Role of the H Helix in Heparin Binding to Protein C Inhibitor*

Protein C inhibitor (PCI) is a plasma serine proteinase inhibitor (serpin) that is a major physiological regulator of activated protein C. Inhibition of its target proteinase is accelerated by heparin in a reaction that involves the binding of both inhibitor and proteinase to heparin to form a ternary complex. This study was undertaken to understand the role of the H helix region (residues 264-278) of PC1 in heparin binding and used (i) a recombi- nant truncated PC1 fusion protein of the first 294 residues, (ii) H helix synthetic peptides containing single ArglLys + Glu substitutions, and (iii) site-directed Ala mutagenesis of 4 basic residues (Arg-269, Lys-270, Lys- 276, and Lys-277) in the H helix region of full-length recombinant PC1 (rPCI) expressed in Buculouirus. The PC1 fusion protein interfered in heparin-accelerated PCI-proteinase inhibition reactions, and it bound to heparin-Sepharose. Compared to the wild-type PC1 fu- sion protein, deletion of the H helix from the fusion protein resulted in a reduction of both heparin-Sepharose binding and the ability to compete for heparin during PCI-proteinase inhibition reactions. Competition assays with H helix synthetic peptides revealed that the R269E altered peptide was the least effective


PO. Box 85500, 3508 GA Utrecht, The Netherlands
Protein C inhibitor (PCI) is a plasma serine proteinase inhibitor (serpin) that is a major physiological regulator of activated protein C. Inhibition of its target proteinase is accelerated by heparin in a reaction that involves the binding of both inhibitor and proteinase to heparin to form a ternary complex. This study was undertaken to understand the role of the H helix region (residues 264-278) of PC1 in heparin binding and used (i) a recombinant truncated PC1 fusion protein of the first 294 residues, (ii) H helix synthetic peptides containing single ArglLys + Glu substitutions, and (iii) site-directed Ala mutagenesis of 4 basic residues (Arg-269, Lys-270, Lys-276, and Lys-277) in the H helix region of full-length recombinant PC1 (rPCI) expressed in Buculouirus. The PC1 fusion protein interfered in heparin-accelerated PCI-proteinase inhibition reactions, and it bound to heparin-Sepharose. Compared to the wild-type PC1 fusion protein, deletion of the H helix from the fusion protein resulted in a reduction of both heparin-Sepharose binding and the ability to compete for heparin during PCI-proteinase inhibition reactions. Competition assays with H helix synthetic peptides revealed that the R269E altered peptide was the least effective at blocking heparin-catalyzed PCI-proteinase inhibition reactions. Compared with full-length active wild-type rPCI, R269A K270A and K276A:K277A rPCI both had reduced heparin-Sepharose binding, but only R269AK270A rPCI showed a loss of heparin-accelerated proteinase inhibition for both activated protein C and thrombin. We conclude that a major heparin-binding site of PC1 is the H helix, unlike its heparin-binding serpin homologues antithrombin and heparin cofactor 11, which bind heparin primarily through the D helix.
Protein C inhibitor (PCI)' is a plasma glycoprotein that is a major physiological regulator of the anticoagulant proteinase activated protein C (APC) (for a review, see Refs. 1 and 2). In addition to activated protein C, PC1 can also inhibit thrombin and several other proteinases involved in coagulation, fibrincan Heart Association-Sanofi Winthrop (to F. C. C.), Research Grant * This work was supported in part by a grant-in-aid from the Ameri-HL-06350 from the National Institutes of Health (to F. C. C.), Grant 92.306 from the Netherlands Heart Foundation (to J. C. M. M.), and a fellowship from the Royal Netherlands Academy ofArts and Sciences (to J. C. M. M.), The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  The abbreviations used are: PCI, protein C inhibitor; rPCI, recombinant PCI; APC, activated protein C; serpin, serine proteinase inhibitor; PEG, polyethylene glycol; BSA, bovine serum albumin; MBP, maltose-binding protein.
