Role of Lysine 173 in Heparin Binding to Heparin Cofactor 11"

Heparin cofactor I1 (HC) is a plasma serine proteinase inhibitor (serpin) that inhibits a-thrombin in a reaction that is dramatically enhanced by heparin and other glycosaminoglycans/polyanions. We investigated the glycosaminoglycan binding site in HC by: (i) chemical modification with pyridoxal 5"phosphate (PLP) in the absence and presence of heparin and dermatan sulfate; (ii) molecular modeling; and (iii) site-directed oligonucleotide mutagenesis. Four lysyl residues (173, 252, 343, and 348) were protected from modification by heparin and to a lesser extent by dermatan sulfate. Heparin-protected PLPHC retained both heparin cofactor and dermatan sulfate cofactor activity while dermatan sulfate-protected PLPHC retained some dermatan sulfate cofactor activity and little heparin cofactor activity. Molecular modeling studies revealed that L Y S ' ~ ~ and L Y S ~ ' ~ are within a region previously shown to contain residues involved in glycosaminoglycan binding. Lys343 and Lys348 are distant from this region, but protection by heparin and dermatan

Heparin cofactor I1 (HC) is a plasma serine proteinase inhibitor (serpin) that inhibits a-thrombin in a reaction that is dramatically enhanced by heparin and other glycosaminoglycans/polyanions. We investigated the glycosaminoglycan binding site in HC by: (i) chemical modification with pyridoxal 5"phosphate (PLP) in the absence and presence of heparin and dermatan sulfate; (ii) molecular modeling; and (iii) site-directed oligonucleotide mutagenesis. Four lysyl residues (173, 252, 343, and 348) were protected from modification by heparin and to a lesser extent by dermatan sulfate. Heparin-protected PLPHC retained both heparin cofactor and dermatan sulfate cofactor activity while dermatan sulfate-protected PLPHC retained some dermatan sulfate cofactor activity and little heparin cofactor activity. Molecular modeling studies revealed that L Y S '~~ and L Y S~'~ are within a region previously shown to contain residues involved in glycosaminoglycan binding. Lys343 and Lys348 are distant from this region, but protection by heparin and dermatan sulfate might result from a conformational change following glycosaminoglycan binding to the inhibitor. Site-directed mutagenesis of L Y S '~~ and Lys343 was performed to further dissect the role of these two regions during HC-heparin and HC-dermatan sulfate interactions. The Lys343 + Asn or Thr mutants had normal or only slightly reduced heparin or dermatan sulfate cofactor activity and eluted from heparin-Sepharose at the same ionic strength as native recombinant HC. However, the L Y S '~~ + Gln or Leu mutants had greatly reduced heparin cofactor activity and eluted from heparin-Sepharose at a significantly lower ionic strength than native recombinant HC but retained normal dermatan sulfate cofactor activity. Our results demonstrate that L Y S '~~ is involved in the interaction of HC with heparin but not with dermatan sulfate, whereas Lys343 is not critical for HC binding to either glycosaminoglycan. These data provide further evidence for the determinants required for glycosaminoglycan binding to HC. Heparin cofactor I1 (HC)l and antithrombin I11 (AT) are members of the family of serine proteinase inhibitors (serpins) in human plasma. Both proteins inhibit proteinases by forming a covalent complex with the proteinase active site in a 1:l molar ratio (1)(2)(3). Although AT inhibits most of the proteinases involved in coagulation, HC exerts its anticoagulant effect by specifically inhibiting thrombin (2, 4).
Heparin is a highly negatively charged glycosaminoglycan composed of alternating residues of glucosamine and uronic acid (5). Heparin increases by several orders of magnitude the i n vitro and ex vivo rates of thrombin inhibition by both AT and HC (1,6,7). Dermatan sulfate is another glycosaminoglycan, consisting of alternating galactosamine and uronic acid residues, which acts specifically on HC inhibition of thrombin and does not accelerate the inhibition of any coagulation proteinase by AT (4). There is extensive evidence that the effect of heparin and dermatan sulfate is mediated through the formation of a ternary complex with both the inhibitor and thrombin binding to the same glycosaminoglycan molecule (8)(9)(10)(11).
Investigations of natural mutations as well as chemical modification studies have identified two regions in AT, from Gly35 to Glusn and LYS"'~ to Lys'", that are important for heparin binding. Included in these regions are several essential lysyl and arginyl residues which are thought to interact with the negatively charged glycosaminoglycans . Molecular modeling studies, using the crystallographic data from the related serpin a,-proteinase inhibitor (al-PI; 29), have shown that these two regions of AT are also close in the tertiary structure of the molecule and may together form the heparin binding site in AT (26).
