CD59 is a 18- to 20-kDa, GPI-anchored membrane protein that functions as a key regulator of the terminal step of the complement activation cascade. It restricts binding of C9 to the C5b-8 complex, thereby preventing the formation of the membrane attack complex (C5b-9 of complement). A single human CD59 gene has been identified, and corresponding genetic homologues from rat, mouse, and pig have been characterized in previous studies. In this study, we report the discovery and functional characterization of a separate cd59 gene in the mouse (referred to as cd59b, the previously characterized mouse cd59 gene as cd59a). Mouse cd59b is 85% and 63% identical to cd59a at the nucleotide and amino acid level, respectively. In cDNA transfection experiments with Chinese hamster ovary cells, peptide-tagged cd59b was detected on the cell surface by flow cytometry and was shown to be susceptible to phosphatidylinositol-specific phospholipase C cleavage. Chinese hamster ovary cells expressing cd59b were significantly more resistant than control cells to human and mouse complement-mediated lysis. These results suggest that cd59b encodes a GPI-anchored protein that is functionally active as a membrane attack complex inhibitor. Northern blot analysis revealed that cd59b is expressed selectively in the mouse testis. In contrast, the major transcript of cd59a was shown to be expressed at high levels in the heart, kidney, liver, and lung, but only minimally in the testis. These results revealed the existence of two distinct cd59 genes in the mouse that are differentially regulated and that may have nonoverlapping physiological functions in vivo.
Activation of the autologous complement system must be carefully controlled to avoid the deleterious effect of complement on autologous cells. This restriction of autologous complement activation is accomplished in part by endogenous complement-inhibitory proteins anchored on the membrane of most animal cells (1, 2, 3). Expression of these inhibitors limits the activation of the complement cascade on autologous/homologous cell surfaces (1, 2, 3). A central and well-characterized member of the family of membrane complement-inhibitory proteins is CD59, a GPI-linked glycoprotein with a molecular mass of approximately 20 kDa (4, 5, 6, 7, 8). CD59 acts at the terminal step of the complement activation cascade, i.e., the binding of C9 to C5b-8. By inhibiting this step, CD59 effectively prevents the formation of the polymeric membrane attack complex (MAC)3 and protects the cells from MAC-induced lysis and nonlytic phenomena (4, 5, 6, 7, 8).
The importance of CD59 in protecting cells within the vascular space from complement attack is illustrated by the acquired hematological disorder, paroxysmal nocturnal hemoglobinuria (PNH) (9, 10). Due to a defect in GPI anchor biosynthesis, the affected blood cells of PNH patients do not possess CD59 (9, 10, 11, 12) or decay-accelerating factor (DAF), a second GPI-anchored membrane complement inhibitor that acts at the C3 convertase step (9, 10, 13). As a result of DAF and CD59 deficiencies, affected erythrocytes and platelets from these patients are highly sensitive to autologous complement-mediated lysis and activation, and thereby suffer hemolytic anemia and thrombosis (9, 10, 11, 12, 13). Studies on the structure of human CD59 revealed that human CD59 is a single copy gene localized on chromosome 11p13 (14, 15, 16, 17, 18, 19). More recently, CD59 cDNAs from rat, mouse, and pig have been cloned and the encoded proteins functionally characterized (20, 21, 22). The study of animal CD59 has not only facilitated our understanding of the evolution and structure-function relationships of CD59, but also paved the way for establishing appropriate animal models to study the physiological functions of CD59 in vivo. In the case of mouse cd59, a cDNA has been cloned and the corresponding gene has been localized to the region E2-E4 of mouse chromosome 2 (21), a region syntenous with human chromosome 11p13, in which the human CD59 gene has been mapped (19). In an effort to characterize the genomic structure of the mouse cd59, we have identified a second mouse cd59 gene, cd59b. In this study, we describe the cDNA and deduced amino acid sequences of the new gene, the functional characterization of its product as a MAC inhibitor, and the tissue expression patterns of the two mouse cd59 genes.
