Biological Activities of Homologous Loop Regions in the Laminin (cid:1) Chain G Domains*

Laminin (cid:1) chains ( (cid:1) 1 – (cid:1) 5 chains) have diverse chain- specific biological functions. The LG4 modules of laminin (cid:1) chains consist of a 14-stranded (cid:2) -sheet (A-N) sand-wich structure. Several biologically active sequences have been identified in the connecting loop regions. Here, we evaluated the biological activities of the loop regions of the E and F strands in the LG4 modules using five homologous peptides from each of the mouse (cid:1) chains (EF-1: DYATLQLQEGRLHFMFDLG, (cid:1) 1 chain 2747–2765; EF-2: DFGTVQLRNGFPFFSYDLG, (cid:1) 2 chain 2808–2826; EF-3: RDSFVALYLSEGHVIFALG, (cid:1) Permafluore (Shandon, Pittsburgh, PA) an Axioplan-2 epifluo-rescence (Zeiss, Tokyo, Japan). Quantix (Photometrics, Mu¨nchen, Germany) (Scanalytics, Fairfax, VA). with HazeBuster software (Vay Tek, Fairfield, IA).

Laminins, multifunctional glycoproteins of basement membranes, have diverse biological activities, including promotion of cell adhesion, migration, neurite outgrowth, tumor metastasis, and angiogenesis (1). Laminins consist of three different subunits, ␣, ␤, and ␥ chains. So far, five ␣, three ␤, and three ␥ chains have been identified, and at least 15 isoforms (laminin-1 to -15) are formed by various combinations of each subunit (2)(3)(4)(5). Laminin-1, which is the most extensively characterized laminin isoform, has been analyzed for biological activity using proteolytic fragments, recombinant proteins, and synthetic peptides (6,7). Previously, we screened for cell-adhesive sequences in laminin-1 (␣ 1 ␤ 1 ␥ 1 ) using 673 overlapping synthetic peptides covering the entire protein (8 -11). Most of the active peptides were localized in the globular domains and found to play a critical role in binding to cell surface receptors in a peptide-and cell type-specific manner (12,13).
The C-terminal globular domains (G domains) of the laminin ␣ chains consist of five laminin G domain-like modules (LG1 to -5) that play a critical role in the biological functions of laminins (14). Several studies have focused on the biological activities of the ␣ 1 chain G domain. E8, a proteolytic fragment containing the LG1-3 modules of the ␣ 1 chain, possesses potent cell binding activity mediated through ␣ 6 ␤ 1 integrin (15,16). Recombinant and reconstitution experiments with the ␣ 1 chain G domain have suggested that this activity is dependent on protein conformation (17,18). Several synthetic peptides derived from the ␣ 1 chain G domain promote heparin binding, cell adhesion, neurite outgrowth, and tumor growth and metastasis (8, 19 -21). Some of the peptides promote acinar-like formation of human salivary gland cells in culture and inhibit acinar development and lung alveolar formation in organ culture (22)(23)(24)(25)(26)(27). Many of these peptides interact with integrins or syndecan-1, a membrane-associated heparan sulfate proteoglycan (8,13,25,27,28).
Biologically active sites in the G domains of the other ␣ chains have begun to be identified (29 -34). The A3G75 site (KNSFMALYLSKG, human laminin ␣ 3 chain 1411-1422) promoted syndecan-2-and syndecan-4-mediated cell attachment and neurite outgrowth (29,30). The A4G82 (TLFLAH-GRLVFM, mouse laminin ␣ 4 chain 1514 -1525) sequence showed heparin binding and cell attachment activity (31,32). These active sequences are in the homologous region of the LG4 modules of the ␣ 3 and ␣ 4 chains. The crystal structure of the laminin ␣ 2 LG5 module has been prepared and found to consist of 14 ␤ strands (A-N strands) (35). The homologous regions of A3G75 and A4G82 are located on the connecting loop region between the E and F strands (14,29,30,32). Additionally, mutagenesis analysis using recombinant LG4 proteins with serine or alanine substitutions of basic residues (Lys 1421 and Arg 1423 , human laminin ␣ 3 chain; His 1519 and Arg 1521 , mouse laminin ␣ 4 chain) in the E-F loop regions suggested that the loop regions are critical for the biological activity of the recombinant proteins (29,31).