The location of the heparin-binding site in PC1 is still unresolved. Heparin-binding proteins are known to contain clusters of basic amino acid residues that form centers of high positive charge density and that form electrostatic interactions with the acidic groups of heparin (19). Three regions in the primary structure of PC1 contain clusters of basic amino acids: residues 1-11, 82-90, and 266-278. Based on PC1 molecular modeling studies, Kuhn et al. (20) proposed that a large positively charged heparin-binding surface is formed by the juxtaposition of the first 15 amino-terminal amino acids (termed the A+ helix)' with the H helix region (residues 264-278).' A role for the A+ helix in heparin binding was supported by demonstrating that an anti-PC1 antibody mapped to the A+ helix blocked heparin-Sepharose binding (20). We provided evidence that the H helix in PC1 binds heparin by showing that a H helix synthetic peptide bound to immobilized heparin and interfered with heparin-catalyzed serpin-proteinase inhibition reactions (12). In contrast, an A+ helix synthetic peptide and a peptide containing PC1 residues 82-90 were less effective a t blocking heparin-catalyzed PCI-proteinase inhibition reactions (12 ylene polymethobromide), and bovine serum albumin (BSA) were from Sigma. The following proteinase chromogenic substrates were used: tosyl-Gly-Pro-Arg-p-nitroanilide (Chromozym TH; Boehringer Mannheim) for thrombin and Lys-benzyloxycarbonyl Pro-Arg-p-nitroanilide (Spectrozyme PCa; American Diagnostica Inc.) for APC. Human a-thrombin was purified as described previously (21), and human APC was purchased from Haematologic Technologies Inc. Human plasma PC1 was purified as described (22) with modifications involving affinity chromatography on a n anti-PC1 monoclonal antibody column instead of DEAE-Sepharose fractionation. Human plasma antithrombin was purified as described previously (23). Unfractionated heparin was obtained from Diosynth (Oss, The Netherlands). Heparin-Sepharose was purchased from Pharmacia Biotech Inc.
Production of Duncated PCI Fusion Protein-PC1 cDNA was obtained by first-strand cDNA synthesis of HepG2 (a hepatoma cell line) total RNA using a random hexameric oligonucleotide primer, followed by polymerase chain reaction amplification with two PCI-specific oligonucleotides that omitted the signal sequence and engineered 5'-and 3'-NcoI restriction sites (5'-TCCTCCATGGCTCACCGCCACCACCCC-CGGGA-3' and 5'"ITITCCATGGGAGAAGCCCCACCTCAGGGGCG-3'). The PC1 cDNA was subcloned into phagemid pSL1190 to add 5'-EcoRI and 3'-BarnHI flanking sequences, which permitted insertion into M13mp19 for mutagenesis. The oligonucleotide-directed mutagenesis method of Kunkel et al. (24) was used. Briefly, uracil-rich singlestranded DNA was obtained by infection of E. coli strain BW313 with recombinant PCI/M13mpl9 phage. The purified DNA was used as a template for in vitro annealing and elongation of a mutagenic oligonucleotide (5'-GCTCCTATCAGT_GAGAGAAAGTCC-3') that incorporated a stop codon at position 987-989 (lo), corresponding to residue 295 in PCI. E. coli strain DH5aF' was transformed with the heteroduplex DNA. A mutant clone was selected and fully sequenced using the dideoxynucleotide chain termination method (33) (Sequenase version 2.0). This modified cDNA(termed PCI,,,) was excised from M13mp19 by NcoI digestion, treated with mung bean exonuclease to remove the sticky ends, and ligated into the XmnI site of pMAL-c2, a prokaryotic vector that expresses foreign DNA as a fusion protein with E. coli maltose-binding protein (MBP). E. coli strain TB1 was transformed by