News & Views | Published:


The Phoenix effect

Nature volume 441, pages 291292 (18 May 2006) | Download Citation


A spore-forming bacterium can escape from social collapse and extinction with a single mutation that has a dramatic effect. Here is evidence that a cooperative system can recover from the very brink of destruction.

Where there is society, there are cheaters that threaten to ruin it. The evolution of such selfish behaviour can destroy cooperation, and may even drive a species to extinction1. When Fiegna, Velicer and colleagues (page 310)2 mixed a cheater strain with a cooperative strain of a bacterium, therefore, tragedy seemed assured. Many populations of the bacteria did indeed die out but, in one, a new social strain arose phoenix-like from the social collapse. This new strain resisted the cheater, produced more spores than the original strain and, most amazingly of all, evolved these abilities through just a single mutational change.

The bacterium concerned is Myxococcus xanthus. It earns its title of ‘social’ bacterium from the ability of starving cells to aggregate and form a fruiting body of hardy spores, in which many of the cells die3,4,5 (Fig. 1). This social behaviour has been put to good use by Velicer and colleagues in an intriguing story that has spanned almost a decade3,4,5. It began with an experiment that selected for mutant strains that grew rapidly under conditions in which they did not need to make spores5. Interestingly, several strains lost the ability to sporulate, and there was a further twist. Some could still form spores if mixed with the ancestral spore-former and, furthermore, produced more spores than the ancestor. Although socially inept when alone, these strains could exploit the shared resources in aggregations with the ancestor and use them to gain a selfish evolutionary advantage. Velicer and colleagues had inadvertently created cheater mutants4.

Figure 1: Fruiting bodies of the social bacterium Myxococcus xanthus emerging from soil.
Figure 1

Each fruiting body is about 100 µm in diameter, and contains a few thousand hardy spores that form through the aggregation of around 100,000 starving cells. (Photo by M. Vos.)

But would the cheaters completely replace the spore-former? To answer this, Fiegna and Velicer3 put mixtures containing one cheater strain and the social ancestor through several fruiting cycles in the lab by taking the spores from one cycle to seed the next. Two cheater strains persisted alongside the spore-former, but the more vigorous cheater rapidly increased in frequency and, as it did so, the results were devastating. By replacing the social ancestor, the cheater removed its own ability to produce spores. The result was population extinction: without spores, there was nothing to seed the next cycle of fruiting. These experiments provided a graphic illustration of what has come to be known as ‘evolutionary suicide’, where natural selection favours strategies that promote extinction. This process is alarming some conservation biologists, who worry that anthropogenic effects may be kick-starting spirals of selfish adaptations in some species that will drive their own destruction (Fig. 2b, overleaf)1.

Figure 2: Four possible outcomes when a cheater evolves in a social species.
Figure 2

A cheater is an organism that exploits a cooperative adaptation for selfish gain. a, Preadapted resistance to cheating. It is typical to assume that social systems arise in such a way that cheaters can have only a limited impact (as shown), or do not succeed at all. Examples of preadaptations include high relatedness and pre-existing constraints that link cheating to a cost to the cheater11. Policing and enforcement systems may evolve later to further constrain cheaters12. b, Extinction. The cheater causes extinction of the social trait, or species (evolutionary suicide1). This selects for species preadapted to resist cheating10. c, Unstable recovery. A social strategy arises that resists the cheater but cannot out-compete the original strategy. The original strategy may reinvade and perpetuate a cycle of reinvasions in a rock–paper–scissors dynamic8,9. d, Stable recovery. Sociality is restored by a strategy that out-competes both the cheater and the original strategy, as occurred with the Phoenix mutant2. The result is a stable adaptation that protects the social system from the cheater. This process may be behind the policing and enforcement systems in other social species12.

