sign in sign up
<<
Home , Collagen

Proc. Nati. Acad. Sci. USA Vol. 86, pp. 9906-9910, December 1989 Cell Biology The human VLA-2 is a receptor on some cells and a collagen/ receptor on others MARIANO J. ELICES AND MARTIN E. HEMLER* Dana-Farber Cancer Institute, Harvard Medical School, Boston, MA 02115 Communicated by Elizabeth D. Hay, September 5, 1989

ABSTRACT The integrin heterodimer VLA-2, previously nor VLA-2 appears to interact with lami- known as a collagen receptor, is now shown also to be a laminin nin. receptor. Adhesion of the human melanoma cell line LOX to laminin was inhibited by anti-VLA a2 antibodies. Because VLA-2-mediated LOX cell attachment to laminin was not METHODS inhibited by digestion with , collagen contamination Antibodies and Cells. The monoclonal antibodies (mAbs) oflaminin was not a factor. In addition, VLA-2 from LOX cells A-lA5 (anti-f81), TS2/7 (anti-a'), 12F1 (anti-a2), J143 (anti- bound to immobilized laminin, and binding was disrupted by a3), and the control mAbs P3 and J-2A2 were obtained as EDTA but not by Arg-Gly-Asp (RGD) . VLA-3 also described (22). The mAbs PlH5 and PlB5 that recognize the bound to lammnin-Sepharose, although less avidly than VLA-2. VLA a2 and a3 subunits, respectively (19, 20), were from E. Thus, at least four separate members of the integrin .,8 Wayner (University of Washington, Seattle). Anti-VLA-2 subfamily serve as laminin receptors-i.e., VLA-2 and VLA-3 mAbs 5E8 (23) and Gil4 (24) were from R. Bankert (Roswell (this study) together with VLA-1 and VLA-6 (other reports). Park Memorial Institute, Buffalo, NY) and S. Santoso (Gies- Whereas LOX and other cell lines used VLA-2 as both a sen, F.R.G.), respectively. The mAbs BlE5 and BlEll laminin and collagen receptor, fibroblast VLA-2 mediated recognize epitopes on VLA aS and f1, respectively, and were collagen but not laminin binding. Likewise, VLA-2 from provided by C. Damsky (University of California, San Fran- did not interact with laminin. Despite this functional cisco). GoH3, a mAb specific for VLA a6 (25), was obtained discordancy, VLA-2 from laminin-binding and nonbinding from A. Sonnenberg (The Netherlands Cancer Institute, sources was indistinguishable by all immunochemical and Amsterdam). Finally, a rabbit heteroantiserum to a synthetic biochemical criteria examined. Thus, functional differences in from the COOH-terminal intracytoplasmic portion of VLA-2 may be due to cell type-specific modulation. VLA a2 (26) was prepared in this laboratory. The human melanoma cell line LOX was obtained from Laminin, a major constituent of basement membranes, pro- Lan Bo Chen (Dana-Farber Cancer Institute). All the cell motes adhesion, growth, migration, and differentiation of lines described in Table 1 were grown in RPMI 1640 medium cells (1). The laminin (Mr 900,000) consists of containing 10o fetal bovine serum. three subunits arranged in the shape of a cross with one long . Laminin from the Engelbreth-Holm- and