In this study, we focused on the E-F loop regions of laminin ␣ chain LG4 modules. We prepared five homologous peptides (EF-1 to EF-5) from the LG4 modules of the ␣ chains (␣ 1 -␣ 5 ) and evaluated their biological activity. These peptides showed chain-specific cell attachment activity with different receptor interactions, including integrins and syndecans, and some of the peptides promoted neurite outgrowth. Further, a cyclized form of the peptide, which interacted with ␣ 2 ␤ 1 integrin, enhanced the biological activity.
A cyclic peptide, cyc-EF-1Xm (CLQLQEGRLHFXFDC, X: norleucine), was synthesized using the synthetic protocol described above with some modifications. Since the methionine residue was easily oxidized during a cyclization step, the methionine residue was replaced with norleucine. Protected peptide resins were prepared on NovaSyn® TGR resin by the Fmoc-based solid-phase peptide synthesis. Treatment of the protected peptide resin with a solution of trifluoroacetic acid/1,2ethanedithiol/H 2 O (95:2.5:2.5, v/v/v) at 0°C for 2 h afforded a linear peptide, which was precipitated with ethyl ether. The linear peptide was cyclized by oxidation in a mixture of H 2 O/acetic acid/dimethyl sulfoxide (45:45:10, v/v/v, 1 mM) at room temperature for 2 days and purified by reverse phase HPLC using a Cosmosil 5C18-ARII column (Nacalai Tesque, Kyoto, Japan) with a gradient of water/acetonitrile containing 0.1% trifluoroacetic acid.
Purity and identity of the peptides were confirmed by analytical HPLC and by a fast atom bombardment mass spectrometer (the GC-MS and NMR Laboratory, Graduate School of Agriculture, Hokkaido University) or an ion-spray mass spectrometer (Kyoto University).
Recombinant Protein-Construction of a plasmid for expression of a recombinant mouse laminin ␣ 1 chain LG4-5 module (rec-␣1LG4-5) and preparation of rec-␣1LG4-5 were carried out using similar procedures for the preparation of recombinant human laminin ␣ 3 chain proteins described previously (29,30). The pEF␣1Gmyc plasmid (31) that encoded the mouse laminin ␣ 1 chain G domain, a gift from Dr. Yasuo Kitagawa, was used as a template in the polymerase chain reaction to amplify the ␣ 1 chain LG4-5 sequence. The primers with the AvrII site were as follows: nucleotides 8158 -8175 of the mouse laminin ␣ 1 chain, 5Ј-GAGCCTAGGGCTACCACAGCCTGAACTG-3Ј (forward); nucleotides 9277-9297 of the mouse laminin ␣ 1 chain, 5Ј-GAGCCTAGGCCCT-CAGGCCCGGGGCAGGAATG-3Ј (backward). The cDNA of the mouse laminin ␣ 1 chain LG4-5 module (2704 -3083) were inserted into the NheI site, which locates between the human laminin ␥ 2 signal and the human IgG Fc sequences in the MO90 plasmid (29). The plasmid was transfected to 293T cells with the calcium phosphate transfection kit (Invitrogen). Conditioned medium with 1.5% fetal bovine serum (Invitrogen) in Dulbecco's modified Eagle's medium (DMEM; Invitrogen) was collected for 3 days at 12-h intervals, and 1 mM phenylmethylsul-fonyl fluoride (Sigma) was added. The conditioned medium was applied to a protein A affinity column (HiTrap; Amersham Biosciences) equilibrated in 10 mM Tris-HCl (pH 7.4) containing 150 mM NaCl, 2 mM EDTA, and 0.5 mM N-ethylmaleimide (buffer A). After washing, the bound protein was eluted with 0.1 M glycine-HCl (pH 2.7) and immediately neutralized by 1 M Tris-HCl (pH 9.0). The buffer was exchanged to buffer A using a desalting column (HiTrap; Amersham Biosciences). The rec-␣1LG4-5 protein was detected with biotinylated anti-human IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) and streptavidin-peroxidase (Sigma) in Western blotting. Protein concentration was determined with the BCA assay kit (Pierce).
Cell Attachment Assay-Cell attachment assays were performed in 96-well plates (Nunc, Inc., Naperville, IL) coated with either various amounts of synthetic peptides or the rec-␣1LG4-5 protein. For peptide coating, various amounts of peptides in 50 l of Milli-Q water were added to the wells and dried overnight at room temperature. For rec-␣1LG4-5 coating, various amounts of rec-␣1LG4-5 in 50 l of buffer A were added to the wells and incubated overnight at 4°C. The substratecoated wells were blocked with 150 l of 1% bovine serum albumin (BSA; Sigma) in DMEM at room temperature for 1 h and then washed with DMEM containing 0.1% BSA. Cells were detached with 0.02% trypsin-EDTA (Invitrogen) and recovered with DMEM containing 10% fetal bovine serum at 37°C for 20 min. After washing with DMEM containing 0.1% BSA, cells (2 ϫ 10 4 cells/100 l) were added to the wells and incubated at 37°C for 1 h. The attached cells were stained with 0.2% crystal violet aqueous solution in 20% methanol for 10 min. After washing with Milli-Q water, 1% SDS (150 l) was used to dissolve the stained cells, and the optical density at 570 nm was measured using a model 550 microplate reader (Bio-Rad).
For inhibition of cell attachment with heparin, EDTA, and peptides, human fibroblasts were incubated for 30 min at 37°C in the presence of either 10 g/ml heparin, 5 mM EDTA, or various concentrations of peptides. For inhibition of cell attachment with anti-integrin antibodies, human fibroblasts were preincubated in suspension with 10 g/ml of the anti-integrin antibody for 15 min at 37°C. Then the cells were added to the wells and incubated for 30 min at 37°C. Attached cells were measured as described above.
Neurite Outgrowth Assay-For neurite outgrowth assays, a 24-well plate was coated with the indicated amounts of peptides as described above. After priming with nerve growth factor (100 ng/ml) for 24 h, PC12 cells were washed with DMEM and seeded into the wells in serum-free DMEM/F-12 (Invitrogen). The cells were then incubated at 37°C for 24 h in 5% CO 2 . The cells were fixed and stained with a 0.2% crystal violet aqueous solution in 20% methanol for 10 min. After the wells were washed, the cells were analyzed under a microscope.
Next, we tested the neurite outgrowth activity of five peptides with PC12 rat pheochromocytoma cells (Fig. 3). EF-4 and EF-3 promoted neurite outgrowth with PC12 cells (Fig. 3, C and D), whereas EF-1, EF-2, and EF-5 did not show activity (Fig. 3, A, B, and E). These results suggest that the homologous EF peptides have cell type-specific cell attachment and neurite outgrowth activities.
Organization of Actin Filaments and Localization of Vinculin-Next, we focused on the EF-1, EF-2, EF-4, and EF-5 pep-tides, which promoted fibroblast attachment, and examined the organization of actin filaments and localization of vinculin on the peptide-coated plates (Fig. 4). On the AG73 peptide, actin filament spikes were associated with membrane ruffles, as previously described (40), and no focal contacts with vinculin were observed (Fig. 4E). The EF-1 peptide induced well organized actin stress fibers and focal contacts containing vinculin (Fig. 4A). In contrast, EF-2 and EF-4 induced a morphology similar to AG73 with filopodia and ruffles, and no focal contact was observed (Fig. 4, B and C). EF-5 did not induce actin stress fibers or focal contacts (Fig. 4D). These results demonstrate that the cellular responses to the peptides are different.
