Laminin Interactions Important for Basement Membrane Assembly Are Promoted by Zinc and Implicate Laminin Zinc Finger-like Sequences*

Laminin is an abundant basement membrane (BM) glycoprotein which regulates specific cellular functions and participates in the assembly and maintenance of the BM superstructure. The assembly of BM is believed to involve the independent polymerization of collagen type IV and laminin, as well as high affinity interactions between laminin, entactin/nidogen, perlecan, and collagen type IV. We report here that Zn 2 (cid:49) can influence laminin binding activity, in vitro . Laminin contains 42 cysteine- rich repeats of which 12 contained nested zinc finger consensus sequences. Recently, the entactin binding site was mapped to one of these zinc finger-containing repeats on the laminin (cid:103) chain (Mayer, U., Nischt, R., Poschl, E., Mann, K., Fukuda, K., Gerl, M., Yamada, Y., and Timpl, R. (1993) EMBO J. 12, 1879–1885). Based on these observations, the effect of a series of essential ions (Ca 2 (cid:49) , Cd 2 (cid:49) , Cu 2 (cid:49) , Mg 2 (cid:49) , Mn 2 (cid:49) , and Zn 2 (cid:49) ) on laminin binding activity was evaluated. Zn 2 (cid:49) was found to be the most effective at enhancing laminin-entactin and lami- nin-collagen type IV binding. Laminin-bound Zn 2 (cid:49) was detected by flame atomic absorption spectroscopy at a maximum of 8 mol/mol of laminin. Furthermore, Ca 2 (cid:49) dependent laminin polymerization was unaffected by Zn 2 (cid:49) , an observation consistent with the lack of zinc finger-containing repeats ), zinc was the most effective at enhancing laminin-entac-tin and laminin-collagen type IV interactions. The zinc effect was saturable, and a maximum of 8 mol of zinc/mol of laminin was detected by flame atomic absorption spectroscopy. Laminin polymerization was not zinc-dependent, consistent with the lack of zinc finger-like sequences in the terminal domains which are required for polymerization. Our results provide biochemical evidence supporting earlier reports (20, 21) that laminin-collagen type IV interactions could occur without en-* -nitrophenyl

Basement membrane (BM) 1 is a distinct type of extracellular matrix, which divides tissue into compartments, provides filtration and structural support, sequesters growth factors, and directly influences cellular behavior (2,3). These functions are believed to be dependent on BM composition and ultrastructure. The assembly of BM involves the synthesis and secretion of the major BM components (laminin, entactin/nidogen, perlecan, and collagen type IV) into a diffusion-limited space where, by a mass action-driven process, they become interconnected through site-specific interactions generating a 50 -200 nm thick network. Little is known of the physical nature of these binding sites or of the regulatory factors which govern their interactions. Deviation in BM metabolism is believed to underlie complications associated with diseases such as Alport's and Goodpasture's syndromes (4,5), diabetes mellitus (6), amyloid (7), and Alzheimer's disease (8,9).
Laminin is a unique and essential component of BM, contributing to its architecture, and providing signals for cell adhesion, migration, and differentiation. The prototype of this family of glycoproteins, laminin-1, is derived from Engelbreth-Holm-Swarm (EHS) tumor and is composed of three different subunits ␣1, ␤1, and ␥1 (previously A, B1, and B2, respectively) which form a multidomain cruciform structure possessing one long and three short arms. Several studies have shown that laminin can exist as a polymer both in vivo and in vitro. The assembly of BM involves primarily the polymerization of 2 independent networks: one of collagen type IV, which becomes covalently stabilized (10 -13), and the other of laminin, in a noncovalent, calcium-dependent process (14 -16). In the BM of EHS tumor, about 80% of the laminin is deposited as an independent polymer, while the remainder is found also anchored, noncovalently, to the collagen type IV network (17). High affinity interactions between laminin-entactin and entactin-collagen type IV have been reported supporting the concept that the 2 networks are interconnected by entactin (18,19). However, direct association between laminin short arms and collagen type IV has also been reported (20,21).
A recent study focusing on Alzheimer's disease presented data showing Zn 2ϩ stimulated laminin binding to the Alzheimer's amyloid precursor protein (22). In the same report, it was observed that of the many cysteine-rich domains in laminin (also known as EGF-like repeats), some contained the zinc finger consensus sequence (23). Of 42 cysteine-rich repeats found on the amino-terminal ends of the ␣1, ␤1, and ␥1 chains, 12 appear to have Cys spacing similar to that observed for certain zinc fingers. More recently, the entactin binding site was localized to a 58-mer corresponding to the 4th repeat of domain III, on the ␥1 chain (1) which, as our observation suggests, contains a nested zinc finger sequence.
