The Journal of Cell Biology, Volume 145, Number 3, May 3, 1999 605–617
The Rockefeller University Press, 0021-9525/99/05/605/13 $2.00
Characterization and Expression of the Laminin
Non-Basement Membrane–associated, Laminin Chain
3 Chain: A Novel,
Manuel Koch,* Pamela F. Olson,* Anne Albus,* William J in,* Dale D. Hunter,
Robert E. Burgeson,* and Marie-France Champliaud*
*The Cutaneous Biology Research Center, Massachusetts General Hospital, and the Department of Dermatology, Harvard
Medical School, Charlestown, Massachusetts 02129; and
The Departments of Neuroscience, A natomy and Cell Biology, and
Ophthalmology, Tufts University School of Medicine, Boston, Massachusetts 02111
William J . Brunken,*
posed of an
tional roles in development and in stabilizing epithelial
structures. Here, we identified a novel laminin, com-
posed of known
chains but containing a novel
3. We have cloned gene encoding this chain,
LA MC3, which maps to chromosome 9 at q31-34. Pro-
tein and cDNA analyses demonstrate that
all the expected domains of a
consensus glycosylation sites and a putative nidogen-
binding site. This suggests that
are likely to exist in a stable matrix.
Studies of the tissue distribution of
that it is broadly expressed in: skin, heart, lung, and the
reproductive tracts. In skin,
Laminins are heterotrimeric molecules com-
, and a
chain; they have broad func-
chain, including two
3 chain show
3 protein is seen within
the basement membrane of the dermal-epidermal junc-
tion at points of nerve penetration. The
a prominent element of the apical surface of ciliated
epithelial cells of: lung, oviduct, epididymis, ductus def-
erens, and seminiferous tubules. The distribution of
3-containing laminins on the apical surfaces of a vari-
ety of epithelial tissues is novel and suggests that they
are not found within ultrastructurally defined basement
membranes. It seems likely that these apical laminins
are important in the morphogenesis and structural sta-
bility of the ciliated processes of these cells.
3 chain is also
laminin • testis • oviduct • lung • chromo-
arising from the coiled-coil interaction of three genetically
distinct polypeptide chains and three NH
arms, each originating from the individual polypeptide
chains (Maurer, 1996). The three subunit chains are
according to the current nomenclature
(Burgeson et al., 1994). The complete primary structure
for each of the nine human laminin subunit chains has
1 (Haaparanta et al., 1991),
teenaho et al., 1994),
3 (Ryan et al., 1994),
et al., 1995a),
1 (Pikkarainen et al., 1987),
1994; Iivanainen et al., 1995b),
1 (Pikkarainen et al., 1988), and
1992). The complete cDNA sequence for a fifth laminin
membranes (Ekblom, 1996). In overall appearance,
laminins are cross-shaped with a single long arm
are large glycoproteins found in all basement
2, (V uol-
2 (Wewer et al.,
3 (Gerecke et al., 1994),
2 (Kallunki et al.,
chain has been determined in mouse (Miner et al., 1995);
partial cDNA sequences of human
1997), and a novel chicken
chain (Y bot-Gonzalez et al.,
1995) have been reported. A ll three chains have globular
domains separated by multiple epidermal growth factor-
like domains within the NH
long arm portions are composed of heptad repeats that are
-helical coiled-coil proteins. In addition, the
COOH terminus of each
chain is composed of five glob-
domains (Engel, 1992).
The many functions ascribed to laminins are thought to
derive from their structural and signal transduction roles,
by which they contribute to the formation and stability of
basement membranes, to the stability of cellular attach-
ments to basement membranes, and to cytoskeletal rear-
rangements mediated by their occupancy of cell surface
receptors (Ryan et al., 1994). These activities at least par-
tially result from: (a) the binding of the COOH-terminal
laminin G domains to integrins in most cells (Deutzmann
5 (Durkin et al.,
-terminal short arms. Their
Dr. Brunken is on leave from the Department of Biology, Boston College.
A ddress correspondence to Dr. Marie-France Champliaud, MGH-East
CBRC, Bldg. 149, 13th Street, Charlestown, MA 02129. Tel.: (617) 724-
8285. Fax: (617) 726-4453. E-mail: firstname.lastname@example.org.
cular dystrophy; FISH, fluorescent in situ hybridization; G, globular.
Abbreviations used in this paper: FCMD, Fukuyama congenital mus-
The Journal of Cell Biology, Volume 145, 1999
et al., 1990; Drago et al., 1991; Goodman, 1992; Matter and
Laurie, 1994; Rousselle et al., 1995; Chen et al., 1997), and/or
to dystroglycan in muscle cells (Henry and Campbell,
1996; Pall et al., 1996; Wewer and Engvall, 1996; Cohen et al.,
1997); (b) from the self-assembly of the laminins into a
pericellular extracellular matrix through interactions of
domains V I, which are present at the ends of the short
arms of the subunit chains (Y urchenco et al., 1992; Y ur-
chenco and Cheng, 1993, 1994); and (c) from assembly of
the pericellular laminin network with a conceptually sepa-
rate network of type IV collagen molecules specifically
mediated by the molecule nidogen, one end of which binds
1 chain, and the other end of which binds
type IV collagen and other basement membrane matrix
components (Fox et al., 1991; Battaglia et al., 1992; A umail-
ley et al., 1993; Reinhardt et al., 1993).
The role of laminin 5 (
lial–stromal interactions is the exception to this general-
ized scheme. While the 5 laminin G domains of
1 on the epithelial basolateral surface
(Niessen et al., 1994), the absence of domains V I on the
truncated short arms of
Ryan et al., 1994), and the absence of a nidogen binding
2 (Mayer et al., 1995) prevent their participation
in the above-described model. Instead, the NH
of the epithelial cell–associated laminin 5 binds type V II
collagen present within the subjacent stromal matrix
(Rousselle et al., 1997). Thus, laminin 5 appears to play a
unique role in epithelial frictional resistance, rather than a
direct role in overall basement membrane structure.
The primary functions of the laminin
above generalized scheme is the contribution of a domain
V I which shares high amino acid sequence identity to do-
mains V I of the laminin
quence required for the binding of nidogen. This se-
quence, NIDPNA V , is present in the fourth EGF-like
repeat of domain III (Mayer et al., 1979; Poschl et al.,
1994). In the
2 chain, the analogous sequence NV DPSA S
is present, but does not support high affinity nidogen bind-
ing (Mayer et al., 1995).