Materials and Methods
Isolation of the cd59b gene
A full-length mouse cd59a cDNA was amplified by standard RT-PCR and used as a probe to screen a 129/SV murine λFixII genomic library (Stratagene, La Jolla, CA). First-strand cDNA was synthesized using 20 μg of total liver RNA and oligo(dT) primer, as previously described (23). Two primers, 5′-AGCACAGTCACTGGCGAT-3′ (upstream) and 5′-GAGGCAAGCTCTTACTATAT-3′ (downstream), were designed according to the published cDNA sequence of mouse cd59a (21). The amplified cDNA (1.25 kb) was purified on an agarose gel and cloned into the pCR2 vector using the TA cloning kit (Invitrogen, Carlsbad, CA). The mouse genomic library was plated and screened according to the manufacturer’s instruction using 32P-labeled cd59a cDNA as a probe. Positive clones were purified through secondary and tertiary screening, and phage DNAs were prepared by the liquid culture method (24). The phage DNA was digested with NotI enzyme to release the DNA insert that was subsequently purified from an agarose gel and cloned into the pBluescript plasmid vector (Stratagene).
cDNA cloning of cd59b
The following cd59b-specific primers were used to amplify by RT-PCR a 1083-bp cDNA fragment that covered over 80% of the full-length cDNA: 5′-CTGTTTAGACCCGGTTTCTT-3′ (P5, upstream), 5′-CCACTATGTAGGTCAGGCT-3′ (P6, downstream). The first-strand cDNA used for this reaction was synthesized with total RNAs isolated from the mouse liver. To obtain the remaining 5′-cDNA sequence, the 5′-RACE (rapid amplification of cDNA ends) method (25) was used with a Marathon cDNA amplification kit from Clontech (Palo Alto, CA). The two specific downstream primers used in the 5′-RACE experiment were: 5′-CGGCTACAGCATAGAGACAGGAAT-3′ (outside) and 5′-AAGAAACCGGGTCTAAACAG-3′ (inside). The full coding region of cd59b was amplified by RT-PCR with the following two primers: 5′-ATGAGAGCTCAGAGGGGA-3′ (upstream), 5′-ATCCAGGATGACTTAGAAGCA-3′ (downstream). All amplified PCR products were first cloned into the pCR2 vector using the TA cloning kit, and their sequences were determined on an ABI model 373 automatic sequencer using the PRISM labeling kit (Applied Biosystems, Foster City, CA).
Construction of cd59b and cd59a expression plasmids
Both cd59b and cd59a (as a control) were expressed as N-terminal peptide-tagged proteins to facilitate their detection on the cell surface. Initially, the cDNAs (coding region for cd59b and full-length for cd59a (21)) were restriction cut from the pCR2 vector at EcoRI site and subcloned into the pCDNA3 vector at the same site. Subsequently, plasmids encoding cd59b and cd59a proteins containing a FLAG peptide tag (DYKDDDDK) were constructed to help visualize surface expression of the proteins by indirect immunofluorescence using a FLAG-specific Ab. A similar approach involving N-terminal peptide tagging has been successfully used in previous functional studies of human CD59 (26, 27). The FLAG sequence (encoded by a 24-nucleotide sequence GACTACAAGGACGACGATGACAAG) was inserted into the coding region of cd59b or cd59a (after the second amino acid (lysine) in the predicted mature cd59b protein, or after the second amino acid (threonine) in the mature cd59a protein) to yield N-terminal tagged cd59 proteins after cleavage of the signal peptides. Construction of these plasmids was achieved by a two-round PCR method. In the first-round PCR, two separate fragments for cd59b or cd59a were amplified using cd59b or cd59a cDNA in pCDNA3 as a template and the following two pairs of primers: pair 1, T7 (forward) and a cd59b- or cd59a-specific primer containing the FLAG sequence, 5′-TTTGAGcttgtcatcgtcgtccttgtagtcTTTGAGACTAACAGCTGTGGAACA-3′ (reverse primer for