Substrate Specificities of Recombinant Mannan-binding Lectin-associated Serine Proteases-1 and -2*

Mannan-binding lectin (MBL)-associated serine pro-teases-1 and 2 (MASP-1 and MASP-2) are homologous modular proteases that each interact with MBL, an oligomeric serum lectin involved in innate immunity. To precisely determine their substrate specificity, human MASP-1 and MASP-2, and fragments from their catalytic regions were expressed using a baculovirus/insect cells system. Recombinant MASP-2 displayed a rather wide, C1s-like esterolytic activity, and specifically cleaved complement proteins C2 and C4, with relative efficiencies 3- and 23-fold higher, respectively, than human C1s. MASP-2 also showed very weak C3 cleaving activity. Recombinant MASP-1 had a lower and more restricted esterolytic activity. It showed marginal activity toward C2 and C3, and no activity on C4. The enzymic activity of both MASP-1 and MASP-2 was specifically titrated by C1 inhibitor, and abolished at a 1:1 C1 inhibitor:protease ratio. Taken together with previous findings, these and other data strongly support the hypothesis that MASP-2 is the protease that, in association with MBL, inhib- was assessed by incubation of the proteases with excess molar ratios of C1 inhibitor for min at 37 in 50 m M 145 m M NaCl, 7.4, followed by SDS-PAGE analysis under reducing conditions. C1 inhibitor-protease complexes were revealed by Western blot analysis using a rabbit polyclonal antibody directed against full-length MASP-2, and a rabbit anti-peptide antibody directed against the serine protease domain of MASP-1. Ti-tration of the enzymatic activity of C1s, MASP-1, and MASP-2 by C1 inhibitor was performed by preincubating these proteases (at concentrations of 0.25 (cid:2) M , 20 n M , and 0.25 (cid:2) M , respectively) with increasing concentrations of C1 inhibitor (10–500 n M ) for 15 min at 37 °C. The residual C2 cleaving activity of MASP-2 and C1s was then measured by incubating the enzymes (3.75 and 7.5 n M , respectively) for 15 min at 37 °C in the presence of 2.25 (cid:2) M C2, followed by SDS-PAGE analysis, as described above. In the case of MASP-1, the residual esterolytic activity was measured on Bz-Arg-OEt, as described above. The reactivity of C1s and MASP-2 (cid:1) 2 M was measured by preincubating the enzymes (0.25 (cid:2) M each) for 30 min at 37 °C in the presence of increasing concentrations (3.7–300 (cid:2) M ) of (cid:1) 2 M. Residual C4 cleaving activity was then measured by incubating the enzymes (1.25 n M each) for 15 min at 37 °C in the presence of 2.25 (cid:2) M C4, followed by SDS-PAGE analysis. Esterolytic activities were measured on the synthetic esters Ac-Gly-Lys-OMe and Tos-Arg-OMe using a spectrophoto- metric assay based on the measurement of methanol released upon hydrolysis on the thioester Z-Gly-Arg-S-Bzl Bz-Arg- trations

MASP-2 cleaves C4 much more efficiently than does C1s, emphasizing the physiological relevance of MASP-2 with respect to complement activation. In contrast, recombinant MASP-1 shows only marginal C3 cleaving activity, raising questions about its involvement in complement activation.
Proteins-MBL was isolated from human plasma according to the procedure described by Tan et al. (26), with a further purification step using ion-exchange chromatography, as described in Thielens et al. (27). Recombinant full-length MASP-1 was expressed using a baculovirus insect cell system and purified by ion-exchange chromatography and affinity chromatography on an UltraLink TM -MBL column, as described previously (27). Activated C1s and complement proteins C4, C2, and C3 were purified from human plasma according to published procedures (28 -31). Partially purified C5 was prepared as described by Al Salihi et al. (31), and partially purified factor B was obtained from the flowthrough fraction of the last C2 purification step on C4b-Sepharose (30). C1 inhibitor was purified from human plasma essentially as described in Ref. 32, except that concanavalin A-agarose (Sigma) was used instead of concanavalin A-Sepharose. Purified human ␣ 2 -macroglobulin (␣ 2 M) was kindly provided by L. Sottrup-Jensen (University of Aarhus). The concentrations of purified proteins were determined using the following absorption coefficients (A 1 cm 1% at 280 nm) and molecular weights: C1s, 14.5 and 79,800; C4, 8.3 and 205,000; C2, 8.9 and 102,000; C3, 10.0 and 185,000 (29,30,33). The absorption coefficients (A 1 cm 1% at 280 nm) used for the recombinant fragments CCP1/2-SP and CCP2-SP of MASP-2 (18.3 and 19.2, respectively) were calculated from the number of Trp, Tyr, and disulfides according to the method of Edelhoch (34), using molecular weight values of 42,710 and 35,330, calculated from the sequence (6). Because of the low amounts of material recovered, the concentrations of full-length MASP-1 and MASP-2 and of the CCP1/ 2-SP fragment of MASP-1 were estimated on the basis of Coomassie Blue staining after SDS-PAGE analysis using appropriate internal standards and respective molecular weights of 82,000, 74,200, and 45,200 (6, 9).