three short arms (1). Distinct functions of laminin, such Swarm (EHS) murine tumor (27) was a gift from H. Kleinman as promotion ofneurite outgrowth, interaction with basement (National Institute of Dental Research, Bethesda, MD). membrane components, and attachment and spreading of was purified from human plasma (28), and col- cells, each reside within separate structural domains (1). In lagens type I and IV were purchased from Telios Pharma- particular, cell adhesion appears to be mediated by at least ceuticals (La Jolla, CA). three different regions in laminin (2-5), and a peptide derived Cell Attachment and mAb Inhibition Assays. To assess cell from one of the cell-binding sites has been shown to inhibit attachment to proteins, cells were incu- bated with 51Cr (0.5 ,Ci; 1 Ci = 37 GBq) overnight, washed experimental metastasis (6). three times, and resuspended in RPMI 1640 supplemented Thus far, specific cell-adhesion receptors for laminin in- with 0.5% fetal bovine serum and 0.1 mM MnC12. Then 1.5 clude a 68-kDa (7-9), as well as other surface mole- x 103 cells per well were plated in triplicate for 20-30 min at cules that belong to the integrin family. are a group 37°C on 96-well microtiter dishes (0.8-cm diameter per well) of af3 heterodimers involved in cell-cell and cell-extracellu- that were previously incubated with laminin, collagen, or lar matrix interactions, some of which occur via recognition fibronectin (1 ,g/well). After unbound cells were aspirated, ofan Arg-Gly-Asp (RGD) tripeptide sequence present in their and the plates were washed twice with RPMI 1640, 51Cr ligands (10, 11). Members ofthe integrin family implicated as present in 0.1% SDS cell lysates was measured by using a y laminin receptors include a cell complex that resembles counter. For mAb inhibition experiments, labeled cells were VLA-3 (12), human VLA-6 from platelets (13), and a rat cell incubated with mAbs for 30 min at 4°C, washed, and plated complex that may be the homolog of human VLA-1 (14). on matrix ligand as described above. Specific attachment was VLA-2, another member of the integrin ,1 subfamily, has calculated by subtracting the radioactivity bound to bovine been shown to act as a cell-surface receptor for collagen in serum albumin-coated controls. platelets (15-18), (19, 20), and melanoma cells Laminin-Sepharose Affinity Chromatography. Human (21). However, in the platelet system VLA-2 was clearly LOX melanoma cells were detached from culture unable to mediate cell binding to laminin (13, 18). In the flasks with 0.03% EDTA in phosphate-buffered saline (PBS), present report, a role for VLA-2 as a laminin receptor in washed twice with PBS, and iodinated with 2 ,uCi of Na1251 human LOX melanoma cells, as well as other cell lines, is (DuPont/NEN) and lactoperoxidase at 0.2 mg/ml (Sigma) for unequivocally demonstrated, despite finding that neither 2-5 x 107 cells. lodinated cells were lysed with 0.1 M octyl

