PERSPECTIVE PERSPECTIVE Evolution of microbial markets Gijsbert D. A. Wernera,1, Joan E. Strassmannb, Aniek B. F. Ivensc,d, Daniel J. P. Engelmoera, Erik Verbruggene, David C. Quellerb, Ronald Noëf,g, Nancy Collins Johnsonh, Peter Hammersteini, and E. Toby Kiersa aDepartment of Ecological Science, Vrije Universiteit Amsterdam, 1081 HV, Amsterdam, The Netherlands; bDepartment of Biology, Washington University in St. Louis, St. Louis, MO 63130; cTheoretical Biology Group, Centre for Ecological and Evolutionary Studies, University of Groningen, 9700 CC, Groningen, The Netherlands; dLaboratory of Insect Social Evolution, The Rockefeller University, New York, NY 10065; eInstitut für Biologie, Plant , Freie Universität Berlin, 14195 Berlin, Germany; fFaculté de Psychologie, Université de Strasbourg et Ethologie Evolutive, Département Ecologie, Physiologie et Ethologie, Centre National de la Recherche Scientifique, 67087 Strasbourg Cedex, France; gNetherlands Institute of Advanced Studies, 2242 PR, Wassenaar, The Netherlands; hSchool of Earth Sciences and Environmental Sustainability and Biological Sciences, Northern Arizona University, Flagstaff, AZ 86011-5694; and iInstitute for Theoretical Biology, Humboldt University, 10115 Berlin, Germany

Edited by Richard E. Lenski, Michigan State University, East Lansing, MI, and approved December 12, 2013 (received for review September 10, 2013)

Biological market theory has been used successfully to explain cooperative behavior in many animal species. Microbes also engage in cooperative behaviors, both with hosts and other microbes, that can be described in economic terms. However, a market approach is not traditionally used to analyze these interactions. Here, we extend the biological market framework to ask whether this theory is of use to evolutionary biologists studying microbes. We consider six economic strategies used by microbes to optimize their success in markets. We argue that an economic market framework is a useful tool to generate specific and interesting predictions about microbial interactions, including the evolution of partner discrimination, hoarding strategies, specialized versus diversified mutualistic services, and the role of spatial structures, such as flocks and consortia. There is untapped potential for studying the evolutionary dynamics of microbial systems. Market theory can help structure this potential by characterizing strategic investment of microbes across a diversity of conditions.

cooperation | | trade | partner choice

A perspective borrowed from another field of important role in animal and plant evolution, as availability of partners or level of external research often offers insights for generating ecology, and development (7–9). Microbial resources. The emphasis is on the relative new hypotheses and approaches (1). Biolog- mutualisms exhibit the same characteristics value of goods and services exchanged and ical market theory argues that exchanges of as other mutualisms, including how this value changes as the biological resources and services among organisms can among multiple partner lineages and a poten- context changes. This includes potential be analyzed in market terms, with individuals tial to cheat (10). An essential component of situations where services or goods are of so making strategic trading investments to max- understanding these systems is determining little value that it is more beneficial to refrain imize market gains (2, 3). However, it is un- what helps to stabilize cooperation between from trade. known how applicable the theory is across partners over evolutionary time. This is a However, although the market analogy is diverse systems or whether it can be of use large question, with answers that vary across appealing, empirical tests of the theory have to predict how cooperative strategies evolve. mutualistic systems. focused on animal systems (e.g., refs. 20, 29, One area where market theory is particu- A market framework can theoretically be and 31; see Box 1 for more examples). This larly useful is in the study of mutualisms: used to study mutualisms if the following focus is a potential problem because we still cooperative interactions between individuals conditions for biological markets are met do not understand how cognitive abilities of different species. A major question in (Box 1 and refs. 2 and 3): (i)commodities contribute to the of market dy- mutualism research is why organisms pro- (goods or services) must be exchanged be- namics (32, 33). Advocates of biological mar- vide benefits to other individuals at a potential tween individuals; (ii) at least two distinct ket theory argue that cognition is not needed cost to themselves. An important variable classes of traders must exist; (iii)individuals and terms of trade can be explained solely in maintaining mutualisms is the degree to from at least one trader class must be able to by the current value of the potential partner iv which an individual can control the of choose or switch partners; and ( )there and the context of the exchange (34). The its partner (4). When both partners retain must be individual differences in commodity expectation in market theory is on im- control and can actively choose whether or price so that there are opportunities for mediate rewards, not anticipated benefits: not to cooperate, exchange processes are more “outbidding” price competition. Addition- analogous to human markets because trade ally, most markets are also characterized (leverage) operates over force (coercion) (5). by (v) temporal variation in supply and de- This paper was conceived at the workshop Cooperation in Multi- Partner Settings: Biological Markets and Social Dilemmas at the Mutualisms involving microbial partners mand of the commodities, which can initi- Lorentz Center in Leiden, The Netherlands, January 16–20, 2012.

are relatively underexplored because actions ate price fluctuations (2, 3). Author contributions: G.D.A.W., J.E.S., A.B.F.I., D.J.P.E., E.V., D.C.Q., and responses can be harder to observe. An important advantage of a market frame- R.N., N.C.J., P.H., and E.T.K. wrote the paper.

However, every species on earth is involved work is that strategic investment on both The authors declare no conflict of interest. directly or indirectly in one or more micro- sides of a mutualism can be studied (mul- This article is a PNAS Direct Submission. bial partnerships; some are involved in tiplayer models), while taking into account 1To whom correspondence should be addressed. E-mail: g.d.a. hundreds (6). These interactions play an the biological context of the exchange, such [email protected].