HC shows little homology to the G1y3' to Glu5" region of AT, but extensive homology to the Lyslo7 to L y P region, which extends from L Y S '~~ t o Phelgs in HC. This region has been suggested as the putative glycosaminoglycan binding site in HC (33)(34)(35). Lysyl and arginyl residues have also been shown to be essential for glycosaminoglycan binding by HC (30,34,36) and specific residues in this region have been shown to be important for these interactions (34,35,37). In this study we used pyridoxal 5'-phosphate (PLP) to selectively modify lysines in the absence and presence of heparin and dermatan sulfate in order to discover lysines in HC possibly involved in binding to these glycosaminoglycans. We evaluated these results and studied the structure of HC by computer-assisted molecular modeling. We then used site- directed mutagenesis to further assess the involvement of L Y S '~~ and Lys343 in glycosaminoglycan binding.

RESULTS
Identification of Lysyl Residues in HC Protected by Heparin and Dermatan Sulfate from Phosphopyridoxylatwn-In agreement with our previous observations (30), when HC was modified in the absence of glycosaminoglycan, an average of 2.7 mol of PLP were incorporated per mol of HC. In the presence of 1 mg/ml heparin or dermatan sulfate an average of 1.9 or 2.5 mol of PLP/mol of HC were incorporated, respectively. Increasing the dermatan sulfate concentration tenfold did not increase the protection (data not shown). Tryptic maps were next produced in order to identify which lysines had been modified. The tryptic peptide map of unprotected PLPHC showed eight major peaks at 325 nm where PLP absorbs (Fig. 1, panel A ) . These eight peaks yielded six peptides, the sequences of which are shown in Table I. When rechromatographed, peaks 2-4 in Fig. 1, panel A ,  other peaks gave one 325-nm peak when rechromatographed.
The tryptic maps of heparin-protected PLPHC and dermatan sulfate-protected PLPHC (Fig. 1, panels B and C, respectively) show that peaks 1, 5, 7, and 8 were reduced in the heparin-protected sample and to a lesser extent in the dermatan sulfate-protected sample. These results also show that L y P and Lys303 were modified to essentially the same extent in all three PLPHC samples. The extent of protection was determined (Table 11). In heparin-protected PLPHC both LYS'?~ and LysZ5' showed only about 20% of the modification that they showed in the unprotected sample, whereas Lys3a3 and Lys348 showed 31 and 57% modification, respectively. The protection was not as extensive in dermatan sulfate-protected PLPHC. L Y S '~~ and LysZ5' showed 81 and 57% modification, respectively, compared with the unprotected sample, whereas Lys343 and Lys348 showed 52 and 74% modification, respectively, compared with unprotected PLPHC. Interaction of Modified HC with Heparin and Dermatan Sulfate-The ability of heparin and dermatan sulfate to accelerate thrombin inhibition by the modified HC species was studied. Control experiments showed that all samples had normal antithrombin activity (inhibition of thrombin by HC in the absence of glycosaminoglycans) consistent with previous results (Ref. 30; data not shown). Fig. 2 shows thrombin inhibition by the various modified HC species as a function of heparin and dermatan sulfate concentration.
Over a range of 0.1-120 pg/ml heparin the reduced control HC sample (treated with sodium borohydride but not PLP) showed a normal pattern of increasing and then decreasing inhibition, with a maximum inhibition at 10 pg/ml heparin ( Fig. 2 A ) . The modified inhibitors also showed this pattern, although none retained the full activity of the reduced control HC. Also, the maximum inhibition was shifted to 40 pg/ml heparin, probably due to the repulsion of the heparin by the negatively charged PLP moiety. The unprotected PLPHC retained -20% heparin cofactor activity, which is consistent with similar chemical modification studies with AT (21).
Heparin-protected PLPHC and dermatan sulfate-protected PLPHC retained -66 and -23% heparin cofactor activity, respectively. Over a dermatan sulfate range of 1-1000 pg/ml, the reduced control HC sample again showed a normal pattern of increasing and then decreasing inhibition, with a maximum inhibition at 250 pg/ml dermatan sulfate (Fig. 2 B ) . Unprotected PLPHC retained -20% dermatan sulfate cofactor activity and dermatan sulfate-protected PLPHC retained -33%, relative to the reduced control HC sample. However, heparin-protected PLPHC retained 91% dermatan sulfate cofactor activity, although twice as much dermatan sulfate was necessary to reach maximum inhibition. Thus heparin (with a greater extent of lysine protection) was able to protect both heparin and dermatan sulfate cofactor activity in HC, whereas dermatan sulfate (with a lesser extent of protection) could only minimally protect dermatan sulfate cofactor activity.