On the basis of these observations, we investigated the influence of different essential ions on 3 laminin interactions important for BM assembly. We report here that, of all the essential ions tested (Ca 2ϩ , Cd 2ϩ , Cu 2ϩ , Mg 2ϩ , Mn 2ϩ , and Zn 2ϩ ), zinc was the most effective at enhancing laminin-entactin and laminin-collagen type IV interactions. The zinc effect was saturable, and a maximum of 8 mol of zinc/mol of laminin was detected by flame atomic absorption spectroscopy. Laminin polymerization was not zinc-dependent, consistent with the lack of zinc finger-like sequences in the terminal domains which are required for polymerization. Our results provide biochemical evidence supporting earlier reports (20,21) that laminin-collagen type IV interactions could occur without en-tactin acting as a bridging molecule. We also provide evidence that the entactin and collagen type IV binding sites on laminin may involve a zinc finger-like secondary structure. To our knowledge, this is the first report of an extracellular zinc finger motif acting directly as, or contributing to, high affinity protein-protein interactions. It also implicates Zn 2ϩ as an important cofactor/modulator of BM assembly.

MATERIALS AND METHODS
Protein Purification-Laminin and entactin were purified from Engelbreth-Holm-Swarm (EHS) mouse sarcoma propagated in nonlathyritic mice (Swiss Webster, Charles Rivers, Montreal, Quebec), harvested at 2-4 cm, frozen in N 2 (l), and stored at Ϫ70°C. Purification was based on a NaCl extraction procedure in Refs. 24 and 25, and from Hynda Kleinman, NIH). 2 Tumor was homogenized and washed 2 ϫ in 3.4 M NaCl, 50 mM Tris, 2 mM EDTA, 1 mM N-ethylmaleimide, 1 mM phenylmethylsulfonyl fluoride, pH 7.5, centrifuged at 10,000 ϫ g for 15 min, discarding the supernatant each time. The residue was extracted in the same buffer but with 0.5 M NaCl, for 4 -8 h, then centrifuged as above. The supernatant was precipitated with 30% saturated ammonium sulfate, and the pellet was redissolved and dialyzed against 50 mM Tris, 150 mM NaCl, pH 7.5 (TBS). NaCl was increased to 1.7 M, precipitating collagen type IV which was removed by centrifugation. The supernatant was applied to a Bio-Gel A-5m (2.5 ϫ 120 cm) gel filtration column eluted with TBS. Fractions containing the lamininentactin complex were pooled, concentrated with Aquacide II (Calbiochem), and dialyzed against TBS/2 M guanidine HCl. The dialysate was applied to a Sephacryl-S400HR (2.5 ϫ 140 cm) column eluted with the same buffer. Laminin and entactin fractions were pooled, dialyzed against TBS, concentrated, frozen in N 2 (l), and stored at Ϫ70°C. The steps in the procedures were monitored by SDS-polyacrylamide gel electrophoresis on a 5-10% acrylamide gradient gel (26).
Solid Phase Binding Assay-An enzyme-linked immunosorbent assay technique was employed to study the interaction between laminin, entactin, and collagen type IV. Polystyrene microtiter plates (Immulon 4, Dynatech Laboratories) were coated with 100 l of 200 ng/ml entactin or 500 ng/ml collagen type IV in 20 mM NaHCO 3 , pH 9.6. After an overnight incubation at 4°C, the plates were washed with 20 mM Tris-HCl, 150 mM NaCl, pH 7.5 (TBS), then incubated with 1% bovine serum albumin in TBS (150 l) for 2 h at 37°C to block residual hydrophobic surfaces. Plates were washed with TBS containing 0.05% (w/v) Tween 20 (TBS-Tween), then laminin at different concentrations was added in the same buffer and left overnight at 4°C to allow maximum binding. Plates were washed again in TBS-Tween and incubated with mouse anti-laminin IgG (Sigma) diluted 1:750 in TBS-Tween, 0.1% bovine serum albumin for 2 h at 37°C, rewashed, then incubated in the same way with goat anti-rabbit IgG conjugated with alkaline phosphatase (Boehringer Mannheim) at a 1:500 dilution. After washing, bound IgG was detected by the addition of alkaline substrate solution containing 2 mg/ml p-nitrophenyl phosphate, 0.1 mM ZnCl 2 , 1 mM MgCl 2 , and 100 mM glycine, pH 10.0. Plates were left at room temperature for 15-30 min, and the reaction was stopped with 50 l of 2 M NaOH. The absorbance due to the released p-nitrophenol was measured at 405 nm with a Titertek Multiscan/MCC 340 (Flow Laboratories). The amount of bound ligand was determined by subtracting the optical density of the blank wells, in which the ligand was omitted. For direct quantitation, ligand standards were precoated onto wells on the same plates as the test proteins in order to generate standard curves.
To measure coating efficiency, laminin, entactin, and collagen type IV were radioiodinated with 125 I, using IODOBEADS (Pierce), and the amount of protein coated onto the microtiter plates was measured by subtracting the counts/min remaining in the coating buffer after coating from the total. Coating efficiency was 90 -100% at the concentrations used in the binding assays.
Binding data were analyzed as in Ref. 22 with a nonlinear curve fit program (SigmaPlot, Jandel Scientific) using Equations 1 for a onebinding site model with nonspecific binding or 2 for a two-binding site model with nonspecific binding, where S is the proportionality constant for nonspecific binding and L is the laminin concentration.
In all cases, the data fit the one-site model the best, and nonspecific binding was very low (S Ͻ 10 Ϫ10 ). Zinc Analysis-Laminin zinc content was assayed by flame atomic absorption spectroscopy using elemental zinc standards (0 -2 ppm). Laminin was either assayed directly after purification or after loading with ZnCl 2 , which involved dialysis against 1 liter of TBS containing 50 M ZnCl 2 overnight, followed by extensive dialysis against TBS, 0.1 mM EDTA, then just TBS to remove unbound Zn 2ϩ . Samples at 0.5 mg/ml protein were dissolved in 2% nitric acid prior to analysis.

Purification of Laminin and Entactin-Laminin
and entactin were purified by NaCl extraction (see "Materials and Methods") and gel filtration chromatography. The laminin-entactin complex (normally in a 1:1 stoichiometry) was isolated from the EHS extract by gel filtration on a Bio-Gel A-5m column eluted with TBS (Fig. 1A). Dialysis against TBS/2 M guanidine HCl dissociated the complex, and laminin and entactin were separated by gel filtration on a Sephacryl-S400HR column (Fig. 1B). Urea (2 M) did not completely dissociate the complex (not shown). Nonlathyritic mice were used to maintain collagen type IV cross-linking and minimize contamination of the laminin. Trace amounts of collagen type IV, possibly complexed with laminin, were resolved in the leading minor peak (Fig.  1B), as determined by polyacrylamide gel electrophoresis (not shown). A yield of 1.96 Ϯ 0.28 mg of laminin and 0.21 Ϯ 0.03 mg of entactin per g of tumor (n ϭ 4) was achieved which is comparable to the yields attained with the EDTA extraction method (25).
Laminin Binding to Entactin-Laminin binding to entactin 2 H. Kleinman, personal communication.  (19). The binding data fit one class of binding site model best supporting the conclusion made by others, that there is only one entactin binding site on laminin. But positive confirmation was not possible with this assay since neither the orientation nor the availability of binding sites could be determined.
The protein denaturant urea at 2 M prevented binding, indicating the interaction was conformation-dependent (Fig. 2). The reduction in binding when the NaCl concentration was increased (0.3 M) indicated that the interaction was also ionic in nature. Free sulfhydryl groups were also implicated by the reduction in binding activity observed after alkylation with N-ethylmaleimide without reduction of disulfide bonds. Also, the inhibition of laminin binding activity with EDTA suggested a divalent metal requirement, not previously observed, and, of a battery of common trace elements tested at their respective normal plasma concentrations, zinc was the most effective at enhancing laminin binding activity (Fig. 3). Furthermore, the effect zinc had was exerted through laminin only since preincubation of the entactin with Zn 2ϩ was no more effective at increasing binding activity than omitting Zn 2ϩ altogether. CuCl 2 was the only other divalent ion that had a measurable effect on laminin binding activity, probably reflecting its similarity in atomic mass to Zn 2ϩ (Cu 2ϩ ϭ 63.55 versus Zn 2ϩ ϭ 65.38). The binding maximum with Cu 2ϩ was much lower than that observed with Zn 2ϩ . Trivalent metals such as Fe 3ϩ and Al 3ϩ were found to cause a nonspecific increase in laminin binding activity and are likely not involved in this interaction (not shown). The suppression of binding by heparin may be caused by steric hindrance, although the heparin binding regions on laminin that have been mapped (VI domain of ␣ short arm and G domain of ␣1 long arm) appear to be clear of the entactin binding domain (28,29). Alternatively, the inhibitory effect of heparin may be due to the sequestration of Zn 2ϩ (30).