www.pnas.org/cgi/doi/10.1073/pnas.1315980111 PNAS | January 28, 2014 | vol. 111 | no. 4 | 1237–1244 Downloaded by guest on September 29, 2021 understanding of microbial markets and of Box 1. Characteristics of biological markets microbial cooperation in general. This includes (i) Exchange of commodities potential practical applications of microbial markets, for instance, in human health or The most basic element of any (biological) market is the exchange of commodities sustainable agriculture (see Applied Microbial between two interacting partners. Commodities can be resources such as nutrients (11) Markets below).Weidentifyanddiscusssix and services such as grooming (12). In some primate species, mating opportunities are economic strategies important for microbes exchanged for grooming services (13, 14). Aphids exchange honeydew, a nutritious re- to optimize their success in potential biolog- source, for ant-protection services (15, 16). Although biological markets typically involve ical markets: (i) avoid bad trading partners; the exchange of different commodities, primate-grooming markets are an example in (ii) build local business ties; (iii) diversify or which the same commodity is provided by both partners (17, 18). specialize; (iv) become indispensable; (v)save for a rainy day; and (vi) eliminate the com- (ii) Distinct classes of traders petition. These strategies are not necessarily Each biological market exchange requires two distinct classes of actors, or trader classes, mutually exclusive: microbes may use multi- both of which have a commodity to offer (2, 3). Trader classes can belong to the same ple strategies and some microbial behaviors species as in mating or cooperative breeding markets (13, 19) or to different species may correspond to aspects of more than one (mutualisms) (20, 21). In mutualistic biological markets, both trader classes have an strategy. In some cases (e.g., becoming indis- advantage over the other class in producing and providing their specialized commodity pensable), it is the coevolution of partners (22, 23). promoting a market, not necessarily an active (iii) Choosing and switching partners strategy driven by one side. Lastly, although the ability of at least one trader class to choose In any market, trading partners are chosen from a number of potential partners. among competing partners is an important Individuals discriminate, choosing a partner based on differences in the price or quality condition for markets to emerge, microbes of the commodity offered (24). Ants choose aphids producing the highest quality of can also use strategies that limit market choice honeydew and remunerate them with high-quality protection services (15, 25). Similarly, in an effort to gain monopolistic control in primate-mating markets, males that offer more grooming are rewarded with more (strategies iv and vi). In all cases, strategies mating opportunities (13, 26). Such economic partner discrimination is the key factor are selected because they benefit the indi- driving market dynamics (24, 27) and distinguishes biological markets from traditional vidual microbe, or a co-occurring clone of two-party, direct-reciprocity models (28). Therefore, only mutualisms that involve identical microbes, not because they opti- discrimination among partners by one or both trader classes can be considered mize the market as a whole: choices bene- biological markets. ficial to individual traders can have the short- term effect of restricting options available to iv ( ) Price differences a trading partner. Even though such strategies Analogous to human markets, competition between individuals within a trading can initially limit market dynamics, they class can drive down commodity prices in so-called “outbidding competition” (3). In might be restored when new competitors human economies, this idea manifests as “price wars.” For instance, cleaner fish pro- offering the same service enter the exchange. vide a higher-quality cleaning service in the presence of other competing cleaner fish Microbial Market Strategies (29, 30). (v) Supply and demand variation Strategy One: Avoid Bad Trading Part- ners. Among the most important require- Supply and demand determine the bartering value of commodities exchanged (2). ments for a biological market to emerge is for Supply and demand are not fixed quantities but are inherently context-dependent and individuals to be able to discriminate among change over time. For instance, if more food-providing vervet monkeys are introduced in a variety of partners (Box 1, condition iii). a population, the grooming services received as payment decrease (31). Discrimination is the act of evaluating and comparing trading partners based on their actual (or perceived) contribution to an in- advantageous responses can evolve without artificially manipulating investment (e.g., teraction (24). This is essential when indi- cognition (35). ref. 38), now allow us to study cooperation, viduals interact with multiple partners because If biological markets do not require cog- and more specifically market dynamics, in it allows for simultaneous comparisons of nition, we predict that the theory can be ef- microbes (10). quality among competitors. For instance, us- fectively extended to organisms without ner- In this Perspective, we explore the evolu- ing a biological market framework, researchers vous systems, like microbes. Have microbes tion of market strategies by microbes. We have demonstrated how client fish will switch evolved strategies that allow them to dis- only consider systems in which at least one of to better quality cleaners if they are cheated by criminate among trading partners? Do mi- the two trading partners is microbial. We a cleaner that feeds on their mucus or scales crobial mutualism markets follow the laws discuss microbial market research questions, (20, 29, 30). of supply and demand? Although some such as the evolution of microbial partner Experimental research in plant–microbe pioneering modeling work has been done discrimination and the role of environmental mutualisms has tested similar ideas, asking in plant–microbial mutualisms (11, 22, 36, variability on microbial cooperation, and whether plant hosts can detect variation in re- 37), these advances have been almost purely propose methods to address these questions sources or services provided and respond ac- theoretical. However, technological advan- (Table 1). Experimental and theoretical study cordingly. Such discrimination mechanisms ces, such as creating cheating mutants or of such questions will drive forward our have been found in legumes, with some

1238 | www.pnas.org/cgi/doi/10.1073/pnas.1315980111 Werner et al. Downloaded by guest on September 29, 2021 Table 1. Research questions relating to the evolution of microbial biological markets PERSPECTIVE Evolution of Question Suggested approaches

Variation in What is individual variation in market prices? Descriptive; study microbial individuality partner quality Is it influenced by level of competition? Manipulate level of competition; study change in partner quality variation Can it be used in an applied way? Identify best genotypes for systems useful to humans; use experimental evolution to generate variation

Partner Can microbes simultaneously evaluate partners and Track nutrients in networks/aggregates of microbes interacting with discrimination discriminate? multiple partners For free-living microbes, can (chemotactic) mobility achieve Choice experiments between high- and low-quality partners discrimination? Is discrimination more likely in multicellular than in Descriptive studies; compare occurrence of partner discrimination single-celled microbes? Does spatial structure promote discrimination? Analyze preference for quality partners in consortia/ vs. in free-living microbes

Markets under Do trading partners respond to volatility according Manipulate resources; track effects on microbial trade volatility to market hypotheses? What (a)biotic factors influence supply/demand of resources Manipulate conditions to increase/decrease value of a resource; track and affect trade patterns? effects on trade

Hoarding Do microbes hoard resources? Track resource storage, particularly in “farmers” Do microbes trade hoarded resources at a later stage? Manipulate market conditions so trade is not beneficial; change supply/demand to see trade increase.

species preferentially supporting rhizobial tions in market conditions such as external the stability of microbial markets depends symbionts that provide more fixed N2 for nutrient availability (51). on spatially structured environments (54). hosts (39–41) and in the arbuscular mycor- An illustration of the importance of spatial rhizal mutualism in which host plants allo- Strategy Two: Build Local Business Ties. structure for a potential microbial market was cate more carbon resources to fungal strains In a market, your local environment deter- found in a recent experiment in which co- providing more nutrient resources (42, 43). mines your trade connections. Simultaneous operation was evolved de novo in cocultures However, can microbes also discriminate comparisons are not always possible to of two bacterial species (Salmonella enterica among their hosts? In the mycorrhizal mu- identify the best partners. Nonfilamentous ser. Typhimurium and )(55). tualism, an individual (i.e., a single organisms, such as most , have fewer In this system, Salmonella facilitated the multicellular genotype) can potentially com- opportunities than filamentous microbes like growth of its partner by secreting a costly pare resource allocation of several hosts si- mycorrhizal fungi to interact simultaneously amino acid, whereas E. coli reciprocated multaneously because it is connected un- with multiple partners. If direct discrimina- through processing accumulating waste. Pref- derground to multiple plants (44, 45). tion among trading partners is more difficult, erential reciprocation of cooperative genotypes Research suggests that mycorrhizal fungi can market conditions still arise? was found to only occur in a spatially struc- can evaluate the quality of these hosts and Theoretical work suggests that spatial tured (solid) environment, leading to in- adjust resource allocation patterns accord- structuring of partners can limit the spread of creased densities of both species within 10 ingly, allowing for market dynamics to less beneficial partners and favor the evolu- generations. In contrast, in unstructured en- emerge (43, 46, 47). It remains an open tion of cooperative exchanges (52, 53). This is vironments (liquid media), cooperation did question how this information is integrated because spatial structure offers a way to keep not evolve. This illustrates how novel syn- over the fungal network and over what dis- benefits close to cooperators, rather than thetic markets can arise, given the appro- tances this evaluation can take place. diffusing to others. Similarly, spatial structure priate selective conditions. Discrimination, which involves choosiness can also potentially promote the origin and If being spatially aggregated is individually and searching, is generally associated with maintenance of microbial markets. Market beneficial, physical structures may evolve in costs (27). This implies that for discrimina- theory predicts that the success and stability which trade is conducted and maintained. tion ability to evolve there must be (i)suffi- of markets will depend on the extent to which In a mutualism between sulfur-reducing, cient variation in trader quality and (ii)a cooperators gain more access to resources chemotrophic bacteria and phototrophic sul- fitness benefit for partnering with higher- than noncooperators (3). Spatial structure phide oxidizers, spatially organized aggregates quality individuals. Additionally, if discrimi- facilitates the creation of physical connections called consortia are formed (56, 57), poten- nation occurs before actual investment in the to profitable partners, limits diffusion of costly tially allowing for more controlled resource interaction, the potential problem faced by substances, and reduces the number of po- flow (56, 58). Likewise, cross-feeding mutual- a discriminating trading partner is that of tential partners to evaluate compared with isms between hydrogen-producing fermenters “false advertising” (i.e., if a candidate partner well-mixed environments. As a result, spatial and methanogens (e.g., in sewage sludge) form honestly reflects its value as partner) (48, 49). structure could increase the effectiveness aggregative flocs to retain shared metabolites Recent research addresses how fungi sense of partner discrimination, particularly for (54).Consortiaandflocsarecommoninmi- differences in quality (50) and has used nonfilamentous microbes, where such dis- crobial communities (58, 59). Such aggrega- phenotypic plasticity models to ask how crimination is otherwise more difficult. tions may even represent the starting point trading strategies are influenced by fluctua- Consequently, a major question is whether for mutualisms in which one organism