Construction of Expression Plasmids Containing the CCP1/2-SP Fragment of MASP-1-A DNA fragment encoding residues 299 -699 of human MASP-1 was amplified by polymerase chain reaction using the Vent R polymerase and the MASP-1b/pDR 2 ⌬EF1␣ vector (25) as a template, according to established procedures. The sequence of the sense primer (5Ј-GAAGATCTCAATGAGTGCCCAGAGCT-3Ј) introduced a BglII restriction site (underlined) and allowed in-frame cloning with the melittin signal peptide of the pNT-Bac expression vector (17). The antisense primer (5Ј-GGAATTCTCAGTTCCTCACTCC-3Ј) introduced a stop codon (bold face) followed by an EcoRI site (underlined) at the 3Ј end of the fragment. The polymerase chain reaction product was digested with BglII and EcoRI, purified, and cloned into the BamHI/ EcoRI sites of the pNT-Bac vector.
Construction of Expression Plasmids Containing Full-length MASP-2 and MASP-2-derived Fragments-The DNA fragment encoding the MASP-2 signal peptide plus the mature protein (amino acids 1-671) was excised from the pBS-MASP-2 vector (6) by digestion with XhoI and EcoRI and cloned into the corresponding restriction sites of the pFast-Bac1 baculovirus transfer vector (Life Technologies, Inc.). DNA fragments encoding residues 298 -686 (CCP1/2-SP construct) and residues 365-686 (CCP2-SP construct) of human MASP-2 were amplified by polymerase chain reaction using the Vent R polymerase and the pFast-Bac MASP-2 vector as a template. The sequences of the sense primers (5Ј-CGGGATCCCCAGCCTTGCCCTTATCC-3Ј for the CCP1/2-SP construct and 5Ј-CGGGATCCCGACTGTGGCCCTCCTG-3Ј for the CCP2-SP construct) introduced a BamHI restriction site (underlined) and allowed in-frame cloning with melittin signal peptide of the pNT-Bac expression vector. The antisense primer (5Ј-GGAATTCTTAAAAA-TCACTAATTATGTTCTC-3Ј) was identical for both constructs and introduced a stop codon (bold face) followed by an EcoRI site (underlined). Both polymerase chain reaction fragments were digested with EcoRI and BamHI, purified, and cloned into the corresponding sites of the pNT-Bac vector. All DNA constructs used in this study were checked for the absence of mutations by double-stranded DNA sequencing (Genome Express, Grenoble).
Cells and Viruses-The insect cells Spodoptera frugiperda (Ready-Plaque Sf9 cells from Novagen) and Trichoplusia ni (High Five TM ) were routinely grown and maintained as described previously (27). Recombinant baculoviruses were generated using the Bac-to-Bac TM system (Life Technologies, Inc.). The bacmid DNA was purified using the Qiagen midiprep purification system (Qiagen S.A., Courtaboeuf, France) and used to transfect Sf9 insect cells utilizing cellfectin in Sf900 II SFM medium (Life Technologies, Inc.) as described in the manufacturer's protocol. Recombinant virus particles were collected 4 days later, titrated by virus plaque assay, and amplified as described by King and Possee (35).
Protein Production and Purification-High Five cells (1.75 ϫ 10 7 cells/175-cm 2 tissue culture flask) were infected with the recombinant viruses at a multiplicity of infection of 2-3 in Sf900 II SFM medium at 28°C for 48 h (MASP-2 fragments) or 72 h (full-length MASP-2 and CCP1/2-SP fragment of MASP-1). Culture supernatants were collected by centrifugation.
The supernatant containing MASP-2 was dialyzed against 50 mM NaCl, 1 mM EDTA, 50 mM triethanolamine hydrochloride, pH 8.5, and loaded onto a Q-Sepharose Fast Flow column (Amersham Pharmacia Biotech) (2.8 ϫ 10 cm) equilibrated in the same buffer. Elution was carried out by applying a 800-ml linear gradient from 50 to 500 mM NaCl in the same buffer. Fractions containing the recombinant protein were identified by Western blot analysis, and dialyzed against 50 mM NaCl, 1 mM EDTA, 50 mM triethanolamine hydrochloride, pH 8.0. Further purification was achieved by ion-exchange chromatography on a Mono-Q HR 5/5 column (Amersham Pharmacia Biotech) equilibrated in the same buffer. Elution was carried out with a linear NaCl gradient from 50 to 500 mM in 90 min.
