The function and structure of the human complement protease, MASP-3

Mannose-binding lectin (MBL)-associated serine proteases (MASPs) activate the lectin pathway of the complement system, an integral part of the human innate immune response. Thus far, three MASP enzymes (MASP-1, -2, and -3) and two non-catalytic derivatives, MAp19 and 44, have been identified. MASP-1, MASP-3 and MAp44 are products of alternative splicing of the same masp-1/3 gene. MASP-1 and MASP-3 enzymes only differ in their C-terminal serine protease domains and thus are identical in their five N-terminal non-catalytic domains. However, the fact that MASP-3 is conserved among all vertebrate species examined, including those in which MASP-1 is not present, indicates that MASP-3 may play a role different to MASP-1 in vertebrate species. MASP-3 is ubiquitously distributed in the human body and exists in human plasma at a relatively high concentration of around 5 μg/ml, while its role in activation of complement is unlikely, since it does not cleave complement components C2, C3 and C4, and it is not inhibited by C1-inhibitor. Recent genetic studies have identified a critical correlation between genetic defects of MASP-3 and an autosomal recessive disorder, the Carnevale, Mingarelli, Malpuech and Michels (3MC) syndrome, indicating a possible role of MASP-3 in embryonic development. However, knowledge of the structure and function of MASP-3 was modest at the outset of this study. MASP-3 is secreted as a single-chain zymogen and is activated by cleavage at the specific R449-I450 activation bond. The lack of knowledge of the enzyme responsible for activation of MASP-3 has cast a great level of uncertainty on previous studies using active MASP-3. This study provides the first demonstration that MASP-2, not MASP-1, activates MASP-3. However, the presence of the R454 N455 peptide bond near the activation bond limits the specificity of this activation. Further, data derived here also first identified a specific in vitro activation of MASP-3 by the human proprotein convertase enzymes, furin and PACE4, which specifically cleave after dibasic residues, present in the activation site of MASP-3. However, the slow rate of cleavage indicates that the conditions of such activation need to be optimized. In order to achieve efficient activation of MASP-3, two MASP-3 mutant proteins, M3EEKQ and M3Q, were constructed, both of which contain sequences N terminal to the activation bond of the classical complement pathway serine protease homologue, C1s, which is specifically activated by C1r. Both M3EEKQ and M3Q were activated effectively and uniformly by recombinant C1r. The active forms of MASP-3 mutants consist of a serine protease domain identical to that of wild type MASP-3. Further, the substrate specificity of MASP-3 was characterized using combinational peptidyl substrate libraries, showing that MASP-3 most prefers the sequences GRIF/GRIY for cleavage, with specificity for R/K residues at the P1 position. The fluorescence-quenched tripeptide substrate, RIF, was found to be 20-fold more effectively cleaved by MASP-3 than the AMC substrate, VPR, which was previously claimed to be among the best substrates for MASP-3. The present study of the 3MC syndrome-associated MASP-3 G666E and G687R mutants revealed that the active forms of these two mutants fail to cleave any of the substrates tested, including the optimal substrate, RIF. This data is therefore the first piece of evidence that conclusively shows that the 3MC syndrome is always related to a lack of MASP 3 activity. Solution of the three-dimensional structure of the zymogen form of the G666E mutant of MASP-3 to 2.6 Å shows that the mutation most likely results in an aberrant ionic bonding in the active site of the enzyme that is unlikely to be altered upon activation of MASP-3, thus providing the first insight into the mechanism whereby the mutation causes enzyme inactivation. As a summary, by identifying the enzyme candidates responsible for activation of MASP-3, developing MASP-3 mutants that can be efficiently activated, identifying the optimal substrate for cleavage by MASP-3, solving the first structure of a 3MC syndrome-associated MASP-3 mutant and illustrating the molecular basis of the inactivation of this mutant, the present study has contributed strongly to knowledge about MASP-3, which had previously been very little characterized. These findings also provide a strong background for further study of the enzyme.