Werner et al. PNAS | January 28, 2014 | vol. 111 | no. 4 | 1239 Downloaded by guest on September 29, 2021 eventually compartmentalizes the other in a workers themselves with broad spectrum Strategy Four: Become Indispensable. In specific structure. protection against infectious fungi, including economics, if a partner is indispensable, the However, the effect of structure on mi- entomopathogenic (e.g., Metarhizium aniso- price it receives is predicted to be high. This crobial exchange is not straightforward (60); pliae) infections (69). This dual service can is because by restricting choice an economic benefits of increased spatial structure can increase ant incentives to maintain bacterial actor can monopolize market exchange (14, sometimes be negated by the negative effects communities even in the absence of specific 73). From a biological market perspective, of competition among relatives (61, 62). A parasitic threats or direct immune challenges. monopolistic control of vital partner func- second problem is that markets can be par- In contrast, specialized trade occurs when tions, such as reproduction, by microbial asitized by third parties that use metabolites a trading partner provides only a single ser- mutualists increases their market value. Part- without contributing to the exchange (63). In vice or good and is predicted to be favored ners are less likely or able to discontinue trade biological markets, traders are predicted to when there is strong competition within because this could result in negative fitness evolve mechanisms to avoid losses to non- a trader class (i.e., when many competing consequences and even death. Although in- trading partners. This is particularly impor- traders are offering the same commodity). dispensability cannot always be considered tant in resource-poor environments (64). This is because organisms often experience indicative of an active strategy used by can potentially be limited if tradeoffs in the acquisition of various re- microbes to command higher prices (i.e., microbes are able to create direct (physical) sources (22, 23). If there is such a tradeoff, it can result from a coevolutionary process connections to trade resources. Metal- a specialized microbe offering only one re- involving both host and microbe or two reducing bacteria like Geobacter sulfurre- source would be able to offer a better price microbes), the end result is a monopolistic ducens form pili that function as microbial for that resource than a microbe offering position. As in economic markets, microbial nanowires (65). These wires probably allow multiple resources. In human economics, this partners still face risks of being undercut the transfer of electrons to other Geobacter “law of comparative advantage” was first by another microbe offering a better price. species over manifolds the size of bacteria proposed two centuries ago by David Therefore, the best examples of microbes (66, 67). Although very little is known Ricardo, arguing that it is beneficial to spe- using this strategy involve tight physical about these transfers, this is intriguing be- cialize in the acquisition of one commodity ties, for example, those associated with en- cause it suggests a potential mechanism to and obtain the other through trade (70). In dosymbiosis (e.g., refs. 74 and 75), which direct resources (e.g., electrons) to specific a recently discovered exchange between a ni- limit the potential for partner switching. If microbial partners. Other forms of (tem- trogen-fixing unicellular cyanobacterium and it is difficult or impossible to switch to an- porary) adhesion between partners may a tiny single-celled prymnesiophytes algal other microbe from the environment, the likewise direct resource transfer to high- partner (71), the cyanobacterium has lost its prevailing partner would occupy a monopo- quality trading partners. More research is ability to carry out its own listic position and be indispensable. needed to understand how the potential for and receives carbon from this specific algal In some cases, hosts have relied on mi- directed transfers depends on the microbial partner in exchange for fixed N. It is un- crobial mutualisms to such an extent that this environment, for example, the presence of known whether there is a tradeoff between N dependency has driven an evolutionary loss dense microbial biofilms vs. simple well- fixation and photosynthesis in this cyano- of host traits (76). For example, several genes mixed environments. bacterium. However, this case raises the in- involved in amino acid biosynthesis have triguing question of whether through the loss been lost in the pea aphid Acyrthosiphon Strategy Three: Diversify or Specialize. of photosynthesis, this cyanobacterium can pisum. These are encoded in the genome of In market terms, diversification in com- fix N more efficiently and consequently offer Buchnera,theaphid’s symbiont (77). Corals modities (i.e., offering multiple services or a better price than other symbiotic cyano- (Acropora sp.) have lost the ability to syn- goods rather than only a single one) is a bacteria or generalist planktonic species com- thesize cysteine, because this is provided by strategy that could reduce exposure to price peting for algal carbon. their endosymbionts (78). Hundreds of ma- volatility. In theory, diversification reduces How can market theory generate hypoth- rine worm species show loss of traits for di- risks associated with volatile market con- eses on the evolution of specialized vs. di- gestive or excretory systems because of their ditions, thereby facilitating more consistent versified services in microbial markets? One evolved dependence on chemosynthetic mi- performance under a wider range of con- idea is that high dependence on a partner crobes for digestion (79). The microbes ditions. In contrast, through commodity spe- promotes diversification. By studying or even benefit from protection and are provided cialization, a trader can usually offer a good at synthetically creating microbial trade markets with a suitable , whereas the host is a lower price, increasing its chances to outbid (e.g., ref. 72) with partners displaying differ- able to survive on inorganic energy and car- competitors (Box 1, condition iv) (11, 22, 23). ent levels of dependency and competition, we bon sources, allowing them to flourish in In microbial markets, the individual ad- maybeabletotestempiricallywhenthe harsh conditions (80). As a result of vital trait vantages of using diversified or specialized benefits of diversification outweigh those of loss, these microbial symbionts become fully strategies will depend strongly on context specialization. A second question is whether indispensable to their hosts. An important (11, 37). Diversification of services or com- diverse services arise as byproducts or whether research question, although empirically chal- modities can be favored if this strategy in- microbial partners are under selection to lenging, is to test whether indispensable creases the range of environmental condi- vary benefits to hosts? The costs of providing microbial partners command higher prices tions over which a trader remains a useful diverse benefits will be affected by tradeoffs compared with symbioses where the host is partner. For instance, the actinobacterial associated with their provisioning. Di- not fully dependent on its microbial partners. symbionts of leaf-cutter ants produce anti- versification is predicted to emerge if com- This effect would only be predicted to occur biotics that inhibit Escovopsis, a parasitic modities provided by one partner are in cases where the potential for switching to fungus invading fungal gardens (68). Recent positively correlated, whereas specialization another microbial partner is limited. Without work suggests these symbionts provide a would be favored by a tradeoff in the ac- such limitations, indispensability of indi- more diversified service by also providing ant quisition of two commodities (22, 23). vidual microbes driving up prices could be