At least for the bacteria, however, all was not doom and gloom. Curiously, in one experiment, spores reappeared and the population recovered2. It turned out that a new super-strain had evolved that could resist the cheater. This was named Phoenix after the mythical burning bird that can rise from its own ashes. In a tour de force of genetics, which involved the marathon task of sequencing the entire genome of the new strain, Fiegna et al.2,6 found that Phoenix arose with just a single mutation. This mutation simply increases the levels of one particular enzyme (an acetyl-transferase), but this is hypothesized to trigger a flood of subsequent changes in gene regulation and to drive a previously unrecognized route to sociality in M. xanthus.

To show that a social adaptation can escape from such a severe cheater, and can do so through a single mutation, makes this a landmark study2,6. As with all studies, however, there are caveats. Most importantly, this was all done in the lab, and strains were selected and mixed in ways that are unlikely to occur in nature. For example, natural aggregations of M. xanthus may only ever contain a single strain, which would mean that the lab cheaters that cannot develop alone would never spread. One must be cautious, therefore, when using these results to draw conclusions about the natural world. Nevertheless, the study represents a notable proof-of-principle that has intriguing implications for sociobiology.

First of all, it is quite amazing to find that a single mutation can drive such a dramatic social recovery, a result that underlines just how little we know about the genetic basis of social traits7. It is true that during a rapid population crash there may be little time for anything other than the most simple of genetic changes. However, a recovery mutation might also have resulted in only a slight initial improvement in spore production and a rather different evolutionary prognosis (Fig. 2c). If a new cooperator evolved to replace the cheater that was a poor sporulator, the way might be open for the original cooperator to re-evolve. This would mean the cheater could then reinvade, and perpetuate a potentially endless evolutionary game of rock–paper–scissors. Although this may seem to stretch credulity, such outcomes are known from both chemical warfare in bacteria8 and male–male competition in lizards9.

This is not the case for Phoenix (Fig. 2d), which can out-compete the original strain and forms spores just fine. Such precipitous recoveries may turn out to be part of the evolutionary process. Extinctions play a key role in the history of life by removing species that are poorly adapted to persist, a process that favours both sexual reproduction and reduced within-species conflict10,11. However, the study by Fiegna et al.2 shows that species may also escape from the very brink of disaster. There is some sense in this: a virulent cheater that threatens a population will necessarily result in strong natural selection for strategies that can re-evolve sociality in its wake. It was this effect that led to the rapid dominance of the Phoenix mutant, and one can speculate that it might also explain adaptations that police cheating in other societies12. The final twist in the tale is that Phoenix actually produces more spores than the original strain. Escape from a virulent cheater is not just possible; it can even improve things.


  1. 1.

    & Oikos 111, 616–619 (2005).

  2. 2.

    , , & Nature 441, 310–314 (2006).

  3. 3.

    & Proc. R. Soc. Lond. B 270, 1527–1534 (2003).

  4. 4.

    , & Nature 404, 598–601 (2000).

  5. 5.

    , & Proc. Natl Acad. Sci. USA 95, 12376–12380 (1998).

  6. 6.

    et al. Proc. Natl Acad. Sci. USA 103, 8107–8112 (2006).

  7. 7.

    , & Nature Rev. Genet. 6, 257–270 (2005).

  8. 8.

    & Nature 428, 412 (2004).

  9. 9.

    & Nature 380, 240 (1996).

  10. 10.

    in Levels of Selection in Evolution (ed. Keller, L.) 238–252 (Princeton Univ. Press, 1999).

  11. 11.

    , , , & Nature 431, 693–696 (2004).

  12. 12.

    Evolution 57, 693–705 (2003).

Download references

Author information


  1. Kevin R. Foster, currently in the Program for Evolutionary Dynamics, Harvard University, will shortly move to the Bauer Center for Genomics Research, Harvard University, Cambridge, Massachusetts 02138, USA.

    • Kevin R. Foster


  1. Search for Kevin R. Foster in:

About this article

Publication history



Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.

Newsletter Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing