Functional Classification of ADAMs Based on a Conserved Motif for Binding to Integrin α9β1
2002; Elsevier BV; Volume: 277; Issue: 20 Linguagem: Inglês
10.1074/jbc.m200086200
ISSN1083-351X
AutoresKoji Eto, Clotilde Huet, Takehiko Tarui, S.V. Kupriyanov, Hai-Zhen Liu, Wilma Puzon-McLaughlin, Xi-Ping Zhang, Dean Sheppard, Eva Engvall, Yoshikazu Takada,
Tópico(s)Biochemical and Structural Characterization
ResumoADAMs (a disintegrin and metalloproteases) are members of the metzincin superfamily of metalloproteases. Among integrins binding to disintegrin domains of ADAMs are α9β1 and αvβ3, and they bind in an RGD-independent and an RGD-dependent manner, respectively. Human ADAM15 is the only ADAM with the RGD motif in the disintegrin domain. Thus, both integrin α9β1 and αvβ3 recognize the ADAM15 disintegrin domain. We determined how these integrins recognize the ADAM15 disintegrin domain by mutational analysis. We found that the Arg481 and the Asp-Leu-Pro-Glu-Phe residues (residues 488–492) were critical for α9β1 binding, but the RGD motif (residues 484–486) was not. In contrast, the RGD motif was critical for αvβ3 binding, but the other residues flanking the RGD motif were not. As the R X6DLPEF α9β1recognition motif (residues 481–492) is conserved among ADAMs, except for ADAM10 and 17, we hypothesized that α9β1 may recognize disintegrin domains in all ADAMs except ADAM10 and 17. Indeed we found that α9β1 bound avidly to the disintegrin domains of ADAM1, 2, 3, and 9 but not to the disintegrin domains of ADAM10 and 17. As several ADAMs have been implicated in sperm-oocyte interaction, we tested whether the functional classification of ADAMs, based on specificity for integrin α9β1, applies to sperm-egg binding. We found that the ADAM2 and 15 disintegrin domains bound to oocytes, but the ADAM17 disintegrin domain did not. Furthermore, the ADAM2 and 15 disintegrin domains effectively blocked binding of sperm to oocytes, but the ADAM17 disintegrin domain did not. These results suggest that oocytes and α9β1 have similar binding specificities for ADAMs and that α9β1, or a receptor with similar specificity, may be involved in sperm-egg interaction during fertilization. As α9β1 is a receptor for many ADAM disintegrins and α9β1 and ADAMs are widely expressed, α9β1-ADAM interaction may be of a broad biological importance. ADAMs (a disintegrin and metalloproteases) are members of the metzincin superfamily of metalloproteases. Among integrins binding to disintegrin domains of ADAMs are α9β1 and αvβ3, and they bind in an RGD-independent and an RGD-dependent manner, respectively. Human ADAM15 is the only ADAM with the RGD motif in the disintegrin domain. Thus, both integrin α9β1 and αvβ3 recognize the ADAM15 disintegrin domain. We determined how these integrins recognize the ADAM15 disintegrin domain by mutational analysis. We found that the Arg481 and the Asp-Leu-Pro-Glu-Phe residues (residues 488–492) were critical for α9β1 binding, but the RGD motif (residues 484–486) was not. In contrast, the RGD motif was critical for αvβ3 binding, but the other residues flanking the RGD motif were not. As the R X6DLPEF α9β1recognition motif (residues 481–492) is conserved among ADAMs, except for ADAM10 and 17, we hypothesized that α9β1 may recognize disintegrin domains in all ADAMs except ADAM10 and 17. Indeed we found that α9β1 bound avidly to the disintegrin domains of ADAM1, 2, 3, and 9 but not to the disintegrin domains of ADAM10 and 17. As several ADAMs have been implicated in sperm-oocyte interaction, we tested whether the functional classification of ADAMs, based on specificity for integrin α9β1, applies to sperm-egg binding. We found that the ADAM2 and 15 disintegrin domains bound to oocytes, but the ADAM17 disintegrin domain did not. Furthermore, the ADAM2 and 15 disintegrin domains effectively blocked binding of sperm to oocytes, but the ADAM17 disintegrin domain did not. These results suggest that oocytes and α9β1 have similar binding specificities for ADAMs and that α9β1, or a receptor with similar specificity, may be involved in sperm-egg interaction during fertilization. As α9β1 is a receptor for many ADAM disintegrins and α9β1 and ADAMs are widely expressed, α9β1-ADAM interaction may be of a broad biological importance. ADAMs 1The abbreviations used are: ADAMa disintegrin and metalloproteaseCHOChinese hamster ovaryGSTglutathione S-transferaseEGFepidermal growth factorBSAbovine serum albumin (a disintegrin andmetalloproteases) or MDC (metalloprotease/disintegrin/cysteine-rich) proteins are a family of transmembrane glycoproteins of more than 30 members (see www.people.virginia.edu/∼jag6n/Table_of_the_ADAMs.html and www.gene.ucl.ac.uk/nomenclature/genefamily/metallo. html). ADAMs have a prodomain, a metalloprotease domain, a disintegrin domain, a cysteine-rich region, an EGF-like domain, a transmembrane domain, and a cytoplasmic tail (1Weskamp G. Blobel C.P. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 2748-2751Crossref PubMed Scopus (75) Google Scholar, 2Wolfsberg T.G. Straight P.D. Gerena R.L. Huovila A.P. Primakoff P. Myles D.G. White J.M. Dev. Biol. 1995; 169: 378-383Crossref PubMed Scopus (360) Google Scholar, 3Black R.A. White J.M. Curr. Opin. Cell Biol. 1998; 10: 654-659Crossref PubMed Scopus (430) Google Scholar). Several ADAMs are involved in crucial biological processes such as fertilization (ADAM1, 2, and 3) (4Wolfsberg T. White J. Dev. Biol. 1996; 180: 389-401Crossref PubMed Scopus (217) Google Scholar, 5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar, 6Evans J.P. Schultz R.M. Kopf G.S. J. Cell Sci. 1995; 108: 3267-3278Crossref PubMed Google Scholar) and muscle cell differentiation (meltrin-α, ADAM12) (7Yagami-Hiromasa T. Sato T. Kurisaki T. Kamijo K. Nabeshima Y. Fujisawa-Sehara A. Nature. 1995; 377: 652-656Crossref PubMed Scopus (439) Google Scholar, 8Gilpin B.J. Loechel F. Mattei M.G. Engvall E. Albrechtsen R. Wewer U.M. J. Biol. Chem. 1998; 273: 157-166Abstract Full Text Full Text PDF PubMed Scopus (284) Google Scholar). Most, but not all, ADAMs have a catalytically active metalloprotease domain, which processes several biologically important cell surface proteins including tumor necrosis factor-α (tumor necrosis factor converting enzyme (TACE), ADAM17) (for reviews, see Refs. 9Blobel C.P. Cell. 1997; 90: 589-592Abstract Full Text Full Text PDF PubMed Scopus (333) Google Scholar and10Werb Z. Yan Y. Science. 1998; 282: 1279-1280Crossref PubMed Google Scholar), Alzheimer protein precursor (ADAM10 and 17) (11Buxbaum J.D. Liu K.N. Luo Y. Slack J.L. Stocking K.L. Peschon J.J. Johnson R.S. Castner B.J. Cerretti D.P. Black R.A. J. Biol. Chem. 1998; 273: 27765-27767Abstract Full Text Full Text PDF PubMed Scopus (839) Google Scholar, 12Lammich S. Kojro E. Postina R. Gilbert S. Pfeiffer R. Jasionowski M. Haass C. Fahrenholz F. Proc. Natl. Acad. Sci. U. S. A. 1999; 96: 3922-3927Crossref PubMed Scopus (986) Google Scholar), Delta (ADAM10) (9Blobel C.P. Cell. 1997; 90: 589-592Abstract Full Text Full Text PDF PubMed Scopus (333) Google Scholar, 13Pan D. Rubin G.M. Cell. 1997; 90: 271-280Abstract Full Text Full Text PDF PubMed Scopus (361) Google Scholar, 14Qi H. Rand M.D., Wu, X. Sestan N. Wang W. Rakic P., Xu, T. Artavanis-Tsakonas S. Science. 1999; 283: 91-94Crossref PubMed Scopus (375) Google Scholar), and heparin-binding EGF (ADAM9) (15Izumi Y. Hirata M. Hasuwa H. Iwamoto R. Umata T. Miyado K. Tamai Y. Kurisaki T. Sehara-Fujisawa A. Ohno S. Mekada E. EMBO J. 1998; 17: 7260-7272Crossref PubMed Scopus (475) Google Scholar). a disintegrin and metalloprotease Chinese hamster ovary glutathione S-transferase epidermal growth factor bovine serum albumin The ADAM disintegrin domains are homologous to snake venom disintegrins and are potential integrin ligands. Snake venom disintegrins bind to the platelet integrin αIIbβ3 in an RGD-dependent manner. Of all known ADAMs, ADAM15 is the only one that has the RGD motif in the disintegrin domain. Indeed, as we have reported, the ADAM15 disintegrin domain binds to integrin αvβ3 in an RGD-dependent manner (16Zhang X.P. Kamata T. Yokoyama K. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1998; 273: 7345-7350Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). However, ADAM15 binds also to α9β1 in an RGD-independent manner (17Eto K. Puzon-McLaughlin W. Sheppard D. Sehara-Fujisawa A. Zhang X.P. Takada Y. J. Biol. Chem. 2000; 275: 34922-34930Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar) and to integrin α5β1 (18Nath D. Slocombe P.M. Stephens P.E. Warn A. Hutchinson G.R. Yamada K.M. Docherty A.J. Murphy G. J. Cell Sci. 1999; 112: 579-587Crossref PubMed Google Scholar). Several laboratories, including ours, have reported that the non-RGD ADAM disintegrin domains interact with several integrins such as ADAM2 and 9 with α6β1 (5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar, 19Nath D. Slocombe P.M. Webster A. Stephens P.E. Docherty A.J. Murphy G. J. Cell Sci. 2000; 113: 2319-2328Crossref PubMed Google Scholar), ADAM12 with α9β1(17), ADAM23 with αvβ3 (20Cal S. Freije J.M. Lopez J.M. Takada Y. Lopez-Otin C. Mol. Biol. Cell. 2000; 11: 1457-1469Crossref PubMed Scopus (108) Google Scholar), and ADAM28 with α4β1 (21Bridges L.C. Tani P.H. Hanson K.R. Roberts C.M. Judkins M.B. Bowditch R.D. J. Biol. Chem. 2002; 277: 3784-3792Abstract Full Text Full Text PDF PubMed Scopus (78) Google Scholar). At least three ADAMs have been shown to participate in fertilization (ADAM1, 2, and 3) (Refs. 5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar and 22Yuan R. Primakoff P. Myles D.G. J. Cell Biol. 1997; 137: 105-112Crossref PubMed Scopus (203) Google Scholar, and for review, see Ref. 23Evans J.P. Bioessays. 