mechanism for diversity in siderophore- Identification of a fungal triacetylfusarinine C engage in competition over public goods must producing bacteria. Ecol. Lett. 15, 119–125 (2012). siderophore transport gene (TAF1) in Saccharomyces 4. Nogueira, T. et al. Horizontal gene transfer of the cerevisiae as a member of the major facilitator therefore have been a factor in the diversifica secretome drives the evolution of bacterial superfamily. Biometals 12, 301–306 (1999). tion of fungal secondary metabolism18,23,101, cooperation and virulence. Curr. Biol. 19, 1683–1691 31. Gore, J., Youk, H. & van Oudenaarden, A. Snowdrift (2009). game dynamics and facultative cheating in yeast. consistent with the proposal by Hamilton 5. Smith, J. The social evolution of bacterial pathogenesis. Nature 459, 253–256 (2009). that selection will act to drive diversification Proc. Biol. Sci. 268, 61–69 (2001). 32. Greig, D. & Travisano, M. The Prisoner’s Dilemma and 6. Berbee, M. L. & Taylor, J. W. Dating the molecular polymorphism in yeast SUC genes. Proc. Biol. Sci. 271 of systems that control access to cooperative clock in fungi – how close are we? Fungal Biol. Rev. (Suppl. 3), S25–S26 (2004). benefits, such as public goods47,48. 24, 1–16 (2011). 33. Crespi, B. J. The evolution of social behavior in 7. Selosse, M. A. & Le Tacon, F. The land flora: a microorganisms. Trends. Ecol. Evol. 16, 178–183 phototroph–fungus partnership? Trends Ecol. Evol. (2001). Conclusions 13, 15–20 (1998). 34. West, S. A., Diggle, S. P., Buckling, A., Gardner, A. & 8. Taylor, J. W. & Berbee, M. L. Dating divergences in the Griffin, A. S. The social lives of microbes. Annu. Rev. Osmotrophic microorganisms engage in fungal tree of life: review and new analyses. Mycologia Ecol. Evol. Syst. 38, 53–77 (2007). public goods interactions as a product of 98, 838–849 (2006). 35. Inglis, R. F., Brown, S. P. & Buckling, A. Spite versus 9. Pirozynski, K. A. & Malloch, D. W. The origin of land cheats: competition among social strategies shapes their mechanism of feeding. Therefore, plants: a matter of mycotrophism. Biosystems 5, virulence in Pseudomonas aeruginosa. Evolution 66, the acquisition of secreted depolymer 153–164 (1975). 3472–3484 (2012). 10. Fisher, M. C. et al. Emerging fungal threats to animal, 36. Riley, M. A. & Wertz, J. E. Bacteriocins: evolution, izing enzymes, transporter proteins and plant and ecosystem health. Nature 484, 186–194 ecology, and application. Annu. Rev. Microbiol. 56, metabolic pathways associated with toxin (2012). 117–137 (2002). 11. Hawksworth, D. L. The fungal dimension of 37. West, S. A., Griffin, A. S. & Gardner, A. Social biogenesis and detoxification is likely to have biodiversity: magnitude, significance and conservation. semantics: altruism, cooperation, mutualism, strong been pivotal in the diversification of osmo Mycol. Res. 95, 641–655 (1991). reciprocity and group selection. J. Evol. Biol. 20, 12. Hibbett, D. S. et al. A higher-level phylogenetic 415–432 (2007). trophic microorganisms. Consequently, the classification of the Fungi. Mycol. Res. 111, 509–547 38. MaClean, R. C., Fuentes-Hernandez, A., Greig, D., acquisition of such traits by HGT is likely (2007). Hurst, L. D. & Gudelj, I. A mixture of “cheats” and 13. James, T. Y. et al. Reconstructing the early evolution “co‑operators” can enable maximal group benefit. to have been under strong selective pres of Fungi using a six-gene phylogeny. Nature 443, PLoS Biol. 8, e1000486 (2010). sure, as it allows recipients to spread to new 818–822 (2006). 