The complement system presents a powerful defense against infection and is tightly regulated to prevent damage to self by functionally equivalent soluble and membrane regulators. We describe complement C2 receptor inhibitor trispanning (CRIT), a novel human complement regulatory receptor, expressed on hemopoietic cells and a wide range of tissues throughout the body. CRIT is present in human parasites through horizontal transmission. Serum complement component C2 binds to the N-terminal extracellular domain 1 of CRIT, which, in peptide form, blocks C3 convertase formation and complement-mediated inflammation. Unlike C1 inhibitor, which inhibits the cleavage of C4 and C2, CRIT only blocks C2 cleavage but, in so doing, shares with C1 inhibitor the same functional effect, of preventing classical pathway C3 convertase formation. Ab blockage of cellular CRIT reduces inhibition of cytolysis, indicating that CRIT is a novel complement regulator protecting autologous cells.
An overview of known human membrane regulators of complement activation (RCA)5 (1, 2) indicates that cell membranes have no equivalent to C1 inhibitor (C1-INH), a serine protease inhibitor that regulates the classical pathway (CP). C1-INH also inhibits the mannan-binding lectin pathway by inhibiting mannan-binding lectin-associated serine proteases (3). Schistosoma complement C2 receptor inhibitor trispanning (CRIT; formerly, trispanning orphan receptor) like C1-INH inhibits C1s-mediated cleavage of C2, so preventing formation of the CP C3 convertase, C4b2a. However, their mode of action differs. C1-INH regulates the CP at the C1 level by binding stoichiometrically to the active sites on both C1r and C1s (4). CRIT binds to C2 in a manner that inhibits its C1s-mediated cleavage. There is no enzymatic inhibition and CRIT is presumably protecting the C1s cleavage site on C2.
Homologs of complement regulator proteins are found on parasites such as trypanosomes and schistosomes living in the mammalian host vasculature (5). Trypanosomes are protected by decay-accelerating factor (DAF) and CD59 homologs with decay-accelerating activity either for the alternative pathway (AP) (6) or else for both the CP and AP convertases (7), whereas schistosomes adsorb host DAF from erythrocytes (8). Schistosoma CRIT having a putative complement-protecting function in schistosomes is found on the surface of the Schistosoma parasite (9). We have now cloned a CRIT homolog from the Trypanosoma parasite, and found close homologs in human, rat, and cod.
We found human (Hu)-CRIT to be expressed on a wide range of tissues as well as on hemopoietic cells. We suggest that this novel receptor with three N-terminal transmembrane (TM) domains represents a novel family of receptors and present several lines of evidence supporting CRIT as receptor for C2. The complement regulatory function of CRIT was suggested previously (10, 11) by using CRIT extracellular domain 1 (ed1) and its 11-aa C terminus (CRIT-H17) to inhibit CP-mediated hemolysis in the presence of normal human serum (NHS). We have now shown that, by blocking CRIT through preincubation with anti-CRIT-ed1, two CRIT-expressing human myeloid cell lines as well as monocytes, can be deprotected, resulting in increased susceptibility to Ab-dependent complement lysis in vitro. When C2 binds to CRIT on the cell surface, it does not form a variant-C3 convertase, because C2 cannot be cleaved by C1s. Having found this previously with synthetic CRIT-ed1 peptide, we now show this to be the case with native membrane-bound CRIT.
Materials and Methods
Abs and peptides
Affinity-purified anti-CRIT-ed1 polyclonal Ab and anti-C2 Abs were described before (10). A rabbit polyclonal Ab against Schistosoma hematobium (Sh)-CRIT-ed2 (NH2-SSTSDIRLVIHTKTGPIYIKST-CO2H (two conservative amino acid changes relative to human CRIT-ed2)) was made and affinity purified as described for anti-CRIT-ed1. Anti-CRIT-intracellular domain 2 (id2) mouse polyclonal Ab was obtained by a standard immunization of mice with the synthetic peptide (NH2-KYEDILKIPTNAYAC-CO2H) based on K195-A208 of the cytoplasmic tail of human CRIT, coupled through a C-terminal cysteine to KLH. HRP-conjugated secondary Abs were obtained from Bio-Rad. All synthetic peptides were synthesized as described before (10). The sequences were as follows: ed1 (NH2-MSPSLVSDTQKHERGSHEVKIKHFSPY-CO2H); ed2 (see above);H17 (NH2-HEVKIKHFSPY-CO2H); H17S (NH2-EKFYHIHSKPY-CO2H); C4β212–232 (NH2-FEVKKYVLPNFEVKITPGKPYCO2H); and C4βS (NH2-FYFNEPVEPKVLPKGVKYTKI-CO2H).
Biotinylation, FITC labeling, and radioiodination of C2
C2 (50 μg) was biotinylated by using NHS-LC-biotin (Pierce) according to previously described methods (10), but using a lower biotin-to-protein ratio of 50 μg of biotin per milligram of protein. After dialysis vs PBS or PBS containing Mg2+ (1 mM), the protein was assayed. For FITC labeling, C2 (10 μg) was incubated with FITC (Sigma-Aldrich) (6 μl of 0.1 mg/ml) and 1 ml of carbonate buffer at room temperature for 2 h in the dark, followed by extensive dialysis against PBS. Radioiodination of C2 was conducted according to a standard protocol (12). Essentially, 100 μg of protein in 0.5 ml was added to 100 μg of IodoGen (Pierce). Iodination was conducted by adding 1 mCi of Na125I (Amersham Biosciences). Unincorporated iodine was separated from iodinated C2 using a PD-10 gel filtration column (Amersham Biosciences) presaturated with 5 mg of BSA. The specific activity of 125I-labeled C2 was 8 × 104 cpm/mg.
Human cells and cell lines
Lymphocytes were obtained from heparinized human blood diluted 1/1 with RPMI 1640 using Histopaque 1077 (Sigma-Aldrich). Erythrocytes and platelets were isolated according to a standard protocol (13). The lymphocytes were washed and resuspended in RPMI 1640. Erythrocytes were recovered, and the pellet was washed with PBS. Monocytes were obtained from heparinized blood buffy coat by density gradient centrifugation using Histopaque-1077 (Sigma-Aldrich). Monocytes were purified by adherence separation (14). Purified monocytes (2 × 105 cells/ml) were dispensed in 100-μl aliquots into the wells of a 96-well microtiter plate and stimulated, if necessary, with 10 U/ml final concentration of human IFN-γ (Sigma-Aldrich) and incubated overnight at 37°C in 5% CO2. To remove the fibroblasts that proliferate on addition of IFN-γ, the monocytes were washed three times with RPMI 1640. The human carcinoma cell lines, Jurkat, Raji, THP-1, U937, ECV-304, HeLaS3, HS1, MRC5, T47D, and rat cell line, C58(NT) (all American Type Culture Collection) were maintained in RPMI 1640 medium with 10% FCS, glutamine (2 mM), sodium pyruvate (1 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml) at 37°C and 5% CO2.
SDS-PAGE, Western and ligand (Far Western) blotting
Protein assays by dye binding using a Bio-Rad protein assay kit were conducted to ensure even loadings for SDS-PAGE. SDS-PAGE and immunoblotting analysis of proteins was conducted as described before (10). Reduced samples (with 20 or 50 mM DTT) and nonreduced samples were prepared in SDS-PAGE loading buffer. Where necessary in experiments to study the covalent oligomerization of CRIT, cell lysis buffer included 10 mM iodoacetamide. For ligand blotting, the protein to be used as probe (usually C2) was biotinylated and detected as described above. For preincubation of biotinylated C2 (C2bio) with peptides, before probing of a blot, a 10,000 M excess of peptide over C2bio was added at 4°C for 2 h.
Tissue extraction for Western analysis was conducted by homogenizing 50 mg of tissue in a Dounce homogenizer in 150 μl of RIPA (1% Nonidet P-40, 0.5% sodium deoxycholate, and 0.1% SDS in PBS) in the presence of protease inhibitors. After centrifugation (14 krpm for 10 min), the supernatant was respun as above.
For lysates of mammalian cells, cells (typically 30 × 106) were washed twice in PBS and resuspended in 0.75 ml of lysis buffer (PBS with 1% (v/v) Nonidet P-40 and 5 mM EDTA (pH 8)). After adding protease inhibitors (0.2 mM PMSF, 4 mM aminoethylbenzenesulfonylfluoride (Pefabloc), 1 mM NaF, 5 μg/ml aprotinin, 4 mM benzamidine) (all Sigma-Aldrich), the cells were mixed gently and incubated on ice for 10 min. After low-speed centrifugation (250 × g; 5 min), the supernatant was respun (25,000 × g; 25 min), and the supernatant was aliquoted and stored (−80°C).
CRIT expression in Escherichia coli was induced with isopropyl-d-thiogalactopyranoside for 1 h. After harvesting, the bacteria were resuspended in 1/10 the original volume of buffer (20 mM Tris-HCl (pH 8.0), 500 mM NaCl, 2 mM EGTA, protease inhibitor mixture as for mammalian cell lysate) and lysed with 0.1 mg/ml lysozyme, 0.1% Triton X-100, 10 mM MgCl2, 50 μg/ml DNase, and 20 μg/ml RNase for 60 min at room temperature. Cell lysates were centrifuged (25,000 × g for 30 min), and supernatants were used freshly.
Fresh surgical tissue (kidney and testis) was fixed in 10% neutral buffered formalin. Tissue was embedded in paraffin, and 4- to 5-μm sections were cut. Tissue for microarrays (BioCat) was similarly prepared. Blocking of endogenous peroxidase was conducted with 0.3% H2O2 in methanol for 30 min followed by three 5-min washes in distilled water and then one wash in PBS. Tissues were digested with 0.05% proteinase K in PBS for 2 min and then washed in PBS. Blocking of nonspecific staining was performed by incubating for 20 min in normal horse serum, excess serum then being blotted off the sections. The primary Ab was anti-CRIT-ed2 (1/100) incubated either overnight at 4°C or for 1 h at room temperature. For detection, the liquid diaminobenzidine substrate-chromogen system (DakoCytomation) was used according to the manufacturer’s instructions. Counterstaining was by hematoxylin Mayer, and dehydration by using decreasing alcohol concentrations. Finally, stained sections were mounted with permanent mounting medium (Eukitt/Fluka Chemical). Immunohistological examination was performed using a Zeiss Axiophot microscope as described before (11).
Essentially, in the hemolytic assay performed, the increasing concentration of CRIT-ed1 being added to NHS competes with C4b for binding to C2. The effect of this competition is measured in this case by the decreased CP C3 convertase assembly and resulting decrease of percent hemolysis. This standard procedure has been described previously (10).