2001; 23: 628-639Crossref PubMed Scopus (135) Google Scholar). The ADAMs on the sperm surface are processed and lack the pro- and metalloprotease domains; thus the disintegrin domains of ADAMs may be the most important in sperm-egg binding. Recent reports showed that several amino acid residues within the putative integrin-binding loop of the ADAM disintegrin domain are critical for sperm-egg interaction (24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar, 25Zhu X. Bansal N.P. Evans J.P. J. Biol. Chem. 2000; 275: 7677-7683Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 26Takahashi Y. Bigler D. Ito Y. White J.M. Mol. Biol. Cell. 2001; 12: 809-820Crossref PubMed Scopus (112) Google Scholar). The identity of the integrin(s) to which the disintegrin domains bind has not been fully established. As antibodies to integrin α6 were shown to block sperm-egg binding, it has been proposed that integrin α6β1 on the egg binds to ADAM2 and 3 on the sperm (26Takahashi Y. Bigler D. Ito Y. White J.M. Mol. Biol. Cell. 2001; 12: 809-820Crossref PubMed Scopus (112) Google Scholar). However, oocytes from α6-null mice can be fertilized in vitro (27Miller B.J. Georges-Labouesse E. Primakoff P. Myles D.G. J. Cell Biol. 2000; 149: 1289-1296Crossref PubMed Scopus (164) Google Scholar), and anti-integrin α6 subunit monoclonal antibody does not always block sperm binding (28Evans J.P. Kopf G.S. Schultz R.M. Dev. Biol. 1997; 187: 79-93Crossref PubMed Scopus (135) Google Scholar). Thus, it is still unclear whether α6β1 is the main or only integrin involved in sperm-egg interaction, and the specificity, occurrence, and significance of integrin-ADAM binding in other cells and tissues are not known (for review, see Ref. 23Evans J.P. Bioessays. 2001; 23: 628-639Crossref PubMed Scopus (135) Google Scholar). In the present study, we analyzed how integrins αvβ3 and α9β1recognize the ADAM15 disintegrin domain by mutating amino acid residues in the putative integrin-binding site of the disintegrin domain (Fig. 1 a). We found that αvβ3 and α9β1 recognize distinct motifs in the disintegrin domain, the RGD (residues 484–486) and the R X6DLPEF (residues 481–492) motifs, respectively. The R X6DLPEF α9β1 recognition motif is conserved among ADAMs, except for ADAM10 and 17, and we provide evidence that α9β1 recognizes several, perhaps all, ADAM disintegrins with this motif. We also found that oocytes and α9β1 have similar binding specificities for ADAM disintegrin domains and propose that α9β1, or a receptor with similar specificity, may be involved in sperm-oocyte interaction during fertilization. Considering that α9β1 and ADAMs are widely expressed, α9β1-ADAM interactions may have a broad significance in many biological and pathological processes such as fertilization, development, and tissue remodeling. Complementary DNA fragments encoding the disintegrin domain of ADAMs were amplified by polymerase chain reaction and cloned in a pGEX-2T vector (Amersham Biosciences) as described previously (16Zhang X.P. Kamata T. Yokoyama K. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1998; 273: 7345-7350Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). The disintegrin domains were derived from mouse ADAM1 (Arg246–Gln334), mouse ADAM2 (Lys388–Pro479), mouse ADAM3 (Gly394–Glu486), human ADAM9 (Ser413–Tyr503), mouse ADAM10 (Gly457–Thr554), human ADAM15 (Met420–Glu510), mouse ADAM15 (Met421–Glu511), and mouse ADAM17 (Ser474–Thr565). GST fusion proteins were produced and purified as described previously (16Zhang X.P. Kamata T. Yokoyama K. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1998; 273: 7345-7350Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). Absorbance at 280 nm was measured to determine the concentration of purified proteins, and the amount of proteins was calculated as described previously (16Zhang X.P. Kamata T. Yokoyama K. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1998; 273: 7345-7350Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). In some experiments, because GST binds to the egg plasma membrane (24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar), disintegrin domains used for bead coating (see below) were released from the GST by incubating with thrombin (1 unit/mg of protein) for 6 h at room temperature, and the free GST was removed by glutathione-agarose affinity chromatography. CHO cells expressing αvβ3, human α6, or human α9 (designated β3-, α6-, or α9-CHO cells, respectively) have been described elsewhere (17Eto K. Puzon-McLaughlin W. Sheppard D. Sehara-Fujisawa A. Zhang X.P. Takada Y. J. Biol. Chem. 2000; 275: 34922-34930Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar). Adhesion assays were performed as described previously (16Zhang X.P. Kamata T. Yokoyama K. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1998; 273: 7345-7350Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). Briefly, wells of 96-well Immulon-2 microtiter plates (Dynatech Laboratories, Chantilly, VA) were coated with substrates in 100 μl of PBS (10 mm phosphate buffer, 0.15 m NaCl, pH 7.4) overnight at 4 °C. Remaining protein-binding sites were blocked by incubating with 1% bovine serum albumin (BSA) (Calbiochem) for 1 h at room temperature. After washing with PBS, CHO cells (105 cells/well) in 100 μl of Dulbecco's modified Eagle's medium supplemented with 1% BSA were added to the wells and incubated at 37 °C for 1 h. After unbound cells were removed by rinsing the wells with Dulbecco's modified Eagle's medium, bound cells were quantified by measuring endogenous phosphatase activity (29Prater C.A. Plotkin J. Jaye D. Frazier W.A. J. Cell Biol. 1991; 112: 1031-1040Crossref PubMed Scopus (188) Google Scholar). Three-week-old B6129PF1/J female mice (Jackson Laboratory) were superovulated using standard hormonal treatment. Oocytes were collected 12–13 h after administration of human chorionic gonadotrophin, and cumulus cells were removed by incubation with hyaluronidase in Flushing and Holding medium (Specialty Media) for 5 min at 37 °C. The zona pellucida were softened in Flushing and Holding medium containing 10 μg/ml α-chymotrypsin (Sigma) and then removed by passing the eggs through a narrow pipette as described previously (5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar). Zona pellucida-free eggs were allowed to recover in fertilization medium composed of Human Tubal Fluid medium (Irvine Scientific) supplemented with 5 mg/ml BSA fraction V (Sigma) for 1–2 h at 37 °C under mineral oil in a 5% CO2 atmosphere. Sperms were collected from 3–6-month-old B6129F1 male mice (Taconic Farms) by placing the cauda of the epididymis and the vas deferens in 1 ml of fertilization medium under mineral oil. Each tissue was slit open with the edge of an injection needle. Sperms were allowed to swim out for 20 min at 37 °C in an atmosphere of 5% CO2. Tissues were removed from the medium, and spermatozoa were capacitated for 2–3 h under the same conditions. The sperm concentration was estimated with a hemocytometer. Fluorescent beads (0.2-μm yellow-green sulfate microspheres, Molecular Probes, Inc.) were coated overnight at 4 °C with purified GST-free disintegrin domains (0.3 mg/ml), washed with PBS, quenched for 1 h at 4 °C in fertilization medium containing 3% BSA, and resuspended by sonication with a water bath sonicator just before use. Zona pellucida-free eggs were incubated in 25-μl drops of 0.1% (v/v) coated beads in 3% BSA in fertilization medium under mineral oil for 1 h at 37 °C in a 5% CO2 atmosphere. Eggs were then washed three times with fertilization medium, fixed in 0.5% glutaraldehyde, and analyzed by confocal microscopy (Bio-Rad, MRC 1024). To confirm that beads were coated with equal amounts of recombinant protein, aliquots of coated beads were boiled in Laemmli SDS sample buffer and subjected to SDS-polyacrylamide gel electrophoresis followed by staining with Coomassie Blue. Zona pellucida-free eggs were preincubated with GST fusion proteins in 100-μl drops of fertilization medium under mineral oil for 30 min at 37 °C in a 5% CO2 atmosphere. The concentration of GST elution buffer (100 mm Tris, 5 mm reduced glutathione) was adjusted to 10% (v/v) final concentration. Sperm was then added into the drops (2.5 × 105 sperm/ml final concentration) and incubated for 1 h. Eggs were washed by removing medium from the drop and adding fresh Flushing and Holding medium three times. Finally eggs were fixed by adding glutaraldehyde to the drops to a final concentration of 0.5%. The number of spermatozoa bound to each egg was counted immediately under a microscope. Aliquots of the GST fusion proteins used were subjected to SDS-PAGE under reducing conditions, transferred onto nitrocellulose membrane, and stained with Ponceau S to confirm the amount and the quality of recombinant proteins at the end of the incubation. Control in vitro fertilization was routinely performed at ∼50 × 106 sperm/ml to ascertain the quality of gametes and to verify that more than 90% of the eggs could