A protein of 33 kDa (p33) that tightly binds to the globular domains of the first complement component, C1q, is thought to serve as the major C1q receptor (gC1qR) on B cells, neutrophils, and mast cells. However, the cellular routing and the subcellular localization of p33/gC1qR are unknown. We have performed confocal laser-scanning microscopy and found that p33/gC1qR is present in intracellular compartments, where it colocalizes with the mitochondrial marker protein, pyruvate dehydrogenase. No surface staining for p33/gC1qR on endothelial EA.hy926 cells was observed. A fusion protein of the p33/gC1qR presequence with green fluorescent protein translocated to the mitochondria of transfected COS-7 cells. Concomitantly, a 6-kDa portion of the fusion protein was proteolytically removed. The 33 amino-terminal residues of the presequence proved sufficient to direct reporter constructs to mitochondria. Association of p33/gC1qR with mitoplasts indicated that the mature protein of 209 residues resides in the matrix and/or the inner membrane of mitochondria. Immunocytochemistry of fetal mice tissues revealed a ubiquitous expression of p33/gC1qR, most prominently in tissues that are rich in mitochondria. Thus, the candidate complement receptor p33/gC1qR of intact cells cannot interact with plasma C1q due to mutually exclusive localizations of the components. The functional role of p33/gC1qR needs to be reconsidered.

The complement system cooperates with Igs in defending vertebrates against bacterial infections. It is made up by a set of mostly functionally dormant plasma proteins that are activated in an amplifying series of reactions by the classical or by the alternative pathway. The cellular proteins that bind complement factors to the outer cell membranes are refered to as complement receptors. In the classical pathway, the assembly of complement factors is triggered by the first component and its subcomponents, i.e., C1q, C1r, and C1s. Among these, C1q is unique in that it has a globular protein head and a collagen-like tail.

Three types of complement receptors that bind to C1q have been described to date; they differ by their ligand recognition specificity: C1qRp, a high molecular mass receptor of 126 kDa (1); cC1qR3 of 60 kDa, which binds to the collagenous portion of C1q (2); and a cell-associated protein of 33 kDa (p33) that interacts with the globular heads of C1q (3), and therefore has been dubbed gC1q receptor (gC1qR). Detailed studies have demonstrated that p33/gC1qR readily complexes with plasma proteins such as H-kininogen (4), factor XII (5), vitronectin (6), thrombin, prothrombin (7), and cC1qR (8). In vitro binding studies revealed that the binding to p33/gC1qR is specific and saturable with apparent dissociation constants (Kd) ranging from 9 nM for H-kininogen to 240 nM for C1q (4).

To serve its postulated function as a cellular C1q-binding protein, p33/gC1qR must be expressed on the cell surface. The mature form of the protein has been isolated and sequenced from total cell extracts of B cells; it comprises 209 amino acid residues and represents a highly acidic protein (28 Glu and 20 Asp residues) that is devoid of a typical hydrophobic transmembrane-spanning region (9). The corresponding cDNA sequence predicts a preprotein of p33/gC1qR that contains additional 73 residues at the amino terminus, and it has been speculated that this presequence might anchor gC1qR to the cell membrane (7, 9). Indeed, immunofluorescence studies have claimed the surface expression of p33/gC1qR for B cells, mast cells, neutrophils, platelets, and endothelial cells (9, 10, 11, 12, 13); however, recent studies with B cells, macrophages, and endothelial cells have challenged these findings (14, 15). Daha and coworkers (14) demonstrated that gC1qR is located in the cytoplasm of B cells, from which it can be released under stress conditions, whereas we have demonstrated that gC1qR is associated with the vesicular fraction of endothelial cells (15). However, the precise cellular location of gC1qR has remained elusive.

To analyze the subcellular distribution of p33/gC1qR and to follow its intracellular routing, we have overexpressed the protein in Sf9 cells. Unexpectedly, we were unable to detect p33/gC1qR on the cell surface of infected Sf9 cells. Rather, we determined that p33/gC1qR was associated mainly with the intracellular particulate fraction. Using fusion constructs with green fluorescent protein (GFP), we demonstrated that the amino-terminal portion of p33/gC1qR represents a targeting sequence that directs the protein to mitochondria. We conclude that p33/gC1qR cannot function as a binding protein for C1q in intact cells.

Materials were purchased from the following manufacturers: HUVEC cDNA library and anti-GFP Ab from Clontech Laboratories (Palo Alto, CA); rabbit reticulocyte lysate from Promega Corp. (Madison, WI); PCR primers from MWG Biotech (Ebersberg, Germany); pAlpha + GFP from Maxygen (Santa Clara, CA); pVL1392 vector and liposome kit from Invitrogen BV (Leek, The Netherlands); standard marker proteins from Pharmacia Biotech (Uppsala, Sweden); [α-35S]dATP and enhanced chemiluminescence (ECL) kit from Amersham Buchler (Braunschweig, Germany); [35S]cysteine/methionine (Tran35S label) and [α-32P]dATP from ICN Biomedicals (Eschwege, Germany); and rhodamine 123 from Molecular Probes Europe BV (Leiden, The Netherlands). H-kininogen was isolated from human plasma (16).

The cloning of the nucleotide sequence encoding the mature form of p33/gC1qR (positions 74 to 282 of the protein sequence) from total RNA of HUVEC by reverse transcriptase-PCR was done as described (4). Using this PCR fragment as a probe, we screened a HUVEC cDNA library and retrieved a full-length p33/gC1qR cDNA clone. The p33/gC1qR cDNA was subcloned into the Bluescript KS+ (Bks) vector and sequenced on both strands. We transcribed p33/gC1qR sense RNA using T3 RNA polymerase and the Bks-p33/gC1qR vector that had been linearized with XbaI. For transcription of a p33/gC1qR antisense RNA, the vector was cut with EcoRI, and the T7 RNA polymerase was used. In vitro translation of 1 μg of p33/gC1qR sense and antisense RNA, respectively, was conducted using a commercial reticulocyte lysate (Promega Corp.). The polypeptides were labeled with [35S]cysteine/methionine.

To generate fusion proteins of GFP and p33/gC1qR, various parts of the p33/gC1qR DNA sequence were amplified by PCR using primers with flanking XbaI restriction sites. The linearized Bks-p33/gC1qR plasmid (3 ng/5 μl) was added to 95 μl of a polymerase mixture that contained 2 U Taq polymerase, 25 pmol each of the 5′ and 3′ primers, 250 μM concentrations of each dNTP, 10 mM Tris-HCl, pH 8.3, 50 mM KCl, and 1.5 mM MgCl2. The individual samples were overlaid with 70 μl of mineral oil and amplified in a thermal cycler for 45 s at 94°C, 45 s at 58°C, and 1 min at 72°C for 40 cycles. Isolated PCR fragments were digested with XbaI and ligated into the XbaI site of pAlpha + GFP vector, thus generating GFP fusions with variable amino-terminal extensions of p33/gC1qR.

