Equivalent binding sites reveal convergently evolved interaction motifs

Andreas Henschel 0 Wan Kyu Kim 0 Michael Schroeder 0 0 Bioinformatics Group, Biotechnological Centre , TU Dresden, Germany Motivation: Much research has been devoted to the characterization of interaction interfaces found in complexes with known structure. In this context, the interactions of non-homologous domains at equivalent binding sites are of particular interest, as they can reveal convergently evolved interface motifs. Such motifs are an important source of information to formulate rules for interaction specificity and to design ligands based on the common features shared among diverse partners. Results: We develop a novel method to identify non-homologous structural domains which bind at equivalent sites when interacting with a common partner. We systematically apply this method to all pairs of interactions with known structure and derive a comprehensive database for these interactions. Of all non-homologous domains, which bind with a common interaction partner, 4.2% use the same interface of the common interaction partner (excluding immunoglobulins and proteases). This rises to 16% if immunoglobulin and proteases are included. We demonstrate two applications of our database: first, the systematic screening for viral protein interfaces, which can mimic native interfaces and thus interfere; and second, structural motifs in enzymes and its inhibitors. We highlight several cases of virus protein mimicry: viral M3 protein interferes with a chemokine dimer interface. The virus has evolved the motif SVSPLP, which mimics the native SSDTTP motif. A second example is the regulatory factor Nef in HIV which can mimic a kinase when interacting with SH3. Among others the virus has evolved the kinase's PxxP motif. Further, we elucidate motif resemblances in Baculovirus p35 and HIV capsid proteins. Finally, chymotrypsin is subject to scrutiny wrt. its structural similarity to subtilisin and wrt. its inhibitor's similar recognition sites. Contact: Supplementary informaton: A database is online at scoppi.biotec. tu-dresden.de/abac/ The Author 2005. Published by Oxford University Press. All rights reserved. For Permissions, please email: - INTRODUCTION Protein interactions underlie all cellular processes and are important to reveal function. The interactions from known three-dimensional structures have been of particular interest in that they allow a number of detailed analyses of interfaces in terms of physico-chemical properties, shape and geometry [Jones and Thornton (1996); Conte To whom correspondence should be addressed. et al. (1999); Bashton and Chothia (2002); Chakrabarti and Janin (2002); Nussinov et al. (1997); Ofran and Rost (2003)]. The rapid growth of multichain and multidomain structures in Protein Databank (PDB) [Berman et al. (2000)] enabled systematic analyses of domaindomain interactions and interfaces [Park et al. (2001); Bolser et al. (2003); Apic et al. (2001); Kim et al. (2004)] and several databases dedicated to the collection of structural domain domain interactions are available [Finn et al. (2005); Stein et al. (2005); Davis and Sali (2005)]. Much work has concentrated on understanding under what circumstances homologous interactions are conserved [Pazos and Valencia (2001); Aloy et al. (2003); Tsai et al. (1996); Torrance et al. (2005)]. Aloy et al. (2003) did an extensive analysis on the relationship between sequence similarity and binding orientation and showed the geometry of interaction tends to be conserved between highly similar pairs. An alternative approach is to investigate how non-homologous proteins bind at equivalent surfaces of homologous proteins [Tsai et al. (1996)]. Such interactions do not necessarily compete in vivo, but they reveal equivalent interaction sites. In some cases, the interactions may be truly competitive and regulated temporally by chemical modification or regulatory factors and spatially by compartmentalization. Independent of competitive or noncompetitive binding, the identification of equivalent interfaces is a pointer to convergently evolved motifs. The motifs help to reveal key features which are necessary for the interaction. A well-known example of a convergently evolved motif is the catalytic triad (Ser, His, Asp) found in both chymotrypsin and subtilisin (e.g. Fig. 4a). The local features of the enzymes catalytic sites are conserved in other enzymes [Torrance et al. (2005)]. Chymotrypsin and subtilisin do not share any sequence or structure similarity. Indeed, both belong to different classes with chymotrypsin consisting only of beta-sheets and subtilisin of betaalphabeta units. Despite this different architecture, there are various inhibitors, which inhibit both enzymes and which use the same interface to do so. Thus, despite non-homology of the enzymes, equivalent binding sites are used. Consider Figure 1a. To elucidate such interfaces with convergently evolved motifs, we screen the known structures in PDB for pairs of interactions A B and A0 C, where B, C are from different superfamilies and A, A0 from the same family. If B and C bind to equivalent sites of A and A0, respectively, we label B and C as interfaces with convergently evolved motifs. To define the equivalence of interfaces we use sequence and structure alignments of the shared domains A and A0. If there is sufficient overlap in the sequence alignment of A and A0s interface residues and if the angle between interfaces of B and C after superimposition of A and A0 is sufficiently small, B and C bind at equivalent sites. In our analysis we use Structural Classification of Proteins (SCOP) domains [Murzin et al. (1995)]. We require A and A0 to be of the same family, since interfaces are known to be more conserved in both sequence and structure within a family [Valdar and Thornton (2001)], but not across the families of a superfamily [Rekha et al. (2005)]. For B and C we require different superfamilies, which ensures that they are evolutionarily not related. The method sketched above identifies all pairs of interfaces with convergently evolved motifs. One application of such a resource is the study of viral proteins, which mimic the interfaces of native proteins and can thus interfere accordingly. We discuss two such cases: the M3 protein, which mimics the chemokine homodimer interface and the regulatory factor Nef found in HIV, which mimics a kinase interface when interacting with SH3. To identify non-homologous domains binding at equivalent sites we proceed as illustrated in Figure 1a: we consider all pairs of interactions A B and A0 C, where A and A0 belong to the same family and B and C to different superfamilies. If B and C bind at equivalent interfaces, we screen them for shared motifs. To define equivalent binding sites, we use a two-stage procedure: first, we scan for a significant interface residue overlap on the aligned sequences; second, the angle and the spatial overlap between the two interfaces are used to further refine the sea (...truncated)