The supernatants containing the CCP1/2-SP and CCP2-SP fragments of MASP-2 were dialyzed against 5 mM EDTA, 20 mM Na 2 HPO 4 , pH 8.6, and loaded onto a Q-Sepharose Fast Flow column (2.8 ϫ 10 cm) equilibrated in the same buffer. Elution was carried out by applying a 700-ml linear gradient from 0 to 350 mM NaCl in the same buffer. Fractions containing the recombinant proteins were identified by Western blot analysis, dialyzed against 1.5 M (NH 4 ) 2 SO 4 , 0.1 M Na 2 HPO 4 , pH 7.4, and further purified by high-pressure hydrophobic interaction chromatography on a TSK-Phenyl 5PW column (Beckman) equilibrated in the same buffer. Elution was carried out by decreasing the (NH 4 ) 2 SO 4 concentration from 1.5 M to 0 in 30 min. Both purified recombinant fragments were dialyzed against 145 mM NaCl, 50 mM triethanolamine hydrochloride, pH 7.4, concentrated up to 0.2 mg/ml by ultrafiltration, and stored at Ϫ20°C.
The supernatant containing the CCP1/2-SP fragment of MASP-1 was dialyzed against 50 mM Na acetate, pH 5.1, and loaded on a SP-Sepharose column (Amersham Pharmacia Biotech) (2.8 ϫ 8 cm) equilibrated in the same buffer. Elution was performed with a 800-ml linear gradient from 0 to 600 mM NaCl. Recombinant proteins were dialyzed against 50 mM triethanolamine hydrochloride, 145 mM NaCl, pH 7.4, concentrated by ultrafiltration to 0.05-0.5 mg/ml, and stored at 0°C.
Polyacrylamide Gel Electrophoresis and Immunoblotting-SDS-PAGE analysis was performed as described previously (36). Western blot analysis and immunodetection of the recombinant proteins were carried out as described by Rossi et al. (17), or using the ECL detection procedure of Amersham Pharmacia Biotech. The antibodies used were the mouse monoclonal anti-MASP-2 antibody 1.3B7, a rabbit polyclonal anti-MASP-2 antibody (37), and a rabbit anti-peptide antibody directed against the serine protease domain of MASP-1 (6).
N-terminal Sequence Determination-N-terminal sequence analyses were performed after SDS-PAGE and electrotransfer, using an Applied Biosystems model 477A protein sequencer as described previously (38).
Proteolytic Assays-The proteolytic activity of MASP-1, MASP-2, and C1s toward C2, C3, C4, C5, and factor B was measured by incubation at different enzyme:protein ratios for varying periods at 37°C, as indicated in the text. The extent of reaction was determined after SDS-PAGE analysis under reducing conditions by gel scanning of the cleavage fragments, as described previously (17). For determination of the kinetic parameters, the concentrations of C2 ranged from 1.0 to 4 M, and those of C4 ranged from 50 to 500 nM (MASP-2 cleavage) or from 0.5 to 4 M (C1s cleavage). Fixed enzyme concentrations of 2 nM were used in all cases, and enzyme dilutions were performed in 50 mM triethanolamine HCl, 145 mM NaCl, pH 7.4, containing 1 mg/ml ovalbumin. The kinetic constants were determined by the Lineweaver-Burk method using linear regression analysis and are based on duplicate or triplicate measurements of initial rates at 6 separate substrate concentrations.
All kinetic analyses were conducted in 150 mM NaCl, 5 mM EDTA, 20 mM sodium phosphate, pH 7.4.
Reactivity Toward Inhibitors-Complex formation between C1 inhibitor and MASP-1 or MASP-2 was assessed by incubation of the proteases with excess molar ratios of C1 inhibitor for 15 min at 37°C in 50 mM triethanolamine-HCl, 145 mM NaCl, pH 7.4, followed by SDS-PAGE analysis under reducing conditions. C1 inhibitor-protease complexes were revealed by Western blot analysis using a rabbit polyclonal antibody directed against full-length MASP-2, and a rabbit anti-peptide antibody directed against the serine protease domain of MASP-1. Titration of the enzymatic activity of C1s, MASP-1, and MASP-2 by C1 inhibitor was performed by preincubating these proteases (at concentrations of 0.25 M, 20 nM, and 0.25 M, respectively) with increasing concentrations of C1 inhibitor (10 -500 nM) for 15 min at 37°C. The residual C2 cleaving activity of MASP-2 and C1s was then measured by incubating the enzymes (3.75 and 7.5 nM, respectively) for 15 min at 37°C in the presence of 2.25 M C2, followed by SDS-PAGE analysis, as described above. In the case of MASP-1, the residual esterolytic activity was measured on Bz-Arg-OEt, as described above.