The following p33/gC1qR-GFP constructs were generated: [p331–33] GFP, positions 1–33 of the p33/gC1qR protein sequence, amplified with the primers 5′-GCATCTAGACGTGTTCGCAGTCGTTTCC-3′ and 5′-GGGTCTAGACTGCAGGAGCTGCCGGAAAGGCG-3′; [p331–81]GFP, positions 1–81 of p33/gC1qR amplified with the primers 5′-GCATCTAGACGTGTTCGCAGTCGTTTCC-3′ and 5′-GGAAATCTAGAAGCTTTGTCTCCGTCGGTGTGC-3′; [p3332–81]GFP, positions 32–81 amplified with the primers 5′-CCTTCTAGAATGGCTCTGCAGCCGGCACCCCGGCTGTGC-3′ and 5′-GGAAATCTAGAAGCTTTGTCTCCGTCGGTGTGC-3′; [p331–282]GFP, full-length p33/gC1qR precursor amplified with the primers 5′-GCATCTAGACGTGTTCGCAGTCGTTTCC-3′ and 5′-GTCTTCTAGACTGGCTCTTGACAAAACTCTTGAGG-3′; and [p3374–282] GFP, maturated form of p33/gC1qR amplified with the primers 5′-CTGTTCTAGAATGGCTCTGCACACCGACGGAGACAAAGC-3′ and 5′-GTCTTCTAGACTGGCTCTTGACAAAACTCTTGAGG-3′.

GFP fusions, with mitochondrial targeting sequences of human cytochrome c1 (cc1) (17) and human Rieske iron-sulfur protein (rfs) (18), were generated by reverse-transcriptase PCR using total RNA from EA.hy926 cells. The following constructs were made: [cc11–88]GFP, amino acid positions 1–88 of the cytochrome c1 sequence, amplified with primers 5′-GGCTCTAGAGAGGCCAAGATGGCGGCAGCTGCG-3′ and 5′-GGGTCTAGACTCCAGGTCACTGGCACTCACAGC-3′; [rfs1–86]GFP, positions 1–86 of rfs, amplified with the primers 5′-GGATCTAGAGTCGCCATGTTGTCGGTAGCAGCC-3′ and 5′-GTATCTAGAGAATTCAGGCACCTTGATGTCTGTG-3′.

The Spodoptera frugiperda (Sf9) cell line was grown as monolayers in ambient atmosphere at 27°C in TC100 medium (Life Technologies, Eggenstein, Germany) supplemented with 10% (v/v) FCS, penicillin (100 IU/ml), and streptomycin (100 μg/ml). EA.hy926 cells (19) were cultured in DMEM containing 4.5 g/L glucose, 10% (v/v) FCS, 100 μM hypoxanthine, 0.4 μM aminopterin, 16 μM TdR, and penicillin/streptomycin in a humidified 5% CO2 atmosphere at 37°C. COS-7 cells were kept under the same conditions, except that RPMI medium with 10% (v/v) FCS and penicillin/streptomycin were used. The organelle fraction of EA.hy926 cells was prepared as described (15).

The full-length p33/gC1qR cDNA was cloned into the EcoRI/XbaI restriction sites of the baculovirus transfer vector pVL1392 (Invitrogen). Recombinant baculoviruses were generated by cotransfection of Sf9 cells with the pVL1392-p33/gC1qR construct and with linearized AcMINPV DNA by the lipofection method, according to the manufacturer’s instructions (Invitrogen). Briefly, Sf9 cells grown to near confluence on a 10-cm dish were washed with PBS (PBS = 0.15 M NaCl, 10 mM sodium phosphate, pH 7.4) and starved for 2 h in TC100 medium lacking supplements and FCS. Recombinant pVL1392-p33/gC1qR plasmid (400 ng), 100 ng linearized AcMINPV DNA, 1 ml TC100 medium without supplements or FCS, and 25 μl cationic liposomes were mixed (“transfection mixture”) and incubated for 20 min at room temperature. The medium was removed from the starved Sf9 cells, and the transfection mixture was added. Following incubation for 4 h at room temperature, 1 ml of complete TC100 medium was added. Cells were incubated at 27°C until morphologic changes due to infection appeared (typically after 3–5 days). Individual baculovirus clones were established by limited dilution of a high titer virus stock and infection of Sf9 cells grown in microwell plates. Baculovirus clones were identified by immunoblotting for expressed p33/gC1qR protein. For bulk production, Sf9 cells were infected with recombinant viruses at a multiplicity of infection of 2 to 5. The cells were harvested 40 to 96 h after infection and used to prepare recombinant p33/gC1qR protein.

H-kininogen from human plasma was radioiodinated as previously described (20). Infected Sf9 cells were washed twice in 0.135 M NaCl, 2.7 mM KCl, 11.9 mM NaHCO3, 0.36 mM NaH2PO4, 14.7 mM HEPES, pH 7.35 (HEPES-Tyrode’s buffer), containing 3.5 mg/ml dextrose, and 50 μM ZnCl2, and suspended in the same buffer including 0.2% BSA. The cells (106/100 μl) were mixed gently for 1 h at 4°C with 10 nM of 125I-H-kininogen in the presence or absence of a 100-fold molar excess of unlabeled H-kininogen and recombinant p33/gC1qR, respectively. Bound 125I-H-kininogen was separated from the free ligand by centrifugation (10,000 × g, 4 min, 4°C) of the cell suspension through 500 μl of a mixture of dibutylphthalate/1-bis(2-ethylhexyl)phthalate (1.1/1, v/v) (21). Cell-bound radioactivity was measured in a gamma counter (Packard Instrument Company, Meriden, CT).

For the indirect ELISA, microtiter plates (MaxiSorb, Nunc, Wiesbaden, Germany) were incubated overnight with 2 μg/ml of H-kininogen, C1q, factor XII, and α2-HS glycoprotein, respectively, in 100 mM sodium acetate, 100 mM NaCl, pH 5.5. The plates were washed with PBS and incubated with serial dilutions (2n; starting concentration 500 μg/ml) of total cellular extracts of infected Sf9 cells. Bound protein was detected by an Ab to p33/gC1qR (0.5 μg/ml), followed by a secondary peroxidase-conjugated Ab to rabbit Ig (Bio-Rad Laboratories, Hercules, CA; 1:5000), and the chromogenic substrate 2,2-azino-bis(3-ethyl-2, 3-dihydrobenzthiazoline-6-sulfonate) (ABTS) for 30 min. The change of absorbance was read at 405 nm. All incubations were done at 37°C, except for the coating step, which was done at 4°C (22).