In this report, we describe a novel human laminin
chain. Its predicted structure indicates the presence of all
the domains homologous to
mologue and a nidogen-binding site containing a single
conservative amino acid substitution. We call this chain
3 chain. The predicted structure of this chain
suggests that it should be capable of associations with
other laminin chains for basement membrane assembly;
however, immunolocalization studies in several tissues in-
dicate the presence of laminin
ultrastructurally identifiable basement membranes.
2) in stabilization of epithe-
3 bind in-
2 (Kallunki et al., 1992;
1 chain in the
chains and a unique se-
1, including a domain V I ho-
3 chains in regions lacking
Materials and Methods
Isolation of a Novel Laminin 12 (
The purification of the novel laminin 12 (
lows. Human placental chorionic villi were frozen in liquid nitrogen,
ground in a Waring blender, and then washed in 1 M NaCl. Unless other-
wise noted, all subsequent steps were performed at 4
pellet (200 g, wet weight) was suspended by stirring for 48 h in 1 liter of
extraction buffer (0.5 M NaCl, 10 mM EDTA , and 625 mg/liter
maleimide, 150 mg/liter phenylmethylsulfonyl fluoride, and 50 mM Tris-
3) was carried out as fol-
C. The final tissue
HCl, pH 7.8). The soluble fraction was collected after centrifugation
g , 60 min) and precipitated with 300 g/liter ammonium sulfate.
The precipitated proteins were collected by centrifugation (30,000
min) and dissolved in chromatography buffer (2 M urea, 25 mM NaCl, 5 mM
EDTA , and 50 mM Tris-HCl, pH 7.8). The sample was then dialyzed
against the same buffer. A fter dialysis, 0.5 vol of buffer-equilibrated
DEA E-cellulose (DE-52; Whatman) was added and the mixture was
shaken overnight. Material not bound to DEA E-cellulose was collected
by filtration on a Buchner funnel (Whatman; filter 4) and precipitated by
addition of 300 g/liter ammonium sulfate. The proteins were collected by
g , 60 min), redissolved in Concanavalin A buffer
(0.5 M NaCl, 5 mM CaCl
, 5 mM MgCl
and dialyzed against the same buffer overnight. The fraction was applied
to a 2.5
5 cm Concanavalin A –Sepharose column (Pharmacia). Un-
bound material was removed by extensive washing while bound proteins
were eluted by successive washing with 10 mM
-glucopyrannoside, and finally 1 M
Chemical Co.). Laminins are typically recovered in the latter two frac-
tions; each fraction was independently concentrated to 10 ml with an A m-
icon concentrator (30-kD membrane) and applied to a 2.5
Sephacryl S-500 column in 0.5 M NaCl, 50 mM Tris-HCl, pH 7.8. The frac-
tions of interest were pooled, dialyzed against Mono-Q buffer (0.1 M
NaCl and 25 mM Tris-HCl, pH 7.8), and applied to a 1
column (Pharmacia). Elution was achieved with a 60-ml 0.1–0.5-M NaCl
gradient. The fraction eluted at 250 mM NaCl was taken for further study.
, and 50 mM Tris-HCl, pH 7.8),
-mannopyrannoside, 1 M
5 cm Mono-Q
Protein sequencing was performed with minor modifications of published
methods (A ebersold et al., 1987). In brief, laminin 12 was resolved on a
polyacrylamide gel in the presence of 2-mercaptoethanol. The bands at
205, 185, and 170 kD were excised separately, digested with trypsin, and
then separated by HPLC and sequenced on an A pplied Biosystem se-
quenator. A nalysis of a trypsin digest of laminin
12 was performed with matrix-assisted laser desorption time-of-flight
mass spectrometry performed on a Finnigan Lasermat 2000 (Chait and
2 isolated from laminin
By comparison of the laminin
cession number, P11047) with the dbEST database (Boguski et al., 1993;
NCBI) using the program BLA ST (A ltschul et al., 1990) one clone (these
data are available from GenBank/EMBL/DDBJ under accession number
A A 297192) was chosen as a possible candidate for a new laminin
To extend the cDNA , specific primers for 5
from a previously published expressed sequence tag (clone, A A 297192).
Nested PCR on placental Marathon-R eady cDNA (Clontech) were
performed following the manufacturer’s instructions using the supplied
nonspecific primers with the following gene-specific primers: for the
extension; in the first round, (5
GCA CTGG); in the second round, (5
CA GTCTTGG); for the 3
extension, in the first round, (5
CA CGGGA CTGCA GCCGCTGCTA CCC); in the second round
-dGCTGCTA CCCTGGCTTCTTCGA CCTCC). For PCR, the Long
Expand PCR Kit (Boehringer Mannheim) was used with the following
conditions: denaturation, 94
C for 3 min; 10 cycles of 94
C per cycle) for 30 s, 68
C for 4 min; 25 cycles of 94
C for 30 s, 68
C for 4 min (
10 s per cycle); a final extension period at
C of 8 min.
The PCR samples from the first round were purified (PCR Purification
Kit; Qiagen) and 2% of the sample volume was used in the second round
of PCR using the same PCR protocol. These PCR products were purified
from an agarose gel (Gel Purification Kit; Qiagen) and either subcloned
(into PCR II or PCR 2.1 vectors; Invitrogen) or directly used for sequenc-
ing. To reconfirm the nucleotide sequence and control for PCR-induced
nucleotide substitutions, gene-specific primers were used to reamplify the
3 cDNA . A first strand cDNA synthesis kit (Clontech) was used to
synthesize cDNA from total placental RNA using oligo dT, random, or
specific human laminin
3 antisense primers following the manufacturer’s
protocol; PCR was used to generate overlapping clones complementary to
the entire human laminin
3 chain. Sequencing of all the obtained PCR
products revealed the nucleotide sequence of laminin
eventually inferred as nt 297 to nt 5020. However, all further 5
failed to extend the sequence further toward the 5
1 amino acid sequence (SWISS-PROT ac-
extension were deduced
-dGCGGCA GGTGCA CTGTC-
C for 30 s, 63
C for 30 s,
3 from what we
Koch et al.
