The increased availability of sequences raises the need for sequence to reactivity algorithms, SRA, which quantitatively predict the reactivity of members of a family of proteins from their sequence alone. An algorithm for predicting the standard free energies of association between six selected serine proteinases and a large subset of Kazal family inhibitors was just developed in our laboratory. It was tested by comparing its predictions to already known results for 92 natural avian ovomucoid third domains and on five other, quite unrelated, Kazal family inhibitors. The SRA allows us to predict the best possible, most specific possible and least specific possible sequences for the six enzymes. We plan to express these and test them as an additional important test. Numerous other design objectives are possible. This should be highly useful in drug design either directly as proteins or as models for peptidomimetics. The SRA is additivity based and the tests confirm that additivity holds very well for most residue pairs. However, the pair P2 and P1' is a clear exception and at this time, we cannot predict for pairs that we did not measure. We propose to remove this restriction thus greatly widening the SRA. A number of other restrictions will also be eliminated. The SRA will be more useful to biologists when it is extended to more enzymes. Additivity within a family is called intrascaffolding additivity. Additivity between families is called as interscaffolding additivity and will continue to be studied. Aside from the standard free energy of association, the enthalpy, entropy, heat capacity change, the on and off rate constants, the pH dependence and the equilibrium constant for the reactive site peptide bond hydrolysis are expected to be largely additive. These will be studied on the existing variant set. About 10 percent of the cases tested are nonadditive. These will be selected for detailed study with the aim of elucidating how additivity fails and incorporating corrections into the SRA for such failures. It is anticipated that the ability to correct will be even more important when defining SRAs for other less additive systems.
Showing the most recent 10 out of 13 publications