Harvested cells were washed two times with PBS and resuspended in sample buffer (63 mM Tris-HCl, pH 6.8, containing 2.5% SDS, 5% glycerol, 5% β-mercaptoethanol, and 0.005% bromphenol blue) for 30 min at 42°C. Proteins were resolved by SDS-PAGE (12% (w/v) total acrylamide) and subsequently electrotransferred to nitrocellulose membranes using established protocols. For immunoprinting, affinity-purified rabbit Abs to the mature form of p33/gC1qR (positions 74–282; As385) recombinantly expressed in Escherichia coli (15) were used at a concentration of 0.5 μg/ml. Alternatively, Abs to synthetic peptide RLC24 derived from the targeting sequence of the p33/gC1qR precursor (positions 37–60; As423) were used at a concentration of 1 μg/ml; the anti-GFP Ab was applied at 1:5000. Bound Ab was detected by a secondary peroxidase-labeled Ab (1:5000) and the chemiluminescent substrate assay.

COS-7 cells were transfected with the fusion constructs of pAlpha + GFP and p33/gC1qR using the DEAE-dextran method. After 60 h, the transfected cells were washed twice with ice-cold PBS and fixed with 4% (v/v) formaldehyde in PBS. For membrane permeabilization, the cells were treated with ice-cold methanol for 10 min. After washing three times with PBS, the cells were incubated with 15 μg/ml anti-p33 diluted in blocking buffer (PBS containing 0.5% BSA) for 30 min at 37°C. The cells were washed three times with PBS and incubated with tetramethylrhodamine isothiocyanate (TRITC)-conjugated goat anti-rabbit Ig (Sigma Chemical Co., St. Louis, MO; 1:100 in blocking buffer) for 30 min at 37°C. After washing three times with PBS, the cells were embedded in n-propyl gallate solution (5% in glycerol) and viewed with an epifluorescence microscope (Axiophot, Zeiss, Jena, Germany). To stain for mitochondria, rhodamine 123 was dissolved in water at 1 mg/ml and diluted with culture medium to a final concentration of 10 μg/ml. EA.hy926 cells grown on chamber slides (Nunc) were incubated with the dye for 30 min in a 5% CO2 incubator at 37°C. Adherent cells were rinsed three times for 5 min each with medium. The medium was removed, and the stained cells were examined by epifluorescence microscopy.

Fixed and permeabilized EA.hy926 cells were stained for 30 min with 15 μg/ml (final concentration) of anti-p33 or with a human autoantibody directed to mitochondrial pyruvate dehydrogenase (1:1000), or both. Anti-p33 was detected by a goat anti-rabbit Ig conjugated to the fluorescent dye Cy3 (Dianova, Hamburg, Germany). The human autoantibody was visualized by FITC-conjugated goat anti-human Ig. In control stainings, no cross-reactivity of the anti-rabbit and anti-human Ab was observed (data not shown). For double labeling, the cells were first stained at room temperature with anti-p33, followed by incubation with autoantibody to pyruvate dehydrogenase. The unbound Abs were removed by washing twice with PBS for 5 min each. The cells were mounted using PBS/glycerol (1:1, v/v). Cells were sliced into horizontal optical sections at an interval of 500 nm, and examined by confocal laser-scanning microscopy (LSM10, Zeiss). Evaluation of the stored images of the horizontal optical sections was done with the LSM10 image-processing unit. For immunocytochemistry on the light-microscopy level, paraffin sections of fetal mice fixed with Bouin’s fluid were dewaxed, rinsed, and stained with 15 μg/ml anti-p33. Bound Ab was detected by the peroxidase-antiperoxidase technique and the chromogenic substrate diaminobenzidine·4HCl. Sections were lightly counterstained with Harris’ hematoxylin solution and viewed in a light microscope (Axiophot, Zeiss).

Using a cDNA fragment of 644 bp that had been generated by reverse-transcriptase PCR from total RNA of human endothelial cells (4), we screened a HUVEC cDNA library and isolated a clone of 1165 bp encoding the full-length p33/gC1qR preprotein of 282 amino acid residues. Sequence analysis established the full identity of the endothelial p33/gC1qR clone with the published gC1qR cDNA sequence from B cells (9). To estimate the size of the corresponding primary translation product, we performed in vitro transcription/translation experiments in the presence of 35S-radiolabeled cysteine/methionine (Fig. 1 A). A 39-kDa protein representing the p33/gC1qR precursor was translated from sense RNA (lane 1), but not from antisense RNA (lane 2). Immunoprinting with an anti-peptide Ab, anti-RLC24, derived from the extreme amino-terminal segment, putative presequence, of the p33/gC1qR precursor, indicated that this 39-kDa preprotein does not accumulate in the total cellular extracts of HUVEC or EA.hy926 cells to detectable amounts (data not shown).

FIGURE 1.

Expression of the p33/gC1qR precursor protein. A, In vitro transcribed RNA (1 μg) was translated in the presence of [35S]cysteine/methionine. The translation products (5 μl each) of p33/gC1qR sense RNA (lane 1) or antisense RNA (lane 2) were resolved by SDS-PAGE, followed by fluorography and autoradiography. B, Recombinant expression of p33/gC1qR precursor in Sf9 cells. Total cellular extracts of Sf9 infected with baculovirus containing the full-length p33/gC1qR-cDNA were separated by SDS-PAGE (30 μg protein per lane), blotted to nitrocellulose, probed with 0.5 μg/ml anti-p33 (lanes 3 and 4) or 1 μg/ml anti-RLC24 (lanes 5 and 6), and detected by a peroxidase-conjugated secondary Ab. As a control, Sf9 cell lysates expressing an unrelated protein were probed with anti-RLC24 (lane 7). The infection periods lasted 40 h (lanes 3 and 5) or 96 h (lanes 4, 6, and 7).

FIGURE 1.

Expression of the p33/gC1qR precursor protein. A, In vitro transcribed RNA (1 μg) was translated in the presence of [35S]cysteine/methionine. The translation products (5 μl each) of p33/gC1qR sense RNA (lane 1) or antisense RNA (lane 2) were resolved by SDS-PAGE, followed by fluorography and autoradiography. B, Recombinant expression of p33/gC1qR precursor in Sf9 cells. Total cellular extracts of Sf9 infected with baculovirus containing the full-length p33/gC1qR-cDNA were separated by SDS-PAGE (30 μg protein per lane), blotted to nitrocellulose, probed with 0.5 μg/ml anti-p33 (lanes 3 and 4) or 1 μg/ml anti-RLC24 (lanes 5 and 6), and detected by a peroxidase-conjugated secondary Ab. As a control, Sf9 cell lysates expressing an unrelated protein were probed with anti-RLC24 (lane 7). The infection periods lasted 40 h (lanes 3 and 5) or 96 h (lanes 4, 6, and 7).

Close modal

The full-length cDNA for p33/gC1qR was inserted into the baculovirus genome. Following infection of Sf9 cells, the expression of p33/gC1qR precursor was monitored by immunoblotting with anti-p33 directed to the mature form of p33/gC1qR, or with anti-RCL24 targeted to positions 37 to 60 of its presequence (see above). After 40 h of infection, a 39-kDa band was detected by both Abs (Fig. 1 B, lanes 3 and 5). Prolongation of the infection periods up to 96 h allowed the detection of additional bands at 33 kDa (major) and 35 kDa (minor) by anti-p33 (lane 4). The major 33-kDa band most likely represents the mature form of p33/gC1qR precursor because it fails to bind the anti-RCL24 (lane 6), whereas the 35-kDa protein might represent an intermediate processing product. Unrelated recombinant proteins failed to produce specific bands with anti-RCL24 (lane 7) or anti-p33 (not shown).

Isolated p33/gC1qR binds to plasma proteins such as C1q (apparent Kd = 240 nM), and even more tightly to H-kininogen (Kd = 9 nM). Therefore, we probed Sf9 cells overexpressing p33/gC1qR for the presence of newly exposed binding sites for H-kininogen (Fig. 2 A). After 40 h of infection with recombinant baculovirus, the Sf9 cells were incubated with 10 nM 125I-H-kininogen in the absence (hatched column) or presence of a 100-fold molar excess of unlabeled H-kininogen (stipled column) or p33/gC1qR (solid column). For control, Sf9 cells expressing an unrelated protein were tested under identical conditions. Even prolonged expression of p33/gC1qR up to 60 h did not increase the number of specific binding sites on the surface of Sf9 cells. Likewise, FACS studies with anti-p33 did not indicate surface expression of p33/gC1qR on infected Sf9 cells (data not shown).

FIGURE 2.

Binding sites exposed by Sf9 cells overexpressing p33/gC1qR. A, Sf9 cells (106 per assay) that had been infected for 40 h with baculovirus containing the full-length cDNA of p33/gC1qR or an unrelated protein were incubated with 10 nM 125I-H-kininogen in the absence (hatched column) or presence of 1 μM H-kininogen (stipled column) or of 1 μM p33/gC1qR (solid column). Cell-bound radioactivity was measured in a gamma counter; means ± SD of at least three independent experiments are presented. B, Microtiter plates coated with 2 μg/ml of H-kininogen (⋄), C1q (○), factor XII (▵), or α2-HS glycoprotein (⬡) were incubated with serial dilutions (2n; starting concentration 500 μg/ml) of total cell lysates of Sf9 cells overexpressing p33/gC1qR (infection period 40 h). Bound p33/gC1qR was detected with 0.5 μg/ml of anti-p33, followed by a peroxidase-labeled secondary Ab (1:5000) and ABTS. For control, extracts of Sf9 cells overexpressing an unrelated protein (endothelin receptor) were analyzed for kininogen binding (□).

FIGURE 2.

Binding sites exposed by Sf9 cells overexpressing p33/gC1qR. A, Sf9 cells (106 per assay) that had been infected for 40 h with baculovirus containing the full-length cDNA of p33/gC1qR or an unrelated protein were incubated with 10 nM 125I-H-kininogen in the absence (hatched column) or presence of 1 μM H-kininogen (stipled column) or of 1 μM p33/gC1qR (solid column). Cell-bound radioactivity was measured in a gamma counter; means ± SD of at least three independent experiments are presented. B, Microtiter plates coated with 2 μg/ml of H-kininogen (⋄), C1q (○), factor XII (▵), or α2-HS glycoprotein (⬡) were incubated with serial dilutions (2n; starting concentration 500 μg/ml) of total cell lysates of Sf9 cells overexpressing p33/gC1qR (infection period 40 h). Bound p33/gC1qR was detected with 0.5 μg/ml of anti-p33, followed by a peroxidase-labeled secondary Ab (1:5000) and ABTS. For control, extracts of Sf9 cells overexpressing an unrelated protein (endothelin receptor) were analyzed for kininogen binding (□).

Close modal

To demonstrate that infected Sf9 cells produce significant amounts of functional p33/gC1qR binding to H-kininogen, C1q, and other plasma proteins, we used the indirect ELISA. Microtiter plates coated with H-kininogen, C1q, factor XII, or α2-HS glycoprotein (negative control) were incubated with increasing concentrations of total cellular extract of Sf9 cells expressing p33/gC1qR. Figure 2,B demonstrates that p33/gC1qR from whole cell lysates of Sf9 cells infected with the virus encoding p33/gC1qR readily bound H-kininogen (Fig. 1,B, diamonds) and C1q (Fig. 1,B, circles), whereas moderate or no binding was found with factor XII (Fig. 1,B, triangles) and α2-HS glycoprotein (Fig. 1,B, pentagons), respectively. No binding of H-kininogen was observed in extracts of Sf9 cells expressing an unrelated protein (Fig. 2 B, squares). Hence, the recombinant p33/gC1qR displays ligand-binding properties as would be expected from a C1q receptor.

It has been proposed, but not demonstrated, that the presequence of 73 residues might serve as a signal sequence (amino-terminal portion) and a membrane anchor (carboxyl-terminal portion) that tethers p33/gC1qR to the cell surface (7, 9). Because we failed to detect surface expression of p33/gC1qR in Sf9 insect cells, we asked whether the presequence is involved in the intracellular targeting of p33/gC1qR in mammalian cells. To this end, we generated fusion constructs of the p33/gC1qR sequence or portions thereof (= amino-terminal part of the fusion protein) and GFP (= carboxyl-terminal part) and expressed them in COS-7 cells: [p331–282]GFP comprising the entire p33/gC1qR sequence; [p331–81]GFP carrying the complete presequence; [p331–33] GFP and [p3332–81]GFP holding the amino- and carboxyl-terminal portion, respectively, of the presequence; and [p3374–282] containing the complete sequence of the mature protein, but lacking the presequence (Fig. 3).

FIGURE 3.

GFP fusion constructs with p33/gC1qR. Top, Full-length p33/C1qR precursor of 282 residues. The segmental structure of the protein is highlighted (presequence = hatched and stipled; mature protein = solid bar). Fusion proteins: the positions of the p33/gC1qR portion preceding the GFP protein are given in brackets.

FIGURE 3.

GFP fusion constructs with p33/gC1qR. Top, Full-length p33/C1qR precursor of 282 residues. The segmental structure of the protein is highlighted (presequence = hatched and stipled; mature protein = solid bar). Fusion proteins: the positions of the p33/gC1qR portion preceding the GFP protein are given in brackets.

Close modal

COS-7 cells were transiently transfected with the constructs, fixed, permeabilized, and stained with anti-p33 to detect endogenous and recombinant p33/gC1qR. Forced expression of the construct [p3374–282]GFP, which lacks the presequence, resulted in a diffuse staining of the cytoplasm (Fig. 4,A) for GFP (green) and p33/gC1qR (red). By contrast, the construct [p331–282]GFP containing the full-length p33/gC1qR precursor was directed to intracellular compartments (Fig. 4,B). Fusion of GFP to the proper presequence, i.e., [p331–81]GFP, targeted GFP to vesicular structures, where it colocalized with endogenous p33/gC1qR of COS cells (Fig. 4,C). Truncation of the carboxyl-terminal portion of the presequence, [p331–33]GFP, did not significantly change this latter distribution pattern (Fig. 4,D), whereas the amino-terminally truncated presequence, [p3332–81]GFP, afforded a cytoplasmic localization (Fig. 4,E). For control, expression of unfused GFP was studied and resulted in diffuse staining of the cytoplasm for GFP, and a discrete vesicular staining for the endogenous p33/gC1qR (Fig. 4 F). The combined results indicate that the presequence represents a functional targeting sequence that directs the reporter protein GFP to the same intracellular compartments that harbor endogenous p33/gC1qR.

