Human Semaphorin K1 Is Glycosylphosphatidylinositol-linked and Defines a New Subfamily of Viral-related Semaphorins
1998; Elsevier BV; Volume: 273; Issue: 35 Linguagem: Inglês
10.1074/jbc.273.35.22428
ISSN1083-351X
AutoresXiaomei Xu, Sheldon Ng, Zhiliang Wu, Dat Nguyen, Sheila A. Homburger, Cynthia Seidel‐Dugan, Allen Ebens, Yuling Luo,
Tópico(s)Neurogenesis and neuroplasticity mechanisms
ResumoThe semaphorin family contains a large number of secreted and transmembrane proteins, some of which are known to act as repulsive axon guidance cues during development or to be involved in immune function. We report here on the identification of semaphorin K1 (sema K1), the first semaphorin known to be associated with cell surfaces via a glycosylphosphatidylinositol linkage. Sema K1 is highly homologous to a viral semaphorin and can interact with specific immune cells, suggesting that like its viral counterpart, sema K1 could play an important role in regulating immune function. Sema K1 does not bind to neuropilin-1 or neuropilin-2, the two receptors implicated in mediating the repulsive action of several secreted semaphorins, and thus it likely acts through a novel receptor. In contrast to most previously described semaphorins, sema K1 is only weakly expressed during development but is present at high levels in postnatal and adult tissues, particularly brain and spinal cord. The semaphorin family contains a large number of secreted and transmembrane proteins, some of which are known to act as repulsive axon guidance cues during development or to be involved in immune function. We report here on the identification of semaphorin K1 (sema K1), the first semaphorin known to be associated with cell surfaces via a glycosylphosphatidylinositol linkage. Sema K1 is highly homologous to a viral semaphorin and can interact with specific immune cells, suggesting that like its viral counterpart, sema K1 could play an important role in regulating immune function. Sema K1 does not bind to neuropilin-1 or neuropilin-2, the two receptors implicated in mediating the repulsive action of several secreted semaphorins, and thus it likely acts through a novel receptor. In contrast to most previously described semaphorins, sema K1 is only weakly expressed during development but is present at high levels in postnatal and adult tissues, particularly brain and spinal cord. The semaphorins constitute a large family of evolutionally conserved glycoproteins that are defined by a characteristic semaphorin domain of approximately 500 amino acids (1Kolodkin A.L. Matthes D.J. Goodman C.S. Cell. 1993; 75: 1389-1399Abstract Full Text PDF PubMed Scopus (790) Google Scholar, 2Puschel A.-W. Adams R.-H. Betz H. Neuron. 1995; 14: 941-948Abstract Full Text PDF PubMed Scopus (333) Google Scholar, 3Luo Y. Shepherd I. Li J. Renzi M.J. Chang S. Raper J.A. Neuron. 1995; 14: 1131-1140Abstract Full Text PDF PubMed Scopus (215) Google Scholar). The first vertebrate semaphorin, collapsin-1 in chick, was identified by its ability to induce growth cone collapse (4Luo Y. Raible D. Raper J.A. Cell. 1993; 75: 217-227Abstract Full Text PDF PubMed Scopus (1013) Google Scholar). Consistent with this function, its mammalian homologue, sema III, has been shown to repel specific subsets of sensory axons (5Messersmith E.-K. Leonardo E.-D. Shatz C.-J. Tessier-Lavigne M. Goodman C.-S. Kolodkin A.-L. Neuron. 1995; 14: 949-959Abstract Full Text PDF PubMed Scopus (457) Google Scholar). As a result of these and other studies, Coll-1/sema III/D has been implicated in the patterning of sensory axon projections into the ventral spinal cord and cranial nerve projections into the periphery (6Fan J. Raper J.-A. Neuron. 1995; 14: 263-274Abstract Full Text PDF PubMed Scopus (224) Google Scholar, 7Kobayashi H. Koppel A.