Effects of Heparin and EDTA on Fibroblast Attachment-Next, we evaluated the effects of heparin and EDTA on fibroblast attachment to the peptides (Fig. 5). We used AG73 and laminin-1 as controls for heparin-dependent and cationdependent cell attachment, respectively. Heparin strongly inhibited attachment on EF-2 and EF-4, whereas cell attachment on EF-1 and EF-5 was not inhibited by heparin. In contrast, EDTA significantly inhibited cell attachment on EF-1, but no effect on EF-2-and on EF-4-mediated attachment was observed. Further, fibroblast attachment to EF-5 was enhanced by EDTA. Fibroblasts attached to the EF-1 peptide in a cationdependent manner, suggesting that EF-1 promoted integrinmediated cell adhesion. However, cell attachment on EF-2 and EF-4 was inhibited by heparin, suggesting that these peptides have the potential to interact with membrane-associated heparin/heparan sulfate proteoglycans.
Attachment of 293T Cells Transfected with Either Syndecan-2 or Glypican-1 to Peptides-Syndecans and glypicans are cell surface heparan sulfate proteoglycans. Human neonatal dermal fibroblasts express syndecan-2 based on reverse transcriptase-PCR and Western blotting analysis (29). To examine whether syndecan-2 was involved in cell adhesion to EF-2 and EF-4, we used 293T cells overexpressing syndecan-2 and glypican-1 (29) and analyzed cell attachment (Fig. 6). 293T cell adhesion to EF-2 and EF-4 was significantly enhanced (more than 2-fold) by overexpression of syndecan-2 and slightly enhanced by overexpression of glypican-1. Overexpression of either proteoglycan had no affect on cell attachment to EF-1.
These results indicate that cell attachment to EF-2 and EF-4 mainly involved syndecan-2.
Effect of Anti-integrin Antibodies on Cell Attachment to Peptides-Since EF-1 promoted cation-dependent fibroblast at- tachment, we examined the effect of anti-integrin subunit antibodies on fibroblast attachment to EF-1 (Fig. 7). Cell attachment to EF-1 was significantly inhibited by anti-␣ 2 and anti-␤ 1 integrin antibodies. These antibodies did not affect cell attachment to EF-2 (data not shown). Other function-blocking anti-integrin antibodies, including anti-␤ 4 , ␣ 1 , ␣ 3 , ␣ 6 , and ␣ V , did not affect cell attachment to EF-1. The data indicate that the EF-1 peptide promotes ␣ 2 ␤ 1 integrin-mediated cell attachment.
Cell Attachment Activity of a Recombinant Laminin ␣ 1 Chain LG4-5 Module and Effect of the EF-1 Peptide on the Attach-ment-The interaction of the laminin ␣ 1 chain LG4-5 module containing the EF-1 sequence with ␣ 2 ␤ 1 integrin was not previously reported. Therefore, we prepared a recombinant laminin ␣ 1 chain LG4-5 module (rec-␣1LG4-5) and examined integrin-mediated cell attachment to the rec-␣1LG4-5 protein.
Fibroblasts attached to rec-␣1LG4-5 in a dose-dependent manner (Fig. 8A). Heparin inhibited 70% of the attachment (Fig.  8B), suggesting that cell surface heparin/heparan sulfate chains are involved in the fibroblast attachment to rec-␣1LG4-5. EDTA also inhibited about 40% of the cell attachment, suggesting that a part of the cell attachment was medi- a For cell attachment assays, various amounts of peptides were coated on 96-well plates as described under "Materials and Methods." Human neonatal dermal fibroblasts, HT1080 human fibrosarcoma cells, and HSG cells were used. In all cases, the biological activities of the peptides were quantitated and evaluated relative to those observed with AG73. Cell attachment was evaluated on the following subjective scale: ϩϩ, adhesion comparable with those on AG73; ϩ, weak adhesion compared with that on AG73; Ϫ, no adhesion.
b For neurite outgrowth assays with PC12 rat pheochromocytoma cells, various amounts of peptides were coated on the wells as described under "Materials and Methods." Neurite outgrowth was evaluated on the following subjective scale: ϩ, promote neurite outgrowth; Ϫ, inactive.