FIGURE 4.

Localization of fusion proteins and endogenous p33/gC1qR. Three days after transfection of COS-7 cells with p33/gC1qR-GFP constructs, cells were fixed with formaldehyde, permeabilized with methanol, and stained with 15 μg/ml anti-p33, followed by incubation with a TRITC-labeled secondary Ab (right panels). GFP fluorescence was monitored at 450- to 490-nm excitation, 510-nm dichroic mirror, and 520-nm barrier filter (left panels). A, [p3374–282]GFP; B, [p331–282]GFP; C, [p331–81] GFP; D, [p331–33]GFP; E, [p3332–81]GFP; and F, unfused GFP. For nomenclature, see Figure 3.

FIGURE 4.

Localization of fusion proteins and endogenous p33/gC1qR. Three days after transfection of COS-7 cells with p33/gC1qR-GFP constructs, cells were fixed with formaldehyde, permeabilized with methanol, and stained with 15 μg/ml anti-p33, followed by incubation with a TRITC-labeled secondary Ab (right panels). GFP fluorescence was monitored at 450- to 490-nm excitation, 510-nm dichroic mirror, and 520-nm barrier filter (left panels). A, [p3374–282]GFP; B, [p331–282]GFP; C, [p331–81] GFP; D, [p331–33]GFP; E, [p3332–81]GFP; and F, unfused GFP. For nomenclature, see Figure 3.

Close modal

The cellular distribution of p33/gC1qR-containing structures was reminiscent of mitochondria that typically show perinuclear and cytoplasmic location. Indeed, staining for mitochondria in EA.hy926 cells by rhodamine 123 fluorochrome (23) yielded staining patterns (not shown) that strongly resembled those produced by anti-p33. To test the hypothesis that p33/gC1qR might reside in mitochondria, we analyzed nontransfected, endothelial EA.hy926 cells by double immunostaining and confocal laser-scanning microscopy. Cells were stained with anti-p33 and with a human autoantibody directed to mitochondrial pyruvate dehydrogenase (Fig. 5). Horizontal sections were scanned, and overlays were prepared from the stored images of the optical sections. A similar staining pattern was obvious for mitochondrial pyruvate dehydrogenase (Fig. 5,A) and endogenous p33/gC1qR (Fig. 5,B). An overlay of the staining patterns produced by the autoantibody and anti-p33 revealed that they were essentially congruent (Fig. 5 C). We conclude that p33/gC1qR colocalizes with mitochondrial pyruvate dehydrogenase of EA.hy926 cells.

FIGURE 5.

Localization of pyruvate dehydrogenase, p33/g1CqR, and GFP. A–C, EA.hy926 cells grown on glass coverslips were fixed with formaldehyde, permeabilized with methanol, and stained with a human autoantibody to pyruvate dehydrogenase (A) or anti-p33 (B), followed by a FITC (A)- or TRITC (B)-labeled secondary Ab. Cells were viewed by confocal laser-scanning microscopy. C, Overlay of A and B. D–G, Transfected COS-7 cells were treated as detailed in the legend to Figure 4. D and F, GFP fluorescence; E and G, TRITC staining for p33/gC1qR. D and E, [cc11–88]GFP; F and G, [rfs1–86]GFP.

FIGURE 5.

Localization of pyruvate dehydrogenase, p33/g1CqR, and GFP. A–C, EA.hy926 cells grown on glass coverslips were fixed with formaldehyde, permeabilized with methanol, and stained with a human autoantibody to pyruvate dehydrogenase (A) or anti-p33 (B), followed by a FITC (A)- or TRITC (B)-labeled secondary Ab. Cells were viewed by confocal laser-scanning microscopy. C, Overlay of A and B. D–G, Transfected COS-7 cells were treated as detailed in the legend to Figure 4. D and F, GFP fluorescence; E and G, TRITC staining for p33/gC1qR. D and E, [cc11–88]GFP; F and G, [rfs1–86]GFP.

Close modal

The colocalization of p33/gC1qR with a mitochondrial enyzme complex prompted us to examine the amino acid composition of the p33/gC1qR presequence. The unusual distribution of charged amino acids, i.e., a single acidic residue (Glu), 11 basic residues (Arg), and the presence of 7 hydroxylated residues, suggested that the leader sequence may in fact represent a mitochondrial targeting sequence (24). To test this hypothesis, we fused the reporter protein GFP with two established mitochondrial targeting sequences, i.e., residues 1–88 of cytochrome c1, [cc11–88]GFP, and residues 1–86 of rfs, [rfs1–86]GFP. We then tested by transiently transfecting COS-7 cells whether the mitochondrial reporter constructs are present in the same compartments as endogenous p33/gC1qR (Fig. 5). Double labeling revealed that the staining patterns for [cc11–88]GFP and [rfs1–86]GFP (Fig. 5, D and F) are congruent with those of endogenous p33/gC1qR in COS-7 cells (Fig. 5, E and G), suggesting that the proteins have been routed to the same intracellular compartments, namely mitochondria.

Mitochondrial proteins carrying cleavable targeting sequences are matured by processing peptidases after their import into mitochondria (25). Using anti-GFP, we followed the intracellular processing of the p33/gC1qR precursor by Western blotting of the various fusion constructs. For control, we transiently expressed unfused GFP in COS-7 cells and obtained a single band of 30 kDa for GFP (Fig. 6,A, lane 1). Both the [p331–81]GFP construct and the [p331–33]GFP construct, which harbor the relevant topogenic segments of the p33/gC1qR targeting sequence, gave rise to mature GFP of 30 kDa (lanes 2 and 3). The additional bands of 36 and 32 kDa, respectively, most likely represent the corresponding precursor proteins indicative of incomplete processing. The amino-terminally truncated form of the presequence, [p3332–81]GFP, generated a single band of 34 kDa, but failed to produce mature GFP protein of 30 kDa (lane 4). Both the [p3374–282]GFP and the [p331–282]GFP fusion constructs gave a 63-kDa protein (lanes 5 and 6). In the latter case, the appearance of an additional 69-kDa band suggested that the [p331–282]GFP undergoes proteolytic conversion into [p3374–282]GFP. Hence, the combined data of Figures 4 and 6 indicate that intracellular routing of the p33/gC1qR precursor is accompanied by the removal of its targeting sequence.

FIGURE 6.

