Gene

Demography

Biofilm

Ecology

Bacteria

Biology

Public good

Microeconomics

Variety (cybernetics)

Cheating

Economics

Computer science

Quorum sensing

Evolutionary biology

Genetics

Population

Sociology

Artificial intelligence

Multicellular organism

Biological dispersal

Current Biology

Communication and cooperation are not restricted to complex, higher organisms. Microbes, too, perform a variety of collective, multicellular behaviors, including biofilm formation, quorum sensing, nutrient acquisition, and dispersal. The products of these microbial cooperative behaviors are generally referred to as public goods. Here we describe the nature of microbial public goods, the associated problem of cheating, and ways in which microbes maintain public goods in the face of cheating. We highlight work in a growing field at the interface of microbiology, evolution, and ecology that combines multiple approaches in experimental evolution, genetics, and mathematical modeling. Communication and cooperation are not restricted to complex, higher organisms. Microbes, too, perform a variety of collective, multicellular behaviors, including biofilm formation, quorum sensing, nutrient acquisition, and dispersal. The products of these microbial cooperative behaviors are generally referred to as public goods. Here we describe the nature of microbial public goods, the associated problem of cheating, and ways in which microbes maintain public goods in the face of cheating. We highlight work in a growing field at the interface of microbiology, evolution, and ecology that combines multiple approaches in experimental evolution, genetics, and mathematical modeling. The concept of public goods is rooted in economics. Examples of economic public goods are national defense, public parks, street lighting, and waste-water treatment plants. They are financed by tax revenues and are freely available to all citizens. Analogously, the defining characteristic of public goods in the context of social evolution is that they are costly to produce for the individual and are beneficial to all members of the local group or population. In microbial cooperation, public goods are loosely defined as compounds or functions that provide a collective benefit, generally through release into the extracellular environment. Microbial public goods take many forms (Figure 1 and Table 1): They range from large proteins to small metabolites; they may be actively secreted or they may passively diffuse from the producing cell. For example, extracellular enzymes degrade polymers into smaller fragments that can be taken up by individual cells, siderophores scavenge iron in depleted environments, biosurfactants promote group motility over surfaces, toxins injure host tissues to release nutrients, and antibiotics harm competing microbes. Exopolysaccharides and proteins form a protective extracellular matrix in bacterial biofilms, and secreted signaling molecules coordinate the behavior of the population in a process termed quorum sensing. A well-studied microbial species exhibiting all these properties is the opportunistic bacterial pathogen Pseudomonas aeruginosa, which is often associated with chronic infections in cystic fibrosis.Table 1A selection of microbial public goods.Public goodRole1The presumed major function is listed here, rather than all possible ascribed functions.Enzymes (for example, proteases)Extracellular nutrient digestionSiderophoresIron acquisitionAntibioticsMicrobial warfareExopolysaccharidesStructure and protection in biofilmsBiosurfactantsBacterial motilityQuorum sensing moleculesCell–cell signalingToxinsDamage of host tissuesLeaky metabolitesCellular growth and nutritionLightAttract prey or distract predators of symbiotic hostFruiting bodyProduction and dissemination of spores1 The presumed major function is listed here, rather than all possible ascribed functions. Open table in a new tab A different type of public good that is not a chemical compound is the light produced by some bioluminescent bacteria. When the bacteria form a mutually beneficial relationship with marine animals, the light serves to attract prey or distract predators. Finally, examples of more complex forms of microbial behavior are the fruiting bodies — multicellular, stalked sporangia that are formed by myxobacteria and social amoebae. Here, the entire structure may be considered the public good, as it is the product of cooperation. Another characteristic of public goods is that their benefit generally increases with population density. This aspect is easy to understand in the context of antibiotics or secreted enzymes. Antibiotics are only effective against competitors above a certain threshold concentration and therefore require an appropriately high density of cells that contribute to antibiotic production. Similarly, as population density increases, secreted enzymes that process nutrients progressively benefit neighboring cells rather than diffuse away. These properties presumably are a major reason why the production of many public goods is controlled by quorum sensing. Quorum sensing can therefore be understood as a mechanism of facultative cooperation, where signaling with relatively cheap public goods serves to limit the production of more expensive public goods to the appropriate social and environmental conditions. In microbial populations, public goods are readily shared, and thus they can be easily exploited by cheaters. Cheaters are non-cooperating individuals in the population that benefit from the public goods produced by the cooperators but do not share the cost of producing those goods (Figure 2A). Cheating in microbes can be demonstrated by straightforward culturing experiments utilizing strains marked with antibiotic resistance or fluorescent tags. Public-good producing and non-producing strains are cultured either separately or together in growth media that favor public goods production. When these strains are grown in separate cultures, producers grow at a faster rate than non-producers. In well-mixed co-culture, in contrast, non-producers outgrow producers, and are thus said to have higher ‘relative fitness’, which is typically determined as the ratio of the average growth rates of the two strains. Moreover, similar types of experiments have revealed both density dependence and frequency dependence as basic properties of public-goods cheating. This simply means that cheaters are better able to exploit cooperators when they are closer to them in dense populations, and when the cheaters themselves are less frequent in the population. Because cooperating populations are susceptible to invasion by cheaters, cooperation is not considered an evolutionarily stable strategy. This outcome is illustrated by the ‘prisoner’s dilemma’ scenario in evolutionary game theory. In the prisoner’s dilemma, two criminals are arrested for a petty crime, but are also suspected of having committed a more serious crime. The two are brought in for questioning individually, with neither prisoner knowing if the other confessed to or denied the more serious crime. Either denial or confession by one or both prisoners will lead to significantly different sentences (Figure 2B). Although it is in the best interest of both prisoners that neither confesses, without knowing whether the other prisoner has confessed, it is always in the best interest of each individual to confess. Hence, cheating on, rather than cooperating with, the fellow prisoner is the dominant strategy in this game. Numerous studies have confirmed this outcome in microbial populations. Cheaters have been shown to evolve in controlled laboratory experiments, in the context of bacterial infections, and even in the natural environment. In bacterial quorum sensing, for example, cheaters evolve in the form of ‘signal-blind’ mutants that no longer recognize the signal to gather food via public goods, but still reap the benefits. Likewise, in siderophore-mediated iron acquisition, cheater mutants deficient in siderophore synthesis, but not siderophore uptake, exploit siderophore-producing cooperators. A scenario related to the prisoner’s dilemma is Garrett Hardin’s “tragedy of the commons”, which considers the effects of exploitation on the community as a whole (Figure 2C). A tragedy of the commons occurs when individuals put their interests above that of the community and exploit a common resource until it is exhausted. An example of this is a public pasture that is being used by four herdsmen to graze their herds of cattle. The pasture can support eight cows, so each herdsman starts out with two cows in their herd. If a herdsman adds a cow to his herd there will not be enough grass in the pasture to support the cows and the growth of all cows will be reduced. It is in the individual herdsman’s best interest to add cows to their herd, knowing that he receives the benefit of having extra cows in the pasture, yet the cost of overgrazing is shared evenly by all herdsmen. Because of this, the herdsmen will keep adding cows to their herds until the pasture is completely overgrazed and unable to support any cows. Evolutionary biologists have applied Hardin’s analogy to a wide range of social dilemmas, including the use of public goods formed by cooperation. Here, the tragedy is exemplified by the burden of invading cheaters on a population cooperating via public goods. As cheaters invade, the fraction of public good producers in the population decreases to the point of population collapse. The tragedy of the commons, as the result of cheater invasion, has been mathematically proven and demonstrated in several microbial systems. In myxobacteria for example, cheating cells that do not invest in building the non-spore parts of the fruiting body will produce more spores than wild-type cells, thereby invading the population and ultimately destroying the fruiting body as the public good. Even though cooperative behavior is threatened by cheating, it persists and is found throughout life. Because selection acts on the levels of individuals and genes, the benefit that cooperation provides at the population level cannot explain its evolution and maintenance. A general theory that describes how cooperation is selectively favored over cheating is kin selection, which considers the fitness consequences of cooperation for all individuals in a local group that share the cooperative trait. This idea of inclusive fitness is encapsulated by Hamilton’s rule. It stipulates that a social trait will spread in a population if m + rn > 0 where m is the direct fitness gained by the actor performing the behavior and n is the indirect fitness gained by the recipient of the behavior, weighted by the relatedness r between actor and recipient. Fitness itself can be defined as the difference between benefits and costs. In this gene-centric theory, relatedness refers to the probability of an actor and a recipient sharing the locus encoding the social behavior. Cooperative behaviors that are mutually beneficial (m > 0) always fulfill this rule, whereas those that are altruistic (m < 0) require relatedness among individuals (r > 0). Several mechanisms have been described in recent years that help stabilize public good cooperation in microbial populations (summarized in Figure 3). In the context of inclusive fitness, these mechanisms generally either increase the direct fitness of the cooperating cell, or they increase the relatedness or indirect fitness to other cooperating cells. Of note, classifications can be made in different ways and a clear distinction between direct and indirect effects is not always possible. While Hamilton’s rule is quite simple and intuitive, more specific models, including game-theoretical ones, are often used to obtain a deeper mechanistic and dynamic understanding of microbial systems. An increase in direct fitness may be attained by an increase in benefits or a decrease in costs. The direct benefit of cooperation can be increased by preferential access to the public good, through either limited diffusion or partial privatization. The latter has been shown for yeast invertase, an enzyme that hydrolyzes sucrose (Figure 3A). As cells do not completely let go of invertase, they are able to capture a small portion of the sucrose before it