electroporation. A clone was selected that contained the PCI,,, insert in the correct orientation and reading frame and that was missing the amino-terminal codon for alanine that resulted from the engineered NcoI site (PC12,,/pMAL-2c) (see Fig. 1). PCIz9, helix deletion mutants were also generated by using the oligonucleotide-directed mutagenesis method of Kunkel et al. (24) on DNA cassettes removed from PCI2,,/pMAL-2c and inserted into M13mp19, a Sad-KpnI DNA fragment for the A+ helix deletion, and a KpnI-BanHI fragment for the H helix deletion. Coding sequence for the A+ helix (amino acids 1-15 of the mature protein) was deleted using a 30-mer composed of the 15 nucleotides 5' and 3' to the 45 bases being deleted (5'-GGGATCGAGGGAAGGCATGTAGGTGCCACG-3'). Coding sequence for PC1 amino acid residues 264-277 (H helix region) was deleted using a similarly designed oligonucleotide (5"GTGGAGAATGGACTGAG-GCAGCTGAGCTT -3'). Mutated cassettes were fully sequenced and subcloned back into the PCI2,,/pMAL-2c expression vector to replace the corresponding wild-type sequence. The AA+ helidAH helix double mutant was constructed by replacing the wild-type Sad-KpnI cassette in AH-PC12,,/pMAL-2c with the mutant AA+skPCI,,, cassette, For large-scale fusion protein production, the expression vector constructs were transformed into E. coli strain PR700. Fusion protein was purified from cell extracts by amylose resin affinity chromatography according to the manufacturer's procedure. The purified fusion proteins were dialyzed into 20 mM HEPES, 10 m~ NaCl, 0.1% PEG, 1 mM NaN,, pH 7.4, and their concentrations were determined by the dye binding assay of Bradford (32) using BSAas the standard. Immunoblots showed that the wild-type and AH helix fusion proteins were detected by a monoclonal antibody to the A+ helix of PCI, while the AA+ helix and PA+ helidAH helix fusion proteins were not (data not shown).
Baculovirus Expression of Full-length rPCI-Wild-type rPCI was generated and characterized as described previously (25)

5'-C'M'AAGATG'I"I'CGCAGCGAGGCAGCTCGAG-3'
was used to replace Lys-276 and Lys-277 with alanines (K276A:K277A rPCI). At the same time, the internal KpnI site from the vector polylinker was modified (made uncuttable) with oligonucleotide 5'-GGATCCCGGGTATATTCTGAATTCC-3' to facilitate cloning back into the PCYpVL1393 transfer vector construct. Mutant cassettes were sequenced and subcloned back into PCVpVL1393 to replace wild-type DNA, and the transfer vector was sequenced again to confirm the presence of the desired substitutions prior to homologous recombination. Generation of high titer recombinant viral stocks and infection of High-Fivem cells (Invitrogen) were performed as described previously (25). Wild-type rPCI was purified by batch absorption to heparin-Sepharose as detailed (25). H helix mutants were purified with the following modifications. Conditioned medium was diluted 1:2 with 20 mM HEPES, 10 mM NaCl, 0.1% PEG, 0.02% NaN,, pH 6.5, to lower the ionic strength before batch absorption to heparin-Sepharose. The resin was washed twice with this HEPES buffer, pH 6.5, containing 150 mM NaCl and once with this buffer containing 250 mM NaCl before elution with 1 M NaCl in HNPN buffer (20 mM HEPES, 150 mM NaCl, 0.1% PEG, 0.02% NaN,, pH 7.4). The proteins were dialyzed into HNPN buffer and quantitated by enzyme-linked immunosorbent assay and thrombin inhibition rate constants in the absence of heparin. Typical protein yields from 50 ml of culture medium (from two T-150 flasks of confluent High-FiveTM cells) were 70 pg of wild-type rPCI and 15-40 pg of the H helix mutants.
Competition Assays-Fusion protein competition in proteinase inhibition assays was performed at ambient temperature in 96-well microtiter plates under pseudo first-order conditions. Increasing concentrations (0.05-5 PM) of fusion protein or MBP alone were added to assay mixtures containing inhibitor, 1 pg/ml heparin, and 2 mg/ml BSA in HNPN buffer. Reactions were started by the addition of proteinase. Inhibitor-proteinase concentrations used were 50 nM PC1 and 5 nM thrombin, 100 nM PC1 and 10 nMAPC, and 10 nM antithrombin and 1 nM thrombin. After incubation, remaining proteinase activity was measured by hydrolysis of 0.15 mM chromogenic substrate with 2 mg/ml Polybrene. Second-order proteinase inhibition rate constants (k,) were calculated as -ln(a)/t[Il, where a is the fractional proteinase activity remaining relative to the uninhibited control, t is the incubation time, and [I] is the inhibitor concentration. After incubation, the remaining chromogenic activity was measured by adding 0.5 mM S-2366 to the assay mixture. The rate of p-nitroaniline formation was linearly related to the free proteinase concentration. Control experiments verified that none of the peptides affected proteinase chromogenic activity or the ability of PC1 to inhibit APC or thrombin in the absence of heparin.
Heparin-Sepharose Affinity Chromatography-The relative affinity of recombinant and plasma PCIs for immobilized heparin was determined using the fast protein liquid chromatography system of Pharmacia Biotech Inc. and a 1-ml heparin-Sepharose column. Samples were loaded in 20 mM HEPES, 10 mM NaC1, 0.1% PEG, pH 7.4, and eluted with a linear 1.5 mumin salt gradient of 10 mM to 1.2 M NaCl. 120 pg of the fusion proteins were loaded, and elution was monitored by continuous absorbance a t 280 nm readings. 0.5 pg of full-length rPCI was loaded, 30 x 1-ml fractions were collected, and the fractions were analyzed by enzyme-linked immunosorbent assay. The salt gradient and elution profile were plotted, and the salt concentration at which peak elution occurred was calculated. The mean of five to seven runs on two separate columns is reported for the fusion proteins, and the mean of three runs is reported for full-length rPCI and plasma PCI.
Heparin-accelerated Proteinase Inhibition by Full-length rPCI and Mutants-Inhibition assays were performed a t ambient temperature in BSA-coated 96-well microtiter plates and contained 5 nM rPCI and 0.5 nM thrombin (or 10 nM rPCI and 1 nM APC), 0.1 mg/ml Polybrene or various concentrations of heparin, and 2 mg/ml BSA in HNPN buffer. Reactions were started with the addition of proteinase, and after incubation, the remaining proteinase activity was measured with 0.15 mM chromogenic substrate containing 2 mg/ml Polybrene. Second-order inhibition rate constants were calculated as described above. Each assay consisted of triplicate determinations, and the mean inhibition rate from three separate assays is reported. For each mutant, preparations from two separate clones were analyzed and gave comparable results.

RESULTS
Recombinant PCI-(1-294) Fusion Proteins-A recombinant PC1 fragment with both A+ and H helices but lacking the reactive-site loop region (therefore possessing no inhibitory activity) was expressed in E. coli as a fusion protein with MBP and is designated "wild-type" PCI-(1-294) fusion protein. Three helix deletion mutants were also constructed in which residues 1-15 of PC1 (A+ helix), residues 264-277 (H helix), or both the A+ and H helices were deleted, and they are designated AA+ helix, AH helix, and AA+ helix/AH helix fusion proteins, respectively (Fig. 1). To measure the effect of helix deletions on heparin binding by PCI-(1-294) fusion protein, increasing concentrations of fusion protein were added to an assay system containing thrombin, plasma-derived PCI, and-heparin. A heparin-binding protein would be expected to compete with plasma PC1 for heparin, thereby decreasing the rate of proteinase in- inhibition assays containing 50 nM PCI, 5 nM thrombin, and 1 pg/ml heparin. After incubation, the remaining thrombin activity was measured by tosyl-Gly-Pro-Arg-p-nitroanilide hydrolysis. The y axis shows thrombin activity relative to the activity in the absence of PCI. Increased thrombin activity indicates decreased inhibition due to competition between PCI and fusion protein for heparin binding. The mean of normalized data from three separate experiments is presented.
