A Conserved Region of Human Vitamin K-dependent Carboxylase between Residues 393 and 404 Is Important for Its Interaction with the Glutamate Substrate*

Certain individuals with combined deficiencies of vitamin K-dependent proteins have a mutation, L394R, in their (cid:1) -glutamyl carboxylase causing impaired glutamate binding. The sequence surrounding Leu 394 is similar in all known carboxylases, suggesting that the region is functionally important. To test this hypothesis we made the following mutant enzymes: W390A, Y395A, S398A, W399A, and H404A. We purified the enzymes and corrected the activity measurements for active enzyme concentration. Carboxylases W390A, S398A, and H404A had activities similar to that of wild type; however, Y395A and W399A had lower activities than did wild type. In the following descriptions we include our previously reported results for L394R. Kinetic studies with the substrate FLEEL, revealed K m values of 0.5 (wild type), 6.5 (L394R), 15 (Y395A), and 24 (W399A) m M . The k cat values relative to wild type were 51% (L394R), 1% (Y395A), and 2% (W399A). The k Five cells Active Enzyme Concentrations of Wild-type and Mutant Carboxylases Active enzyme concentrations of wild-type mutant carboxylases FLEEL Carboxylation Studies Carboxylase activities of the wild- type, W390A, L394R, Y395A, S398A, W399A, and H404A enzymes were initially determined by measuring 14 CO 2 incorporation into the peptide substrate FLEEL as Reactions Subsequent kinetic studies of FLEEL carboxylation the wild type, L394R, Y395A, and W399A enzymes by measuring 14 CO 2 incorporation at varying concentrations of FLEEL. KH 2 was used at 222 (cid:3) M except in W399A reactions, which contained 28 (cid:3) M KH 2 for reasons described In FLEEL carboxylation reactions, the free factor IX propeptide was also included: 5 (cid:3) M in wild type and L394R reactions, and 11 (cid:3) M and 28 (cid:3) M in Y395A and W399A reactions, re- spectively. Reactions were performed for 30 min with the wild-type and L394R enzymes. Due to their low activities, the reaction times were extended to 1 h with the Y395A and W399A enzymes. substrates (22, 25). We examined the effect of ProFIX19 on the carboxylation of FIXproGla for wild-type, L394R, and Y395A enzymes. We could not do this experiment with W399A because of its low activity and high K m for FIXproGla. We used 0.5 (cid:3) M FIXproGla in wild-type and Y395A reactions and 4 (cid:3) M in L394R reactions. Data were analyzed using a competitive inhibition model. L394R

Vitamin K-dependent carboxylase, or ␥-glutamyl carboxylase, is an integral membrane protein residing in the rough endoplasmic reticulum. It catalyzes the post-translational modification of specific glutamic acid residues to ␥-carboxyglutamic acid (Gla), 1 a modification necessary for the activity of certain proteins involved in blood coagulation (1), bone metabolism (2), and cell proliferation (3). Four other Gla protein sequences of unknown function have also been identified (4,5). In addition to the glutamate substrate, ␥-carboxylation requires carbon dioxide, oxygen, and KH 2 , the reduced form of vitamin K as co-substrates (6). Precursors of the physiological substrates of carboxylase contain a propeptide sequence, which anchors the substrate to the enzyme while multiple Glu residues are carboxylated. The conversion of glutamic acid to Gla is coupled with the oxygenation of KH 2 to vitamin K 2,3-epoxide and has been referred to as vitamin K epoxidase activity (7,8).
Two natural mutants of ␥-glutamyl carboxylase have been identified to date: L394R (9) and W501S (10). Individuals with either mutation have bleeding disorders, due to reduced activities of all clotting factors, which are presumably only partially carboxylated, since they have impaired calcium binding properties (11). In a subsequent characterization of recombinant L394R carboxylase we showed that the presumed undercarboxylation is primarily due to a defect in glutamate-substrate binding (12).
Comparison of the known carboxylase sequences among different species shows that leucine 394 is located in a highly conserved region in the molecule (Fig. 1), between residues 374 and 405 of the human sequence (13,14). The region between residues 393 and 404 is totally conserved not only among the known vertebrate carboxylase sequences but also in the known invertebrate sequences of Drosophila, Anopheles, 2 and Conus (13)(14)(15)(16).
In this study, we have expressed and examined the properties of mutants of some residues in this highly conserved region of carboxylase. Our results show that the primary defect in Y395A, W399A, and L394R is impaired binding of the glutamate substrate. In addition, residues Leu 394 and Trp 399 appear to be involved in linkage between the glutamate binding active site and the propeptide binding site of carboxylase.

EXPERIMENTAL PROCEDURES
Materials-FLEEL was purchased from Bachem (Philadelphia, PA). FIXproGla, a 59-residue recombinant peptide whose sequence is derived from the propeptide and the first 41 residues of the Gla domain of human factor IX, was prepared and purified as described (17). The peptide ProFIX19, was synthesized by Dr. Frank Church (University of North Carolina at Chapel Hill). The peptide Boc-mEEV was synthe-sized as described (18). Unlabeled and fluorescently labeled (with 5(6)carboxyfluorescein at the amino terminus) peptides of the human factor IX propeptide, ProFIX18, and the consensus sequence, pCon (19), were synthesized by Chiron Mimotopes (Clayton, Victoria, Australia). Vitamin K 1 was from Abbott Laboratories (Chicago, IL). NaH 14 CO 3 (specific activity, 54 mCi/mmol) was from ICN Pharmaceuticals (Costa Mesa, CA). L-␣-Phosphatidylcholine (Type X-E), CHAPS, and pepstatin were from Sigma. Leupeptin, aprotinin, and phenylmethylsulfonyl fluoride were from Roche Applied Science. The pSK Ϫ vector was from Stratagene (La Jolla, CA). The pVL1392 vector was from BD Pharmingen (San Diego, CA  (20), cloned into the pSK Ϫ vector, was modified by site-directed mutagenesis to construct the W390A, L394R, Y395A, S398A, W399A, and H404A mutants. All constructs contain the FLAG epitope (DYKDDDDK sequence) at their amino termini and the HPC4 tag containing the EDQVDPRLIDGK sequence (21) at their carboxyl termini. The constructs were subcloned into the pVL1392 vector, and the proteins expressed in baculovirus-infected High Five cells (22). The protein was purified using the HPC4 antibody affinity resin as described (12).
Determination of Active Enzyme Concentrations of Wild-type and Mutant Carboxylases-Active enzyme concentrations of the wild-type and mutant carboxylases were determined as described before by titrating the protein sample with fluorescent-labeled consensus propeptide (23).
Subsequent kinetic studies of FLEEL carboxylation were performed with the wild type, L394R, Y395A, and W399A enzymes by measuring 14 CO 2 incorporation at varying concentrations of FLEEL. KH 2 was used at 222 M except in W399A reactions, which contained 28 M KH 2 for reasons described below. In FLEEL carboxylation reactions, the free factor IX propeptide was also included: 5 M in wild type and L394R reactions, and 11 M and 28 M in Y395A and W399A reactions, respectively. Reactions were performed for 30 min with the wild-type and L394R enzymes. Due to their low activities, the reaction times were extended to 1 h with the Y395A and W399A enzymes.
Inhibition of FLEEL Carboxylation by Boc-mEEV-The tripeptide Boc-mEEV, a substrate analog in which the amino-terminal glutamic acid is modified to (2S,4S)-4-methylglutamic acid, is a strong competitive inhibitor of wild-type carboxylase for FLEEL carboxylation (18,24). We examined the effect of Boc-mEEV on the inhibition of FLEEL carboxylation, for the wild-type and L394R, Y395A, and W399A mutant enzymes. FLEEL concentrations of 0.8, 10, 12, and 16 mM were used in the wild-type, L394R, Y395A, and W399A reactions, respectively. These studies were performed in the presence of the factor IX propeptide. Data were analyzed using a competitive inhibition model. Effect of Free ProFIX19 Concentration on FLEEL Carboxylation Activity-In these studies, the effect of varying concentrations of Pro-FIX19 on carboxylation of FLEEL was measured. The reactions were carried out as described above for FLEEL carboxylation, except that the FLEEL concentration was kept constant and the propeptide concentration was varied. FLEEL concentrations of 1, 12, 24, and 32 mM were used in wild-type, L394R, Y395A, and W399A reactions, respectively.
Inhibition of FIXproGla Carboxylation by Free Propeptide-The free propeptide of vitamin K-dependent proteins is a competitive inhibitor for carboxylation of propeptide-linked substrates (22,25). We examined the effect of ProFIX19 on the carboxylation of FIXproGla for wild-type, L394R, and Y395A enzymes. We could not do this experiment with W399A because of its low activity and high K m for FIXproGla. We used 0.5 M FIXproGla in wild-type and Y395A reactions and 4 M in L394R reactions. Data were analyzed using a competitive inhibition model.
Concentration Dependence of Vitamin K Hydroquinone on Carboxylation-The effect of varying concentrations of KH 2 on FIXproGla carboxylation for wild-type, L394R, Y395A, and W399A was examined. The reactions were performed at fixed FIXproGla concentrations of 1.2 M in wild-type and Y395A reactions, and at 10 and 20 M in L394R and W399A reactions, respectively. The reactions were performed for 30 min with wild type and L394R and for 1 h with Y395A and W399A mutants, because of their low activities.
Vitamin K Epoxidase Activity Assay-Vitamin K epoxide formation by wild-type, L394R, and Y395A enzymes during their FLEEL carboxylation reactions was measured. FLEEL carboxylations in the presence of free propeptide were performed as described above, except that unlabeled NaHCO 3 was used. FLEEL concentrations of 1, 12, and 28 mM were used in the wild-type, L394R, and Y395A reactions, respectively. Vitamin K epoxide formation was quantified as described (26).

RESULTS
Carboxylation of FLEEL-We created 6 mutations within the highly conserved region of the carboxylase molecule. Of these, three had activity similar to that of wild type: the activities were 774 Ϯ 39 pmol of 14 CO 2 /h/pmol enzyme for WT, 590 Ϯ 14 for W390A, 699 Ϯ 15 for S398A and 822 Ϯ 21 for H404A. In contrast, the activities of L394R, Y395A, and W399A were dramatically reduced and were too low to be measured except in the presence of propeptide; therefore, we did all further characterizations at saturating propeptide concentrations. The K m values of L394R, Y395A, and W399A for FLEEL were significantly higher, 12-, 27-, and 45-fold than wild type (Table I). Furthermore, their k cat values were reduced ( Table I).
The k cat values of Y395A and W399A, were only 1-2% of wild type. The catalytic efficiencies, k cat /K m , were 24-fold lower for L394R and Ͼ2000-fold lower for Y395A and W399A than WT (Table I). Substrate inhibition with FLEEL has been widely observed and occurs at about 9 mM with the wild-type enzyme. We observed substrate inhibition at Ͼ18 mM for L394R (12). However, consistent with their lower affinity for FLEEL, neither Y395A nor W399A exhibited substrate inhibition up to FLEEL concentrations of 30 and 50 mM, respectively.
Inhibition of FLEEL Carboxylation by Boc-mEEV-The tripeptide Boc-mEEV was previously shown to be an active sitespecific, competitive inhibitor of FLEEL carboxylation (18,24). The K i of Boc-mEEV for wild-type carboxylase was 0.013 Ϯ  Carboxylation of FIXproGla-The K m of substrates with a propeptide is 1000-fold less than substrates lacking the propeptide (17,27,28). FIXproGla, a peptide substrate based on the Gla domain of human factor IX covalently-linked to its propeptide, appears to mimic a physiological substrate of vitamin K-dependent carboxylase (17,25). While the k cat of Y395A for FIXproGla is less than 10% of that of wild-type, its K m is similar to that of the wild type (Table II). On the other hand, L394R and W399A both have higher K m values for FIXproGla than does wild type; 9-fold and 24-fold respectively. The k cat values for FIXproGla carboxylation by Y395A and W399A were only 7 and 11%, respectively, of that of wild type. In contrast, the k cat for L394R is 2.7-fold higher than that for wild type, which we discussed before (12). Compared with wild type, the k cat /K m values for FIXproGla carboxylation are lower in the mutants, the most prominent effect observed in W399A (Table II).
Affinity of Wild-type, L394R, Y395A, and W399A Carboxylases for ProFIX19 -The free propeptides of vitamin K-dependent coagulation proteins enhance the rate of carboxylation of small glutamate peptide substrates (29,30). Wild-type carboxylase shows significant activity toward FLEEL carboxylation even in the absence of propeptide. In the presence of saturating propeptide, the rate of carboxylation by wild-type enzyme increases 5-6-fold. Because of their low activity in the absence of propeptide, it was difficult to estimate the fold increase with propeptide for Y395A, W399A, and L394R, but the increases were about 8-, 11-, and Ͼ100-fold, respectively. From the dependence of FLEEL carboxylation activity on propeptide concentration (Fig. 3, A and B), the apparent affinities (K d ) of wild-type, L394R, Y395A, and W399A carboxylases for Pro-FIX19 were determined by assuming that the increase in activity is proportional to the occupancy of the propeptide binding site. The apparent K d values of L394R, Y395A, and W399A were 17-, 2-, and 23-fold higher than that of wild type respectively (Table III).

Inhibition of FIXproGla Carboxylation by Free
Propeptide-In a further comparison of the relative affinities of wild-type, L394R, Y395A, and W399A, we measured the K i values of the factor IX propeptide as it competitively inhibits carboxylation of FIXproGla (22,25) (Fig. 4 and Table III). The affinity of Y395A was similar to that of wild type. On the other hand, the L394R affinity was about 7-fold lower than that of wild type, and that of W399A could not be determined, due to its high K m for FIXproGla and its low activity for FIXproGla carboxylation. The relative affinities of wild type, L394R, and Y395A for the propeptide determined by this approach and the method examining the effect of propeptide on stimulation of FLEEL carboxylation discussed above were comparable (Table III).
Effect of Vitamin K Hydroquinone on Carboxylation-We measured the dependence of FIXproGla carboxylation on KH 2  concentration for wild type, L394R, Y395A, and W399A. Since L394R and W399A have higher K m values for FIXproGla than wild-type and Y395A (Table II), we used a higher FIXproGla concentration in L394R and W399A reactions. The K m values for KH 2 were: 7.0 Ϯ 1.2 M, 32.8 Ϯ 5.4 M, and 38.5 Ϯ 9.6 M for WT, L394R, and Y395A enzymes, respectively (Fig. 5, A and  B). The K m of W399A for KH 2 could not be determined accurately since higher KH 2 concentrations inhibited its activity (Fig. 5B); the K m was estimated to be about 10 M from the initial phase of the kinetic curve. Therefore, all kinetic studies of the W399A mutant were performed at 28 M KH 2 . The reason for onset of inhibition at a lower KH 2 concentration for W399A is not clear, but similar phenomena have been reported in some other carboxylase mutants (31). The rate of carboxylation measured at 28 M KH 2 for W399A is about 70% of the V max predicted from extrapolating the initial non-inhibitory phase of the kinetic curve.
Vitamin K Epoxide Formation-Rates of vitamin K epoxide (KO) formation were measured as 48.8 Ϯ 1.5, 27.6 Ϯ 0.4, and 9.8 Ϯ 0.1 pmol of KO/min/pmol of enzyme for wild type, L394R, and Y395A. The amount of KO formed in W399A reactions was too low to be quantitated precisely, due to the low KH 2 concentration (28 M) used in W399A reactions (see above). As was observed with their reduced rates of carboxylation, L394R and Y395A show reduced rates of epoxide formation relative to the wild-type enzyme, suggesting that these mutant carboxylase and epoxidase activities remain coupled (8,32). DISCUSSION The objective of this study was to investigate the hypothesis we suggested in our previous work (12), that the region around amino acid 394 of human vitamin K-dependent carboxylase is important for substrate binding. This mutation was found in the carboxylase from an individual with combined deficiencies of vitamin K-dependent proteins (9,11). Leucine 394 is located in a highly conserved region spanning residues 374 -418 of the human carboxylase. In this region of homology, identical sequences are present in human (20), bovine (33), mouse, rat (34), and beluga whale (14) while 91% homology is observed with the sequence of toadfish (14). In addition to the species listed above, there are presently three known invertebrate carboxylase sequences: Drosophila (13,15), Conus (14,16) and Anopheles. When the sequences of these invertebrate species are added to the alignment, a smaller region, 390 -404, of high homology (only an asparagine changed to glutamine at 392 in cone snail) is observed. This extreme sequence conservation implies a structural and/or functional role for this region of the carboxylase molecule.