Molecular Modeling of HC-We utilized computer-assisted molecular modeling in order to visualize and further compare the putative glycosaminoglycan binding regions of AT and HC. As has been done with AT (26), we modeled HC based on its serpin homologue al-PI. Shown in Fig. 3 are the 4 lysyl residues protected in the presence of heparin and dermatan sulfate from modification by PLP. Also shown are several residues in the LyP5 to the Phe'" region of HC known to be involved in glycosaminoglycan binding. This region includes L Y S '~~ and is homologous with the proposed heparin binding region in AT. Lys2'* is nearby with respect to tertiary structure. These residues form a region of positive charge on the surface of the molecule, which could interact with glycosaminoglycans and other polyanions. Lys"' and Lys348 are not near the putative binding site, but are in a region homologous to a region in AT in which lysines were protected by heparin, but not found to be essential for heparin binding (23). These residues may be protected by a conformational change upon glycosaminoglycan binding or by nonspecific interactions of these residues with the glycosaminoglycans.
Expression of Recombinant HC-Mutated recombinant HC (rHC) molecules were created by two distinct mutagenesis techniques resulting in a Leu or Gln substitution for L Y S '~~ and Thr or Asn substitution for Lys343. Each of the mutations Shown in black are the four lysyl residues protected by heparin and dermatan sulfate from modification by PLP and residues known to be involved in glycosaminoglycan binding to HC. eliminated the positive charge on the side chain of the amino acid. The final construct was expressed in Escherichia coli as product of the vector pMON-HCII. Besides the absence of the signal peptide, the protein product lacked the NHp-terminal 18 amino acid residues and the post-translational modifications of plasma HC and has been characterized previously (34, 35). The rHC variants appeared to be identical in size to native rHC and had an apparent molecular weight of 55,000 determined by immunoblot analysis of concentrated lysate subjected to sodium dodecyl sulfate-polyacrylamide gel electrophoresis (not shown).
Binding of rHC to Heparin Sephurose-Native rHC and each of the variants were subjected to heparin-Sepharose chromatography to determine their relative affinities for heparin. Native rHC and both of the Lys343 substitutions were eluted from the column with a peak at -0.35 M NaCl suggesting that the positive charge of this residue is not necessary for the interaction with heparin (Fig. 4A). With native rHC, -10% of the immunoreactive protein was detected in the column flow-through, but -25-30% was identified in the flow through of the Lys343 --., Thr or Asn variants (data not shown). Under the same conditions, we have previously demonstrated that the column capacity had not been exceeded and that the flow through may represent denatured rHC (35). In contrast, L Y S '~~ + Leu or Gln resulted in protein products that were eluted from heparin-Sepharose at a lower ionic strength (Fig.  4B). The peak of native rHC eluted at -0.38 M NaCl, whereas rHC ( L Y s '~~ + Gln) and rHC ( L Y s '~~ + Leu) eluted at -0.26 and 0.15 M NaC1, respectively. For both of these rHC variants, a small fraction of immunoreactive protein (110%) was detected in the column flow-through.
Glycosaminoglycan-independent Thrombin Inhibition by rHC Variants-Thrombin inhibition was measured in the absence of glycosaminoglycan for native rHC and each of the rHC variants (Table 111), as described previously (47). Native rHC had a second-order rate constant of 1.1-1.6 X 10' M" min" which is -4-fold lower than previously reported by Tollefsen et al. (7). Using the same truncated form of native rHC, Blinder and Tollefsen  The mutation Lys343 + Thr produced only modest changes in the second-order rate constant providing evidence that the reactive site of this mutant was intact. In contrast, the Lys343 + Asn substitution appeared to decrease the second-order rate constant -7-fold, suggesting that this molecule had an altered ability to inhibit proteinases in the absence of glycosaminoglycans. The L Y S '~~ + Leu and L Y S '~~ + Gln substitutions resulted in proteins with very similar rate constants compared with native rHC and therefore retained their ability to inhibit thrombin.