diffuses away. The resulting evolutionary dynamic is akin to the ‘snow drift game’, a game that is similar in structure to the prisoner’s dilemma. However, an important difference in the snow drift game is that cooperation, rather than cheating, is advantageous if the opponent cheats, ultimately leading to the coexistence of cooperators and cheaters.The same outcome is predicted by the Black Queen Hypothesis. This evolutionary hypothesis, named after the queen of spades in the card game Hearts, explains selective gene loss in microbial species (essentially a form of cheating) as an adaptation to inadvertently leaky cellular products provided by other community members. Here, the producer species also has preferential access to the leaky function compared to a non-producing cell. The cost of public goods that imposes a burden on cooperation may be minimized as well, depending on molecular properties and environmental context. Nutrient limitation, for example, is an important determinant of these costs (Figure 3B). If the particular nutrient that limits cell growth is also required for the production of the public good, then cooperation is costly; otherwise, it is not. Consequently, public goods may be regulated according to specific nutritional context. In P. aeruginosa, certain public goods, such as biosurfactants used for swarming motility, are produced only when metabolic costs are low. There are several mechanisms that increase the relatedness of cooperating cells or, in other words, increase the assortment of similar individuals such that cooperators are more likely to interact with other cooperators. These include spatial structuring and limited population mixing, as well as kin discrimination and facultative cooperation. Spatial structure and limited mixing can lead to populations where all cells in the immediate area are highly related (Figure 3C). When dispersal is limited, cooperator and cheater lineages tend to segregate such that exploitation is limited. The sessile growth of many microbes in biofilms spatially structures the population. A microbial life cycle of colonial growth and dispersal can further stabilize cooperation, as repeated dispersal of colonies into small founder populations stochastically excludes cheaters, thereby improving the fitness of cooperating cells. Kin discrimination enables individuals to distinguish kin from non-kin and selectively direct aid towards them. A type of good that can provide exclusive benefits to related individuals in a population is a ‘club good’. For example, many microbes secrete highly specific siderophores that vary from strain to strain. These siderophores can only be taken up by cells that have the cognate receptor. This mechanism helps restrict the benefit of siderophore secretion to the producer itself and to relatives in the immediate area (Figure 3D), although, as mentioned above, it is circumvented by siderophore-deficient cheaters that still produce a suitable receptor. Kin discrimination also allows microbes to direct harmful behaviors at non-kin, often by means of toxins. Some toxins act at a distance, via secretion into the extracellular space, whereas others require direct cell–cell contact. The defining feature is that the cells secreting the toxin also produce an immunity factor, thus only harming non-kin that do not carry the immunity factor. The Type VI secretion system is an example of a widespread contact-dependent toxin delivery system. Bacteria such as P. aeruginosa, Vibrio cholerae, and Burkholderia thailandensis utilize Type VI secretion systems to attack and directly inject toxic effector proteins into neighboring cells. The opposing effects of toxin delivery on kin versus non-kin cells can result in spatial structuring, increasing the relatedness of neighbors in bacterial communities. The tight coupling of these toxin–immunity systems to public-goods production even allows cooperators to ‘police’ cheaters (Figure 3E). In B. thailandensis, for example, quorum sensing regulates a large number of functions, including public goods and a Type VI secretion system. Cheater mutants deficient in quorum sensing have not only lost the ability to secrete toxins via the Type VI secretion system, but they are also no longer resistant to them. Thus, they are killed upon encounter with toxin-secreting wild-type cells. A general feature of this and related cases is the pleiotropic linkage of public and private functions, where loss of the latter is costly for cheaters. Quorum sensing is a mechanism that not only allows facultative cooperation but also kin discrimination (Figure 3F). As mentioned above, quorum sensing restricts the production of public goods to times when it would be the most beneficial to the cooperating population. Diffusible chemical signals assess the density of cooperating individuals in a community. When the density of cooperating kin is high enough to overcome diffusion and guarantee that public-good secretion aids relatives, quorum-sensing bacteria begin expressing cooperative behaviors. In addition to the social factors described above, non-social adaptation to a particular environment can favor cooperators over cheaters. In Pseudomonas bacteria and yeast, for example, mutations improve the intracellular uptake of cooperatively acquired nutrients (Figure 3G). In P. aeruginosa, this can be achieved by the inactivation of a repressor of an amino-acid and peptide-transport system. Non-social adaptation works best if cooperators are initially more abundant than cheaters, because a larger population is more likely to sample beneficial mutations. Fast growth, ease of handling, and excellent genetics make many microbes ideal experimental models for studying the evolution and maintenance of cooperation via public goods. Social interactions within genetically defined populations and communities can be studied under controlled conditions in the laboratory. These studies have revealed the many forms of public goods in microbes, the dynamics of cooperation and cheating, and a range of mechanisms that help optimize and stabilize cooperative behavior.

Public goods and cheating in microbes