Identification of cell-binding sites on the Laminin alpha 5 N-terminal domain by site-directed mutagenesis.

The newly discovered laminin alpha(5) chain is a multidomain, extracellular matrix protein implicated in various biological functions such as the development of blood vessels and nerves. The N-terminal globular domain of the laminin alpha chains has an important role for biological activities through interactions with cell surface receptors. In this study, we identified residues that are critical for cell binding within the laminin alpha(5) N-terminal globular domain VI (approximately 270 residues) using site-directed mutagenesis and synthetic peptides. A recombinant protein of domain VI and the first four epidermal growth factor-like repeats of domain V, generated in a mammalian expression system, was highly active for HT-1080 cell binding, while a recombinant protein consisting of only the epidermal growth factor-like repeats showed no cell binding. By competition analysis with synthetic peptides for cell binding, we identified two sequences: S2, (123)GQVFHVAYVLIKF(135) and S6, (225)RDFTKATNIRLRFLR(239), within domain VI that inhibited cell binding to domain VI. Alanine substitution mutagenesis indicated that four residues (Tyr(130), Arg(225), Lys(229), and Arg(239)) within these two sequences are crucial for cell binding. Real-time heparin-binding kinetics of the domain VI mutants analyzed by surface plasmon resonance indicated that Arg(239) of S6 was critical for both heparin and cell binding. In addition, cell binding to domain VI was inhibited by heparin/heparan sulfate, which suggests an overlap of cell and heparin-binding sites. Furthermore, inhibition studies using integrin subunit monoclonal antibodies showed that integrin alpha(3)beta(1) was a major receptor for domain VI binding. Our results provide evidence that two sites spaced about 90 residues apart within the laminin alpha(5) chain N-terminal globular domain VI are critical for cell surface receptor binding.

Laminins are heterotrimeric basement membrane proteins that exert multiple biological functions through interactions with other matrix molecules and cell surface receptors. The regulation of these interactions is critical to many biological processes, including cell adhesion, migration, angiogenesis, tumor progression, and neurite outgrowth (for review, see Ref. 1). The laminin family contains at least 11 chains (five ␣, three ␤, and three ␥ chains). The newly discovered ␣ 5 chain is the most widely expressed member of the laminin ␣ chain family (2).
The N-terminal globular domain (domain VI, ϳ250 -270 residues) is specific for laminin and is the most conserved (ϳ60% sequence identity) among the laminin domains (12). Studies with function-blocking antibodies and cell binding assays have indicated that domain VI of both the laminin ␣ 1 and ␣ 2 chains contains binding sites for the ␣ 1 ␤ 1 and ␣ 2 ␤ 1 integrins (13)(14)(15). Furthermore, domain VI is also capable of binding heparin and heparan sulfate chains of perlecan (13)(14)(15). In addition to binding functions for cell surface receptor and for matrix proteins, domain VI is also essential for the self-assembly of laminins (13,16,18). Previous studies showed that the synthetic peptide RQVFQVAYIIIKA (A-13), derived from domain VI of the laminin ␣ 1 chain, binds ␤ 1 subunit-containing integrins (19). The active core sequence (VAYI) of this peptide is conserved in the ␣ 5 chain, but the functional importance of this site within domain VI has not yet been examined.
In the present work, we studied cell binding functions of mouse laminin ␣ 5 domain VI using site-directed mutagenesis and synthetic peptides. We found that two sites, spaced by ϳ90 amino acids, are required for cell binding. We also identified four residues within the two sites that are essential for the binding. In addition, we demonstrated that an arginine residue of one of the sites is critical for both heparin and cell binding. Our findings suggest that the protein conformation surrounding these sites is important for cell binding through integrin and heparan sulfate-containing cell surface receptors.