The reactivity of C1s and MASP-2 toward ␣ 2 M was measured by preincubating the enzymes (0.25 M each) for 30 min at 37°C in the presence of increasing concentrations (3.7-300 M) of ␣ 2 M. Residual C4 cleaving activity was then measured by incubating the enzymes (1.25 nM each) for 15 min at 37°C in the presence of 2.25 M C4, followed by SDS-PAGE analysis.

Production and Characterization of Recombinant MASP-2 and C-terminal Fragments from MASP-1 and MASP-2-
The modular structures of MASP-1, MASP-2, and of the truncated fragments used in the present study are depicted in Fig. 1. The recombinant baculoviruses used for expression of each construct were obtained as described under "Experimental Procedures" and used to infect High Five insect cells for various periods at 28°C. The amount of recombinant material recovered in the culture medium, as estimated by SDS-PAGE and Western blot analysis, ranged from 0.15 (MASP-2) to 2 g/ml (CCP1/2-SP fragment of MASP-1). Protein purification was performed as described under "Experimental Procedures," using an initial ion-exchange fractionation step in all cases. Because of the low amounts of material recovered, protein detec-tion and analysis was performed routinely using Western blot analysis rather than Coomassie Blue staining.
Due to its very low recovery, recombinant MASP-2 could not be purified to homogeneity since attempts to completely remove contaminant proteins resulted in a nearly complete loss of material. SDS-PAGE analysis of the partially purified MASP-2 fraction under nonreducing conditions revealed two bands reactive with specific antibodies: (i) a major, 80-kDa species corresponding to the full-length protease, yielding two sequences, Thr-Pro-Leu-Gly-Pro-Lys-Trp-Pro-Glu-Pro . . . and Ile-Tyr-Gly-Gly-Gln-Lys-Ala-Lys-Pro-Gly . . . , corresponding to the N-terminal ends of the mature protein and of the serine protease domain, respectively; (ii) a 45-kDa species yielding only the first sequence above, and corresponding to a truncated fragment derived from the N-terminal end of the protein. Analysis under reducing conditions indicated that full-length MASP-2 was recovered in a partially activated form, since only 20 -30% of the protein migrated as a single chain, proenzyme species, whereas the remainder yielded two bands at 45 and 28 kDa, corresponding to the N-terminal A chain and the serine protease domain, respectively. Based on Coomassie Blue staining after SDS-PAGE analysis, the relative amount of fulllength MASP-2 in the partially purified fraction averaged 10% of the total protein contents.
Both the CCP1/2-SP and CCP2-SP fragments of MASP-2 could be purified to homogeneity. On SDS-PAGE analysis, the CCP1/2-SP fragment migrated under nonreducing conditions as a 44-kDa band, which upon reduction split into two bands corresponding to the serine protease domain (27 kDa) and the CCP1/2 segment (17 kDa) (Fig. 2, lanes 1 and 2), indicating complete activation of the protease. In contrast, the shorter CCP2-SP fragment migrated as a single band of about 35 kDa both under nonreducing and reducing conditions (Fig. 2, lanes  3 and 4), indicating that this species had retained a single chain, proenzyme structure. N-terminal sequence analysis of the latter fragment yielded a single sequence Asp-Pro-Asp-Cys-Gly-Pro-Pro . . . corresponding to the expected N-terminal end of the fragment, whereas the activated CCP1/2 SP fragment yielded equivalent amounts of two sequences: Asp-Pro-Gln-Pro-Cys-Pro-Tyr . . . (N-terminal end) and Ile-Tyr-Gly-Gly-Gln-Lys-Ala . . . (serine protease domain). Protein staining with Coomassie Blue (not shown) yielded the same pictures as observed by Western blot analysis, indicating that no major fragment or contaminant was present in the purified preparations. As in the case of full-length MASP-2, the CCP1/2-SP fragment of MASP-1 could not be purified to homogeneity, mainly because of the low amounts of recombinant material available. SDS-PAGE analysis indicated that the fragment migrated under nonreducing conditions as a single band of about 50 kDa. Upon reduction, part of the material (40%) still migrated as a single chain, proenzyme species, whereas the remainder yielded a doublet at 32 and 34 kDa, reactive with antibodies directed to the C-terminal end of the molecule, and corresponding to the serine protease domain. Sequence analysis confirmed these findings, as the proenzyme species yielded a single sequence Asp-Leu-Val-Glu-Leu-Pro-Glu-Leu-Gln-Pro . . . corresponding to the N-terminal end of the fragment, the doublet at 32-34 kDa yielding the expected sequence Ile-Phe-Asn-Gly-Arg-Pro-Ala-Gln-Lys-Gly . . . characteristic of the serine protease domain. Thus, as previously observed in the case of fulllength MASP-1 expressed in the same system, the recombinant CCP1/2-SP fragment was recovered in a partially activated form. As judged from Coomassie Blue staining, the MASP-1 CCP1/2-SP fragment was estimated at 20 -30% of the total protein contents of the partially purified fraction, depending on the preparation.
