C1 Inhibitor: Analysis of the Role of Amino Acid Residues Within the Reactive Center Loop in Target Protease Recognition1 Rana Zahedi,2* Ryan C. MacFarlane,* Jeffrey J. Wisnieski,† and Alvin E. Davis III3* Previous analysis of a naturally occurring C1 inhibitor P2 mutant (Ala4433 Val) indicated a role for P2 in specificity determination. To define this role and that of other reactive center loop residues, a number of different amino acids were introduced at P2, as well as at P6 (Ala439) and P8ⴕ/9ⴕ (Gln452Gln453). Ala4393 Val is a naturally occurring mutant observed in a patient with hereditary angioedema. Previous data suggested that Gln452Gln453 might be a contact site for C1s. Reactivity of the inhibitors toward target (C1s, C1r, kallikrein,  factor XIIa, and plasmin) and nontarget proteases (␣-thrombin and trypsin) were studied. Substitution of P2 with bulky or charged residues resulted in decreased reactivity with all target proteases. Substitution with residues with hydrophobic or polar side chains resulted in decreased reactivity with some proteases, but in unaltered or increased reactivity with others. Second order rate constants for the reaction with C1s were determined for the mutants with activities most similar to the wild-type protein. The three P2 mutants showed reductions in rate from 3.35 ⴛ 105 Mⴚ1sⴚ1 for the wild type to 1.61, 1.29, and 0.63 ⴛ 105 for the Ser, Thr, and Val mutants, respectively. In contrast, the Ala4393 Val and the Gln452Gln4533 Ala mutants showed little difference in association rates with C1s, in comparison with the wild-type inhibitor. The data confirm the importance of P2 in specificity determination. However, the P6 position appears to be of little, if any, importance. Furthermore, it appears unlikely that Gln452Gln453 comprise a portion of a protease contact site within the inhibitor. The Journal of Immunology, 2001, 167: 1500 –1506.
T
he C1 inhibitor (C1INH)4 is the only inhibitor of the complement proteases C1s and C1r, and is the primary inhibitor of the proteases of the contact phase of coagulation: plasma kallikrein, factor XIa, and factor XIIa (1– 6). C1INH belongs to the superfamily of serine protease inhibitors, or serpins, many members of which are involved in the regulation of proteolytic cascades, including those participating in inflammation, coagulation, and fibrinolysis. Serpins share homologous structures and a common mode of action. They are composed of 400 – 450 aa residues organized into three  sheets and eight or nine ␣ helices (7). Serpins act as suicide substrates in which the amino acid sequence of the reactive center loop mimics that of an ideal natural substrate of target proteases (8). Reaction with target protease results in formation of an initial Michaelis complex. After this initial interaction, an intermediate complex forms that either develops into a stable complex or dissociates to release the inactive cleaved serpin and the active enzyme (8). The potency of an inhibitor de-
*Center for Blood Research, Boston, MA 02115; and †Department of Veterans Affairs Medical Center, Department of Medicine, Case Western Reserve University, Cleveland, OH 44106 Received for publication March 21, 2001. Accepted for publication May 30, 2001. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 1 This work was supported by National Institutes of Health Grants HD22082 (to A.E.D.) and HD33727 (to A.E.D.). 2 Current address: Lawrence Berkeley National Laboratories, 1 Cyclotron Road, Mail Stop 83-109, Berkeley, CA 94720. 3 Address correspondence and reprint requests to Dr. Alvin E. Davis III, Center for Blood Research, 800 Huntington Avenue, Boston, MA 02115. E-mail address:
[email protected] 4 Abbreviations used in this paper: C1INH, C1 inhibitor; HAE, hereditary angioedema.
Copyright © 2001 by The American Association of Immunologists
pends on the relative rates of these two reactions. The rate of dissociation of the stable complex is extremely slow (9). The primary recognition element for serpin-protease association is a 15 aa residue exposed segment known as the reactive center loop. This loop varies in sequence and length; it is mobile and can adopt different conformations. Cleavage of the reactive center loop by nontarget proteases results in complete insertion of the loop into the five-stranded  sheet A, thereby converting it to a six-stranded  sheet (10). In the complex with protease, the reactive center loop is similarly inserted; the protease is thereby displaced from the site of initial interaction at one pole of the inhibitor molecule to the other pole (11–14). The amino acid sequence of the reactive center loop from approximately P5 through P5⬘ varies among the serpins and, to a great extent, determines the specificity of the inhibitor (15–21). The P1 residue is the major determinant of serpin specificity, as has been illustrated by analysis of both naturally occurring serpins and serpins with new unique engineered mutations (15, 20). Because a serpin with a single P1 residue may inhibit a variety of target proteases and because different serpins with the same P1 residue inhibit different proteases, other residues must be involved in specificity determination. Analyses of mutant serpins, which indicate that the P1 residue is not the sole determinant of specificity, support this assumption (16 –20, 22–24). Other residues within the reactive center region also play a significant role. However, the roles of different residues appear to vary with different serpin-protease combinations. We described a variant C1INH in which the P2 residue (Ala443) was substituted with Val (25, 26). This mutant had a decreased rate of reactivity toward C1s and C1r, and had acquired weak inhibitory activity toward trypsin, but retained normal activity against the target proteases factor XIIa and plasma kallikrein. A patient with hereditary angioedema (HAE) has been described in which the P6 residue (Ala439) was substituted with a Val (27). It was not clear whether this resulted in absolute deficiency of C1INH or in 0022-1767/01/$02.00
The Journal of Immunology a dysfunctional inhibitor. An identical mutation at the P6 position of antithrombin, also without functional analysis, previously had been reported (28). Analysis of autoantibodies to C1INH from patients with acquired angioedema led to the identification of the P8⬘/P9⬘ (Gln452Gln453) region as a potential contact site with C1s during complex formation. To further characterize the roles of these reactive center region residues in complex formation and in specificity determination, we have created mutant proteins with substitutions at each of these positions. In the case of P2, a variety of substitutions varying in side chain size, charge, and hydrophobicity have been constructed. We have analyzed the reactivity of these mutant proteins with target proteases, and have determined their sensitivity to heat denaturation as a measure of conformation. In addition, we have analyzed the reactivity of the mutant inhibitors with plasmin and thrombin.
