(Received for publication, January 29, 1997)
From Gilead Sciences, Foster City, California 94404
Mutation of 79 highly exposed amino acids that comprise approximately 62% of the solvent accessible surface of thrombin identified residues that modulate the inhibition of thrombin by antithrombin III, the principal physiological inhibitor of thrombin. Mutations that decreased the susceptibility of thrombin to inhibition by antithrombin III in the presence and absence of heparin (W50A, E229A, and R233A) also decreased hydrolysis of a small tripeptidyl substrate. These residues were clustered around the active site cleft of thrombin and were predicted to interact directly with the "substrate loop" of antithrombin III. Despite the large size of antithrombin III (58 kDa), no residues outside of the active cleft were identified that interact directly with antithrombin III. Mutations that decreased the susceptibility of thrombin to inhibition by antithrombin III in the presence but not in the absence of heparin (R89A/R93A/E94A, R98A, R245A, K248A, K252A/D255A/Q256A) in general did not also affect hydrolysis of the tripeptidyl substrate. These residues were clustered among a patch of basic residues on a surface of thrombin perpendicular to the face containing the active site cleft and were predicted to interact directly with heparin. Three mutations (E25A, R178A/R180A/D183A, and E202A) caused a slight enhancement of inhibition by antithrombin III.
Vascular injury or inflammation results in the activation of thrombin (1, 2). Thrombin cleaves and activates multiple substrates to mediate hemostasis through fibrin clot formation and platelet aggregation (3-7) and also regulates blood coagulation by activating the anticoagulant protein C pathway (8). The predominant mechanism for the clearance of thrombin is through inhibition by antithrombin III (ATIII),1 a member of the family of serine protease inhibitors known as serpins (1, 9-11). The physiological importance of this process is illustrated by the observation that individuals possessing inherited ATIII deficiency or insufficiency are subject to recurrent thrombosis (12).
ATIII is a 58-kDa glycoprotein that inhibits thrombin in two kinetically distinct steps: the formation of a weak initial complex followed by rapid conversion to a stable complex (13). Based on the crystal structures of ATIII (14) and other serpins in cleaved (15, 16) and uncleaved (17) forms, inhibition has been proposed to proceed through the recognition of the Arg393-Ser394 bond in an exposed loop of ATIII as a substrate. However, upon cleavage, a conformational change in the inhibitor has been proposed to result in the trapping of the enzyme either as a tetrahedral intermediate or as an acyl-enzyme. Unlike the intermediate typically formed with substrates, the intermediate formed between thrombin and ATIII is stable and resistant to hydrolysis perhaps due to the conformational change which may exclude water from the active site (9).
The rate-limiting step in the inhibition of thrombin by ATIII is the
formation of the initial complex (~2 × 105
M1 min
1) which can be
accelerated by 3 orders of magnitude (~2 × 108
M
1 min
1) in the presence of the
glycosaminoglycan, heparin. Acceleration by heparin appears to be
mediated by two processes. The formation of a binary complex between
ATIII and heparin induces a conformational change in ATIII that
enhances association with thrombin and the interaction of heparin with
binding sites on both thrombin and ATIII in the ternary complex slows
the dissociation of the initial complex.
The analysis of naturally occurring mutants of ATIII has allowed extensive assignment of structure-function relationships. Mutations that affect inhibition of thrombin appear to fall into three functional classes that are spatially distinct (9, 10, 12, 16, 18). Mutations that map to the substrate loop of ATIII perturb direct interactions with thrombin and result in diminished potency or altered specificity of inhibition. Mutation of a cluster of positively charged residues that stretches from the base of the A helix across the D helix and extends toward the substrate loop affects heparin binding. Mutations in the sequences flanking the substrate loop limit the conformational change that accompanies inhibition and cause a loss of inhibitory activity or convert ATIII into a substrate. Mutations involving residues outside of the substrate loop of ATIII that are directly involved in interactions with thrombin have not been described.