hibition. Wild-type PCI-(1-294) fusion protein blocked heparinaccelerated thrombin inhibition in a dose-dependent manner (IC5,, = 1 VM) (Fig. 2). Free MBP alone or wild-type fusion protein in the absence of heparin did not have this property (data not shown). The AA+ helix fusion protein had a doseresponse curve much like that of the wild-type fusion protein, but both the AH helix and AA+ helix/AH helix fusion proteins had a drastically reduced ability to compete with plasma PC1 for heparin binding (Fig. 2).
The fusion proteins were also assayed for their effect on heparin-accelerated APC inhibition by PC1 and thrombin inhibition by antithrombin. For both inhibitor-proteinase pairs tested at a fixed concentration of 3 p~ fusion protein and 1 yg/ml heparin (k, = 18.9 x lo4 M -~ min-' for PCI-APC and 18.4 x lo8 M-' min" for antithrombin-thrombin in the absence of fusion proteins), the addition of the wild-type or AA+ helix fusion protein reduced the rate of proteinase inhibition by both serpins to -9 or -16%, respectively, of the inhibition rate in the absence of fusion protein. In contrast, the AH helix fusion protein competed less effectively, reducing the rate of heparinaccelerated proteinase inhibition by both serpins to -53% of the rate without fusion protein present. The wild-type fusion protein and all three helix deletion fusion proteins also bound to heparin-Sepharose. In contrast, free MBP did not bind and was recovered in the flow-through fraction. The sodium chloride concentration required for peak elution was determined to provide a measure of relative affinity. The fusion proteins bound to heparin-Sepharose with highest to lowest relative affinity in the following order: wild-type (590 mM NaC1) > AA+ helix (500 mM NaC1) > AH helix (410 mM NaC1) > AA+ helix/AH helix (260 mM NaC1).
Altered PCI H Helix Synthetic Peptides-The PC1 fusion protein study directed our focus to the H helix of PCI. To test if certain basic residues in the H helix may play a more prominent role than others in the PCI-heparin interaction, synthetic peptides were assembled corresponding to PCI-(264-283) (SEKTLLKWL_KMFmQLELY) in which each Lys and Arg was substituted individually with Glu. The heparin binding properties of the peptides were assessed in competition assays and shown to decrease the rate of proteinase inhibition in a dose-dependent manner (Fig. 3A), and they generated sigmoidal dose-response curves like those obtained for the PC1 fusion proteins (data not shown). All the altered peptides were less seven altered peptides, in which each Lys and Arg was individually substituted with Glu, were tested for their ability to interfere in heparin-accelerated PCI-thrombin and PCI-APC inhibition assays as described under "Experimental Procedures." The peptide concentrations required for 50% reduction (1C5J of the inhibition rate constant are given on the y axis. The mean of three separate APC experiments and two thrombin experiments is reported. B , peptides were loaded onto a heparin-Sepharose column (2-ml bed volume) equilibrated in 50 mM phosphate buffer, pH 7.4. They were eluted with a 20-ml gradient of 0-750 mM NaCl in phosphate buffer at 0.5 mumin and at room temperature with detection at 214 nm. Conductivity (microsiemens (rnS)) measurements were made for the peak elution fraction, and the mean of two to eight separate runs for each peptide is shown. competitive, requiring higher concentrations than the wildtype H helix peptide (IC5o = 5 p~) to get the same degree of interference (Fig. 3A). In particular, H helix peptide R269E was the least competitive for both APC (ICEo = 45 PM) and thrombin (ICEo = 30 p~) inhibition by PCI-heparin. Furthermore, a similar trend was seen with the altered H helix peptides when assessed for heparin-Sepharose binding (Fig. 3B).