Materials and Methods Site-directed mutagenesis and expression in COS-1 cells Mutations at the C1INH P2 position were introduced using the QuickChange site-directed mutagenesis kit (Stratagene, La Jolla, CA). Primers corresponding to nucleotides 16,773 through 16,802 of the C1INH genomic sequence that varied in the nucleotides at positions 16,785 through 16,787 were designed. These substitutions led to replacement of Ala443 with Asn, Gly, Leu, and Thr. The same approach was used to replace the P6 Ala residue with Val and both the P8⬘ and P9⬘ residues (Gln452 and Gln453) with Ala. Mutated and wild-type proteins then were expressed transiently in COS-1 cells. For some experiments, newly synthesized proteins were metabolically radiolabeled with [35S]Met (Amersham Life Sciences, Arlington Heights, IL). Medium was harvested 72 h after transfection and dialyzed into PBS.
Thermal stability COS-1 cell culture supernatants (100 l) containing recombinant inhibitors were incubated with and without trypsin (5 g) at 40, 50, 60, 70, and 80°C for 120 min, following which they were subjected to centrifugation at 14,000 ⫻ g for 30 min. Quantitation of residual Ag in the supernatant was performed using an ELISA as described by Aulak et al. (29).
Complex formation between C1INH proteins and proteases Radiolabeled recombinant wild-type or mutant C1INHs were incubated for 1 h at 37°C with C1s (8.4 g), C1r (6.4 g) (Advanced Research Technologies, San Diego, CA), kallikrein (15 g),  factor XIIa (3.5 g) (Enzyme Research Laboratories, South Bend, IN), plasmin (0.2, 0.5, 1, 2, 5, and 10 g/l) and ␣-thrombin (0.2, 0.5, 1, 2, 5, and 10 g/l) (plasmin and thrombin provided by Dr. V. Donaldson, Children’s Hospital Research Foundation, Cincinnati, OH), or trypsin (0.5 g; Sigma, St. Louis, MO). Each protease was incubated in molar excess relative to recombinant C1INH. After incubation, PMSF (final concentration, 1 mM) was added, and Triton X-100 (0.5%), deoxycholic acid (0.25%), SDS (0.5%), and EDTA (5 mM) were added to each sample. The IgG fraction of rabbit anti-human C1INH (3 l) then was added, and samples were incubated at 4°C overnight. A suspension of fixed Staphylococcus aureus (6 l, IgG sorb; The Enzyme Center, Boston, MA), sonicated and washed three times (with 1% Triton X-100, 1% SDS, 0.5% deoxycholate, and 5 mg/ml BSA in 0.02 M sodium phosphate and 0.15 M NaCl, pH 7.4), was added to each sample and incubated at 4°C for 1 h. The pellets (14,000 ⫻ g for 3 min) were washed once with the above washing solution and three times with a washing solution lacking BSA. Each precipitate then was dissolved in SDS sample buffer without reducing agent (15 l; 4.2% SDS, 20% glycerol, 0.1 M Tris-HCl, pH 6.5, and 0.01% bromophenol blue), vortexed, boiled for 3 min, and subjected to brief centrifugation and electrophoresis on SDS7.5% polyacrylamide gel. Gels were fixed, dried, and exposed to x-ray film (Kodak XAR-5; Eastman Kodak, Rochester, NY) at ⫺70°C.
