Expression and Characterization of the Naturally Occurring Mutation L394R in Human γ-Glutamyl Carboxylase*

Patients with mutation L394R in γ-glutamyl carboxylase have a severe bleeding disorder because of decreased biological activities of all vitamin K-dependent coagulation proteins. Vitamin K administration partially corrects this deficiency. To characterize L394R, we purified recombinant mutant L394R and wild-type carboxylase expressed in baculovirus-infected insect cells. By kinetic studies, we analyzed the catalytic activity of mutant L394R and its binding to factor IX's propeptide and vitamin KH2. Mutant L394R differs from its wild-type counterpart as follows: 1) 110-fold higher K i for Boc-mEEV, an active site-specific, competitive inhibitor of FLEEL; 2) 30-fold lower V max/K m toward the substrate FLEEL in the presence of the propeptide; 3) severely reduced activity toward FLEEL carboxylation in the absence of the propeptide; 4) 7-fold decreased affinity for the propeptide; 5) 9-fold higher K m for FIXproGla, a substrate containing the propeptide and the Gla domain of human factor IX; and 6) 5-fold higher K m for vitamin KH2. The primary defect in mutant L394R appears to be in its glutamate-binding site. To a lesser degree, the propeptide and KH2 binding properties are altered in the L394R mutant. Compared with its wild-type counterpart, the L394R mutant shows an augmented activation of FLEEL carboxylation by the propeptide.

Vitamin K-dependent carboxylase, also known as ␥-glutamyl carboxylase, an integral membrane protein residing in the rough endoplasmic reticulum, catalyzes the posttranslational modification of specific glutamic acid residues to ␥-carboxyglutamic acid (Gla) 1 in vitamin K-dependent proteins (1). Glacontaining proteins are involved in blood coagulation (2), bone metabolism (3), and regulation of cell proliferation (4). The Gla domains of blood coagulation and anticoagulation proteins mediate calcium-dependent interactions between the protein and phospholipid membranes (5), a process necessary for the biological activity of these proteins. In addition to the glutamate substrate, ␥-glutamyl carboxylation requires carbon dioxide, oxygen, and the essential cofactor vitamin K hydroquinone (KH 2 ), which is the reduced form of vitamin K (6). The formation of Gla from glutamate is coupled with the conversion of vitamin KH 2 to vitamin K 2,3-epoxide. Both of these activities occur in the vitamin K-dependent carboxylase (7,8). The warfarin-sensitive microsomal enzyme vitamin K epoxide reductase recycles the epoxide back to vitamin KH 2 (9), thus completing the vitamin K cycle.
There have been only a few cases of combined deficiencies of vitamin K-dependent coagulation factors reported (10 -18). Patients with this disorder suffer from a bleeding diathesis due to deficiencies of prothrombin and factors VII, IX, and X. In addition, the anticoagulation activities of proteins C and S are decreased. Brenner et al. (16) reported that patients' coagulation activities were severely reduced, while their antigen levels were only moderately decreased. Abnormality of the vitamin K epoxide reductase was ruled out because the patients had undetectable serum levels of vitamin K epoxide, and the hepatic vitamin K intake and skeletal development were both normal (18). Because the coagulation activities of these patients were partially corrected by administration of vitamin K and because prothrombin with impaired Gla-mediated calcium binding was identified, a defect in the vitamin K-dependent carboxylase was suspected (16). Genetic analysis of the proband (16) and her three affected siblings (18) revealed a mutation in the ␥-glutamyl carboxylase gene (18). All were homozygous for a T 3 G point mutation in exon 9 of the carboxylase gene. This mutation results in the substitution of arginine for leucine at residue 394, a conserved residue found in the ␥-glutamyl carboxylase of human (19), bovine (20), rat (21), Drosophila (22), mouse, whale, and toadfish (23). Furthermore, leucine 394 resides in a conserved region with 90% sequence identity spanning residues 374 -405 (22,23), which implies a functional or structural importance for this region. In this report, we compare the functional properties of the L394R mutant ␥-glutamyl carboxylase to those of its wild-type counterpart. Our results indicate that the major defect of mutation L394R is at or near its glutamate-binding site. Furthermore, an augmented allosteric effect between the propeptide-and glutamate-binding sites is observed in the L394R mutant.

EXPERIMENTAL PROCEDURES
Materials-FLEEL was purchased from Bachem (Philadelphia, PA). L-␣-Phosphatidylcholine (type V-E), CHAPS, and pepstatin were from Sigma. Vitamin K 1 was from Abbott. FIXproGla, a 59-residue recombinant peptide containing the propeptide and first 41 residues of the Gla domain of human factor IX, was prepared and purified as described (24). The peptide ProFIX19, which contains the sequence AVFLD-HENANKILNRPKRY, was synthesized by Dr. Frank Church (University of North Carolina, Chapel Hill). The peptide Boc-mEEV was synthesized as described (25). NaH 14 CO 3 (specific activity, 56 mCi/mmol) was from ICN Pharmaceuticals (Costa Mesa, CA). Leupeptin, aprotinin, and phenylmethylsulfonyl fluoride were from Roche Molecular Biochemicals. The pSK Ϫ vector was from Stratagene (La Jolla, CA). The pVL1392 vector was from Pharmingen (San Diego, CA). The BacVector 3000 baculoviral DNA was from Novagen (Madison, WI). Sf9 (Spodoptera frugiperda) insect cells were obtained from the Lineberger Cancer Center at the University of North Carolina (Chapel Hill, NC). High Five (Trichoplusia ni) insect cells were provided by Dr. Thomas Kost of Glaxo Wellcome. HPC4 antibody affinity resin was provided by Dr. Charles Esmon (Oklahoma Medical Research Foundation, Oklahoma City, OK). The anti-FLAG M2 monoclonal antibody and Met-FLAGbacterial alkaline phosphatase were from Sigma. Peroxidase-conjugated goat anti-mouse immunoglobulin was from Jackson ImmunoResearch (West Grove, PA). ECL Western blotting detection reagents were from Amersham Pharmacia Biotech.
Expression and Purification of Recombinant Wild-type and L394R Mutant Carboxylase-The cDNA for human ␥-glutamyl carboxylase (19), cloned into the pSK Ϫ vector, was modified by site-specific mutagenesis (26) to make the L394R mutant. Both wild-type and mutant constructs contain the FLAG epitope (DYKDDDDK) attached to their amino termini and the HPC4 tag containing the sequence EDQVD-PRLIDGK (27) at their carboxyl termini. The engineered DNA constructs, coding for wild-type and mutant carboxylase, were subcloned into the pVL1392 vector, and the proteins were expressed in baculovirus-infected High Five cells as described (28).
Isolation of microsomes from High Five cells was performed as described (29) with minor modifications. Cells (8 -10 ϫ 10 8 ) harvested from 1 liter of a culture were washed with 300 ml of cold buffer containing 25 mM Tris, pH 7.4, 150 mM NaCl, and 15% glycerol. The washed cell pellet was resuspended in 120 ml of cold buffer containing protease inhibitor mixture, which contains 2 mM dithiothreitol, 0.5 g/ml leupeptin, 1 g/ml pepstatin, 2 g/ml aprotinin, and 0.1 mg/ml phenylmethylsulfonyl fluoride, and disrupted by sonication (with 80 pulses, 1.5 s each, using a Heat Systems XL2020 sonicator at a power output of 6). The homogenate was centrifuged at 4300 ϫ g avg at 4°C for 15 min. The supernatant was recovered and centrifuged at 150,000 ϫ g avg for 1 h. at 4°C. Solubilization of the microsomal pellet and the subsequent purification of carboxylase using the HPC4 antibody affinity resin were performed as described (28).

SDS-Polyacrylamide Gel Electrophoresis and Western Blot
Analysis-Purified carboxylase was analyzed by silver-stained SDS-polyacrylamide gel electrophoresis (10% polyacrylamide gels; Bio-Rad) and by Western blot analysis. For Western blot analysis, the proteins transferred to a polyvinylidene difluoride membrane were probed with the anti-FLAG M2 monoclonal antibody (1.4 g/ml) and then with the peroxidase-conjugated secondary antibody (0.07 g/ml). The FLAG tagcontaining proteins were detected by chemiluminescence following incubation of the membranes with ECL reagents (Amersham Pharmacia Biotech) and autoradiography on Hyperfilm (Amersham Pharmacia Biotech). The wild-type and mutant carboxylases were quantitated by dot blot analysis on polyvinylidene difluoride membranes using known amounts of Met FLAG-bacterial alkaline phosphatase as standards. The FLAG tag-containing carboxylase and the standard were detected using anti-FLAG M2 antibody. Quantification was made based on the analysis of autoradiographs using an ImageQuant densitometer (Molecular Dynamics, Inc., Sunnyvale, CA).
Vitamin K Epoxidase Activity Assay-Carboxylations of FLEEL in the presence of 5 M ProFIX19 were performed as described above, except that unlabeled NaHCO 3 was used. FLEEL concentrations of 1 and 12 mM were used in the wild-type and mutant reactions, respec-tively. Vitamin K epoxide formation was quantitated as described (30).
Kinetic Studies-For the kinetic studies of FLEEL carboxylation, 5 M ProFIX19 was included. All components used in the assay were premixed except vitamin KH 2 and NaH 14 CO 3 , which were added to start the reaction. The assays were carried out for 30 min for carboxylation of FLEEL and 1 h for carboxylation of FIXproGla. In vitamin KH 2 kinetic studies, the concentration of dithiothreitol was kept constant (6 mM) in all reactions, which were performed for 30 min.
Inhibition of FIXproGla Carboxylation by Free Propeptide-We examined the effect of the free propeptide on the carboxylation of FIXpro-Gla, which contains a propeptide covalently linked to the Gla domain of factor IX. ProFIX19 was used to compete with 0.5 M FIXproGla for the wild-type carboxylase and with 4.0 M FIXproGla for the mutant L394R.
Stimulation of FLEEL Carboxylation by ProFIX19 -The reactions were carried out as described above, using 1 mM of FLEEL for the wild-type carboxylase and 12 mM of FLEEL for the mutant L394R and varying the ProFIX19 concentration up to 2.4 M.

Expression of the Recombinant Wild-type and L394R Mutant
Carboxylase-The recombinant wild-type and L394R mutant carboxylase, expressed in baculovirus-infected High Five cells, contain FLAG tags at their amino termini and HPC4 tags at their carboxyl termini. These epitopes, which lie outside the coding region of the carboxylase, facilitate the identification, purification, and quantification of the recombinant proteins. The recombinant carboxylase was affinity-purified using an HPC4 antibody column. The purification method employed provides carboxylase of high purity (Fig. 1). The estimated molecular mass of the proteins is around 95 kDa, similar to that reported for the purified carboxylase (32). Western blot analysis revealed a single antibody-reactive band at 95 kDa in both the wild-type and mutant carboxylase preparations (Fig. 1) and indicates the absence of proteolysis in either preparation. Effect of ProFIX19 on FLEEL Carboxylation-The free propeptide of the vitamin K-dependent coagulation proteins has been shown to enhance the rate of carboxylation of small glutamate-containing peptide substrates that lack a covalently linked propeptide sequence (33,34). We therefore used Pro-FIX19, the propeptide sequence of factor IX, to compare its effect on the carboxylation of FLEEL by the wild-type and mutant carboxylases. In the absence of the free propeptide, the wild-type carboxylase exhibits significant activity toward FLEEL carboxylation. As shown in Fig. 2, the rate of FLEEL carboxylation by the wild-type carboxylase increased 5-6-fold when saturating amounts of ProFIX19 were added. In contrast, mutant L394R showed almost undetectable activity toward FLEEL in the absence of the free propeptide. However, this deficiency can be ameliorated by adding ProFIX19. Compared with its wild-type counterpart, a higher concentration of Pro-FIX19 was required to achieve maximal stimulation of FLEEL carboxylation by mutant L394R. The concentrations of pro-FIX19 determined for half-maximal stimulation were 26 Ϯ 3 and 409 Ϯ 27 nM for the wild type and mutant L394R, respectively. These results demonstrate that mutant L394R has a decreased affinity for ProFIX19 when compared with the wildtype enzyme.
FLEEL Carboxylation-Kinetic constants for FLEEL carboxylation were determined in the presence of saturating concentrations of ProFIX19 (Table I). As shown in Fig. 3, substrate inhibition was apparent at high concentrations of FLEEL. This phenomenon was previously seen in the purified bovine carboxylase (8). Inhibition occurred when FLEEL concentrations were greater than 9 mM for the wild type and 18 mM for the mutant L394R. All kinetic constants were derived from the Michaelis-Menten equation using data below the inhibitory concentrations of FLEEL. Comparison of V max for FLEEL carboxylation showed that the reaction rate of mutant L394R was 2.5-fold slower than that of the wild-type (Fig. 3, Table I). The K m of FLEEL for mutant L394R is 6.49 Ϯ 0.72 mM, which is 12-fold higher than that of the wild-type (0.54 Ϯ 0.10 mM). It was impossible to determine the K m of FLEEL for mutant L394R in the absence of the propeptide, even when FLEEL concentrations were increased up to 60 mM, because its activity is almost undetectable under such conditions.
Active Site Inhibition Studies-To examine the glutamate binding properties of mutant L394R, we used the active sitespecific, competitive inhibitor Boc-mEEV (25). As shown in Fig.  4, by fitting the data into the competitive inhibition equation, the K i of Boc-mEEV is 0.013 Ϯ 0.002 mM for wild-type carboxylase compared with 1.44 Ϯ 0.16 mM for mutant L394R. These results revealed a 110-fold difference in the apparent affinities of wild type and L394R for Boc-mEEV.
Vitamin K Epoxide Formation-The rate of vitamin K epoxide formation was 20.3 Ϯ 0.5 pmol of KO/min/pmol of mutant L394R versus 36.6 Ϯ 3.9 pmol of KO/min/pmol of wild-type carboxylase. This 1.8-fold reduction in epoxide formation observed in mutant L394R compared with the wild type carboxylase parallels the 2.5-fold difference observed in FLEEL carboxylation, suggesting that the L394R mutant's carboxylase and epoxidase activities remain coupled (8,36).
Carboxylation of FIXproGla-FIXproGla, a peptide substrate containing the propeptide covalently linked to the Gla domain of human factor IX, mimics the physiological substrate of vitamin K-dependent carboxylase (24,35). Carboxylation of FIXproGla by the wild-type and mutant L394R carboxylase follows Michaelis-Menten kinetics. The K m of FIXproGla by wild-type carboxylase was 0.23 Ϯ 0.03 M, which is 3 orders of magnitude lower than that for FLEEL and is in agreement with previous reports (24,37,38). As shown in Fig. 5 and Table  I, the K m of FIXproGla is 2.08 Ϯ 0.23 M for mutant L394R, which is 9-fold higher than that of the wild-type carboxylase. Interestingly, the V max of FIXproGla carboxylation by mutant L394R was 2-fold higher than that by the wild-type carboxylase.

Inhibition of FIXproGla Carboxylation by Free Propeptide-
The propeptide of factor IX has been shown to be a competitive inhibitor of FIXproGla carboxylation (28,35). To obtain a better estimate of their relative affinities for the propeptide, we compared the K i values of the wild type and L394R mutant for the inhibition of FIXproGla carboxylation by ProFIX19. The K i values determined for wild-type and mutant L394R carboxylases were 63 Ϯ 13 and 459 Ϯ 28 nM, respectively (Fig. 6). This observed decrease in the relative affinity toward ProFIX19 by the mutant L394R agrees with our results obtained from studies of ProFIX19 activation of FLEEL carboxylation.
Effect of Vitamin K Hydroquinone on Carboxylation-Since mutant L394R has a higher K m for FIXproGla than that of the wild-type carboxylase (Table I), 10.0 and 1.2 M FIXproGla were used for mutant L394R and for the wild-type carboxylase, respectively, in this study; the experiment was designed so that the substrates were 5-fold greater than the K m for each enzyme. As shown in Fig. 7 and Table I, the K m of vitamin KH 2 was 7.0 Ϯ 1.2 M for the wild-type carboxylase and 32.8 Ϯ 5.4 M for mutant L394R. The V max of vitamin KH 2 for the carboxylation of FIXproGla was 1.9-fold higher for the mutant L394R than that for the wild-type carboxylase. DISCUSSION The purposes of this study were to identify the mechanistic defect of mutant L394R ␥-glutamyl carboxylase (16,18) and to relate the functional properties of the mutant enzyme to the clinical symptoms of patients bearing this mutation. Our data suggest that the major defect of L394R is its ability to bind glutamate-containing substrates. The most convincing evidence for this conclusion is that the K i of the competitive inhibitor of FLEEL carboxylation, Boc-mEEV, is 110-fold higher for L394R than that for the wild-type carboxylase (Fig.  4). Furthermore, our results demonstrate that, in the absence of the propeptide, the catalytic activity of L394R toward FLEEL carboxylation is less than 1% of the wild type's activity. It is striking that saturating concentrations of the factor IX propeptide increased carboxylation of FLEEL by L394R 200 - 300 times (compared with 5-6 times for wild-type carboxylase). Nevertheless, primarily because of its 12-fold increased K m , the catalytic efficiency (V max /K m ) of L394R toward the small sub-strate FLEEL was, in the presence of propeptide, still 30-fold lower than wild-type carboxylase (Table I). While the K m can only be used as an approximation of substrate affinity, the increased K m for FLEEL is consistent with our results with the competitive inhibitor Boc-mEEV. Thus, the mutation of leucine    In contrast to FLEEL, the more physiological substrate, FIX-proGla, which has 12 Glu residues, has a slightly higher (ϳ2fold) V max and a 9-fold higher K m for the mutant enzyme than for the wild-type ␥-glutamyl carboxylase. Other carboxylase mutations with higher V max and K m have been reported (39). The increased V max for the mutant is consistent with our previous suggestions that product release is the rate-limiting step for propeptide-containing substrates (the propeptide of the vitamin K-dependent proteins is the primary binding site for the carboxylase-substrate interaction (35,40)) and that all carboxylations occur during a single binding event (41). Since the assay measures only CO 2 incorporation and not complete carboxylation of a 12-Glu substrate, any decrease in peptide affinity can increase the V max , due to an increased off-rate, and may also result in undercarboxylated products. For L394R, FLEEL's V max is slower because CO 2 incorporation, rather than off-rate, is the rate-limiting step.
L394R also has smaller but significant effects on the propeptide and vitamin K interactions. The propeptide concentration required for half-maximal stimulation of FLEEL carboxylation was 15-fold higher, while the K i for the inhibition of CO 2 incorporation into FIXproGla by FIX's propeptide was 7-fold higher for L394R than for the wild-type enzyme. L394R's K m for vitamin K was also 5-fold higher than that of the wild-type enzyme. These results seem consistent with previous studies suggesting that the propeptide, vitamin K, and glutamate binding site are functionally linked (33,34,42). Thus, while the primary defect in L394R is glutamate substrate binding, linkage of this site to the other substrate sites results in a complex, multifaceted effect on enzyme activity.