The publication costs of this article were defrayed in part by page charge Abbreviations: mAb, monoclonal antibody; PBS, phosphate- payment. This article must therefore be hereby marked "advertisement" buffered saline. in accordance with 18 U.S.C. §1734 solely to indicate this fact. *To whom reprint requests should be addressed. 9906 Downloaded by guest on September 24, 2021 Cell Biology: Elices and Hernler Proc. Natl. Acad. Sci. USA 86 (1989) 9907 glucoside (Sigma) and 0.1 M octyl thioglucoside (Calbio- laminin-coated surfaces, with half-maximal binding at chem) in PBS containing 0.1 mM MnCI2, at 10 0.5 ,ug oflaminin per well (data not shown). To investigate the ,ug/ml, 10 juM , and 1 mM phenylmethylsulfonyl contribution of VLA heterodimers to melanoma-cell adhe- fluoride for 4 hr at 40C. Detergent lysates were clarified by sion on laminin, a panel of mAbs was tested for inhibition of centrifugation at 12,000 x g, supplemented with bovine cell binding (Fig. 1A). Cell attachment to laminin could be serum albumin (0.5 mg/ml), and loaded onto 2 ml oflaminin- completely abolished by incubation with the mAb BlEll Sepharose columns (3.5 mg of laminin per ml of packed (anti-VLA p1), thus indicating that f1 integrins had a major beads) preequilibrated with 25 mM each of octyl glucoside role in LOX cell adhesion to laminin. Ofthe mAbs against the and octyl thioglucoside in the same buffer as above (buffer VLA proteins expressed on the surface ofLOX cells (VLA-1, A). After equilibration for 2 hr, columns were washed with -2, -3, -5, and -6) only anti-VLA a2 mAb 5E8 (Fig. 1A) and buffer A until eluted radioactivity decreased to background P1H5 (data not shown) substantially inhibited melanoma-cell levels (-20 column volumes). Stepwise elution was con- adhesion to laminin. In contrast, neither mAbs PlB5 (anti- ducted with 0.5 M NaCi, 10 mM EDTA, and 4 M urea, each VLA a 3) and BlE5 (anti-VLA a5) (Fig. 1A) nor TS2/7 in buffer A. Fractions (0.8 ml each) were collected and (anti-VLA a') and GoH3 (anti-VLA a 6) were inhibitory (data analyzed for radioactivity with a y counter. Immunoprecip- not shown). The prototype anti-a2 mAb 12F1 (31) did not itation of labeled proteins was done as described (22). block cell binding to either laminin or collagen type I, NH2-Terminal Sequencing of a2 Subunit. VLA-2 was pu- although material immunoreactive with either 5E8 or 12F1 rified from LOX (1 x 109 cells) by using lectin affinity was reciprocally precleared by incubation with the other chromatography, followed by Gil4-Sepharose chromatogra- mAb (Fig. 1B). Thus, 12F1 and 5E8 both bound to VLA-2 but phy with described methods (29). The mAb Gil4 bound might differ in epitope recognition. Upon titration with in- quantitatively to all VLA-2, whether from laminin binding or creasing concentrations of mAb 5E8 (Fig. 1C), a maximal nonbinding cells. After purification, VLA-2 subunits were inhibition of 70% was obtained when laminin was saturating separated by SDS/PAGE and then transferred to polyvinyl- for cell adhesion. At lower levels oflaminin, LOX cells bound idene difluoride membrane (Immobilon, Millipore), as de- to laminin-coated plates with less avidity, and thus inhibition scribed (30). The a2 polypeptide chain was identified by by mAb 5E8 approached 100% (data not shown). with Coomassie blue, cut from the membrane, and Comparison of Matrix Adhesion for LOX and Other Cells. the NH2-terminal sequence was determined at the To determine whether various human cell lines may use Harvard Microsequencing Facility, Cambridge, MA. VLA-2 to adhere to laminin, their binding to laminin-coated plates was analyzed. Table 1 shows that the interaction ofthe adherent cell lines EJ, LOX, and SK-N-SH with laminin was RESULTS blocked by mAb 5E8, whereas cell lines of hematopoietic Inhibition of LOX Cell Attachment to Laminin. The LOX origin such as HL-60, HPB-ALL, and K-562, not expressing human melanoma cell line was found to attach efficiently to VLA-2, did not bind to laminin. RD cells bound to laminin