1501 in each C1INH preparation was determined by incubation of varying quantities of mutant inhibitor with a fixed known quantity of fully active C1s for a prolonged period, following which residual C1s esterolytic activity was determined by hydrolysis of the chromogenic substrate S-2314 (Chromogenix, Molndal, Sweden). The active concentration of C1s was determined by incubation of varying molar ratios of C1s and plasma-derived C1INH of known concentration (determined using an extinction coefficient of 3.7) (30) followed by determination of the amount of complex formation by SDS-PAGE. For the kinetic assays, equimolar concentrations of C1s and recombinant wild-type or mutant C1INH were incubated at 37°C in PBS. At different times (0 –30 min), aliquots of the reaction mix were removed and the reaction stopped by dilution in the chromogenic substrate S2314 (0.5 mM in PBS). Residual enzyme activity was quantitated by monitoring hydrolysis of the substrate (increase in absorption over time at a wavelength of 405 nm). The second order rate constants were determined using the halftime of the reaction (t1/2), which is calculated by plotting the inverse of the enzymatic rate vs time; this is related to the (Ka) through the equation t1/2 ⫽ (Ka ⫻ E 0)⫺1, where E 0 is the total enzyme concentration.
Results Thermal denaturation profiles of the P2 mutant C1INHs Thermal stability provides a simple, but reliable measure of the conformation of serpins (31). Normal intact serpins multimerize at temperatures of 50 – 60°C. This results in a decrease in epitopes detected with polyclonal anti-C1INH antisera. Therefore, with multimerization the apparent concentration of C1INH decreases. Serpins that have been cleaved near the reactive center are more stable and remain monomeric at temperatures as high as 80 –90°C. Thermal denaturation curves of all the P2, the P6 Ala3 Val, and the P8⬘/P9⬘ Gln3 Ala mutant C1INHs were indistinguishable from that of the wild-type protein (Fig. 1). After cleavage by trypsin, all of the recombinant proteins were stable at elevated temperatures (50 –90°C). These data suggest that the mutations did not induce any major conformational changes, and that each was able to undergo the conformational rearrangement with reactive center loop insertion that takes place in normal inhibitor serpins following reactive center cleavage or complex formation (10, 11). Therefore, it is likely that each of the mutants would be capable of complex formation with an appropriate protease.
Inhibitor assay and calculation of second order rate constants COS-1 cell supernatants (40 ml) containing recombinant C1INH proteins were collected and concentrated to 2–5 ml. The C1INH concentrations of the samples were measured by ELISA (29). The concentrations of the recombinant mutant C1INH proteins after concentration were: wild type (5.5 g/ml), Ala4433 Val mutant (5 g/ml), Ala4433 Ser mutant (6.5 g/ ml), Ala4433 Thr mutant (5.5 g/ml), Ala4393 Val mutant (2.5 g/ml), and Gln452Gln4533 Ala mutant (1.6 g/ml). The amount of active inhibitor
FIGURE 1. Thermal denaturation profiles of recombinant C1INH proteins. Supernatant from transfected COS-1 cells were incubated in the presence or absence of trypsin (5 g) at different temperatures for 2 h. Samples then were subjected to centrifugation for 30 min, and residual Ag was quantitated by ELISA.
1502 Reactivity of the P2 mutant proteins with target proteases C1s. The reactions of each of the recombinant inhibitors with the different proteases in excess were assessed by SDS-PAGE following incubation for 1 h at 37°C. Recombinant wild-type C1INH formed a stable complex with C1s, accompanied by the appearance of a relatively small amount of cleaved C1INH, similar to previous observations (Fig. 2A) (25, 26). Compared with the wild type, recombinant C1INH with an Asn at the P2 position showed less complex formation with C1s in the 60-min incubation period (Fig. 1A). An appreciable quantity of the native form of the inhibitor was present even after 1 h of incubation. Furthermore, no cleavage of the Ala4433Asn mutant was observed. It should be pointed out that even with the same inhibitor-protease pair, in experiments performed at different times, the amount of cleaved vs complexed inhibitor, as assessed by SDS-PAGE, may vary somewhat. This likely is a result of subtle differences in experimental conditions. Therefore, over-interpretation of moderate changes in the amounts of cleaved inhibitor should be avoided. However, as seen in Fig. 2A, the complete absence of cleavage of this mutant clearly differs from the amount observed with the recombinant wild-type protein. The results with the Ala4433Asn mutant are similar to those observed previously with the P2 Asp mutant (26). The results described here, together with previously published data (26), are summarized in Table I. Substitution of P2 with Gly, Leu, or Thr did not significantly affect the amount of complex formation with C1s or the amount of resulting cleaved inhibitor, in comparison with the wild-type inhibitor (Fig. 2, A and B). The reactivities of these mutants were similar to those of the wild type and the pre-
C1INH TARGET PROTEASE RECOGNITION Table I. Complex formation between recombinant C1INH mutant proteins and proteases following incubation at 37°C for 60 min as assessed by SDS-PAGEa
Ala4433Val Ala4433Asn Ala4433Asp Ala4433Gly Ala4433Leu Ala4433Ser Ala4433Thr Ala4393Val Gln452,4533Ala
C1s
C1r
Kall
FXIIa
Try
Pl
Th
7 2 2 7 7 7 7 7 7
2 2 2 ?2 ?2 7 2 7 7
7 2 2 7 7 7 7 7 7
7 2 0 7 7 7 7 7 7
1 7 7 7 7 7 7 ND ND
7* 0 ND 0 7* ND 2 7 2
1* 0 ND 0 1* ND 0 0* 0
a Kall, Plasma kallikrein; FXIIa, factor XIIa; Try, trypsin; Pl, plasmin; Th, thrombin; 7, complex formation equivalent to that of the WT; 1, complex formation greater than, or more rapid than, with WT; 2, complex formation less than, or slower than, with WT; 0, no complex formation; ⴱ, an increase in reactive center cleaved inhibitor. The reactivities of Ala4433 Val with all of the proteases except plasmin and thrombin, and the reactivities of Ala4433 Asp and Ser with C1r, C1s, kallikrein and factor XIIa from (Ref. 26).
viously described P2 Ser substitution mutant (26). Neither the P6 Ala3Val nor the P8⬘/P9⬘ Gln3Ala substitutions showed any significant alteration in complex formation with C1s in comparison with the wild-type protein. Second order association rate constants were determined, as described in Materials and Methods, for some of those mutants that appeared similar to the wild type in the above experiments (Ala4433Val, Ala4433Ser, Ala4433Thr, Ala4393Val, and Gln452Gln4533AlaAla) (Table II). The rate for the wild-type protein is at the high end of the range of values reported in the literature for normal plasma-derived or recombinant C1INH (19, 32– 34). Previous data, based on time course analysis of complex formation, as observed on SDS-PAGE in which recombinant C1INH was incubated with a vast molar excess of C1s, indicated that the Ala4433Val mutant was much less active than the wildtype inhibitor (25, 26). The wild-type inhibitor was fully complexed within 0.5 min, whereas the mutant required ⬃5 min before complex formation was complete. As expected from these data, the association rate for the Ala4433Val mutant with C1s was substantially lower than that of the wild-type protein ( p ⫽ 0.04). In addition, the association rate constants of the Ser and Thr mutants were reduced in comparison with the wild-type inhibitor, although to a lesser extent than that of the Ala4433Val mutant (50–60% reduction in rate as opposed to an 80% reduction for Ala4433Val). C1r. Recombinant wild-type C1INH reacts more slowly with C1r than with C1s (25, 33, 34); however, in a 1-h incubation period this difference is not apparent (Fig. 3, A and B) (25). As with C1s, the reaction of the P2 Asn mutant with C1r was very slow, as indicated by the presence of residual native C1INH after a 1-h incubation at 37°C. The P2 Thr mutant did not show any alteration in reactivity with C1s following a 1-h incubation, even though the association rate constant for this reaction was slower than the reaction of the
Table II. Association rate constants for the interaction of C1INH mutants with C1s C1INH
FIGURE 2. Complex formation between recombinant inhibitors and C1s. COS-1 cell supernatants (100 l) containing the different recombinant inhibitors intrinsically labeled with [35S]Met were incubated with C1s for 1 h at 37°C. Samples then were immunoprecipitated with polyclonal antiserum to human C1INH and subjected to SDS-PAGE. Autoradiography was performed as described in Materials and Methods.
Wild type Ala4433Ser Ala4433Thr Ala4433Val Ala4393Val Gln452,4533Ala
Ka (⫻105 M⫺1s⫺1)
3.35 ⫾ 0.73 1.61 ⫾ 0.44 1.29 ⫾ 0.32 0.63 ⫾ 0.40 2.14 ⫾ 0.98 2.99 ⫾ 1.07
The Journal of Immunology
FIGURE 3. Complex formation between recombinant inhibitors and C1r. All conditions and analyses were as described for Fig. 1.
wild-type inhibitor with C1s. The slow reaction was more apparent with C1r; less complex was formed and small amounts of uncomplexed inhibitor remained after a 1-h incubation (Fig. 3B). The P2 Gly and Leu mutants also showed very small amounts of native uncomplexed inhibitor at the end of the incubation period (albeit less than with the Ala4433Thr mutant), which may suggest a small effect on the rate of complex formation with C1r (Fig. 3, A and B). As with C1s, the P6 and P8⬘/P9⬘ mutants each complexed with C1r to the same extent as did the wild-type inhibitor (Fig. 3C). Kallikrein and factor XIIa. In the reaction with  factor XIIa, recombinant wild-type C1INH showed nearly complete complex formation with minimal amounts of cleaved inhibitor or of remaining uncleaved inhibitor. However, a significant quantity of cleaved inhibitor is released during the reaction with plasma kallikrein (Fig. 4, A–C). This is in agreement with previous observations (8, 26, 35, 36). For plasma-derived C1INH, Patston et al. (8) determined a stoichiometry of 1.05:1 and 2.5:1 for the reactions with C1s and kallikrein, respectively. Only recombinant C1INH with an Asn at the P2 position showed altered reactivity with kallikrein or factor XIIa (in addition to the previously reported decreased reactivities with P2 Asp mutant (Fig. 4A). This mutant completely lost its ability to form a complex with factor XIIa, and had diminished
1503
FIGURE 4. Complex formation between recombinant inhibitors and plasma kallikrein (K) or coagulation factor XIIa. All conditions and analyses were as described for Fig. 1.
reactivity with kallikrein (Fig. 4A). These findings are essentially identical with the reactivities of the Ala4433Asp mutant with these proteases (26) (Table I). Each of the other mutants, including the P6 and the P8⬘/P9⬘ mutants, showed a pattern of reactivity with both plasma kallikrein and factor XIIa that was virtually identical with the reactivity of the wild-type inhibitor (Fig. 4, A–C). Plasmin. Normal recombinant wild-type C1INH reacts slowly with plasmin; complex formation was incomplete after the 1-h incubation (Figs. 5, A and C, and 6). No cleaved wild-type inhibitor was observed in this reaction mixture. The P2 Val and Leu mutants also complexed with plasmin (Figs. 5, A and B, and 6). However, there was a large quantity of cleaved C1INH present in each of these reactions. Very little complex formation was observed between plasmin and the P2 Asn, Gly, and Thr mutants (Fig. 5, A and B). The Ala4393Val mutant complexed with plasmin to an extent similar to that with the wild-type C1INH (Fig. 5C). In contrast, the P8⬘/P9⬘ mutant appeared to form less complex with plasmin than did the wild type during the 60-min incubation period. Thrombin. No significant stable complex formation was observed between wild-type C1INH and thrombin, but a moderate
1504
C1INH TARGET PROTEASE RECOGNITION
FIGURE 6. Complex formation between recombinant wild type (WT) and Ala4433 Val C1INH and plasmin. COS-1 cell supernatants (100 l) containing recombinant inhibitor intrinsically labeled with [35S]Met were incubated with the indicated quantities of plasmin for 1 h at 37°C. Immunoprecipitation, SDS-PAGE, and autoradiography were performed as for Fig. 1.
Discussion Substitution at the P2 position clearly has effects on reactivity with each of the proteases tested, including both target (C1r, C1s, plasma kallikrein, factor XIIa, plasmin) and nontarget proteases (thrombin, trypsin). This is in distinction to the observations with P2 substitutions in ␣1-antichymotrypsin, in which reactivities with factor Xa and activated protein C were markedly altered whereas that with trypsin was relatively unaffected (18). Several characteristics of the reactivities of these mutants are notable. As might be expected, bulky and/or charged amino acids were poorly tolerated at this position, and likely interfered with access of protease to the reactive center. The mutants with P2 Ala3 Asp and Asn substitutions reacted extremely poorly with all of the proteases tested
FIGURE 5. Complex formation between recombinant inhibitors and thrombin (Th) or plasmin (Pl). All conditions and analyses were as described for Fig. 1.
amount of cleaved inhibitor was generated during the 1-h incubation (Figs. 5A and 7). The P2 Val and P2 Leu mutants each formed a small amount of stable complex with thrombin but each was cleaved more efficiently than was the wild type (Figs. 5, A and B, and 7). The Asn, Gly, and Thr mutants did not complex with, and were resistant to cleavage by, thrombin. The P6 mutant formed little or no complex with thrombin, but was cleaved to a significant extent, whereas the P8⬘/P9⬘ mutant formed minimal or no complex, and was not cleaved. Trypsin. We previously demonstrated that the Ala4433Val mutant protein was relatively resistant to cleavage by trypsin in comparison with the wild-type protein and that it had acquired the ability to complex with trypsin at a very slow rate (26). The reactivities of each of the other mutants appeared much more similar to that of the wild-type protein in that they were readily cleaved and showed little or no complex formation with trypsin (data not shown).
FIGURE 7. Complex formation between recombinant wild type (WT) and Ala443 3 Val C1INH and thrombin. All conditions and analyses were as described for Fig. 6, except that the protease was thrombin rather than plasmin.
The Journal of Immunology (Table I). This observation is similar to previous findings with antithrombin and protein C inhibitor (37, 38). Also, the reactivities of the inhibitors with C1r and C1s tended to parallel one another, as did the reactivities with factor XIIa and kallikrein. Normal wildtype C1INH reacts more slowly with C1r than with C1s (32, 34). This difference in rate explains the observation that on SDSPAGE, incomplete complex formation is more apparent with C1r than with C1s. Given the similar substrate specificities of C1r and C1s and of factor XIIa and kallikrein, these results might have been expected. Virtually any change in side chain from the wild-type Ala resulted in decreased inhibitory activity against C1s (and C1r), whereas reactivity with kallikrein and factor XIIa was less sensitive to these substitutions. The sensitivity of reactivity with C1s to changes at P2 was apparent on SDS-PAGE for the Ala4433Asn and Asp mutants, and for the Ser, Thr, and Val mutants on kinetic analysis. Unfortunately, there were insufficient quantities of the Ala4433Gly and Leu variants for kinetic studies with C1s and of any of the recombinant inhibitors for kinetic analyses with proteases other than C1s. The slower rate of reactivity of each of these variants with C1r was apparent on SDS-PAGE. This high degree of specificity in the reactivity with C1s (and C1r) is consistent with data from the recently reported crystal structure of the catalytic domain of C1s (39). This demonstrated that access to the C1s binding subsites, particularly S2, is partially obstructed in comparison with many other proteases, such as trypsin. This restricted access, as suggested by Gaboriaud et al. (39), very likely explains the high degree of substrate specificity of C1s and the fact that its only inhibitor in plasma is C1INH. The observations related to the reactivity of the P2 substitution mutants with thrombin are also of interest in relation to specificity determination. The increased reactivity of the P2 Val and Leu mutants with thrombin, as demonstrated by both increased complex formation and increased cleavage in comparison with wild type, correlates with thrombin specificity toward fibrinogen. Fibrinopeptide B, which has an Ala at the P2 position, is released more slowly than is fibrinopeptide A, which has a Val at the P2 position (40). In general, it was observed that more hydrophobic residues in the P2-P4 positions resulted in substrates that were better recognized by thrombin (41). The somewhat increased complex formation with, and cleavage of, the P2 Val and Leu mutants are consistent with these earlier observations. However, based on the previous data (41), it might have been expected that the P2 Gly mutant also would have been recognized, but it neither was complexed with nor cleaved by thrombin. The Ala4393 Val (P6) substitution mutant, identified originally in a family with HAE (27), revealed only a moderate decrease in activity against C1s in comparison with the wild-type protein (Ka 2.14 vs 3.35 ⫻ 105 M⫺1s⫺1). Its reactivity with all of the proteases, including C1s (after a 1-h incubation), was indistinguishable from normal on SDS-PAGE. Current information suggests that angioedema is mediated by bradykinin (25, 42, 43). Therefore, prevention of attacks of angioedema very likely is dependent on the inhibition of plasma kallikrein (which catalyzes the cleavage of high m.w. kininogen to release bradykinin) by C1INH. Determination of rate constants for this mutant (when available in larger quantities) with kallikrein will be required to determine whether a relatively mild dysfunction of this inhibitor is responsible for the development of angioedema in this family. In addition, this family reportedly presented with laboratory values consistent with type 1 HAE (decreased C1INH antigenic and functional levels) as opposed to type 2 (decreased C1INH functional levels with relatively normal antigenic levels). This could be explained if the mutation resulted in either a decrease in secretion
1505 of newly synthesized protein or in enhanced intracellular degradation. The amount of recombinant mutant protein synthesized by COS cells did not dramatically differ from the amount of wild-type inhibitor synthesized, which argues against this explanation. Unfortunately, these hypotheses cannot be tested directly because this family is not currently available. In the original analysis of this family, this substitution was the only mutation identified within the coding region or at intron-exon junctions (27). In addition, the argument that this mutation is sufficient to result in HAE is supported by the observation that an identical mutation in antithrombin was described in a family with recurrent venous thromboses (28). The Gln452Gln4533 Ala substitution was analyzed to test the hypothesis that these residues make up part of a contact site for C1s. This was based on the observations that autoantibodies to C1INH react with a synthetic peptide corresponding to aa 448 – 459 and that the epitope included residues Gln452, Gln453, Pro454, and Phe455 (44). These authors then went on to report experiments that indicated that this peptide bound directly to C1s and prevented its complex formation with C1INH (45). Substitution of Gln452, Gln453, or Phe455 in this peptide abrogated C1s binding (46). This led to the conclusion that this region at the distal end of the reactive center loop constitutes a C1s contact site. The results of the experiments reported here show that substitution of both Gln452 and Gln453 with Ala residues has only a relatively marginal effect on the inhibition of C1s (or, of the other proteases, with the exception of plasmin) by C1INH. It seems unlikely that a contact site dependent on the side chains of two adjacent Gln residues would be adequately replaced by Ala residues. Therefore, it seems likely that these residues, at least, are not part of a major contact site for C1s in its recognition of C1INH. Recently, it was demonstrated that C1INH can be cross-linked to fibrin by tissue transglutaminase via Gln453, and that this immobilized C1INH retained its inhibitory activity toward kallikrein and C1s (47). The Gln452Gln453 mutant may prove to be valuable to analyze the potential biologic role of this cross-linking reaction. In summary, these studies have further characterized the roles of amino acids within the reactive center loop of C1INH in specificity determination. In particular, the results emphasize the importance of the amino acid at the P2 position in recognition by the highly specific proteases C1r and C1s. Conversely, plasma kallikrein and factor XIIa are much more tolerant of variation at this position.