-M. Luo Y. Raper J.-A. J. Neurosci. 1997; 17: 8339-8352Crossref PubMed Google Scholar, 8Puschel A.-W. Adams R.-H. Betz H. Mol. Cell. Neurosci. 1996; 7: 419-431Crossref PubMed Scopus (142) Google Scholar, 9Behar O. Golden J.-A. Mashimo H. Schoen F.-J. Fishman M.-C. Nature. 1996; 383: 525-528Crossref PubMed Scopus (509) Google Scholar, 10Shepherd I.-T. Luo Y. Lefcort F. Reichardt L.-F. Raper J.A. Development ( Camb. ). 1997; 124: 1377-1385PubMed Google Scholar, 11Taniguchi M. Yuasa S. Fujisawa H. Naruse I. Saga S. Mishina M. Yagi T. Neuron. 1997; 19: 519-530Abstract Full Text Full Text PDF PubMed Scopus (474) Google Scholar). Several other semaphorins have also been implicated as repulsive and/or attractive cues in axon guidance, axon fasciculation, and synapse formation (1Kolodkin A.L. Matthes D.J. Goodman C.S. Cell. 1993; 75: 1389-1399Abstract Full Text PDF PubMed Scopus (790) Google Scholar, 12Kolodkin A.L. Matthes D.J. O'Connor T.P. Patel N.H. Admon A. Bentley D. Goodman C.-S. Neuron. 1992; 9: 831-845Abstract Full Text PDF PubMed Scopus (315) Google Scholar, 13Matthes D.J. Sink H. Kolodkin A.L. Goodman C.S. Cell. 1995; 81: 631-639Abstract Full Text PDF PubMed Scopus (179) Google Scholar, 14Wong J.T. Yu W.T. O'Connor T.P. Development ( Camb. ). 1997; 124: 3597-3607PubMed Google Scholar, 15Adams R.-H. Betz H. Puschel A.-W. Mech. Dev. 1996; 57: 33-45Crossref PubMed Scopus (170) Google Scholar, 16Feiner L. Koppel A.-M. Kobayashi H. Raper J.-A. Neuron. 1997; 19: 539-545Abstract Full Text Full Text PDF PubMed Scopus (181) Google Scholar, 17Yu H.-H. Araj H.-H. Ralls S.-A. Kolodkin A.-L. Neuron. 1998; 20: 207-220Abstract Full Text Full Text PDF PubMed Scopus (142) Google Scholar). In addition, members of the semaphorin family have been implicated in functions outside the nervous system, including bone skeleton and heart formation (9Behar O. Golden J.-A. Mashimo H. Schoen F.-J. Fishman M.-C. Nature. 1996; 383: 525-528Crossref PubMed Scopus (509) Google Scholar), immune function (18Bougeret C. Mansur I.G. Dastot H. Schmid M. Mahouy G. Bensussan A. Boumsell L. J. Immunol. 1992; 148: 318-323PubMed Google Scholar, 19Hall K.-T. Boumsell L. Schultze J.L. Boussiotis V.-A. Dorfman D.-M. Cardoso A.-A. Bensussan A. Nadler L.-M. Freeman G.-J. Proc. Natl. Acad. Sci. U. S. A. 1995; 93: 11780-11785Crossref Scopus (228) Google Scholar), tumor suppression (20Xiang R.-H. Hensel C.-H. Garcia D.-K. Carlson H.-C. Kok K. Daly M.-C. Kerbacher K. van den Berg A. Veldhuis P. Buys C.-H. Naylor S.-L. Genomics. 1996; 32: 39-48Crossref PubMed Scopus (135) Google Scholar, 21Roche J. Boldog F. Robinson M. Robinson L. Varella-Garcia M. Swanton M. Waggoner B. Fishel R. Franklin W. Gemmill R. Drabkin H. Oncogene. 1996; 12: 1289-1297PubMed Google Scholar, 22Sekido Y. Bader S. Latif F. Chen J.-Y. Duh F.-M. Wei M.-H. Albanesi J.-P. Lee C.-C. Lerman M.-I. Minna J.-D. Proc. Natl. Acad. Sci. U. S. A. 1996; 93: 4120-4125Crossref PubMed Scopus (221) Google Scholar), and conferring drug resistance to cells (23Yamada T. Endo R. Gotoh M. Hirohashi S. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 14713-14718Crossref PubMed Scopus (91) Google Scholar). Recent studies have identified the first semaphorin receptor as a member of the neuropilin family. Neuropilin-1 is a high affinity receptor for sema III, E, and IV, whereas neuropilin-2 binds differentially to the subfamily of secreted semaphorins (24He Z. Tessier-Lavigne M. Cell. 1997; 90: 739-751Abstract Full Text Full Text PDF PubMed Scopus (965) Google Scholar, 25Kolodkin A.L. Levengood D.V. Rowe E.G. Tai Y.T. Giger R.J. Ginty D.D. Cell. 1997; 90: 753-762Abstract Full Text Full Text PDF PubMed Scopus (997) Google Scholar, 26Chen H. Chedotal A. He Z. Goodman C.-S. Tessier-Lavigne M. Neuron. 