ABSTRACT: The persistence of genetic variation in master regulators of gene expression, such as quorum-sensing systems, is hard to explain. Here, we investigated two alternative hypotheses for the prevalence of polymorphic quorum sensing in Gram-positive bacteria, i.e., the use of different signal/receptor pairs ('pherotypes') to regulate the same functions. First, social interactions between pherotypes or 'facultative cheating' may favor rare variants that exploit the signals of others. Second, different pherotypes may increase fitness in different environments. We evaluated these hypotheses in the invertebrate pathogen Bacillus thuringiensis, using three pherotypes expressed in a common genetic background. Facultative cheating could occur in well-mixed host homogenates provided there was minimal cross talk between competing pherotypes. However, facultative cheating did not occur when spatial structure was increased in static cultures or in naturalistic oral infections, where common pherotypes had higher fitness. There was clear support for environment-dependent fitness; pherotypes varied in responsiveness to signals and in mean competitive fitness. Notably, competitive fitness varied with group size. In contrast to typical social evolution models of quorum sensing which predict higher response to signal at larger group size, the pherotype with highest responsiveness to signals performed best in smaller hosts where infections have a lower pathogen group size. In this system, low signal abundance appears to limit fitness in hosts, while the optimal level of response to signals varies in different host environments.IMPORTANCE Quorum sensing describes the ability of microbes to alter gene regulation according to their local population size. Some successful theory suggests that this is a form of cooperation, namely, investment in shared products is only worthwhile if there are sufficient bacteria making the same product. This theory can explain the genetic diversity in these signaling systems in Gram-positive bacteria, such as Bacillus and Staphylococcus sp. The possible advantages gained by rare genotypes (which can exploit the products of their more common neighbors) could explain why different genotypes can coexist. We show that while these social interactions can occur in simple laboratory experiments, they do not occur in naturalistic infections using an invertebrate pathogen, Bacillus thuringiensis Instead, our results suggest that different genotypes are adapted to differently sized hosts. Overall, social models are not easily applied to this system, implying that a different explanation for this form of quorum sensing is required.