References 1. Ratnoff, O., and I. Lepow. 1957. Some properties of an esterase derived from preparations of the first component of complement. J. Exp. Med. 106:327. 2. Schreiber, A. D., A. P. Kaplan, and K. F. Austen. 1973. Inhibition by C1-INH of Hageman factor fragment activation of coagulation, fibrinolysis, and kinin generation. J. Clin. Invest. 52:1402. 3. Ratnoff, O., J. Pensky, D. Ogston, and G. Naff. 1969. The inhibition of plasmin, plasma kallikrein, plasma permeability factor, and the C1⬘r subcomponent of complement by serum C1⬘ esterase inhibitor. J. Exp. Med. 129:315. 4. Gigli, I., J. W. Mason, R. W. Colman, and K. F. Austen. 1970. Interaction of plasma kallikrein with the C1 inhibitor. J. Immunol. 104:574. 5. Forbes, C., J. Pensky, and O. Ratnoff. 1970. Inhibition of activated Hageman factor and activated plasma thromboplastin antecedent by purified C1 inactivator. J. Lab. Clin. Med. 76:809. 6. Wuillemin, W. A., M. Minnema, J. C. Meijers, D. Roem, A. J. Eerenberg, J. H. Nuijens, H. ten Cate, and C. E. Hack. 1995. Inactivation of factor XIa in human plasma assessed by measuring factor XIa-protease inhibitor complexes: major role for C1-inhibitor. Blood 85:1517. 7. Carrell, R. W., and D. R. Boswell. 1986. Serpins: the superfamily of serine proteinase inhibitors. In Proteinase Inhibitors. A. Barrett, and G. Salvesen, eds. Elsevier, Amsterdam, p. 403. 8. Patston, P. A., P. Gettins, J. Beechem, and M. Schapira. 1991. Mechanism of serpin action: evidence that C1 inhibitor functions as a suicide substrate. Biochemistry 30:8876. 9. Olson, S. T., P. E. Bock, J. Kvassman, J. D. Shore, D. A. Lawrence, D. Ginsburg, and I. Bjork. 1995. Role of the catalytic serine in the interactions of serine proteinases with protein inhibitors of the serpin family: contribution of a covalent
1506
10.
11. 12.
13. 14. 15. 16. 17.
18.
19.
20. 21. 22.
23.
24. 25.
26.
27.
interaction to the binding energy of serpin-proteinase complexes. J. Biol. Chem. 15:30007. Loebermann, H., R. Tokuoka, J. Deisenhofer, and R. Huber. 1984. Human ␣1proteinase inhibitor: crystal structure analysis of two crystal modifications, molecular model and preliminary analysis of the implications for function. J. Mol. Biol. 177:531. Huntington, J. A., R. J. Read, and R. W. Carrell. 2000. Structure of a serpinprotease complex shows inhibition by deformation. Nature 407:923. Fa, M., F. Bergstrom, P. Hagglof, M. Wilczynska, L. B. Johansson, and T. Ny. 2000. The structure of a serpin-proteinase complex revealed by intramolecular distance measurements using donor-donor energy migration and mapping of interaction sites. Structure Fold Des. 8:397. Wright, H. T., and J. N. Scarsdale. 1995. Structural basis for serpin inhibitor activity. Proteins 22:210. Stratikos, E., and P. G. W. Gettins. 1997. Major proteinase movement upon stable serpin-proteinase complex formation. Proc. Natl. Acad. Sci. USA 94:453. Matheson, N., I. Bathurst, and J. Travis. 1989. The primary role of the P1 residue (Ser359) of ␣-1-proteinase inhibitor. Biochem. Biophsy. Res. Commun. 159:271. Owen, M. C., S. O. Brennan, J. H. Lewis, and R. W. Carrell. 1983. Mutation of antitrypsin to antithrombin. N. Engl. J. Med. 309:694. Patston, P. A., N. Roodi, J. A. Schifferli, R. Bischoff, M. Courtney, and M. Schapira. 1990. Reactivity of ␣1-antitrypsin mutants against proteolytic enzymes of the kallikrein-kinin, complement and fibrinolytic systems. J. Biol. Chem. 265:10786. Djie, M. Z., B. F. Le Bonniec, P. C. R. Hopkins, K. Hipler, and S. R. Stone. 1996. Role of the P2 residue in determining the specificity of serpins. Biochemistry 35:11461. Eldering, E., C. C. Haijbregts, J. H. Nuijens, A. J. Verhoeven, and C. E. Hack. 1993. Recombinant C1 inhibitor P5/P3 variants display resistance to catalytic inactivation by stimulated neutrophils. J. Clin. Invest. 92:1035. Stephens, A. W., A. Siddiqui, and C. H. W. Hirs. 1988. Site-directed mutagenesis of the reactive center (serine 394) of antithrombin III. J. Biol. Chem. 263:15849. York, J. D., P. Li, and S. J. Gardele. 1991. Combinatorial mutagenesis of the reactive site region in plasminogen activator inhibitor. J. Biol. Chem. 266:8495. Bottomley, S. P., and S. R. Stone. 1998. Protein engineering of chimeric serpins: an investigation into effects of the serpin scaffold and reactive centre loop length. Protein Eng. 11:1243. Chaillan-Huntington, C. E., P. G. W. Gettins, J. A. Huntington, and P. A. Patston. 1997. The P6 –P2 region of serpins is critical for proteinase inhibition and complex stability. Biochemistry 36:9562. Chaillan-Huntington, C. E., and P. A. Patston. 1998. Influence of the P5 residue on ␣1-proteinase inhibitor mechanism. J. Biol. Chem. 273:4569. Zahedi, R., J. J. Bissler, A. E. Davis III., C. Andreadis, and J. J. Wisnieske. 1995. Unique C1 inhibitor dysfunction in a kindred without angioedema. II. Identification of an Ala443-Val substitution and functional analysis of the recombinant mutant protein. J. Clin. Invest. 95:1299. Zahedi, R., J. Wisnieski, and A. E. Davis III. 1997. Role of the P2 residue of complement 1 inhibitor (Ala443) in determination of target protease specificity. J. Immunol. 159:983. Bissler, J. J., K. S. Aulak, V. H. Donaldson, F. S. Rosen, M. Cicardi, R. A. Harrison, and A. E. Davis III. 1997. Molecular defects in hereditary angioneurotic edema. Proc. Assoc. Am. Physicians 109:164.