Complement-dependent cytotoxicity on CRIT-expressing cells
Cells (0.25 × 106) were incubated with 100 μl of anti-human lymphocyte serum (1/30) or rabbit prebleed/normal rabbit IgG as a control followed by incubation with anti-CRIT-ed1. After 1 h at 4°C and two washes in isotonic veronal buffered saline with divalent cations (GVB++) (15), 100 μl of 10% NHS in GVB++ was added for 30 min at 37°C. The percent lysis was determined as the percentage of cells permeable to trypan blue.
To detect the binding of C2 to the Jurkat cell surface, the C2 was biotinylated as described above. Cells were treated according to the manufacturer’s instructions, for 5 min at room temperature, with an avidin/biotin blocking reagent (Vector Laboratories). Cells were then washed in RPMI 1640 (containing 1% BSA and 0.01% sodium azide), resuspended at a concentration of 1 × 106 cells/ml, and incubated on ice for 30 min with 50 μl of C2bio (2.75 μg/ml) made up to 100 μl with FACS buffer (1× PBS, 1% BSA, 0.02% sodium azide). After washes and incubation with 5-μl per reaction streptavidin-PE, the cells were fixed with 1% paraformaldehyde for 30 min at room temperature and analyzed on a FACScan flow cytometer (BD Biosciences). CellQuest software was used for data acquisition and analysis. A total of 10,000 events was analyzed. FITC-labeled C2 (25 μl of 10 μg/ml) was similarly incubated with 1.5 × 106 Jurkat cells.
To examine CRIT expression on the Jurkat cell surface, 1 × 106 cells incubated with avidin/biotin blocking reagent, as above, were treated as for labeling with C2bio but instead with biotinylated anti-CRIT-ed1 (20 μl of 50 μg/ml).
To investigate the C2 binding to CRIT-ed1 or C4b, first CRIT-ed1 peptide (0.5 μg/well) or C4b (2 μg/well) in carbonate buffer (pH 9.6) (with or without 5 mM MgCl2) were coated on the plate (overnight at 4°C). After washing in PBS (with 0.05% Tween 20) and blocking with 200 μl of PBS (containing 2% BSA), increasing amounts of C2bio (with or without 1 mM Mg2+) were added. To see whether CRIT-ed1 can block C2-C4b binding in this system, C4b in carbonate buffer was coated on the plate and incubated with C2bio (150 ng) in the presence or absence of 1.2 μM ed1 (300 ng) in a final volume of 100 μl. Finally, the procedure to detect C2bio binding in our ELISA requiring incubation with streptavidin-HRP (1/2500) followed by development with o-phenylenediamine was essentially as described before (16). A490 was measured using a microplate reader (spectroMAX 190; Molecular Devices).
Apparent dissociation constants (KD(app)) were calculated using nonlinear regression, which was performed by plotting A490 against the concentration of C2 (rather than concentration of bound C2 against concentration of free C2 as required for Scatchard analysis and KD calculations) and fitting with the one-binding class model (a better fit than the two-binding class model) using the GraphPad Prism software (GraphPad Software).
DNA isolation and Southern blotting
Rat genomic DNA was purchased (Stratagene). Human genomic DNA for Southern blotting was prepared from 500 μl of packed Jurkat cells (17). Southern blotting (17) using the full-length Sh-CRIT cDNA as a probe was conducted at low stringency by incubating at 45°C for 16 h, followed by two washes with 3× SSC and 0.1% SDS at room temperature, rinsing in 2× SSC. For higher stringency, blots were incubated at 65°C and then washed with 1× SSC and 0.1% SDS at 65°C. Blots were exposed to x-ray film with an intensifying screen for 48 h.
Cloning of CRIT homologs and in vitro transcription/translation
A rat homolog of Sh-CRIT was obtained by PCR from rat genomic DNA, using degenerate oligonucleotides based on the Sh-CRIT N-terminal and C-terminal nucleotide sequence and, where possible, based on a rat codon usage. The primers were as follows: RaF (sense), 5′-CGCGATGTC(C/T)CC(A/C/G/T)I(C/G)ICTIGTITC-3′; and RaR (antisense), 5′-GCGTTA(G/A)CAAGAIGA(C/G)TG(A/C/G/T)GC(A/G)TT-3′. The PCR was conducted at 94°C (30 s), 55°C (60 s), and 72°C (60 s) for 35 cycles. The purified product (using GeneClean (Bio 101)) was then cloned into pGEM-T (Promega), according to the manufacturer’s instructions, and fully sequenced in both strands. Human genomic DNA was extracted from whole blood using a blood and cell culture DNA mini-kit (Qiagen), and human testis and pancreas cDNA was purchased from Ambion. Based on the sequence alignment of Schistosoma and rat CRIT, oligonucleotides were designed with the introduction of a HindIII restriction site at the 5′-end and a KpnI restriction site at the 3′-end to enable subcloning into the pcDNA3.1 vector in the correct orientation (HuF (sense), 5′-CGAAGCTTATGTCTCCAAGTCTAGTG3′; and HuR (antisense), 5′-CGGGTACCTTAGCAAGAGAGTGAGC-3′). Pfu ultra high-fidelity DNA polymerase (Stratagene) was used, and the amplified PCR product was digested with HindIII and KpnI, and ligated into the pcDNA3.1 vector precut with the same enzymes. The ligation was transformed into DH5α competent cells (17). In vitro transcription/translation from the pcDNA3.1-HuCRIT template was conducted using the TNT-coupled reticulocyte lysate system (Promega) according to the manufacturer’s instructions. The trypanosome and cod homologs were obtained similarly by PCR using oligonucleotides based on the Sh-CRIT sequence.
