Báo cáo khoa học: Domain IV of mouse laminin b1 and b2 chains Structure, glycosaminoglycan modi®cation and immunochemical analysis of tissue contents

Eur. J. Biochem. 269, 431–442 (2002) (cid:211) FEBS 2002

Domain IV of mouse laminin b1 and b2 chains Structure, glycosaminoglycan modification and immunochemical analysis of tissue contents

Takako Sasaki1, Karlheinz Mann1, Jeffrey H. Miner2, Nicolai Miosge3 and Rupert Timpl1

1Max-Planck-Institut fu¨r Biochemie, Martinsried, Germany; 2Renal Division and Department of Cell Biology and Physiology, Washington University School of Medicine, St Louis, MO, USA; 3Center of Anatomy, Department of Histology, University of Go¨ttingen, Germany

showed no cross-reaction with each other and allowed establishment of b chain-specific radioimmunoassays and light and electron microscopic immunostaining of tissues. This demonstrated a 5–25-fold lower content of b2 com- pared with b1 chains in various tissue extracts of adult mice. Tissues derived from b2-deficient mice failed to react with the b2-specific antibodies but showed a twofold higher content of b1 than heterozygotes. The antibodies to b2 showed broader tissue staining than reported previously, including in particular a distinct reaction with the extra- synaptic endomysium of skeletal muscle. Immunogold staining localized both b chains primarily to basement membranes of kidney, muscle and various other tissues.

Keywords: basement membranes; chondroitin sulfate; laminin domain; radioimmunoassay; recombinant pro- duction.

Domain IV, consisting of about 230 residues, represents a particular protein module so far found only in laminin b1 and b2 chains. Both domains were obtained by recombi- nant production in mammalian cells. They showed a globular structure, as expected from electron microscopic examination of laminins. Fragment b1IV was obtained as a monomer and a disulfide-bonded dimer, and both were modified to (cid:25) 50% by a single chondroitin sulfate chain attached to Ser721 of an SGD consensus sequence. Dimerization is caused by an odd number of cysteines, thiol character. them having a partial with three of Whether both modifications also occur in tissue forms of laminin remains to be established. Fragment b2IV was only obtained as a monomer, as it lacked one crucial cysteine and the SGD sequence. It required, however, the presence of two adjacent LE modules for proper folding. raised against both fragments Polyclonal antibodies

[6–8]. These two b chains consist of about 1760 residues, have an identical modular structure, and show about 50% sequence identity [9,10]. Their mRNAs are expressed at different levels in a large number of tissues [10,11], produced by a variety of cultured cells, and, as shown by antibody staining, encode proteins found in various basement mem- branes [12–16]. Their distribution can change during embryonic development, particularly in aorta, kidney and skeletal muscle. The b2 chain was actually discovered in a search for proteins that are specific for neuromuscular synapses [9] and this restriction has been confirmed in subsequent studies of mouse [15] but not human tissues [17–19].

Laminins represent a family of heterotrimeric proteins (abc) which are mainly localized in basement membranes and involved in cell–matrix and various other protein interac- tions. Fifteen different isoforms are so far known, laminin-1 to laminin-15, based on the assembly of different a1 to a5, b1 to b3 and c1 to c3 chains [1–3]. These chains share a 600- residue domain II,I which oligomerizes into a rod-like coiled-coil structure forming the long arm of laminins. The N-terminal short arms and C-terminal G domains, however, are composed of laminin-type LE, L4, LN and LG modules, which form rod-like or globular structures in various combinations [1,4]. Most of these modules are also shared by several other extracellular proteins such as the proteoglycans perlecan and agrin. A few other domains are so far unique to laminins and include domain IV of the b1 and b2 chain. Only b1IV has so far been obtained as a proteolytic fragment [5].

Ten of the laminin isoforms contain either the b1 or b2 chain in combination with c1 and one of the five a chains

Little evidence is, however, available on potential func- tional differences between b1 and b2 chains. Laminins containing either variant were shown to promote in an equivalent manner cell adhesion [6] and neurite outgrowth [7]. These activities are attributed to the shared a chains and apparently not modulated by different b chains. The analysis of chimeric b chains provided evidence that a 16-residue sequence close to the C-terminus of domain I is directing the synaptic localization of b2-containing laminins [20]. A Leu-Arg-Glu sequence in a similar region of b2 chain was also claimed to provide a specific stop signal for neurons [21]. This sequence seems, however, not to be active when installed in a native coiled-coil conformation [22]. A particular role for the b2 chain was also emphasized

Correspondence to R. Timpl, Max-Planck-Institut fu¨ r Biochemie, Am Klopferspitz 18A, D-82152 Martinsried, Germany. Fax: + 49 89 8578 2422, Tel.: + 49 89 8578 2440, E-mail: timpl@biochem.mpg.de (Received 8 august 2001, revised 31 October 2001, accepted 7 November 2001)

432 T. Sasaki et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

by gene knock-out studies, which demonstrated in mice distinct deficiences in neuromuscular synapses and glomer- ular filtration [23,24].

In this study, we have used recombinant production of domain IV in mammalian cells to explore further structural and tissue differences of the b1 and b2 chain. This revealed glycosaminoglycan modification and disulfide-dependent dimerization of b1IV but not of b2IV, which could be readily explained by sequence differences. This also allowed development of sensitive and specific immunological assays for both b chains that are useful for quantitative analyses and examination of their distribution in tissues.

were separated by size-exclusion chromatography on a Superdex Peptide column (Amersham-Pharmacia; HR 10/ 30) in 0.1% trifluoroacetic acid/25% acetonitrile. Further cleavage at enzyme/substrate ratios of 1 : 100 with trypsin or endoproteinase Lys-C in 0.2 M NH4HCO3 or with pepsin in 0.1 M glycine/HCl, pH 1.9 (23 (cid:176)C, 15 h) was followed by Superdex chromatography and/or HPLC on a C18 column [31]. Differential alkylation with 4-vinylpyridine or iodoac- etate after partial and complete reduction with dithiothreitol in 0.05 M Tris/HCl, pH 8.5, and 6 M guanidinium hydro- chloride/0.05 M Tris/HCl, pH 8.5, respectively, followed a previously used procedure [32]. Digestion with chondroitin- ases ABC, AC or B (Sigma) followed a previous protocol [33].

M A T E R I A L S A N D M E T H O D S

Sources of proteins

Analytical methods

Laminin-1, in complex with nidogen-1 and its elastase fragment E10, were obtained from a mouse tumor basement membrane [5,25]. Other recombinant fragments of mouse laminins included a1IVa [26], b1VI//V and b3VI/V (unpub- lished) which were prepared by established procedures [27].

Construction of expression vectors and cell transfections

Hydrolysis with 6 M or 3 M HCl (16 h, 110 (cid:176)C) was used to determine protein and hexosamine contents on an LC3000 analyzer (Biotronic). Edman degradation was performed on Applied Biosystems sequencers 473 and 492 following the manufacturer’s instructions. The recovery of pyridylethylcy- steines and carboxymethylcysteines was estimated from the height of the respective phenylthiohydantoin derivatives in the corresponding sequencer runs. Electron microscopy of rotary shadowed proteins [34], CD, and matrix-assisted laser desorption ionization MS [35] followed standard protocols.

Immunological methods

The templates used were a complete mouse b1 cDNA (plasmid no. 1609) provided by Y. Yamada (NIDR, Bethseda, MD, USA), and for b2IV we used RNA from embryonic mouse endothelial cells provided by A. Hatzo- poulos (GSF, Munich, Germany). The sense and antisense primers for b1 were GTCAGCTAGCTAACGAGGTGG AGTCCGGTTAC and GTCACTCGAGCTAAAGGCC CGTCTGGTGAATCAAG, respectively, and for b2 GTC AGCTAGCCCGTCCCTGTGACTGTGATG and GTC ACTCGAGCTAGGCTTGACAGCCTGCAGGG, respectively. They were used for amplification by PCR and RT-PCR, respectively. These primers introduced at the 5¢ end an NheI site and at the 3¢ end a stop codon followed by an XhoI site to allow insertion into the episomal expression vector pCEP-Pu containing the BM-40 signal peptide [28]. A Ser721Ala mutation was introduced into b1IV by fusion PCR [29].These vectors were used to transfect 293-EBNA cells, and serum-free medium was collected from these cells [28].

Generation of rabbit antisera, affinity purification of antibodies, ELISA titration, and radioimmunoassays fol- lowed established protocols [36]. A radioimmuno-inhibition assay specific for recombinant mouse laminin fragment c1III3-5 using antiserum against laminin fragment P1 has been described [37]. Immunoblotting using ECL Western- blotting reagents followed a previous procedure [38]. Mouse Engelbreth-Holm-Swarm (EHS) tumor and tissues from adult mice and 18-day-old pups heterozygous for (b2 +/–) or lacking (b2 –/–) the laminin b2 chain gene [23] were extracted with EDTA-containing buffers and then deter- gents; this was followed by digestion with bacterial colla- genase [39]. These extracts were then used at dilutions of 1 : 10 and higher for radioimmuno-inhibition assays.

Purification of recombinant proteins

Immunofluorescence staining of frozen tissue sections folllowed previously used procedures [8,15]. Gold particles (16 nm) were used to label affinity-purified rabbit antibodies as described [40]. Tissue sections on nickel grids were incubated for 15 min with NaCl/Pi, pH 7.2, and then labeled gold diluted 1 : 200 for 16 h at room temperature. Sections were rinsed with water, stained with uranyl acetate (15 min) and lead citrate (5 min), and examined with a Zeiss EM 109 electron microscope. Controls included colloidal gold alone or samples coated with goat anti-(rabbit Ig) IgG or anti-(rat Ig) IgG and were all negative.

R E S U L T S

Conditioned medium was passed over a DEAE-cellulose column equilibrated in 0.05 M Tris/HCl, pH 8.6, and eluted with a linear 0–0.6 M NaCl gradient. The effluent was monitored at 280 nm and by a carbazole assay [30]. Fragment b1IV was eluted at 0.2–0.3 M and 0.3–0.4 M NaCl, whereas most of fragment b2IV did not bind to the column. They were both purified on a Superose 12 column (HR 16/50) equilibrated in 0.2 M ammonium acetate, pH 6.8. Fragment b2IV was then bound to a Mono Q column in 0.02 M Tris/HCl, pH 8.0, and eluted with 0.03– 0.05 M NaCl.

Expression and purification of recombinant domain IV from mouse laminin b1 and b2 chain

Characterization of protein structures

Domain IV of the b1 (position 541–771) and b2 (position 556–782) chains correspond to the central globular domain

Cleavage with CNBr was performed in 70% formic acid (24 h at 23 (cid:176)C) in the dark under nitrogen. The fragments

(cid:211) FEBS 2002

Laminin b1 and b2 chains (Eur. J. Biochem. 269) 433

Fig. 1. Sequence alignment of domain IV (A) from mouse laminin b1 (top) and b2 (bottom) chains and domain structure of the laminin b1 and b2 chains (B). (A) Both sequences show 41% identity (bars) and 29% conservative changes (dots). Cysteines are numbered 1–5 in b1 and 1–4 in b2, and carbohydrate attachment motifs (SGD, NYT) are high- lighted in bold. Asterisks mark corrections to the published genomic sequence of mouse b2 [43]. An arrowhead indicates the start of frag- ment E10 [5]. The numbering includes the signal peptides. (B) Laminin b1 and b2 chains have the same domain structures consisting of LN and LE (circles) modules, domain IV, and a coiled-coil (cc) domain II,I.

in the short arm of these laminin chains [2] (Fig. 1B). They are of similar size and share 41% identical residues including four out of five cysteines and 29% conservative replace- ments (Fig. 1A). Features unique to b1IV are single potential acceptor sites for N-glycosylation and glycosami- noglycan attachment, respectively. Episomal expression vectors were prepared for both domains in order to obtain them as recombinant fragments from serum-free culture medium of transfected human kidney 293-EBNA cells. Production and secretion of fragment b1IV occurred at high rates (150–170 lgÆmL)1Æday)1) as shown by radioimmuno- assay (see below), whereas no protein could be detected after transfection with the b2IV expression vector corresponding to the sequence shown in Fig. 1A. This indicated a protein- folding problem, which was overcome by adding two laminin-type epidermal growth factor-like (LE) modules to the N-treminus and C-terminus of b2IV (LE5, position 523– 555; LE6, position 783–831, see Fig. 1B), allowing produc- tion of this longer fragment at a moderate level (5 lgÆmL)1Æ day)1). This fragment will be referred to as b2IV.

