From Protein Structure to Function with Bioinformatics.pdf

6 Function Diversity Within Folds and Superfamilies 155within CATH superfamilies showed that not only do inserted secondary structurestend to co-locate in the fold but that the resulting embellishments often occur closeto functionally important regions such as enzyme catalytic sites or protein-proteininterfaces (Reeves et al. 2006); this observation indicates a correlation betweenstructural and functional changes.Insertions of new elements of secondary structure near the active site will mostlikely change the function, but more subtle changes such as residue substitutions ofimportant catalytic residues will also result in functional differences, even thoughthese are more likely to distinguish relatively close homologues than remote homologuesat the superfamily-level discussed here. Changes in domain context can alsoresult in drastic changes in the role of proteins, so that even if some aspect ofmolecular function is conserved, it can hardly be said of the proteins that their functionis the same (Hegyi and Gerstein 2001; Todd et al. 2001). This is the case of thePBP-like domains of eukaryotic and prokaryotic glutamate receptors, which bindthe same ligand in a similar topological location, but widely differ in their functionat the cellular level (see Fig. 6.2).The long-term evolutionary processes via which function can diverge betweenhomologues are numerous and difficult to summarise. Nevertheless, in a recentattempt to understand and categorise such processes, Bashton and Chothia havedescribed and illustrated a subset of these to understand how the function of homologousdomains can change depending on whether they are found in the context of single-domainproteins or combined with other domains in multi-domain proteins(Bashton and Chothia 2007). Examples of the processes identified include caseswhere the domain function is modified by its combination with other domains thatmodify its substrate specificity, or cases where the fusion of domains results in multifunctionalproteins in which each domain is responsible for a particular function.The above-mentioned occurrence of structural changes in the vicinity of functionalregions points at the resulting functional diversity that is to be expected between superfamilymembers. And indeed, results from several studies indicate that remote homologueswithin superfamilies often perform very different functions (Todd et al. 2001).Most of these studies are focused on the evolution of function within particular superfamiliesthat generally show exceptional functional diversification, and prominentexamples of which include haloacid dehalogenases (Burroughs et al. 2006), shortchaindehydrogenases/reductases (Favia et al. 2008), enolases (Gerlt and Babbitt2001), HUP domains (Aravind et al. 2002) or “Two dinucleotide binding domains”flavoproteins (tDBDF’s) (Ojha et al. 2007). The study of these different groups of proteinshas revealed a large variety of processes by which function diverges between relatives,and these processes will now be considered separately with examples.Mechanistically Diverse SuperfamiliesA subset of much studied superfamilies constitute the core of the data in theStructure-Function Linkage Database (Pegg et al. 2006) and in spite of their functionaldiversity, and respecting the criteria of inclusion in SFLD (see Section