The predominant structural features of the thrombin molecule include a deep canyon-like active site cleft formed between extended loops Leu46-Asn57 and Leu144-Gly155 that sterically restricts the substrate specificity of thrombin and two positively charged surfaces termed exosites I and II that are involved in supplementary interactions with many macromolecular substrates and inhibitors of thrombin (19, 20). In analogous fashion to the characterization of ATIII, the analysis of the interaction of thrombin with ATIII has focused on the recognition of the substrate loop of ATIII by the active site cleft (21-25) and the binding of heparin to exosite II (26-30). No studies have identified direct interactions between residues outside of the active site cleft of thrombin and ATIII.
We were interested in characterizing the interaction of thrombin with ATIII because naturally occurring mutations within the functional epitope may predispose individuals to thrombotic disease just as analogous functionally deficient variants of ATIII are prothrombotic (10, 12). In addition, the in vivo potency of a novel class of anticoagulant agents, where thrombin has been genetically engineered to function selectively as an activator of endogenous protein C (31-33), can be modulated by altering clearance by ATIII. To identify residues on thrombin that are important for inhibition by ATIII, including residues outside of the active site cleft, we constructed a library of site-directed mutants of thrombin in which a total of 79 residues that are most exposed on the surface of thrombin were mutated to alanine (34). This library of mutants was screened for its sensitivity to inhibition by ATIII in the presence and absence of heparin and the functional requirements for ATIII inhibition were compared with those required for the hydrolysis of a small peptidyl substrate that interacts only with the active site cleft.
The human prothrombin gene was inserted into the expression vector pRc/CMV (Invitrogen). Codons for the 79 polar and charged amino acids that displayed greater than 35% fractional accessibility to a solvent probe of radius 1.4 Å were mutated to alanine by oligonucleotide-directed mutagenesis. Mutant prothrombin in conditioned medium from transiently transfected COS-7 cells was concentrated by ultrafiltration, processed to thrombin by Echis carinatus venom, and quantitated by enzyme-linked immunosorbent assay. All these methods have been described in detail previously (32, 34).
Steady-state Kinetics of S-2238 Hydrolysis by Thrombin MutantsThrombin mutants in conditioned media were first assayed for amidolytic activity at 100 µM S-2238 (H-D-Phe-Pip-Arg-p-nitroanilide) in 20 mM Tris, pH 8.3, 150 mM NaCl. An amount of mutant thrombin that gave a change in absorbance of 0.01 A405/min was then used for each substrate concentration point for the Km determination. The final concentration of thrombin in the reaction was ~0.74 nM for most mutants. The initial rates of S-2238 hydrolysis were measured, by monitoring the absorbance change at 405 nm, at the following substrate concentrations: 2, 4, 6, 8, 10, 15, 20, 25, 35, 50, 70, and 100 µM. Kinetic parameters, kcat and Km were determined by fitting the Michaelis-Menten equation to the data.
Inhibition of Thrombin Mutants by Antithrombin III in the Absence of HeparinThrombin mutants were rapidly screened for resistance to inhibition by ATIII by diluting 0.5 pmol of thrombin mutant in selection buffer (20 mM Tris acetate, pH 7.5, 140 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2) plus 0.1 µg/ml bovine serum albumin and 4.5 µg/ml Polybrene (to neutralize any contaminating heparinoids) in quadruplicate wells of a 96-well microtiter plate. The final reaction volume was 100 µl. Purified human ATIII (Enzyme Research Laboratories) was added to duplicate wells to a final concentration of 1.0 µM. The inhibition reaction was incubated at 25 °C for 30 min and then stopped by the addition of 100 µM of the chromogenic substrate S-2238. The chromogenic substrate served a dual purpose, halting the inhibition reaction by competition and providing a means of quantitating the residual activity due to uninhibited thrombin. This was accomplished by monitoring the change in absorbance at 405 nm over a 2-min interval. Percent inhibition was calculated from the ratio of activity remaining in the presence versus the absence of ATIII. Under these conditions, wild-type thrombin was inhibited by 99%.