60

'I0 , 40 B ._; Preclear: 5E8 m None 12 F1 zun L

__ C4JN Wi (%4 Wj CM

mAb C 100 OA O'

._ 50 a1)

0 0.1 1.0 10.0 100.0 [mAb], jg/ml a b c d e f

FIG. 1. Inhibition of melanoma-cell adhesion to laminin by anti-VLA-2 antibodies. (A) Inhibition of LOX-cell binding to laminin was done by using the following mAbs, each at -10 ,ug/ml: J-2A2 (control), 5E8 (anti-a2), PlB5 (anti-a3), BlE5 (anti-a5), and BlEll (anti-PD). (B) Immunoprecipitation of VLA-2 by mAb 5E8. Radiolabeled material from LOX-cell extracts either was not (lanes a and b) or was precleared with mAbs 12F1 (lanes c and d) and 5E8 (lanes e and f), respectively, before reprecipitating with either mAb 12F1 (lanes a, c, and e) or 5E8 (lanes b, d, and f). (C) Titration of LOX-cell attachment to laminin with increasing concentrations of mAbs J-2A2 (o), 5E8 (0), and BlEll (A). Cell adhesion assays were conducted as described. Downloaded by guest on September 24, 2021 9908 Cell Biology: Elices and Hemler Proc. Natl. Acad. Sci. USA 86 (1989) Table 1. Binding of cell lines to extracellular matrix proteins and inhibition with mAbs Collagen type I Fibronectin Laminin Inhibited Inhibited Inhibited Cell line Binding, % by 5E8, % Binding, % by BlE5, % Binding, % by 5E8, % MRC-5 45 65 20 80 <1 EJ 88 74 79 82 69 67 LOX 87 66 76 81 65 70 RD 13 0 82 86 38 0 SK-N-SH 80 61 55 84 72 73 HL-60 <1 16 96 <1 HPB-ALL <1 - 10 89 <1 K-562 <1 - 12 98 <1 MRC-5 (normal fibroblast), EJ (bladder carcinoma), LOX (melanoma), SK-N-SH (neuroblastoma), HL-60 (promyelo- cytic leukemia), HPB-ALL (T lymphoblastoma), and K-562 (erythroleukemia) cells were all tested for attachment to ligand-coated plates as described. The mAbs used for inhibition of cell attachment were J-2A2 (control), 5E8 (anti-a 2), PlB5 (anti-a3), and BlE5 (anti-a5). even though they did not express marked levels of VLA-2, loss of LOX-cell adhesion. In similar fashion, when LOX and as expected this binding was not inhibited by anti-VLA-2 cells in suspension were incubated with collagenase, cell mAb. For the MRC-5 fibroblast line, attachment to collagen binding to laminin did not diminish in subsequent attachment type I was mediated by VLA-2 because mAb 5E8 blocked assays. Thus, laminin binding was not an indirect phenom- binding; yet these cells did not adhere to laminin. This enon mediated by a collagen bridge. observation suggested that VLA-2 may function as a receptor Direct Demonstration of VLA Proteins Binding to Laminin. exclusively for collagen in some cell types but may see both To directly confirm that VLA-2 interacts with laminin, as collagen and laminin in others. For all cell types analyzed, suggested by the inhibition results obtained above, affinity- binding to fibronectin was inhibited (80-98%) by the anti- chromatography separations were conducted. Fig. 3 shows a VLA-5 mAb BlE5, regardless of whether laminin and/or representative elution profile of an octyl glucoside LOX-cell collagen binding was observed. Also, attachment of all cell lysate fractionated on a laminin-Sepharose column in the lines listed in Table 1 to collagen, laminin, or fibronectin was presence of MnCl2 (33, 34). Although most label passed blocked by the anti-,31 mAb BlEll (data not shown). unretarded through the column, sequential elution with buff- Laminin Binding Not Influenced by Collagen Contamina- ers containing 0.5 M NaCl, 10 mM EDTA, and 4 M urea, tion. When iodinated laminin was subjected to SDS/PAGE respectively, resulted in selective enrichment for discrete and autoradiography, the corresponding autoradiogram (Fig. polypeptide species. Notably, elution with EDTA yielded 2A) showed laminin subunits A, B1, and B2, as well as 10% three distinct bands of40, 120, and 150 kDa, the last two with of entactin/ (32). However, the possibility remained mobilities resembling VLA / and a subunits. Immunoprecip- that laminin contamination by collagen type IV could infu- itation of column eluates (Fig. 4) showed that VLA-2 was the ence the results described above. Thus, the effect ofbacterial major component in the EDTA eluate, together with a lesser collagenase upon LOX binding to laminin and collagen type amount of VLA-3, whereas the NaCI fraction contained IV was determined. Fig. 2B shows that melanoma-cell at- mostly VLA-3. When the same laminin affinity-column ex- tachment to collagen type IV was abrogated after digestion periments were done using platelets, no detectable VLA with collagenase, whereas treatment oflaminin resulted in no proteins were obtained (data not shown). In addition, neither A B

