Commentary on: M. Ackermann, B. Stecher, N. E. Freed, P. Songhet, W.-D. Hardt, and M. Doebeli (2008) Self-destructive cooperation mediated by phenotype noise. Nature 454:987-9

One of the most exciting developments in microbial population biology over the past few years is the recognition that high levels of phenotypic noise – in which genetically identical microbes express different genes and manifest different phenotypes despite a common environment – is widespread in bacterial populations and that this noise plays an important role in bacterial evolutionary ecology (e.g. Elowitz et al. 2002, Balaban et al. 2004, Rosenfeld et al. 2005, Acar et al. 2008, Veening et al. 2008). I have always thought that the best explanations for this phenomenon involve bet hedging in uncertain environments (Seger and Brockmann 1987), and indeed this bet-hedging perspective has been well supported by mathematical modeling (e.g. Thattai and van Oudenaarden 2004, Kussell et al. 2005).

But in this week’s issue of Nature, Martin Ackermann and colleagues propose an alternative explanation that may explain some important cases, though probably not all case, of adaptive phenotypic noise in bacteria. Using experiments and mathematical models, Ackermann and his collaborators studied the process of pathogenesis in Salmonella typhimurium. In this species, a fraction of the bacterial cells in the gut lumen invade the gut tissue and express a set of virulence factors that trigger a large-scale gut inflammatory response. This response eliminates competing species from the gut – but it also results in near-certain death for the invading cells. The act of invading the gut and provoking an inflammatory response to the benefit of other Salmonella cells remaining behind in the lumen is what Ackermann et al. describe as “self-destructive cooperation.”

How could this have evolved and what does phenotypic noise – in this case, expression of the virulence factors by only a subset of the Salmonella cells in the gut – have to do with that evolution? Ackermann’s argument, which is bolstered by a clear and simple mathematical model and experimental tests of the underlying assumptions, is straightforward but ingenious. It is well known that costly cooperation can evolve when there is positive assortment, such that cooperators are more likely than defectors to be around other cooperators. If colonization occurs by a relatively small founding population, as can be the case for Salmonella, this can provide the necessary degree of positive assortment to explain costly cooperation. But a problem remains for explaining self-destructive cooperation. If a gene deterministically induces the self-destructive cooperative behavior, no individuals carrying that gene will be left behind to reap the benefits. The only way that self-destructive cooperation can evolve is if there is heterogeneity in expression of the trait. Ackermann et al. propose that this evolutionary argument may be the reason for phenotypic noise in expression of the Salmonella virulence factors. Moreover, a similar argument may apply to many other examples of bacterial virulence, such as the toxins released by cell lysis in Streptococcus pneumoniae and Clostridium difficile.

If Ackermann et al. are correct in this explanation – and I think that they are likely to be – this example highlights the necessity of understanding a pathogen’s evolutionary ecology if we hope to understand its mechanisms of pathogenesis. It also hints at ways to combat certain bacterial infection. If Salmonella typhimurium virulence is the result of an evolved solution to a tricky cooperative dilemma, then methods of disease control that interfere with that cooperation could leave the pathogen in the position of having to re-evolve cooperation – and this challenge can be far more difficult and take far longer than the challenge of evolving resistance to a conventional antibiotic (André and Godelle 2005).

References

Acar et al. (2008) Nature Genetics 40:471-575

Ackermann et al. (2008) Nature 454:987-990

André and Godelle (2005) Ecology Letters 8:800-810

Balaban et al. (2004) Science 305:1622-1625

Elowitz et al (2002) Science 297:1183-1186

Kussell et al. (2005) Genetics 169, 1807-1814

Seger and Brockmann (1987) Oxford Surveys In Evolutionary Biology 4:182-211

Thattai and van Oudenaarden (2004) Genetics, 167: 523-530

Veening et al. (2008) Proceedings of the National Academy of Sciences USA 105:4393-4398;