1997; 19: 547-559Abstract Full Text Full Text PDF PubMed Scopus (568) Google Scholar, 27Kitsukawa T. Shimizu M. Sanbo M. Hirata T. Taniguchi M. Bekku Y. Yagi T. Fujisawa H. Neuron. 1997; 19: 995-1005Abstract Full Text Full Text PDF PubMed Scopus (559) Google Scholar). The vertebrate semaphorin family can be classified into several phylogenetically distinct subfamilies (Ref. 15Adams R.-H. Betz H. Puschel A.-W. Mech. Dev. 1996; 57: 33-45Crossref PubMed Scopus (170) Google Scholar, Fig. 1 B). Each subfamily has a unique structural arrangement of protein domains. The secreted members of the semaphorin family contain a characteristic semaphorin domain at the NH2 terminus, followed by an immunoglobulin (Ig) domain and a stretch of basic amino acids in the carboxyl-terminal region. Between the NH2-terminal semaphorin domain and the transmembrane spanning region, the transmembrane semaphorins contain several alternative structural motifs, including either an Ig domain, a stretch of thrombospondin repeats, or a sequence with no obvious domain homology. Interestingly, semaphorin-like sequences have been identified in the genomes of poxviruses (1Kolodkin A.L. Matthes D.J. Goodman C.S. Cell. 1993; 75: 1389-1399Abstract Full Text PDF PubMed Scopus (790) Google Scholar) and alcelaphine herpesvirus-1 (28Ensser A. Fleckenstein B. J. Gen. Virol. 1995; 76: 1063-1067Crossref PubMed Scopus (41) Google Scholar). These virally encoded semaphorins occupy unique branches of the semaphorin phylogenetic tree. Although it has been suggested that these viral semaphorins must have vertebrate homologues, to date none have been reported. Here we report the identification of a GPI 1The abbreviations used are: GPIglycosylphosphatidylinositolAHValcelaphine herpesvirusAPhuman placenta alkaline phosphatasePI-PLCphosphatidylinositol-specific phospholipase CPCRpolymerase chain reactionPAGEpolyacrylamide gel electrophoresisTNFtumor necrosis factorTNFRtumor necrosis factor receptorESTexpressed sequence tags.1The abbreviations used are: GPIglycosylphosphatidylinositolAHValcelaphine herpesvirusAPhuman placenta alkaline phosphatasePI-PLCphosphatidylinositol-specific phospholipase CPCRpolymerase chain reactionPAGEpolyacrylamide gel electrophoresisTNFtumor necrosis factorTNFRtumor necrosis factor receptorESTexpressed sequence tags.-linked human semaphorin, semaphorin K1, which is highly homologous to the semaphorin encoded by alcelaphine herpesvirus-1. The expression and binding properties of sema K1 suggest that it could play an important role in both adult nervous system and in modulating immune function. glycosylphosphatidylinositol alcelaphine herpesvirus human placenta alkaline phosphatase phosphatidylinositol-specific phospholipase C polymerase chain reaction polyacrylamide gel electrophoresis tumor necrosis factor tumor necrosis factor receptor expressed sequence tags. glycosylphosphatidylinositol alcelaphine herpesvirus human placenta alkaline phosphatase phosphatidylinositol-specific phospholipase C polymerase chain reaction polyacrylamide gel electrophoresis tumor necrosis factor tumor necrosis factor receptor expressed sequence tags. Four human ESTs, R33537, W47265, R33439,H03806, and one mouse EST, AA260340, were identified that show highest homology with the semaphorin gene in alcelaphine herpesvirus-1 (AHV sema). Oligonucleotides corresponding to the sequences of human ESTs were used to amplify by PCR a cDNA fragment from a human placenta cDNA library (CLONTECH). This PCR fragment corresponds to the central portion of sema K1. The 3′ end was cloned by rapid amplification of cDNA ends using human placenta Marathon-Ready cDNA from CLONTECH (29Frohman M.A. Methods Enzymol. 