All sequencing was conducted on both strands of the pGEM-T clones using an ABI Prism BigDye terminators, version 3.0, cycle sequencing kit and detected on an ABI 3100 Avant Genetic Analyzer, automated DNA sequencer, according to the manufacturer’s protocols (Applied Biosystems). Vector-specific primers SP6 and T7 were used, as well as primers based on the established insert sequence.
Assay for cleavage of C2 by C1s
This was essentially conducted as described previously (10). To test for the effect of CRIT as expressed on the Jurkat cell surface (5 × 103 cells) as well as the CRIT-ed1 synthetic peptide (105 μM), on C1s cleavage of C2, they were preincubated with C2 (2 μg) in 20 mM Tris at 20°C for 30 min. Then, 0.2 μg of C1s was added and incubated for 1 h at 37°C before analysis by immunoblotting.
Human and parasite CRIT
CRIT is found on the larval and adult worm stage of the Schistosoma parasite (9). Under the previous designation of “trispanning orphan receptor,” it has also been found recently on schistosome eggs (18). Schistosomes live in veins near the urinary bladder or intestines of the host and the CRIT receptor for C2 is able to limit CP C3 convertase formation on the parasite surface (10, 19, 20, 21). Southern blotting of human genomic DNA under low stringency conditions, using the Sh-CRIT open reading frame (ORF) as a probe, revealed a Hu-CRIT homolog (Fig. 1 A). It was also found that the Ab against the N-terminal extracellular domain, ed1, of Sh-CRIT recognized a 32-kDa protein in various human cells and tissues (described below).
The partial sequence of the CRIT homolog in Gadus morhua (Cod-CRIT), and full-length sequences of CRIT in Trypanosoma cruzi (Tc)-CRIT, rat (Ra)-CRIT, and Hu-CRIT were obtained by PCR, the latter from testis and pancreas cDNA. Alignment at the amino acid level (Fig. 1,B) reveals near identical TM domains and ligand-binding ed1 in all species, the latter accounting for the cross-immunoreactivity, between schistosomes and humans in ed1. The high interspecies homology at 82% between full-length Hu- and Ra-CRIT also occurs in the G protein-coupled receptor (GPCR) family which can show 85–98% identity between species (22). At the nucleotide level, as shown in the phylogenetic tree in Fig. 1,C, the parasite cluster (Sh-CRIT and Tc-CRIT) is as closely related to Hu-CRIT as rat, suggesting that these human parasites acquired the CRIT gene from their host. The appropriation of host genes by parasites accounts for the many genes with high degrees of identity between parasite (especially schistosome) and host (23, 24), and such molecular mimicry in which host proteins are expressed by parasites helps disguise the parasite as “self” in terms of the host immune system. We found the nucleotide sequence of the Hu-CRIT cDNA to be identical with that of the genomic sequence, implying that the ORF of Hu-CRIT is encoded by a single exon, although we cannot yet rule out an intron in the untranslated regions. In this study, we present the Hu-CRIT gene structure only as the predicted ORF based on that of Schistosoma CRIT, thus excluding 5′ and 3′ noncoding regions and other structures such as promoters and transcription regulatory elements. The presence of a CRIT gene in cod of which we present a partial sequence in Fig. 1 B, indicates that CRIT genes may have evolved from a common ancestral gene, at least present in the earliest teleosts.
CRIT membrane topology and sequence motifs
Fig. 1,D shows how, for five of six topology algorithms, Hu-CRIT, representative of the other CRIT sequences, is expected to lie in the membrane. Empirical evidence for the internal/external orientation of CRIT has been described before (9) and is supported by the fluorescence staining of CRIT-positive cells incubated with anti-CRIT-ed1 (Fig. 2,A). The majority of substitutions between the species occur within the first 60 residues of the 161-residue cytoplasmic tail, id2, proximal to the third TM domain. Within the remainder of the cytoplasmic tail, as illustrated in Fig. 1,D, and shared by parasite and mammalian sequences, are three predicted tyrosine-based YXXØ endocytosis motifs (where X is any amino acid and Ø is a strongly hydrophobic amino acid), two di-leucine endocytosis motifs, and, illustrated schematically in Fig. 1 D, an amphipathic α helix (putative sorting signal toward degradation) (25). Also within the cytoplasmic tail, there is a consensus binding motif for the Src homology 2 domain of the Src family cytoplasmic tyrosine kinases (26), in particular, Fyn, Fps/Fes, and Syk, as well as consensus phosphorylation sites for various tyrosine kinases.
CRIT forms covalently linked dimers
Having first detected CRIT receptor by flow cytometry on the surface of Jurkat cells (Fig. 2 A), we found, by immunoblotting of a cell lysate, that CRIT exists as a dimer under nonreducing conditions (B). Comparison of a nonreduced Jurkat extract (NR, lane 1) with a reduced extract (R, lane 2) shows, under reducing conditions, the disappearance of the dimeric form and a slight increase in intensity (which was frequently more prominent) of the monomeric form. The presence of the sulfhydryl trapping agent iodoacetamide (10 mM) in the cell lysis buffer did not prevent the appearance of the covalently linked dimeric CRIT (not shown), meaning that disulfides were not formed artifactually as a result of prevailing non-denaturing and oxidizing conditions during cell lysis. At this stage, we cannot categorically say whether homo- or heterodimerization is occurring. However, because dimers occurred in Jurkat cells that do not make C2, regardless of whether they were maintained in complete or serum-free medium, it appears that the dimerization is constitutive and not dependent on ligand. Furthermore, heterodimerization could conceivably occur between CRIT and a molecule other than C2. To resolve this, studies are ongoing to cotransfect cells with CRIT possessing different tags to see whether (homo-)dimers can be detected with both tag-specific Abs. If homodimerization is occurring, then disulfide bonds could be formed between any of the three cysteines in TM1, one each in TMs 2 and 3 and three in the cytoplasmic tail.
