Available online at www.sciencedirect.com Studies in History and Philosophy of Biological and Biomedical Sciences Stud. Hist. Phil. Biol. & Biomed. Sci. 38 (2007) 820–833 www.elsevier.com/locate/shpsc From quorum to cooperation: lessons from bacterial sociality for evolutionary theory Pamela Lyon University of Adelaide/Australian National University, 20 Wellesley Avenue, Evandale SA 5069, Australia Abstract The study of cooperation and altruism, almost since its inception, has been carried out without reference to the most numerous, diverse and very possibly most cooperative domain of life on the planet: bacteria. This is starting to change, for good reason. Far from being clonal loners, bacteria are highly social creatures capable of astonishingly complex collective behaviour that is mediated, as it is in colonial insects, by chemical communication. The article discusses recent experiments that explore different facets of current theories of the evolution and maintenance of cooperation using bacterial models. Not only do bacteria hold great promise as experimentally trac- table, rapidly evolving systems for testing hypotheses, bacterial experiments have already raised interesting questions about the assump- tions on which our current understanding of cooperation and altruism rests. Ó 2007 Elsevier Ltd. All rights reserved. Keywords: Cooperation; Altruism; Kin selection; Group selection; Communication; Cell–cell signalling When citing this paper, please use the full journal title Studies in History and Philosophy of Biological and Biomedical Sciences 1. Introduction One of the enduring unsettled issues of evolutionary Given the intrinsic selfishness assumed to underlie Dar- biology is the paradox of collateral altruistic behav- winian competition, cooperation should be rare in nature. iour—that is, when some individuals subordinate their But it is not. In fact, cooperation ‘pervades all levels of bio- own interests and those of their immediate offspring in logical organization’ (Sachs, 2006, p. 1415), in the sense order to serve the interests of a larger group beyond off- that individual entities act to produce an effect the cost of spring. (E. O. Wilson, 2005, p. 159) which to the agent is not immediately—or even ever—com- pensated. Not only do some organisms cooperate with During the process of aggregation and early mound for- other organisms (both of their own kind and of different mation [in Myxococcus xanthus], 65 to 90% of the cells kinds) and groups cooperate with groups, so also do spa- lyse [commit suicide] ... (M. Dworkin, 1996, p. 81) tially distant transcriptional units of DNA (‘‘genes’’) coop- erate to produce proteins; proteins cooperate to catalyze reactions and to transduce signals within cells; and cells cooperate with one another (Wingreen & Levin, 2006). E-mail address: [email protected] 1369-8486/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.shpsc.2007.09.008 P. Lyon / Stud. Hist. Phil. Biol. & Biomed. Sci. 38 (2007) 820–833 821 Without cooperation most of the ‘major landmarks in the generally and within particular contexts. For the past four diversification of life and the hierarchical organization of decades, however, kin selection and the game theoretic the living world’ would have been impossible, including interpretation of reciprocity have been the dominant theo- the transitions from nonlife to life, networks of cooperating retical paradigms (Sachs et al., 2004). Both approaches genes to the first functioning cell, prokaryotes to eukary- assess acts in terms of costs and benefits, mainly to individ- otes, unicellular to multicellular organization, asexual to uals, where cost and benefit are typically calculated in sexual reproduction, and so on to the development of com- terms of reproductive fitness, its reduction or enhancement. plex ecosystems (Michod & Herron, 2006, p. 1406). In Both approaches are also formally tractable—see, for short, the more nature yields her secrets, the more ubiqui- example, Nowak’s five mathematical ‘rules’ for the evolu- tous cooperation appears to be. tion of cooperation (Nowak, 2006)—and thus yield elegant A growing number of reviews identifies multiple avenues computer simulations. to the evolution of cooperation and altruism in a selfish Kin selection and game theoretic reciprocity have world (Dugatkin, 2002; Kerr et al., 2004; Sachs et al., always had their critics, but the limitations of these 2004; Fletcher & Zwick, 2006; Lehmann & Keller, 2006; approaches have mounted with growing concern for eco- van Baalen & Jansen, 2006). How the paths are parsed var- logical validity. A persistent complaint is that these ies among the different authors but, roughly speaking, they abstract formal models often fail to connect with empirical are individuated