Proteolytic Activity of Recombinant MASP-2-Recombinant MASP-2 and its CCP1/2-SP fragment both cleaved C4 very efficiently, as shown by their ability to convert the C4␣ chain into the smaller C4-␣Ј species (Fig. 3A). In both cases, almost complete cleavage of C4 was achieved upon incubation for about 15 min at 37°C at an enzyme:protein molar ratio of 1:2000. Both recombinant proteases also readily and specifically cleaved C2, as shown by their ability to split the protein into its characteristic C2a and C2b fragments (Fig. 3B). Again, C2 cleavage was essentially complete after 20 min at 37°C at an enzyme:protein molar ratio of 1:1000. Under these conditions, no significant C4 or C2 cleavage was observed when these proteins were incubated in the presence of the proenzyme CCP2-SP fragment of MASP-2. The ability of MASP-2 to cleave C3 was also tested using the same methodology. As illustrated in Fig. 3C, incubation of C3 with increasing amounts of either full-length MASP-2 or its catalytic fragment CCP1/2-SP led to an increased production of the C3␣Ј fragment. However, the cleavage reaction was very inefficient, as complete conversion of C3␣ into C3␣Ј required overnight incubation at 37°C at an enzyme:protein molar ratio of 1:6 (Fig. 3C, lane 5). For comparison, complete cleavage of C3␣ by trypsin was achieved at an enzyme:protein molar ratio of 1:13 in only 5 min at 37°C (Fig. 3C, lane 1). Using partially purified preparations of C5 and factor B as substrates, no significant cleavage of these complement proteins by either MASP-2 or its CCP1/2-SP fragment could be detected after overnight incubation at 37°C at enzyme:protein molar ratios up to 1:10 (not shown).
The kinetic parameters for C2 and C4 cleavage by MASP-2 were determined using the full-length protease and its CCP1/ 2-SP fragment, and the values were compared with those obtained with active C1s purified from human serum. In the case of C2, all three enzymes showed comparable k cat values, whereas both MASP-2 species exhibited K m values slightly, but consistently lower than C1s (Table I and Fig. 4). As a result, the C2 cleavage efficiency of both MASP-2 species, as measured by the k cat /K m ratio, was slightly higher than that of C1s. With respect to C4 cleavage, again, all three enzymes exhibited similar k cat values. In contrast, both MASP-2 and its CCP1/ 2-SP fragment showed K m values in the nanomolar range, i.e. 26 -32 times lower than the value of 1.92 Ϯ 0.5 M determined for C1s. As a result, the k cat /K m ratios for MASP-2 and its CCP1/2-SP fragment were 12-and 23-fold higher, respectively, than that for C1s, indicating a much higher efficiency in the case of C4 cleavage. In this respect, it should be mentioned that the k cat values for MASP-2 were significantly lower than those determined for the CCP1/2-SP fragment. This is likely because of the fact that the concentration of MASP-2 could only be roughly estimated (see "Experimental Procedures") and, we believe, was probably overestimated rather than the opposite. As a consequence, the k cat values determined for the CCP1/ 2-SP fragment should be regarded as the most representative. It may be estimated therefore that MASP-2 exhibits C2 and C4 cleaving efficiencies about 3 and 23 times higher, respectively, than C1s (see Table I). Proteolytic Activity of Recombinant MASP-1-The proteolytic activity of recombinant MASP-1 was tested as described above for MASP-2, using both the full-length protease and its C-terminal CCP1/2-SP fragment. For either species, no detectable activity was observed on human C4, C5, and factor B, even after overnight incubation at 37°C using enzyme:substrate molar ratios up to 1:10 (not shown). Using C3 as a substrate, a low but significant activity was consistently detected using either full-length MASP-1 or the CCP1/2-SP fragment. However, partial C3 cleavage (about 13%) required overnight incubation at 37°C at an enzyme:substrate molar ratio as high as 1:10 (Fig. 5A, lane 3). This faint activity was nevertheless clearly attributable to MASP-1, since it was blocked by pretreatment with C1 inhibitor, and by DFP (Fig. 5A, lanes 4 and  5). In a control experiment, complete C3 cleavage was achieved upon incubation for 10 min at 37°C with trypsin using a 1:10 enzyme:substrate molar ratio (Fig. 5A, lane 6). As shown in Fig.  5B, C2 was also found to be cleaved by recombinant MASP-1. Again, however, this reaction was quite inefficient since significant cleavage (about 28%) required overnight incubation at 37°C in the presence of either full-length MASP-1 or its CCP1/ 2-SP fragment at a 1:10 enzyme:substrate ratio (Fig. 5B, lane  3). Pretreatment of MASP-1 with C1 inhibitor or DFP reduced the extent of cleavage to about 5% (Fig. 5B, lanes 4 and 5), i.e. a value comparable with that observed when C2 alone was incubated overnight at 37°C (Fig. 5B, lane 2). The observed C2 cleavage was therefore clearly mediated by MASP-1, but was quite inefficient compared with the C2 cleaving activity of MASP-2 or C1s.