c,d Actin and vinculin were detected in the fibroblasts attached on peptides as described under "Materials and Methods." Formation of focal contacts was evaluated on the following subjective scale: ϩ, focal contact observed; Ϫ, not observed. e For cell attachment inhibition assays, human fibroblasts were incubated in the presence of 10 g/ml heparin or 5 mM EDTA as described under "Materials and Methods." f Receptors for cell attachment to EF-1, EF-2, and EF-4, which were inhibited by heparin or EDTA, were analyzed. Using 293T cells, which overexpressed syndecan-2, enhanced cell attachment to EF-2 and EF-4. Anti-␣ 2 and anti-␤ 1 integrin antibodies inhibited cell attachment to EF-1. ated by cation-dependent receptor(s), including integrins. An anti-␤ 1 integrin subunit antibody slightly inhibited fibroblast attachment to rec-␣1LG4-5, whereas an anti-␣ 2 integrin subunit antibody did not (Fig. 8B). The same concentration of anti-␤ 1 integrin antibody inhibited 90% of fibroblast attachment to laminin-1 (data not shown). Fibroblast attachment to rec-␣1LG4-5 was inhibited in the presence of both anti-␣ 2 and anti-␤ 1 integrin antibodies (Fig. 8B). The EF-1 peptide was also able to inhibit attachment to the recombinant protein (Fig. 8B). Taken together, interactions with ␣ 2 ␤ 1 integrin and with cell surface proteoglycan(s) are involved in fibroblast attachment to the laminin ␣ 1 chain LG4-5 module, and the EF-1 site is important for the interaction with ␣ 2 ␤ 1 integrin.
Active Core Sequences of EF-1-To determine the active core sequence of EF-1, we prepared systematically truncated N-and C-terminal peptides and tested their cell attachment activity using fibroblasts (Table II). EF-1d (LQLQEGRLHFMFDLG), an N-terminal truncated peptide of EF-1, with 4 amino acids removed, still retained full activity, whereas the activity of EF-1e, with a deletion of the N-terminal leucine and glutamine residues from EF-1d, was weaker than that of EF-1. When the amino-terminal LQ (EF-1e), LQLQE (EF-1f), or LQLQEG (EF-1g) were removed from EF-1d, a significant reduction in cell attachment activity was observed. Further, the removal of LQLQEGR (EF-1h) from EF-1d resulted in total loss of activity. EF-1j (DYATLQLQEGRLHFMFD), a C-terminal truncated peptide of EF-1, still retained full activity, whereas the activity of EF-1k, with a deletion of the C-terminal aspartate residue from EF-1j, showed no activity. EF-1m (LQLQEGRLHFMFD), an N-terminal and C-terminal truncated peptide of EF-1, showed weak cell attachment activity. These results suggest that the 13-amino acid sequence LQLQEGRLHFMFD is criti- cal for the cell attachment activity of EF-1. The cause of the reduced activity of EF-1m may be due to conformational changes, which are stabilized by the additional N-terminal amino acids.
Cell Attachment Activity of Cyclic EF-1Xm Peptide-Next, we prepared a cyclic EF-1Xm peptide to evaluate the role of conformation in the biological response (Table II). The cyc-EF-1Xm peptide significantly enhanced fibroblast attachment activity relative to the linear form (Fig. 9A). Fibroblast attached on cyc-EF-1Xm showed well organized actin stress fibers and focal contacts (Fig. 9B). Further, fibroblast attachment to cyc-EF-1Xm was significantly inhibited by EDTA and by anti-␣ 2 and anti-␤ 1 integrin antibodies (Fig. 9C), whereas the other integrin antibodies and heparin did not affect cyc-EF-1Xmmediated fibroblast attachment. These results indicate that the cyc-EF-1Xm peptide interacts with ␣ 2 ␤ 1 integrin similar to that of a linear peptide EF-1.
