The workhorse of cell biology, yeast, is a surprisingly cooperative organism. It uses an unusual means of identifying partners — a 'green-beard gene', which encodes a tag that must match among cooperating cells.
Everyone knows how a glass or two of beer can act as social glue, making even misanthropes amiable. Oddly, the production of beer has a similarly convivial effect on the tiny brewers that make it, cells of the yeast Saccharomyces cerevisiae. As alcohol content rises, the normally solitary cells begin to adhere to each other in clumps called flocs. Work on these flocs, just published by Smukalla et al. in Cell1, shows that the yeast cells face a familiar social dilemma, but that they solve it in an exotic fashion.
In addition to its humble jobs as brewer and baker, yeast has added a high-tech career. It has become the principal laboratory model organism for studying the biology of eukaryotes — organisms such as plants and animals that have a membrane-bound nucleus. We have therefore come to understand yeast as well as any organism. Yet, although brewers have known about flocculation for centuries — sedimentation of the flocs provides an easy way to remove the yeast and keep the beer from tasting like the yeast paste Marmite — biologists have been slower off the mark. The problem is that flocculation was lost during the domestication of yeast. Smukalla et al.1 therefore chose to study feral strains. They show that expression of one of five flocculation cell-adhesion genes, FLO1, explains much of the variation in flocculation (Fig. 1). They establish causation by expressing FLO1 in the domesticated strain, resurrecting flocculation, and putting FLO1 into Saccharomyces paradoxus, inducing flocs even in this non-flocculating species.
Flocculation is a true cooperative trait that poses the classical social dilemma of how to sustain cooperation in the face of cheaters. Smukalla et al. show that cells on the inside of flocs are protected against damage from chemicals, including alcohol, partly because of physical shielding provided by cells on the outside. But those gains come at a cost; cells that express FLO1 grow more slowly, even if they are prevented from flocculating. So, is it possible for cheaters to gain the protection of flocs without paying the cost?
The answer seems to be no, according to Smukalla et al., because the ticket for admission to the floc is the ability to make the adhesion protein. It is just as if the bonhomie of the beer hall extended only to those drinking the same brand of beer or, on St Patrick's Day, only to those drinking green beer. In behavioural ecology, such tokens of inclusion are named not for green beer but for green beards. The name is fanciful because, until recently, such genes were purely imaginary. It is well known that genes for helping can spread by benefiting relatives, who share the gene with a specifiable probability. The alternative is to identify and help actual bearers of the gene, whether they are relatives or not2. Richard Dawkins3 argued that such green-beard genes were unlikely to occur because they would need to cause three traits: a label, such as a green beard; recognition of others with the label; and nicer behaviour towards those with the label.
Green-beard systems have, however, begun to leave the realm of the imaginary. The initial examples involved nasty behaviour towards those lacking the gene. In the fire ant Solenopsis invicta, a linked set of alleles, including one encoding an olfactory receptor, causes workers to kill queens that lack the green-beard allele4. Microbes are proving a richer source of green-beard genes. Many bacteria act rather like fire ants, using linked poison-antidote genes to kill members of the same species that don't possess the antidote gene5. The social amoeba Dictyostelium discoideum provided the first single-gene example and the first altruistic one. Its csaA gene codes for a homophilic adhesion protein, one that grabs on to the same protein on other cells, so excluding strains that do not express the protein from the benefits of later altruism within the group6.
The yeast FLO1 green-beard gene adds a new level of interest because it is highly variable, and because the variation — in the number of repeats of a 100-base sequence — exerts a major influence on protein-binding strength. Conceivably, such variation could lead to multiple recognition tags (green beards, red beards, yellow beards and so on), although this possibility remains to be tested. Alternatively, the variation might be the product of highly complex adaptive dynamics, which have been observed in simulation models as various green-beard cooperators and cheaters rise and fall in frequency7.
The discovery of FLO1 also extends the reach of single-gene green beards. It is easy to see how homophilic adhesion molecules such as the product of csaA could recognize each other. In fact, it was predicted that they could be green beards8. But FLO1 makes a heterophilic adhesion protein, one that binds not to itself but to oligosaccharides on the walls of other cells. Cells without the adhesion protein can still be bound by those that have it, but the binding is weaker than proper two-way binding, so such cells usually remain outside flocs. Even worse, the flocless cells that do get into flocs are exploited — they tend to end up in the outer layer, where they can be damaged in protecting the floc cells inside.
The discovery of a green beard in an organism as mundane as yeast suggests that such genes might prove to be quite common, at least in microbes. More generally, microbial sociality itself is probably more common than lab studies might indicate. In model systems, domestication can easily lead to loss of social behaviours9. Another example is the bacterium Bacillus subtilis, in which multicellular fruiting bodies were seen only when wild strains were studied10. Much like the entrepreneur who rescued Marmite from the discards of the brewery, scientists can find value in the traits discarded by the domesticators.