Heparin-and Dermatan Sulfate-dependent Thrombin Inhibition by rHC Variants-The effects of heparin and dermatan sulfate on the thrombin inhibition by rHC and rHC variants are shown in Fig. 5. In these experiments, native rHC inhibited 50% of the thrombin with -3-7 pg/ml of heparin and -4-10 pg/ml of dermatan sulfate. The Lys343 + Thr variant required similar heparin concentrations and only 2-%fold more dermatan sulfate than the native rHC to inhibit 50% of the thrombin, suggesting that the charge at Lys343 is probably not critical for glycosaminoglycan-dependent inhibition. However, the Lys343 + Asn variant required -10-fold higher concentrations of each glycosaminoglycan for 50% inhibition, which may be accounted for by the decreased rate of thrombin inhibition by this variant in the absence of glycosaminoglycan (Table 111).
In contrast, the rHC molecules containing substitutions of LYS'?~ required at least a 30-fold higher concentration of heparin for 50% thrombin inhibition. With dermatan sulfate, however, the L Y S '~~ + Leu variant showed normal thrombin inhibition and the L Y S '~~ + Gln variant required only a 2fold higher concentration for 50% inhibition. These observations suggest that the charge a t L Y S '~~ is required for heparinbut not dermatan sulfate-dependent thrombin inhibition.

DISCUSSION
The glycosaminoglycans heparin and dermatan sulfate exert their anticoagulant effect in human plasma by increasing the inhibition rate of proteinases by the serpins HC and A T by as much as 1000-fold (1, 6, 7). In both cases, binding of the glycosaminoglycan to the inhibitor is necessary for activity, and this binding is thought to occur by ionic interactions between the negatively charged glycosaminoglycan and positively charged amino acids of the inhibitor. Chemical modification, natural mutation, and site-directed mutagenesis of these basic amino acid residues result in inhibitors with decreased binding to glycosaminoglycans . The binding site of AT is very specific for a pentasaccharide sequence in heparin, and this sequence contains a unique 3-0-sulfated glucosamine at the third position. Glycosaminoglycans and other polyanions that do not contain these structures have almost no ability to accelerate proteinase inhibition by AT (48). HC does not require this specific pentasaccharide sequence for acceleration of thrombin inhibition; in fact, the HC/thrombin inhibition reaction can be accelerated by a wide range of polyanions (33,(49)(50)(51).
Even though they have different specificities, it is thought that the glycosaminoglycan binding sites in both HC and AT occur in similar regions. Several residues have been shown to be specifically involved in heparin binding to AT, and as a result two regions, Gly35 to Glu5' and Lyslo7 to LYS'~~, have been designated as forming the heparin binding domain of AT (16-19, 21, 23, 25, 27, 52). HC contains a region, L y P 5 to Phelg5, that shows extensive homology to the Lyslo7 to L Y S '~~ region of AT (Fig. 6). This region is densely populated with positively charged amino acids that could interact with various polyanions and is the putative glycosaminoglycan binding site in HC. Indeed, recent studies have shown the involvement of: (i) Lys'= and possibly Argla4 and Arglg3 in heparin binding and (ii) LyslS5, ArgIa9, Argl'', ArgIg3, and possibly Arg" in dermatan sulfate binding to HC (34,35,37).
In this study we have utilized several approaches to define further the determinants of the glycosaminoglycan binding domain in HC. First, we used chemical modification in the absence and presence of heparir. or dermatan sulfate to identify lysyl residues that might be involved in glycosaminoglycan binding. We identified 4 lysyl residues (173,252,343, and 348) in HC that we were protected from modification in the presence of glycosaminoglycan.
We then used computer-assisted molecular modeling to interpret the above results. L Y S '~~ was very well protected from chemical modification by heparin and less so by dermatan sulfate. L Y S '~~ is homologous to a residue in AT that has been found to be involved in heparin binding to that inhibitor (21). Our model shows it to be on the surface of the molecule and thus accessible for interaction with glycosaminoglycans. Amphipathic a-helices are thought to be a possible secondary structure through which proteins bind to glycosaminoglycans (33,53). LYS'?~ is located just before the start of such an a-helix (D-helix using the nomenclature of al-PI; Refs. 26 and 29) containing other residues in HC important in glycosaminoglycan binding. These results implicate L Y S '~~ in glycosaminoglycan binding to HC, especially in heparin interactions.