Construction of Expression Vectors and Site-directed Mutagenesis-
Mouse kidney cDNA was used as a template in polymerase chain reaction to amplify sequences encoding domain VI and the first four EGF 1 -like repeats of domain V of the laminin ␣ 1 and ␣ 5 chain. Polymerase chain reaction was performed with PfuTurbo DNA polymerase (Stratagene, La Jolla, CA) using the following primers: 1, GAGAGAA-AGCTTCGCACTCCCGGGGGCGATGGC; 2, GAGAGACTCGAGAGT-* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  GGCAGCTAGGCCCGTGGAATC; 3, GAGAGAAAGCTTCAGCAGAGA-GGCTTGTTCCCTGC; 4, GAGAGACTCGAGAGGAGCAGCCCTCGGG-GTTTCG. Primers 1 and 2 were used for laminin ␣ 5 domain VI (Arg 1 -His 514 ), and primers 3 and 4 for the laminin ␣ 1 domain VI (Gln 1 -Ser 484 ). For polymerase chain reaction amplification of the EGF-like repeats of laminin ␣ 5 domain V (Arg 273 -His 514 ), primers 2 and 5, GAGAGAAAG-CTTCGCTGTGTCTGTCATGGCCACG, were used. In addition to the coding sequences, these primers contained either a HindIII or a XhoI restriction site. The polymerase chain reaction products were digested with HindIII and XhoI and ligated into the expression vector pSecTag2/ Hygro B (Invitrogen, Carlsbad, CA). The resulting expression vectors encode the Ig chain leader sequence, a laminin domain, and a c-myc epitope followed by a hexahistidine affinity tag sequence. Site-directed mutagenesis was performed using the Quickchange kit method (Stratagene, La Jolla, CA) and the N-terminal globular domain VI and the first four EGF-like repeats of domain V of the mouse laminin ␣ 5 chain, which has been cloned into pBluescript II SK(ϩ) (Stratagene). The following laminin ␣ 5 residues were individually substituted with alanine: Tyr 130 , Arg 225 , Lys 229 , Arg 234 , Arg 236 , and Arg 239 . All expression constructions and mutations were verified by DNA sequencing.
Expression and Purification of Recombinant Proteins-The expression vectors were transiently transfected into monkey kidney COS-7 cells (CRL-1651, ATCC) using FuGENE 6 (Roche Molecular Biochemicals, Indianapolis, IN). The cells were maintained at 37°C in a humid atmosphere with 5% CO 2 in Dulbecco's modified Eagle's medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. Generation of secreted recombinant His-tagged laminin domains was confirmed by Western blotting of serum-free conditioned medium using an anti-His(C-terminal)-horseradish peroxidase monoclonal antibody (Invitrogen). For recombinant production, conditioned medium (Dulbecco's modified Eagle's medium containing 1% fetal calf serum) was collected 7 days after transfection. The medium was cleared of detached cells and debris through centrifugation, and 1 M Tris-HCl, pH 8.0, and 5 M NaCl were added to final concentrations of 50 mM and 0.5 M, respectively. In addition, 0.2 M phenylmethanesulfonyl fluoride and 3 M imidazole, pH 6.0, were added to concentrations of 1 and 10 mM, respectively. Nickel-charged agarose resin (Probond, Invitrogen) was equilibrated with 50 mM Tris-HCl, 0.3 M NaCl, 10 mM imidazole, pH 8.0, and incubated with the conditioned medium. After incubation at 4°C for 1 h, the resin was transferred to a column and washed with 50 mM Tris-HCl, 0.3 M NaCl, 20 mM imidazole, pH 8.0. The His-tagged proteins were eluted with 50 mM Tris-HCl, 0.3 M NaCl, 250 mM imidazole, pH 8.0. Purified proteins were dialyzed against phosphate-buffered saline and quantified using a BCA protein assay kit with bovine serum albumin as a standard (Pierce, Rockford, IL). Purity was determined by SDS-polyacrylamide gel electrophoresis followed by colloidal Coomassie G-250 Blue (GelCode Blue, Pierce, Rockford, IL) staining and judged to Ͼ95%.
