Biol. Chem., Vol. 390, pp. 503–507, May/June 2009 • Copyright  by Walter de Gruyter • Berlin • New York. DOI 10.1515/BC.2009.064 2009/270 Article in press - uncorrected proof Short Communication Cooperative effects in the substrate specificity of the complement protease C1s Sarah E. Boyd 1 , Felicity K. Kerr 2 , David W. Albrecht 1 , Maria Garcia de la Banda 1 , Natasha Ng 2 and Robert N. Pike 2, * 1 Clayton School of Information Technology, Monash University, Clayton, Victoria 3800, Australia 2 Department of Biochemistry and Molecular Biology, Monash University, Clayton, Victoria 3800, Australia * Corresponding author e-mail: rob.pike@med.monash.edu.au Abstract Complement is a key component of the immune system, but can contribute to inflammatory diseases. The sub- strate specificity of C1s protease has been successfully investigated using a combinatorial approach, while a positional scanning method failed. The lack of success of the latter approach is possibly due to cooperativity in the active site, which could confound such analyses. With a panel of peptides devised using factorial design, we show pronounced cooperativity between the S 4 and S 1 9 subsites in the active site of the enzyme, and weaker cooperativity between the S 1 9 and S 3 9 subsites. The use of factorial design has promise as a methodology for determining cooperativity in protease active sites. Keywords: active site; C1s; complement; cooperativity; protease; substrate specificity. Complement is an integral effector of the immune response in both the innate and adaptive immune sys- tems (Janeway et al., 2001). However, unregulated acti- vation of this system can cause unwanted and harmful inflammation (Morgan and Harris, 2003). Regulation of the complement pathway is therefore an attractive approach to preventing or treating related inflammatory diseases. The complement pathway is a system of proteins tight- ly regulated by proteases (Janeway et al., 2001), and it has been proposed that effective control of complement could be achieved by inhibiting the serine proteases of this system (Morgan and Harris, 2003). In the classical pathway, there are two candidate serine proteases, C1r and C1s, both of which function in the initiating complex and are therefore logical targets for efficiently inhibiting the pathway. The C1s protease is of particular interest for therapeutic purposes, not only because of its role in the initiation of the pathway but also because it appears to interact only with complement components C2 and C4, and with the serine protease inhibitor (serpin) C1-inhibi- tor. This narrow range of physiological substrates and inhibitors for C1s suggests that the substrate specificity of the enzyme is restricted. The substrate specificity of a protease is mediated by the active site of the enzyme, which contains a fixed number of subsites, each of which recognises a consecutive residue in the substrate sequence (Schechter and Berger, 1967). Thus, C1s may be an attractive target for highly selective inhibition, and exploiting the substrate specificity of this protease could allow the development of potent inhibitors of complement. While approaches using combinatorial methods, such as phage display of peptides (Kerr et al., 2005), have suc- cessfully been used to map the specificity of C1s, posi- tional scanning approaches (O’Brien et al., 2003 and unpublished data) were less successful. Computational modelling in the PoPS program (Boyd et al., 2005) indi- cates that even though the active site specificity of C1s is less restricted than expected, it nevertheless appears to be an important factor in the selection of the physio- logical substrates observed to date (Kerr et al., 2005, 2007). A possible explanation for these observations may lie in the common assumption in specificity analysis that subsites act independently in the selection of substrates. Generally, the design and analysis of specificity studies assumes that each subsite contributes independently to the binding of residues and cleavage of the substrate. However, there is evidence to show that this assumption is incorrect in at least some instances, as illustrated in the accompanying review (Ng et al., 2009). It is clear that for at least some proteases, specific changes in the res- idue at one position in a substrate have a significant, non-linear effect on the specificity for residues at other positions. This phenomenon is commonly referred to as cooperativity. It should be noted that the term coopera- tivity could be used to describe both effects within the active site, as well as behaviour exhibited by allosteric enzymes. Here, cooperativity is used only in the former sense, i.e., to describe specificity effects that come from within the active site. We hypothesised that cooperativity in the active site of C1s was significantly influencing the selection of sub- strates. If confirmed, this hypothesis would help explain the highly selective nature of C1s and be a very important factor in the design of therapeutic inhibitors. To test our hypothesis, we used the statistical method of factorial design (Box et al., 1978) to design a peptide library for pairwise analysis of cooperativity in C1s specificity. Fac- torial design is a long-established standard experimental design, which provides particular features that make it ideal for this study. In the first instance, a naive library