be fertilized. Human ADAM15 is the only ADAM with the RGD motif within the putative integrin-binding sites of disintegrin domains (Fig. 1 a). We have reported that αvβ3 and α9β1 recognize the ADAM15 disintegrin domain in an RGD-dependent and an RGD-independent manner, respectively (17Eto K. Puzon-McLaughlin W. Sheppard D. Sehara-Fujisawa A. Zhang X.P. Takada Y. J. Biol. Chem. 2000; 275: 34922-34930Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar). To identify how different integrins recognize the ADAM15 disintegrin domain, we introduced mutations around the RGD motif. The mutant ADAM15 disintegrin domains were synthesized in bacteria as GST fusion proteins. The ability of the mutant disintegrin domains to bind to αvβ3 and α9β1 was determined in cell adhesion assays with CHO cells expressing recombinant αvβ3or α9β1 (designated β3-CHO cells and α9-CHO cells, respectively). CHO cells transfected with expression vector (designated mock-CHO cells) were used as controls. CHO cells express endogenous αvβ1, αvβ5, and α5β1 (30Takagi J. Kamata T. Meredith J. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1997; 272: 19794-19800Abstract Full Text Full Text PDF PubMed Scopus (112) Google Scholar), but these integrins do not bind to ADAM15 (16Zhang X.P. Kamata T. Yokoyama K. Puzon-McLaughlin W. Takada Y. J. Biol. Chem. 1998; 273: 7345-7350Abstract Full Text Full Text PDF PubMed Scopus (197) Google Scholar). Fig. 1 b shows that mutating the RGD motif to SGA completely blocked the binding of αvβ3 but had no detectable effect on the binding of α9β1 to the ADAM15 disintegrin domain, consistent with the previous report (17Eto K. Puzon-McLaughlin W. Sheppard D. Sehara-Fujisawa A. Zhang X.P. Takada Y. J. Biol. Chem. 2000; 275: 34922-34930Abstract Full Text Full Text PDF PubMed Scopus (202) Google Scholar). We found that the R481A, C487A, D488A, L489A, E491, and F492A mutations negatively affected adhesion of α9-CHO cells to the ADAM15 disintegrin domain (Fig. 1, c and d). In contrast, the same mutations flanking the RGD motif did not significantly affect adhesion of β3-CHO cells to the ADAM15 disintegrin domain (Fig. 1, e and f). Mutating the Asp488 and Glu491 simultaneously did not further affect the adhesion of α9- or β3-CHO cells to the ADAM15 disintegrin domain (Fig. 1,d and f). Altogether these results show that the αvβ3-ADAM15 interaction requires the RGD motif but not the surrounding residues. Thus, αvβ3 and α9β1recognize ADAM15 in distinct manners. Interestingly the residues critical for α9β1 binding in the disintegrin domain, including Asp488, have also been identified as critical for sperm-egg binding in ADAM2 and 3 (24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar, 25Zhu X. Bansal N.P. Evans J.P. J. Biol. Chem. 2000; 275: 7677-7683Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 26Takahashi Y. Bigler D. Ito Y. White J.M. Mol. Biol. Cell. 2001; 12: 809-820Crossref PubMed Scopus (112) Google Scholar). The sperm-egg binding studies were performed without reference to any specific integrin, but the results are compatible with α9β1 being the oocyte receptor. The α9β1 binding motif in ADAM15 is conserved in many ADAMs. The alignment of the sequences of the putative integrin-binding regions of several ADAM disintegrin domains is shown in Fig. 2 a. The Arg481, Asp488, Leu489, Pro490, Glu491, and Phe492 residues that are critical for α9β1 binding in ADAM15 (designated the R X6DLPEF motif) are conserved among the disintegrin domains of all mammalian ADAMs except for ADAM10, which has the sequence R X5AREGI, and ADAM17, which has the sequence Q X7KGVSY, in this region. We thus hypothesized that α9β1 is a common receptor for many, if not all, ADAM disintegrin domains with this motif. To address this hypothesis, we tested the disintegrin domains of several ADAMs for interaction with α9β1 in cell attachment assays (Fig. 2,b and c). We found that α9-CHO cells adhered to the disintegrin domains from ADAM1, 2, 3, and 9 as well as 15 used as a positive control. α9-CHO cells do not spread very well on these disintegrins. The α9-CHO cells did not significantly adhere to the disintegrin domains from ADAM10 or 17. Control mock-transfected CHO cells did not significantly adhere to any of the ADAMs tested. We further found that adhesion of α9-CHO cells to the ADAMs was completely blocked by anti-α9 monoclonal antibody Y9A2 (data not shown), showing that the adhesion was α9-specific. These results indicate that α9β1 may be a common receptor for ADAM disintegrin domains with the R X6DLPEF-like motif. Since α6β1 has been implicated in ADAM2 and ADAM3 binding (26Takahashi Y. Bigler D. Ito Y. White J.M. Mol. Biol. Cell. 