Single-site substitutions may result in a change of the overall fold of domain VI. Gel filtration analysis of mutants, including Y130A and R239A, were therefore performed in 0.05 M phosphate buffer with 0.15 M NaCl, pH 7.0, using a Superose 12 HR 10/30 column on an Ä KTA EXPLORER design system (Amersham Pharmacia Biotech). The majority of the mutants and the wild-type domain VI eluted at a volume corresponding to a molecular weight of ϳ65,000 (data not shown). The results indicate that these molecules are monomers without mis-folding.
Synthetic Peptides-Peptide synthesis was performed on ABI model 433A peptide synthesizers at the Facility for Biotechnology Resources (U. S. Food and Drug Administration). All the peptides were prepared with a C-terminal amide group. The peptides were purified by reversephase high performance liquid chromatography and characterized by mass spectrometry.
Real-time Heparin Binding Kinetics of Recombinant Proteins Measured by Surface Plasmon Resonance-Biotinylated heparin (Celsus Laboratories, Inc., Cincinnati, OH) at 40 g/ml in running buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, including 0.005% surfactant P20) was immobilized on a streptavidin-coated sensor chip (Sensor Chip SA, BIAcore, Inc., Piscataway, NJ) at 10 l/min for 4 min to an immobilization level of 300 resonance units. In the affinity measurements, recombinant proteins at different concentrations (50 -200 nM) were injected on the heparin-coated surface at 30 l/min in the running buffer at 25°C, and the binding and dissociation were registered (2 min each) in a BIAcore TM 1000 instrument (BIAcore, Inc.). The streptavidinheparin surface was regenerated at the end of each run by two successive injections of 30 l of 20 mM NaOH containing 1 M NaCl. In control experiments with the same concentrations of recombinant proteins, but with a blank streptavidin sensor chip, no binding was seen. The sensorgrams obtained were analyzed by nonlinear least square curve fitting using BIAevaluation 2.1 software assuming single-site association and dissociation models.

Expression and Cell Binding Activity of Recombinant Laminin ␣ 5 N-terminal Globular Domain VI-Constructs
were generated for expression of mouse laminin ␣ 5 domain VI and the first four EGF-like repeats of domain V in mammalian cells (Fig. 1). In addition, a similar construct was also generated for laminin ␣ 1 domain VI, which corresponds to that described previously (15). The secreted recombinant proteins were purified by Ni-agarose chromatography from the conditioned medium of transfected COS-7 cells with a yield of 1-2 g/ml medium. The purity of the protein preparations was more than 95% as judged from Coomassie Blue-stained gels after SDSpolyacrylamide gel electrophoresis (data not shown). Recombinant fragments containing domains VI through IV of the laminin ␣ 1 and ␣ 2 chains have previously been shown to promote binding of HT-1080 fibrosarcoma cells (13,14). This cell line was therefore used to analyze the adhesive properties of the recombinant laminin ␣ 5 domain VI. HT-1080 cells showed strong binding to domain VI, while the recombinant domain V, consisting of the EGF-like repeats, showed no cell binding, demonstrating that the cell binding activities reside in domain VI (Fig. 1).