LysZ5' was well protected from chemical modification by heparin and fairly well by dermatan sulfate. LysZ5' is not within the putative glycosaminoglycan binding of site of HC, but from our model it is close to this region in the tertiary structure of the molecule. LYS'~', ArglS9, Arg'", ArgIg3, and possibly Arg'% are known to be involved in glycosaminoglycan binding in HC (34, 35, 37) and can be seen in Fig. 3 to form a surface of positive charge, near LysZs2, that could interact with the negatively charged glycosaminoglycans. Lys' ' ' may be involved in glycosaminoglycan binding, but it is also possible that interactions of other residues in the putative glycosaminoglycan binding site lead indirectly to protection of this lysyl residue from phosphopyridoxylation.
Lys343 and Lys348 are also not in the putative glycosaminoglycan binding site of HC. These residues were moderately well protected by both heparin and dermatan sulfate, with Lys343 being somewhat more protected by each glycosaminoglycan. In our model Lys343 and Lys348 are also shown to be on the surface of the HC molecule but removed from the other residues known to be involved in glycosaminoglycan binding. These residues are probably not involved in glycosaminoglycan binding and may be protected from chemical modification by a conformational change caused by the glycosaminoglycan binding to HC or by nonspecific ionic interactions between the glycosaminoglycan and HC.
Finally, we performed site-directed mutagenesis based on the information from the chemical modification and modeling studies. We chose L Y S '~~ as the most probable residue involved in glycosaminoglycan binding and we chose to mutate Lys343 in order to investigate the possible involvement of the region around this residue in glycosaminoglycan binding.
The L Y S '~~ variants both bound with less affinity to heparin-Sepharose than did the native rHC. These variant rHC species also showed greatly reduced heparin cofactor activity but essentially normal dermatan sulfate cofactor activity. This is in agreement with the chemical modification data for LYS"'~. Heparin was better able to protect L Y S '~~ from chemical modification than dermatan sulfate was, implying that this residue interacted more with heparin than with dermatan sulfate. Additionally, the heparin-protected PLPHC retained heparin cofactor activity, whereas the dermatan sulfate-protected PLPHC (with 60% less protection of L Y S '~~) retained almost no heparin cofactor activity. These results suggest that L Y S '~~ is involved in the binding of heparin to HC but is not critical for the binding of dermatan sulfate to HC. We believe the effects of chemical modification and mutation of L Y S '~~ on heparin binding are due to the loss of the positive charge of this residue and therefore the elimination of an electrostatic interaction between the side chain of L Y S '~~ and heparin. However, we cannot rule out the possibility that the effects on binding result from alteration of the tertiary structure of the glycosaminoglycan binding domain in HC.
Different results were obtained with the Lys343 variants. Both of these variants bound to heparin-Sepharose almost identically to native rHC. The heparin cofactor and dermatan sulfate cofactor activities of the Lys343 Thr variant were very similar to those of the native rHC. The Lys343 + Gln variant had less heparin and dermatan sulfate cofactor activity than rHC, but this can be accounted for by the decreased antithrombin activity of this variant. These results indicate that Lys343 is not critical for the binding of either glycosaminoglycan to HC.
A thorough understanding of the interaction of glycosaminoglycans with AT and HC is necessary because of the widespread use of heparin as an anticoagulant. Various residues in the LydGs to Phelg5 region of HC have been shown to effect both heparin and dermatan sulfate binding and the mutation of Arg"' results in deficient dermatan sulfate binding but normal heparin binding (34, 35, 37). In this study we showed that mutation of L Y S '~~ leads to deficient heparin binding but normal dermatan sulfate binding. Thus we can state that although heparin and dermatan sulfate bind to the same region of HC, they have specific residues on the protein with which they interact. Further and precise determination of the nature of these interactions between glycosaminoglycans and AT or HC may ultimately lead to the development of better and more specific anticoagulants. M NaCi in order to remove any remaining reagents Concentration of P1.P bound to IIC was prepared errentially as described (23). The PLPHC specin were denatured by dialysis overnight Regions lacking coordinates, such as the aminoterminus, insertions, or deletmns. were ignored in the simulation. The mutated protein was minimized using the force-fieldr in SYBYL/MENDYL. The drawings were made using the Tripos NITRO program on a Macintosh n.