The sequence of the 5
genomic P1 clone DMPC-HFF#1-1461F2, which was obtained from a
PCR-based library screen performed at Genome Systems, Inc. The oligo-
nucleotide primers that were provided to Genome Systems specifically
amplify exon 2 of the human
3 gene (sense primer, 5
CA GGGGA A GGCGGGTCCTG; antisense primer, 5
GA GA TCA CGTA TGTGA G). To obtain the sequence of the missing 5
end, the genomic clone was sequenced (in 4% DMSO) with gene-specific
antisense primers. The sequence of the laminin
was confirmed by RT-PCR from placental RNA in 4% DMSO, (sense
-dCGCGCGGCGTCGGTGCCCTTGA CC; antisense primer
-dGCTTGTA GA TGGCA A A GCTCTCA GG).
end of the cDNA was determined from the human
Nucleotide sequences were determined with a Thermo Sequenase cycle
sequencing kit and
M13 forward or reverse primers or gene-specific primers synthesized in
our laboratory. A 1:1 ratio of inosine to guanosine was included in the se-
quencing mix. Sequence data were assembled and manipulated using
Genetyx-Max 8.0 and Genestream-1 at http://genome.eerie.fr/home.html
(Software Development Co., Ltd.). The signal peptide cleavage site was
predicted using http://genome.cbs.dtu.dk/services/SignalP/ (Nielsen et al.,
33P-ddNTP (A mersham Pharmacia) using either the
Northern Blot Analysis
A 956-bp PCR product (nt 1316–2271) was generated (Long Expand PCR
Kit; Boehringer Mannheim) from placental cDNA , purified (PCR purifi-
cation kit; Qiagen), and labeled with [33P]dCTP (NEN) using the redi-
prime DNA labeling system (A mersham). Without further purification,
the probe was denatured in the same buffer containing 1/10 (vol/vol) hu-
man Cot-1 DNA (Boehringer Mannheim), and 1/10 (vol/vol) sheared
salmon testes DNA (GIBCO BRL) at 94?C for 5 min then chilled before
use. Northern blots (Clontech) were prehybridized in 50% formamide, 5?
SSPE, 1? Denhardt’s, 1% SDS, 10% Dextran-sulfate, 0.1 mg/ml salmon
sperm DNA (GIBCO BRL) at 42?C for 2 h the probe was added and hy-
bridized for 20 h. The blot was washed three times in 2? SSC, 1% SDS at
42?C and two times in 0.1? SSC, 1% SDS at 42?C. Blots were placed on a
BioMax MR film (Kodak) with a BioMax TranScreen-LE intensifying
screen (Kodak) for 20 h at ?70?C.
Recombinant Expression of Domain I of the ?3 Chain
A cDNA encoding the COOH terminus of human ?3 was cloned into the
HisTrx and pPEP-T vectors (kindly provided by Richard Kammerer,
Biozentrum, Basel, Switzerland; based on the pET system; Novagen). The
HisTrx vector has a histidine-tagged bacterial thioredoxin cDNA as a car-
rier in front of the cloning site; pPEP-T has a piece of the coiled-coil do-
main of mouse ?1 in front of the cloning site. The ?3 cDNA fragment used
was amplified by PCR from human placenta cDNA (see cDNA cloning)
using primers that include the EcoRI adapters: forward, 5?-GCGGA TC-
CGA GGA A GCTGA GCGGGTGGGTGCTG-3?; reverse, 5?-GCGA A -
TTCTTA CTGCCA GCTGGCA CA GTTCTCGGG-3?. The resultant plas-
mids were transformed into BL21(DE3) pLysS bacteria (Novagen) and
fusion proteins were isolated according to the pET System manual
(Novagen). A recombinant fragment containing only histidine-tagged
thioredoxin was similarly expressed and purified.
The 170-kD band (i.e., ?3 chain) was excised from the reducing SDS-
PA GE gel described above and injected into a rabbit for antibody produc-
tion following standard procedures (Harlow and Lane, 1988). The result-
ing serum (R16) was evaluated by Western analysis and shown to react
with the 170-kD ?3 chain, and showed minor cross-reactivity with other
laminin chains at high antibody concentrations. A ll antibody-related stud-
ies presented in this communication were conducted at concentrations
well below those where cross-reactivity was observed. The histidine-
tagged, thioredoxin-?3 fusion protein was used for the production of a
second rabbit antiserum (R21) which reacted with a single band in West-
ern blots of placental extracts.
The R16 antiserum was affinity-purified by binding to gel-purified ?3 that
had been transferred to nitrocellulose and then eluted with 1 M acetic acid
followed by immediate neutralization. The R21 serum was purified by
binding to the histidine-tagged, mouse ?1-human ?3 fusion protein cou-
pled to activated CNBr-Sepharose; the bound antibodies were eluted with
2 M urea in PBS, or with 1 M acetic acid which was immediately neutral-
ized. The immunofluorescent patterns produced by these two affinity-
purified antibody pools were indistinguishable, and were similar to whole
R21 serum with reduced background staining. Only affinity-purified prep-
arations of R21 serum were used for these studies. A ntibodies made
against histidine-tagged thioredoxin were similarly isolated by affinity
chromatography from R21 serum; immunofluorescent patterns with these
controls were blank.
Most tissues were obtained from various colleagues using specimens for
other purposes: these include tissues from male and female rats; from nor-
mal human tissues discarded after surgery; and from rhesus monkey,
Macaca mulatta. Bovine tissues were purchased from a local slaughter-
house. Dissected and blocked tissues were placed directly in embedding
compound (O.C.T.; Sakura Finetek) and frozen by immersion in liquid ni-
trogen-cooled isopentane. 10-?m sections were made on a Leica CM 3000
or 3050 and collected on Superfrost slides (Fisher Scientific). Sections
were air dried and stored at ?20?C until use. Just before use, sections
were immersed in acetone at ?20?C and then rinsed three times in PBS at
room temperature. Sections were incubated with primary antibodies di-
luted in PBS containing: 2% normal goat serum, 0.25% sodium azide, and
0.1% Triton X -100. Sections were incubated overnight at 4?C; they were
washed in three changes of PBS (5 min per wash) and then incubated for
45–60 min with secondary antibody coupled to either Cy3, FITC, or Texas
red. A fter incubation, sections were washed and coverslipped in Prolong
(Molecular Probes). The sections were imaged on a Leica confocal laser
scanning microscope (Leica TCS-NT). The gain was adjusted in each
channel of the confocal to assure that there was no bleeding across the
channels; this adjustment is performed at the outset of each confocal ses-
sion. Images were transferred to A dobe Photoshop and cropped for re-
production. The brightness and contrast were adjusted to make printed
images similar to that obtained on the microscope monitor.