CRIT cellular and tissue distribution
We expressed Hu-CRIT both in E. coli (Fig. 2,C) and by in vitro transcription translation (D) as a ∼31- to 32-kDa protein. Full-length CRIT1–280 was also detected in Jurkat cell lysates using anti-ed1 or anti-ed2 as well as an Ab, anti-id2, against the cytoplasmic tail of CRIT, specifically against a peptide based on the region K195–A208. Anti-ed1 also revealed the rat CRIT homolog as a ∼31- to 32-kDa protein in the rat T cell line, C58(NT). Comparing CRIT sequences available, the predicted size of human CRIT is not dissimilar to CRIT from T. cruzi, S. hematobium/S. mansoni/S. japonicum, or rat. Compared with CRIT monomer electrophoresed under nonreducing conditions (Fig. 2 D), 1 mM DTT decreased the electrophoretic mobility of CRIT monomer (as viewed on an 8% SDS-PAGE gel to maximize the resolution of proteins between 20 and 40 kDa), indicating a likely reduction of an intramolecular disulfide bond.
Western blotting (Fig. 2,E) showed expression of CRIT in various human hemopoietic cell lines including Jurkat (T lymphocyte), Raji (B lymphocyte), THP-1 (myeloid), U937 (myeloid), and ECV304 (endothelial), as well as several nonleukocyte cells (Table I) such as T47D (epithelial breast ductal) and weakly on HeLaS3 (epithelial) and MRC5 (fibroblast). In addition, CRIT was detected by immunoblotting in human platelets, monocytes, dendritic cells, and lymphocytes, but not in erythrocytes, neutrophils, nor HS1 (liver fibroblast-like cell line). CRIT was also found (Fig. 2,E and Table I) in testis, tonsil, kidney, thymus, and liver (latter not shown), but not in colon.
|Cell Type/Tissue .||CRIT .|
|T47D (epithelial breast ductal carcinoma)||+|
|MRC5 (fibroblast, passage 31)||+/−|
|HS1 (liver fibroblasts)||−|
|HeLa S3 (cervix uteri epithelial carcinoma)||+/−|
|Cell Type/Tissue .||CRIT .|
|T47D (epithelial breast ductal carcinoma)||+|
|MRC5 (fibroblast, passage 31)||+/−|
|HS1 (liver fibroblasts)||−|
|HeLa S3 (cervix uteri epithelial carcinoma)||+/−|
Localization of CRIT in 38 normal, fresh (biopsy material) human tissues was determined by immunohistochemistry (summarized in Fig. 3) using anti-CRIT-ed2, which unlike anti-CRIT-ed1, worked well in staining of paraffin-embedded sections. We were able to establish the following distribution for the CRIT receptor: CRIT was found to be strongly to moderately positive in almost all tissues in smooth muscle cells (SMCs) and related cells (pericytes) in vessels. According to the literature, complement regulators DAF, CD59, membrane cofactor protein (MCP), and Crry are expressed at the mRNA level by vascular SMCs in rat (27), whereas C3 and C4 are found in human SMCs (28). Interestingly, CRIT stained strongly in glandular epithelial cells in the endometrium (Fig. 3,A) during the proliferating but not secretory phase. Likewise, MCP, DAF, and CD59 are expressed on an endometrial epithelial cell line (29) and, together with complement receptor 1 (CR1), expressed in normal endometrial tissue in all phases of the menstrual cycle (30). Other CRIT-positive cells include pancreatic islet cells (Fig. 3,B) (which, according to their percentage and distribution, are suggestively insulin-producing β cells), cells which also express CD59 and MCP (31). Keratinocytes in the esophagus (which express DAF, MCP, and CD59) (32) also stained positive for CRIT (Fig. 3,C). In the esophagus, SMCs were also clearly CRIT positive (Fig. 3,D). In the kidney, anti-CRIT-ed2 showed a strong reaction with podocytes in kidney glomeruli (Fig. 3,E) (as for CR1) (33) as well as endothelia and SMCs of blood vessels (F) but not in fibrocytes. However, DAF, MCP, and CD59 are expressed on glomerular epithelial, endothelial, and mesangial cells (34). In the testis (Fig. 3, G and H), there was a positive reaction of scattered and partly elongated cells attached to the basement membrane of the testicular tubules, which most likely are Sertoli cells. CRIT was also found in the myoepithelia of some exocrine glands (breast) and in stromal cells in breast and prostate gland. Finally, in the placenta, there were groups of decidual cells weakly positive for CRIT.