1993; 218: 340-356Crossref PubMed Scopus (465) Google Scholar). The remaining 5′ end was cloned by PCR amplification from aCLONTECH human brain λgt11 cDNA library using an internal primer from sema K1 and an anchor primer corresponding to the λgt11 vector sequence. A specific PCR product corresponding to the 5′ end was identified by Southern blot using sema K1 oligonucleotides as probes. Repeated effort has been made, but failed to yield the remaining 5′ end of cDNA. The full-length cDNA of human sema K1 except the region corresponding to the signal peptide sequence was independently cloned from CLONTECHhuman placenta λgt10 library by high fidelity PCR amplification and its DNA sequence reconfirmed. Three expression constructs were made that allow the expression of recombinant proteins tagged with either a Myc-His tag at the carboxyl terminus (pEX-mh), an alkaline phosphatase tag at the amino terminus and a Myc-His tag at the carboxyl terminus (pEX-AP), or an Fc domain of human immunoglobulin at the carboxyl terminus (pEX-Fc). Similar expression constructs have been made for collapsins and semaphorins, and the resulting fusion proteins were shown to be fully functional (7Kobayashi H. Koppel A.-M. Luo Y. Raper J.-A. J. Neurosci. 1997; 17: 8339-8352Crossref PubMed Google Scholar, 10Shepherd I.-T. Luo Y. Lefcort F. Reichardt L.-F. Raper J.A. Development ( Camb. ). 1997; 124: 1377-1385PubMed Google Scholar, 23Yamada T. Endo R. Gotoh M. Hirohashi S. Proc. Natl. Acad. Sci. U. S. A. 1997; 94: 14713-14718Crossref PubMed Scopus (91) Google Scholar, 24He Z. Tessier-Lavigne M. Cell. 1997; 90: 739-751Abstract Full Text Full Text PDF PubMed Scopus (965) Google Scholar, 30Koppel A.-M. Feiner L. Kobayashi H. Raper J.-A. Neuron. 1997; 19: 531-537Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 31Eickholt B.J. Morrow R. Walsh F.S. Doherty P. Mol. Cell Neurosci. 1997; 9: 358-371Crossref PubMed Scopus (28) Google Scholar). The multiple cloning site of pSecTagA (Invitrogen) was excised with PmeI and NheI and cloned into pcDNA3.1 (Invitrogen) to make Myc-His tagged vector pEX-mh. This expression vector contains a signal peptide sequence from immunoglobulin κ chain for protein secretion. The cDNA for human placental alkaline phosphatase was PCR-amplified from pSEAP (CLONTECH) and cloned into theSfiI site of pEX-mh maintaining the original reading frame to make the AP-tagged vector pEX-AP. The Fc domain of human IgG1 and an enterokinase cleavage site were PCR-amplified from Signal-pIgplus (Novagen) and cloned into the ApaI to PmeI sites of pEX-mh maintaining the original reading frame to make the Fc-tagged vector pEX-Fc. Various cDNAs for full-length sema K1, extracellular domain of sema K1 (residues starting from Gly-612 to the carboxyl-terminal end were deleted), sema III, and neuropilin-1 were PCR-amplified from cDNA libraries and subcloned into these expression vectors. The neuropilin-2 expression construct was kindly provided by Dr. Marc Tessier-Lavigne (25Kolodkin A.L. Levengood D.V. Rowe E.G. Tai Y.T. Giger R.J. Ginty D.D. Cell. 1997; 90: 753-762Abstract Full Text Full Text PDF PubMed Scopus (997) Google Scholar). COS-7 cells were transiently transfected with the full-length sema K1 in pEX-AP vector using LipofectAMINE (Life Technologies, Inc.). Two days after transfection, cells were washed and treated with or without PI-PLC (Boehringer Mannheim) at 250 milliunits/ml for 1 h at 37 °C. Cells were then fixed in 4% paraformaldehyde for 10 min at room temperature. After phosphate-buffered saline wash, cells were incubated with a rabbit anti-AP antibody (Accurate Antibodies) at a dilution of 1:500 for 1 h followed by a Cy3-anti-rabbit antibody at a dilution of 1:200. The fluorescent images of the transfected cells were photographed in a Zeiss microscope using a 40× lens. COS-7 cells were transiently transfected with the full-length sema K1 in pEX-AP vector with