cd59b, FLAG sequence in lower case) or 5′-TGTGAGcttgtcatcgtcgtccttgtagtcTGTGAGGCTAACAGCTGTGGAACA-3′ (reverse primer for cd59a, FLAG sequence in lower case); pair 2, Sp6 (reverse) and a cd59b- or cd59a-specific primer containing the FLAG sequence, 5′-gactacaaggacgacgatgacaagCTCAAATGCTACAACTGTTTAGACCCGGTTTTCT-3′ (forward primer for cd59b, FLAG sequence in lower case) or 5′-gactacaaggacgacgatgacaagCTCACATGCTACCACTGTTTCCAACCGGTGGTT-3′ (forward primer for cd59a, FLAG sequence in lower case). The two amplified cDNA fragments were purified on a 1% agarose gel. To obtain FLAG-containing cd59b or cd59a cDNA, the purified cDNA fragments were mixed and used in a second-round PCR, with T7 as a forward and Sp6 as a reverse primer. The resulting PCR product was purified, digested with KpnI and NotI, and cloned into KpnI- and NotI-digested pAlter-MAX vector (abbreviated as pAlter below) (Promega, Madison, WI). The recombinant plasmid was confirmed to contain the authentic signal peptide, FLAG, and the correct mouse cd59b or cd59a sequence by sequence analysis before being used in transfection experiment.
Cell transfection experiment
Plasmid containing FLAG-cd59b or FLAG-cd59a, or the control vector (pAlter) was transfected into Chinese hamster ovary cells (CHO; American Type Culture Collection, Bethesda, MD) to assess the MAC-inhibitory activity of the encoded protein. CHO cells were cultured in F12 medium supplemented with 10% FBS, 4 mM glutamine, and 100 U/ml each of penicillin and streptomycin. DNA transfection was conducted using Lipofectamine (Life Technologies, Grand Island, NY) by following the manufacturer’s instructions. To confer G418 resistance, cells were cotransfected with pAlter or pAlter-FLAG-cd59b or cd59a and pCDNA3 (5 μg pAlter or pAlter-FLAG-cd59 plasmid and 100 ng pCDNA3). Two days after transfection, G418 (Life Technologies) was added to the medium (800 μg/ml) to select for transfected cells. Drug-resistant cells began to form small colonies after 2 wk of G418 addition. Individual colonies were picked into a 24-well plate and propagated. Total RNAs were subsequently prepared from the cloned cd59-transfected CHO cells and analyzed by Northern blot analysis to detect cd59b or cd59a mRNA. The clone that contained the highest level of cd59b or cd59a mRNA was chosen for protein expression and MAC-inhibiting activity assays. A drug-resistant clone from pAlter-transfected cells was used as a control in these assays.
Detection of cell surface expression of cd59b and phosphatidylinositol-specific phospholipase C (PIPLC) cleavage experiment
Cell surface expression of cd59b and cd59a in cloned CHO cells was assessed by FACScan analysis using an anti-FLAG mAb. Cells were physically dislodged from the plate in calcium-free PBS (followed by gentle vortex to disperse cell aggregate), washed three times with PBS/BSA (3%), and resuspended at 2 × 106 cells/ml in the same buffer. Cells were incubated with 10 μg/ml anti-FLAG mAb (Sigma, St. Louis, MO; catalogue F-3165) in PBS/BSA (3%) for 30 min at room temperature, washed three times with PBS/BSA (3%) buffer, and incubated for 30 min with a FITC-conjugated rabbit anti-mouse IgG secondary Ab (Sigma; catalogue F-9137). Cells were washed in PBS for three times before being analyzed for fluorescence intensity using a Becton Dickinson FACScan (San Jose, CA). To examine the effect of treatment with PIPLC, cells were washed and resuspended in PBS at 4 × 106 cells/ml. They were then incubated with PIPLC (1 U/ml final concentration; Sigma; catalogue P-8804) for 30 min at 37°C. Cells were washed with PBS/BSA (3%) for three times, stained with anti-FLAG Ab and the secondary Ab, and analyzed by FACScan, as described above.