The initial purification of fragment b1IV on DEAE- cellulose showed it to be present in about equal proportions

in two pools eluted at 0.2–0.3 M NaCl (pool 1) and 0.3– 0.4 M NaCl (pool 2). Pool 1 consisted mainly of a 34-kDa band, as expected for a b1IV monomer, and about equal amounts of a disulfide-bonded b1IV dimer (66 kDa), which could be separated by molecular sieve chromatography as shown by electrophoresis (Fig. 2, lanes 1 and 3). Pool 2 coincided with a strong carbazole reaction for uronic acids,

Fig. 2. SDS/PAGE of purified recombinant fragments containing domain IV of b1 or b2 chain under nonreducing (A) and reducing (B) conditions. Lanes were loaded with equal amounts of b1IV monomer (1), b2IV (2), b1IV dimer (3), b1IV substituted with chondroitin sulfate before (4) and after (5) digestion with chondroitinase ABC, and b1IV mutant S721A monomer form (6). Positions of calibrating proteins are shown in kDa on the left.

434 T. Sasaki et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

Glu-C (data not shown). CD spectra of the b1IV monomer showed a strong negative ellipticity at 209–220 nm, indicat- ing about 30% a helix and only a small amount of b structure (Fig. 4). The b2IV structure had only half the a-helical content, which may have become obscured through the presence of two additional LE modules, which, as shown previously, have only a small amount of secondary structure [41].

indicating the presence of a proteoglycan. The correspond- ing b1IV fragment was eluted in front of the b1IV dimer from a molecular sieve and showed a broad electrophoretic band mainly in the range 70–110 kDa, which could be converted into the monomer and dimer bands after treatment with chondroitinase ABC (Fig. 2, lanes 4 and 5). Recombinant fragment b2IV did not bind to DEAE- cellulose and showed, after molecular sieve chromatogra- phy, a single band of 33 kDa (Fig. 2, lanes 2), indicating its monomeric nature and lack of glycosaminoglycan modifi- cation. Edman degradation of the purified recombinant fragments showed a major single N-terminal sequence, which was APLANEVESG for the b1 variants and APLARPXD for b2IV, as expected from the vector constructs. The first four residues, respectively, were derived from the foreign signal peptides introduced to allow secretion of the proteins.

Structural characterization of the fragments

Hexosamine analysis of b1IV monomer and dimer showed 5–6 residues of glucosamine and less than one galactosamine. That Edman degradation of various prote- failed to identify Asn677, olytic peptides (see below) indicating full occupation of the single N-linked acceptor site. The proteoglycan form of b1IV showed the same glucosamine content, but a large increase in galactosamine content (48 residues), which corresponds to about 19 kDa of a glycosaminoglycan chain. Furthermore, digestion with chondroitinases ABC (Fig. 2) and A,C but not by chon- droitinase B or heparitinase (not shown) yielded the b1IV monomer and dimer bands, demonstrating the exclusive substitution by chondroitin sulfate. The mean (SD) length of the side chain was estimated from electron micrographs of 30 particles to be 40 ((cid:139) 10 nm), which is in good agreement with a molecular mass of 20 kDa. Mutation of Ser721 to Ala did not interfere with recombinant produc- tion of b1IV monomers (Fig. 2, lane 6) and dimers but prevented completely the modification by glycosaminogly- cans, consistent with the absence of any other strong acceptor site within the b1IV sequence (Fig. 1). Fragment b2IV lacked any of these substantial post-translational modifications, as determined by MS, which yielded a molecular mass of 34 369 Da, in good agreement with a mass of 34 357 calculated from the sequence.

The proper folding of the recombinant fragments was examined by several methods. Rotary shadowing followed by electron microscopy showed small compact globular shapes for b1IV monomers and dimers (not shown) and b2IV (Fig. 3B). The same globule was also observed for the proteoglycan form of b1IV, but with some of the particles being connected to a long faint thread-like structure (Fig. 3A). As these threads were not observed for b1 monomers and dimers, they probably represent the glyco- saminoglycan side chain. Proteolysis studies of b1IV monomer showed no significant change after trypsin treatment (1–24 h), while elastase caused degradation to a 22-kDa fragment similar to the E10 fragment obtained from elastase digests of laminin-1 [5]. Fragment b2IV, however, was readily cleaved within 1–4 h by trypsin into fragments smaller than 20 kDa but was resistant to endoproteinase

The odd number of cysteines in b1IV indicated the presence of a free thiol group, which is probably responsible for dimerization. The native b1IV dimer was therefore first modified by partial reduction and pyridylethylation to protect the thiol followed by complete reduction under denaturing conditions and carboxymethylation. Peptides were then generated by cleavage with endoproteinase Lys-C followed by cleavage with some other proteases and determination of the sequence around the five cysteines

Fig. 3. Rotary shadowing electron microscopy of chondroitin sulfate- substituted fragment b1IV (A) and fragment b2IV (B). Bar: 100 nm. Fig. 4. CD spectra of recombinant domain IV from laminin b1 (full line) and b2 (dashed line) chain.

(cid:211) FEBS 2002

Laminin b1 and b2 chains (Eur. J. Biochem. 269) 435

(C1 to C5, Fig. 1A) by Edman degradation. The data obtained showed complete carboxymethylation for C1 and C2, while for C3, C4 and C5 there was also 20–25% pyridylethylation. Dimeric b1IV was also cleaved with CNBr giving rise to three single-chain peptides starting at the N-terminus and positions 609 and 677, respectively. This indicated C1–C2 connectivity, which was confirmed by Edman degradation of smaller pepsin fragments. The fourth CNBr peptide showed two sequences starting as expected in front of C3 and C5 and should also contain C4 (Fig. 1A). A smaller pepsin fragment showed a connection between C3 and C5. Yet because of the variable thiol content of C3 to C5 (see above), we cannot exclude the possibility that the disulfide connectivity is flexible, and we suggest that partial formation of C3–C4 and C4–C5 bridges occur as well.

Immunochemical assays for domains b1IV and b2IV and their application to the analysis of tissues

Many previous studies based mainly on immunohistology have shown a complex and in part overlapping expression pattern of laminin b1 and b2 chains in adult tissues, during embryonic development and in cultured cells (reviewed in [2,16]). Several of the monoclonal antibodies against b2 chain were generated in mice and failed to react with the corresponding mouse antigens [9]. To circumvent these limitations and to allow quantitative assays, we generated rabbit antisera against recombinant mouse b1IV and b2IV. These antisera had high titers (half-maximal binding) at dilutions of 1 : 4000 (anti-b1) and 1 : 20 000 (anti-b2) in ELISA and radioimmunoassays and did not cross-react (titer less than 1 : 100) with the homologous (b2 or b1IV) antigen or recombinant N-terminal fragments b1VI/V and b3VI/V of mouse laminin. They were also clearly distin- guishable in immunoblots of several biological samples, with anti-b1 reacting mainly with a 220-kDa band and anti- b2 with a 190-kDa band after electrophoresis under reducing conditions (data not shown).