1240 | www.pnas.org/cgi/doi/10.1073/pnas.1315980111 Werner et al. Downloaded by guest on September 29, 2021 counteracted by outbidding competition by smaller one (e.g., refs. 82 and 86). Future The mutualism between insect-killing PERSPECTIVE microbes from the environment. work should determine whether volatility in Xenorhabdus bacterial symbionts and ento- market demand or in resource mopathogenic Steinernema nematodesisa Strategy Five: Save for a Rainy Day. are indeed factors promoting hoarding of potentially useful system to study competi- Careful management of resources is impor- resources for future trade. tion in a market framework (95–97). Xen- tant for traders engaged in market exchange. orhabdus bovienii use toxins to kill insects This can mean saving resources for future Strategy Six: Eliminate the Competition. in which the nematode reproduces. When trade, rather than consuming or trading im- Although direct elimination of competition is nematodes migrate to find new insect prey, mediately. Humans and some other animals a deviation from biological market assump- they take only a few clones of X. bovienii practice such management in agriculture or tions, antagonism against other individuals in (97), and thus competition for migrating food storage where a portion of material is one’s trader class can help secure favorable nematode hosts is intense. X. bovienii maintained rather than consumed (81, 82), trade conditions for the individual using clones use bacteriocins to ensure access so it can be propagated for subsequent gen- them (2, 3). For example, monopolistic con- to migrating host resources (95–97). Initial erations, or sold later at a better price. We trol of chacma baboon mating markets work on field populations suggests that, expect such hoarding of market resources for through competitor suppression by alpha in line with predictions (e.g., ref. 89), future trade if three conditions apply: (i)the males has been observed (14). When costs bacteriocins most strongly mediate intra- supply of the resource is volatile to such an of directly suppressing competitors are small specific competition but that they also extent that there are periods of luxury supply or competition is intense, these types of an- play a role in mediating coexisting spe- (i.e., the trader can afford not to use it all); tagonistic strategies are likely to be favored cies (97). Because of intense within-trader (ii) there is volatility in demand from the over outbidding competition. Coercion rather class competition in this system, competi- other trader class; and (iii)itisstorable. than leverage becomes the dominant force tion is expected to be strong and antag- Microbes can also practice prudent man- driving trade. However, if competitors are onistic strategies to control market access agement of resources and even increase their completely eliminated, market dynamics will arelikelytoevolve. capital, like some animals. For example, there disappear because the only trader left be- is evidence that rather than providing hosts comes a monopolist. Discussion and Perspectives with phosphorus, some mycorrhizal fungal To eliminate competitors directly and se- Although research on microbial social inter- strains can store the phosphorus nutrients cure greater market control, microbes could actions has greatly expanded in recent years, they have collected or keep them in a form use a variety of strategies, including bacte- there has been a limited attempt to study inaccessible to hosts (43, 83). Such specula- riocin production (52, 87, 88). Bacteriocins microbial exchange systems using a market tion with resources potentially allows a fungal are a class of toxins produced by bacteria to framework. Studies on microbial markets partner to gain a better carbon price for inhibit competing strains. Microbes produce have either been purely theoretical (11, 22, their nutrients, as soil nutrient levels be- bacteriocins that typically target closely re- 23, 36, 37), whereas more empirical work has come depleted. lated rather than distantly related strains only tested these ideas in plant root mutual- The social amoeba Dictyostelium dis- within the same species, because these are isms (40, 42, 43, 49, 98). We argue that bi- coideum has been shown to farm bacteria most likely to compete for similar resources ological market theory can help structure before consuming them. For most of their (89, 90). In this way, competing strains can how we interpret and explain the wide life cycle, D. discoideum consume bacteria via be eliminated, and the bacteriocin producer range of collaborative behaviors performed phagocytosis and divide by binary fission, but can achieve a resource monopoly. To avoid by microbes, generating new hypotheses and under starvation conditions, they aggregate targeting related clones, kin-recognition sys- approaches in social . At the into a multicellular slug that differentiates to tems, like toxin–antidote systems, are often same time, we recognize that it is still un- contain a sterile stalk, as well as hardy spores used (91). For example, E. coli deploys toxins known whether a substantial number of (84). About one-third of D. discoideum clones to eliminate competing strains, while using microbial mutualisms fully conform to the carry bacteria through the normally sterile immunity genes specific to these toxins to conditions for biological markets (Box 1). spore stage and then release the bacteria to protect itself (92). Consequently, it will remain difficult to assess propagate them in an environment where In leaf-cutter fungal gardens, resident ba- the importance of microbial markets before they proliferate. Once the bacteria reach sidiomycete fungi produce incompatibility more research explicitly evaluating trade in sufficient densities in the favorable habitat, compounds that effectively kill intruding fun- microbes is performed. Below we address the amoeba farmers consume the bacterial gal strains, and this antagonistic interac- some specific questions that help answer crop (85). Farming strains devoured fewer tion increases with increasing genetic distance this question. than half the bacteria present, whereas (93).Workerantsalsohelpdistributetoxins Regardless of the actual importance in nonfarmers consumed nearly all of the that prevent alien fungi from establishing. natural conditions, microbial systems are bacteria (85). This illustrates that microbes This is possible because antagonistic com- incredibly amenable to study markets: (i) (whether or not they are active traders) are pounds of fungal origin are ingested by ants resources can be easily manipulated; (ii)re- able to prudently save resources for future when feeding on the resident fungus and be- source exchange can be quantitatively tracked; use, which may be important for microbial come part of the fecal droplets that ants use to (iii) and short generation times mean that markets to evolve. This strategy is most fertilize fungus gardens (94). Application of market changes can be studied on an evo- likely to evolve in microbial markets con- such droplets to alien tissue generally elimi- lutionary time scale (experimental evolu- sisting of a larger partner using a smaller nates alien fungi (93). By eliminating com- tion). These advantages also mean that one (e.g., amoeba–bacteria interactions). peting alien strains, the resident fungus thus is microbial systems can potentially be used to More generally, farming strategies are likely able to monopolize the ant host leaf supply, test longstanding questions in biological to evolve when there is a larger partner that securing a greater share of the resources pro- market theory, including how prices are is dependent on harvesting or milking a vided to the fungal gardens. set in nature, how environmental variation

