A series of 17 new analogues of trypsin inhibitor SFTI-1 were designed and synthesized to obtain matriptase-2 inhibitors. A number of the modified bicyclic peptides displayed much higher affinity towards matriptase-2 than towards the highly homologous matriptase-1. Replacement of Lys5 by Arg in the wild-type SFTI-1 led to an 11-fold increase in the matriptase-2 inhibitory activity. Replacement of Arg2 by its enantiomer (D-arginine) slightly lowered the inhibition of matriptase-2, but almost completely abolished the affinity towards matriptase-1, thus yielding the most selective matriptase-2 inhibitor. This is the first report describing inhibitors of the recently discovered matriptase-2 based on the SFTI-1 structure. The results showed that SFTI-1 is a promising scaffold for the design of potent and selective inhibitors of this enzyme.

No abstract

This study examines peptide splicing catalyzed by serine proteinases. A series of two-peptide-chain analogs of trypsin inhibitor SFTI-1 were designed and synthesized via the solid-phase method. All consisted of two peptide chains (also called N-and C-terminal fragments) joined together by one disulfide bridge. The analogs were incubated with bovine b-trypsin or bovine a-chymotrypsin. Analysis of MS data analysis showed that, after enzyme-catalyzed degradation of the single peptide bond between the Lys and Ser residues located at the C-terminus of the C-terminal peptide chain, a new peptide bond was formed. This bond brought together the separated peptide chains, and, as a result, monocyclic SFTI-1 was recovered. This proteolytic route of peptide rearrangement appears to be similar to peptide splicing catalyzed by proteasomes. However, the proteasome is much more complex than 'classical' serine proteinases.

Serine-proteinase-catalyzed peptide splicing was demonstrated in analogues of the trypsin inhibitor SFTI-1: both single peptides and two-peptide chains (C- and N-terminal peptide chains linked by a disulfide bridge). In the second series, peptide splicing with catalytic amount of proteinase was observed only when formation of acyl-enzyme intermediate was preceded by hydrolysis of the substrate Lys-Ser peptide bond. Here we demonstrate that with an equimolar amount of the proteinase, splicing occurs in all the two-peptide-chain analogues. This conclusion was supported by high resolution crystal structures of selected analogues in complex with trypsin. We showed that the process followed a direct transpeptidation mechanism. Thus, the acyl-enzyme intermediate was formed and was immediately used for a new peptide bond formation; products associated with the hydrolysis of the acyl-enzyme were not observed. The peptide splicing was sequence- not structure-specific.

Protease inhibitors of the Bowman-Birk (BBI) family are commonly found in plants and animals where they play a protective role against invading pathogens. Here, we report an atomic resolution (1Å) crystal structure of a peptide inhibitor isolated from a skin secretion of a Chinese bamboo odorous frog Huia versabilis (HV-BBI) in complex with trypsin. HV-BBI shares significant similarities in sequence with a previously described inhibitor from a diskless-fingered odorous frog Odorrana graham (ORB). However, the latter is characterized by more than a 16,000 fold higher Ki against trypsin than HV-BBI. Comparative analysis of trypsin cocrystal structures of HV-BBI and ORB and additionally that of Sunflower Trypsin Inhibitor (SFTI-1) together with accessory information on the affinities of inhibitor variants allowed us to pinpoint the inhibitor moiety responsible for the observed large difference in activity and also to define the extent of modifications permissible within the common protease-binding loop scaffold of BBI inhibitors. We suggest that modifications outside of the inhibitory loop permit the evolution of specificity toward different enzymes characterized by trypsin-like specificity.