From Protein Structure to Function with Bioinformatics.pdf

6 Function Diversity Within Folds and Superfamilies 153By definition, a duplication event gives rise to homologous genes. But furtherdistinctions can be made. Genes that descend from a common ancestor gene viaduplication within a given genome and in the absence of an accompanying speciationprocess are known as paralogues. Genes that descend from a common ancestorgene via duplication of the genome itself during speciation are known as orthologues.It is generally assumed that orthologous genes tend to preserve the functionof the ancestor gene, due to a strong selective pressure to ensure that the ancestralfunction is still performed in both descendant species (Tatusov et al. 1997). Basedon this assumption, some authors even define orthologues as homologues that havethe same function in different species. Several databases have been set up to defineorthologous genes from different sets of organisms (Dolinski and Botstein 2007).On the contrary, the presence of multiple copies of a given gene within a genome,i.e. paralogues, could arguably often result in one of the copies being under strongselective pressure to maintain the original function thus allowing more freedom fordivergence for the other copies. The process by which one copy of a duplicatedgene conserves the function of the ancestor gene, whereas the other copies evolvealternative functions is known as neofunctionalisation. In the absence of selectivepressure on these additional copies, a frequent outcome of evolutionary divergenceis the loss of some of them into pseudo-genes, which are gene relics no longerexpressed (Harrison and Gerstein 2002). This evolutionary process is callednonfunctionalisation. Subfunctionalisation is a third evolutionary process whichrefers to cases where multiple functions of an ancestral gene are divided betweenthe paralogues. In any case, paralogues are often considered to be more functionallydiverse than orthologues because of their larger freedom to diverge.In whatever order they occurred, the subsequent events of duplication into orthologuesor paralogues that took place during biological history have resulted in thecurrent protein superfamilies. Not all superfamilies seem to have been equally successfulin this process, as some of them are known to account for disproportionatelylarge numbers of genes in fully sequenced genomes (Marsden et al. 2006). To date,reasons for evolutionary success disparity of the different superfamilies are not clear,and arguments relating to structural and functional properties, or evolutionary dynamicshave been proposed (Goldstein 2008). It can be expected that older superfamilies,having had more time to diverge and explore different functions, should generally bemore extended in present time. For example, the HUP superfamily (CATH code3.40.50.620), which on account of phylogenetic considerations is believed to traceback to the RNA world, displays a very wide array of seemingly unrelated functions(Aravind et al. 2002); whereas several recent superfamilies that are observed exclusivelyin eukaryotic species are often restricted to very specific sets of functions.However, the age doesn’t seem to be the main factor explaining the varying sizes ofsuperfamilies. In a recent analysis of evolutionary dynamics of gene families thatcontain genes with essential functions (termed E-families) and gene families that donot contain such genes (termed N-families), it was proposed that paralogues in E-families are more likely to evolve new functions than those in N-families thus suggestingthat the function of ancestral genes in a family is a key determinant of itsevolutionary success (Shakhnovich and Koonin 2006). As will be shown in the nextsection, other arguments to explain the variable success of protein superfamilies mayderive from the mechanisms that have been proposed to explain function evolution.