39. Galeote, V. et al. FSY1, a horizontally transferred gene 14. Richards, T. A., Jones, M. D. M., Leonard, G. & in the Saccharomyces cerevisiae EC1118 wine yeast environments and utilize new food sources. Bass, D. Marine fungi: their ecology and molecular strain, encodes a high-affinity fructose/H+ symporter. A review of published reports of HGT can diversity. Annu. Rev. Mar. Sci. 4, 495–522 (2012). Microbiology 156, 3754–3761 (2011). 15. Buée, M. et al. 454 Pyrosequencing analyses of forest 40. Slot, J. C. & Hibbett, D. S. Horizontal transfer of a didates demonstrates that genes encoding soils reveal an unexpectedly high fungal diversity. New nitrate assimilation gene cluster and ecological secreted enzymes (28 HGT events or genes), Phytol. 184, 449–456 (2009). transitions in fungi: a phylogenetic study. PLoS ONE 2, 16. Jumpponen, A. & Jones, K. L. Massively parallel 454 e1097 (2007). transporters (24 HGT events or genes) and sequencing indicates hyperdiverse fungal communities 41. Brown, C. J., Todd, K. M. & Rosenzweig, R. F. toxin biosynthesis systems (28 HGT events in temperate Quercus macrocarpa phyllosphere. New Multiple duplications of yeast hexose transport Phytol. 184, 438–448 (2009). genes in response to selection in a glucose-limited transferring a total of 69 genes) have all 17. O’brien, H. E., Parrent, J. L., Jackson, J. A., Moncalvo, environment. Mol. Biol. Evol. 15, 931–942 (1998). been subjected to HGT into osmotrophic J.‑M. & Vilgalys, R. Fungal community analysis by 42. Wei, H. et al. A putative high affinity hexose large-scale sequencing of environmental samples. transporter, hxtA, of Aspergillus nidulans is induced in eukaryotic microorganisms, such as fungi Appl. Environ. Microbiol. 71, 5544–5550 (2005). vegetative hyphae upon starvation and in ascogenous and oomycetes (see Supplementary informa 18. Keller, N. P., Turner, G. & Bennett, J. W. Fungal hyphae during cleistothecium formation. Fungal Genet. Biol. 41, 148–156 (2004). REFS 87,88 secondary metabolism — from biochemistry to tion S1–S3 (tables) and ). This genomics. Nature Rev. Microbiol. 3, 937–947 (2005). 43. Javelle, A., Andre, B., Marini, A.‑M. & Chalot, M. High- group of 121 HGT candidates predicted to 19. Stajich, J. E. et al. The fungi. Curr. Biol. 19, R840–R845 affinity ammonium transporters and nitrogen sensing (2009). in mycorrhizas. Trends Microbiol. 11, 53–55 (2003). function in osmotrophic and public goods 20. Soanes, D. M. et al. Comparative genome analysis 44. Smith, E. E., Sims, E. H., Spencer, D. H., Kaul, R. & phenotypes represents more than 32% of of filamentous fungi reveals gene family expansions Olson, M. V. Evidence for diversifying selection at the 76,87,88 associated with fungal pathogenesis. PLoS ONE 3, pyoverdine locus of Pseudomonas aeruginosa. the 370 collated HGTs into fungi and e2300 (2008). J. Bacteriol. 187, 2138–2147 (2005). oomycetes89,96. We therefore propose that 21. Cornell, M. J. et al. Comparative genome analysis 45. Kruckeberg, A. L. The hexose transporter family of across a kingdom of eukaryotic organisms: Saccharomyces cerevisiae. Arch. Microbiol. 166, HGT has played a part in the generation, specialization and diversification in the fungi. 283–292 (1996). protection and acquisition of public goods Genome Res. 17, 1809–1822 (2007). 46. Koschwanez, J. H., Foster, K. R. & Murray, A. W. 22. Chiang, Y. M., Lee, K. H., Sanchez, J. F., Keller, N. P. & Sucrose utilization in budding yeast as a model for the by osmotrophic eukaryotes. This hypothesis Wang, C. C. Unlocking fungal cryptic natural products. origin of undifferentiated multicellularity. PLoS Biol. 9, and the observations reported here fit into a Nat. Prod. Commun. 4, 1505–1510 (2009). e1001122 (2011).