LipofectAMINE (Life Technologies, Inc.). Cells transfected with the full-length CD100 in pEX-AP served as a control. Two days after transfection, cells were incubated with or without 250 milliunits/ml of PI-PLC (Boehringer Mannheim) for 1 h at 37 °C. Supernatants and cell lysates were collected and run on a 4–20% SDS-PAGE gel and the AP-tagged sema K1 protein was detected with a horseradish peroxidase-conjugated anti-alkaline phosphatase antibody. Stable 293-EBNA cell lines secreting Myc-His-tagged, AP-tagged, or Fc-tagged sema K1 and sema III were derived from transfection of various expression plasmids followed by G418 selection. Conditioned media from stably transfected cell lines were collected and were confirmed for the expression and integrity of recombinant proteins by Western blot using anti-AP, anti-Fc, or anti-Myc antibodies. SDS-PAGE gel demonstrated that sema K1-Fc fusion protein migrates as a dimer linked by the disulfide bonds in the Fc region, whereas the sema K1-mh and AP-sema K1 are monomeric (not shown). Approximately equal amount of AP- or Fc-tagged sema III and sema K1 fusion proteins as judged by Western blot were used in the ligand binding experiments. The amount of active sema III used for the ligand binding experiment was further quantified by a growth cone collapse assay and estimated to be over 80 collapsing units/ml (4Luo Y. Raible D. Raper J.A. Cell. 1993; 75: 217-227Abstract Full Text PDF PubMed Scopus (1013) Google Scholar,7Kobayashi H. Koppel A.-M. Luo Y. Raper J.-A. J. Neurosci. 1997; 17: 8339-8352Crossref PubMed Google Scholar). COS-7 cells were transiently transfected with full-length neuropilin-1 or neuropilin-2 expression constructs with LipofectAMINE (Life Technologies, Inc.). The expression of neuropilin-1 or -2 was confirmed using a monoclonal antibody 9E10 against the Myc tag at the carboxyl-terminal ends of both receptors. After 2 days of transfection, the cells were then incubated with supernatants containing approximately equal amount of sema III-Fc or sema K1-Fc for 1 h. After post-fixing in 1% paraformaldehyde for 10 min, the cells were heat-inactivated at 65 °C for 1 h to destroy the endogenous alkaline phosphatase activity. Cells were then incubated with alkaline phosphatase-conjugated anti-Fc antibody at 1:500 dilution for 1 h and processed for chromogenic AP enzymatic reaction. For the immune cell staining experiment, P388D1 or RBL-2H3 cells were fixed in 1% paraformaldehyde for 10 min. The suspension cells (A20 and Jurkat) were washed in phosphate-buffered saline once and fixed in 1% paraformaldehyde for 10 min and then cytospun onto glass slides. After blocking for 30 min, AP-sema K1- or AP-sema III-containing supernatants were added to each well and incubated for 1 h. The cells were then post-fixed in 100% methanol for 10 min, and the endogenous AP activity was heat-inactivated at 65 °C for 1 h. Cells were then processed for chromogenic AP enzymatic reactions. AP alone was used as a negative control. For experiments in which sema K1-mh or sema III-mh was used to compete with AP-sema K1 or AP-sema III binding, respectively, sema K1-mh or sema III-mh was incubated with different cell lines for 30 min at room temperature prior to AP-sema K1 or AP-sema III incubation. A 298-base pair DNA fragment corresponding to the sequence of mouse EST AA260340was PCR-amplified from a mouse cDNA library. This DNA fragment is predicted to encode a mouse homologue of human sema K1 based on the fact that it shares over 95% amino acid identity with the corresponding region of human sema K1. It was used as a probe in the Northern blot analysis and the in situ hybridization experiments. Northern blot was either purchased fromCLONTECH (Fig. 4 A) or prepared as follows (Fig. 4 B). Adult mouse tissues were homogenized