Site-directed Mutagenesis of H Helix Basic Residues in Fulllength rPCI-The role of specific basic amino acids in the H helix region of PC1 was further assessed using full-length active rPCI produced by a Baculovirus expression system. Two H helix double mutants were generated by replacing Arg-269 and Lys-270 with Ala (R269A:K270A rPCI) and Lys-276 and Lys-277 with Ala (K276AK277A rPCI). The H helix mutants were compared with wild-type rPCI for their heparin binding properties. Wild-type rPCI eluted from heparin-Sepharose at 630 mM NaC1. Both R269AR270A and K276A:K277A rPCI showed Protein C Inhibitor 28693 reduced relative heparin-Sepharose binding, both eluting at 460 mM NaCl (Table I).
Second-order rate constants for APC and thrombin inhibition by rPCIs were determined in the absence and presence of increasing concentrations of heparin. Wild-type rPCI and both H helix substitution mutants yielded typical bell-shaped heparin template curves (Fig. 4). APC inhibition in the absence of heparin yielded rate constants of 4.71 x lo4, 4.80 x lo4, and 3.90 X lo4 M" min" for wild-type, R269A:K270A, and K276AK277A rPCI, respectively. In the presence of optimal heparin concentrations, APC inhibition rates for wild-type, R269A:K270A, and K276AK277A rPCI were accelerated 32-, 1 5 , and 39-fold, respectively ( Fig. 4 and Table I). In the absence of heparin, thrombin inhibition rates for wild-type, R269A:K270A, and K276A:K277ArPCI were 1.11 x lo6, 0.97 x lo6, and 0.79 x lo6 M-' min-I, respectively. In the presence of optimal heparin concentrations, thrombin inhibition rates for wild-type, R269A: K270A, and K276AK277A rPCI were stimulated 27-, 11-, and 32-fold, respectively ( Fig. 4 and Table 1). For both APC and thrombin inhibition, R269A:R270A and K276A:K277A rPCI required more heparin t o reach the maximal inhibition rate than did wild-type rPCI. DISCUSSION We utilized three approaches to assess the role of the H helix as a putative heparin-binding site in PCI. A recombinant truncated PC1 fusion protein was shown to interfere in heparinaccelerated proteinase inhibition and to bind heparin-Sepharose. Deletion mutagenesis showed that the H helix played a greater role than the A+ helix in heparin binding by the PC1 fusion protein and supports the hypothesis that the H helix is an important heparin-binding site in PCI. PC1 H helix peptides were also shown to bind heparin in competition assays. Substitution of specific basic residues with Glu reduced heparin binding by the peptides compared with the wild-type peptide. The peptide study further implies the importance of electrostatic interactions in PCI-heparin binding and indicates in particular that Arg-269 may be especially important. These results refine and extend previous reports that implicated the H helix as a putative heparin-binding site in PCI.