Functional assays of cd59b
Vector-transfected and FLAG-cd59b- or FLAG-cd59a-transfected CHO cells were tested for their sensitivity to human and mouse complement-mediated lysis. Briefly, cells (2 × 104 cells/well) were seeded in 96-well plates. After reaching 90% confluence, they were loaded with a fluorescent dye, 2′,7′-bis-(2-carboxyethyl)-5-(and -6)-carboxyfluorescein, acetoxymethyl ester (BCEF-AM) (Molecular Probes, Eugene, OR), by incubating with 20 μM BCECF-AM at 37°C for 30 min. After washing several times to remove the unincorporated fluorescent dye, the cells were sensitized with 50 μl of an IgG fraction of rabbit anti-hamster lymphocyte (4 mg/ml in HBSS/1% BSA; Intercell Technologies, Hopewell, NJ). The Ab-sensitized cells were then exposed to different concentrations of human or mouse complement in HBSS/1% BSA for 30 min at 37°C. Protection against mouse complement was tested using mouse plasma. Mouse blood was collected by cardiac puncture into EDTA-containing tubes, and plasma was prepared by centrifugation at 3000 rpm on a microcentrifuge. To activate the mouse classical complement pathway, CaCl2 (2 mM final) was added to the assaying wells immediately after mouse plasma addition. At the end of the incubation, the supernatants in the wells were collected to determine the fluorescent dye content as a measure of cell lysis. To the remaining cell pellet, Triton X-100 (1% final concentration) was added to release fluorescent dye in surviving cells. Percentage of complement-mediated cell lysis was calculated by comparing the amount of fluorescent dye released into the supernatant and the total amount of fluorescent dye (supernatant plus cell pellet).
Northern blot analysis
Total RNAs from various mouse tissues were isolated using the Trizol Reagent (Life Technologies). RNA samples (10 μg each lane) were separated on a 1% formaldehyde-agarose gel and transferred onto a nylon membrane (Hybond-N; Amersham, Arlington Heights, IL) via capillary action overnight in 20× SSC. The membrane was cross-linked under UV and hybridized first with a 32P-labeled cd59a probe. After probing with the cd59a probe, the membrane was stripped by boiling in 0.1× SSC-0.1% SDS and rehybridized with a 32P-labeled cd59b-specific probe. The cd59a probe used corresponded to the 5′ 370 bp in the cd59a cDNA (21) and was prepared by BamHI digestion of the full-length cDNA. The 124-bp cd59b-specific probe was generated by PCR using the full coding region of cd59b as a template and the following oligonucleotides as primers: 5′-TCCAACTATATTATGAGCCG-3′ (upstream, nucleotide 304–323), 5′-TCAATGAGGAAGTTTCTGCG-3′ (downstream, nucleotide 408–427). Both cDNA probes were labeled with 32P using random primers. Northern hybridizations were conducted in QuikHyb solution (Stratagene, La Jolla, CA) at 68°C for 1 h. The membrane was washed, first in 2× SSC-0.1% SDS at 55°C for 15 min and then in 0.1× SSC-0.1% SDS at 55°C, and exposed to x-ray film.