2001; 12: 809-820Crossref PubMed Scopus (112) Google Scholar), we tested whether α6β1expressed in CHO cells bound to the ADAM2 and 3 disintegrin domains. We found that α6β1 did not mediate adhesion of α6-CHO cells to ADAM2 and 3 under the conditions in which α9-CHO cells adhered to these ligands (data not shown). Several residues in the putative integrin-binding region of ADAM1, 2, and 3 have been identified as critical for sperm-egg binding (24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar, 25Zhu X. Bansal N.P. Evans J.P. J. Biol. Chem. 2000; 275: 7677-7683Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar, 26Takahashi Y. Bigler D. Ito Y. White J.M. Mol. Biol. Cell. 2001; 12: 809-820Crossref PubMed Scopus (112) Google Scholar), but it has not been established which receptor(s) may be involved. Since ADAM1, 2, and 3 disintegrins bind avidly to α9β1, we wanted to test whether the binding of sperm to the oocyte may have a specificity similar to disintegrin binding to α9β1. To address this hypothesis, we first studied whether the recombinant ADAM disintegrin domains interact with oocytes. Although ADAM15 and 17 have not been shown to be present on sperm, they were used to test specificity. We tested whether fluorescent beads coated with disintegrin domains of ADAM2, 15, and 17 bind to zona pellucida-free eggs. As GST alone binds to such eggs (24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar), the GST portion of the recombinant fusion proteins was cleaved off, and the recombinant GST-free disintegrin domains were purified prior to coating of the beads. Fig. 3 shows that the beads coated with the ADAM15 disintegrin domain bound efficiently to the oocytes, the binding being largely concentrated to the microvillar region of the egg (Fig. 3 b, inset). ADAM2 disintegrin domain-coated beads also bound to the egg microvillar region (Fig. 3 a) but to a lesser extent than ADAM15-coated beads. Beads coated with the ADAM17 disintegrin domain (Fig. 3 c) and BSA-coated beads (not shown) did not bind to eggs. We next tested whether recombinant disintegrin domains, used as GST fusion proteins, would inhibit the binding of sperm to oocytes. The disintegrin domains from ADAM2 and ADAM15 inhibited the binding of sperm to mouse eggs in a dose-dependent manner, whereas that from ADAM17 did not affect sperm-egg binding (Fig. 4). The ADAM15 disintegrin domain was a more potent inhibitor than the ADAM2 disintegrin domain, blocking sperm-egg binding completely at 3 μm. Consistent with previous reports, the ADAM2 disintegrin domain at 3 μminhibited sperm-egg binding to about 50%. Altogether, these results indicate that the binding specificity of sperm for the oocyte is similar to that of ADAM disintegrin for integrin α9β1. We have shown here that the conserved R X6DLPEF motif flanking the RGD motif in the ADAM15 disintegrin domain, but not the RGD motif itself, is important for binding of ADAM15 to α9β1. The ADAM disintegrin domains with this motif, e.g. those of ADAM1, 2, 3, 9, 12, and 15, bind avidly to α9β1, but those lacking this motif, ADAM10 and 17, do not. The R X6DLPEF motif is conserved among all ADAM disintegrin domains except ADAM10 and 17 (Fig. 2 a), suggesting that α9β1 is a receptor for all ADAMs with this recognition motif and that binding of ADAM disintegrin domains to α9β1 through this motif may have critical biological functions. ADAM10 and 17 represent a subfamily within the ADAM family with several distinct structural (and functional) features. The disintegrin domains of ADAM10 and 17 contain only 13 of the 15 cysteine residues characteristic of a typical type III snake venom disintegrin domain, indicating that the disintegrin function of ADAM10 and 17 may be different from those of ADAMs with the conserved R X6DLPEF motif as well as from that of snake venom disintegrins and may not bind to any integrin at all. The RGD motif is not important for α9β1binding to ADAM15. Thus, α9β1 recognizes ADAM15 in a manner distinct from that of αvβ3. Human ADAM15 is the only ADAM with the RGD motif in the disintegrin domain, and the RGD motif is not conserved among ADAM disintegrin domains (even mouse ADAM15 does not have this motif), suggesting that RGD-dependent interaction of ADAM15 with integrins may have only limited importance. This is in stark contrast to many snake venom disintegrins in which the RGD motif is highly conserved. Snake venom disintegrins bind to αIIbβ3 or αvβ3integrins and block thrombosis and hemostasis (for reviews, see Refs. 2Wolfsberg T.G. Straight P.D. Gerena R.L. Huovila A.P. Primakoff P. Myles D.G. White J.M. Dev. Biol. 1995; 169: 378-383Crossref PubMed Scopus (360) Google