Proteolytic processing of fused targeting sequences. A and B, Three days after transfecting COS-7 cells, total cellular extracts were prepared, resolved by SDS-PAGE (20 μg protein each), blotted to nitrocellulose, and probed by anti-GFP (1:5000), followed by the peroxidase-linked chemiluminescence assay. A, Unfused GFP (lane 1); [p331–81]GFP (lane 2); [p331–33]GFP (lane 3); [p3332–81]GFP (lane 4); [p3374–282]GFP (lane 5); [p331–282]GFP (lane 6). B, [cc11–88]GFP (lane 7); [rfs1–86]GFP (lane 8). C, Western blots of total protein (20 μg per lane) from the organelle fraction of EA.hy926 cells (lane 9), from total cellular extracts of bovine aortic endothelial cells (BAEC; lane 10), and from bovine liver mitoplasts (lane 11). Anti-p33 (0.5 μg/ml) was used as the probe.

FIGURE 6.

Proteolytic processing of fused targeting sequences. A and B, Three days after transfecting COS-7 cells, total cellular extracts were prepared, resolved by SDS-PAGE (20 μg protein each), blotted to nitrocellulose, and probed by anti-GFP (1:5000), followed by the peroxidase-linked chemiluminescence assay. A, Unfused GFP (lane 1); [p331–81]GFP (lane 2); [p331–33]GFP (lane 3); [p3332–81]GFP (lane 4); [p3374–282]GFP (lane 5); [p331–282]GFP (lane 6). B, [cc11–88]GFP (lane 7); [rfs1–86]GFP (lane 8). C, Western blots of total protein (20 μg per lane) from the organelle fraction of EA.hy926 cells (lane 9), from total cellular extracts of bovine aortic endothelial cells (BAEC; lane 10), and from bovine liver mitoplasts (lane 11). Anti-p33 (0.5 μg/ml) was used as the probe.

Close modal

Using authentic mitochondrial targeting sequences, we found that fusions of prototypical mitochondrial targeting sequences to GFP, i.e., [cc11–88]GFP and [rfs1–86]GFP, produced bands of 30 and 36 kDa for mature GFP and the corresponding precursor proteins, respectively (Fig. 6,B, lanes 7 and 8). The presence of an intermediate-size form of 32 kDa in the case of [cc11–88]GFP is in agreement with the bipartite targeting sequence of human cytochrome c1, of which the first part represents a mitochondrial import signal that is cleaved in the matrix, whereas the second part directs the protein to the intermembrane space, where processing to the mature form occurs (25). By contrast, no such intermediate-size form was detected in the case of [rfs1–86]GFP (Fig. 6,B, lane 8) and [p331–81]GFP (Fig. 6,A, lane 2). Notably, the targeting sequence of human rfs undergoes a single cleavage and directs the protein to the outer surface of the inner mitochondrial membrane (26). Therefore, we concluded that p33/gC1qR may be associated with the matrix and/or the inner membrane of mitochondria. To further test this hypothesis, we examined the presence of p33/gC1qR in bovine liver mitoplasts, i.e., mitochondria devoid of an intact outer membrane (Fig. 6,C). Immunoprint analysis revealed that Ab to human p33/gC1qR, anti-p33, readily detected its bovine homologue of 32 kDa in whole cell extracts of bovine aortic endothelial cells (EA.hy926, Fig. 6 C, lane 9; BAEC, lane 10). Likewise, liver mitoplasts gave a strong band of 32 kDa (lane 11), thus supporting the notion that p33/gC1qR resides in the mitochondrial matrix and/or the inner mitochondrial membrane.

We asked whether p33/gC1qR is expressed predominantly in tissues that are rich in mitochondria (Fig. 7). Initial experiments demonstrated that anti-p33 readily cross-reacts with its mouse homologue (data not shown). Immunostaining of paraffin sections of fetal mice indicated that p33/gC1qR is expressed ubiquitously in fetal mice tissues (Fig. 7,A); the corresponding preimmune serum stained negative (Fig. 7,B). Most prominent staining was observed in brown fat tissue (Fig. 7,A′), liver (Fig. 7,B′), kidney (Fig. 7,C′), hindgut (Fig. 7,D′), pancreas, and salivary glands (not shown at higher magnification). The cellular distribution was cytoplasmic, particulate (Fig. 7, a′–d′). Polarized cells in secretory structures such as hindgut (Fig. 7 d′) and pancreas (not shown at higher magnification) had the strongest staining in their apical parts. This pattern is compatible with and indicative of p33/gC1qR location in mitochondria of metabolically highly active, secretory tissues.

FIGURE 7.

Localization of p33/gC1qR in fetal mouse tissues. Fetal mice (E16) were fixed in Bouin’s fluid. Paraffin sections were prepared, dewaxed, and stained with 15 μg/ml of preimmune Ab (B) or anti-p33 (all other panels). Bound Ab was detected by the peroxidase-antiperoxidase method and diaminobenzidine. Low magnification views (×4) show sagittal sections of E16 mice, including part of the placenta (A, B). Medium magnification views (×100) show p33 distribution in brown fat tissue (A′), liver and adjacent diaphragm (B′), kidney (C′), and intestine (D′). High magnifications of boxed areas (×600) show individual cells of brown fat (a′), liver parenchyme (b′), renal collecting duct (c′), and intestine mucosa (d′).

FIGURE 7.

Localization of p33/gC1qR in fetal mouse tissues. Fetal mice (E16) were fixed in Bouin’s fluid. Paraffin sections were prepared, dewaxed, and stained with 15 μg/ml of preimmune Ab (B) or anti-p33 (all other panels). Bound Ab was detected by the peroxidase-antiperoxidase method and diaminobenzidine. Low magnification views (×4) show sagittal sections of E16 mice, including part of the placenta (A, B). Medium magnification views (×100) show p33 distribution in brown fat tissue (A′), liver and adjacent diaphragm (B′), kidney (C′), and intestine (D′). High magnifications of boxed areas (×600) show individual cells of brown fat (a′), liver parenchyme (b′), renal collecting duct (c′), and intestine mucosa (d′).

Close modal

Binding to the cell surface receptors of peripheral blood cells or vascular endothelial cells is a salient feature of the components of proteolytic cascades such as the coagulation, fibrinolytic, kinin, or complement system. In agreement with this notion, the complement receptors for C3b(i), i.e., CR1, CR3, and CR4, are translocated from intracellular compartments of neutrophils to the cell surface after stimulation (27). Unlike these CR receptors that represent integral membrane proteins, the putative receptor for the globular heads of C1q is a highly hydrophilic protein lacking transmembrane-spanning regions (9). Because p33/gC1qR was immunodetected on the surface of various cells such as neutrophils, platelets, mast cells, endothelial cells, and B cells (9, 10, 11, 12, 13), it has been hypothesized that p33/gC1qR might anchor to the cell membrane via its presequence. However, the sequence analysis of p33/gC1qR from various human cells has revealed that only the fully processed protein of 33 kDa is present in total cellular extracts (4, 5).