Southern blot analysis
Southern hybridization of plasmid and genomic DNAs was conducted using the same general protocol. Plasmid DNAs (1–2 μg DNAs, 10–20 U of restriction enzymes) were digested for 2–3 h. Mouse and rat genomic DNAs were prepared from C57/B6 strain mice and Wistar strain rats, respectively. Human genomic DNA was prepared from liver biopsy samples obtained through the Corporate Human Tissues Network (CHTN; Eastern Division, Philadelphia, PA). Genomic DNAs were digested with different restriction enzymes overnight (3 μg DNAs and 60–90 U of restriction enzyme in 20–30 μl total volume). Samples were electrophoresed on 1% agarose gels and transferred onto nylon membranes (Hybond-N; Amersham) via capillary action in 6× SSC. Membranes were cross-linked under UV and hybridized with the appropriate 32P-labeled probes, as specified. For the genomic Southern blotting, the mouse probe used was an approximately 720-bp PstI-EcoRI fragment corresponding to the 3′-untranslated region of the mouse cd59b cDNA. This cDNA segment is contained within a single exon (exon 4). There is high homology between cd59a and cd59b in this region (95%), and the probe is expected to hybridize to both genes. The rat probe used was a 364-bp fragment (U48255; nucleotide 1029–1392) corresponding to the 3′ portion of the rat cd59 cDNA (20). The human probe used was a 373-bp fragment (M34671.1; nucleotide 534–906) corresponding to the 3′ portion of the human CD59 cDNA (28). cDNA probes were labeled with 32P using random primers. Southern hybridization was conducted in QuikHyb solution (Stratagene, La Jolla, CA) at 68°C for 1 h. The membrane was washed, first in 2× SSC-0.1% SDS at 55°C for 15 min and then in 0.1× SSC-0.1% SDS at 55°C, and exposed to x-ray film.
The initial objective of this study was to determine the structural organization of the mouse cd59a gene. Two oligonucleotide primers, corresponding to the 5′ and 3′ ends of the published mouse cd59a cDNA sequence (21), were synthesized and used to amplify by RT-PCR a 1.2-kb cDNA fragment, using first-strand cDNAs prepared from mouse liver RNA. Restriction digestion analysis confirmed that this fragment corresponded to the previously cloned cd59a cDNA (21). Using this cDNA as a probe, we screened a murine 129/SV genomic library and obtained four positive clones. Restriction characterization of a clone with the strongest hybridization signal revealed a 16-kb insert. The insert was removed from the phage DNA by NotI restriction enzyme digestion, gel purified, and cloned into the plasmid vector P-Bluescript (Stratagene). Subsequently, both Southern blotting and sequence analysis were conducted to identify putative CD59 exon sequences contained within this genomic fragment.
Initially, we focused on a 2.2-kb XbaI and a 1.6-kb XbaI-KpnI restriction fragment. Both fragments gave positive signals on Southern blot analysis. The sequences of these fragments were determined after subcloning into the P-Bluescript plasmid vector. A 79-bp region with 100% homology to the published mouse cd59a cDNA sequence (21) was identified from the 2.2-kb XbaI fragment (Table I). Based on the known genomic structure of human CD59 (14), it was concluded that this segment represented exon 2 of the mouse cd59 gene. Unexpectedly, sequence analysis of the 1.6-kb XbaI-KpnI fragment failed to identify a consensus sequence with the published mouse cd59a cDNA (21). Instead, a 102-bp segment showing a high degree of homology (86% identity) with the cd59a cDNA was located (Table I). This result suggested that our cd59 genomic clone represented either a separate CD59 gene or a pseudogene.
|Exon .||Sequencea .||Length of Exon (bp)b .||Identity with cd59a (%) .|
|Exon .||Sequencea .||Length of Exon (bp)b .||Identity with cd59a (%) .|
Three segments of exonic sequence (capital letters) with homology to cd59a cDNA were located. The exon/intron boundaries for exon 2 and 3 were determined, and their sequences are shown (lower case letters for intron).
Only a partial sequence of exon 4 (762 bp, representing a PstI-HindIII restriction fragment) was determined from the genomic DNA.
To further characterize this genomic clone, a third 0.8-kb PstI-HindIII restriction fragment, which cross-hybridized strongly on Southern blot with the 3′ half (noncoding region) of the cd59a cDNA, was cloned and sequenced. This revealed a 762-bp partial exon sequence that was highly homologous (95% identity) to the 3′-untranslated region of the cd59a cDNA (21) (Table I). By analogy with the human CD59 genomic structure (14), it was concluded that this PstI-HindIII fragment corresponded to exon 4 of mouse cd59. Together, the above results identified three putative mouse cd59 exons (2, 3, 4). Exon 2 was identical to and exons 3 and 4 had high homology with the cd59a cDNA sequence (Table I) (21).