CRIT binds complement C2
We previously demonstrated C2 binding to Sh-CRIT-ed1 by affinity purification of C2 from NHS using the Sh-CRIT-ed1 peptide (19). Furthermore, by using C2 hemolytic assays in which limiting amounts of C2 (just sufficient to restore complement activity) were added to C2-deficient serum, we showed that preincubation of C2 with CRIT-ed1 peptide inhibited complement activation (10). We now confirm the C2 interaction with Hu-CRIT by ligand blotting. For this, total protein lysates of Jurkat cells, immunoblotted onto nitrocellulose, were probed with C2bio (Fig. 4,A). C2, like anti-ed1, bound to monomeric and dimeric CRIT. C2 still bound when preincubated with either CRIT-ed2 peptide or CRIT-H17S peptide. However, CRIT-ed1 peptide abrogated C2 binding. Identical results were obtained with 125I-labeled ligand. In work to be reported in detail elsewhere, we showed that a recombinant von Willebrand factor A (vWFA) domain of C2, binds CRIT via ed1, and that this binding can be blocked by prior incubation of vWFA with the C2 mAb Hyb-5050 (35). In this study, we show that the vWFA domain of C2 expressed and purified from E. coli, binds in vitro-expressed CRIT. After stabilizing the protein interaction by cross-linking, both unbound CRIT and CRIT-vWFA C2 are detected by anti-CRIT-ed1 in Western blots (Fig. 4,B). The binding of biotinylated C2 or FITC-labeled C2 (not shown) to Jurkat cells via Hu-CRIT-ed1 was also shown by FACS analysis (Fig. 4 C) (and immunofluorescence microscopy). Binding could be eliminated by prior incubation of the cells with anti-CRIT-ed1 blocking Ab. Factor B and serum albumin showed no binding to the Jurkat cell surface.
The dependence on C2bio concentration for the ELISA signal of C2bio interacting with C4b or CRIT-ed1 and that this interaction is a saturable phenomenon are shown in the graph of binding data or A490 (y) as a function of ligand concentration (X) (Fig. 4,D). Analysis of the results indicated that the one-binding class model was a better fit than the two-binding class model. This dose dependency and saturable binding is also shown in the graph inset. Nonlinear regression of the equation describing one-site binding of ligand to receptor that follows the law of mass action (y = BmaxX/KD(app) + X (Equation 1), where Bmax is the maximal binding, and KD(app) is the concentration of ligand required to reach half-maximal binding) onto the data shown in Fig. 4 D gave an estimate for KD(app) of 0.0152 μM for C4b/C2bio. This is in close agreement with the KD estimate of 0.015 μM for C4b/C2 by plasmon surface resonance (36) and indicates that the biotinylation of C2 had not compromised its protein-binding ability. In our study, the KD from a Scatchard analysis plot of concentration of C2 bound (vbound) against concentration of free C2, was not calculated.
Although we have no data for the affinity of C2 for native CRIT, we have been able to estimate the affinity for the ligand-binding region, CRIT-ed1 in the form of a synthetic peptide. Comparison of data for C2bio binding CRIT-ed1 or C4b (Fig. 4,D) indicates that an ELISA reading for CRIT-ed1/C2bio binding ∼3-fold lower than for C4b/C2bio corresponds to a KD(app) for CRIT-ed1/C2bio (0.022 μM) 2-fold lower than between C4b/C2bio (0.0152 μM) (KD of 0.015 μM (36)) (Table II). According to our estimations, C2 has an equal to 2-fold lower affinity for CRIT than for C4b and therefore could not prevent assembly of the CP C3 convertase but is more likely reducing excessive activation on a target. The concentration of C2 (ligand) in plasma is 0.25 μM. This is ∼10 times the average (using two different methods; Table II) equilibrium dissociation constant of C2-CRIT receptor complexes, KD(app) of 0.0235 μM, implying (Equation 1) that C2 will occupy 90% of CRIT receptors at equilibrium.
|Assay .||C2-Interacting Protein .||.||.|
|.||CRIT-ed1 .||CRIT-H17 .||C4β212–232 .|
|Surface plasmon resonance (36 )||0.015 (C4b/C2) (KD)|
|Assay .||C2-Interacting Protein .||.||.|
|.||CRIT-ed1 .||CRIT-H17 .||C4β212–232 .|
|Surface plasmon resonance (36 )||0.015 (C4b/C2) (KD)|
Micromolar concentration. Mean of four independent experiments.
Again, maximum blockage of the C4b/C2bio interaction was possible (Fig. 4,E) by preincubation of C2bio with a concentration of ed1 (1.2 μM) ∼80 times that of the KD for C4b/C2 (0.015 μM) and ∼51 times the KD for CRIT-ed1/C2 (0.0235 μM). This is in agreement with predictions for ligand/receptor associations that the ligand concentration needs to be 99 times the KD to occupy 99% of the receptors at equilibrium. C4b/C2bio binding could not be inhibited by preincubating C2bio with ed2 (Fig. 4 E).
As an additional test of the binding of C2 to CRIT-ed1, it was possible to inhibit the binding of C2 to C4 in the assembly of the CP C3 convertase with increasing concentrations of CRIT- or C4-based peptides. Plotting percent inhibition of hemolysis, as a measure of the extent of binding of CRIT- or C4-based peptides to C2, resulted in Fig. 4,F. Similar nonlinear regression calculations of the data gave estimates of KD(app) for CRIT-ed1/C2 of 0.025 μM and C4β212–232/C2 (where C4β212–232 is the C4 β-chain peptide described before (10)) of 0.011 μM. These data are in good agreement with other empirical data summarized in Table II, and are not compromised by the many proteins in NHS, because the CRIT-ed1/C2 vs C4b/C2 interactions are highly specific. CRIT-H17 representing the 11-aa C-terminal part of CRIT-ed1 has a 10-fold order of magnitude greater affinity for C2 than CRIT-ed1. Interestingly, in preliminary unreported results before the study of Inal and Schifferli (10), CRIT-H17 alone, but not CRIT-ed1, gave significant reductions in various parameters of complement-mediated inflammation in the reversed passive Arthus reaction in mice (11). In all current experiments, C2 binding to Hu-CRIT-ed1 was not Mg2+ dependent.
