Protein C is a multidomain protein. Mutations influencing protein C activity have been identified in each domain.32,33 Mature protein C has a mass of approximately 62,000 Daltons and circulates at approximately 65 nM. Protein C is synthesized with a leader peptide followed by a propeptide. The propeptide is involved in recognition by the vitamin K-dependent carboxylase and, hence, mutations in this region could result in incomplete γ-carboxylation of Glu residues with a resultant decrease in biologic activity. Both the leader and propeptide are removed proteolytically during maturation.80,81 A schematic representation of protein C, protein S, TM and EPCR, in relationship to the membrane surfaces at which they function, is depicted in Fig. 170-3. The sequence of human protein C is shown in Fig. 170-4. The mature protein C molecule consists of an N-terminal Gla domain, a hydrophobic stack region that connects the Gla domain to the two EGF domains, and a protease domain with homology to trypsin. In general, at the gene level, these domains are separated by introns.82 Human protein C circulates primarily as a two-chain zymogen, but approximately 10 percent of the protein C circulates as a single chain.83 The thrombin-TM complex is the major physiological activator of human protein C.84 This activation occurs by cleavage at Arg 169, releasing a 12-residue peptide.85 In the single chain protein C, no peptide is released during activation.
Representation of protein C, protein S, thrombomodulin, and EPCR. The nature and overall domain structure of protein C, protein S, thrombomodulin, and EPCR are illustrated in addition to a representation of the nature of the membrane interactions. Specific domains of each protein are identified. Gla, γ-carboxyglutamic acid; Th-sens., thrombin-sensitive. (Modified from Esmon C: The roles of protein C and thrombomodulin in the regulation of blood coagulation. J Biol Chem 264:4743, 1989. Used with permission.)
Amino acid sequence of human protein C. Amino acids are numbered from the amino-terminus of the mature protein. Y depicts γ-carboxylation, and an oval depicts hydroxylation of an amino acid. Diamonds represent sites of N-linked glycosylation. Residues within the two EGF-like domains are shaded. The serine, aspartic acid, and histidine residues that constitute the catalytic active site are identified in black. > denotes the location of an intron in the protein C gene, and the roman numeral identifies the following exon. The dipeptide proteolytically removed during the post-translational processing of most protein C molecules is marked by the small arrows. The site of proteolytic cleavage during protein C activation is identified by the large arrow.
There are Ca2+ binding sites located in the Gla domain,15 the first EGF domain,20,86,87 and the protease domain of protein C.88 The metal-binding sites in the Gla domain are required for binding to phospholipid15 and to EPCR.89 Mutational studies of the Gla domain suggest that the N-terminal half of the domain is critical for membrane interaction.15 Based on site-directed mutagenesis studies, it is clear that mutation of Gla residues 7, 20, 26, and 29 results in nearly complete loss of APC anticoagulant activity, mutation of Gla 25 decreases activity 75 percent, and that Gla 6, 14, and 19 are not critical for any APC functions tested to date90– 93 (reviewed in reference 15). Mutation of the hydrophobic residues near the N-terminus also disrupts phospholipid binding, particularly at Leu 5.94 In addition to the importance of Gla residues in membrane binding, complete carboxylation is required for protein C binding to EPCR.89
APC functions much better on phospholipid vesicles containing phosphatidylethanolamine,50,95 and protein C activation proceeds more rapidly on these vesicles.96 In contrast, the prothrombin activation complex demonstrates a lower phosphatidylethanolamine dependence. Chimeric protein C in which the Gla domain (first 46 residues) is replaced with that of prothrombin lacks the phosphatidylethanolamine-dependence characteristic of the normal protein C.50 Plasma anticoagulant activity of this chimera is also protein S-independent. It is the C-terminal 24 residues of the Gla domain that are responsible for the phosphatidylethanolamine dependence.50 This same region is necessary for protein S-dependent enhancement of factor Va inactivation and for the participation of protein S as an anticoagulant in plasma. These results suggest that missense mutations in the C-terminal half of the Gla domain could have subtle influences on protein C activity that would require relatively sophisticated assays to detect.
The Ca2+ binding site in the first EGF domain aids in stabilizing the functional conformation of the protein C Gla domain.97,98 This site is not critical for activation by the thrombin-TM complex in solution.99 Protein C contains a β hydroxy aspartic acid residue100 in this domain. If this residue is mutated to Glu101 or Ala,102 then Ca2+ binding to APC is impaired, which results in reduced plasma anticoagulant activity. Therefore, it is likely that other mutations that interfere with Ca2+ binding would result in reductions of functional APC activity and probably protein C activation, if the assay system involves any phospholipid- or membrane-dependent steps.