Further puzzling is the fact that p33/gC1qR is similar or even identical to putative nuclear factors. A 32-kDa protein (SF2p32) that copurifies with the pre-mRNA splicing factor SF2 from HeLa cells is identical with p33/gC1qR (28). Yu et al. (29) reported the sequence of trans activator (Tat)-associated protein (TAP) of 32 kDa that facilitates the binding of HIV-1-encoded Tat to the transcription factor TFIIB. Comparison of the TAP sequence with that of p33/gC1qR reveals total sequence identity for positions 1 to 23 and 67 to 282 of the proteins. Complete sequence identity of the entire protein sequence (positions 1–282) is readily obtained by the correction of two potential sequencing errors or cloning artifacts that cause frameshifts at the points of divergency (data not shown). By the two-hybrid approach, a mouse homologue of TAP (“YL2”) that associates with the HIV-1 trans activator Rev has been isolated from T cells (30). The corresponding cDNA sequence that covers only positions 74–282 of p33/gC1qR is 92% identical to the human homologue (30). Unfortunately, the subcellular localization of the putative nuclear factors SF2p32, TAP, and YL2 has not been reported to date.

Given the wide array of plasma, viral, and nuclear proteins that interact with p33/gC1qR-like proteins (4, 5, 6, 7, 8, 9, 28, 29, 30), we suggest that p33/gC1qR represents a multiligand-binding protein. Indeed, the unique amino acid composition with clusters of highly charged sequence segments makes p33/gC1qR a prime candidate for promiscuous interaction with many different proteins, including cC1qR (8). Therefore, the identification of p33/gC1qR in affinity-purified or immunoprecipitated protein isolates from whole cell extracts or in yeast two-hybrid systems warrants a careful analysis of the physiologic relevance of the observed protein-protein interactions.

Considering its mitochondrial localization of p33/gC1qR described in this work, the proposed functions as an extracellular or cytoplasmic anchor for plasma, nuclear, or viral proteins cannot be fulfilled by p33/gC1qR, unless it is released from mitochondria. The secretion of minor amounts of p33/gC1qR has been claimed for lymphocytes (14, 31); however, our own studies do not support such a notion. Using Abs against distinct epitopes of p33/gC1qR, we were unable to detect significant amounts of the protein on the surface or in the nucleus of intact endothelial cells (15). Likewise, the stimulation of EA.hy926 cells with bradykinin, epidermal growth factor, PMA, FMLP, IL-1β, TNF-α, ionophore A23187, or shear stress failed to induce significant secretion or translocation of p33/gC1qR (data not shown). Our experiments do not rule out the possibility that p33/gC1qR is released from mitochondria following cell apoptosis; however, such a mechanism is incompatible with its postulated function as a major complement receptor of circulating or resident cells of the host defense systems. Rather, cellular protein(s) other than p33/gC1qR must account for the observed binding sites for the globular heads of C1q (3). Other mechanisms that could be invoked for an extramitochondrial location of mitochondrial proteins include alternative splicing of the primary transcript (32) and/or redistribution of differentially folded translation products (33); however, the failure to immunodetect p33/gC1qR Ag in the cytoplasm denies such possibilities.

At present, one can only speculate about the true cellular function(s) and ligand(s) of p33/gC1qR-like proteins in mammalian cells. In this context, it is most noteworthy that the yeast homologue of p33/gC1qR has been identified recently in Saccharomyces cerevisiae (34). Mitochondrial acidic matrix protein of 33 kDa (Mam33p) shares 53% sequence similarity and 24% sequence identity with human p33/gC1qR. The Mam33p precursor comprises a mitochondrial targeting sequence of 47 residues, which directs the protein to the mitochondrial matrix. Disruption of the Mam33p gene does not produce a major phenotype, e.g., cell growth is near normal (34). Hence, the role(s) of Mam33p in yeast remains obscure. Given the evolutionary conservation of p33/gC1qR-like proteins, it will be interesting to learn about the effects of a targeted deletion of the p33/gC1qR gene in multicellular species such as mouse.

Note added in proof: A recent study by Muta et al. (35) on the subcellular localization of SF2p32 protein arrives at similar conclusions.

1

This work was supported by grants from Deutsche Forschungsgemeinschaft (Mu 598/5-2) and from Fonds der Chemischen Industrie (to W.M.-E.).

3

Abbreviations used in this paper: cC1qR, receptor for the collagenous domains of C1q; ABTS, diammonium 2,2′-azino-bis(3-ethyl-2, 3-dihydrobenzthiazoline-6-sulfonate); cc1, cytochrome c1; gC1qR, receptor for the globular domains of C1q; GFP, green fluorescent protein; H-kininogen, high molecular mass kininogen; 125I-H-kininogen, 125I-labeled high molecular mass kininogen; rfs, Rieske iron-sulfur protein; TRITC, tetramethylrhodamine isothiocyanate.

