Letter to the Editor

LETTER TO THE EDITOR Dear Editors: In our paper, ''Adaptive evolution ofhighlymutable loci in pathogenic bacteria,''l we reviewed the evidence that some bacteria have highly mutable loci that influence the expression ofa subpopulation ofchromosomal genes. We argued that the coexistence of these highly mutable genes with those that have much lower mutation rates is the result of adaptive evolution. High rates ofmutation in such genes facilitate the efficient exploration ofphenotypic solutions to unpredictable aspects ofthe environment . We also suggested that similarly hyper-mutable genes might be found in other organisms, including the hosts that bacteria parasitise. Concerning this latter comment , we have been gently taken to task in a recent essay by Neil Greenspan ( "Genomic logic, allelic inference and the functional classification ofgenes").2 The crux of Greenspan's objections is our use of the term 'contingency' to identify genes with high, sequence dependent, rates of mutability. In his essay, Greenspan opts for a definition that is based on the functions ofmolecules encoded by genes, a perfectly reasonable preference, albeit imparting a different meaning. He discussed eucaryotic genes that apparently do not have higher than usual mutation rates, but which interact with numerous different microbes, and considers that it would be ". . . preferable to consider the C3 alleles to be contingency genes, at least with respect to their functions in host defense." We think it is unreasonable for Greenspan to charge us with inadequacies ofa scheme ofgene classification that, ipso facto, we did not use in the first place! We accept his conclusion that "... the extent of sequence variation in a gene does not logically determine the class of function," but not with his assertion that this is what our paper was proposing. Indeed, based on our subsequent discussions conducted by e-mail, there appear to be few, if any, fundamental differences in our views. Rather, our views are complementary and we would be disappointed if mere semantics (differences in definition) were to undermine our respective efforts to understand how organisms evolve mechanisms for processing information in the environment and respond to its myriad challenges. Greenspan questions the legitimacy and utility of classifying genes into those with high and low mutation rates. A gene may evolve higher rates of mutation or change its primary function and, therefore, this will lead to an imperfect correspondence between function and mutability. This is certainly true, but what we emphasised in our paper is that, at any particular point in time, there can be a very strong, although imperfect, association between function and mutability. That this association has predictive power was subsequently demonstrated when the hypothesis was put to the test through the availability ofwhole genome sequences ofpathogenic bacteria. The results provide strong support for the utility of sequence motifs (tandem DNA repeats) , as a method of identifying novel bacterial genes involved in host-microbial interactions, examples being genes for lipopolysaccharide and flagellin biosynthesis , haemoglobin-binding genes and outer membrane proteins.3,4 The high mutation rates associated with repetitive DNA are one, but of course not the only, determinant of an organism's capacity for variation, one that has implications for adaptation ofmicrobes because it facilitates "... the efficient exploration of pheno154 Letter to the Editor typic solutions to unpredictable aspects of the host environment while minimising the deleterious effects of hypermutability on an organism's fitness."1 The entity that we sought to emphasise is intergenic variation in mutation rates, that is the de novo production ofvariation and not the standing level ofvariability in a population. We used contingency in the context of localised hypermutability of procaryotes; the somatic hypermutability of B lymphocytes in the generation of antibody diversity could be cited as an example in eucaryotes. In contrast, the significant and major focus of Greenspan's essay is the importance of gene variation in human populations, exemplified by allelic variations in host defence genes (e.g., C3, globin) . For example, haemoglobin genes are highly polymorphic in the presence ofmalaria because of diversifying selection (having to do with gene frequency), not because they are hypermutable (having to do with rate) . Yet, both hypermutability and allelic variation are clearly relevant to the interactions of microbes and hosts. As the human genome project makes available...