To determine whether the partially characterized cd59 genomic fragment represented a functional cd59 gene or a cd59 pseudogene, RT-PCR was performed to detect mRNA potentially transcribed from this gene. Two specific oligonucleotide primers (P5 and P6) were designed based on the putative exonic sequences. Because the sequences of these two primers were not present in the previously characterized cd59a cDNA (21), they were not expected to amplify the known cd59a cDNA. A 1083-bp cDNA fragment was amplified using first-strand cDNAs synthesized from the mouse liver. The cDNA was cloned and sequence analysis showed that it corresponded to the 3′-cDNA of a new mouse cd59 gene, referred to below as cd59b (the previously characterized mouse cd59 gene is referred to as cd59a). Subsequently, 5′-RACE was performed to amplify a 199-bp cDNA, corresponding to the remaining 5′ sequence of cd59b. The full-length cd59b sequence was obtained from these two overlapping cDNA clones (Fig. 1). All putative exonic sequences (exons 2, 3, and 4) determined from genomic DNA clones could be located within the full-length cDNA (Fig. 1). There is a high degree of homology between cd59a and cd59b, except at the extreme end of the 5′-untranslated region. The most 5′ regions in the two cd59 cDNAs (63 bp in cd59a, 78 bp in cd59b) are divergent and are encoded by nonhomologous exons (exon 1).
The full-length cd59b cDNA was found to contain a single open reading frame that encodes a polypeptide chain of 129 aa (Fig. 1). Alignment with the amino acid sequence of cd59a showed a 63% sequence identity. Noticeably, there is a high degree of homology between cd59a and cd59b at both the N-terminal and C-terminal end (Fig. 2). These regions are likely to represent signal peptides (N-terminal signal peptide for protein trafficking and C-terminal GPI anchor attachment signal peptide). Hydrophobicity analysis revealed these areas to be highly hydrophobic (Fig. 2 B). Comparison of the amino acid sequence of cd59b with that of human and rat CD59 showed that cd59b has a slightly higher homology with the human and the rat homologue (44% and 68%, respectively) than does cd59a (41% and 59%, respectively).
To confirm that cd59b is a GPI-anchored membrane protein, a plasmid encoding an N-terminal FLAG peptide-tagged cd59b was constructed and used in CHO cell transfection experiments. Stable CHO cell lines were generated, and cell surface expression of cd59b was assessed by FACS analysis using a FLAG peptide-specific Ab. A similar plasmid encoding a peptide-tagged cd59a was also constructed and used as a positive control for this experiment. Previous studies have already established that cd59a attaches to the cell membrane via a GPI anchor (21). Fig. 3 shows that both FLAG-tagged cd59b and cd59a were expressed on the surface of CHO cells transfected with FLAG-cd59-containing plasmids. In contrast, no FLAG expression was detected on pAlter vector-transfected CHO cells. Furthermore, after PIPLC treatment, both FLAG-cd59b and FLAG-cd59a were removed from the cell surface, suggesting that cd59b, like cd59a, is a GPI-anchored membrane protein. To determine whether cd59b encodes a protein with MAC-inhibiting activity, the sensitivity of vector or FLAG-cd59b-transfected CHO cells to human and mouse complement-mediated lysis was compared. Fig. 4 shows that cd59b- and cd59a-expressing CHO cells were significantly more resistant than vector-transfected CHO cells to human and mouse complement-mediated lysis. This result established that cd59b is functionally active as a MAC inhibitor.
To explore the tissue expression patterns of cd59a and cd59b in vivo, Northern blot analysis was conducted using total RNAs isolated from the mouse liver, lung, spleen, kidney, testis, and heart. When the membrane was hybridized with a cd59a cDNA probe (corresponding to the 370-bp 5′ sequence), a major 1.3-kb transcript was detected abundantly in the heart, kidney, lung, and liver, but only weakly in the spleen and testis (Fig. 5,A). In addition, there was a 0.6-kb transcript that was detectable in all tissues, but more abundant in the testis and the heart (Fig. 5,A). In contrast, when a cd59b-specific probe was used on the same blot, abundant cd59b mRNAs (mainly 0.5-kb, also a 2.7-kb species) were detected only in the testis and not in any other tissues (Fig. 5 B). Thus, the two cd59 genes appear to be differentially expressed in these tissues.