The antisera allowed the development of specific and sensitive radioimmuno-inhibition assays, with half-maximal inhibition being achieved at 0.1–0.2 nM b1IV and b2IV, respectively (Fig. 5). The b1IV assay was inhibited in nearly identical manner by monomeric and dimeric b1IV and the proteoglycan form and by laminin-1 (a1b1c1) derived from the mouse EHS tumor (Fig. 5A). Laminin-1 fragment E10 showed a less steep and incomplete inhibition profile, indicating the loss of some antigenic epitopes. A more than 1000-fold excess of b2IV showed no inhibition (Fig. 5A), and the same was found for recombinant laminin fragments b1VI/V, b3VI/V and a1IVa (data not shown). Similarly, the assay for b2IV could not be inhibited by a large excess of b1IV (Fig. 5B). Both assays could, however, be inhibited by tissue extracts, as shown for mouse heart and kidney (Fig. 5), with inhibition curves similar in slope to that of the reference inhibitors b1IV and b2IV, respectively. Together the data showed that these assays were highly specific and suitable for the quantitation of b1 and b2 chain epitopes in biological samples.

used this tumor and also mouse kidney and muscle for successive extraction with EDTA and detergents followed by digestion with chondroitinase/heparitinase and bacterial collagenase, which solubilized nearly all of the tissue. The extracts were analysed by the b1IV and b2IV assay and also by an assay specific for the laminin c1 chain [37]. This demonstrated that 71–93% of the extractable laminin c1, b1 and b2 chain were already solubilized by EDTA and detergent. Furthermore, the content of extractable lami- nin-1 was calculated based on the c1 and b1 chain assay to be 5.5–5.8 mgÆg)1 EHS tumor, in agreement with previous data. A larger number of adult mouse tissues were then extracted with EDTA and detergent only and examined by these three assays (Table 1). A variable content of laminin b1 chain [17–422 pmolÆ(g wet tissue))1] and lower amounts of b2 chain (4–20% of b1) were found. As expected, the highest amounts of b1 chain were found in the EHS tumor, while the content of b2 chain did not exceed 0.2%. The amounts of c1 chain were within the usual range of

The extracellular matrix of mouse EHS tumor has previously been shown to consist exclusively of basement membrane, and its major component, laminin-1, could be solubilized in decent amounts (5 mgÆg)1 wet tissue) by extraction with EDTA and 6 M guanidine [25]. We now

Fig. 5. Radioimmuno-inhibition assays specific for laminin b1 (A) and b2 (B) chains. The assay consisted of 1 ng 125I-labeled fragment b1IV monomer or b2IV and fixed amounts of the corresponding antiserum. Inhibitors used were b1IV monomer (n), chondroitin sulfate-substi- tuted b1IV (,), b2IV (m), mouse laminin-1 (s) and its E10 fragment (h) at the concentrations shown at the bottom. EDTA extracts of mouse heart (d) and kidney (j) were used at the dilutions shown at the top.

436 T. Sasaki et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

Table 1. Contents [pmolÆ(g wet tissue))1] of laminin c1, b1 and b2 chains in tissue extracts from normal adult and mutant (b2 +/–; b2 –/–) mice as determined by radioimmuno-inhibition assays. Tissues were extracted with EDTA-containing bu(cid:128)er and detergent and analyzed by assays specific for recombinant fragments c1III3-5, b1IV and b2IV.

One reason for the discrepancies between published reports showing a restricted distribution of b2 and our results showing a more widespread distribution could be that the antiserum used here cross-reacts with another laminin chain in immunohistochemical assays. To investi- gate this possibility, we immunostained tissues from Lamb2 mutant mice (Fig. 7). These mice have a mutation that has been shown to prevent any accumulation of laminin b2 in basement membranes [23,24]. No significant fluorescence was found on staining mutant tissues with the anti-b2IV serum, whereas tissues from a littermate control were well stained. This demonstrates that the antiserum reacts with laminin b2 but with no other laminin chain and no other basement membrane component.

Tissue c1 b1 b2

Results from immunohistochemical assays using anti-b1 serum (data not shown) were mostly consistent with previously published reports showing a widespread (although not ubiquitous) expression pattern for laminin b1. For example, in skeletal muscle, b1 was detected in the extrasynaptic muscle fiber basement membranes and in endoneurial basement membranes. It was not observed at synapses or in the perineurium, sites where b2 is known to be concentrated [12]. In kidney, b1 was detected in all tubular basement membranes and in the glomerular mesangium. In addition, it was detected at a low level in the glomerular basement membrane, a site where b2 is concentrated (Fig. 6). Together with our b2 immunohisto- chemical results, these data suggest that many basement membranes contain both laminin b1 and b2 chains. How- ever, in most cases one appears to be much more prevalent than the other.

analytical error ((cid:139) 20%) of such inhibition assays [36], in most cases in good agreement with the sum of b1 and b2 chains. It served therefore as an internal control for the quality of the data. Tissue extracts of mice being deficient in the laminin b2 chain [23,24] were used as further controls and failed to inhibit the b2 assay (content less than 0.5 pmolÆg)1). Interestingly, these extracts showed a twofold increase in the b1 chain content when compared with heterozygous (b2 +/–) controls (Table 1), indicating a significant increase in biosynthesis or mRNA stability.

Immunolocalization in tissues

Affinity-purified antibodies specific for b1IV or b2IV were also used for ultrastructural localization of both b chains in adult mouse tissues by immunogold staining (Figs 8 and 9). This demonstrated for tubular and glomer- ular basement membranes of kidney a distinct reaction with both antibodies of about equal intensity (Fig. 8A–D) and of basement membranes of collecting ducts, blood vessels and Bowman’s capsule. In the heart, some weaker staining was observed in basement membranes adjacent to cardiomyo- cytes and endothelial cells (Fig. 8E,F) and also in basement membranes of the endocardium and pericardium. Staining of skeletal muscle identified both b chains in endothelial and muscle cell basement membranes, whereas the adjacent interstitial region did not react (Fig. 9A,B). A distinctly stronger reaction for b2 was observed in the synaptic clefts of neuromuscular junctions compared with extrasynaptic regions (Fig. 9C,D) and at myotendineous junctions. Other basement membranes containing both b chains included the dermal–epidermal junction and testis, whereas the dermal connective tissue could not be stained.