Other primary laminin reactive primaries used were: polyclonal anti-
EHS-laminin-1 (Sigma Chemical Co.); monoclonal anti-laminin ?2 chain
(mA b 1922, Chemicon); polyclonal anti-laminin ?4 (Miner et al., 1997);
polyclonal anti-laminin ?5 chain (Miner et al., 1995); two monoclonal anti-
laminin ?1 chain (545, Marinkovich et al., 1992; clone C21, Green et al.,
1992); monoclonal anti-laminin ?2 chain (V errando et al., 1987). Mono-
clonal anti-PGP 9.5 (Ultraclone, Ltd.) was used to identify nerves in skin.
Secondary antibodies used were: goat anti-rabbit FITC (ICN Pharmaceu-
ticals); goat anti–rabbit-Cy3 (Jackson ImmunoResearch Laboratories).
In Situ Hybridizations
Paraffin sections were processed for in situ hybridizations as previously
described in detail (Libby et al., 1997). In brief, cRNA probes for the lami-
nin ?3 chain were generated from human ?3 clones; cRNA s were labeled
during transcription by the incorporation of digoxigenin-UTP (Boeh-
ringer Mannheim); ?1 ?g/ml of cRNA was used for hybridization; hybrid-
izations were performed at high stringency (50% formamide and 5? SSC,
60?C; see Libby et al., 1997 for complete details). A fter overnight hybrid-
ization, sections were washed (50% formamide, 1? SSC, for 30 min at
60?C) and the unhybridized probe was destroyed by RNase A . The hy-
brids were detected with an anti-digoxigenin antibody coupled to alkaline
phosphatase (Boehringer Mannheim). Sections were incubated overnight
with anti-digoxigenin diluted 1:1,000 in blocking solution (Boehringer
Mannheim). A fter washing to remove unbound antibody, endogenous al-
kaline phosphatase activity was blocked by washing in levamisole for 10
min; the alkaline phosphatase reaction was carried out overnight at room
SDS-PA GE (Laemmli, 1970) and electrophoretic transfer of proteins to
nitrocellulose with immunoblot analysis were performed essentially as
previously described (Lunstrum et al., 1986). For the FISH analysis, a
The Journal of Cell Biology, Volume 145, 1999
1217 bp cDNA fragment was generated by RT-PCR from placental RNA ,
using the sense primer 5?-dA GTGCCA CTA TA A CGGCA CA TGCG and
antisense primer 5?-dCTCGTGTCTGCA A GGA GTCTGTCA . The gel
band was purified and subcloned (PCR II vector; Invitrogen). A fter
the sequence of the fragment was verified, the resultant plasmid was used
for the fluorescent in situ localization of the LA MC3 gene (SeeDNA Bio-
Characterization of Laminin 12 (?2?1?3)
Laminin 12 was extracted from human chorionic villi using
EDTA and partially purified by a combination of DEA E-
cellulose, Concanavalin A , Sephacryl S-500, and Mono-Q
chromatography (see Materials and Methods). The final
fraction of interest resulting from the above protocol con-
tains multiple laminins. Laminin 12 was resolved from this
mixture by SDS-PA GE (3–5% polyacrylamide) under
nonreducing conditions. Six bands were resolved (Fig. 1,
Unreduced). Only the bands at ?560 kD and at the top of
the gel were reactive with a polyclonal anti-laminin antise-
rum (Sigma Chemical Co.; not shown). Therefore, the re-
solved band at 560 kD was excised, reduced in 10% 2-mer-
captoethanol SDS-PA GE sample buffer, and resolved by
5% SDS-PA GE. Three bands were observed with masses
of ?205, 185, and 170 kD (Fig. 1, Reduced). The band at
185 kD reacted with a monoclonal antibody (clone 545;
Marinkovich et al., 1992) specific to the laminin ?1 chain
(Fig. 1, Western blot). Each of the three bands was di-
gested with trypsin, the peptides were resolved by HPLC,
and selected resolved peptides were sequenced. The se-
quences obtained are shown in Table I. The 205-kD chain
contained three peptides sequence identical to human
laminin ?2 (published residues 536, 70, 1367; V uolteenaho
et al., 1994). On that basis, the band was identified as hu-
man laminin ?2, despite our observation that the 205-kD
band did not react with anti-?2 mA b (mA b 1922; Chemi-
con). The band at 185 kD produced two peptides identical
to human ?1, and was thereby confirmed as human ?1.
In contrast to the easy identification of the other two
bands, the band at 170 kD contained three sequences not
contained within any known laminin chain. The NH2-ter-
minal sequence of the 170-kD chain was determined, and
it also was novel; i.e., nonidentical to known laminin se-
quences. A s these four sequences from the 170-kD band
were derived from an unknown laminin and we had identi-
fied the laminin ? and ? chains, we assumed these se-
quences were derived from a novel laminin ? chain that we
The apparent molecular masses for the 205- and 185-kD
bands are not consistent with the literature values pub-
lished for the ?2 and ?1 chains, respectively. Thus, these
bands are indicated in Fig. 1 as ?2t, ?1t, and ?3t to indicate
that they have been processed (truncated). Laminin 2 and
laminin 4 were also present in these preparations; when
characterized by similar procedures (not described here in
detail) they showed molecular masses consistent with liter-
ature predictions, suggesting that our preparations were
not extensively and nonspecifically degraded. Together
these observations suggest that the truncations observed
for the ?3-containing molecules may be physiologically
Characterization of the ?3 cDNA
The cDNA sequences of human ?1 and ?2 were used to
probe the National Center for Biomedical Information
(NCBI) expressed-sequence-tag database (dbEST), and a
clone was identified that was homologous, but not identi-
cal, to ?1 and ?2. The sequence of this clone was used to
design PCR primers for extensions at 3? and 5? ends (see
Materials and Methods) using human placental cDNA ,
and additional sequence information was obtained by a
combination of genomic DNA and placental cDNA se-
quencing. The resulting sequence is shown in Fig. 2. The
Figure 1. Identification of laminin 12 isolated from human pla-
centa. A partially purified preparation of laminins isolated from
an EDTA extract of placenta was subjected to 3–5% SDS-PA GE
(Unreduced). The gel band in the indicated position (bar) was
excised from the gel and resolved by 5% SDS-PA GE after disul-
fide-bond reduction. Three bands were seen from the single un-
reduced band (Reduced). The 185-kD band was recognized by an
anti-?1 antibody, while the 170-kD band was specifically recog-
nized by a polyclonal antibody (R16; which used the same band
as an antigen; Western blot). Each of the three bands (Reduced
gel) was subsequently identified as laminin ?2, ?1, and ?3 (see
text for details). A s they have faster electrophoretic mobilities
than expected for these chains they are marked ?2t, ?1t, and ?3t
to indicate that they are truncated.