Werner et al. PNAS | January 28, 2014 | vol. 111 | no. 4 | 1241 Downloaded by guest on September 29, 2021 influences trade strategies, or the impor- most attractive host with the highest number microbial market: when demand is higher, a tance of false vs. honest advertising. An- of microbial partners to choose from. Such higher price can be asked for that service. other valuable contribution will be to study chemotactic partner choice is probably more Microbial systems are useful to study the microbial markets and compare them to difficult to achieve in dense polymicrobial intricacies of this problem. This is of par- animal markets (Box 1). This could help us biofilms than in less dense liquid environ- ticular relevance for supply and demand to determine which, if any, market features ments with clear concentration gradients. To variations induced by climate change and are specific to cognitive agents (34, 35). test market-based partner discrimination in theresponseofmutualismstothesechanges Experimental evolution approaches can be chemotactic mutualisms, choice experiments (112). Another example is in agriculture: the particularly useful for studying microbial could test whether individuals move toward application of fertilizer can reduce the mar- markets because they can circumvent the high-quality trading partners. Experiments ket value of microbial mutualists providing difficulties in inferring adaptations from could run for several generations to see nutrients to plant host and thus drive a re- behaviors exhibited by naturally evolved whether greater discrimination and migration duction in their reliance on mutualistic mi- microbes. Below, we propose some ques- begin to evolve and under which conditions crobial partners (113, 114). tions that warrant particular attention and (for instance, in biofilms vs. free-living con- may be successfully approached using ditions or in filamentous, multicellular mi- The Evolution of Microbial Markets from a microbial market framework (Table 1). crobes vs. in single-celled microbes). Noncooperative Interactions. Evolutionary If partner discrimination becomes so ef- biologists are interested in understanding the Evolution of Variation in Partner Quality. fective that only high-quality genotypes can transitions that drive facultative exchanges An important assumption of biological mar- survive, variations in partner quality could into obligate ones. Microbial markets offer ket theory is that there is sufficient individual ultimately be lost (107, 108). In these fixed a unique opportunity to reenact major tran- variation in partner quality to make choosi- populations, there is no longer variability in sitions using experimental evolution ap- ness worthwhile. Hence, a key step to un- prices, theoretically ending market dynamics. proaches. For example, the recently intro- derstanding the evolution of microbial mar- However, in natural systems, which are duced Black Queen Hypothesis (115, 116) kets is the notion of individual variation in constantly exposed to environmental vol- suggests that the inherent leakiness of many partner quality. Advances in the study of atility, the best-quality genotype can vary microbial public goods results in microbes microbial individuality have revealed that the depending on market conditions. Such con- being selected to lose the ability to produce typical population-level focus in microbiol- text dependence could cause variation in certain goods themselves, as long as these ogy can mask the existence of different phe- preferential allocation of resources to trading goods are provided by helper microbes in the notypes in microbial communities (99, 100). partners, potentially counteracting this loss . This dependence is predicted to Individuality (i.e., cell-to-cell phenotypic var- in diversity. drive the evolution of cooperative iability) of bacteria is an important factor In contrast, there might be microbial trad- in which the beneficiary benefits the helper, driving and evolution (101, ing interactions where the microbial partner sustaining its population (115). In this way, 102). Future research should aim to study cannot achieve partner choice. This could, for obligate microbial cooperative interactions, this phenotypic variation between microbial instance, be the case in endosymbionts that and potentially rudimentary microbial individuals, particularly in terms of trading have no (or only limited) active horizontal markets, can evolve from previously non- partner quality. The study of such variation, transmission. In these cases, the microbial cooperative interactions. Experimental evo- particularly the identification of beneficial market would reduce to a host-controlled lution approaches (e.g., ref. 117), can be strains, is also of practical application for monopoly (e.g., ref. 14). Partner choice useful to understand the loss of traits in mutualistic systems (e.g., rhizobial and my- models might still be useful in these cir- microbial communities and what selection corrhizal mutualisms) that can be better cumstances because variation in host de- pressures favor an increased dependence optimized for human benefit (103). mand for particular microbial commodities on Black Queen markets. may be influenced by environmental con- How Does Partner Discrimination Evolve? ditions, and this could be analyzed with Hoarding and Microbial Prudence. Pru- A key element of any market is that dis- market theory. dence is often associated with advanced cog- crimination among partners is based on the nitive capacities, but it is not a requirement benefit they provide. In mycorrhizal mutu- How Do Markets Evolve Under Fluctua- if direct physiological responses to varying alisms, there is a potential for this to evolve tions in Supply and Demand? One market resource levels are used to achieve real- because individuals are able to simulta- prediction that could be tested in microbial time collaboration (35). In the legume– neously monitor and compare resource flows systems is that the benefit of trade depends rhizobia mutualism, some rhizobial strains to and from different trading partners in not only on the interacting partners but also use a chemical called rhizobitoxine that can a single trading network (43, 46). In free- on the available supply of commodities from substantially increase the accumulation of living microbes, there is less potential for other sources. Analogous to primate markets, carbon [as poly-3-hydroxybutyrate (PHB)] such simultaneous comparisons. However, where the number of infants and fertile while reducing the growth of the legume host microbes can exert control over their location females influences the value they represent and competing rhizobial strains. This helps through chemotaxis (104). If free-living mi- (109, 110), microbial experiments are possi- increase rhizobial reproduction and long- crobes migrate to locations closer to high- ble in which the abundance of a resource is term survival under starvation conditions quality trading partners and then locally re- externally manipulated to change the balance (118, 119). Recent work suggests that the lease their commodity, this can facilitate the of trade costs and benefits. We predict that manner in which PHB resources are hoarded evolution of partner choice. For example, biotic or abiotic conditions that influence the is conditional, with a rhizobial strain pro- chemotactic choice could play a role in cases demand for a particular service [e.g., a rise ducing fewer high-PHB cells when low where microbes migrate toward in temperature if the service is heat resis- competitor density predicts only short-term host root exudates (105, 106), providing the tance (111)] will influence its value on the starvation (118) and can thus be understood

1242 | www.pnas.org/cgi/doi/10.1073/pnas.1315980111 Werner et al. Downloaded by guest on September 29, 2021 as a hoarding strategy in response to and identification of the market conditions standing of microbiology and increase our PERSPECTIVE competition. This system, and others like that drive specialization vs. diversification, knowledge of cooperation in general. Dictyostelium farming (85) and the ecto- while simultaneously taking into account the ACKNOWLEDGMENTS. We thank Kevin Foster and two mycorrhizal soil fungus Morchella crassipes biological context of exchange. In the current Pseudomonas putida anonymous reviewers for very constructive feedback. This farming bacteria as a era of synthetic biology, a microbial market work was supported by US National Science Foundation carbon source (86), can be used to study perspective can increase our understanding of Grants 1204352, 1256416, and 0842327 (to J.E.S., D.C.Q., under what conditions scarce resources and N.C.J.); by a Ubbo Emmius Fellowship, University of the complex feedbacks among partners and Groningen, and a Niels Stensen Fellowship (to A.B.F.I.); are held, so they can be either used later inspire the engineering of novel interactions. and by the Netherlands Organization for Scientific Re- or traded for a greater profit in the future. search Meervoud Grant 836.10.001 and Vidi Grant This will both drive forward our under- 864.10.005 (to E.T.K.). Applied Microbial Markets. Lastly, there is