Esterolytic Activities of MASP-1 and MASP-2-The esterolytic activity of recombinant MASP-1 and MASP-2 was measured on various synthetic substrates and compared with that of human C1r and C1s. For both MASP-1 and MASP-2, comparable results were obtained using either the full-length proteases or their catalytic CCP1/2-SP fragment. Full-length MASP-1 and the MASP-2 CCP1/2-SP fragment were routinely used for determination of the kinetic constants shown in Table  II. MASP-2 and C1s efficiently cleaved the four substrates tested in this study, and both enzymes exhibited similar k cat and K m values. However, C1s showed a significantly better efficiency on Ac-Gly-Lys-OMe, Bz-Arg-Oet, and Tos-Arg-OMe, whereas MASP-2 was more efficient on the thioester Z-Gly-Arg-S-Bzl, because of a remarkably low K m value (Table II). In keeping with previous reports (39), C1r hydrolyzed Ac-Gly-Lys-OMe and Z-Gly-Arg-S-Bzl, but had no activity on Bz-Arg-OEt and Tos-Arg-OMe. Likewise, MASP-1 also exhibited a restricted esterolytic activity, as it was not reactive toward Ac-Gly-Lys-OMe and Tos-Arg-OMe, and cleaved Z-Gly-Arg-S-Bzl and Bz-Arg-OEt to significant extents. Interestingly, of the four proteases used, MASP-1 showed the least K m value toward Bz-Arg-OEt (Table  II). All of the esterolytic activities measured with recombinant MASP-1 and MASP-2 were blocked after pretreatment of the proteases with excess C1 inhibitor.
Reactivity of MASP-1 and MASP-2 Toward Inhibitors-After incubation with a molar excess of C1 inhibitor, full-length MASP-1 and MASP-2, as well as their CCP1/2-SP fragments, all specifically reacted with this inhibitor, as shown by the formation of stable complexes migrating on SDS-PAGE analysis under reducing conditions as species of over 100 kDa, corresponding to C1 inhibitor-serine protease domain complexes (not shown). To determine the stoichiometry of the reaction between MASP-2 and C1 inhibitor, the CCP1/2-SP fragment of MASP-2 was preincubated with increasing molar ratios of C1 inhibitor and then the residual protease activity was measured using the C2 cleavage assay. As shown in Fig. 6B, increasing the C1 inhibitor:MASP-2 ratio led to a linear decrease of enzymatic activity, with complete inhibition at a molar ratio of about 1:1. In a comparative experiment, the same stoichiome-  try was determined in the case of C1s, in agreement with previous reports (32). The stoichiometry of the reaction between MASP-1 and C1 inhibitor was determined in a similar way, by measuring the esterolytic activity of the protease on Bz-Arg-OEt. As shown in Fig. 6C, preincubation with increasing concentrations of C1 inhibitor led to a linear decrease of MASP-1 esterolytic activity, with again complete inhibition at a C1 inhibitor:MASP-1 molar ratio of about 1:1. Thus, as determined earlier in the case of C1r and C1s (32,41), both MASP-1 and MASP-2 reacted with C1 inhibitor in a 1:1 stoichiometry.