C1INH TARGET PROTEASE RECOGNITION 28. White, D., G. Abraham, C. Carter, V. V. Kakkar, and D. N. Cooper. 1992. A novel missense mutation in the antithrombin III gene (Ala387-Val) causing recurrent venous thrombosis. Hum. Genet. 90:472. 29. Aulak, K. S., E. Eldering, C. E. Hack, Y. P. T. Lubbers, R. A. Harrison, A. Mast, M. Cicardi, and A. E. Davis III. 1993. A hinge region mutation in C1-inhibitor (Ala436Thr) results in nonsubstrate-like behavior and in polymerization of the molecule. J. Biol. Chem. 268:18088. 30. Harrison, R. A. 1983. Human C1 inhibitor: improved isolation and preliminary structural characterization. Biochemistry 22:5001. 31. Stein, P. E., D. A. Tewkesbury, and R. W. Carrell. 1989. Ovalbumin and angiotensinogen lack serpin S-R conformational change. Biochem. J. 262:103. 32. Nilsson, T., and B. Wiman. 1983. Kinetics of the reaction between human C1esterase inhibitor and C1r or C1s. Eur. J. Biochem. 129:663. 33. Lennick, M., S. A. Brew, and K. C. Ingham. 1986. Kinetics of interaction of C1 inhibitor with complement C1s. Biochemistry 25:3890. 34. Sim, R., G. Arlaud, and M. Colomb. 1980. Kinetics of reaction of human C1inhibitor with the human complement system proteases C1r and C1s. Biochim. Biophys. Acta 612:433. 35. Coutinho, M., K. S. Aulak, and A. E. Davis III. 1994. Functional analysis of the serpin domain of C1 inhibitor. J. Immunol. 153:3648. 36. Zahedi, R., K. S. Aulak, E. Eldering, and A. E. Davis III. 1996. Characterization of C1 inhibitor-Ta: a dysfunctional C1 inhibitor with deletion of lysine-251. J. Biol. Chem. 271:24307. 37. Phillips, J. E., S. T. Cooper, E. E. Potter, and F. C. Church. 1994. Mutagenesis of recombinant protein C inhibitor reactive site residues alters target proteinase specificity. J. Biol. Chem. 269:16696. 38. Sheffield, W. P., and M. A. Blajchman. 1994. Site-directed mutagenesis of the P2 residue of human antithrombin. FEBS Lett. 339:147. 39. Gaboriaud, C., V. Rossi, I. Bally, G. J. Arlaud, and J. C. Fontecilla-Camps. 2000. Crystal structure of the catalytic domain of human complement C1s: a serine protease with a handle. EMBO J. 19:1755. 40. Hanna, L. S., H. A. Scheraga, C. W. Francis, and V. J. Marder. 1984. Comparison of structures of various human fibrinogens and a derivative thereof by a study of the kinetics of release of fibrinopeptides. Biochemistry 23:4681. 41. Chang, J. Y. 1985. Thrombin specificity: requirement for apolar amino acids adjacent to the thrombin cleavage site for polypeptide substrate. Eur. J. Biochem. 151:217. 42. Shoemaker, L. R., S. J. Schurman, V. H. Donaldson, and A. E. Davis III. 1994. Hereditary angioneurotic edema: characterization of plasma kinin and vascular permeability-enhancing activities. Clin. Exp. Immunol. 95:22. 43. Nussberger, J., M. Cugno, C. Amstutz, M. Cicardi, A. Pellacani, and A. Agostoni. 1998. Plasma bradykinin in angio-oedema. Lancet 351:1693. 44. He, S., S. Tsang, J. North, N. Chohan, and R. B. Sim. 1996. Epitope mapping of C1 inhibitor autoantibodies from patients with acquired C1 inhibitor deficiency. J. Immunol. 156:2009. 45. He, S., R. B. Sim, and K. Whaley. 1997. A secondary C1s interaction site on C1-inhibitor is essential for formation of a stable enzyme-inhibitor complex. FEBS Lett. 405:42. 46. He, S., J.-C. Yang, S. Tsang, R. B. Sim, and K. Whaley. 1997. Role of the distal hinge region of C1-inhibitor in the regulation of C1s activity. FEBS Lett. 412: 506. 47. Hauert, J., P. A. Patston, and M. Schapira. 2000. C1 inhibitor cross-linking by tissue transglutaminase. J. Biol. Chem. 275:14558.