Effect of Synthetic Peptides on Cell Binding to Laminin ␣ 5 Domain VI-To localize potential cell-binding sites on laminin ␣ 5 domain VI, we screened synthetic peptides for their inhibitory effect on HT-1080 cell binding to domain VI. We synthesized 15 peptides that corresponded to possible integrin recognition sequences (for review, see Ref. 20) within laminin ␣ 5 domain VI and to active sequences previously identified in laminin ␣ 1 domain VI (Table I). Four peptides (S1, S2, S6, and S7) were found to inhibit HT-1080 cell binding to domain VI (Fig. 2), whereas the other peptides showed no or only small effects on the cell-domain interaction. A GRGDS peptide, which is reported to block the function of various integrins, had no effect on cell binding. The sequences of peptides S1 and S2 overlap with four residues, but only S2 showed strong activity when tested in a direct cell binding assay (data not shown). The S6 and S7 peptides overlap by eight residues and showed no activity when tested for direct cell binding (data not shown). None of the peptides inhibited cell binding to laminin-10/11, suggesting that other cell-binding sites are available on the intact molecule (data not shown). Taken together, these results suggest that two sequences: S2, 123 GQVFHVAYVLIKF 135 and S6, 225 RDFTKATNIRLRFLR 239 , separated by ϳ90 residues, directly interact with cell surface receptors within domain VI.
Cell Binding Activity of Laminin ␣ 5 Domain VI Mutants-We next examined the functional role of these two sites within domain VI by alanine substitution mutagenesis. Our laboratory has previously identified several synthetic peptides active for cell binding within domain VI of the mouse laminin ␣ 1 chain (19). The S2 peptide corresponds to the highly active peptide A-13 (RQVFQVAYIIIKA) of laminin ␣ 1 domain VI. Deletion analysis revealed that the VAYI sequence within A-13 is critical for high cell binding activity (19). The active core sequence VAYI is conserved in the laminin ␣ 5 chain with the exception of a substitution of Ile with Val (Fig. 3). To test the importance of the tyrosine in the VAYI site, we generated a single substitution mutant by replacing tyrosine with alanine (Y130A). This showed that Tyr 130 plays a significant role for cell binding, as the mutant domain VI with Y130A (A5VI-Y130A) reduced cell binding to ϳ25% of that of wild-type domain VI (Fig. 4).
The S6 peptide overlaps with two previously reported active synthetic peptides (A-24 and A-25) derived from the laminin ␣ 1 chain (19). The S6 sequence is characterized by five basic residues (four Arg and one Lys). Notable, three of the four Arg residues are conserved in domain VI of all the laminin ␣ chains (Fig. 3). To analyze the contribution of the individual Arg/Lys residues within the S6 site to cell binding, we generated five single substitution mutants by replacing Arg/Lys residues with Ala (R225A, K229A, R234A, R236A, and R239A). All the mutants were expressed at a level similar to wild-type domain VI, suggesting that the mutations did not cause unstable synthesis of domain VI due to unfolding of the proteins (data not shown). Cell binding to A5VI-K229A was very poor, while neither A5VI-R234A nor A5VI-R236A had a significant effect on cell binding (Fig. 4). R225A and R239A also showed a significant reduction in cell binding, by ϳ60% of the wild type (Fig. 4). These data indicate that four residues, Tyr 130 , Arg 225 , Lys 229 , and Arg 239 , within these two regions are crucial for cell binding.

Kinetic Analysis of Interactions between Heparin and Laminin ␣ 5 Domain VI Mutants using Surface Plasmon Resonance-
Since heparin binding activity has been localized to domain VI of the laminin ␣ 1 and ␣ 2 chains (13-15), cell binding to recombinant laminin ␣ 5 domain VI was examined in the presence of heparin and other heparin-like glycosaminoglycans. Heparin and heparan sulfate inhibited HT-1080 cell binding to domain VI, whereas keratan sulfate was much less inhibitory (Fig. 5). These results indicate that the heparin-binding site overlaps with the cell-binding sites.