in UltraspecTM RNA reagent (Biotecx), and total RNA was extracted and precipitated. Poly(A)+ RNA was further selected using FastTrack 2.0 mRNA isolation kit (Invitrogen). Two-microgram poly(A)+ RNA from each tissue was run in formaldehyde gels and blotted onto positively charged nylon membranes (Boehringer Mannheim). DNA probes were labeled with [32P]dCTP (Amersham Pharmacia Biotech) using random prime labeling kit of Life Technologies, Inc. Hybridization was performed in NorthernMax Buffer (Ambion) following the manufacturer's instructions. In situhybridization procedure was performed on cryostat sections of E11, E15 mouse embryos, and on brain and spinal cord sections of P3 and 5-week-old mice (C57BL/6J) as described (32Schaeren-Wiemers N. Gerfin-Moser A. Histochemistry. 1993; 100: 431-440Crossref PubMed Scopus (1084) Google Scholar). Tissues were fixed in 4% paraformaldehyde for 4 h at 4 °C and embedded in OCT embedding compound. 20-μm sections were cut and were treated with 1.0 μg/ml proteinase K for 15 min at 37 °C, 0.2 m HCl for 20 min, and then acetylated for 10 min with 0.1 m triethanolamine and 0.25% acetic anhydride. Sections were prehybridized for 1 h at 65 °C, then hybridized with digoxigenin-labeled probes (2 μg/ml) overnight at 55 °C. The hybridization buffer consists of 50% formamide, 5 × SSC, 10% dextran sulfate, 1 × Denhardt's, 0.25 mg/ml tRNA, 0.1 mg/ml single-stranded DNA. After hybridization, slides were washed with 0.2 × SSC for 60 min at 65 °C and detected with an AP-conjugated anti-digoxigenin antibody at a dilution of 1:2000. In an effort to identify veterbrate homologues of viral semaphorins, we searched existing EST data bases against semaphorin-like sequences found in vaccinia virus and in alcelaphine herpesvirus-1 (AHV sema) using the BLAST algorithm (33Altschul S.-F. Gish W. Miller W. Myers E.-W. Lipman D.-J. J. Mol. Biol. 1990; 215: 403-410Crossref PubMed Scopus (69631) Google Scholar). Four human and one mouse ESTs were identified, which encode novel sequences that were most homologous to AHV sema (28Ensser A. Fleckenstein B. J. Gen. Virol. 1995; 76: 1063-1067Crossref PubMed Scopus (41) Google Scholar). PCR primers were designed based on the EST sequences and were used to obtain a 2.5-kilobase pair cDNA that encodes a candidate semaphorin gene. The cDNA contains all the human EST sequences and predicts a protein of 634 amino acids with a molecular mass of 71.5 kDa (Fig. 1 A). This protein is named semaphorin K1 (sema K1), since sema A to J has been designated for other known semaphorins in mouse genome data base and sequence analysis indicates that sema K1 is the first member of a new semaphorin subfamily (see below). Hydropathy analysis of the predicted sema K1 sequence (34Kyte J. Doolittle R.F. J. Mol. Biol. 1982; 157: 105-132Crossref PubMed Scopus (17086) Google Scholar) indicates that the sema K1 sequence is slightly incomplete, lacking approximately half of the signal peptide sequence required for protein secretion (35von Heijne G. J. Mol. Biol. 1985; 184: 99-105Crossref PubMed Scopus (1529) Google Scholar). Consistent with this predication, the alignment between AHV sema and sema K1 also showed an eight-amino acid difference at the amino-terminal end of sema K1 (Fig.1 A). The hydropathy analysis of sema K1 also identified a long stretch of hydrophobic residues at the carboxyl-terminal end, which resembles a signal peptide sequence required for GPI anchorage (36Udenfriend S. Kodukula K. Annu. Rev. Biochem. 