The zinc effect on laminin-entactin binding activity was saturable with optimal binding occurring at physiological Zn 2ϩ concentration (15 M) (Fig. 4). However, as the zinc concentration was raised above 15 M, the amount of nonspecific binding increased (Fig. 4B), suggesting the influence of Zn 2ϩ on specific binding activity was saturated at 15 M. Laminin (0.5 mg/ml) dialyzed against an excess of ZnCl 2 (50 M, 1 liter) followed by TBS to remove free metal, was found to contain 8.4 Ϯ 1.7 mol of Zn 2ϩ /mol of laminin (Fig. 4C). A small amount of Zn 2ϩ , 1.3 Ϯ 0.8 mol of Zn 2ϩ /mol of laminin, was also detected for laminin dialyzed against TBS only. Incubation at higher ZnCl 2 concentrations (Ͼ50 M) was avoided since it caused the, albeit reversible, precipitation of laminin.
Inspection of the laminin amino acid sequence revealed 12 cysteine-rich repeats which contain nested zinc finger consensus sequences not previously reported (Fig. 5). Furthermore, the entactin binding site was recently mapped to a zinc fingercontaining Cys-rich repeat on the laminin ␥1 chain (1).
Laminin Binding to Collagen Type IV-Laminin did not bind collagen type IV (Collaborative Research) in the absence of divalent metals (2 mM EDTA) (Fig. 6). The addition of CaCl 2 had little effect on the binding. However, with the addition of 15 M ZnCl 2 , a high affinity interaction was detected, involving a single class of binding sites, with an affinity of K d ϭ 5.4 nM, B max ϭ 39.0 ng. At higher NaCl concentrations, this binding was reduced, indicating that there was an ionic component to the interaction. Alkylation with N-ethylmaleimide, without reduction of the disulfide bonds, decreased the binding affinity (K d ϭ 11.7 nM) suggesting free sulfhydryls could be important for the binding site. When entactin was included, at a 2 M excess of laminin, the affinity of the laminin-collagen type IV interaction was increased but with a reduced B max (K d ϭ 2.4 nM, B max ϭ 15.0 ng). This binding characteristic most likely reflects the preferential formation of laminin-entactin-collagen type IV ternary complexes. The affinities are within the ranges previously reported for entactin-collagen type IV (K d ϭ 2-10) and for laminin-entactin-collagen type IV (K d ϭ 9 -20 nM) (18). Recombinant entactin bound collagen type IV and laminin with a higher affinity (K d Յ 1.0 nM) (19,34).
Laminin Polymerization Is Independent of Zn 2ϩ -The polymerization of laminin has been found to be dependent on calcium ions (35) and could not be substituted with Zn 2ϩ (Fig.  7). When laminin (0.3 mg/ml) was incubated in TBS ϩ 1 mM CaCl 2 , 40 -45% of the monomer polymerized. ZnCl 2 could not replace CaCl 2 , and its effect on polymerization was not significantly different from that observed with EDTA (Student's t-test, p Ͻ 0.05). Furthermore, ZnCl 2 did not interfere with CaCl 2 mediation of the process, since incubation with CaCl 2 , or with CaCl 2 ϩ ZnCl 2 , showed no difference in polymerization (Student's t-test, p Ͻ 0.05). This is consistent with the lack of putative Zn 2ϩ finger sequences in the domains involved in polymerization. DISCUSSION Basement membrane formation involves the secretion of a small set of high molecular weight proteins/proteoglycan which spontaneously interact to generate a supermolecular matrix. The identification and characterization of the specific binding sites directing this assembly is an area under active study. Its been proposed that the "backbone" for BM consists of two independent polymers of laminin and collagen type IV which become interconnected by either direct laminin-collagen type IV associations (20,21) or by the actions of a bridging protein, entactin (18).