Selected thrombin mutants were analyzed in more detail to determine the
second-order rate constants for inhibition by ATIII alone. Between 2 and 5 pmol of each thrombin mutant was diluted into 200 µl of
selection buffer plus 0.1 µg/ml bovine serum albumin and 4.5 µg/ml
Polybrene containing 0.5-5 µM ATIII. The reaction was
incubated at 25 °C and at regular intervals over a 30-min time
course 20-µl samples were withdrawn and diluted into 180 µl of 100 µM S-2238 to terminate the reaction. The rate of
hydrolysis of S-2238 due to residual uninhibited thrombin was
determined by monitoring the change in absorbance at 405 nm. Under
these pseudo-first order conditions the time course of inhibition was fitted to the equation: Et = E0ek
t, where
E0 is the thrombin activity at time
t = 0, Et is the thrombin activity
at time t, and k
is the observed pseudo-first order rate constant. The second-order rate constant of inhibition (k2) was calculated by dividing k
by
the ATIII concentration.
Inhibition by ATIII in the presence of
heparin was performed in the presence of competing substrate (S-2238)
as described previously (30). Briefly, 150 nM ATIII, 0.2 unit/ml heparin, and 150 µM S-2238 were premixed in 590 µl of reaction buffer containing: 20 mM Tris acetate, pH
7.5, 140 mM NaCl, 5 mM KCl, 1 mM
MgCl2, 1 mM CaCl2, and 0.1%
PEG-8000. This mixture was added to a spectrophotometric cuvette
containing 10 µl of 300 nM mutant thrombin (final
concentration of 5 nM), and hydrolysis of S-2238 was
immediately measured for 2 min. The pseudo-first order rate constant of
inhibition k was determined by curve fitting as described
previously (32). The second-order rate constant,
k2, was determined from k
after
correcting for substrate competition using the Km
value of S-2238 hydrolysis (32).
Most of the thrombin mutants hydrolyzed the tripeptidyl
chromogenic substrate S-2238 with catalytic efficiencies
(kcat/Km) comparable to that
of wild-type thrombin (~40 s1
µM
1). Only 3 mutants (W50A, N53A/T55A, and
E229A) displayed catalytic efficiencies that were reduced by greater
than 5-fold relative to wild-type thrombin (Table I).
Asn53 is the sole site of N-linked glycosylation
in thrombin and its substitution may affect the thermal stability of
mutant N53A/T55A (data not shown).
|
All 56 mutants were screened
for inhibition by ATIII both in the presence and absence of heparin.
Heparin-independent ATIII inhibition was determined at a single time
point and expressed as percent inhibition. For ATIII inhibition in the
presence of 0.2 unit/ml heparin, the second-order rate constant of
inhibition k2 was determined (Fig.
1). Three mutants, W50A, E229A, and R233A were more
resistant to inhibition by ATIII in the absence of heparin compared
with wild-type thrombin. Predictably, these same three mutants were
also refractory to heparin-dependent ATIII inhibition (Fig.
1). In addition to the three mutants above, there were five additional
mutants (R89A/R93A/E94A, R98A, R245A, K248A, and K252A/D255A/Q256A) that were refractory to ATIII inhibition only in the presence of
heparin. Three other mutants (E25A, R178A/R180A/D183A, and E202A)
displayed slightly increased sensitivity toward
heparin-dependent ATIII inhibition (Fig. 1).
Mutant Thrombins Defective in Inhibition by Antithrombin III
The eight mutants found to be refractory to ATIII inhibition in the absence or presence of heparin were further analyzed in a time course assay for inhibition by ATIII alone. The second-order rate constant of inhibition in the absence of heparin, k2, was determined and reported together with the k2 for heparin-dependent ATIII inhibition (Table II). The k2 of mutants W50A, E229A, and R233A was decreased by 6.3-14.7-fold for heparin-independent ATIII inhibition and similarly by 8.3-14.4-fold for heparin-dependent inhibition (Table II). These three mutations most probably weakened direct interactions between thrombin and ATIII.
|
The k2 for mutants R89A/R93A/E94A, R98A, R245A, K248A, and K252A/D255A/Q256A was similar to that of the wild-type for heparin-independent ATIII inhibition but was decreased by 7.7-43-fold for heparin-dependent ATIII inhibition. These mutants were most probably weakened in their interactions with heparin. Heparin acceleration of ATIII inhibition of mutant K248A was unmeasureable at 0.2 unit/ml heparin. However, heparin-dependent inhibition of K248A by ATIII was detectable when the concentration of heparin was increased by 5-fold (1.0 units/ml) although reduced 14.6-fold relative to wild-type thrombin at the same increased concentration of heparin (data not shown).