The antiserum against recombinant mouse b2IV was tested in immunohistochemical assays on a panel of mouse tissues using indirect immunofluorescence (Fig. 6). The antiserum revealed a more widespread distribution of laminin b2 than has previously been reported. In the kidney, it was previously shown to be confined to glomerular and vascular smooth muscle basement membranes [9,12]. Here, it was detected in these basement membranes, but also in most intertubular capillaries and segmentally in some tubular basement membranes (Fig. 6A). In skeletal muscle fibers, it was concentrated at synapses but was also found extrasy- naptically. In peripheral nerve, it was found in the perineurium, as previously reported [9,12], but also in the endoneurium. Blood vessels throughout skeletal muscle also contained the b2 chain (Fig. 6B).

EHS tumor Placenta Heart Skeletal muscle Intestine Stomach Thymus Skin Lung Lung, b2 +/– Lung, b2 –/– Kidney Kidney, b2 +/– Kidney, b2 –/– 6935 510 567 133 97 143 44 16 225 169 313 84 101 205 6647 422 343 117 104 125 66 17 176 153 348 106 140 244 14 14 35 23 4 20 6 3 12 11 < 0.5 8 9 < 0.4

D I S C U S S I O N

Domain IV of laminin b1 and b2 represents a globular structure in the central portion of the short arm region [1,2] and has not yet been characterized at the structural and functional level. We have now obtained these domains as recombinant products from mammalian cells and demon- strated their proper folding by electron microscopy, CD, protease resistance, and immunochemical analysis. The data also show that domain b1IV represents an autono- mous folding unit and was produced at high rate. This was

In the small intestine, b2 was detected in smooth muscle and blood vessels and at a low level in the epithelial basement membrane of the crypts (Fig. 6C). In lung, it was present in basement membranes associated with bronchial epithelium, bronchial smooth muscle, and alveoli through- out the parenchyma (Fig. 6D). In the retina, b2 was detected in the inner limiting membrane, Bruch’s mem- brane, and in capillary basement membranes, but was not detected in the interphotoreceptor matrix or in other layers of the retina (Fig. E). In heart, b2 was found in cardio- myocyte and vascular basement membranes throughout (Fig. 6F).

(cid:211) FEBS 2002

Laminin b1 and b2 chains (Eur. J. Biochem. 269) 437

surprisingly not the case for b2IV which needed two additional LE modules for reasonable expression and secretion. Similar observations have been reported for the N-terminal globular LN module (domain VI) of the laminin a1 chain, which could only be produced in mammalian cells after it had been joined to a tandem of

four LE modules. The latter, however, folded properly on its own [27]. Another interesting observation is the relatively high content of a helix in b1IV and probably b2IV. A high helix content has so far only been demon- strated for the coiled-coil domains I and II of laminins [1,2], whereas other domains, such as those composed of

Fig. 6. Immunohistochemical localization of laminin b2 chain in basement membranes of adult mouse tissues by antibodies against b2IV. (A) In kidney, b2 is detected in basement membranes of glomerulus (g), large blood vessels (v), and most intertubular capillaries (arrows). Some segments of tubular basement membranes are also stained. (B) In skeletal muscle, b2 is detected in basement membranes all around the muscle fiber (f) as well as the neuromuscular junction (arrowhead). In peripheral nerve (n), b2 is detected in both perineurial and endoneurial basement membranes. (C) In small intestine, b2 is detected primarily in basement membranes of smooth muscle (sm), large blood vessels (v), and capillaries (arrows). Low levels of b2 are detectable in basement membranes associated with crypts (c). (D) In lung, b2 is detected in basement membranes of the bronchial epithelium (e), bronchial smooth muscle (sm), and alveolar epithelium throughout the parenchyma (p). (E) In retina, b2 is detected in the inner limiting membrane (ilm), Bruch’s membrane (bm), and in capillaries (arrows). (F) In 3-week-old heart, b2 is detected in basement membranes surrounding all cardiomyocytes, large vessels (v), and capillaries (arrows). Bar, 100 lm for A–C, F; 50 lm for D; 75 lm for E.

438 T. Sasaki et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

LE or LG modules, essentially lack helical segments, as shown by crystallographic analyses [41,42]. This suggests novel folding for domain IV, which can now be investi- gated by X-ray analysis of recombinant b1IV monomers and dimers.

b2IV of rat, mouse and human [9,43] to SGGD and preceded by a four-residue gap (Fig. 1), which may explain the failure to detect any glycosaminoglycan substitution in recombinant b2IV. Yet an SGG site in perlecan domain V [44] and three SGD sites in domain I [33,45] can serve as acceptor sites for heparan sulfate/chondroitin sulfate, and more remote sequences may regulate their complete or partial occupation [45]. Special features of such regulations may now be unravelled by site-directed mutagenesis of the corresponding regions in b1IV and b2IV domains. Our data also indicate that tissue laminins containing b1 chains may

Additional differences between domain IV of b1 and b2 chains are related to single amino-acid substitutions. A single serine within an SGDG consensus sequence was shown by site-directed mutagenesis (S721A) to be partially substituted by a 20-kDa chondroitin sulfate chain in both the b1IV monomers and dimers. This sequence is changed in

Fig. 7. Antiserum to laminin b2IV does not react with basement membranes in Lamb2 mutant tissues. Kidney (A and B) and skeletal muscle (C and D) from 3-week-old Lamb2 +/+ and –/– littermates were stained with anti-b2 serum. In the control, basement membranes throughout the kidney and skeletal muscle were stained. Basement membranes at neuromuscular junctions (arrows in C) were more immunoreactive than were extrasynaptic basement membranes. No immunoreactivity was detected in mutant tissues, demonstrating the specificity of the antiserum. Neuromuscular junctions were identified by double labeling with rhodamine-a-bungarotoxin (C¢ and D¢). Bar, 100 lm.

(cid:211) FEBS 2002

Laminin b1 and b2 chains (Eur. J. Biochem. 269) 439

comparable low rate of modification of two tissue forms of perlecan [44].

An extra cysteine (C3 at position 710) in domain b1IV not conserved in b2IV was a candidate to explain the substantial dimerization of recombinant b1IV. We there- fore used stepwise reduction under nondenaturing and denaturing conditions as previously used to identify a single cysteine responsible for fibulin-2 dimerization [32]. This provided clear evidence for a C1–C2 connectivity within b1IV but failed to identify unequivocally C3, C4 or C5 as being responsible for dimerization. This could indicate the existence of various disulfide isomers, but we

be, at least in part, converted into proteoglycans. This is underscored by conservation of the SGD sequence in the human laminin b1 chain [46], whereas the laminin b chains of Drosophila [47] and Caenrhabditis elegans (accession no. AAB 94193) lack this sequence and also differ in cysteine patterns. Preliminary unpublished studies to identify such forms in tissue extracts have, however, failed so far and indicated only a low level of substitution or a specific restriction to a few tissue sites. Yet we do not think that the partial modification of b1IV is an artefact of recombinant production because a similar low rate of substitution of recombinant perlecan domain V correlated well with a

Fig. 8. Immunogold staining of laminin b1 (A,C,E) and b2 (B,D,F) chain in basement membranes of kidney and heart of adult mice. Both antibodies reacted equally well with the basement membrane of a proximal tubule (A,B; arrows) and with the glomerular basement membrane (C,D; asterisks). In the heart (E,F) both chains could be detected in the basement membranes of cardiomyocytes (arrows) as well as of endothelial cells of an adjacent capillary (asterisks). The capillary lumen (l) and the interstitial tissue between both basement membranes showed insignificant staining. Bars, 0.25 lm.