Mulogenesis of HC cDNA-Single nucleotide mutagenesis of the Lys343 codon was performed in the M13mp18 containing the nonmding strand of HC by the method of Nakamaye and Ecktein as described previously 141). A Bru 36 I-Xho I fragment of the DNA containing the mutation was ligated into the expression vector pMON-HCII and the final plasmid c o n s t r~~l was confirmed by sequencing using the chain termination method of Sanger et nl 142). 143). In bnef, sequential SilP-directed mutagenesis was performed in Ml3mpl8 containing the HC Random mutations were generated in the Lys173 codon by the procedure of Wells el d cDNA to create a Sac I site at nucleotides 595-600 followed by the formation of an R5r II site in nucleotides 683-689 (numbered according to Blinder 11 a1 Fig 2; 391. The predicted amino acid sequence was not altered by the nucleotide substitutions A 660 bp Bsu 36 I-Xho I fragment of the cDNA which contained both new restriction sites was isolated from the replicative form and ligated into similarly digested pMON-HCII that had been digested with mung bean nuclease to eliminate I Soc I site in the 3' polylinker region of the HC insert (44). Proper ligation was verified by digestion with RSI II and Ssc 1 and the flnal plasmid ~onstruct was confirmed by didwxy requendng of both strands. Approximately 5 pg of the pMON-HCII plasmid was digested wlth both restriction enzymes and the vector product was isolated from the Small Sac I-Rrr II fragment by agarose gel electrophoresis and purified from the gel using "glass m i l k beads (Gene Clean. Biolnl). Five oligonucleotides were synthesized lo span the Sac I-Rsr I1 domain. TWO of the oligonucleotider, a 44mcr and a 4lmer. formed the coding strand and three complementary oligonudwtider (34, 28, and 30 nucleotides in length) comprised the non-coding strand. The sequence of the oligonucleotides varied from the HC cDNA by replacement of all three positions in the codon for ~~$ 1 7 3 with any nucleotide, allowing for generation of random changer at those positions. In addition, the codon for Arg184 was changed from CGT to CGG to create I Bsp El restriction site that was used for screening. Oligonucleotides were pooled at a final mncentralion of 10 pM in H z 0 and 2.5 pl of the pool was phosphorylated with T4 polynucleotide kinase.
Approximately 25 fmol of the phosphorylated oligonucleotide were mixed with -12 (mol of the digested DNA in a buffer containing 25 mM Trir-HCI. 10 mM MgCi2, 10 mM DIT, 4 mM ATP, 74 DNA ligase was added (1 Weinr unit) and the reaction was incubated at 14 OC for 16 hours pH 7.5 and annealed by heating to 68 OC for 5 min and cooled to rmm temperature over 2 hours.
Escherichia cola ]MI09 were transformed with a portion of the ligation mix, 24 colonies were isolated and amplified, and the plasmid was purified (451. A 448 bp DNA fragment spanning the oligonucleotide insert was amplified using the Taq polymerare chain reaction as described by the manufacturer and the fragments were digested with Bsp El to screen for the inserted oligonucleotides. Seven plasmids containing the ssp Ei site were sequenced between the Rsr II-Sac I sites in both orient st ion^ and of those, two clones were detected that contained only mutations in the codon far ~~$ 1 7 3 .

Critical Lysyl Residue of Heparin Cofactor I1
Erpresrion of Rrcombinant HC ond f l C v u r~n l s -For each recombinant HC preparsuon, transformed Erchrrtchm rol! JMlOl were grown to an optical density of -1 2 at 550 nrn in 500 ml of medium, and expression was induced wilh nalidixic acid according 10 Li el nf (46) The cells wcre washed, reruspended in phosphate-buffered salme. and lysed by son~calion The producl was rubpcled to ammonium sulfate precipitatron and the concentration of the rHC was determmed by a quant~lative immunoblot as dermibed previously (35). There was no a p p a~n l difference in the recovery oi the expressed rHC cornparfd lo Ihe variant rHC preparations.
Funclimsl Anofyrir of Recombinant HC and HC uorionlr-Binding of r l I C to hepann was determined using a 2-ml heparin-Sepharore column as described previously (35). Thrombin inhibition was determined in the absence of glycosaminoglycans by incubating 900 pi o i 40-168 nM rHC. 50 pl of 280 nM thrombin and 50 pl of H z 0 at room temperature At specific tunc$ 100 HI aliquots W P~F added lo 400 p1 of 100 pM Chrornozym TH and the rate of change of absorbance at 405 n m was monitored contmuously for 100 seconds The rate of change oi absorbance WBF piOporliOnd lo Ihe Concentration of active thrombin remaining in the incubation The eifect of glycosaminog!ycanr was determined using thrombin and rHC at the concentrations used above and adding various concentrations of heparin or dermatan sulfate to a final volume a i 100 pl Absorbance at 405 nm was determined after Ihe addition of Chromozym TH as described above.