Scholarand 31McLane M.A. Marcinkiewicz C. Vijay-Kumar S. Wierzbicka-Patynowski I. Niewiarowski S. Proc. Soc. Exp. Biol. Med. 1998; 219: 109-119Crossref PubMed Scopus (187) Google Scholar). The three-dimensional structures of the snake venom disintegrins echistatin (Protein Data Bank code 2ECH) and kistrin (1KST) are available. The “disintegrin loops” of echistatin and kistrin are protruding loops (approximately 15 Å long and 4 Å wide) with the RGD motif at the tip of the loops. The three-dimensional structure of the ADAM disintegrin is not available. We generated a molecular model of the “disintegrin loop region” of the ADAM15 disintegrin domain based on the echistatin structure (2ECH) using the SWISS-MODEL protein modeling server (Fig. 5) assuming that the ADAM disintegrin and snake venom disintegrins are similar in structure. In this model oppositely charged residues (Arg481 and Glu491, Arg484 and Asp488) and hydrophobic residues (Pro482 and Pro490) are close to each other, stabilizing the loop. This model predicts that the Arg481 and the DLPEF motif in the α9β1 binding motif are located on the opposite side of the loop and that Arg481 and Glu491 are close to each other in space, although they are distant from each other in the primary structure. The RGD motif is present at the tip of the loop. It is likely that the entire loop sequence may be required for synthetic peptides of the ADAM disintegrins to be properly folded and effectively bind to α9β1. α9β1 has been shown to mediate cell adhesion and migration but not cell spreading in vitro. The association of an adhesion receptor, such as α9β1, with a metalloprotease, such as an ADAM, would be expected to facilitate cell migration in vivo in that the protease may digest and modify extracellular matrices and other tissue barriers in the immediate area of adhesion. The association between α9β1 and ADAM would further be expected to occur between molecules on the same cell,i.e. in cis, as cell-cell interaction is not normally a feature of cell migration. In contrast, it has been proposed that shedding of growth factors occurs with the growth factor and ADAM expressed on different cells, i.e. in trans. ADAM9 is responsible for processing of heparin-binding EGF (15Izumi Y. Hirata M. Hasuwa H. Iwamoto R. Umata T. Miyado K. Tamai Y. Kurisaki T. Sehara-Fujisawa A. Ohno S. Mekada E. EMBO J. 1998; 17: 7260-7272Crossref PubMed Scopus (475) Google Scholar) and binds with high affinity to integrin α9β1 (the present report). It will be interesting to know whether integrin α9β1 is involved in the processing of heparin-binding EGF. Our identification of mutations that obliterate the binding of α9β1 to ADAMs can now be used to test the role of α9β1-ADAM binding on cell migration as well as on processing of heparin-binding EGF and other potential ADAM substrates. As the prominent sheddases ADAM10 and 17 do not bind to α9β1, it is unlikely that α9β1 is involved in the processing of the proteins they target, including tumor necrosis factor-α and the Alzheimer precursor protein. Few ADAMs other than ADAM10 and 17 have been tested for capacity to process these proteins. It is possible that the specificities of the ADAM domains other than the metalloprotease domains and the cell type-specific expression of ADAM ligands, such as integrin α9β1, will determine which proteins will be digested and when. The digestion may then occur either in cis or in trans depending on the situation. The metalloprotease domains themselves may be rather nonspecific. The relative nonspecificity of metalloproteases agrees with the lack of success in developing specific metalloprotease inhibitors despite a large effort in this area by the pharmaceutical companies. As many ADAMs that have the R X6DLPEF motif are present on sperm (5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar) and ADAMs have been implicated in sperm-egg binding during fertilization, we tested whether α9β1-ADAM interaction may be involved in sperm-egg binding. Two recent studies demonstrate that the presence of acidic residues surrounding the cysteine in the disintegrin loop of ADAM2 (X(C/D)ECD) is essential for the binding of ADAM2 to the egg plasma membrane (24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar, 25Zhu X. Bansal N.P. Evans J.P. J. Biol. Chem. 2000; 275: 7677-7683Abstract Full Text Full Text PDF PubMed Scopus (73) Google Scholar). The underlined residues correspond to the RGD sequence in ADAM15. ADAM3, which also plays a role in sperm-egg binding, has the same characteristics in the sequence of its disintegrin loop (KSDCD). It has also been reported that synthetic peptides of the ADAM2 and 3 disintegrin domain