440 T. Sasaki et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

into networks, which occurs

cannot exclude the possibility that they represent artefacts of the treatments used. Nevertheless, we also cannot dismiss the possibility of the disulfide-dependent dimeriza- tion of laminins in which domain b1IV participates. Such complexes between laminin-5 (a3b3c2) and laminin-6 (a3b1c1) are known to exist in certain basement mem- branes [48] or may be generated upon self-assembly of laminins through their N-terminal regions including the b1 chain [49].

particularly high in skeletal muscle, consistent, as shown here and previously, for developing and adult human muscle [17], by distinct staining of extrasynaptic areas of muscle basement membranes. The radioimmunoassay data are also inconsistent with restriction to synapses, as shown in rats by monoclonal antibodies [9,15], considering the low density of synaptic junctions (usually accounting for less than 0.1% of the total basement membranes). The radio- immunoassays also indicated a twofold higher content of b1 chain in lung and kidney of b2-deficient mice compared with littermate controls. Such upregulation was previously predicted from immunostaining of kidneys, but it did not lead to functional compensation [24].

This is the first extensive examination of b2 distribu- tion in mouse tissues, most other studies having utilized rat, rabbit and human tissues. Thus, there could be differences in b2 distribution among species. However, restricted distribution of b2 in mouse kidney and skeletal muscle has been reported [15,23,24], and we have observed widespread expression of b2 in rat muscle using anti-b2IV serum (data not shown). One possible explanation is that many laminin b2 monoclonal anti- bodies that have been produced may recognize a specific form of b2 found primarily in glomerular, synaptic, perineural, and smooth muscle basement membranes. This could result from selective glycosylation, conforma- tional alteration, or association with a specific a chain. laminin a5 is also found in these basement Indeed,

Laminin b1 and b2 chains have been localized by immunohistology at the light microscopy level to a large variety of tissues, but not necessarily to the same subana- tomical regions, which has generated several controversial observations (reviewed in [16]). We have now prepared potent polyvalent antisera against the natively folded fragments b1IV and b2IV in order to re-examine some of the previous data and to establish more quantitative and sensitive assays. The rabbit antisera obtained showed a high specificity for b1 and b2 chains, respectively, as demon- strated by ELISA, radioimmunoassay, and immunoblot. This was underscored by the failure to show any reactivity of antibodies to b2 with tissues derived from b2-deficient mice. Specific radioimmuno-inhibition assays of high sen- sitivity (IC50 (cid:136) 0.1–0.2 nM) were developed and showed both b chains and equivalent levels of laminin c1 chains in EDTA/detergent extracts of various mouse tissues. The content of b1 chains in these extracts exceeded the content of b2 by a factor of 5–20. The relative amounts of b2 were

Fig. 9. Immunogold staining of laminin b1 (A) and b2 (B–D) in extrasynaptic and synaptic regions of skeletal muscle from adult mice. Both antibodies (A,B) stained the basement membrane around myocytes (arrows) and the endothelium of a capillary (asterisks). The arrowhead indicates some staining close to a pericyte. l, Capillary lumen. (C,D) shows some stronger staining for b2 of the synaptic cleft (asterisks) than for the extrasynaptic part (arrowheads). Bars, 0.25 lm.

(cid:211) FEBS 2002

Laminin b1 and b2 chains (Eur. J. Biochem. 269) 441

expression in adult and developing retinae: evidence of two novel CNS laminins. J. Neurosci. 20, 6517–6528.

membranes and is thought to be associated with b2 in the laminin-11 trimer. Alternatively, these antibodies may only recognize the highest concentrations of b2. Consis- tent with this, anti-b2IV serum stains basement mem- branes recognized by the other antibodies more intensely than it stains other basement membranes.

4. Engel, J. & Bozik, D. (1999) Modules in ECM and adhesion molecules. In Guidebook to the Extracellular Matrix, Anchor and Adhesion Proteins (Kreis T. & Vale R., eds), 2nd edn, pp. 344–351. Oxford University Press, Oxford.

Our results agree with those of Wewer and colleagues [17], who found extrasynaptic deposition of b2 in human skeletal muscle basement membranes using several mon- oclonal and polyclonal antibodies. Furthermore, in embryonic rat, Durbeej et al. [50] identified a subset of b2 protein that was detectable on immunoblots but not in immunohistochemical assays, suggesting that a subset of b2 in tissue cannot be recognized by some antibodies known to react with b2. However, in retina, b2 has been shown to have a more widespread distribution than we observed here, as it was also found in the interphotore- ceptor matrix and in the outer plexiform layer [3,51]. As these layers do not contain basement membranes, it is possible that our anti-b2IV serum did not recognize b2 in these layers because of a specific supramolecular organi- zation which may mask most of the antigenic epitopes of domain b2IV.

5. Mann, K., Deutzmann, R. & Timpl, R. (1988) Characterization of proteolytic fragments of the laminin-nidogen complex and their a(cid:129)nity in ligand binding assays. Eur. J. Biochem. 178, 71–80. 6. Brown, J.C., Wiedemann, H. & Timpl, R. (1994) Protein binding and cell adhesion properties of two laminin isoforms (AmB1eB2e, AmB1sB2e) from human placenta. J. Cell Sci. 107, 329–338. 7. Brandenberger, R. & Chiquet, M. (1995) Distinct heparin-binding and neurite-promoting properties of laminin isoforms isolated from chick heart. J. Cell Sci. 108, 3099–3108.