C.,

o c 0 'E x 4 E -j 0. 6 0 -* 0 2 -aCu

- BpB2 oL ff Ii,1 I I I III 0.01 0.1 1.0 40. Ligand concentration, - En/Nd jug/well

FMG. 2. Properties of laminin used for cell-attachment studies. (A) SDS/PAGE analysis of iodinated laminin used in this study. En/Nd, entactin/nidogen. (B) Effect of collagenase treatment of laminin on LOX-cell attachment. Binding of LOX cells to laminin (o, *) and collagen type IV (o, *) was analyzed with (open symbols) or without (closed symbols) prior digestion with bacterial collagenase (type VII, Sigma) at 50 milliunits/ml in 50 mM Tris, pH 7.5/2.5 mM CaC12/0.02% NaN3 for 16 hr at 15°C. Incubations were terminated by adding EDTA, and then both treated and untreated samples were used to coat microtiter plates. LOX-cell attachment was done as described. Downloaded by guest on September 24, 2021 Cell Biology: Elices and Hemler Proc. Natl. Acad. Sci. USA 86 (1989) 9909

50C )-t ~~~~A~ ~~~~~B c

II.._- I _ I | _ I r~~~~~MW 30C

X00 - -93G .ss 0 cho A. FIG. 3. Affinity chromatography of 0 A41E LOX-cell extracts on laminin-Sepharose. x A radiolabeled extract from human LOX E melanoma cells was passed over laminin- 0Q C Sepharose as described. Fractions (0.8 ml each) were collected and analyzed for radioactivity by using a y counter. All four peak fractions and a sample of the K- a0 column wash (arrows) were analyzed by 1S* \ SDS/PAGE under reducing conditions, and labeled bands were visualized by F- \ autoradiography^ (between brackets). Peaks A, B, and C resulted from stepwise I elution with 0.5 M NaCl, then 10 mM ______EDTA, and finally 4 M urea, respec- 0 10 20 30 40 P 50 60 tively.sidase; Myo,BSA,myoglobin;bovine serum83G, albumin;galacto- Fraction Ova, ovalbumin. VLA-2 nor VLA-3 could be eluted from laminin-Sepharose through 15 positions was Tyr-Asn-Val-Gly-Leu-Pro-Glu-Ala- columns using RGD peptides (data not shown). Lys-Ile-Phe-Ser-Gly-Pro-Ser. This sequence was 100% iden- Structural Similarity in VLA-2 from Funtionally Distinct tical to that reported for platelet or a2 subunit at the Sources. Because VLA-2 from LOX cells did bind to laminin NH2-terminal 15 positions (26). In addition, this NH2- and VLA-2 from fibroblasts and platelets did not, it was terminal sequence has been consistently obtained by using important to determine whether there were any obvious three different anti-a2 antibodies for purification (data not structural differences in VLA-2 isolated from various cell shown), and also it exactly matched the amino acid sequence sources. Thus, VLA-2 preparations from LOX cells, plate- derived from a 2 cDNA from fibroblast and endothelial cell lets, and/or fibroblasts were compared. NH2-terminal amino libraries (26). acid sequencing was carried out for purified a subunit from In immunoprecipitation and/or flow cytometry experi- LOX VLA-2 (see Methods), and the unequivocal result ments, VLA-2 from LOX cells, platelets, and fibroblasts was recognized by three distinct mAbs (5E8, P1H5, Gil4) and by L a m i n n rabbit anti-intact 2 antiserum (data not shown). In addition, VLA-2 protein immunoprecipitates from each cell source NaCI E DTA were identical in size (both reduced and nonreduced) with no F owthrough e u t o n e u t i o n apparent differences in posttranslational modification. Fi- I ~~~~~~~~~~~~II nally, a2 mRNA from