Biology Reference
In-Depth Information
Blunt, Liar, Ignorant]. It is important to note that in all cases some degree of social
cheating occurs in the form of exploitation or parasitism. Thus our analysis predicts
the large-scale occurrence of social cheating in microbial populations.
Two of the polymorphisms mentioned above merit more elaborate discussion.
The Janus Head of QS
In cases where the cooperation gene is (almost) fixed, one might at first sight expect
a monomorphic unconditionally cooperating (“Blunt”) population, because Blunt is
the cooperator with the lowest metabolic costs, and in a fully cooperating popula-
tion QS is not needed to obtain information about the potential level of cooperation.
However, we find next to “Blunt” appreciable frequencies of fully QS (“Honest”) and
partially QS (“Vain”) cells. The reason appears to be that here QS is selected because
it allows exploitation of Blunt strains, which unconditionally cooperate. As soon as a
quorum of Blunts is present, the other cells need not cooperate themselves in order to
profit from the cooperation benefit. Adoption of the QS machinery allows them to do
precisely this, since in such cases the level of signal is too low to trigger their coopera-
tive behavior. This phenomenon is an unexpected and novel result, showing that QS
not only prevents wasting resources when too few potential cooperators are around,
but also allows cells to parasitize on unconditionally cooperating neighbors, when a
sufficient number of those are present. It may occur at a large scale when the gene for
cooperation is (almost) fixed in the population, due to a favorable benefit/cost ratio of
the cooperative behavior and the quorum threshold relatively high. As explained more
fully in the Results section, 100% cooperating populations seem to evolve in most
cases to a [Blunt, Honest, Vain] mixture, characterized by a cyclic interaction pattern
(Blunt>Vain>Honest>Blunt) reminiscent of the rock-scissors-paper (RSP) game [20,
21].
Spiteful Behavior
The second polymorphism we want to call attention to is the coexistence of Honest,
Liar, and Ignorant, which occurs for example, at n
q
= 4 and n
q
= 5 for certain diffusion
values. Clearly, the non-cooperative Ignorant and Liar cells exploit the Honest cells
which provide cooperation benefits. Here the selective advantage of Liar is at first
sight remarkable, since it pays a higher metabolic cost than Ignorant, and can only
expect the same share as Ignorant from the cooperation benefit made available by the
Honest cells. The only effect of Liar is to sometimes induce Honest cells to cooperate
when actually less than the quorum of Honest is present. Nothing is gained, except that
in such cases Honest is paying the cooperation cost without getting the benefit. Thus
this coexistence is a clear example of spiteful behavior. Liar lowers the relative fitness
of Honest, but also pays a fitness penalty (the cost of signaling) itself.
How do our results relate to previous theoretical and empirical work on the evo-
lution of QS and cooperation? They confi rm the basic result from earlier theoretical
analyses of the evolution of QS [15], which predicted a stable polymorphism in mi-
crobial populations between QS and non-QS strains. The conclusion of Brown and
Johnstone that the highest level of QS signal expression is expected for intermediate