To examine the relationships between the heparin and cellbinding site, the binding kinetics between immobilized biotinylated heparin and the laminin ␣ 5 domain VI mutants were measured directly by real-time biomolecular interaction analysis using surface plasmon resonance on a BIAcore TM system. Similar equilibrium dissociation constants (K d ϭ 9 -16 nM) were obtained for the interactions of wild-type domain VI of the laminin ␣ 5 chain and five of the mutants with heparin (Table  II). Less than a 2-fold difference was observed for the kinetic binding constants between wild-type domain VI and five of the mutants including Y130A, R225A, K229A, R234A, and R236A. Statistically significant differences in kinetic constants derived from BIAcore experiments are generally considered to be at least 5-10-fold. This demonstrates that the heparin binding activity was unchanged, indicating that the structural integrity of domain VI was maintained in these mutants and that these positions were not part of the heparin-binding site. In contrast, no binding of the R239A mutant to heparin was observed even at a high protein concentration (200 nM). Muta-  tion of Arg 239 was also found to be critical for cell binding, reducing activity by ϳ60% compared with wild-type domain VI (see above). Consistent with the cell binding results, no heparin binding was observed for a recombinant protein consisting of the first four EGF-like repeats of domain V. These results demonstrate that the heparin-binding site is located within domain VI. Interestingly, the binding affinity of domain VI of the laminin ␣ 1 chain to heparin was about 4-fold weaker (K d ϭ 45 nM), demonstrating differences in affinity for heparin between the laminin domain VI isoforms. Taken together, these results indicate that Arg 239 within the S6 site is critical for both heparin and cell binding.
Effect of Integrin Monoclonal Antibodies on Cell Binding to Laminin ␣ 5 Domain VI-Cell binding to laminin ␣ 5 and ␣ 1 domain VI and to laminins containing either the ␣ 5 chain (laminin-10/11) or the ␣ 1 chain (laminin-1) was examined in the presence of different anti-integrin antibodies to identify integrin receptors involved in domain VI binding. The ␤ 1 integrin antibody demonstrated partial inhibition of HT-1080 cell binding to laminin-10/11 but strong inhibition of cell binding to laminin-1 (Fig. 6). The monoclonal antibody against integrin ␣ 6 inhibited cell binding to laminin-1 but not laminin-10/11, while a monoclonal antibody against integrin ␣ 3 showed a weak inhibitory effect on cell binding to laminin-10/11. Other antibodies against integrin subunits, including ␤ 4 , ␣ 1 , ␣ 2 , ␣ 4 , ␣ 5 , and ␣ V , had no effect on cell binding. Cell binding to both laminin-1 and laminin-10/11 was dependent on divalent cations, since it was abolished by 5 mM EDTA (Fig. 6), which supports the role of integrins or ␣-dystroglycan as receptors. These results suggest that different integrin receptors mediate HT-1080 cell binding to laminin-1 and laminin-10/11 and that the ␣ 3 ␤ 1 integrin binds laminin-10/11, agreeing with previous reports (6, 7).
The monoclonal antibody against the ␤ 1 integrin subunit inhibited HT-1080 binding to recombinant domain VI of both laminin ␣ 5 and ␣ 1 (Fig. 7, A and B). These results show that ␤ 1 subunit containing integrins is critical for cell binding to domain VI of both chains. Furthermore, 5 mM EDTA was also found to inhibit cell binding, supporting cation and integrin dependence of the interaction (Fig. 5). We then tested several monoclonal antibodies against integrin ␣ subunits to identify partners for the ␤ 1 integrins. Anti-integrin ␣ 3 antibodies strongly inhibited HT-1080 cell binding to laminin ␣ 5 and ␣ 1 domain VI (Fig. 7, A and B). Monoclonal antibodies against ␣ 1 , ␣ 5 , and ␣ V had no or only small effects on cell binding, indicating that these integrins were not major mediators of cell binding to laminin ␣ 5 and ␣ 1 domain VI (Fig. 7, A and B). HT-1080 cell binding to laminin ␣ 5 and ␣ 1 domain VI was reduced to about 30 -50% by anti-integrin ␣ 2 and ␣ 4 antibodies. Integrin ␣ 6 antibody also weakly reduced cell binding to laminin ␣ 5 domain VI but not to laminin ␣ 1 domain VI (Fig. 7B). Taken together, these results identify ␣ 3 ␤ 1 integrin as a major receptor to ␣ 5 domain VI. DISCUSSION The newly discovered laminin ␣ 5 chain has been implicated in various biological activities, such as angiogenesis and nerve regeneration. Previously, we identified a cell-binding site within the C-terminal G domain of the mouse laminin ␣ 5 chain using recombinant proteins and synthetic peptides (21). In the present study, we identified cell-binding sites on the laminin ␣ 5 N-terminal globular domain VI.