There are several features of this mutation that make it interesting. First, the mutated residue is one of 20 contiguous amino acids that are identical in human (19), bovine (20), rat (21), Drosophila (22), mouse, whale, toadfish, and chicken (23). Second, the mutation is a drastic one, a hydrophobic residue to a charged residue. This almost certainly means that the mutation is a surface residue. If it were not, we would expect a very unstable protein with little activity. However, mutant L394R is stable and, under appropriate conditions, has substantial activity. Surface hydrophobic amino acids are often found to be important for protein-protein or protein-substrate interactions (43); therefore, it is likely that leucine 394 is a surface residue that contributes to substrate binding. Thus, although not previously implicated in glutamate binding (44), we postulate that this region of the carboxylase is a functionally important part of the glutamate-binding site.
The most interesting question, then, is how to relate our observations on the purified carboxylases to the symptoms of the patients carrying the mutation L394R. The original reports (16,18) demonstrated that vitamin K administration partially ameliorated the patients' bleeding problems. Therefore, we originally hypothesized that the defect in L394R is in the vitamin K interaction. As described here, however, reduced interactions with vitamin K are apparently not the principal defect. As mentioned above, it has been shown that the propeptide of the vitamin K-dependent proteins provides the primary binding site for the enzyme-substrate interaction (35,40). We have also shown that patients who synthesize carboxylationdeficient factor IX, when the vitamin K concentration is reduced by warfarin therapy, have a factor IX propeptide with reduced affinity for the ␥-glutamyl carboxylase (45). Furthermore, carboxylation appears to occur processively and, during a single binding event, modifies all of the glutamic acids that will be carboxylated (41). Thus, anything that reduces the affinity of the substrate for the carboxylase will result in a shorter residence time of the substrate on the carboxylase, resulting in its more rapid release and leading, potentially, to partially carboxylated, inactive products. Similarly, anything that decreases the rate of CO 2 incorporation may result in the normal residence time being insufficient for complete carboxylation before the product is released. Thus, any mutation that reduces the affinity of the substrate for ␥-glutamyl carboxylase or reduces the rate of CO 2 incorporation has the potential to result in undercarboxylated products.
To understand how the effects listed above might explain the therapeutic efficacy of vitamin K, we examined the effect of vitamin K on the rate of carboxylation at a low FIXproGla concentration, near the K m of the FIXproGla substrate for the wild-type enzyme. It is often assumed that for maximum control of rates, substrates are present in vivo at concentrations near their K m . Results (Fig. 8) show that, in contrast to high FIXproGla concentrations, L394R has a lower V max than wild- type carboxylase at lower concentrations of substrate, but the activity is increased from a low level to approximately 33% of wild type by increasing vitamin K. Based on our results, we conclude that the effect of vitamin K is merely an increased rate of carboxylation, driving the processive carboxylation of the vitamin K-dependent protein toward completion. After all, vitamin K is the only substrate whose in vivo concentration can be clinically manipulated. It should also be noted that vitamin K only partially corrects this deficiency, and, consistent with results shown in Fig. 8, the coagulation activity of these individuals remains below the normal range after vitamin K treatment. (18) Thus, the defect can be explained by a reduced rate of carboxylation, which is a result of a defective glutamate binding site and increased K m for vitamin K. This effect would be exacerbated by the apparent decreased affinity of the carboxylase for the propeptide, resulting in a reduced residence time for the substrate on the carboxylase. The combination of these two effects would result in partially carboxylated vitamin Kdependent proteins.
In summary, we have presented a plausible scenario to explain the amelioration of the patients' symptoms in response to vitamin K therapy. Our data show that the L394R mutant ␥-glutamyl carboxylase has an impaired glutamate-binding site that is more pronounced in the absence of the propeptide. We also note that, to a lesser degree, the mutation affects the propeptide and KH 2 binding properties of the enzyme. These apparent defects in binding propeptide and KH 2 probably arise because of linkage between different binding sites on the carboxylase.