according to the relation of cooperator to observation because they do not account for complex and beneficiary and/or the allocation of costs and benefits asso- dynamic ecological factors, namely, ‘the ‘‘real world’’ of ciated with cooperative acts. existing biological organisms’ (Leimar & Hammerstein, In kin selection (genic selectionism) cooperator and ben- 2006, p. 1403). Calculating costs and benefits to individuals eficiary are genetically closely related. Widely regarded as is not always straightforward, especially in complex social ‘one of the most important developments in evolutionary arrangements such as class-structured2 populations and biology’ (Griffin & West, 2002, p. 15), W. D. Hamilton’s multi-species consortia (Wild & Taylor, 2006), and neither theory of inclusive fitness holds that nature favours a type is divining ‘‘direct’’ and ‘‘indirect’’ fitness components sub- of reproductive fitness that ‘‘includes’’ both the fitness of ject to selection (Wenseleers, 2006). Recent kin selection an individual and the fitness of the individual’s close rela- modelling using non-linear cost/benefit functions believed tives (Hamilton, 1964a,b). Kin selection asserts that the better to mirror natural conditions yielded paradoxical genetic resources enabling cooperative or altruistic behav- results, for example, that there can be selection simulta- iour will evolve in groups where individuals are highly neously for more and less cooperation (Doebeli & Hauert, related and the cost of the cooperative action to the coop- 2006). erator is relatively small while the benefit to kin is large (Maynard Smith, 1964). In short, an organism is more 1.1. Bacteria as experimental models of cooperation likely to subordinate its own selfish interests to effective group behaviour if those with whom it cooperates share If bridging the gap between theoretical and empirical its genes and the personal fitness costs are not too high. research is, as Doebeli and Hauert (2005) suggest, ‘a major Theoretically, the closer the relation (e.g. offspring, sib- challenge for further progress in understanding the evolu- lings), the more likely cooperative behaviour will evolve. tion of cooperation’ (ibid., p. 761), then identifying model In group selection cooperator and beneficiary may or organisms of varying physiological and social complexity may not be genetically closely related but are part of a pop- should be a high priority. Ideally, such models will be ulation that collectively exploits an ecological niche. Reci- not only experimentally manageable but also well under- procity, as the name suggests, means that the cost of a stood behaviourally, physiologically and genetically (Mik- cooperative act is likely to be recompensed, directly or indi- los, 1993). Colonial insects, such as ants and bees, are rectly, by the beneficiary at a future time. Trivers (1971) well studied systems of cooperative sociality. In recent proposed the concept of ‘‘reciprocal altruism’’ in the con- years, however, interest has grown in microorganisms, text of the iterated Prisoner’s Dilemma of economic game including bacteria, as models of social evolution, coopera- theory to explain the evolution of cooperation among dis- tion and altruism (e.g. Crespi, 2001; Velicer, 2003; Travisa- tantly related or unrelated organisms. Finally, mutual no & Velicer, 2004; Strassmann et al., 2000; Velicer et al., advantage resulting as an incidental benefit of the ordinary 2000; Vulic & Kolter, 2001; Griffin et al., 2004; Brockhurst selfish behaviour of individuals is called by-product et al., 2006). 1 mutualism. On one hand, it could be argued such attention is over- Each of these paths has proponents and critics concern- due. Bacteria and Archaea comprise the most numerous ing its relative importance to the evolution of cooperation, kingdoms of life and may account for the largest proportion 1 Indirect reciprocity and by-product mutualism are often listed as the same phenomenon. However, some authors (e.g. Nowak, 2006) identify ‘indirect reciprocity’ as a form of cooperation based on reputation, which is different from by-mutualism, so I have separated them here. 2 A class-structured population is one in which phenotypically distinct types (for example reproductives, workers, sentinels) perform different tasks within the group, such as in ant colonies. 822 P. Lyon / Stud. Hist. Phil. Biol. & Biomed. Sci. 38 (2007) 820–833 of biomass on this planet, yet most theoretical biology and (Molin & Tolker-Neilsen, 2003), which can provide bacte- philosophy of biology has proceeded without reference to ria with innovative functional solutions to novel environ- them (O’Malley & Dupre´, 2007). Certainly Darwin’s the- mental problems (Ochman & Moran, 2001). Cooperative ory of evolution by