CRIT inhibits CP-mediated complement cytolysis
To ascertain whether CRIT is able to protect cells from complement-mediated attack, cells sensitized with Abs against human lymphocytes (or whole human serum) were incubated with increasing concentrations of anti-CRIT Ab, before exposure to NHS as a source of complement. A titration was first conducted with increasing NHS (Fig. 5,A) to show that lysis was being achieved with this system and to find a convenient serum concentration (chosen to be 10%) to be able to see easily any increase in percent lysis of cells. The results (Fig. 5,B) show a significant increase, from 8% lysis without Ab to 28% lysis with 12.5 μg/ml anti-CRIT-ed1 in the promonocytic cell line, U937. With the macrophage-like THP-1 cells (Fig. 5,C), and monocytes (D), there were significant increases from 10 and 15% lysis, respectively, to 43 and 52%, respectively, with 10 μg/ml. As noted previously, macrophages are negative for CRIT (see Table I). Monocytes treated with IFN-γ, thereby expressing lower levels of CRIT, showed a nonsignificant increase in cytolysis from ∼20% without Ab to 34% with 10 μg/ml anti-CRIT-ed1. To control for a possible activation of complement through the presentation of additionally bound anti-CRIT and so additional Fc by the neutralizing anti-CRIT Ab, we incubated cells with increasing anti-CRIT-ed1 but without prior sensitization with the complement-fixing anti-lymphocyte serum. Under these conditions shown in Fig. 5, B and C (▴), there was no increase in cytolysis with increasing anti-CRIT-ed1 concentration, indicating that the blocking Ab did not activate complement.
C2 bound to CRIT is rendered resistant to C1s cleavage
We and others have shown previously that CRIT-ed1 inhibits the complement-mediated lysis of sheep erythrocytes (10, 19, 20, 21). Having also shown that cleavage of C2 by C1s was inhibited in the presence of ed1 peptide, we now wanted to see whether this inhibition could be provided by native CRIT as found on the surface of Jurkat cells. As illustrated in Fig. 5 E, CRIT-ed1 peptide completely inhibited C1s cleavage of C2. This is in contrast to cleavage to C2a and C2b by C1s in the presence of sham peptide (C4βS). Interestingly, in the presence of Jurkat cells (and therefore CRIT), there is almost no cleavage of C2. When the receptor is blocked with anti-CRIT-ed1, the inhibition by CRIT of C1s-mediated cleavage of C2 is lifted.
The complement system is a rapidly activated and self-amplifying system that, to avoid extensive consumption and self-depletion, requires control at several levels. Interestingly, there is often symmetry of function between inhibitors of complement in the fluid phase and on cell surfaces (37), although there is an exception: C1-INH is a major regulator of C4 and C2 cleavage in the fluid phase, preventing excessive formation of the C3 (C4b2a) convertase.
Hu-CRIT was cloned from both genomic and cDNA. Thus, although the 5′ and 3′ untranslated regions may yet be found to possess introns, the high level sequence identity suggests that the coding region of Hu-CRIT could be coded for within a single exon and therefore be intronless, as for example are many GPCR genes. Whether the entire CRIT gene is shown to possess introns or not, we have presented evidence that it encodes a functional receptor (of a size equivalent to that predicted from the ORF) that has affinity for the human complement serum protein, C2. Using an Ab raised against the N-terminal extracellular domain, CRIT was found on the plasma membrane in a wide variety of human cell types. We show that the Schistosoma (and Trypanosoma) parasite and its human host share a receptor for C2, and we postulate that they share a complement regulatory function. The high sequence homology between the mammalian (human and rat)/parasite (Schistosoma and Trypanosoma) and cod CRIT genes is also indicative that they may have evolved from a common ancestral gene.
CRIT regulates the formation of the CP C3 convertase by reducing the association of C2 with C4b. The KD(app) of CRIT-ed1/C2 is of the same order as the KD for C4b/C2 (0.015 μM) (36) and also CR1 for C3b dimers (0.01–0.02 μM) (38, 39). Overall, CRIT-ed1 appears to compete with C4b for binding to C2 with an apparent moderate affinity, but we do not yet know whether CRIT accelerates the decay of the C4b2a convertase. In this study, we have found that anti-CRIT-ed1 blockage of the CRIT receptor increases the complement-mediated killing of two human myeloid cell lines (U937 and THP-1) and of monocytes. However, the relative contribution of CRIT to protecting against complement-dependent cytotoxicity in the THP-1 cells, for example, cannot be known, because certain other complement regulatory proteins are expressed, such as CD59 (40). Once C2 is bound to CRIT (unlike C2 bound to C4b), it cannot be cleaved by C1s to yield C2a and C2b and thus no longer partakes in C3 convertase formation. Of the complement regulators that control the C3 convertase (C4b2a) and C5 convertase (C4b2a3b), three of them, DAF, CR1, and C4BP, compete with C2 for binding to C4b and also bind to C3 or C5 convertases, causing C2 dissociation. These proteins are composed mainly or only of complement control protein repeats, which contain the C4b (and C3b) binding sites and are all encoded by genes in the RCA cluster on chromosome 1q31–32. CRIT, like CD59, is another member of the complement control family of proteins that lacks structural resemblance to the RCA group of proteins.
In humans, CRIT is widely distributed. It is present on many hemopoietic cells, but not neutrophils, and notably, like MCP, not on erythrocytes. CRIT is found in many tissues, and interestingly, like DAF (41), its expression appears to be hormonally regulated during the menstrual cycle. CRIT is also found on endothelial and epithelial cells. Besides its complement-regulating function concomitant with its wide distribution in the body, we speculate that its particular expression in highly differentiated cell types, such as glomerular podocytes (42), Sertoli cells (43), and keratinocytes (44, 45), could suggest additional roles.