There is a Ca2+ binding site in the protease domain88 that potently modulates protein C activation. With thrombin alone, Ca2+ reduces the activation rate more than tenfold primarily by reducing protein C affinity for thrombin.103 In contrast, activation by the thrombin-TM complex depends on the presence of Ca2+.104 The Ca2+ binding site in the protease domain is similar to that in trypsin.88,99,105
The crystal structure of APC lacking the Gla domain has been determined to 2.8 Å (Fig. 170-5),105 and a molecular model of the protease domain of APC has been published.106 Regions of APC thought to be functionally important are discussed in the context of the APC structures in these two papers. Several features that are likely to account for the specificity of APC are revealed by the structure. APC has a groove located between the active site and the Ca2+ site that is a candidate for engaging the residues on the carboxyl side (P′) of the cleavage site of substrates in a manner analogous to anion binding exosite 1 of thrombin. Both grooves are located in similar regions of the protease domains of their respective molecules. Anion binding exosite 1 in thrombin is the docking site for TM,107,108 the protease-activated thrombin receptor 1,109– 113 fibrinogen,114– 116 factor V, and factor VIII.117 Therefore, this groove in APC, when occupied by ligands, is a candidate for eliciting conformational changes and/or interacting with receptors. Of interest, factors V and VIII have highly conserved sequences on the P′ side of Arg 506118 of factor V and Arg 562 in factor VIII, sites rapidly cleaved by APC in factors Va and VIIIa. These sequences are rich in acidic and hydrophobic residues similar to the corresponding domain in the thrombin receptor.110
Space-filling model of the activated protein C molecule. The model is based on the crystal structure of activated protein C lacking the Gla domain.105 The location of a chlormethylketone inhibitor located in the active site is shown in black wire. Basic residues are in black, acidic residues are gray, and other residues are off-white. The protease domain is essentially round and the EGF domains form the “neck” extension at the base of the protease domain. The three basic residues critical for activation by the thrombin-TM complex are at the top of the figure. The exosite runs just underneath these residues. The Ca2+ binding site is to the far right of the protease domain near the acidic residue that is almost completely hidden by the basic residues.
Highly variable glycosylation of protein C results in a complex pattern when analyzed electrophoretically. Further complicating the pattern, plasma contains both two-chain (90 percent) and single-chain (10 percent) protein C. Both forms exist with 4, 3, and 2 N-linked carbohydrate chains.83 The extent of glycosylation alters the anticoagulant activities of APC and the rate of activation of protein C.119 Glycosylation at Asn 329 is probably responsible for the highest molecular weight (α) form of protein C. Mutation of the glycosylation sites increases APC anticoagulant activity two- to threefold. Glycosylation at 313 results in a 2.5-fold reduction in activation rate due to a corresponding increase in the Km . Expression of protein C in human kidney 293 cells leads to decreased sialic acid and increased GalNAc and fucose120 compared to plasma protein C. A polylactosamine on protein C also has been implicated in inhibiting cell adhesion through E-selectin,121 a potential anti-inflammatory function of protein C. Thus, alteration of the glycosylation of protein C can significantly influence the biologic activity of the molecule and contribute to discrepancies between the antigen and activity ratios. These ratios are strongly influenced by the nature of the functional assays employed.
Organization of the Protein C Gene
The protein C gene is located in the q13-q14 region of chromosome 2.122,123 The gene is approximately 11 kb in length and consists of 9 exons and 8 introns (GenBank M11228).82,124 The first exon codes for the N-terminal portion of the pre-pro leader sequence. The second exon codes for the remainder of this sequence, the basic propeptide involved in recognition by the vitamin K-dependent carboxylase and Gla domain. The third exon codes for the aromatic stack of hydrophobic residues that separates the Gla domain from the EGF domain. The fourth and fifth exons code for the first and second EGF-like domains. The sixth exon encodes for the section between the protease domain and the EGF domain, the activation peptide region and the N-terminal portion of the APC protease domain. The seventh exon codes for the middle portion of the protease domain and contains the active site His, while the eighth exon codes for the C-terminal half of the protease domain and contains the active site Asp and Ser residues. Two major mRNA species are observed that appear to be due to different polyadenylation sites.125,126 There are several polymorphisms in the protein C gene, none of which result in amino acid changes. The nucleotide polymorphisms are T/A at −1476, C/T at 3204, T/G at 3342, A/G at 6181 and T/C at 7228.32
Clinical Manifestations of Protein C Deficiency
The frequency of protein C deficiency (i.e., about a 50 percent reduction in protein C levels) in the general population has been estimated at 1/30034,127 to 1/500. The majority of heterozygous protein C deficient individuals have no history of thrombosis.127 However, analysis of patients with thrombosis clearly indicates that protein C deficiency is a risk factor.128 Examination of a large family with two mutations in the Gla domain129 has clearly demonstrated that the protein C deficiency is associated with an increased risk of thrombosis.130 Clinical manifestations of protein C deficiency include deep vein thrombosis and pulmonary embolism.32 Although several case reports of protein C deficiency and arterial thrombosis have been published,129,131,132 protein C deficiency seems to be, at most, a weak risk factor for arterial thrombosis in the general population.133
Early studies indicated that patients with protein C deficiency are at an increased risk of developing warfarin-induced skin necrosis.134– 136 Direct involvement of protein C depletion in this process is supported by the observation that the progression of the skin necrosis appears to be prevented by protein C infusion.27,135 Protein C levels often decrease in patients with septic shock or liver failure, complicating the diagnosis.78,137
The Genetic Basis of Protein C Deficiency
The genetic basis of protein C deficiency has been examined in considerable depth.32,33 The results indicate that the mutations are scattered throughout the gene. Identification of the site of the mutation has yet to aid in the treatment or prediction of future thrombotic complications. In the original review of 142 entries,32 there were three promoter mutations, 10 splice-site abnormalities, 2 in-frame deletions, 6 frameshift deletions, 1 in-frame insertion, 3 frameshift insertions, 26 nonsense mutations, 93 missense mutations, and 1 silent mutation. The protein C gene contains a relatively high percentage (0.60 percent) of guanine plus cytosine. Of the single base-pair substitutions in the protein C gene, 43 percent occurred in CpG dinucleotides that resulted in C6T or G6A transitions, probably indicating a methylation-mediated deamination. Most of the CpG mutations in the protein C gene occur in exon 7.32 Protein C deficiency is classified as type I (protein C antigen and activity are decreased equivalently) and type II (antigen levels are near normal, but the activity levels are reduced). Characterization of the type II mutations should involve assays of protein C activation and APC anticoagulant activity. A database of protein C mutations can be found at Web site www.xs4all.nl/~reitsma/Prot_C_home.htm.