The organization of BM is not uniform, and both developmental and tissue-specific heterogeneity in its structure is likely influenced by the expression of different laminin and collagen type IV isoforms (3,36). Biochemical factors, however, may also contribute to the final structure and function of BMs, and our data suggest Zn 2ϩ may be added to the short list of putative effectors governing BM organization. These include phospholipid, Ca 2ϩ , and heparin. The critical laminin concentration required for polymerization is lower on lipid bilayer surfaces such as plasma membranes (37) and may explain why BM form in close proximity to the cells which synthesize it. At physiological concentrations, both Ca 2ϩ (35) and heparin (28) promote laminin polymerization. Conversely, heparin is an inhibitor of collagen type IV polymerization (38). We have shown that both heparin and Ca 2ϩ block laminin-entactin binding, and, in light of their effects on the polymerization process, they may serve to reduce the cross-linking of the laminin-collagen type IV networks favoring the formation of laminin-rich BM.
The positive effect of zinc on laminin binding activity suggests that it could be a potential metal co-factor for BM assembly and organization. Preincubating entactin or collagen type IV with ZnCl 2 did not enhance laminin binding activity, indicating Zn 2ϩ was affecting laminin only. However, entactin has been shown previously to bind both zinc-and cobalt-loaded columns equally well before and after complete alkylation of the protein (34). This indicated that metal binding was not dependent on protein conformation and likely occurred via the His-Xaa-His sites, which are known to bind certain metals with high affinity (57). In addition, Reinhardt et al. (34) reported that treating entactin with 2 mM EDTA had little effect on its binding to either the laminin P1 pepsin proteolytic fragment or collagen type IV, consistent with our proposal of the role of Zn 2ϩ as a laminin-specific co-factor. They also found that Zn 2ϩ at 50 M inhibited binding of entactin which again is in agreement with our observation that zinc concentrations above physiological (15 M) interfered with laminin-entactin interactions Laminin-collagen type IV binding was enhanced by zinc generating a binding maxima which was significantly higher than for laminin ϩ entactin-collagen type IV, indicating that entactin redirected laminin binding to a smaller number of binding sites. The apparent increase in affinity for the laminin-collagen type IV interactions, when entactin was included in the assay, was likely caused by the formation of stable ternary complexes (19). The number of binding sites to which laminin was limited to was about one-third that seen with Zn 2ϩ alone possibly because of steric blocking by entactin of the collagen type IV binding sites on the laminin ␣1 and ␥1 chains. Aumailley et al. (18) also found that laminin-entactin complex bound collagen type IV with high affinity while little or no binding was detected with isolated laminin, as we observed in the absence of zinc.
Laminin binding sites have been mapped to sites approximately 80 nm, 200 nm, and 300 nm distal of the carboxylterminal globular domain of collagen type IV (21), while the entactin binding sites were located at the first two sites only (18). Our binding assay data suggest that at least 2 types of mutually exclusive interactions cross-linking the laminin and collagen type IV networks are possible; one that is solely Zn 2ϩmediated and the other of higher affinity involving Zn 2ϩ ϩ entactin as the bridging molecule. When we investigated the laminin sequence for potential Zn 2ϩ binding regions structural motifs rich in cysteine residues caught our interest. These motifs of unknown function are common for basement membrane components. Two types have been recognized, based primarily on the number and spacing of the cysteine residues within a 40-amino acid or 60-amino acid domain (Fig. 8). One type has 6 Cys with a spacing similar to the active domain of the pancreatic secretory trypsin inhibitors of the Kazal family (39). The other more common type, also with 6 Cys, is similar to the Cys-rich domains in epidermal growth factor (EGF) (46). This domain aligns best with the protein products from the Drosophila melanogaster neurogenic locus Notch (41) and the Lin-12 protein from Caenorhabditis elegans (47). Also included are other proteins which have fewer copies of the repeat such as the low density lipoprotein receptor (48), factor IX, factor X, and protein C (49,50), and plasminogen activator (51).
The inner rod-like regions of the laminin short arms also contain many Cys repeats which have mitogenic activity for a variety of cell types, that is independent of the EGF receptor (52,53). The nonapeptide GDPGYIGSR and the shorter pentapeptide YIGSR are both found in Cys-rich repeats in domain III of the ␥1 chain and have been shown to promote cell attachment and migration (54,55).
On closer examination of these Cys-rich repeats, both laminin and perlecan contain 60-amino acid repeats of 8 Cys exclusively that exhibit much lower similarity (particularly in Cys spacing) to the EGF-like repeat and may constitute a third type of repeat (Fig. 8). Within the 42 Cys-rich repeats found on laminin, 12 appear to harbor the consensus sequence for Cysrich zinc fingers, a proven zinc binding domain in many systems (Fig. 5). Based on our data and the aforementioned sequence information, we propose that as many as 12 zinc fingercontaining repeats on laminin could potentially be coordinated by Zn 2ϩ atoms. These may generate one high affinity entactin binding site and possibly one or more collagen type IV binding sites.