fibroblasts, LOX cells, and other cells N ~- was same i-- l1 the size (-8 kb) when probed with full-length a2 44 cDNA probe (data not shown). -J J -j j (-)> ,. (-)-:I >'>i :>,. >,. (-)> > > > DISCUSSION VLA-2 as a Laminin Receptor. The present study expands the range of VLA-2 activities to include a laminin receptor function because (i) anti-VLA-2 mAbs inhibit cell attachment to laminin and (ii) VLA-2 binds directly to laminin-Sepharose 0 *: columns. The laminin receptor described herein is clearly distinct from other laminin-binding integrins, such as VLA-6 (13) and the VLA-3-like integrin from Rugli glioblastoma (12). Whereas the a subunits of those integrins drop upon reduc- tion due to of a disulfide-linked C-terminal peptide, VLA a2 subunit is not cleaved. In the cases of the laminin- binding integrins from the chicken CSAT complex (35) and the rat PC-12 pheochromocytoma cell line (36, 37), multiple integrin complexes appear to be involved, and their relation- ship, if any, to human VLA-2 remains unclear. The VLA- a c d b e f g h i i k I m n o 1-like laminin receptor from rat PC-12 cells (14) differs from VLA-2 in that it has a a FIG. 4. Immunoprecipitation of VLA heterodimers from laminin larger subunit. and collagen type IV columns. The material that passed unretarded VLA-3 as a Laminin Receptor. Although no inhibition of through the laminin-Sepharose column, as well as the corresponding LOX-cell binding to laminin was seen with the anti-VLA-3 NaCl and EDTA eluates, were immunoprecipitated with the follow- mAb P1B5 (Fig. 1A), VLA-3 bound and was subsequently ing mAbs: control J-2A2 (-), 5E8 (anti-VLA a2), J143 (anti-VLA a3), eluted from laminin affinity columns (Fig. 3). Previously, B1E5 (anti-VLA a5), and A-lA5 [anti-VLA ,81, total VLA proteins mAbs to VLA-3 have been suggested to block cell adhesion (VLA-T)]. to laminin (19, 20). Therefore, it is possible that anti-VLA-3 Downloaded by guest on September 24, 2021 9910 Cell Biology: Elices and Hemler Proc. Natl. Acad. Sci. USA 86 (1989) blocking of LOX-cell binding to laminin was not seen in this 1. Martin, G. R. & Timpl, R. (1987) Annu. Rev. Cell Biol. 3, 57-85. study because either (i) the mAb recognized an inappropriate 2. Graf, J., Iwamoto, Y., Sasaki, M., Martin, G. R., Kleinman, H. K., Robey, F. A. & Yamada, Y. (1987) Cell 48, 989-996. a3 epitope or (ii) the role of VLA-3 may be overshadowed by 3. Aumailley, M., Nurcombe, V., Edgar, D., Paulsson, M. & Timpl, the contribution of VLA-2. In support of this, elution of R. (1987) J. Biol. Chem. 262, 11532-11538. VLA-3, but not VLA-2, from laminin columns was accom- 4. Goodman, S. L., Deutzmann, R. & Von der Mark, K. (1987) J. Cell plished with 0.5 M NaCl, suggesting (but not proving) that Biol. 105, 589-598. 5. Charonis, A. S., Skubitz, A. P. N., Koliakos, G. G., Reger, L. A., VLA-3 may display less avidity than VLA-2 toward laminin. Dege, J., Vogel, A. M., Wohlhueter, R. & Furcht, L. T. (1988) J. Functional Diversity for VLA-2. For LOX and other cell Cell Biol. 107, 1253-1260. lines, adhesion to laminin appeared chiefly mediated by 6. Iwamoto, Y., Robey, F. A., Graf, J., Sasaki, M., Kleinman, H. K., VLA-2. However, VLA-2 was clearly not a laminin receptor Yamada, Y. & Martin, G. R. (1987) Science 238, 1132-1134. 