726 | OCTOBER 2013 | VOLUME 11 www.nature.com/reviews/micro

47. Hamilton, W. D. The genetical evolution of social 72. Hastrup, A. C. S. et al. Differences in crystalline 93. Jeong, H., Mason, S. P., Barabasi, A. L. & Oltvai, Z. N. behaviour, I. J. Theor. Biol. 7, 1–16 (1964). cellulose modification due to degradation by brown and Lethality and centrality in protein networks. Nature 48. Hamilton, W. D. The genetical evolution of social white rot fungi. Fungal Biol. 116, 1052–1063 (2012). 411, 41–42 (2001). behaviour, II. J. Theor. Biol. 7, 17–52 (1964). 73. Maclean, C. R. & Brandon, C. Stable public goods 94. Stark, C. et al. BioGRID: a general repository for 49. Griffin, A. S., West, S. A. & Buckling, A. Cooperation cooperation and dynamic social interactions in yeast. interaction datasets. Nucleic Acids Res. 34, and competition in pathogenic bacteria. Nature 430, J. Evol. Biol. 21, 1836–1843 (2008). D535–539 (2006). 1024–1027 (2004). 74. Garcia-Vallve, S., Romeu, A. & Palau, J. Horizontal 95. Cotton, J. A. & McInerney, J. O. Eukaryotic genes 50. Nadell, C. D., Foster, K. R. & Xavier, J. B. Emergence gene transfer of glycosyl hydrolases of the rumen of archaebacterial origin are more important than of spatial structure in cell groups and the evolution fungi. Mol. Biol. Evol. 17, 352–361 (2000). the more numerous eubacterial genes, irrespective of cooperation. PLoS Comput. Biol. 6, e1000716 75. Gojkovic, Z. et al. Horizontal gene transfer promoted of function. Proc. Natl Acad. Sci. USA 107, (2010). evolution of the ability to propagate under anaerobic 17252–17255 (2010). 51. Nowak, M. A. & May, R. M. Evolutionary games and conditions in yeasts. Mol. Genet. Genom. 271, 96. Belbahri, L., Calmin, G., Mauch, F. & Andersson, J. O. spatial chaos. Nature 359, 826–829 (1992). 387–393 (2004). Evolution of the cutinase gene family: evidence for 52. Hauert, C. & Doebeli, M. Spatial structure often 76. Richards, T. A., Leonard, G., Soanes, D. M. & lateral gene transfer of a candidate Phytophthora inhibits the evolution of cooperation in the snowdrift Talbot, N. J. Gene transfer into the fungi. Fungal Biol. virulence factor. Gene 408, 1–8 (2008). game. Nature 428, 643–646 (2004). Rev. 25, 98–110 (2011). 97. de Crecy-Lagard, V., El Yacoubi, B., de la Garza, R. D., 53. Kummerli, R., Griffin, A. S., West, S. A., Buckling, A. & 77. Jain, R., Rivera, M. C., Moore, J. E. & Lake, J. A. Noiriel, A. & Hanson, A. D. Comparative genomics Harrison, F. Viscous medium promotes cooperation in Horizontal gene transfer accelerates genome of bacterial and plant folate synthesis and salvage: the pathogenic bacterium Pseudomonas aeruginosa. innovation and evolution. Mol. Biol. Evol. 20, predictions and validations. BMC Genomics 8, 245 Proc. Biol. Sci. 276, 3531–3538 (2009). 1598–1602 (2003). (2007). 54. Kummerli, R., Gardner, A., West, S. A. & Griffin, A. S. 78. Gabaldon, T. & Huynen, M. A. Reconstruction of the 98. Lehmann, L., Bargum, K. & Reuter, M. An Limited dispersal, budding dispersal, and cooperation: proto-mitochondrial metabolism. Science 301, 609 evolutionary analysis of the relationship between an experimental study. Evolution 63, 939–949 (2003). spite and altruism. J. Evol. Biol. 19, 1507–1516 (2009). 79. Lawrence, J. G. Microbial evolution: enforcing (2006). 