Recombinant proteins of laminin ␣ 5 domain VI and the first four EGF-like repeats of domain V were generated by mammalian expression. The recombinant protein was highly active for HT-1080 fibrosarcoma cell binding, while a recombinant protein consisting of only the EGF-like repeats showed no binding, indicating that domain VI contributed to the cell-binding site. This result is similar to previous reports from studies on recombinant fragments consisting of domains VI through IV, of the laminin ␣ 1 and ␣ 2 chains (13)(14)(15). In addition, recombinant laminin ␣ 5 domain VI was also found to bind other cell lines including mouse B16-F10 melanoma and mesangial cells, while weak binding was observed for a human submandibular gland cell line, indicating cell type-specific interaction of this domain (data not shown).
To localize cell-binding sites on laminin ␣ 5 domain VI, we screened domain VI-derived synthetic peptides for their effect on HT-1080 cell/domain VI interactions. Four peptides: S1, 115 EVNVTLDLGQVFH 127 ; S2, 123 GQVFHVAYVLIKF 135 ; S6, FIG. 6. Inhibition of cell binding by anti-integrin antibodies and EDTA to laminin-10/11 and laminin-1. Cell binding assays using HT-1080 cells were performed on controls without inhibitors (control), or in the presence of rat preimmune IgG (10 g/ml) (IgG), functionblocking anti-integrin monoclonal antibodies (10 g/ml), and EDTA (5 mM). Wells were coated with 2 g/ml laminin-10/11 or laminin-1. The results are expressed as percent binding of cells without inhibitors. Each value represents the mean of three separate determinations Ϯ S.D. Duplicate experiments gave similar results. 225 RDFTKATNIRLRFLR 239 ; and S7, 232 NIRLRFLRTNTL 243 , inhibited cell binding. These peptides corresponded to sequences in the laminin ␣ 1 chain (A-12, A-13, A-24, and A-25) previously reported to be highly active for cell binding (19). The S1 and S2 peptides overlap, as do S6 and S7, thus the results indicate that at least two sequences, 123 GQVFHVAYVLIKF 135 and 225 RDFTKATNIRLRFLR 239 , spaced by ϳ90 residues, directly interact with cell surface receptors within domain VI (Fig. 3). Alanine mutagenesis of recombinant domain VI identified four positions within these two sequences as critically involved in cell binding: Tyr 130 , Arg 225 , Lys 229 , and Arg 239 . A comparison with the sequences of domain VI of the laminin ␣ 1 , ␣ 2 , and ␣ 3 B chains reveals that Tyr 130 and Arg 239 are conserved, while Arg 225 and Lys 229 are conserved between the ␣ 3 and ␣ 5 chains, indicating that the position of the cell-binding sites varies. These differences may be important for regulation and specificity of receptor interactions within domain VI. The site-directed mutagenesis data also demonstrate that the two sites together contribute to a cell binding epitope and imply that these cell-binding sites are highly dependent on the conformation of domain VI. The importance of the three-dimensional structure of cell-binding sites has also been shown for other integrin ligands, including fibronectin and vascular cell adhesion molecule-1, where multiple contacts, involving several different ligand peptide segments, are formed between ligand and receptor (22,23).