The identification of a second cd59 gene in the mouse has raised the question as to whether there might be two cd59 genes in other species. To address this question, Southern blot analysis was conducted on genomic DNAs isolated from the mouse, rat, and human. Genomic DNAs were digested with a panel of six restriction enzymes (EcoRI, BamHI, HindIII, XbaI, PstI, NdeI) that were chosen based on the fact that they are not present within the mouse cd59b cDNA probe used for subsequent Southern blot hybridization. All three cDNA probes (mouse, rat, and human) corresponded to areas in the 3′-untranslated regions of the respective CD59 cDNAs and are either known (human and mouse) or likely (rat) to be contained within a single exon (exon 4). The mouse cd59b probe is 95% homologous to the corresponding cd59a sequence, and was therefore expected to hybridize with both cd59 genes. Southern blot of digested mouse genomic DNA showed two restriction fragments with the enzymes EcoRI, XbaI, PstI, and NdeI (Fig. 6). This result confirmed the existence of two separate mouse cd59 genes, as established above. In contrast to the two gene patterns observed in the mouse genomic Southern blot analysis, only one band was observed with each enzyme in the Southern blot analysis of rat and human genomic DNAs (Fig. 6). The only exception was with human DNA digested with the enzyme PstI (Fig. 6). This was due to the presence of a PstI restriction enzyme site within the human CD59 cDNA probe used (28). These data suggest that the rat and human genomes contain only one CD59 gene.
CD59 is a crucial membrane complement-regulatory protein that prevents host cells from MAC-induced cytolysis. The protein plays a particularly important role in the vasculature, as indicated by the severe hematological abnormalities associated with genetic or somatic mutations leading to CD59 dysfunction in human patients (9, 12). Due to DAF and CD59 deficiencies, affected erythrocytes of PNH patients are highly susceptible to autologous complement-mediated lysis (9). The possibility that CD59 may also protect nucleated cells such as endothelial cells from the sublytic effect of MAC has received much attention recently (29, 30). Impaired endothelial CD59 activity may render vascular cells susceptible to MAC-induced proliferation and contribute to the development of vascular proliferative disorders (29, 31). Characterization of animal homologues of human CD59 is an obligatory step in establishing proper animal models for studying the in vivo functions of CD59.
A mouse homologue of human CD59 (cd59a) has previously been characterized (21). In this study, we have identified a second mouse CD59 gene (cd59b) that is structurally related to cd59a but seems to be regulated differently. One complete exon (putative exon 2) of cd59b is 100% identical to cd59a, whereas another exon (putative exon 3) shares 86% identity with cd59a (Table I). By analogy of the human CD59 genomic structure (14), the remaining cd59b cDNA sequence is likely to be contained in two additional exons (exons 1 and 4). The putative exon 4 of cd59b is 84% identical to cd59a, whereas exon 1 has no homology at all with cd59a. These data suggested that exons 2–4 are homologous between the two cd59 genes, whereas exon 1 in the two genes is not related. The existence of two distinct but homologous cd59 genes in the mouse is further supported by data of Southern blot analysis of mouse genomic DNA (Fig. 6). Two separate bands were detected on Southern blot analysis with four different restriction enzymes (Fig. 6).