55. Brockhurst, M. A., Buckling, A. & Gardner, A. cooperation by partial kin selection. Curr. Biol. 19, 99. Harrison, F. & Buckling, A. Siderophore production Cooperation peaks at intermediate disturbance. Curr. R943–R945 (2009). and biofilm formation as linked social traits. ISME J. 3, Biol. 17, 761–765 (2007). 80. Richards, T. A., Dacks, J. B., Jenkinson, J. M., 632–634 (2009). 56. Verbruggen, E. et al. Spatial structure and interspecific Thornton, C. R. & Talbot, N. J. Evolution of filamentous 100. Williams, P., Winzer, K., Chan, W. C. & Camara, M. cooperation: theory and an empirical test using the plant pathogens: gene exchange across eukaryotic Look who’s talking: communication and quorum mycorrhizal mutualism. Am. Nat. 179, E133–E146 kingdoms. Curr. Biol. 16, 1857–1864 (2006). sensing in the bacterial world. Phil. Trans. R. Soc. B (2012). 81. Richards, T. A. et al. Phylogenomic analysis 362, 1119–1134 (2007). 57. Doolittle, W. F. Microbial evolution: stalking the wild demonstrates a pattern of rare and ancient horizontal 101. Fox, E. M. & Howlett, B. J. Secondary metabolism: bacterial species. Curr. Biol. 18, R565–567 (2008). gene transfer between plants and fungi. Plant Cell 21, regulation and role in fungal biology. Curr. Opin. 58. Boenigk, J., Ereshefsky, M., Hoef-Emden, K., Mallet, J. 1897–1911 (2009). Microbiol. 11, 481–487 (2008). & Bass, D. Concepts in protistology: species definitions 82. Noll, K. M. & Thirangoon, K. Interdomain transfers 102. Trejo-Estrada, S. R., Paszczynski, A. & Crawford, D. L. and boundaries. Eur. J. Protistol. 48, 96–102 (2012). of sugar transporters overcome barriers to gene Antibiotics and enzymes produced by the biocontrol 59. Taylor, J. W. et al. Phylogenetic species recognition expression. Methods Mol. Biol. 532, 309–322 (2009). agent Streptomyces violaceusniger YCED‑9. J. Ind. and species concepts in fungi. Fungal Genet. Biol. 31, 83. Hellborg, L., Woolfit, M., Arthursson-Hellborg, M. & Microbiol. Biotechnol. 21, 81–90 (1998). 21–32 (2000). Piskur, J. Complex evolution of the DAL5 transporter 103. Friesen, T. L. et al. Emergence of a new disease as a 60. Trivers, R. L. The evolution of reciprocal altruism. family. BMC Genomics 9, 164 (2008). result of interspecific virulence gene transfer. Nature Q. Rev. Biol. 46, 37–57 (1971). 84. Gomolplitinant, K. M. & Saier, M. H. Jr. Evolution of Genet. 38, 953–956 (2006). 61. Axelrod, R. & Hamilton, W. D. The evolution of the oligopeptide transporter family. J. Membr. Biol. 104. Patron, N. J. et al. Origin and distribution of cooperation. Science 211, 1390–1396 (1981). 240, 89–110 (2011). epipolythiodioxopiperazine (ETP) gene clusters in 62. Clutton-Brock, T. H. & Parker, G. A. Punishment in 85. Nesbo, C. L., Nelson, K. E. & Doolittle, W. F. filamentous ascomycetes. BMC Evol. Biol. 7, 174 animal species. Nature 373, 209–216 (1995). Suppressive subtractive hybridization detects (2007). 63. Fehr, E. & Gächter, S. Altruistic punishment in extensive genomic diversity in Thermotoga maritima. 105. Slot, J. C. & Rokas, A. Horizontal transfer of a large humans. Nature 415, 137–140 (2002). J. Bacteriol. 