8. Miner, J.H., Patton, B.L., Lentz, S.I., Gilbert, D.J., Snider, W.D., Jenkins, N.A., Copeland, N.G. & Sanes, J.R. (1997) The laminin a chains: expression, developmental transitions, and chromosomal locations of a1–5, identification of heterotrimeric laminins 8–11 and cloning of a novel a3 isoforms. J. Cell Biol. 137, 685–701. 9. Hunter, D.D., Shah, V., Merlie, J.P. & Sanes, J.R. (1989) A laminin-like adhesive protein concentrated in the synaptic cleft of the neuromuscular junction. Nature (London) 338, 229–234. 10. Iivanainen, A., Vuolteenaho, R., Saino, K., Eddy, R., Shows, T.B., Sariola, H. & Tryggvason, K. (1994) The human laminin b2 chain (s-laminin): structure, expression in fetal tissues and chromosomal assignment of the LAMB2 gene. Matrix Biol. 14, 489–497.

11. Wewer, U.M., Gerecke, D.R., Durkin, M.E., Kurtz, K.S., Mattei, M.-G., Champliaud, M.F., Burgeson, R.E. & Albrechtsen, R. (1994) Human b2 chain of laminin (formerly s chain): cDNA cloning, chromosomal localization, and expression in carcinomas. Genomics 24, 243–252.

12. Sanes, J.R., Engvall, E., Butkowski, R. & Hunter, D.D. (1990) Molecular heterogeneity of basal laminae: isoforms of laminin and collagen IV at the neuromuscular junction and elsewhere. J. Cell Biol. 111, 1685–1699.

13. Wewer, U.M., Engvall, E., Paulsson, M., Yamada, Y. & Albr- echtsen, R. (1992) Laminin A, B1, B2, S and M subunits in the postnatal rat liver development and after partial hepatectomy. Lab. Invest. 66, 378–389.

We have also extended the tissue localization of laminin b1 and b2 chains to the ultrastructural level by using immunogold staining, which has not been examined in previous studies. This was particularly important for skeletal muscle, which showed immunogold staining for the laminin a2 chain in muscle and capillary basement membranes, whereas most of the laminin a4 chain was localized to adjacent interstitial regions of the endomysium [52]. A similar restriction to basement membranes of muscle and capillaries was here found for the b1 and b2 chain. Comparable staining for both b chains was also found for tubular, glomerular, and some other renal basement mem- branes. An interesting finding was that staining for b2 was distinctly more intense at the junctional folds of synapses than at the extrasynaptic basement membranes. This indicates a higher epitope density or accessibility and may explain why certain monoclonal antibodies against b2 show restricted synaptic staining [9,15]. A higher synapse-specific accumulation of b2 chain-containing laminins based on a particular recognition sequence [20] would also be compat- ible with a neuromuscular phenotype found in b2-deficient mice [23].

14. Glukhova, M., Koteliansky, V., Fondacci, C., Marotte, F. & Rappaport, L. (1993) Laminin variants and integrin laminin receptors in developing and adult human smooth muscle. Dev. Biol. 157, 437–447.

15. Patton, B.L., Miner, J.H., Chiu, A.Y. & Sanes, J.R. (1997) Dis- tribution and function of laminins in the neuromuscular system of developing, addult and mutant mice. J. Cell Biol. 139, 1507–1521. 16. Gullberg, D., Tiger, C.F. & Velling, T. (1999) Laminins during muscle development and in muscular dystrophies. Cell. Mol. Life Sci. 56, 442–460.

A C K N O W L E D G E M E N T S

17. Wewer, U.M., Thornell, L.E., Loechel, F., Zhang, X., Durkin, M.E. & Amano, S. (1997) Extrasynaptic location of laminin b2 chain in developing and adult human muscle. Am. J. Pathol. 151, 621–631.

We are grateful for the excellent technical assistance of Hanna Wiedemann, Mischa Reiter, Vera van Delden, Christa Wendt, Cong Li, and Albert Ries, and thank Y. Yamada and A. Hatzopoulos for providing reagents. The study was supported by grants from DFG (Ti 95/8-1) and the EC project QLK3-CT2000-00084. 18. Cohn, R.D., Herrmann, R., Wewer, U.M. & Voit, T. (1997) Changes of laminin b2 chain expression in congenital muscular dystrophy. Neuromuscul. Disord. 7, 373–378.

R E F E R E N C E S

19. Cohn, R.D., Herrmann, R., Sorokin, L., Wewer, U.M. & Voit, T. (1998) Laminin a2 chain-deficient congenital muscular dystrophy: variable epitope expression in severe and mild cases. Neurology 51, 94–100.

1. Sasaki, T. & Timpl, R. (1999) Laminins. In Guidebook to the Extracellular Matrix, Anchor and Adhesion Proteins (Kreis T. & Vale R., eds), 2nd edn, pp. 434–443. Oxford University Press, Oxford. 2. Colognato, H. & Yurchenco, P.D. (2000) Form and function: the 20. Martin, P.T., Ettinger, A.J. & Sanes, J.R. (1995) A synaptic localization domain in the synaptic cleft protein laminin b2 (s-laminin). Science 269, 413–416. laminin family of heterotrimers. Dev. Dyn. 218, 213–234. 21. Porter, B.E., Weis, J. & Sanes, J.R. (1995) A motoneuron-selective stop signal in the synaptic protein s-laminin. Neuron 14, 549–559. 3. Libby, R.T., Champliaud, M.-F., Claudepierre, T., Koch, M., Burgeson, R.E., Hunter, D.D. & Brunken, W.J. (2000) Laminin

442 T. Sasaki et al. (Eur. J. Biochem. 269)

(cid:211) FEBS 2002

39. Miosge, N., Sasaki, T. & Timpl, R. (1999) Angiogenesis inhibitor endostatin is a distinct component of elastic fibers in vessel walls. FASEB J. 13, 1743–1750.

40. Miosge, N., Quondamatteo, F., Klenczar, C. & Herken, R. (2000) Nidogen-1: expression and ultrastructural localization during the onset of mesoderm formation in the early mouse embryo. J. Histochem. Cytochem. 48, 229–237.

22. Brandenberger, R., Kammerer, R.A., Engel, J. & Chiquet, M. (1996) Native chick laminin-4 containing the b2-chain (s-laminin) promotes motor axon growth. J. Cell Biol. 135, 1583–1592. 23. Noakes, P.G., Gautam, M., Mudd, J., Sanes, J.R. & Merlie, J.P. (1995) Aberrant di(cid:128)erentiation of neuromuscular junctions in mice lacking s-laminin/laminin b2. Nature (London) 374, 258–262. 24. Noakes, P.G., Miner, J.H., Gautam, M., Cunningham, J.M., Sanes, J.R. & Merlie, J.P. (1995) The renal glomerulus of mice lacking s-laminin/laminin b2: nephrosis despite molecular com- pensation by laminin b1. Nat. Genet. 10, 400–406. 41. Stetefeld, J., Mayer, U., Timpl, R. & Huber, R. (1996) Crystal structure of three consecutive laminin-type epidermal growth factor-like (LE) modules of laminin c1 chain harboring the nid- ogen binding site. J. Mol. Biol. 257, 644–657.