Mutant Thrombins with Enhanced Inhibition by ATIIIThree thrombin mutants (E25A, R178A/R180A/D183A, and E202A) which displayed increased sensitivity to heparin-dependent ATIII inhibition were also further analyzed in a time course assay for inhibition with ATIII alone. The second-order rate constant of inhibition, k2, were reported in Table III. Whether in the absence or presence of heparin, the increase in k2 was 1.5-2.2-fold, suggesting that these mutations probably enhanced the interaction of thrombin with ATIII.
|
A library of thrombin mutants was screened to probe the structural requirements for the interaction of thrombin with antithrombin III, the most physiologically significant inhibitor of thrombin. The 79 amino acid residues mutated are in the upper quartile of thrombin residues in terms of their exposure to solvent and account for approximately 62% of the total solvent accessible surface area of thrombin. Therefore this study represents a comprehensive analysis because it focuses on the residues most likely to be available for interactions with other macromolecules and benefits from the supposition that the surface residues of proteins are more tolerant to sequence variation (35) and thus their substitution is less likely to cause nonspecific structural perturbations.
Mutations that influenced the recognition of ATIII had three distinct
phenotypes and the residues substituted in each class were clustered in
three distinct spatial locations on the surface of thrombin. Three
mutations decreased the susceptibility of thrombin to inhibition by
ATIII in both the presence and absence of heparin. The residues
substituted in these mutants (Trp50, Glu229,
and Arg233) were predicted to be involved in direct
interactions with ATIII. These residues were clustered around the
active site cleft of thrombin and most likely contact the exposed
substrate loop of ATIII (Fig. 2). This interpretation is
supported by the observation that the mutation of these residues also
significantly decreased the recognition of a small tripeptidyl
substrate, S-2238. Notably, there were no residues outside of the
active site cleft that were identified as interacting directly with
ATIII. This is notable in light of the well known propensity of
thrombin to utilize exosite interactions for the recognition of
macromolecular substrates and inhibitors including fibrinogen, platelet
thrombin receptor, protein C, thrombomodulin, and hirudin (19, 20, 34,
36). ATIII is a large protein (58 kDa) yet thrombin recognizes it as it
recognizes a tripeptidyl substrate, utilizing the active site cleft
alone to interact with the exposed substrate loop.
In previous studies, deletions of three residues in the two loops that define the active site cleft of thrombin, Pro48-Trp50 (22, 23) and Glu146-Trp148 (21) decreased the rate of association with ATIII by 100- and 350-fold, respectively. These deletions also had large effects on the hydrolysis of peptidyl and other substrates suggesting the possibility that some proportion of their effects may be attributable to nonspecific structural perturbations caused by the introduction of a relatively unwieldy mutation in close proximity to the active site. Subsequent substitution of Trp50 with alanine decreased the on-rate of ATIII by 13-fold in the absence of heparin and 7-fold in the presence of heparin (25) and single substitutions of Trp148 and Lys154 in the Leu144-Gly155 loop had insignificant effects on the recognition of ATIII (24), which is in close agreement with this study.
The substrate specificity of thrombin has been shown to respond allosterically to Na+ binding (37) and the Na+ free form is less sensitive to inhibition by ATIII both in the presence and absence of heparin (38). X-ray crystallography revealed that the Na+ ion is coordinated by the carbonyl oxygens of Tyr190, Arg233, and Lys236 and mutation of Arg233 and Lys236 and flanking residues Asp232, Asp234, and Tyr237 (33, 38) decreased the response to Na+ and decreased sensitivity to inhibition by ATIII. Therefore the effects of substituting Arg233 and Glu229 (which is salt-bridged to Lys236) on inhibition by ATIII observed in our study may be partially mediated through undefined conformational changes that result from perturbing Na+ binding. Alternatively, the effects of Na+ binding may be mediated by perturbing the conformation of the side chains of Glu229 and Arg233 which may be in direct contact with ATIII.