putative integrin-binding site block sperm-egg interaction (5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar, 22Yuan R. Primakoff P. Myles D.G. J. Cell Biol. 1997; 137: 105-112Crossref PubMed Scopus (203) Google Scholar, 24Bigler D. Takahashi Y. Chen M.S. Almeida E.A. Osbourne L. White J.M. J. Biol. Chem. 2000; 275: 11576-11584Abstract Full Text Full Text PDF PubMed Scopus (100) Google Scholar,26Takahashi Y. Bigler D. Ito Y. White J.M. Mol. Biol. Cell. 2001; 12: 809-820Crossref PubMed Scopus (112) Google Scholar). It is intriguing that the R X6DLPEF α9β1 recognition motif is overlapping with these peptide sequences. All studies to date show that the Asp residue is essential. This residue is not conserved in ADAM10 and 17. We have shown here that the disintegrin domains of ADAM2 and 15, which effectively bind to integrin α9β1, bound to murine oocytes and blocked sperm-egg binding, but the ADAM17 disintegrin domain, which does not bind to α9β1, did not. Although ADAM15 has not been shown to be present on sperm, it shares the specificity with the sperm ADAM2; in fact, in our experiments, ADAM15 was more effective in egg binding and inhibition of sperm-egg binding that ADAM2, although these differences may be primarily technical. These binding studies suggest that the receptor for sperm on the oocytes is similar in specificity to α9β1. These results are well explained if eggs express integrin α9β1 on their surface. Although we confirmed the expression of α9β1 in many mouse tissues (32Palmer E.L. Ruegg C. Ferrando R. Pytela R. Sheppard D. J. Cell Biol. 1993; 123: 1289-1297Crossref PubMed Scopus (217) Google Scholar, 33Wang A. Patrone L. McDonald J.A. Sheppard D. Dev. Dyn. 1995; 204: 421-431Crossref PubMed Scopus (43) Google Scholar, 34Taooka Y. Chen J. Yednock T. Sheppard D. J. Cell Biol. 1999; 145: 413-420Crossref PubMed Scopus (239) Google Scholar), our efforts to identify α9 on oocytes, using immunostaining in mouse ovary with anti-α9 cytoplasmic peptide antibody 1057 or in guinea pig ovary with either antibody 1057 or monoclonal antibody Y9A2, were not successful. The lack of detection of α9 in the ovary suggests that either α9 is not present or it is not accessible to antibodies. Another integrin(s) or other receptor with specificity similar to α9β1 may exist in eggs. It has been proposed that α6β1 is the oocyte receptor for sperm (5Almeida E.A. Huovila A.P. Sutherland A.E. Stephens L.E. Calarco P.G. Shaw L.M. Mercurio A.M. Sonnenberg A. Primakoff P. Myles D.G. White J.M. Cell. 1995; 81: 1095-1104Abstract Full Text PDF PubMed Scopus (471) Google Scholar). Recent findings that sperm-egg interaction is normal in α6-null mice (27Miller B.J. Georges-Labouesse E. Primakoff P. Myles D.G. J. Cell Biol. 2000; 149: 1289-1296Crossref PubMed Scopus (164) Google Scholar) and that the antibody to α6, GoH3, does not always affect sperm-egg interaction (28Evans J.P. Kopf G.S. Schultz R.M. Dev. Biol. 1997; 187: 79-93Crossref PubMed Scopus (135) Google Scholar, 35Evans J.P. Front. Biosci. 1999; 4: D114-D131Crossref PubMed Google Scholar) have raised questions about the role of α6β1 in sperm-egg binding and fusion. We thus tested the ability of α6β1 to bind to the ADAM disintegrin domains. We did not detect adhesion of α6-expressing CHO cells to the ADAM disintegrins under the conditions in which α9-CHO cells adhere to the same ligands. However, it cannot be ruled out that α6β1 may interact with the ADAM disintegrin domains under certain conditions in oocytes. Since we used CHO cells expressing human α9 or α6/hamster β1 hybrid in this study, it also cannot be ruled out that human and hamster β1 provide different ligand specificities. The biological significance of the interaction between ADAMs and α9β1 integrin is not known. α9β1 is present in many developing and mature tissues including airway epithelia, the basal layer of squamous epithelia, smooth muscle, skeletal muscle, hepatocytes, neutrophils, and monocytes (32Palmer E.L. Ruegg C. Ferrando R. Pytela R. Sheppard D. J. Cell Biol. 1993; 123: 1289-1297Crossref PubMed Scopus (217) Google Scholar, 33Wang A. Patrone L. McDonald J.A. Sheppard D. Dev. Dyn. 1995; 204: 421-431Crossref PubMed Scopus (43) Google Scholar, 34Taooka Y. Chen J. Yednock T. Sheppard D. J. Cell Biol. 1999; 145: 413-420Crossref PubMed Scopus (239) Google Scholar). Considering the wide distribution of both ADAMs and α9β1, ADAM-α9β1 interactions may play a role in many physiological and pathological situations including muscle development and regeneration; vascular, cartilage, and other tissue remodeling; and perhaps fertilization. It remains a challenge to determine what this role may be.
Referência(s)