Complement results in the release of the anaphylatoxins C3a and C5a, which in turn bring about the proinflammatory responses of complement, including an increase in vascular permeability, mast cell degranulation, and smooth muscle contraction. C5a is then able to further increase the proinflammatory response by inducing chemotaxis of macrophages and neutrophils and by the activation of neutrophils. These inflammatory responses are a natural reaction of host tissue to injury. However, an unregulated inflammatory response may injure host tissues and is important in many complement-mediated diseases (46). As complement and its activation products, in particular, C5a, can aggregate, stimulate, and cause degranulation of neutrophils, any anti-inflammatory agents should, in certain diseases, preferably interrupt the complement cascade at an early stage. In myocardial infarction, for example, it might be important to block CP and/or lectin pathway but leave the AP and lytic pathway intact. By competing with C4b for binding to C2, and inhibiting its activation, CRIT prevents the formation of the CP C3 convertase and is thus an obvious target for diminishing the proinflammatory response of unregulated complement activation. A soluble form of CRIT in the form of the CRIT-ed1 peptide has been shown to inhibit CP-mediated hemolysis by human serum. Indeed, recombinant soluble forms of the membrane complement inhibitors have been proven to inhibit complement-mediated inflammation (46). If soluble CRIT in the form of the CRIT-ed1 peptide or the smaller CRIT-H17 derivative peptide is not antigenic in humans, then it could potentially be used as a therapeutic agent against certain complement-mediated human diseases. The first promising tests of the in vivo efficacy of CRIT-H17 in blocking immune complex-mediated tissue injury in the classical inflammation model, the reversed passive Arthus reaction (11), and in Forsman shock (20) should be further addressed.
It is well documented that conformational changes can occur within the secondary structure of a receptor upon reduction (47, 48). Essentially, reduction of the intramolecular disulfide bond(s) of a protein, as we believe occurs with CRIT, results in its unfolding from a globular state to a more rod-like conformation, which migrates more slowly in an SDS-PAGE gel. Whether CRIT folding or dimerization affects its ability to bind ligand, as for example with the TNFR, BAFF-R (49), will be the subject of future work.
If the dimerization of CRIT occurs through the formation of intermolecular disulfide bonds then candidate cysteine residues would be Cys31,40,41 of TM1, Cys69 of TM2, Cys110 of TM3, or Cys126,214,280 of the cytoplasmic tail. Disulfide linkages are not usually found in the cytoplasm because of the reducing environment there. Because there are no cysteines in the extracellular domains, it seems probable that the cysteines in the TM domains are involved in dimerization. Disulfide-bonded dimers, mediated through TM domains have been reported for some receptors such as CD44 (50) and the insulin receptor (51). Although reducing agents can dissociate receptor homodimers suggesting disulfide-bonded dimerization, other intermolecular hydrophobic interactions, between TM domains, have been implied, at least for the GPCRs (52). We previously noted that the C2-binding CRIT-H17 motif [F/H]EVKX4/5P, which in the C4β chain occurs contiguously, separated by a β turn, in CRIT, occurs only once, and also speculated that CRIT might be found to exist as a dimer (10). Our current study has shown Hu-CRIT to form dimers (and that C2 binds to both CRIT monomers and dimers). If these dimers are shown to be homodimers, then it is conceivable that two CRIT-H17 motifs (forming part of two ed1 regions) could be brought into close proximity for C2 binding, as in the C4β chain. Whether dimeric CRIT binds C2 with greater affinity and represents the functional form of the receptor we cannot say, but already the work of Oh et al. (20) indicates that a recombinant homodimeric CRIT-H17 peptide has an increased binding capacity for C2 and 5-fold higher anti-complementary activity.
In addition to genetic abrogation and mutagenesis experiments in cells, which together with structural studies are revealing details of the CRIT-ed1 interaction with C2, it is envisaged that CRIT gene knockout experiments will place the relative contribution of CRIT toward complement regulation in context with other CP regulators.
At its inception, this work benefited from stimulating discussions with David Meyer, Andrew Falconar, and Robert Sim.
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 U.S.C. Section 1734 solely to indicate this fact.
This work was supported by the Swiss National Foundation, the Novartis Research Foundation, the Stiftung für Medizinische und Biologische Forschung, and the Icelandic Research Council.
The complement C2 receptor inhibitor trispanning (CRIT) sequence data are available from European Molecular Biology Laboratory/GenBank under accession nos. AY464186 (Homo sapiens), Q9JJ42 (Rattus norvegicus), AY464185 (Trypanosoma cruzi), Q9BLM6 (Schistosoma haematobium), and AY464184 (Gadus morhua).
Abbreviations used in this paper: RCA, regulator of complement activation; C1-INH, C1 inhibitor; CP, classical pathway; CRIT, complement C2 receptor inhibitor trispanning; DAF, decay-accelerating factor; AP, alternative pathway; ed, extracellular domain; id, intracellular domain; H17 or CRIT-H17, the 11-aa C terminus of ed1; NHS, normal human serum; Sh, Schistosoma hematobium; Tc, Trypanosoma cruzi; Ra, rat; Hu, human; C2bio, biotinylated C2; KD(app), apparent dissociation constant; ORF, open reading frame; TM, transmembrane; GPCR, G protein-coupled receptor; SMC, smooth muscle cell; CR, complement receptor; MCP, membrane cofactor protein; vWFA, von Willebrand factor A.