potential applied value in using biological 1 Burt RS (2004) Structural holes and good ideas. Am J Sociol 25 Fischer MEK, Völkl WO, Hoffmann KLH (2005) Honeydew market theory to help identify the conditions 110(2):349–399. production and honeydew sugar composition of polyphagous black under which trade is maximized for partic- 2 Noë R, Hammerstein P (1994) Biological markets: Supply and bean aphid, Aphis fabae (Hemiptera: Aphididae) on various host demand determine the effect of partner choice in cooperation, plants and implications for ant-attendance. Eur J Entomol 102(2): ular partners. This could have benefits in mutualism and mating. Behav Ecol Sociobiol 35(1):1–11. 155–160. agriculture and medicine. In agriculture, 3 Noë R, Hammerstein P (1995) Biological markets. Trends Ecol Evol 26 Norscia I, Antonacci D, Palagi E (2009) Mating first, mating more: favors microbes that max- 10(8):336–339. Biological market fluctuation in a wild prosimian. PLoS ONE 4(3): 4 Edwards DP, Hassall M, Sutherland WJ, Yu DW (2006) Selection for e4679. imize their own fitness and gain of resources protection in an ant-plant mutualism: Host sanctions, host 27 McNamara JM, Leimar O (2010) Variation and the response to over that of their host or competitors. This modularity, and the principal-agent game. Proc Biol Sci 273(1586): variation as a basis for successful cooperation. Philos Trans R Soc can conflict with the goals of agriculture 595–602. Lond B Biol Sci 365(1553):2627–2633. 5 Johnstone RA, Bshary R (2008) Mutualism, market effects and 28 Nowak MA (2006) Five rules for the evolution of cooperation. or other industries, where humans favor partner control. J Evol Biol 21(3):879–888. Science 314(5805):1560–1563. microbes that maximize yields of their crop 6 Bronstein JL (1994) Our current understanding of mutualism. 29 Adam TC (2010) Competition encourages cooperation: Client hosts or other products (120). Once we have Q Rev Biol 69(1):31–51. fish receive higher-quality service when cleaner fish compete. Anim 7 McFall-Ngai M, et al. (2013) Animals in a bacterial world, a new Behav 79(6):1183–1189. a better understanding of the conditions un- imperative for the life sciences. Proc Natl Acad Sci USA 110(9): 30 Bshary R, Schäffer D (2002) Choosy reef fish select cleaner fish – der which microbes transfer more resources, 3229–3236. that provide high-quality service. Anim Behav 63(3):557 564. we can use experimental manipulations to 8 Berendsen RL, Pieterse CMJ, Bakker PAHM (2012) The rhizosphere 31 Fruteau C, Voelkl B, van Damme E, Noë R (2009) Supply and and plant health. Trends Plant Sci 17(8):478–486. demand determine the market value of food providers in wild vervet – select for microbes offering better market 9 Desbrosses GJ, Stougaard J (2011) Root nodulation: A paradigm monkeys. Proc Natl Acad Sci USA 106(29):12007 12012. prices to their hosts (103). Likewise in medi- for how plant-microbe influences host developmental 32 Suchak M, de Waal FBM (2012) Monkeys benefit from reciprocity pathways. Cell Host Microbe 10(4):348–358. without the cognitive burden. Proc Natl Acad Sci USA 109(38): cine, an increased understanding of trade – 10 West SA, Diggle SP, Buckling A, Gardner A, Griffin AS (2007) The 15191 15196. by pathogenic microbes might be useful social lives of microbes. Annu Rev Ecol Evol Syst 38:53–77. 33 Cheney DL (2011) Extent and limits of cooperation in animals. – to disrupt these markets and promote host 11 Kummel M, Salant SW (2006) The economics of mutualisms: Proc Natl Acad Sci USA 108(Suppl 2):10902 10909. 34 Noë R (2006) Cooperation experiments: Coordination through Optimal utilization of mycorrhizal mutualistic partners by plants. health. Recent work suggests that host part- communication versus acting apart together. Anim Behav 71(1):1–18. Ecology 87(4):892–902. 35 Brosnan SF, Salwiczek L, Bshary R (2010) The interplay of ner choice is important in shaping gut 12 Xia D, et al. (2012) Grooming reciprocity in female tibetan cognition and cooperation. Philos Trans R Soc Lond B Biol Sci composition (121), which has a macaques macaca thibetana. Am J Primatol 74(6):569–579. 365(1553):2699–2710. 13 Gumert MD (2007) Payment for sex in a macaque mating key role in shaping human and animal health 36 Cowden CC, Peterson CJ (2009) A multi-mutualist simulation: market. Anim Behav 74(6):1655–1667. Applying biological market models to diverse mycorrhizal and behavior (122, 123). Do hosts use mar- 14 Clarke PMR, Halliday JEB, Barrett L, Henzi SP (2010) Chacma communities. Ecol Modell 220(12):1522–1533. baboon mating markets: Competitor suppression mediates the ket principles to guide their selection 37 Grman E, Robinson TMP, Klausmeier CA (2012) Ecological potential for intersexual exchange. Behav Ecol 21(6):1211–1220. of microbial partners and optimize their specialization and trade affect the outcome of negotiations in 15 Völkl W, Woodring J, Fischer MK, Lorenz MW, Hoffmann KH mutualism. Am Nat 179(5):567–581. gut microbiota? (1999) Ant-aphid mutualisms: The impact of honeydew production 38 Gore J, Youk H, van Oudenaarden A (2009) Snowdrift game and honeydew sugar composition on ant preferences. Oecologia dynamics and facultative cheating in yeast. Nature 459(7244): Conclusion 118(4):483–491. 253–256. 16 Stadler B, Dixon AFG (2005) Ecology and evolution of aphid-ant Biological market theory offers a potentially 39 Kiers ET, Rousseau RA, West SA, Denison RF (2003) Host interactions. Annu Rev Ecol Evol Syst 36:345–372. sanctions and the legume-rhizobium mutualism. Nature 425(6953): valuable framework for studying microbial 17 Port M, Clough D, Kappeler PM (2009) Market effects offset the 78–81. reciprocation of grooming in free-ranging redfronted lemurs, Eulemur cooperation among species. We predict that 40 Simms EL, et al. (2006) An empirical test of partner choice – its primary contribution will be to generate fulvus rufus. Anim Behav 77(1):29 36. mechanisms in a wild legume-rhizobium interaction. Proc Biol Sci 18 Fruteau C, Lemoine S, Hellard E, van Damme E, Noë R (2011) new experimental questions and hypotheses 273(1582):77–81. When females trade grooming for grooming: Testing partner control 41 Oono R, Anderson CG, Denison RF (2011) Failure to fix nitrogen in the field of social microbiology. Also, ap- and partner choice models of cooperation in two primate species. by non-reproductive symbiotic rhizobia triggers host sanctions that – plying market theory to microbial mutual- Anim Behav 81(6):1223 1230. reduce fitness of their reproductive clonemates. Proc Biol Sci 19 Kutsukake N, Clutton-Brock TH (2010) Grooming and the value 278(1718):2698–2703. isms will be an important test of the ro- of social relationships in cooperatively breeding meerkats. Anim 42 Bever JD, Richardson SC, Lawrence BM, Holmes J, Watson M – bustness of market-based principles as a more Behav 79(2):271 279. (2009) Preferential allocation to beneficial symbiont with spatial general principle of social behavior. We 20 Bshary R, Grutter AS (2002) Experimental evidence that partner structure maintains mycorrhizal mutualism. Ecol Lett 12(1):13–21. choice is a driving force in the payoff distribution among cooperators 43 Kiers ET, et al. (2011) Reciprocal rewards stabilize cooperation in identified six strategies microbes use or benefit or mutualists: The cleaner fish case. Ecol Lett 5(1):130–136. the mycorrhizal symbiosis. Science 333(6044):880–882. from to optimize their success in markets. 21 Akçay E, Simms EL (2011) Negotiation, sanctions, and context 44 Giovannetti M, Sbrana C, Avio L, Strani P (2004) Patterns of This list will no doubt grow as researchers dependency in the legume-Rhizobium mutualism. Am Nat 178(1): below-ground plant interconnections established by means of 1–14. arbuscular mycorrhizal networks. New Phytol 164(1):175–181. uncover new ways in which microbes ma- 22 Schwartz MW, Hoeksema JD (1998) Specialization and resource 45 Mikkelsen BL, Rosendahl S, Jakobsen I (2008) Underground nipulate trade in their favor. Because they are trade: Biological markets as a model of mutualisms. Ecology 79(3): resource allocation between individual networks of mycorrhizal fungi. – typically easy to manipulate, microbial mar- 1029 1038. New Phytol 180(4):890–898. 23 Hoeksema JD, Schwartz MW (2003) Expanding comparative- 46 Lekberg Y, Hammer EC, Olsson PA (2010) Plants as resource kets will be useful systems for testing ques- advantage biological market models: Contingency of mutualism on islands and storage units—adopting the mycocentric view of tions about biological markets in general partners’ resource requirements and acquisition trade-offs. Proc Biol arbuscular mycorrhizal networks. FEMS Microbiol Ecol 74(2): Sci 270(1518):913–919. 336–345. (Table 1). These include the evolution of 24 Leimar O, Hammerstein P (2010) Cooperation for direct fitness 47 Selosse M-A, Rousset F (2011) Evolution. The plant-fungal partner choice, responses to price fluctuations, benefits. Philos Trans R Soc Lond B Biol Sci 365(1553):2619–2626. marketplace. Science 333(6044):828–829.