7. Terranova, V. P., Rao, C. N., Kalebic, T., Margulies, I. M. & in all cell types in which it was present. For example, Liotta, L. A. (1983) Proc. Natl. Acad. Sci. USA 80, 444-448. antibody-blocking studies showed that fibroblasts used VLA- 8. Malinoff, H. L. & Wicha, M. S. (1983) J. Cell Biol. 96, 1475-1479. 2 to bind collagen but not laminin. Also, VLA-2 from platelets 9. Lesot, H., Kuhl, U. & Von der Mark, K. (1983) EMBO J. 2, did not appear to be a laminin receptor because it did not 861-865. interact with laminin-Sepharose columns. This observation 10. Ruoslahti, E. & Pierschbacher, M. D. (1987) Science 238, 491-497. 11. Hynes, R. 0. (1987) Cell 48, 549-554. was consistent with previous results by other investigators 12. Gehlsen, K. R., Dillner, L., Engvall, E. & Ruoslahti, E. (1988) (13, 18). In one study, even though VLA-2 was present and Science 24, 1228-1229. fully functional as a collagen receptor, antibodies to VLA-2 13. Sonnenberg, A., Modderman, P. W. & Hogervorst, F. (1988) Na- did not block platelet attachment to laminin (13). In another ture (London) 360, 487-489. bound to 14. Ignatius, M. J. & Reichardt, L. F. (1988) Neuron 1, 713-725. report, platelet VLA-2 incorporated into liposomes 15. Nieuwenhuis, H. K., Akkerman, J. W. N., Houdijk, W. P. M. & collagen but not to laminin (18). Sixma, J. J. (1985) (London) 318, 470-472. Despite functional discordancy among various VLA- 16. Santoro, S. A. (1986) Cell 46, 913-920. 2-expressing cell lines, VLA-2 samples from different 17. Santoro, S. A., Rajpara, S. M., Staatz, W. D. & Woods, V. L. sources did not exhibit any obvious physical or immuno- (1988) Biochem. Biophys. Res. Commun. 153, 217-223. a2 se- 18. Staatz, W., Rajpara, S. M., Wayner, E. A., Carter, W. G. & chemical differences. In particular, NH2-terminal Santoro, S. S. (1989) J. Cell Biol. 108, 1917-1924. quences were identical regardless of source, whereas for the 19. Wayner, E. A. & Carter, W. G. (1987) J. Cell Biol. 105, 1873-1884. 10 other integrin a subunits, the NH2-terminal sequences of 20. Takada, Y., Wayner, E. A., Carter, W. G. & Hemler, M. E. (1988) which are known, there are 8-14 differences between those J. Cell. Biochem. 37, 385-393. and the a2 subunit in the first 15 positions (38). Also, all a2 21. Kramer, R. H. & Marks, N. (1989) J. Biol. Chem. 264, 4684-4688. 22. Hemler, M. E., Huang, C. & Schwarz, L. (1987) J. Biol. Chem. 262, subunits tested were similarly reactive with anti-a2 COOH- 3300-3309. terminal peptide sera, whereas the COOH termini currently 23. Zylstra, S., Chen, F. A., Ghosh, S. K., Repasky, E. A., Rao, U., known for eight different integrin a subunits (39) plus a3 Takita, H. & Bankert, R. B. (1986) Cancer Res. 46, 6446-6451. subunit (Y. Takada, M. H., unpublished work) display little 24. Kiefel, V., Santoso, S., Katzmann, B. & Mueller-Eckhardt, C. or no . (1989) Blood 73, 2219-2223. 25. Sonnenberg, A., Janssen, H., Hogervorst, F., Calafat, J. & Hilgers, Because there was no obvious VLA-2 structural diversity, J. (1987) J. Biol. Chem. 262, 10376-10383. other factors such as cellular signaling or activation events 26. Takada, Y. & Hemler, M. E. (1989) J. Cell Biol. 109, 397-407. may contribute to modify VLA-2 function in a cell type- 27. Timpl, R., Rohde, H., Robey, P. G., Rennard, S. I., Foidart, J. M. specific manner. For example, lIb/IIIa, another integrin & Martin, G. R. (1979) J. Biol. Chem. 254, 9933-9937. heterodimer, was expressed on unactivated platelets but was 28. Ruoslahti, E., Hayman, E. G., Pierschbacher, M. & Engvall, E. (1982) Methods Enzymol. 