1995; 64: 563-591Crossref PubMed Scopus (434) Google Scholar). Thus, sema K1 is predicted to be the first GPI-linked membrane protein in the semaphorin family. The sequence of sema K1 is closely related to that of AHV sema. Whereas the sema domain of sema K1 shares 50% amino acid identity with that of AHV sema, it shares less than 30% identity with that of other known semaphorins. In addition, 17 out of 18 cysteine residues and 4 out of 5 potential N-linked glycosylation sites are conserved (Fig.1 A). The homology extends throughout the entire amino acid sequences of sema K1 and AHV sema except at the carboxyl-terminal end, where only sema K1 contains the signal peptide sequence for GPI anchorage. Thus, sema K1 is predicted to be a GPI-anchored membrane protein, whereas AHV sema is likely to be a secreted protein. The unique structural arrangement of sema K1 suggests that it may define a new subfamily of vertebrate semaphorins. Consistent with this hypothesis, protein sequence homology analysis showed that sema K1 and AHV sema belong to the same branch of the dendrogram tree and this branch is distinct from that of other semaphorins (Fig. 1 B). Sequence alignment with other semaphorins also revealed that members of the viral-related semaphorin subfamily lack three tryptophan residues conserved in other semaphorins, suggesting that viral-related semaphorins may contain a structurally distinct sema domain. To confirm that sema K1 is a GPI-anchored membrane protein, we have transfected COS-7 cells with a sema K1 expression construct and determined the localization of the expressed sema K1 protein. To track sema K1 protein expression, an AP-tagged version of sema K1 was engineered in which the human placenta alkaline phosphatase was fused to the full-length sema K1 at the NH2 terminus. This fusion protein can be detected with an anti-AP antibody or AP activity. Upon transfection of the expression construct into COS-7 cells, the sema K1 fusion protein was detected on the surface of those transfected cells (Fig.2 A). Treatment with PI-PLC resulted in a complete removal of the fusion protein from cell surfaces (Fig. 2 B). To examine whether the release of sema K1 fusion protein from cell surfaces is a specific action of PI-PLC rather than the result of random proteolysis, we compared the presence of this fusion protein in the supernatant and lysate of transfected COS-7 cells with or without PI-PLC treatment. Supernatants and lysates from PI-PLC treated or untreated cells were subjected to Western blot analysis. A 150-kDa protein corresponding to the predicted size of the fusion protein was detected with the anti-AP antibody. When the transfected COS-7 cells were not treated with PI-PLC, most, if not all, of the fusion protein was found to be associated with the cell lysate (Fig.2 C). Treatment of these cells with PI-PLC resulted in significant release of the fusion protein from the cell lysate into the supernatant, without apparent proteolysis (Fig. 2 C). In a control experiment, PI-PLC treatment did not release the transmembrane semaphorin CD100 into the cell supernatant (not shown). Furthermore, when a stop codon was introduced immediately NH2-terminal to the predicted signal peptide sequence for GPI linkage (residues starting from Gly-612 to the carboxyl-terminal end were deleted), the resultant sema K1 protein was released to the cell supernatant in the absence of PI-PLC (not shown). Thus, we conclude that sema K1 is attached to the cell membrane via a GPI linkage. Neuropilin-1 and neuropilin-2 have recently been identified as receptors or components of a receptor complex for sema III and other secreted semaphorins (24He Z. Tessier-Lavigne M. Cell. 1997; 90: 739-751Abstract Full Text Full Text PDF PubMed Scopus (965) Google Scholar, 25Kolodkin A.L. Levengood D.V. Rowe E.G. Tai Y.T. Giger R.J. Ginty D.D. Cell. 1997; 90: 753-762Abstract Full Text Full Text PDF PubMed Scopus (997) Google Scholar, 26Chen H. Chedotal A. He Z. Goodman C.-S. Tessier-Lavigne M. Neuron. 