Table I. Comparison of Peptide Sequences Obtained from
Placental Laminin 12 205- and 185-kD Bands and Published
Sequences of ?2 and ?1 Residues
Human ?2* res. no. 536
Peptide from 205-kD band
Human ?2* res. no. 70
Peptide from 205-kD band
Human ?2* res. no. 1367
Peptide from 205-kD band
Human ?1‡ no. 171
Peptide from 185-kD band
Human ?1‡ res. no. 558
Peptide from 185-kD band
*Vuolteenaho et al., 1994.
‡Pikkarainen et al., 1987.
Koch et al. Laminin ?3 Chain
deduced amino acid sequence contains regions with 100%
identity to all three of the peptide sequences obtained
from the 170-kD band (underlined in Fig. 2). The nucle-
otide sequence reported in this paper has been submitted
to GenBank/EMBL Data Bank with the accession number
The DNA sequence contains an open reading frame
predicting 1620 amino acids, including a 19–amino acid-
long putative signal peptide that closely meets the criteria
described by Nielsen et al. (1997). The predicted cleavage
site was confirmed by protein sequencing of the ?3 NH2
terminus; this sequence exactly matched the predicted
amino acid sequence following the signal peptide. The
overall sequence of ?3 is most similar to that of ?1, sharing
52% amino acid similarity with human ?1 (Pikkarainen et
al., 1988). In addition, the amino acid sequence predicted
by the ?3 cDNA contains a domain distribution most like
that of the ?1 chain. A ll six domains are represented.
Overall, the ?3 chain has 43.6% amino acid identity with
the ?1 chain and 34% identity with the ?2 chain. The high-
est conservation is seen between domains ?1V I and ?3V I
(Fig. 3). Domains ?3V and III also show considerable sim-
ilarity to domains ?1V and III and ?2V and III.
The predicted ?3 sequence contains nine potential gly-
cosylation sites (Fig. 2, boxed), only two of which (Fig. 2,
boxed and underlined) are conserved in both human and
mouse ?1. A s these conserved sites are contained within
the globular domains IV and V I, it is likely that these sites
are used physiologically. There is a single RGD sequence
(boxed, hatched) within domain II, but this site is not con-
served in either human or mouse ?1 and ?2 proteins. The
sequence NV DPNA V (Fig. 2, double boxed) occurs
within the fourth E GF-like repeat of domain III and is
a homologue of the nidogen binding site (NIDPNA V )
within the same domain of ?1. These sequences differ by
only a single conservative amino acid substitution.
LAMC3 Maps to Chromosome 9q31-q34
The ?3 chromosomal location was determined by search-
ing the NCBI Human Genomic Sequencing Index data
base with the ?3 cDNA sequence. The sequence is identi-
cal to a database, Sequence Tagged Sites (clone WI-14302),
that has been localized to chromosome 9q33-q34. A 1.2-kb
?3 cDNA probe within domains I and II of the predicted
Figure 2. The complete amino acid sequence of human ?3 as pre-
dicted from the corresponding cDNA sequence. The amino acid
positions are numbered in accordance to the homologous loca-
tions within laminin ?1 (Pikkarainen et al., 1988). A 19–amino
acid signal sequence precedes the ?3 NH2 terminus (arrow marks
the signal peptide cleavage site). Peptide sequences obtained
from Edman analyses of the fragmented ?3t gel band (Fig. 1, Re-
duced) are underlined. Potential glycosylation sites are boxed,
and those conserved between ?1 and ?3 are also underlined. The
nidogen binding consensus sequence is boxed in bold type. The
sequence RGD is boxed and hatched.
Figure 3. Comparison of the predicted domain structures and the
percent amino acid sequence identity of ?3 with ?1 and with ?2.
The ?3 cDNA predicts a full-length laminin chain where all the
domains found in ?1 are represented. On the left side a diagram
of the domain structure of both ?1 and ?3 is given; on the right,
the domain structure of ?2. A bove each diagram is noted by do-
main the percent amino acid identity between ?1 or ?2 and ?3.
The degree of amino acid identity of ?3 to ?1 is higher than that
of ?3 to ?2, suggesting that ?3 has a closer evolutionary relation-
ship with ?1 than ?2.
The Journal of Cell Biology, Volume 145, 1999
protein, the regions of least homology among the ? chains,
was used to localize LA MC3 by fluorescent in situ hybrid-
ization (FISH) analysis (SeeDNA Biotech, Inc.). The re-
sults confirm the localization to chromosome 9q31-q34
Laminin ?3 Associates with ?2?1 to Form Laminin 12
which, in Placenta, Is Lacking Part of the I/II Domains
of All Three Chains
To determine the domains present within ?2t, the 205-kD
gel band, purified from placenta, was fragmented with
trypsin and the resulting peptides were fractionated by
HPLC; the masses of the eluted peptides were determined
by mass spectroscopy. The ion chromatograms were then
evaluated relative to the masses predicted from the pub-
lished amino acid sequence for ?2 in order to determine
the NH2- and COOH-terminal peptides present within the
digest. The results identified a number of tryptic peptides;
among these, the peptide LV EHV PGQP(V R), beginning
at residue 70 within domain V I, was the most NH2-ter-
minal; the peptide GTTMTPPA DLIE K , beginning at
residue 1367 within domain III, was the most COOH-
terminal. These results indicate that ?2t is a fragment con-
taining the short arm of the laminin ?2 chain. This conclu-
sion is consistent with the observation that the initial
peptide sequence identified from ?2t was within the short
arm domains (above).
?1t and ?3t are also short arm fragments, as all the pep-
tide sequences determined for both species are present
within the short arm domains. However, the masses of ?2t,
?1t, and ?3t are greater than predicted for the short arms
alone. In addition, ?2t, ?1t, and ?3t are not separable by
gel electrophoresis without the reduction of disulfide
bonds. Therefore, this truncated laminin 12 molecule is
very likely to contain portions of domain II of all three
chains, as the interchain disulfide bonds should lie be-
tween these domains. It is of interest to note that domain
II of ?3 contains three cysteinyl residues whose bonding
partners are not readily identified and are not present in
domain II of other laminin chains. These three cysteinyl
residues are conserved in mouse ?3 (A lbus, A ., and R.E.
Burgeson, unpublished observation). Whether these cys-
teinyl residues could form intrachain, interchain, or inter-
molecular disulfide bonds that in some way contribute to
the cleavage of ?3 chain-containing laminins is unknown.
Tissue Distribution of ?3 Expression by
Tissue RNA blots and Master Blot dot blots (Clontech)
were probed with a ?3 nucleotide probe (nt 1316 to 2277).