184, 4475–4488 (2002). and highly toxic secondary metabolic gene cluster 64. Kiers, E. T. et al. Reciprocal rewards stabilize 86. Slot, J. C., Hallstrom, K. N., Matheny, P. B. & Hibbett, between fungi. Curr. Biol. 21, 134–139 (2011). cooperation in the mycorrhizal symbiosis. Science D. S. Diversification of NRT2 and the origin of its fungal 106. Powell, M. J., Letcher, P. M. & Longcore, J. E. 333, 880–882 (2011). homolog. Mol. Biol. Evol. 24, 1731–1743 (2007). Pseudorhizidium is a new genus with distinct zoospore 65. Kiers, E. T., Rousseau, R. A., West, S. A. & 87. McDonald, T. R., Dietrich, F. S. & Lutzoni, F. Multiple ultrastructure in the order Chytridiales. Mycologia Denison, R. F. Host sanctions and the legume– horizontal gene transfers of ammonium transporters 105, 496–507 (2013). rhizobium mutualism. Nature 425, 78–81 (2003). ammonia permeases from prokaryotes to eukaryotes: 66. Hauert, C., Holmes, M. & Doebeli, M. Evolutionary towards a new functional and evolutionary Acknowledgements games and population dynamics: maintenance of classification. Mol. Biol. Evol. 29, 51–60 (2012). T.A.R. is a European Molecular Biology Organization Young cooperation in public goods games. Proc. Biol. Sci. 88. Coelho, M. A., Gonçalves, C., Sampaio, J. P. & Investigator and a Leverhulme Early Career Fellow, and his 273, 2565–2570 (2006). Gonçalves, P. Extensive intra-kingdom horizontal gene research group is supported by grants from the Moore 67. Archetti, M. & Scheuring, I. Coexistence of transfer converging on a fungal fructose transporter Foundation, the UK Natural Environment Research Council cooperation and defection in public goods games. gene. PLoS Genet. 9, e1003587 (2013). and the Biotechnology and Biological Sciences Research Evolution 65, 1140–1148 (2011). 89. Richards, T. A. et al. Horizontal gene transfer Council (BBSRC). N.J.T. is a European Research Council 68. Raihani, N. J. & Bshary, R. The evolution of facilitated the evolution of plant parasitic mechanisms Advanced Investigator and receives funding from the BBSRC, punishment in n‑player public goods games: a in the oomycetes. Proc. Natl Acad. Sci. USA 108, the Bill and Melinda Gates Foundation and the Halpin Trust. volunteer’s dilemma. Evolution 65, 2725–2728 15258–15263 (2011). The authors thank A. Buckling and the anonymous reviewers (2011). 90. Jain, R., Rivera, M. C. & Lake, J. A. Horizontal gene for comments on this manuscript. 69. Archetti, M. et al. Economic game theory for mutualism transfer among genomes: the complexity hypothesis. Competing interests statement and cooperation. Ecol. Lett. 14, 1300–1312 (2011). Proc. Natl Acad. Sci. USA 96, 3801–3806 (1999). The authors declare no competing financial interests. 70. Hauert, C., De Monte, S., Hofbauer, J. & Sigmund, K. 91. Cohen, O., Gophna, U. & Pupko, T. The complexity Volunteering as Red Queen mechanism for hypothesis revisited: connectivity rather than function cooperation in public goods games. Science 296, constitutes a barrier to horizontal gene transfer. Mol. 1129–1132 (2002). Biol. Evol. 28, 1481–1489 (2011). SUPPLEMENTARY INFORMATION 71. Floudas, D. et al. The Paleozoic origin of enzymatic 92. Uetz, P. et al. A comprehensive analysis of protein– See online article: S1 (table) | S2 (table) | S3 (table) lignin decomposition reconstructed from 31 fungal protein interactions in Saccharomyces cerevisiae. ALL LINKS ARE ACTIVE IN THE ONLINE PDF genomes. Science 336, 1715–1719 (2012). Nature 403, 623–627 (2000).

NATURE REVIEWS | MICROBIOLOGY VOLUME 11 | OCTOBER 2013 | 727