1
Nepomuceno, R. R., A. H. Henschen-Edman, W. H. Burgess, A. J. Tenner.
1997
. cDNA cloning and primary structure analysis of C1qRp, the human C1q/MBL/SPA receptor that mediates enhanced phagocytosis in vitro.
Immunity
6
:
119
2
Erdei, A., K. B. Reid.
1988
. The C1q receptor.
Mol. Immunol.
25
:
1067
3
Bordin, S., R. C. Page.
1989
. Detection of a high-affinity binding site for the globular head regions of C1q complement protein on a human diploid fibroblast subtype.
Mol. Immunol.
26
:
677
4
Herwald, H., J. Dedio, R. Kellner, W. Müller-Esterl.
1996
. Isolation and characterization of the kininogen-binding protein p33 from endothelial cells: identity with the gC1q receptor.
J. Biol. Chem.
271
:
13040
5
Joseph, K., B. Ghebrehiwet, E. I. B. Peerschke, K. B. M. Reid, A. P. Kaplan.
1996
. Identification of the zinc-dependent endothelial cell binding protein for high molecular weight kininogen and factor XII: identity with the receptor that binds to the globular “heads” of C1q (gC1q-R).
Proc. Natl. Acad. Sci. USA
93
:
8552
6
Lim, B.-L., K. B. M. Reid, B. Ghebrehiwet, E. I. B. Peerschke, L. E. A. Leigh, K. T. Preissner.
1996
. The binding protein for globular heads of complement C1q, gC1qR: functional expression and characterization as a novel vitronectin binding factor.
J. Biol. Chem.
271
:
26739
7
Ghebrehiwet, B., P. D. Lu, W. Zhang, B.-L. Lim, P. Eggleton, L. E. A. Leigh, K. B. M. Reid, E. I. B. Peerschke.
1996
. Identification of functional domains on gC1q-R, a cell surface protein that binds to the globular “heads” of C1q, using monoclonal antibodies and synthetic peptides.
Hybridoma
15
:
333
8
Ghebrehiwet, B., P. D. Lu, W. Zhang, S. A. Keilbaugh, L. E. A. Leigh, P. Eggleton, K. B. M. Reid, E. I. B. Peerschke.
1997
. Evidence that the two C1q binding membrane proteins, gC1q-R and cC1q-R, associate to form a complex.
J. Immunol.
159
:
1429
9
Ghebrehiwet, B., B.-L. Lim, E. I. B. Peerschke, A. C. Willis, K. B. M. Reid.
1994
. Isolation, cDNA cloning, and overexpression of a 33-kD cell surface glycoprotein that binds to the globular “heads” of C1q.
J. Exp. Med.
179
:
1809
10
Ghebrehiwet, B., R. R. Kew, B. L. Gruber, M. J. Marchese, E. I. B. Peerschke, K. B. M. Reid.
1995
. Murine mast cells express two types of C1q receptors that are involved in the induction of chemotaxis and chemokinesis.
J. Immunol.
155
:
2614
11
Eggleton, P., B. Ghebrehiwet, K. N. Sastry, J. P. Coburn, K. S. Zaner, K. B. M. Reid, A. I. Tauber.
1995
. Identification of a gC1q-binding protein (gC1q-R) on the surface of human neutrophils: subcellular localization and binding properties in comparison with the cC1q-R.
J. Clin. Invest.
95
:
1569
12
Peerschke, E. I. B., K. B. M. Reid, B. Ghebrehiwet.
1994
. Identification of a novel 33-kDa C1q-binding site on human blood platelets.
J. Immunol.
152
:
5896
13
Peerschke, E. I. B., S. S. Smyth, E. I. Teng, M. Dalzell, B. Ghebrehiwet.
1996
. Human umbilical vein endothelial cells possess binding sites for the globular domain of C1q.
J. Immunol.
157
:
4154
14
Van den Berg, R. H., F. Prins, M. C. Faber-Krol, N. J. Lynch, W. Schwaeble, L. A. van Es, M. R. Daha.
1997
. Intracellular localization of the human receptor for the globular domains of C1q.
J. Immunol.
158
:
3909
15
Dedio, J., W. Müller-Esterl.
1996
. Kininogen binding protein p33/gC1qR is localized in the vesicular fraction of endothelial cells.
FEBS Lett.
399
:
255
16
Hasan, A. A. K., D. B. Cines, J. Zhang, A. H. Schmaier.
1994
. The carboxyl terminus of bradykinin and amino terminus of the light chain of kininogens comprise an endothelial cell binding domain.
J. Biol. Chem.
269
:
31822
17
Nishikimi, M., S. Ohta, H. Suzuki, T. Tanaka, F. Kikkawa, M. Tanaka, Y. Kagawa, T. Ozawa.
1988
. Nucleotide sequence of a cDNA encoding the precursor to human cytochrome c1.
Nucleic Acids Res.
16
:
3577
18
Nishikimi, M., Y. Hosokawa, H. Toda, H. Suzuki, T. Ozawa.
1990
. The primary structure of Rieske iron-sulfur protein of mitochondrial cytochrome bc1 complex deduced from cDNA analysis.
Biochem. Int.
20
:
155
19
Edgell, C.-J. S., C. C. McDonald, J. B. Graham.
1983
. Permanent cell line expressing human factor VIII-related antigen established by hybridization.
Proc. Natl. Acad. Sci. USA
80
:
3734
20
Abd Alla, S., J. Buschko, U. Quitterer, A. Maidhof, M. Haasemann, G. Breipohl, J. Knolle, W. Müller-Esterl.
1993
. Structural features of the human bradykinin B2 receptor probed by agonists, antagonists, and anti-idiotypic antibodies.
J. Biol. Chem.
268
:
17277
21
Schreurs, J., K. Arai, A. Miyajima.
1990
. Evidence for a low-affinity interleukin-3 receptor.
Growth Factors
2
:
221
22
Haasemann, M., J. Buschko, A. Faussner, A. A. Roscher, J. Hoebeke, R. M. Burch, W. Müller-Esterl.
1991
. Anti-idiotypic antibodies bearing the internal image of a bradykinin epitope: production, characterization, and interaction with the kinin receptor.
J. Immunol.
147
:
3882
23
Johnson, L. V., M. L. Walsh, L. B. Chen.
1980
. Localization of mitochondria in living cells with rhodamine 123.
Proc. Natl. Acad. Sci. USA
77
:
990
24
Neupert, W..
1997
. Protein import into mitochondria.
Annu. Rev. Biochem.
66
:
863
25
Hartl, F.-U., W. Neupert.
1990
. Protein sorting to mitochondria: evolutionary conservations of folding and assembly.
Science
247
:
930
26
Brandt, U., L. Yu, C.-A. Yu, B. L. Trumpower.
1993
. The mitochondrial targeting presequence of the Rieske iron-sulfur protein is processed in a single step after insertion into the cytochrome bc1 complex in mammals and retained as a subunit in the complex.
J. Biol. Chem.
268
:
8387
27
Sengelov, H..
1995
. Complement receptors in neutrophils.
Crit. Rev. Immunol.
15
:
107
28
Kreiner, A. R., A. Mayeda, D. Kozak, G. Binns.
1991
. Functional expression of cloned human splicing factor SF2: homology to RNA-binding proteins, U1 70K, and Drosophila splicing regulators.
Cell
66
:
383
29
Yu, L., Z. Zhang, P. M. Loewenstein, K. Desai, Q. Tang, D. Mao, J. S. Symington, M. Green.
1995
. Molecular cloning and characterization of a cellular protein that interacts with the human immunodeficiency virus type 1 Tat transactivator and encodes a strong transcriptional activation domain.
J. Virol.
69
:
3007
30
Luo, Y., H. Yu, B. M. Peterlin.
1994
. Cellular protein modulates effects of human immunodeficiency virus type 1 Rev.
J. Virol.
68
:
3850
31
Peterson, K. L., W. Zhang, P. D. Lu, S. A. Keilbaugh, E. I. Peerschke, B. Ghebrehiwet.
1997
. The C1q-binding cell membrane proteins cC1q-R and gC1q-R are released from activated cells: subcellular distribution and immunochemical characterization.
Clin. Immunol. Immunopathol.
84
:
17
32
Natsoulis, G., F. Hilger, G. R. Fink.
1986
. The HTS1 gene encodes both the catoplasmic and mitochondrial histidine tRNA synthetases of S. cerevisiae.
Cell
46
:
235
33
Stein, I., Y. Peleg, S. Even-Ram, O. Pines.
1994
. The single translation product of the FUM1 gene (fumarase) is processed in mitochondria before being distributed between the cytosol and mitochondria in Saccharomyces cerevisiae.
Mol. Cell. Biol.
14
:
4770
34
Seytter, T., F. Lottspeich, W. Neupert, and E. Schwarz. 1998. Mam33p, an oligomeric, acidic protein in the mitochondrial matrix of Saccharomyces cerevisiae is related to the human complement receptor gC1q-R. Yeast. In press.
35
Muta, T., D. Kang, S. Kitajama, T. Fujiwara, N. Hamasaki.
1997
. p32 protein, a splicing factor 2-associated protein, is localized in the mitochondrial matrix and is functionally important in maintaining oxidative phosphorylation.
J. Biol. Chem.
272
:
24363