Alternative mRNA is favored by the A3 haplotype of the EPCR gene PROCR and generates a novel soluble form of EPCR in plasma

Abstract

The endothelial cell protein C receptor also exists in soluble form in plasma (sEPCR), resulting from ADAM17 cleavage. Elevated sEPCR levels are observed in subjects carrying the A3 haplotype, which is characterized by a Ser219Gly substitution in the transmembrane domain, rendering the receptor more sensitive to cleavage. Because sEPCR production is not completely blocked by metalloprotease inhibition, we looked for another mechanism. Comparing mRNA expression patterns and levels in A3 and non-A3 cells from 32 human umbilical cord veins, we detected a truncated mRNA in addition to the full-length mRNA. This truncated mRNA was 16 times more abundant in A3 human umbilical vein endothelial cells than in non-A3 human umbilical vein endothelial cells and encoded a protein lacking the transmembrane domain. We stably expressed a recombinant form of this protein (rEPCRisoform) and a protein mimicking the plasma sEPCR (rEPCRsol). Functional studies of the purified recombinant proteins revealed that the rEPCRisoform bound to recombinant protein C with similar affinity than rEPCRsol and that it also inhibited the anticoagulant activity of APC. Trace amounts of the EPCR isoform were found in the plasma of A3 subjects. These results suggest that the sEPCRisoform could contribute to the regulatory effect of sEPCR in plasma.

Figure 1

Polyacrylamide gel electrophoresis of the amplified 5′ parts and 3′ parts of PROCR cDNA derived from HEK293 and EA.hy 926 cells, and of the amplified 3′ part derived from HUVECs from umbilical cords with different PROCR genotypes. (A) Schematic representation of the EPCR cDNA structure and the location of the primers used to amplify the 5′ and 3′ parts. The translated region of the full-length mRNA is in grey. (B) PCR product from HEK 293 EPCR cDNA and EA.hy 926 EPCR cDNA obtained using EPCR2347Fr and EPCR6970Rv as amplification primers. (C) PCR product from HEK 293 EPCR cDNA and EA.hy 926 EPCR cDNA obtained using EPCR4993Fr and EPCR7320Rv as amplification primers. (D) PCR product from HUVECs EPCR cDNA obtained using EPCR4993Fr and EPCR7320Rv as amplification primers.

Figure 2

Sequences of the 2 amplification products of the 3′ part of EA.hy 926 PROCR cDNA, and location in the mature mRNA. The sequences shown here were obtained using EPCR4993Fr as sequencing primer and are restricted to sequences corresponding to the normal junction (upper electrophoregram) and the unusual junction (lower electrophoregram) of the full-length and truncated mRNAs, respectively. The schema indicates the exonic organization of the full-length and truncated mRNAs. Numbers indicate the first and last nucleotides of the exons of the genomic sequence (according to AF106202 numbering). In the truncated mRNA, exon 3 is linked to nt 7272 of the untranslated region of PROCR exon 4, which leads to a new stop codon at position 7441.

Figure 3

Sequences of the C-terminal parts of transcripts corresponding to the full-length mRNA and the truncated mRNA. Nucleotide sequence and predicted protein sequence of the EPCR isoform (B) compared with those of membrane EPCR (A). Only the C-terminal parts of the proteins are shown. The new exon 3/exon 4 junction generates a C-terminal part with 56 new amino acids (indicated in bold). A potential N-glycosylation site is underlined. The sequence corresponding to the recombinant sEPCR lacking the transmembrane domain prepared here ends at amino acid T 209 (A).

Figure 4

Recombinant proteins prepared in this study and comparison with membrane EPCR and plasma sEPCR resulting from shedding by ADAM17. The amino acid numbering starts at Met + 1. The signal peptide (hatched stretch) that is cut before protein addressing to the membrane (EPCR and sEPCR) or before protein secretion (rEPCRsol and rEPCRisoform) comprises amino acids 1 to 17. The C-terminal amino acid of plasma EPCR (sEPCR) is unknown (↔). The gray stretch denotes the putative transmembrane domain predicted by TMHMM Server, version 2.0 (panel below Membrane EPCR). The hatched gray stretch corresponds to the new C-terminal part of EPCRisoform (relative to membrane EPCR and sEPCR) lacking the transmembrane domain, owing to mRNA truncation.

Figure 5

SDS-PAGE and Western blot analysis of native forms of rEPCRsol and rEPCRisoform, and SDS-PAGE of deglycosylated rEPCRsol and rEPCRisoform. (A) A total of 2.5 μg rEPCRsol and rEPCRisoform was analyzed by SDS-12% PAGE in reducing conditions. (B) Western blot was performed using 0.5 μg recombinant protein per lane. Recombinant proteins were transferred to nitrocellulose membranes on a semidry apparatus, and proteins were detected with a goat polyclonal biotinylated antibody raised against human EPCR; binding was revealed with HRP-labeled streptavidin. (C) SDS-PAGE of rEPCRsol and rEPCRisoform deglycosylated by PNGase F, and determination of polypeptide chains MW. The theoretical MWs based on the number of amino acids are indicated in parentheses.

Figure 6

Binding of rWT PC to rEPCRsol and rEPCRisoform, and effect of rEPCRsol and rEPCRisoform on the plasma clotting time. The results in each panel are the mean of at least 3 separate experiments. (A) Binding of rWT PC to rEPCRsol (■) and to rEPCRisoform (□) in the presence of CaCl₂ and MgCl₂. REPCRsol and rEPCRisoform were captured by 400 ng of immobilized polyclonal antibody to EPCR. Various dilutions of rWT PC were added to the wells in the presence of 1.3 mM Ca²⁺ and 0.6 mM Mg²⁺, and bound rWT PC was detected with a peroxidase-conjugated anti-PC polyclonal antibody. Solid curves were obtained by nonlinear regression analysis. They yielded Kd_app values of 142 nM and 134 nM for rEPCRsol and rEPCRisoform, respectively. The data represent values (mean ± SD) of 4 separate experiments. The standard deviation is not apparent at all points because they are sometimes smaller than the symbols. (B) Increasing concentrations of rEPCRsol (■) and rEPCRisoform (□) were incubated for 2 minutes at 37°C with APC (1.8 nM, final concentration) in the presence of phospholipid vesicles, RVV-X, and PC-deficient plasma. Clotting was initiated by adding 14 mM CaCl₂ (final concentration), and the clotting time was recorded by an ST4 analyzer. The APC-inhibiting effect of the 2 recombinant proteins was expressed as the following ratio: Clotting time in the presence of APC alone/Clotting time in the presence of both APC and rEPCRsol or rEPCRisoform. The standard deviations are not visible at all points because they are sometimes smaller than the symbols.

Figure 7

Detection of sEPCR in plasma of A3-carrying subjects. Western blot of 2 JRK-1496 column elution fractions, each corresponding to the plasma EPCR forms contained in 2.5 mL of a pool of plasma from A3-carrying subjects (lanes 1 and 2) and of 2 JRK-1496 column elution fractions each corresponding to the plasma EPCR forms contained in 5 mL of pooled plasma from A3-noncarrying subjects (lanes 3 and 4) The 2 lanes on the left were loaded with recombinant rEPCRsol and rEPCRisoform, respectively, used as controls. The membrane were blocked, treated with goat anti-EPCR biotinylated antibody, washed, and then incubated with HRP-conjugated streptavidin. Bands were visualized with enhanced chemiluminescence (Pierce Chemical, Rockford, IL). Vertical lines have been inserted to indicate repositioned gel lanes, and the arrows point to the plasma EPCR isoform.