To identify other potentially important residues in this region, we expressed and purified mutants W390A, Y395A, S398A, W399A, and H404A, and examined their functional properties. We recently developed a method for measuring the active fraction of the carboxylase (23), so we also present here the data for wild type and L394R corrected for active site concentrations (12). Three of these five mutant enzymes, W390A, S398A, and H404A, exhibited normal activity. To date,  only certain cysteine residues have been identified as part of the carboxylase active site (35,36). In addition to cysteines, the participation of other amino acids in catalysis by a multisubstrate enzyme, such as carboxylase, seems likely. We had hypothesized that tryptophan, serine, and histidine were possible candidates for catalytic or substrate binding residues. Indeed Trp 399 does appear to be important, but since changing these other three residues has no obvious effect on activity or stability, it is unlikely that they are important either for catalytic function or for structural integrity. If important for maintaining the enzyme structure, one would predict that changing a bulky residue like tryptophan to alanine would create a cavity making the enzyme less stable (37). Furthermore, the serine to alanine mutation would remove a potential hydrogen bond to devastating effect if structural, but the mutation has no effect on stability or activity. Finally, the conserved histidine at position 404 in this region of substrate binding, a possible candidate for an active site residue, has no effect on the enzyme activity.
On the other hand, the changes of Y395A and W399A, both affected carboxylase activity. Like L394R, which we characterized earlier, these two mutants' primary defect appears to be substrate binding, specifically the region containing the glu residues. L394R, Y395A, and W399A all have high K m values for FLEEL relative to that of wild type. However, K m may or may not be a good estimate for K S (substrate affinity). In the simplest case, if k cat is small relative to the rate constant for substrate dissociation (k Ϫ1 ), then K m is similar to K S . So how do we determine if k cat is small relative to k Ϫ1 ? If k cat is large relative to k Ϫ1 then one can show that k cat /K m is an estimate of the substrate association rate constant (k 1 ). One would expect k 1 for a small molecule to be greater than 10 5 M Ϫ1 s Ϫ1 , however, for FLEEL carboxylation, k cat /K m is about 4 ϫ 10 2 M Ϫ1 s Ϫ1 for the wild-type enzyme and even lower for L394R, Y395A, and W399A mutants. Therefore, k cat /K m is probably not close to k 1 . We conclude that for FLEEL, k cat is small relative to k -1 , and therefore, K m is a reasonable estimate for K S . The weak interactions of L394R, Y395A, and W399A with the glutamate substrate are also reflected in the k cat /K m values for FLEEL carboxylation (Table I) which are 24-fold lower in L394R and Ͼ2,000-fold lower in Y395A and W399A, compared with wild type.
The most convincing evidence that this region of the carboxylase is part of a substrate-binding site is provided by inhibition studies of FLEEL carboxylation by the substrate analog, Boc-mEEV. This active site-specific, competitive inhibitor of FLEEL carboxylation (18,24,38) had K i values for L394R, Y395A and W399A 110-to Ͼ400-fold higher than for the wildtype enzyme. Therefore, we conclude that the primary effect of these mutations, as with L394R, is on substrate binding.
The effect on substrate binding may involve the carboxylatable glu residues and/or other substrate residues involved in binding to carboxylase. The Leu 394 , Tyr 395 , and Trp 399 residues that play a role in substrate carboxylation, show hydrophobic properties: the aromatic structure of tyrosine provides a certain degree of hydrophobicity although its hydroxyl group has polar characteristics. This observation may have some significance in terms of understanding interactions between residues on the enzyme and the substrate during binding events of the carboxylatable glutamates at the active site of carboxylase. Direct interactions between these hydrophobic residues on carboxylase with substrate glutamates are unlikely, but they might contact the methylene groups on the glutamates. Alternatively, they may interact with hydrophobic amino acids on the substrate that flank the carboxylatable glutamates. In this regard it is interesting that peptides without a bulky hydrophobic group preceding the glutamate residue are less active substrates than those that do have such a group (39).
In parallel with their severely reduced activities for carboxylation of FLEEL, Y395A, and W399A, have low k cat values for carboxylation of the propeptide-linked substrate FIXproGla (Table II). It should be noted that this is in contrast to L394R, which has a high k cat for FIXproGla; moreover, while W399A and L394R have higher K m values than wild type, the K m of Y395A is similar to that of wild type. These results are harder to interpret than those with FLEEL, since the kinetic constants for FIXproGla are dependent on both the interaction of substrate at the Glu binding site and at the propeptide site. In addition there are multiple carboxylations (up to 12) for every binding event. It appears that, for wild-type enzyme, the high affinity of the propeptide for the enzyme results in the rate limiting step being product release (40). A plausible explanation of this is that the dissociation rate of a propeptide-containing substrate is slow relative to its rate of carboxylation. FLEEL, on the other hand, is carboxylated and released with a relatively fast dissociation rate, so carboxylation is probably the rate-limiting step. We suggested in our L394R report that that increased V max of the enzyme was due to its low affinity for FIXproGla, which, in turn, is the result of the low affinity of the enzyme, relative to wild type, for the propeptide. These effects cause a faster release of product with L394R, and thus, a higher V max . Like L394R, W399A has low affinity for the propeptide, but its catalytic efficiency is low so the overall effect on V max is negative. Y395A also has low catalytic efficiency, but unlike the other two mutants, its affinity for propeptide is similar to that of wild type. So for Y395A carboxylation is still the rate-limiting step, and V max is reduced.