A single major transcript of ?5 kb, consistent in size with
other laminin ? chains, is present in several of the tissues
examined (Fig. 5 A ). A small amount of a second larger
transcript can also be detected. This larger transcript is
most likely due to differences in polyadelylation or due to
inefficient splicing. The ?3 chain RNA is abundant in
spleen, testis, placenta, lung, and liver; lesser amounts are
seen in kidney and ovary (Fig. 5 A ). The predominance of
a single transcript allowed use of the RNA Master Blot
(Clontech) to determine expression in a large number of
other tissues. On this dot blot, tissue RNA concentrations
have been normalized to housekeeping genes. The Master
Blot (Fig. 5 B) confirms the abundant presence of ?3 tran-
scripts in placenta, adrenal gland, testis, lung, and fetal
kidney, but also shows detectable levels of ?3 transcripts in
numerous additional tissues, including brain and skeletal
Characterization of the Immunospecificity of
Anti-Laminin ?3 (R16; R21)
A polyclonal antiserum, R16, was made in a rabbit to the
?3 chain excised from a reduced SDS-PA GE gel similar to
that shown in Fig. 1. A nother, R21, was made to recombi-
nant ?3 protein (see Methods). The R16 antiserum recog-
nizes the ?3 chain on immunoblots of placental extracts,
but at very high antibody concentrations, it shows some re-
activity with the ?1 and ?1 chains as well. Thus, as a con-
trol, human neonatal foreskin was immunostained with
anti-laminin ?1 (polyclonal anti-laminin 1; Sigma Chemi-
cal Co.), anti-laminin ?2 (GB3, V errando et al., 1987), and
with anti-laminin ?3 (R16). Crisp, brilliant fluorescence
was observed along the dermal-epidermal junction, and
around capillaries with the anti-?1 antibodies (data not
shown), and in the basement membrane at the dermal-epi-
dermal junction with anti-?2 (data not shown); in contrast,
no signal above background was detected using the anti-?3
reagent (R16) when it was applied at dilutions of 1:250 or
more (data not shown). The antigen could not be un-
masked by treatment of the cryosections with 2, 4, or 6 M
urea, or with 2 M guanidinium-HCl (data not shown). A s
all known laminin chains have been detected in skin within
either the epithelial basement membranes or the vascular
basement membranes, these results indicate that the cross-
reactivity detected by Western blot analyses using the
polyclonal anti-?3 (R16) antibody was either not apparent
by immunohistochemistry, or was below detection at the
Figure 4. Localization of LA MC3 to chromosome 9, band q31-
q34 by FISH. The position of LA MC3 was probed using a 1.3-kb
cDNA probe within predicted protein domains I and II of ?3.
The FISH signal (A ) was superimposed over the DA PI-banded
chromosomes (B) to identify the location of the ?3 gene. Both al-
leles of the ?3 gene are labeled (A , arrow) at the end of the long
arm of chromosome 9 (B, 9).
Koch et al. Laminin ?3 Chain
antibody concentrations used. For the subsequent anatom-
ical experiments (below), R16 was diluted 1:250 or greater
to assure no cross-reactivity was occurring. R21, the affin-
ity-purified antiserum to recombinant ?3, was also tested
on sections of neonatal foreskin. A s with R16, no immu-
noreactivity was seen (data not shown); thus we conclude
that this antiserum has no cross-reactivity with other
known ? chains. Neither R21 nor R16 antiserums label the
blood vessel basement membranes (see below) consistent
with a lack of cross-reactivity to other ? chains.
?3-containing Laminins Are Localized to Peripheral
Nerves and to Ciliated Epithelial Apical Surfaces
Unlike the lack of anti-laminin ?3 chain immunoreactivity
seen in neonatal foreskin, laminin ?3 chain immunoreac-
tivity was detected in human leg skin. A s shown in Fig. 6
A , and consistent with published results, laminin ?1 chain
reactivity is seen at the dermal-epidermal junction and
within the basement membranes of the vasculature, while
laminin ?2 chain immunoreactivity is restricted to the der-
mal-epidermal junction (Fig. 6 B). The laminin ?3 chain
immunoreactivity is further restricted to distinct patches
widely spaced along the dermal-epidermal junction (Fig. 6
C). In experiments not shown, the immunoreactivity did
not correlate positively or negatively with sites of cell pro-
liferation, nor did it correlate with fixed positions relative
to the rete ridges. However, there is a direct correlation of
the laminin ?3 chain immunoreactivity (Fig. 6 D) with sites
where nerves cross the dermal-epidermal junction as de-
tected by an antibody to the neuronal marker PGP9.5
(Fig. 6 E), which reacts with ubiquitin COOH-terminal hy-
drolase (Day et al., 1990). The results in skin suggest that
?3-containing laminins may be deposited into the dermal–
epidermal junction by nerve or nerve associated cells, or
that its expression by epithelial cells is induced by the ad-
Laminin ?3 is also expressed in the neural retina at the
apical surface of the retina and in the outer synaptic layer
(Libby, R.T., Y . X u, E.P. Gibbons, M.-F. Champliaud, M.
Koch, R.E. Burgeson, D.D. Hunter, and W.J. Brunken,
manuscript submitted for publication); in the retina, the ?3
chain is coexpressed with the ?4, ?3, and ?2 chains. Native
?3-containing laminins have not been isolated as yet from
Figure 5. Laminin ?3 chain is ex-
pressed widely in human tissues. The
same cDNA probe described in Fig. 4
was used to probe tissue blots (A )
and a dot blot (B). Only a single ma-
jor RNA species is detected on the
tissue blots; strong signals are ob-
tained from testis, spleen, placenta,
and lung. Less intense signals are ob-
tained from ovary, kidney, liver, and
skeletal muscle. Dot blots indicate a
more widespread tissue distribution,
including thyroid, gut, and some fetal
The Journal of Cell Biology, Volume 145, 1999
the retina; however, they have from another region of the
central nervous system, the cerebellum, from which we
have obtained two novel laminins, ?3?2?3 and ?4?2?3
(Champliaud, M.-F., unpublished observations). In addi-
tion, anatomical methods (immunohistochemistry and in
situ hybridization), demonstrate the expression of ?3 in
cerebellum and forebrain structures (Brunken, W.J., un-
published observations). It seems likely that ?3-containing
laminins will be a general feature of the matrix in the CNS.