As mentioned above, the proteolytic activity of MASP-1 toward C3 and C2 was prevented in the presence of 5 mM DFP. Because of the very low activity of MASP-1 on these substrates, it was not possible to further analyze its sensitivity with respect to DFP. In the case of MASP-2, preincubation of the protease with 2 mM DFP for 30 min at 37°C resulted in 86% inhibition of its C4 cleaving ability, and nearly complete inhibition was obtained when this treatment was performed twice in a row. Thus, MASP-2 exhibited a DFP sensitivity comparable with that previously determined in the case of C1r and C1s (42). We also checked the effect of ␣ 2 M on the C4 cleaving activity of MASP-2 and C1s. Both proteases were used at a concentration of 0.5 M in the presence of increasing molar ratios of ␣ 2 M up to 120:1. Under these conditions, this protein exerted no detectable inhibitory effect on the proteolytic activ-

DISCUSSION
Full-length MASP-2, as well as fragments from the C-terminal catalytic regions of MASP-1 and MASP-2 were expressed using a baculovirus/insect cell system. As observed previously in the case of full-length MASP-1 expressed in the same system (27), the recombinant material was produced at very low yields (0.15-2.0 g/ml). This rendered isolation of the recombinant material difficult, but nevertheless allowed us to achieve complete purification of MASP-2 fragments CCP1/2-SP and CCP2-SP, and partial purification of the MASP-1 CCP1/2-SP fragment and of full-length MASP-2. As judged from all functional assays performed in this study, the latter two preparations were functionally pure, i.e. devoid of detectable contaminant esterolytic or proteolytic activities. Thus, full-length MASP-1 on the one hand, and full-length MASP-2 on the other hand, exhibited esterolytic and proteolytic activities comparable with those of their CCP1/2-SP fragments, both in terms of efficiency and specificity. In addition, these activities were all blocked by C1 inhibitor, providing further evidence that they were not due to contaminant proteases arising from the insect cells. The recombinant material used in this study was therefore appropriate for a precise assessment of the enzymic properties and specificity of MASP-1 and MASP-2.
As previously observed for the full-length recombinant protease (27), the CCP1/2-SP fragment of MASP-1 was recovered in a partially (about 60%) activated form, suggesting that either autolytic activation or extrinsic proteolytic cleavage occurred during the synthesis process. Similarly, in the case of MASP-2, the full-length species was also recovered in a partially activated form. Interestingly, the short catalytic fragment CCP2-SP of MASP-2 was recovered in a fully proenzyme state, whereas the larger fragment CCP1/2-SP was secreted by the insect cells as a partially activated form, and underwent complete activation during the purification process. These latter findings appear not consistent with an activation process mediated by an extrinsic protease, as access to the Arg 444 -Ile 445 activation site of MASP-2 is not expected to be made easier in the larger CCP1/2-SP fragment than in the shorter CCP2-SP fragment. On the other hand, this strikingly different behavior of the two fragments would be consistent with an autolytic activation mechanism of MASP-2 involving interaction between two molecules through their C-terminal catalytic region, if this process requires the first CCP module. This latter hypothesis appears plausible, in light of the observation that dimerization of the catalytic domains of the homologous protease C1r involves its first CCP module (44). In the same way, the fact that recombinant full-length MASP-2 only undergoes partial activation may be explained by analogy with previous findings on C1r (48,49) indicating that the N-terminal, Ca 2ϩbinding CUB-EGF-CUB region of C1r exerts a negative control on the activation of the C-terminal catalytic region of the protease. The above observations support the hypothesis of an intrinsic ability of MASP-2 to self-activate, in full agreement with previous reports by Vorup-Jensen et al. (45) indicating that reconstitution of MBL with MASP-2 alone is sufficient to trigger complement activation.
Although recombinant MASP-2 and its CCP1/2-SP fragment displayed a detectable C3 cleaving activity, this was only observed upon overnight incubation at elevated enzyme:substrate ratios. It is likely therefore that this faint activity is nonspecific and has no biological relevance. In contrast, in agreement with previous reports (6,21,45), this work confirms that MASP-2 specifically cleaves and activates C4 and C2 and provides the first detailed analysis of the kinetic parameters of these reactions. These data indicate that (i) MASP-2 cleaves C2 slightly more efficiently (2-3 times) than does C1s; (ii) MASP-2 exhibits a C4-cleaving efficiency that is 20 -25 times higher than that of C1s, arising mainly from a much lower K m value (about 60 nM versus about 2 M). These findings have direct functional implications, since C4 cleavage, the initial event in the formation of the classical pathway C3-convertase C4b,2a, occurs in the fluid phase and is the limiting step of this process. Thus, the relative efficiency of C4 cleavage by MASP-2 and C1s in vivo will directly depend on their respective K m values for C4 (about 60 nM and 2 M) relative to the C4 concentration in serum (about 3 M) (46). It can be deduced from these figures that, whereas C1s will cleave C4 in serum at a rate well below the V m value, MASP-2 will always function at or close to maximal velocity. C2 cleavage, the second step of the assembly of the C4b,2a-convertase, is postulated to take place after prior binding of C2 to membrane-bound C4b, in close vicinity of either C1 or the MBL-MASP complex. Therefore the C2 cleavage efficiency of MASP-2 or C1s is not directly related to their K m values for this substrate, which in both cases are well above the C2 concentration in serum, i.e. 0.13-0.43 M (47).