82, 803-831. dependent on platelet activation for function (40). Similarly, 29. Takada, Y., Strominger, J. L. & Hemler, M. E. (1987) Proc. Natl. triggering of T cells through the T-cell receptor or with Acad. Sci. USA 84, 3239-3243. phorbol ester activated the integrin LFA-1 without changing 30. Matsudaira, P. (1987) J. Biol. Chem. 262, 10035-10038. its surface expression levels (41). Additional factors that 31. Pischel, K. D., Hemler, M. E., Huang, C., Bluestein, H. G. & merit consideration as possible means of modulating VLA-2 Woods, V. L. (1987) J. Immunol. 138, 226-233. 32. Timpl, R., Paulsson, M., Dziadek, M. & Fujiwara, S. (1987) activity include variations in divalent cations (33, 34), asso- Methods Enzymol. 145, 363-391. ciation with glycolipids (42), and changes in phosphorylation 33. Gailit, J. & Ruoslahti, E. (1988) J. Biol. Chem. 263, 12927-12932. (43). However, regardless of the mechanistic details, this 34. Hautanen, A., Gailit, J., Mann, D. M. & Ruoslahti, E. (1989) J. Biol. report suggests that the dual laminin/collagen surface recep- Chem. 264, 1437-1442. tor functions of VLA-2 can be dissociated, and cells may use 35. Horwitz, A., Duggan, K., Greggs, R., Decker, C. & Buck, C. (1985) J. Cell Biol. 101, 2134-2144. this property to diversify their capability. 36. Tomaselli, K. J., Damsky, C. H. & Reichardt, L. F. (1987) J. Cell The presence of multiple receptors for laminin within the Biol. 105, 2347-2358. integrin family (VLA-1, VLA-2, VLA-3, and VLA-6) resem- 37. Tomaselli, K. J., Damsky, C. H. & Reichardt, L. F. (1988) J. Cell bles the situation encountered for another extracellular ma- Biol. 107, 1241-1252. trix component, fibronectin, which is recognized by at least 38. Hemler, M. E., Crouse, C., Takada, Y. & Sonnenberg, A. (1988) J. In will be Biol. Chem. 263, 7660-7665. five different integrins (44). future studies, it 39. Takada, Y., Elices, M. J., Crouse, C. & Hemler, M. E. (1989) important to correlate the multiple cell-binding regions in EMBO J. 8, 1361-1368. laminin (2-5) with its specific integrin receptors. 40. Phillips, D. R., Charo, I. F., Parise, L. V. & Fitzgerald, L. A. (1988) Blood 71, 831-843. Note Added in Proof. Results similar to those reported in this paper 41. Dustin, M. L. & Springer, T. A. (1989) Nature (London) 341, have recently been obtained by Languino et al. (45). 619-624. 42. Cheresh, D. A., Pierschbacher, M. D., Herzig, M. A. & Mujoo, K. The authors gratefully acknowledge Drs. R. Bankert, C. Damsky, (1986) J. Cell Biol. 102, 688-696. R. Kantor, H. Kleinman, L. Old, S. Santoso, A. Sonnenberg, E. 43. Horwitz, A., Duggan, K., Buck, C., Beckerle, M. C. & Burridge, Wayner, and V. Woods for providing reagents used in this study. K. (1987) Nature (London) 320, 531-533. M.J.E. is a fellow ofthe Cancer Research Institute, Inc., New York. 44. Hemler, M. E. (1990) Annu. Rev. Immunol., in press. This work was supported by a grant from the National Institutes of 45. Languino, L. R., Gehlsen, K. R., Carter, W. G., Engvall, E. & Health (GM38903). Ruoslahti, E. (1989) J. Cell Biol., in press. Downloaded by guest on September 24, 2021