Based on hydrophobicity analysis and consensus GPI anchor attachment residues (32), we predict that the last 20 aa in cd59b will be removed in the mature protein, with Ser109 being the residue for GPI anchor attachment (Fig. 2). The N-terminal 24 aa in cd59a and cd59b are identical (Fig. 2). Assuming that the first 23 residues correspond to a signal peptide that is cleaved in the mature protein as proposed for cd59a (21), the mature cd59b protein should consist of 86 aa, 11 aa residues longer than the predicted length of mature cd59a (21). The extra amino acids are located primarily at the C terminus, immediately above the putative GPI anchor (Fig. 2). This region has been proposed to provide a stalk between the GPI anchor and the globular CD59 protein and is thought to have little functional relevance (21). In cell transfection experiments using N-terminal peptide-tagged protein constructs, expression of cd59b and cd59a was detected on the cell surface of transfected CHO cells. Furthermore, we demonstrated that cd59b, like cd59a, was susceptible to PIPLC cleavage from the cell surface (Fig. 3). Finally, we showed that in a cell lysis test, CHO cells expressing cd59b were significantly more resistant than vector-transfected cells to human or mouse complement-mediated killing (Fig. 4). Together, these results established that cd59b is a GPI-anchored membrane protein and is functionally active as a MAC inhibitor.
Contrary to a previous study in which cd59a message was detected by RT-PCR in most mouse tissues but could not be detected by Northern blot analysis in any of the tissues examined (21), our Northern blot analysis showed prominent cd59a expression in the mouse heart, kidney, liver, and lung (Fig. 5). In contrast to this wider tissue distribution pattern of cd59a, cd59b mRNAs were only detected by Northern blot in the mouse testis. Interestingly, the predominant form of cd59b mRNA in the testis is a 0.5-kb species. This is substantially shorter than the full-length cd59b cDNA sequence deduced from two overlapping cDNA clones generated by RT-PCR (Fig. 1). It is possible that this shortened form of cd59b mRNA derived from the use of an alternative polyadenylation site. Examination of the 3′-untranslated sequence of the full-length cd59b cDNA identified a putative polyadenylation site (AATAAA, underlined in Fig. 1) approximately 100 bp downstream of the translation stop codon.
The identification of a second mouse cd59 gene prompted us to investigate whether there might also be two CD59 genes in the rat or in the human. Southern blot analysis with a panel of six different restriction enzymes confirmed the existence of two mouse cd59 genes, but suggested that only one CD59 gene is likely to exist in the rat or in the human (Fig. 6). The findings of two mouse cd59 genes and the testis-specific expression of cd59b are reminiscent of mouse daf (33, 34). DAF is a single copy gene both in the human (35, 36) and in the rat (37), whereas two daf genes have been identified in the mouse (33, 34). One mouse gene encodes a GPI-anchored daf protein (GPI-daf), and the other encodes a transmembrane form of daf (TM-daf). Like cd59b, the TM-daf gene is expressed exclusively in the mouse testis (33, 34). The two mouse daf genes are believed to have arisen by gene duplication (33). This mechanism of evolution may apply also to the two mouse cd59 genes.
The prominent expression of cd59b in the mouse testis suggests that cd59b may play an important role in male reproduction. Although it is not yet known within which compartment of the mouse testis cd59b is expressed, immunohistochemical and Western blotting studies using mAbs against human CD59 have revealed that CD59 is expressed abundantly on mature human sperm and on differentiating spermatocyte (condensing spermatids) (38, 39, 40). It has also been demonstrated in vitro that Ab neutralization of CD59 rendered human sperm susceptible to MAC-induced damage, as indicated by increased immobility and membrane permeability (39). Since functional complement components are abundant in the female reproductive tract (41, 42), it has been speculated that CD59 may be involved in protecting the sperm from complement-mediated damage that might be initiated by anti-sperm Abs present in the female reproductive tract (38, 39, 40). With the identification of cd59b, this hypothesis, as well as the precise pattern of CD59 distribution within the testis and the possible roles of CD59 in other aspects of testicular function can be addressed using the mouse as an animal model.
We thank Zhongzhou Yang for technical assistance in the construction of cd59 expression plasmids.
Abbreviations used in this paper: MAC, membrane attack complex; CHO, Chinese hamster ovary; DAF, decay-accelerating factor; PIPLC, phosphatidylinositol-specific phospholipase C; PNH, paroxysmal nocturnal hemoglobinuria; RACE, rapid amplification of cDNA ends.