The Northern analysis indicated that the laminin ?3
chain was most strongly expressed in placenta, testis, lung,
liver, spleen, and ovary. Therefore, we examined the local-
ization of ?3 chains within testis, lung, and ovary. The re-
activity within the epididymis and the fallopian tube were
particularly striking. Thus, the distribution of ?3 in these
tissues was extensively studied. In the female reproductive
system, the oviduct was strongly reactive. Cryosections of
the bovine (Fig. 7, A –F) or rat (Fig. 7, G–I) ampulla re-
acted for ?3 using R16 (Fig. 7, A , D, and G–I), or R21
(Fig. 7, B and E) showed brilliant immunoreactivity at the
apical surfaces of the tubal mucosa. Double immunofluo-
rescent studies performed with laminin ?3 and either lami-
nin ?2 (Fig. 7, A and D) or laminin ?5 antibodies (Fig. 7
E) demonstrated that both of these ? chains are restricted
to the basement membranes of the tubal epithelial and the
subjacent endothelium whereas ?3 is expressed at the api-
cal surface. The pre-immune serum from rabbit 21 (Fig. 7
C) was negative, as was the reactivity of anti-thioredoxin
antibodies purified from the R21 serum by immunoaffinity
(Fig. 7 F).
The pattern of immunoreactivity for R16 in the rat ovi-
duct was identical to that seen in bovine tissue (Fig. 7 G).
Higher magnification micrographs of the epithelial apical
surface of the rat ampulla (Fig. 7, H and I) show the ?3
chain to be localized to the apical surface of the epithelial
cells at the base of the cilia.
It should be noted that the labeling pattern of R21 dif-
fered somewhat from that of R16. In general, the pattern
with R21 was somewhat punctate, showing large deposits
of immunoreactivity at the apical surface, and increased
cytoplasmic labeling of the tubal epithelium, whereas the
R16 immunoreactivity was more restricted to the apical
extracellular surface. These observations suggest that the
R21 antiserum, made to recombinant domain I, may rec-
ognize the unfolded ?3 chain better than the R16 antise-
rum, which should recognize primarily short arm domains.
The male monkey reproductive tract was examined also.
Like the fallopian tube, the epithelium in the epididymis is
a single columnar epithelium (Fig. 8, A , H, and E). In situ
hybridization performed on adjacent sections of the mon-
key epididymis (Fig. 8 B; ?3) localized transcripts for the
?3 chain to the apical region of the epithelial cells (com-
pare Fig. 8, A with B). R16 (data not shown) and R21 (Fig.
8, C and D, R21) sera gave similar patterns, reacting with
both the basal and apical surfaces of the epithelial cells.
The R21 antiserum reacted with apparently intracellular
stores of ?3, as was seen in the bovine fallopian tube. The
preimmune control serum from R21 showed only punctate
autofluorescence (Fig. 8 E, Pre).
Potential chains partners were explored by examination
of the same tissue with antibodies specific for a variety of
other laminin chains: ?2 (Fig. 8 F), ?4 (G), ?1 (H), and ?2
(I). We used two monoclonal antibodies to test for the
presence of ?1 at the apical surface (clones 545; and C21)
both gave the same pattern of immunolabeling; only the
results with clone 545 are shown. A s can be seen readily,
?2 and ?2 were restricted to the basal surface of the epi-
thelial cells, while staining for ?4 and ?1 were also seen at
the apical surface. Thus, in contrast to the results from pla-
cental extracts, ?4 (and not ?2) appears to be a candidate
chain partner for ?3 in the epididymis. These observations
suggest that a wide variety of ?3-containing laminins will
be expressed in a tissue-specific pattern.
Expression of laminin ?3 chain was examined in the rat
as well and the tissue distribution of ?3 in the rat epididy-
mis was similar to that described for the monkey (data not
shown); namely, ?3 immunoreactivity was localized to the
apical surface of the epithelium. We also studied other
regions of the rat reproductive system. Unlike laminin-1
immunoreactivity, which is localized to the basement
membrane of the seminiferous tubules (Fig. 9 A ), ?3 im-
munoreactivity is not present within the basement mem-
brane of the seminiferous tubules nor is it found around
the interstitial cells (Fig. 9 A , arrows; Fig. 9 B, asterisk).
Figure 6. Laminin ?3 chain is expressed in restricted sites, corre-
lated with nerve terminations, in human skin. Laminin ?1, ?2,
and ?3 chains and nerve-specific ubiquitin COOH-terminal hy-
drolase (PGP 9.5) in human leg skin were localized by immuno-
histochemistry and confocal laser microscopy. Laminin ?1 (A )
immunoreactivity is present throughout the dermal-epidermal
junctional basement membrane, as well as in the basement mem-
branes of the glands and hair follicles. Laminin ?2 immunoreac-
tivity (B), on the other hand, is restricted to the basement mem-
brane of the dermal-epidermal junction and that underlying the
epithelial cells of the outer root sheath of the hair follicles. Fi-
nally, laminin ?3 immunoreactivity (C) is most restricted and is
found in distinct patches within the dermal-epidermal junction
basement membrane. Laminin ?3 immunofluorescence (D) and a
nerve-specific marker (E, anti-PGP 9.5) are colocalized when
both primaries are applied to the same section.
Koch et al. Laminin ?3 Chain
Within the seminiferous tubules, only the occasional tu-
bule reacted strongly with the laminin ?3 reactive serum
(R16, Fig. 9 B); it was our impression that those tubules
identified by the antibody contained nearly mature sper-
matids. Further along the male reproductive system, in the
ductus deferens, laminin-1 immunoreactivity (Fig. 9 C; ar-
rows mark the apical surface of the epithelium) was seen
along the epithelial basement membrane, in the lamina
propria and ensheathing the smooth muscle cells of the
muscular layer. In contrast, ?3 immunoreactivity (R16)
was found at the apical and basal surfaces of the epithelial
cells, as well as intracellularly (Fig. 9 D).
