Biology Reference
In-Depth Information
levels of relatedness between interacting strains [16] is confi rmed in our study only
for cases when the cooperation cost is low and the required quorum threshold is also
low. Then the benefi t of cooperation is relatively easy to obtain, and the QS machin-
ery is too costly to operate. Only when the spatial population mixing becomes more
intensive, causing the predictability of the neighborhood composition to go down,
QS becomes profi table. The situation is different for costly cooperation and/or a high
quorum threshold. Apparently, then QS is profi table even at a very low rate of disper-
sal (i.e., at high relatedness between interacting cells) because of the lower level of
cooperation in the population and/or the greater sensitiveness to increased dispersal.
The available empirical observations on natural occurrence of QS cheats are main-
ly from work on
Pseudomonas aeruginosa
[13, 27-31]. Although these experimental
studies cannot yet be informative with respect to the full spectrum of possible mutants
and only focus on one or two QS mutants, they report a considerable level of social
cheating, which is in agreement with our study.
Finally, we mention an experimental result that may be of relevance with respect to
QS evolution but is not included in this model. It is related to the feedback-regulation
of QS signal production: “signal deaf” signaler mutants (in our notation:.Sr genotypes)
are shown to produce an excess of signal molecules compared to signal responsive
ones, because in the latter signal production is downregulated by the extracellular con-
centration of the signal itself, which response-defi cient mutants cannot sense [32, 33].
The effect of signal over-expression on the dynamics of QS evolution require further
theoretical work.
In conclusion, we predict that the evolution of QS as a communication system
regulating cooperative behavior such as the production of a public good has two strik-
ing effects. First, it enables the cooperative behavior to attain a higher frequency in the
population, and second, it opens up a remarkably rich repertoire of social interactions
in which cheating and exploitation are common place.
RESULTS
The Evolution of Cooperation Without Quorum Sensing
We first have performed simulations with the QS functions disabled (mutation rates
in both ways set to 0.0 at the S and the R loci). Without QS allowed, the only possible
genotypes are the “Ignorant” (no cooperation) and the “Blunt” (unconditional coop-
eration), of course.
(1) Cooperation is Relatively Costly (m
c
= 30)
The left column in Figure 1 summarizes the results. Cooperation is only selected under
a very low degree of dispersal. This confirms the essential role played by kin selection
in the evolution of cooperation, since low dispersal in a microbial population implies
that most social interactions are among related individuals. With a low quorum thresh-
old (only few cooperators are necessary to provide the benefit to all the immediate
neighbors) there is much scope for non-cooperators to parasitize, because sufficiently
often they can enjoy the benefit from cooperative neighbors without paying the cost
of cooperation themselves; therefore, with n
q
= 2 and n
q
= 3, only a minority of the
