In a recent, mainly theoretical review, Laurent Keller and Mike Surette provide a cost–benefit analysis of cell–cell communication in bacteria1. The authors argue that it is costly for bacteria to produce signal molecules, as signal production represents a metabolic burden. Furthermore, the response to a signal might also incur a cost to individual cells. The question then arises as to whether it is advantageous for bacterial populations to consist purely of signalling-proficient cells or whether there might be situations in which 'cheaters' would be favoured. Cheaters are defective for signalling and they benefit from the metabolic activities of signalling-proficient strains, for example, from the expression of extracellular enzymes that degrade macromolecules.

There is a solid body of experimental evidence demonstrating that cheaters of the species Pseudomonas aeruginosa have a place in natural environments. P. aeruginosa naturally uses N-acyl-homoserine lactones (AHLs) for cell–cell communication (called quorum sensing). However, roughly 20% of all environmental and clinical isolates of P. aeruginosa are quorum-sensing-deficient — using the above terminology they could be called cheaters. Interestingly, ten recently published papers (reviewed in Ref. 2) have shown that cheaters of P. aeruginosa are not defective in signal production, they are signal-blind. That is, a vast majority of quorum-sensing-negative isolates of P. aeruginosa that have been analyzed by molecular tools are defective in the master regulator of quorum sensing, the transcription factor LasR. Without LasR function, cells cannot respond to AHL signalling. By contrast, defects in the signal-producing enzymes LasI and RhlI are rare in natural isolates, and when such defects do occur they seem to be associated with a primary lasR mutation2. It is easy to see the reason behind this. Only 0.01% of the total cellular amount of ATP is needed to make the AHL signal molecules N-(3-oxo-dodecanoyl)-homoserine lactone and N-butanoyl-homoserine lactone at biologically active (sub-micromolar) concentrations. This estimate is based on the ATP content that Escherichia coli uses to drive biosynthetic pathways and macromolecular syntheses3. By contrast, when the AHL signals activate the quorum-sensing response in P. aeruginosa, this activity consumes at least 5% of the total energy supply, as at least 5% of all P. aeruginosa genes are induced during this response4. So, for cheaters, it does not pay to abandon signal synthesis, it is much more rewarding not to react to the signals.

In vitro, quorum-sensing-proficient cells of P. aeruginosa have a selective metabolic advantage over quorum-sensing-deficient cells, for example, when casein or adenosine is the carbon source, because the degradation of these substrates is ensured by quorum-sensing-controlled enzymes5,6. However, opposite selective forces can also be demonstrated: in a nutrient-rich, alkaline environment lasR-negative mutants have a better chance to survive in stationary phase in comparison with the quorum-sensing-proficient parental cells, which are more vulnerable to cell lysis. Again, these conditions do not select for signal-deficient cells but for signal-blind cells6.