Effect of Peptides on Cell Attachment to the EF-1 Peptide-Next, we examined the effect of various peptides on EF-1mediated fibroblast attachment (Fig. 10). The EF-1 peptide showed inhibitory activity, as expected (Fig. 10). EF-1m slightly inhibited cell attachment. The cyc-EF-1Xm peptide strongly inhibited fibroblast attachment to EF-1 relative to the linear form. FIB-1, an RGD (Arg-Gly-Asp) sequence-containing peptide derived from fibronectin that interacts with ␣ V ␤ 3 integrins (8,41), did not affect EF-1-mediated cell attachment. The data suggest that peptide conformation is important for the interaction between the EF-1 sequence and ␣ 2 ␤ 1 integrin and that the interaction is specific. DISCUSSION Previously, we identified the A3G75 and A4G82 sequences as cell attachment and heparin binding sites in the G domains of the human laminin ␣ 3 and the mouse laminin ␣ 4 chains, respectively (29 -32). Based on the alignment of the LG4 sequences, the A3G75 and A4G82 sequences are homologous sites located in the connecting loop between the E and F strands (14,35). Mutagenesis using recombinant LG4 proteins with serine or alanine substitution mutations of basic residues (Lys 1421 and Arg 1423 , human laminin ␣ 3 chain; His 1519 and Arg 1521 , mouse laminin ␣ 4 chain) in the E-F loop regions has suggested that the loop regions are critical for biological activity of the recombinant proteins (29,31). Here, we have described five homologous peptides (EF-1 to EF-5) from the E-F loop regions of mouse laminin ␣ 1 -␣ 5 chain LG4 modules. The EF peptides showed chain-and cell type-specific cell attachment and neurite outgrowth activity. EF-2 and EF-4 promoted cell attachment of all of the cell types tested. These results indicate that the E-F loop sites in the ␣ 2 and ␣ 4 chains may play a critical role in diverse tissues where they are expressed. EF-1 showed cell attachment activity with fibroblasts and HT-1080 cells, whereas EF-5 promoted only fibroblast attachment. The ␣ 1 and ␣ 5 chains are expressed in the early embryo and in diverse adult tissues, respectively (1). They may interact with mesenchymal cells, but not with HSG cells, via the E-F loop sites in each. Further, PC12 cells promoted neurite outgrowth on EF-3 and EF-4. Laminin ␣ chains are expressed in various neural tissues (1,42). It is suggested that the E-F loop regions in the ␣ 3 and ␣ 4 chains are important for the interaction with neural cells, whereas the other sites may function in the ␣ 1 , ␣ 2 , and ␣ 5 chains.
Syndecans, a family of heparan sulfate proteoglycans, interact with extracellular matrices and various growth factors (43,44). Previously, we defined the A3G75 site as a syndecan-2 and -4 binding site in fibroblasts (29) and the AG73 sequence as a syndecan-1 binding site in HSG cells (13,25). In syndecanmediated cell attachment to the heparin binding domain of fibronectin and to the AG73 peptide, actin filament spikes were associated with membrane ruffles, and no focal contacts with vinculin accumulation were observed (40,45). In this study, actin organization and vinculin localization of fibroblasts on EF-2 and EF-4 were similar to that on AG73. Furthermore, cell attachment to EF-2 and EF-4 was inhibited by heparin and was significantly enhanced by overexpression of syndecan-2. These results suggest that cell attachment to EF-2 and to EF-4 is mainly mediated by syndecan-2.
CLQLQEGRLHFXFDC ϩϩ a Sequences of the synthetic peptides are given in the single-letter code. All peptides have C-terminal amides. The single letter X denotes the norleucine residue.
b Activity was scored on the following subjective scale: ϩϩ, cell attachment comparable with that on EF-1; ϩ, weak adhesion compared with that on EF-1; Ϫ, no adhesion. Active core sequences are shown in boldface type.
tively inhibited fibroblast attachment to this protein. Taken together, EF-1 is a new peptide specifically interacting with ␣ 2 ␤ 1 integrin, and this sequence in the laminin ␣ 1 chain is a novel ␣ 2 ␤ 1 integrin-binding site in laminin-1. The AG73 peptide (RKRLQVQLSIRT, mouse laminin ␣ 1 chain 2719 -2730), which is derived from the LG4 module and binds to syndecans, significantly inhibited cell attachment to the rec-␣1LG4-5 pro-tein (data not shown) and E3 (25). These results suggest that cell surface syndecans and integrins interact with the laminin ␣ 1 chain LG4-5 module via the AG73 and EF-1 sites, respectively.