42. Tisi, D., Talts, J.F., Timpl, R. & Hohenester, E. (2000) Structure of the C-terminal LG domain pair of the laminin a2 chain har- bouring binding sites for a-dystroglycan and heparin. EMBO J. 19, 1432–1440.

25. Paulsson, M., Aumailley, M., Deutzmann, R., Beck, K. & Engel, J. (1987) Laminin–nidogen complex: extraction with chelating agents and structural characterization. Eur. J. Biochem. 166, 11–19. 26. Schulze, B., Mann, K., Po¨ schl, E., Yamada, Y. & Timpl, R. (1996) Structural and functional analysis of the globular domain IVa of the laminin a1 chain and its impact on an adjacent RGD site. Biochem. J. 314, 847–851.

43. Durkin, M.E., Gantam, M., Loechel, F., Sanes, J.R., Merlie, J.P., Albrechtsen, R. & Wewer, U.M. (1996) Structural organization of the human and mouse laminin b2 chain genes, and alternative splicing at the 5¢ end of the human transcript. J. Biol. Chem. 271, 13407–13416. 27. Ettner, N., Go¨ hring, W., Sasaki, T., Mann, K. & Timpl, R. (1998) The N-terminal globular domain of the laminin a1 chain binds to a1b1 and a2b1 integrins and the heparan sulfate-containing domains of perlecan. FEBS Lett. 430, 217–221.

44. Friedrich, M.V.K., Go¨ hring, W., Mo¨ rgelin, M., Brancaccio, A., David, G. & Timpl, R. (1999) Structural basis of glycosamino- glycan modification and heterotypic interactions of perlecan domain V. J. Mol. Biol. 294, 259–270. 28. Kohfeldt, E., Maurer, P., Vannahme, C. & Timpl, R. (1997) Properties of the extracellular calcium binding module of the proteoglycan testican. FEBS Lett. 414, 557–561.

45. Dolan, M., Horchar, T., Rigatti, B. & Hassell, J.R. (1997) Iden- tification of sites in domain I of perlecan that regulate heparan sulfate synthesis. J. Biol. Chem. 272, 4316–4322. 29. Vallejo, A.N., Pogulis, R.J. & Pease, L.R. (1994) In vitro synthesis of novel genes: mutagenesis and recombination by PCR. PCR Methods Applic. 4, 123–130. 30. Bitter, T. & Muir, H. (1962) A modified uronic acid carbazole reaction. Anal. Biochem. 4, 330–334.

46. Pikkarainen, T., Eddy, R., Fukushima, Y., Byers, M., Shows, T., Pihlajaniemi, T., Saraste, M. & Tryggvason, K. (1987) Human laminin b1 chain. A multidomain protein with gene (LAMB1) locus in the q22 region of chromosome 7. J. Biol. Chem. 262, 10454–10462. 31. Mayer, U., Aumailley, M., Mann, K., Timpl, R. & Engel, J. (1991) Calcium-dependent binding of basement membrane protein BM- 40 (osteonectin, SPARC) to basement membrane collagen type IV. Eur. J. Biochem. 198, 141–150.

47. Montell, D.J. & Goodman, C.S. (1988) Drosophila substrate adhesion molecule: sequence of laminin b1 chain reveals domains of homology with mouse. Cell 53, 463–473.

32. Sasaki, T., Mann, K., Wiedemann, H., Go¨ hring, W., Lustig, A., Engel, J., Chu, M.-L. & Timpl, R. (1997) Dimer model for the microfibrillar protein fibulin-2 and identification of the connecting disulfide bridge. EMBO J. 16, 3035–3043.

48. Champliaud, M.-F., Lunstrum, G.-P., Rousselle, P., Nishiyama, T., Keene, D.R. & Burgeson, R.E. (1996) Human ammion con- tains a novel laminin variant, laminin 7, which like laminin 6, covalently associates with laminin 5 to promote stable epithelial– stromal attachment. J. Cell Biol. 132, 1189–1198. 33. Costell, M., Mann, K., Yamada, Y. & Timpl, R. (1997) Charac- terization of recombinant perlecan domain I and its substitution by glycosaminoglycans and oligosaccharides. Eur. J. Biochem. 243, 115–121. 34. Engel, J. (1994) Electron microscopy of extracellular matrix 49. Yurchenco, P.D. & Cheng, Y.-S. (1993) Self-assembly and calci- um-binding sites in laminin. A three-arm interaction model. J. Biol. Chem. 268, 17286–17299. components. Methods Enzymol. 245, 469–488.

50. Durbeej, M., Fecker, L., Hjalt, T., Zhang, H.-Y., Salmivirta, K., Klein, G., Timpl, R., Sorokin, L., Ebendal, T., Ekblom, P. & Ekblom, M. (1996) Expression of laminin a1, a5 and b2 chains during embryogenesis of the kidney and vasculature. Matrix Biol. 15, 397–413. 35. Mann, K., Mechling, D.E., Ba¨ chinger, H.-P., Eckerskorn, C., Gaill, F. & Timpl, R. (1996) Glycosylated threonine but not 4- hydroxyproline dominates the triple helix stabilizing positions in the sequence of a hydrothermal vent worm cuticle collagen. J. Mol. Biol. 261, 255–266. 36. Timpl, R. (1982) Antibodies to collagens and procollagens. Methods Enzymol. 82, 472–498.

51. Hunter, D.D., Murphy, M.D., Olsson, C.V. & Brunken, W.J. (1992) S-laminin expression in adult and developing retinae: a potential cue for photoreceptor morphogenesis. Neuron 8, 399– 413.

37. Mayer, U., Nischt, R., Po¨ schl, E., Mann, K., Fukuda, K., Gerl, M., Yamada, Y. & Timpl, R. (1993) A single EGF-like motif of laminin is responsible for high a(cid:129)nity nidogen binding. EMBO J. 12, 1879–1885.

52. Talts, J.F., Sasaki, T., Miosge, N., Go¨ hring, W., Mann, K., Mayne, R. & Timpl, R. (2000) Structural and functional analy- sis of the recombinant G domain of the laminin a4 chain and its proteolytic processing in tissues. J. Biol. Chem. 275, 35192– 35199. 38. Sasaki, T., Wiedemann, H., Matzner, M., Chu, M.-L. & Timpl, R. (1996) Expression of fibulin-2 by fibroblasts and deposition with fibronectin into a fibrillar matrix. J. Cell Sci. 109, 2895–2904.