Werner et al. PNAS | January 28, 2014 | vol. 111 | no. 4 | 1243 Downloaded by guest on September 29, 2021 48 Fraser B (2013) Cooperation and Its Evolution, eds Sterelny K, 74 Partida-Martinez LP, Monajembashi S, Greulich K-O, Hertweck C 98 Simms EL, Taylor DL (2002) Partner choice in nitrogen-fixation Joyce R, Calcott B, Fraser B (MIT Press, Cambridge, MA), pp 153–174. (2007) Endosymbiont-dependent host reproduction maintains mutualisms of legumes and rhizobia. Integr Comp Biol 42(2): 49 Gubry-Rangin C, Garcia M, Béna G (2010) Partner choice in bacterial-fungal mutualism. Curr Biol 17(9):773–777. 369–380. Medicago truncatula-Sinorhizobium symbiosis. Proc Biol Sci 75 Lackner G, Moebius N, Partida-Martinez LP, Boland S, 99 Amantonico A, Urban PL, Fagerer SR, Balabin RM, Zenobi R 277(1690):1947–1951. Hertweck C (2011) Evolution of an endofungal lifestyle: Deductions (2010) Single-cell MALDI-MS as an analytical tool for studying 50 Fellbaum CR, et al. (2012) Carbon availability triggers fungal from the Burkholderia rhizoxinica genome. BMC Genomics 12:210. intrapopulation metabolic heterogeneity of unicellular organisms. nitrogen uptake and transport in arbuscular mycorrhizal symbiosis. 76 Ellers J, Kiers ET, Currie CR, McDonald BR, Visser B (2012) Anal Chem 82(17):7394–7400. – Proc Natl Acad Sci USA 109(7):2666 2671. Ecological interactions drive evolutionary loss of traits. Ecol Lett 100 Ghosh S, et al. (2011) Phenotypic heterogeneity in 51 Behm JE, Kiers ET A phenotypic plasticity framework for assessing 15(10):1071–1082. mycobacterial stringent response. BMC Syst Biol 5:18. intraspecific variation in arbuscular mycorrhizal fungal traits. J Ecol, 77 Richards S, et al.; International Aphid Genomics Consortium 101 Davidson CJ, Surette MG (2008) Individuality in bacteria. Annu in press. (2010) Genome sequence of the pea aphid Acyrthosiphon pisum. Rev Genet 42:253–268. 52 Frank SA (1994) Spatial polymorphism of bacteriocins and other PLoS Biol 8(2):e1000313. 102 Ackermann M (2013) Microbial individuality in the natural – allelopathic traits. Evol Ecol 8(4):369 386. 78 Shinzato C, et al. (2011) Using the Acropora digitifera genome to environment. ISME J 7(3):465–467. 53 Lion S, Baalen Mv (2008) Self-structuring in spatial evolutionary understand coral responses to environmental change. Nature 103 Verbruggen E, Kiers ET (2010) of ecology. Ecol Lett 11(3):277–295. 476(7360):320–323. mycorrhizal functional diversity in agricultural systems. Evol Appl 54 Wintermute EH, Silver PA (2010) Dynamics in the mixed microbial – 79 Dubilier N, Bergin C, Lott C (2008) Symbiotic diversity in marine 3(5-6):547 560. concourse. Genes Dev 24(23):2603–2614. animals: The art of harnessing . Nat Rev Microbiol 104 Adler J (1975) Chemotaxis in bacteria. Annu Rev Biochem 55 Harcombe W (2010) Novel cooperation experimentally evolved – 6(10):725–740. 44:341 356. between species. Evolution 64(7):2166–2172. 80 Kleiner M, et al. (2012) Metaproteomics of a gutless marine 105 De Weert S, et al. (2002) Flagella-driven chemotaxis towards 56 Overmann J, Schubert K (2002) Phototrophic consortia: Model worm and its symbiotic microbial community reveal unusual exudate components is an important trait for tomato root systems for symbiotic interrelations between prokaryotes. Arch colonization by Pseudomonas fluorescens. Mol Plant Microbe Microbiol 177(3):201–208. pathways for carbon and energy use. Proc Natl Acad Sci USA – Interact 15(11):1173–1180. 57 Müller J, Overmann J (2011) Close interspecies interactions 109(19):E1173 E1182. 81 Brodin A (2005) Mechanisms of cache retrieval in long-term 106 Bay G, et al. (2013) Boreal feather mosses secrete chemical between prokaryotes from sulfureous environments. Front Microbiol signals to gain nitrogen. New Phytol 200(1):54–60. hoarding birds. J Ethol 23(2):77–83. 2:146. 107 Song Z, Feldman MW (2013) Plant-animal mutualism in 58 Overmann J (2004) Symbiosis Mechanisms and Model Systems, 82 Ivens AB, Kronauer DJ, Pen I, Weissing FJ, Boomsma JJ (2012) — biological markets: Evolutionary and ecological dynamics driven by ed Sekbach J (Kluwer, Dordrecht, The Netherlands), pp 241–255. Ants farm subterranean aphids mostly in single clone groups an non-heritable phenotypic variance. Theor Popul Biol 88:20–30. 59 Boetius A, et al. (2000) A marine example of prudent husbandry for carbohydrates and proteins? BMC 108 Foster KR, Kokko H (2006) Cheating can stabilize cooperation apparently mediating anaerobic oxidation of methane. Nature Evol Biol 12:106. in mutualisms. Proc Biol Sci 273(1598):2233–2239. 407(6804):623–626. 83 Hammer EC, Pallon J, Wallander H, Olsson PA (2011) Tit for tat? 109 Koyama NF, Caws C, Aureli F (2012) Supply and demand 60 Verbruggen E, et al. (2012) Spatial structure and interspecific A mycorrhizal fungus accumulates phosphorus under low plant predict male grooming of swollen females in captive chimpanzees, cooperation: Theory and an empirical test using the mycorrhizal carbon availability. FEMS Microbiol Ecol 76(2):236–244. Pan troglodytes. Anim Behav 84(6):1419–1425. mutualism. Am Nat 179(5):E133–E146. 84 Kessin RH (2001) Dictyostelium: Evolution, Cell Biology, and 110 Wei W, et al. (2013) Supply and demand determine the market 61 Taylor PD (1992) in viscous populations - an inclusive the Development of Multicellularity (Cambridge Univ Press, value of access to infants in the golden snub-nosed monkey fitness model. Evol Ecol 6(4):352–356. Cambridge, UK). (Rhinopithecus roxellana). PLoS ONE 8(6):e65962. 