We found that EF-5 showed only weak attachment activity with fibroblasts which was enhanced by EDTA. Actin organization and vinculin localization within the fibroblasts on EF-5 differed from that of cells adhering via syndecans or integrins. Recently, we found that A5G77 (LVLFLNHGHFVA, mouse laminin ␣ 5 chain 3307-3318) peptide, contained within the EF-5 sequence, inhibited branching morphogenesis (50). However, the homologous 12-mer peptides of A5G77 from the ␣ 1 chain were not active. Further, the inhibition of branching morphogenesis by A5G77 resulted in a different morphology from that observed with AG73, a syndecan-binding peptide from the laminin ␣ 1 chain G domain (27). These results suggest that the EF-5 region may interact with different receptor(s), resulting in ␣ 5 chain-specific biological activity.
EF-3 promoted neurite outgrowth with PC12 cells. Previously, we described that a recombinant human laminin ␣ 3 chain LG4 module and a peptide, A3G75, located on the E-F loop region of the module also promote neurite outgrowth with PC12 cells (30). The A3G75 peptide inhibited PC12 cell attachment to the recombinant human laminin ␣ 3 chain LG4 module (30). These results indicate that this region of the ␣ 3 chain is important for neurite outgrowth.
Previously, several cell-adhesive peptides were cyclized and showed increased biological activity over their noncyclized form. A cyclic YIGSR (Tyr-Ile-Gly-Ser-Arg) peptide enhanced cell attachment activity and showed increased inhibition of tumor metastasis over that observed with linear peptide (51). Cyclic RGD peptides also showed increased inhibition in their blocking activity (52). Moreover, cyclic RGD peptides interact with different integrins, including ␣ IIb ␤ 3 , ␣ V ␤ 3 , and ␣ 5 ␤ 1 integrins (53). In this study, cyclization of the linear EF-1Xm peptide significantly enhanced cell attachment activity and the FIG. 9. A, fibroblast attachment activity of the cyc-EF-1Xm peptide. 96-Well plates were coated with various amounts of peptides. The cell attachment assay of fibroblasts was carried out as described in the legend to Fig. 2. Triplicate experiments gave similar results. B, organization of actin filaments and localization of vinculin on the cyc-EF-1Xm-coated plates. The cell attachment assay and staining of actin and vinculin were carried out as described in Fig. 4. The pictures of staining of actin (B-I) and vinculin (B-II) were merged (B). Yellow, red, and blue pseuedocolors were given for images for fluorescein isothiocyanate, rhodamine, and DAPI, respectively. The arrows represent focal contacts. Bar, 20 m. C, effect of anti-integrin subunit antibodies on fibroblast attachment to the cyc-EF-1Xm-coated plate. 96-Well plates were coated with cyc-EF-1Xm (0.5 g/well). The inhibition assay using 10 g/ml heparin, 5 mM EDTA, and 10 g/ml anti-integrin antibodies was performed as described in the legends to Figs. 5 and 7. Each value represents the mean of three separate determinations Ϯ S.D. Duplicate experiments gave similar results. *, p Ͻ 0.01. inhibitory effect on cell attachment to EF-1. The cyc-EF-1Xm peptide induced actin stress fiber and focal contact formation and interacted with ␣ 2 ␤ 1 integrin similar to EF-1. These results indicate that integrin specificity of the EF-1 sequence is enhanced by the cyclization. Additionally, the cyclic A4G82 peptide, involved in the EF-4 sequence, significantly increased heparin binding and cell attachment activity when compared with the linear peptide. 2 Taken together, the loop conformation of the E-F regions is critical for specific binding to cellular receptors.
Here, we demonstrated the biological activities of five homologous peptides derived from the E-F loop regions of the laminin ␣ chain LG4 modules. These homologous peptides showed chain-specific cell attachment and neurite outgrowth activities and interacted with different cell surface receptors, such as ␣ 2 ␤ 1 integrin and syndecan-2. These results suggest that the E-F loop regions are important in various biological activities of the laminin isoforms in tissue-specific locations.