The 7 basic residues in the H helix region can be grouped into two clusters that have homology t o two heparin-binding consensus sequences (19), and Arg-269 is in the first cluster. The peptide data do not provide compelling evidence for the importance of one basic cluster or consensus sequence over the other. To address that question, we generated full-length active rPCI and two H helix double substitution mutants corresponding to adjacentbasicresiduesineachcluster,R269A:K27OAandK276A K277ArPCI. Arginine 278 was not mutated in rPCI because ( a ) a basic residue is conserved in that position in many serpins (111, (b)Arg-278 is at the boundary of the next exon, and (c) an a-helical wheel plot of PCI-(264-283) predicts that Arg-278 lies on the opposite face of a helix to the other basic residues (12). In heparin-accelerated APC and thrombin inhibition assays, R269A:K270A and K276A:K277A rPCI required more heparin to reach the maximal inhibition rate compared with wild-type rPCI, consistent with their reduced relative affinity for heparin-Sepharose. However, K276A:K277A rPCI achieved essentially the same maximal heparin acceleration as wild-type rPCI, whereas R269AK270A rPCI exhibited a markedly reduced heparin stimulation. These results indicate that while both Arg-269 and/or Lys-270 and Lys-276 and/or Lys-277 appear to contribute to overall heparin binding by PCI, only Arg-269 and/or Lys-270 is required for the maximal heparin-accelerated proteinase inhibition response. These full-length rPCI mutants suggest that reduced heparin affinity does not necessarily result in reduced heparin-accelerated proteinase inhibi-Heparin-accelerated proteinase inhibition by rPCI and H helix mutants Relative heparin-Sepharose affinity is given as the NaCl concentration required for peak elution (n = 3). The heparin optimum is the concentration at which the maximum inhibition rate occurs. The rate increase is calculated as the ratio of the maximum rate to the rate in the absence of heparin. tion. PC1 binding to heparin may include random association events, not just the productive binding of serpin to heparin that catalyzes proteinase inhibition. Therefore, it is possible that some basic residues are required for the "optimal alignment" of PC1 on the heparin molecule, while others only contribute to the overall "tightness" of binding. A role for the H helix in heparin binding to PC1 is in sharp contrast to that in antithrombin and heparin cofactor 11. Using a combination of chemical modification studies, analysis of natural human variants, and site-directed mutagenesis, the heparin-binding site in antithrombin and heparin cofactor I1 has been localized primarily to the D helix, with additional contributions by residues in the A helix for antithrombin (26-31). The conserved distribution of basic residues in the D helices of antithrombin and heparin cofactor I1 is conspicuously absent in PC1 (Fig. 5). PCI's C1 and D helices collectively contain 4 basic residues, but a synthetic peptide corresponding to this region (residues 80-93) did not readily bind to heparin Amino acid sequence alignment of the putative heparin-binding sites in heparin-binding serpins. Selected sequences of PCI, antithrombin (AT), and heparin cofactor I1 (HCII) are aligned with the sequence of cy,-proteinase inhibitor ( a l P I ) , a non-heparin-binding serpin, as described by Huber and Carrel1 (11). Basic residues are circled for emphasis. Residues implicated in heparin binding by antithrombin (reviewed in Refs. 26 and 27) and in hepariddermatan sulfate binding by heparin cofactor I1 (28-31) are shaded. (12). The basic charge density in the H helix of PC1 is greater than that in antithrombin, heparin cofactor 11, and qproteinase inhibitor (Fig. 5). Interestingly, acidic residues in antithrombin align with most of the basic residues in the H helix of PCI.
The location of PCI's primary heparin-binding site in a position different from the heparin-binding sites in antithrombin and heparin cofactor I1 may help explain PCI's more modest response to heparin. Under identical thrombin inhibition assay conditions, optimal heparin concentrations elicited a 9000-fold rate increase for heparin cofactor 11, a 2500-fold rate increase for antithrombin, but only a 15-fold rate enhancement for PC1 (18). This difference in rate enhancement does not correlate with serpin-heparin affinity (18). Since heparin acts as a template to bind both serpin and proteinase, the "orientation" of the serpin reactive site and the proteinase active site could be governed, at least in part, by how the serpin binds to heparin. Binding heparin through the H helix instead of the D helix may alter the reaction geometry such that PC1 is put in a less favorable orientation to react with heparin-bound proteinase.
In summary, we have shown that the H helix is an essential component of the heparin-binding site in PCI. It remains to be seen what contribution other structural regions, such as the A+, C1, and D helices, or other basic residues in the H helix make to both heparin binding and heparin-accelerated proteinase inhibition by PCI.

Heparin
Binding to Protein C Inhibitor 28695