It is unknown if disulfide bonding takes place in all of the Cys repeats of laminin and perlecan, but a hypothetical disulfide bonding pattern involving all 8 Cys has been proposed by Appella et al. (46) and Engel (52). However, their proposal is based on the secondary structure for an EGF repeat from human pro-EGF (1-48) determined by NMR which agrees with a predicted disulfide bonding pattern between all 6 Cys (1-3, 2-4, and 5-6) (56). In laminin, only 5 of 8 Cys in the repeats align with those of the EGF domain (31,46). It is possible that in the absence of Zn 2ϩ free sulfhydryls could form disulfide bonds but our data suggests that the Cys-4, -5, -6, and -7 in at least 8 of 42 repeats may be coordinated by Zn 2ϩ generating a zinc finger-like motif. High resolution studies to determine the conformation of these Cys repeats in the absence or presence of Zn 2ϩ is needed.
Laminin isolated from EHS tumor was found to have about 1 mol of Zn 2ϩ /mol of laminin which, after incubation with 50 M ZnCl 2 , could be increased to about 8 mol/mol of laminin, an amount consistent with the predicted number of zinc finger sequences. In further support of this idea, the entactin binding site was recently mapped to a cysteine-rich repeat on the laminin ␥1 chain (1), which happens to contain a zinc finger-like sequence. Three out of twelve Cys-rich repeats on this chain have nested zinc finger consensus sequences. Mayer et al. (1) observed that iodination of the B2III-4 peptide (58-mer), which contained 2 Tyr residues in the nested zinc finger sequence, substantially reduced binding activity. Furthermore, reduction and alkylation of the Cys residues abolished its ability to inhibit laminin-entactin binding. We have confirmed that alkylation of Cys reduced laminin binding activity for both entactin and collagen type IV. Laminin polymerization was found to be Ca 2ϩ -dependent as reported elsewhere and could not be reproduced with Zn 2ϩ , nor did Zn 2ϩ interfere with the effect of Ca 2ϩ , of which 2-3 are required to bind to the terminal globular domain of ␥1 chain to facilitate maximal laminin polymerization (35). Most likely, the two ion species bind to different sites FIG. 7. Laminin polymerization proceeded independent of zinc. Polymerization was assayed based on the method in Ref. 17. Laminin at 0.3 mg/ml was incubated under the four different conditions shown at 37°C for 4 h, centrifuged for 15 min at 12,000 ϫ g, and the polymer fraction was calculated by subtracting the supernatant concentration from the total. Percent laminin polymerized was plotted, based on the mean and standard deviation of 3 experiments. Analysis of the data by Student's t test indicated that EDTA versus ZnCl 2 and CaCl 2 versus ZnCl 2 /CaCl 2 (p Ͻ 0.05) were not significantly different.  (43), agrin (44), laminin (31)(32)(33), and perlecan (45). Cysteine spacing similar to PSTI is found in SPARC and to EGF in BM-90, agrin, entactin, and also SPARC. C, cysteine; X, any residue. on laminin consistent with the lack of zinc finger-like sequences in the terminal globular domains required for polymerization. Hence, laminin zinc fingers may function mainly in the lateral associations interconnecting the laminin and collagen type IV networks. One or more may also be involved in the mitogenic activity of laminin directly, or indirectly, by influencing the accessibility of the RGD or YIGSR peptides found on 2 different Cys-rich domains between zinc finger-containing repeats.
These putative zinc finger motifs on laminin are highly conserved between human (58 -60), mouse (31)(32)(33), and Drosophila (61)(62)(63). For mouse and human, the number and relative location of the zinc finger sequences are identical. In Drosophila, the most distantly related species examined, 7 out of 9 of its zinc finger sequences, including the one at the entactin binding site, have maintained their relative locations in the laminin sequence, with an amino acid sequence identity of about 60% between Drosophila and mammals. Perlecan (mouse), a BM proteoglycan, also contains 6 Cys-rich repeats (45), of which 3 contain zinc finger-like sequences (Fig. 9). Overall sequence similarity between laminin ␣1 chain and perlecan is high, particularly at the amino end suggesting a common evolutionary origin (64,65). Perlecan has also been reported to bind entactin (66), but whether any of the zinc finger-containing repeats act as the entactin binding site remains to be established.