Genomic Logic, Allelic Inference, and the Functional Classification of Genes

GENOMIC LOGIC, ALLELIC INFERENCE, AND THE FUNCTIONAL CLASSIFICATION OF GENES NEIL S. GREENSPAN* Introduction An integral task of any branch of science is the classification of the entities that are the focus ofthat discipline. Attempts to assign genes to various categories are based on the expectation that such categories will assist biologists in thinking about relationships among genes and between genotypes and phenotypes. For example, it has been proposed that the genes ofsome bacterial pathogens can be classified by reference to two categories: (1) housekeeping genes, which encode gene products engaged in predictable interactions necessary for survival and which manifest low rates of mutation , and (2) contingency genes, which deal with threats to survival of an unpredictable nature and which exhibit relatively high rates of mutation [I]. This scheme derives in large part from the discovery of eis elements that increase the probability of alterations in nucleotide sequence or gene expression at particular loci in pathogenic bacteria such as Haemophilus influenzae and Neisseria gonorrhoeae. These elements are found primarily at loci that encode products involved in interactions with host defense molecules . Examples of pertinent gene products include fimbriae, enzymes involved in lipopolysaccharide synthesis, and proteins involved in capsular polysaccharide production and transport. Moxon, et al., speculate that it may be possible to apply the same scheme to "any other organism whose genome can be crudely divided into contingency and housekeeping functions," and they specifically raise the prospect of application to eukaryotes [I]. This speculation, in conjunction with both well-established and recent findings, prompts consideration of the suitabilThe author wishes to Lhank Mitch Drumm, Robert Elston, David Kaplan, Dennis Templeton , and Hunt Willard for critical review of the manuscript, andJames V. Neel for encouragement . * Institute of Pathology, Case Western Reserve University, 10900 Euclid Avenue, Cleveland, OH 44106.© 1998 by The University of Chicago. All rights reserved. 0031-5082/98/4103-1038$01.00 Perspectives in Biology and Medicine, 41, 3 ¦ Spring 1998 | 409 ity of such a scheme of classification for eukaryotic genes, such as mammalian genes involved in host defense. The main thesis to be presented is that the logic of functionally classifying eukaryotic genes, as represented by mammalian genes related to host defense, is more challenging than is implicitly recognized in frameworks of the sort proposed by Moxon, et al. Several facets of this complexity of gene classification are described below. Limitations of the Dichotomy Between Housekeeping and Contingency Genes Conventionally, the immune response is divided into adaptive, antigenspecific and innate, antigen-nonspecific arms [2]. The adaptive immune response relies on, among other molecules, diverse immunoglobulins, aß T cell receptors, and major histocompatibility complex (MHC) class I and class II molecules. Immunoglobulin and T cell receptor genes are encoded at many loci and are subject to diversification during ontogeny, through gene rearrangements and other mechanisms operating in B and T lymphocytes , respectively [2]. MHC genes are highly diverse on the basis of unremarkable rates of mutation coupled with selection for variants [3] . In contrast , the innate immune response relies on types of molecules that are encoded by genes exhibiting relatively modest levels of true polymorphism or rates of mutation [2, 4]. Selective forces acting on these genes can be presumed to largely favor conservation of nucleotide and amino acid sequences . Examples of this class of molecules are the components of the classical, alternative, and lectin complement pathways. While it is customary to distinguish between the genes and molecules involved in mediation of antigen-specific and innate immune responses, these two sets of genes and gene products interact extensively [4], and both sets can be regarded as existing primarily to mediate the common function of host defense. Based on the criterion of predictability of gene product interactions, functions related to host defense would, a priori, be more satisfactorily classified as contingency functions than as housekeeping functions, since the precise challenges to the host defense system vary unpredictably over time. On a given day, a host organism cannot prospectively determine which (if any) particular pathogens or antigens will be encountered. However, a difficulty arises in classifying the genes encoding components of the innate immune response if the chief criterion is the extent of predictability of interactions with the environment. Do...