The second class of mutations was composed of those that displayed resistance to inhibition by ATIII in the presence of heparin but not in the absence of heparin. The residues substituted in these mutants (Arg89, Arg93, Glu94, Arg98, Arg245, Lys248, Lys252, Asp255, Glu256) are candidates for those involved in direct interactions with heparin. These residues are clustered on a surface of thrombin perpendicular to the face containing the active site cleft (Fig. 2). Many of these residues are included in the cluster of basic amino acids that is referred to as thrombin exosite II (19). Previous studies (26, 29) have implicated several of these residues (Arg89, Arg93, Arg98, Arg245, Lys248, Lys252) in the recognition of heparin, including two studies (27, 28) where positively charged amino acids in exosite II were substituted with negatively-charged glutamic acid. Because some of the effects of these non-conservative mutations may be due to the detrimental effects of introducing a residue of opposite charge, the observation of similar defects following replacement with the small neutral amino acid alanine in our study confirms the likely role of these residues in electrostatic interactions with the negatively charged glycosaminoglycan. Several additional basic residues (Arg123, Arg170, Lys174, Arg178, Arg180) have been included in the structural definition of exosite II (19) due to their spatial proximity to the residues listed above. Lys174 was protected from chemical modification by heparin (26) and non-conservative substitution of Arg180 with glutamic acid decreased the efficiency of heparin catalyzed inhibition by ATIII (28). However, none of these residues were implicated in the recognition of heparin following substitution with alanine (Fig. 1).
The third class of mutations was composed of those that displayed slightly enhanced sensitivity to inhibition by ATIII both in the presence and absence of heparin (Table III). The residues substituted in these mutants (Glu25, Arg178, Arg180, Asp183, Glu202) may be involved in unfavorable interactions with ATIII perhaps through steric hindrance and the substitution of these residues may enhance the binding energy of the thrombin-ATIII complex. Consistent with our study, the substitution of Glu202 with glutamine was previously shown to increase by 1.5-fold the second-order rate constant for inhibition by ATIII (39) and the mutant thrombin was proposed to exhibit a preference for substrates, such as ATIII, with glycine at position P2.
Inhibition by ATIII is the primary mechanism for the clearance of active thrombin in vivo (11) and we have previously shown that mutations in thrombin that confer resistance to ATIII in vitro may correlate with a prolonged half-life in plasma and decreased clearance rate in vivo (31, 32). By analogy with naturally occurring mutations in ATIII (10, 12), mutations in thrombin that modulate inhibition by ATIII, such as those described in this study, may predispose individuals to an increased risk of thrombosis. Mutations involving residues in the heparin-binding site (R89A/R93A/E94A, R89A, R245A, K248A, K252A/D255A/Q256A) that are resistant to glycosaminoglycan-catalyzed inhibition by ATIII yet retain activity toward peptidyl substrates (Table I) and substrates in vivo (34) are functionally equivalent to type IIc mutations in ATIII (mutations in the heparin-binding site of ATIII) (10). Individuals possessing such mutations may be prone to thromboembolic disease. In contrast, the mutations identified in this study involving residues that are implicated in direct interactions with ATIII (Trp50, Glu229, Arg233) would not be expected to be prothrombotic. Because of their proximity to the active site, the mutation of these residues also modulates the recognition of other substrates and alters the specificity of thrombin to favor the anticoagulant substrate protein C (31, 32). Thus naturally occurring mutations in this class may have a net anticoagulant effect and may be protective of thrombosis.
We thank Drs. Lisa R. Paborsky, Mark D. Matteucci, and Larry L. K. Leung for helpful discussions regarding this project and valuable contributions to this manuscript.