To further illustrate this point we offer a comparison of the k cat values for FLEEL and FIXproGla (Table IV). The k cat of the wild-type enzyme for FIXproGla is only 11% of that for FLEEL. In contrast to the wild-type enzyme, Y395A, which has an affinity for the propeptide similar to that of wild type, has a k cat for FIXproGla 88% of that for FLEEL. This supports our interpretation that, for Y395A, which has only 1% of the wild-type enzyme's activity for FLEEL, the glutamate carboxylation step, rather than the product (propeptide) release step, is rate-limiting.
L394R and W399A have reduced affinities for the factor IX propeptide. As we suggested in our report on L394R, we believe this is because of linkage between the glutamate and propeptide binding sites. These mutations do not disrupt linkage, since all the enzymes respond to increasing concentrations of propeptide with increased activity toward FLEEL. Perhaps the most important observation in this regard is that Y395A binds normally to the propeptide, while it, like L394R and W399A, has reduced glutamate substrate binding. This result provides further evidence that this region of the carboxylase is primarily involved in the interaction with the Gla domain of the substrate.
We and others have shown that vitamin K inhibits the carboxylase at high concentrations relative to K m for that substrate (8). In addition, certain mutant enzymes are inhibited at lower vitamin K concentrations (31). W399A is inhibited at lower vitamin K concentrations than either wild type or the other mutants we have studied. We do not know why this is true, but we postulate that one explanation is an ordered addition of substrates of the enzyme. The results from a previous study point toward a mechanism with the ordered addition of substrates to carboxylase: the exact order of binding, however, could not be clearly shown (41). In addition, the finding that no significant epoxide formation occurs in the absence of the glutamate substrate (42), suggests it is possible that the glutamate must bind to carboxylase before vitamin K for normal function. This order may be reversed at high concentrations of vitamin K, inhibiting the reaction. However, if enzyme affinity for the glutamate substrate is lower than that of wild type, but the affinity for vitamin K is normal, as is the case with W399A (assuming K m is a measure of affinity), then inhibition may set in at a lower vitamin K concentration. The L394R and Y395A mutants, on the other hand, which have higher affinities for glutamate than W399A, but appear to have lower affinities for vitamin K, do not show inhibition at low vitamin K concentrations. Confirmation of this hypothesis needs further study, which is beyond the scope of this current work.
Others have reported various regions/residues in the carboxylase molecule to be involved in interactions with the glutamate substrate: a region in the amino terminus (43), regions including the amino and carboxyl domains (44), residues Lys 218 , Arg 359 , and His 360 (31,45), Cys 343 and Tyr 345 (46), Cys 99 and Cys 450 (36). Of these residues proposed as functional in substrate binding or activity, four, Cys 343 , Tyr 345 , Arg 359 , and His 360 are predicted to be located in the cytoplasm according to the currently available topological study of the vitamin K-dependent carboxylase (47). On the other hand, lysine 218, and cysteines 99 and 450 are predicted to be in the lumen of the endoplasmic reticulum where catalytic activity is expected to reside (48). Since very little is known about the tertiary structure of the carboxylase, the proximity of these regions to the region focused in our study is unknown. However, this conserved region we have studied is located in the lumen of the endoplasmic reticulum, as would be expected for the substrate binding site(s).
In conclusion, our study has demonstrated that the conserved sequence between residues 393-404 of human vitamin K-dependent carboxylase defines an important region for interactions with the glutamate substrate. In addition to playing a major role in the binding of the glutamate substrate, some residues in this region, but not all, appear to be involved in linkage between the glutamate substrate binding site and propeptide binding site of the enzyme.