According to the above figures, MASP-2 is expected to be about 20 -25 times more efficient than C1s with respect to C4 cleavage. As a consequence, despite the low amounts of MASP-2 in serum, a complex utilizing MBL as a recognition unit and MASP-2 as a catalytic unit is probably as active as C1 in terms of complement activation. These considerations provide strong support to the biological relevance of the lectin pathway of complement activation as a major innate defense mechanism. Together with experiments suggesting that MASP-2 has the ability to self-activate (Ref. 45 and this study), and binds individually to MBL with high affinity (27), the above results strongly suggest that the entity responsible for triggering the MBL-mediated lectin pathway of complement is an MBL⅐MASP-2 complex, in which MASP-2 combines the ability to self-activate as well as to cleave C4 and C2, and hence fulfills the roles of both C1r and C1s in C1. In keeping with previous data (21), our data also demonstrate that the proteolytic activity of MASP-2 is specifically titrated by C1-inhibitor, with a 1:1 C1 inhibitor:MASP-2 stoichiometry. As well documented in the case of the C1 enzymes (41), it may be anticipated therefore that the proteolytic activity of MASP-2 in serum is tightly and specifically controlled by C1 inhibitor.
This work also provides the first analysis of the esterolytic activities and specificities of MASP-1 and MASP-2. In keeping with the fact that MASP-2 and C1s specifically cleave the same protein substrates, they exhibit similar esterolytic activities on the various synthetic substrates tested in this study, which they cleave with comparable k cat and K m values. However, as measured by the k cat /K m ratio, the efficiency of MASP-2 was found consistently lower than that of C1s, with the exception of the thioester Z-Gly-Arg-S-Bzl, which, because of a remarkably low K m value, is the best known substrate of MASP-2. Thus, the shared ability of MASP-2 and C1s to cleave C4 and C2 likely arises in part from common structural features at or in the vicinity of their active sites. On the other hand, the fact that MASP-2 has a much lower K m value for C4 is probably related to structural determinants located outside the active site area, very likely in the CCP modules, as previously demonstrated in the case of C1s (17). MASP-1, in contrast, exhibits a lower esterolytic activity, that is restricted to two of the substrates tested in this study, Bz-Arg-OEt and Z-Gly-Arg-S-Bzl. Nevertheless, it is noteworthy that MASP-1 cleaves Bz-Arg-OEt with an efficiency comparable with that of C1s, because of a rather low K m value (see Table II). Also, the fact that MASP-1 does not hydrolyze Ac-Gly-Lys-OMe suggests that its activity may be restricted to arginyl bonds only. As judged from its rather low and narrow esterolytic activity, MASP-1 appears functionally much closer to C1r than to MASP-2 and C1s, which display broader activities and are more efficient enzymes.
In keeping with data obtained by Matsushita et al. (21) using material purified from human serum, both recombinant fulllength MASP-1 and its catalytic CCP1/2-SP fragment showed detectable C2 and C3 cleaving activities. However, these reactions were quite inefficient, since in both cases only partial cleavage could be obtained after overnight incubation at elevated enzyme:substrate ratios. In this respect, it should be stressed that the C2 cleaving activity of recombinant MASP-1 is quite negligible compared with that of recombinant MASP-2 (ϳ20,000-fold less). In the same way, the C3 cleaving activity of MASP-1 was consistently found to be less than that of MASP-2, which itself is extremely low compared the MASP-2 activities on C4 and C2. With respect to C2 cleavage by MASP-1, it is difficult to compare quantitatively our data with those of Matsushita et al. (21), since these authors used a hemolytic assay to test C2 activation. As for C3 cleavage, as judged from the data presented by Matsushita et al. (21), the activity exhibited by the MASP-1 fraction purified from serum appears comparable with that obtained with our recombinant material, since only partial cleavage was obtained under the conditions used. In this respect, it should also be mentioned that other studies have reported no detectable C3 cleavage by the total MASP fraction (22,24). In summary, although our data certainly validate previous findings by Matsushita et al. (21) that MASP-1 exhibits detectable C2 and C3 cleaving activities, they also clearly show that these activities are quite negligible compared with the very efficient C4 and C2 cleaving activities of MASP-2. It appears very unlikely, therefore, that the physiological function of MASP-1 is to cleave C3 (which would be a very inefficient means of triggering complement, compared with C1s or MASP-2), and/or to cleave C2 (which, alone, would be physiologically not relevant). If this hypothesis is correct, then the physiological substrate(s) of MASP-1 in serum, and hence its physiological role, remain to be established.