62 Wilson D, Pollock G, Dugatkin L (1992) Can altruism evolve in 85 Brock DA, Douglas TE, Queller DC, Strassmann JE (2011) Primitive 111 Wernegreen JJ (2012) Mutualism meltdown in insects: Bacteria purely viscous populations? Evol Ecol 6(4):331–341. agriculture in a social amoeba. Nature 469(7330):393–396. constrain thermal adaptation. Curr Opin Microbiol 15(3):255–262. 63 Boyle RA, Williams HTP, Lenton TM (2012) Natural selection for 86 Pion M, et al. (2013) Bacterial farming by the fungus Morchella 112 Toby Kiers E, Palmer TM, Ives AR, Bruno JF, Bronstein JL (2010) costly nutrient recycling in simulated microbial metacommunities. crassipes. Proc Biol Sci 280(1773):20132242. Mutualisms in a changing world: An evolutionary perspective. Ecol J Theor Biol 312C:1–12. 87 Chao L, Levin BR (1981) Structured and the evolution of Lett 13(12):1459–1474. 64 Nakajima T, Sano A, Matsuoka H (2009) Auto-/heterotrophic anticompetitor toxins in bacteria. Proc Natl Acad Sci USA 78(10): 113 Kiers ET, West SA, Denison RF (2002) Mediating mutualisms: endosymbiosis evolves in a mature stage of development 6324–6328. Farm management practices and evolutionary changes in symbiont in a microcosm composed of an alga, a bacterium and a ciliate. 88 Riley MA (1998) Molecular mechanisms of bacteriocin evolution. co-operation. J Appl Ecol 39(5):745–754. Biosystems 96(2):127–135. – Annu Rev Genet 32:255 278. 114 Johnson NC (2010) Resource stoichiometry elucidates the 65 Reguera G, et al. (2005) Extracellular electron transfer via 89 Gardner A, West SA, Buckling A (2004) Bacteriocins, spite and structure and function of arbuscular mycorrhizas across scales. New microbial nanowires. Nature 435(7045):1098–1101. – virulence. Proc Biol Sci 271(1547):1529 1535. Phytol 185(3):631–647. 66 Summers ZM, et al. (2010) Direct exchange of electrons within 90 Schoustra SE, Dench J, Dali R, Aaron SD, Kassen R (2012) aggregates of an evolved syntrophic coculture of anaerobic bacteria. 115 Sachs JL, Hollowell AC (2012) The origins of cooperative Antagonistic interactions peak at intermediate genetic distance in – Science 330(6009):1413–1415. bacterial communities. MBio 3(3):e00099 e12. clinical and laboratory strains of . BMC 67 Morita M, et al. (2011) Potential for direct interspecies electron 116 Morris JJ, Lenski RE, Zinser ER, Loss AG (2012) The Black Queen Microbiol 12:40. transfer in methanogenic wastewater digester aggregates. MBio 2(4): Hypothesis: Evolution of dependencies through adaptive gene loss. 91 Strassmann JE, Gilbert OM, Queller DC (2011) Kin discrimination – e00159–e11. MBio 3(2):e00036 e12. and cooperation in microbes. Annu Rev Microbiol 65:349–367. 68 Little AEF, Murakami T, Mueller UG, Currie CR (2006) Defending 117 Maughan H, et al. (2006) The population genetics of 92 Cascales E, et al. (2007) Colicin biology. Microbiol Mol Biol Rev against parasites: Fungus-growing ants combine specialized phenotypic deterioration in experimental populations of Bacillus 71(1):158–229. – behaviours and microbial symbionts to protect their fungus gardens. subtilis. Evolution 60(4):686 695. 93 Poulsen M, Boomsma JJ (2005) Mutualistic fungi control crop Biol Lett 2(1):12–16. 118 Ratcliff WC, Denison RF (2010) Individual-level bet hedging in diversity in fungus-growing ants. Science 307(5710):741–744. 69 Mattoso TC, Moreira DDO, Samuels RI (2012) Symbiotic bacteria the bacterium Sinorhizobium meliloti. Curr Biol 20(19):1740–1744. 94 Schiøtt M, Rogowska-Wrzesinska A, Roepstorff P, Boomsma JJ on the cuticle of the leaf-cutting ant Acromyrmex subterraneus 119 Ratcliff WC, Denison RF (2011) Bacterial persistence and bet subterraneus protect workers from attack by entomopathogenic (2010) Leaf-cutting ant fungi produce cell wall degrading pectinase hedging in Sinorhizobium meliloti. Commun Integr Biol 4(1):98–100. fungi. Biol Lett 8(3):461–464. complexes reminiscent of phytopathogenic fungi. BMC Biol 8:156. 120 Denison RF, Kiers ET, West SA (2003) Darwinian agriculture: 70 Ricardo D (1817) On the Principles of Political Economy and 95 Hawlena H, Bashey F, Lively CM (2010) The evolution of spite: When can humans find solutions beyond the reach of natural Taxation (John Murray, London, UK). Population structure and bacteriocin-mediated antagonism in two selection? Q Rev Biol 78(2):145–168. 71 Thompson AW, et al. (2012) Unicellular cyanobacterium natural populations of xenorhabdus bacteria. Evolution 64(11): 121 Schluter J, Foster KR (2012) The evolution of mutualism in gut symbiotic with a single-celled eukaryotic alga. Science 337(6101): 3198–3204. microbiota via host epithelial selection. PLoS Biol 10(11):e1001424. 1546–1550. 96 Hawlena H, Bashey F, Mendes-Soares H, Lively CM (2010) 122 Bäckhed F, Ley RE, Sonnenburg JL, Peterson DA, Gordon JI 72 Kim HJ, Boedicker JQ, Choi JW, Ismagilov RF (2008) Defined Spiteful interactions in a natural population of the bacterium (2005) Host-bacterial mutualism in the human intestine. Science spatial structure stabilizes a synthetic multispecies bacterial Xenorhabdus bovienii. Am Nat 175(3):374–381. 307(5717):1915–1920. community. Proc Natl Acad Sci USA 105(47):18188–18193. 97 Bashey F, Young SK, Hawlena H, Lively CM (2012) Spiteful 123 Ezenwa VO, Gerardo NM, Inouye DW, Medina M, Xavier JB 73 Hammerstein P, Hagen EH (2005) The second wave of interactions between sympatric natural isolates of Xenorhabdus (2012) Microbiology. Animal behavior and the microbiome. Science evolutionary economics in biology. Trends Ecol Evol 20(11):604–609. bovienii benefit kin and reduce virulence. J Evol Biol 25(3):431–437. 338(6104):198–199.

1244 | www.pnas.org/cgi/doi/10.1073/pnas.1315980111 Werner et al. Downloaded by guest on September 29, 2021