1997; 19: 547-559Abstract Full Text Full Text PDF PubMed Scopus (568) Google Scholar). To determine whether sema K1 could use neuropilin-1 or -2 as its receptor, we tested the ability of sema K1 to bind COS-7 cells transfected with neuropilin expression constructs. Soluble sema K1 fusion proteins containing either an AP tag at the NH2 terminus (AP-sema K1), an Fc domain of human IgG1 at the COOH terminus (sema K1-Fc), or a Myc-His tag at the COOH terminus (sema K1-mh) were produced and were used in the ligand binding assay. Similarly arranged AP-sema III, sema III-Fc, and sema III-mh fusion proteins were prepared as controls. To test for interactions with neuropilin-1 or -2, sema K1-Fc, or AP-sema K1 were incubated with neuropilin-expressing COS-7 cells, and ligand binding was detected using an anti-Fc antibody or a chromogenic AP enzymatic reaction. Under conditions where sema III-Fc binds to COS-7 cells expressing neuropilin-1 or -2, the dimerized sema K1-Fc does not bind to either (Fig. 3, A–D, note that sema III binds to neuropilin-2 with lower affinity than to neuropilin-1). In this experiment, sema III-Fc does not bind to mock-transfected COS-7 cells and Fc-tagged MUC18, an adhesion molecule served as a control, does not bind to the neuropilin-expressing COS-7 cells (not shown). Both served as negative controls. Similarly, under conditions when AP-sema III can bind to COS-7 cells expressing neuropilin-1 or -2, the monomeric AP-sema K1 does not bind to these cells (data not shown). Thus, sema K1 does not bind neuropilin-1 or -2 with high affinity and may not act through these receptors. To determine whether or not the soluble sema K1 fusion proteins are competent to bind a cognate receptor and to provide an entry point for investigating the role of sema K1 in modulating immune function, we analyzed several immune cell lines for the presence of sema K1 binding sites. AP-sema K1 or AP-sema III were incubated with Jurkat T cells, A20 B cells, P388D1 macrophages, and RBL-2H3 mast cell lines and the bound ligands were detected with chromogenic AP enzymatic reaction (Fig. 3, E–J). AP-sema K1 bound only to the cell surfaces of P388D1 macrophage and RBL-2H3 mast cell lines. This binding is specific, since AP alone does not bind to any of the cell lines, and the binding could be competed by preincubation with sema K1-mh. In comparison, AP-sema III binding was detected on cell surfaces of all four immune cell lines tested. This binding is also specific, since preincubation of these cells with sema III-mh blocks the binding (not shown). The ability of sema III-Fc or sema K1-Fc to bind these four cell lines was also tested and similar results obtained (not shown). We conclude that sema III can bind the four immune cell lines tested, which contrasts with the more selective binding of sema K1 to macrophage and mast cell lines, suggesting the existence of a specific receptor for sema K1 in these cell lines. To help define the biological role of sema K1, we examined the expression of sema K1 by Northern blot analysis andin situ hybridization. A 298-base pair cDNA corresponding to the mouse homologue of human sema K1 was used as a probe in these studies. This probe does not cross-hybridize with the mRNA of other semaphorins. Northern blot analysis of mRNA isolated from adult mouse tissues revealed a single sema K1 transcript at 4.4 kilobase pairs (Fig. 4). The sema K1 transcript is highly expressed in brain, spinal cord, lung, and testis; moderately expressed in heart, muscle, adrenal gland, lymph nodes, thymus, and intestine; weakly expressed in spleen and kidney; and not detectable in liver, bone marrow, and stomach. To examine the distribution of sema K1 mRNA in detail, in situ hybridization analysis was performed on tissue sections of embryonic day 11 and day 15 embryos, and on the
Referência(s)