Kinetic data obtained here by surface plasmon resonance analysis for six single-site alanine substitution mutants of laminin ␣ 5 domain VI demonstrate that one position (Arg 239 ) within the cell-binding site is also critical for heparin interactions. Cell binding to domain VI was sensitive to inhibition by heparin/heparan sulfate, demonstrating overlap of cell and heparin-binding sites. Interestingly, the binding constants for binding to heparin show a 4-fold difference between domain VI of the laminin ␣ 5 and ␣ 1 chain, with the ␣ 5 domain demonstrating highest affinity. Strong heparin binding affinity suggests the potential to bind heparan sulfate containing matrix molecules or cell surface receptors (21,24,25); accordingly, these molecules may interact mainly with the basic residues within the S6 site. Our results suggest that heparan sulfate-containing cell surface receptor interaction is required for efficient cell binding to laminin ␣ 5 domain VI. This is in accordance with previous studies, which indicate that heparan sulfate-containing cell surface receptors can function as co-receptors for integrins and that these co-receptors are essential for cell binding to some ligands, including the heparin III domain of fibronectin and the angiogenic inducer Cyr61 (17,26). The heparan sulfate-containing cell surface receptors may include syndecan-1 or ␣-dystroglycan; the latter has been shown to bind laminin-10/11 and domain VI of the laminin ␣ 1 chain (8,24). The role of these interactions may be important for cell type-specific binding or signaling (17,26). The heparin-binding site on laminin ␣ 5 domain VI may also function as a binding site for other matrix molecules, since it has been reported that laminin ␣ 1 FIG. 7. Inhibition of cell binding by anti-integrin antibodies to laminin ␣ 1 and ␣ 5 domain VI. Cell binding assays were carried out without inhibitor (control) or in the presence of function-blocking anti-integrin antibodies (10 g/ml). Wells were coated with 10 g/ml recombinant domain VI of the laminin ␣ 1 (panel A) or ␣ 5 chain (panel B). Each value represents the mean of three separate determinations Ϯ S.D. Duplicate experiments gave similar results.
domain VI binds to heparan sulfate chains of perlecan (24). Accordingly, binding through the heparin-binding site may be a mechanism for the regulation of interactions with cells and matrix assembly.
Several integrins have previously been implicated as receptors for laminin-10/11, including ␣ 3 ␤ 1 , ␣ 6 ␤ 1 , and ␣ 6 ␤ 4 (7). In this study, we demonstrate that domain VI of the laminin ␣ 5 chain is a binding site for integrins ␣ 3 ␤ 1 , ␣ 2 ␤ 1 , ␣ 4 ␤ 1 , and ␣ 6 ␤ 1 . HT-1080 cell binding was completely blocked by antiintegrin ␣ 3 or ␤ 1 antibodies, while function-blocking antibodies against the ␣ 2 , ␣ 4 , and ␣ 6 integrins showed a weaker effect, suggesting that the ␣ 3 ␤ 1 integrin is a major mediator of cell binding. Small or no effects were observed with antibodies against ␣ 1 , ␣ 5 , and ␣ V integrins. Comparison with a recombinant protein of laminin ␣ 1 domain VI showed similar integrin specificity except for ␣ 6 , where no effect was observed for laminin ␣ 1 domain VI. The inhibition results are in agreement with the reported integrin specificity of recombinant fragments of domains VI through IV, of the laminin ␣ 1 and ␣ 2 chains (13,14). The previous studies used the same cell line as here, but only results using anti-integrin ␣ 1 and ␣ 2 antibodies were reported. Our results using various monoclonal antibodies suggest that several integrins bind domain VI of the laminin ␣ 5 and ␣ 1 chains and that these receptors bind similar recognition sites within domain VI.
In conclusion, the present data represent the first mapping of sites within the N-terminal globular domain VI of the mouse laminin ␣ 5 chain responsible for cell binding. Our results suggest that heparan sulfate-containing receptors and integrins recognize domain VI. We found that two sequences, spaced by ϳ90 residues within laminin ␣ 5 domain VI, are critical for cell surface receptor binding and that at least four residues within these two regions together form a binding site(s) critical for receptor binding.