The apical distribution of the ?3 chain is not confined to
the reproductive system; in rat lung, the ciliated epithelial
cells lining the bronchi were also strongly reactive with the
anti-laminin ?3 antiserum, R16 (Fig. 9 E). A gain, the fluo-
rescence was apparent along the apical surface, as deter-
mined by differential interference contrast microscopy
(Fig. 9 F). No ?3-immunoreactivity was seen in respiratory
epithelium nor in the pulmonary capillary bed (not
The laminin ?3 chain described here is the eleventh lami-
nin subunit to be identified. The predicted primary and
secondary structure of this chain suggests that ?3 is more
closely related to human ?1 than ?2. Unlike ?2, the ?3
cDNA sequence predicts a laminin subunit without the
short-arm truncations predicted for ?2. Perhaps more sig-
nificantly, ?3 contains a ?1-like nidogen binding motif with
only a single conservative amino acid substitution, suggest-
ing that ?3-containing laminins should be capable of asso-
ciating with other basement membrane molecules through
nidogen interactions (Mayer et al., 1995; Poschl et al.,
1996). In addition, domain V I of ?3 shares the highest se-
quence identity with domain V I of the ?1 chain. A s this
latter domain has been shown to support laminin self-
Figure 7. Laminin ?3 is ex-
pressed on the apical surface of
the ciliated epithelium of the bo-
vine and rat fallopian tube. Con-
focal laser microscope images of
freshly frozen tissue sections of
fallopian tube; in each image, five
optical sections were superim-
posed. The laminin ?3-reactive
antiserum R16 (A and D, red)
and anti-laminin ?2 chain (A and
D, green) were applied to bovine
tissue together; ?3 immunoreac-
tivity is confined to the apical sur-
face of the epithelium, whereas ?2
is present in the basement mem-
brane; D is a higher magnification
of one of the folia in A . A pplica-
tion of the ?3-reactive R21 serum
alone (B) also labels the apical
surface of the bovine epithelium,
but in plaque-like structures. R21
pre-immune serum is nonreactive
(C). Simultaneous visualization of
R21 immunoreactivity (E, red)
and ?5 immunoreactivity (E,
green) shows that the ?3 is found
on the apical side of the epithe-
lium and in cytoplasmic stores in
the long processes of these colum-
nar tubal cells whereas ?5-immu-
noreactivity is confined to the
basement membrane. A ntibodies
purified from the R21 serum with
the carrier protein, His-thiore-
doxin are negative (F). ?3 shows a
similar distribution in rat fallopian
tube (G and H): R16 reactivity is
present at the apical surface over
the whole of the ampulla (G) and
in a higher power view of the epi-
thelium (H); comparison of the fluorescent image in H with the differential interference contrast image in I demonstrates that ?3 immu-
noreactivity is associated with the ciliated surface of the epithelium.
The Journal of Cell Biology, Volume 145, 1999
assembly (Y urchenco and Cheng, 1994), it seems reason-
able to suggest that domain V I of the ?3 chain may also
Two of the predicted glycosylation sites in the ?3 chain
are also found in human and in mouse ?1; these are within
short-arm globular domains, i.e., V I and IV . Interestingly,
the glycosylation site in domain IV is also found in human
(Kallunki et al., 1992) and in mouse (Sugiyama et al.,
1995) ?2. This remarkable conservation of glycosylation
sites among these three chains and between these species
suggests that these sites are indeed important and glycosy-
lated; they are likely to be critical in the folding of this re-
gion or may play another important function.
The RGD sequence within domain II of the ?3 chain is
found in neither ?1 nor in ?2. Moreover, it seems likely
that this sequence is not functional within native ?3-con-
taining laminins as it is located within the coiled-coil
region of ?3. However, it very well may promote integrin-
mediated recognition of non-native molecules or of pro-
In placenta, the ?3 chain can combine with the laminin
?2 and ?1 chains. This observation suggests that, unlike ?2
which pairs preferentially with ?3, ?3 may pair with any ?
chain, with the possible exception of ?3, and with any of
the known ? chains. This prediction suggests the existence
of an additional 10 laminins with the following chain com-
positions: ?1?1?3, ?1?2?3, ?2?1?3, ?2?2?3, ?3?1?3,
?3?2?3, ?4?1?3, ?4?2?3, ?5?1?3, and ?5?2?3. In both
the epididymis and the fallopian tube, ?3 is not combined
with ?2. In the epididymis, the ?4 and ?1 chains appear to
be potential partners. Given that the total number of hu-
man laminins is not known, at least one additional ? chain
has been identified in chicken (Y bot-Gonzalez et al., 1995;
Liu et al., 1998) and in mammals (Olson, P.F., unpublished
observations), assigning a final laminin numerical identi-
fier to these laminins is premature. However, as we have
shown ?2?1?3 to be the twelfth laminin to be identified,
we provisionally call this heterotrimer, laminin 12.
The masses of the chains of laminin 12 as approximated
by electrophoretic mobility are considerably less than pre-
dicted by the amino acid sequences and from prior experi-
ence with the ?2 and ?1 chains. They are also less than the
?2, ?1, and ?2 chains present in laminins 2 and 4 obtained
from the same preparations. The reason for these more
rapid electrophoretic migration rates appears to be pro-
teolysis within the domains II of the chains comprising this
molecule. In placenta, this proteolysis may be physiologi-
cal, since laminins 2 and 4 isolated from the same prepara-
tions are apparently intact. The significance of this obser-
vation awaits considerable additional experimentation
before it is understood. However, we have observed three
cysteinyl residues within domain II of the ?3 chain that are
not present in other human ?, ?, or ? chains. It is possible
that a disulfide bond between two of these residues dis-
torts the coiled-coil conformation making molecules con-
taining this chain more susceptible to proteolysis. A t this
time, we do not know if truncation of laminins containing
the ?3 chain can be generalized to tissues other than pla-
centa; however, this seems unlikely in that our isolation
of ?3-containing laminins from the CNS do not show
Figure 8. Laminin ?3 is expressed on the
surface of male reproductive epithelium.
Light (A and B) and confocal (C–I) images
of the epididymis from monkey (rhesus). In
A , a section was fixed and stained with
H&E for orientation; the epididymis is
composed of a columnar epithelium, with
considerable interstitial matrix between ad-
jacent folds. (B) By in situ hybridization, a
?3-specific probe is localized to the epithe-
lium; the reaction product (dark blue) is
concentrated above the nuclear region of
the cell; this section is not counter-stained.
(C and D) ?3 immunoreactivity (R21) is
present at the surface of the epithelium; (E)
pre-immune (Pre) serum is nonreactive in
this tissue. (F–I) Immunolocalization of
other laminin chains; both ?2 and ?2 are
confined to the basement membranes of the
epithelium and blood vessels, whereas both
?4 and ?1 are seen at the apical surface of
the epithelium; ?4 is also expressed in the
interstitial matrix and the basement mem-
brane. Bars, 100 ?m.
Koch et al. Laminin ?3 Chain
the same truncation (Champliaud, unpublished observa-
tions). The COOH-terminal truncation of ?2t explains its
lack of reactivity with an anti-?2 antibody, mA b 1922,
which is specific for the ?2 G domain (Engvall et al., 1990).
The antiserum to the recombinant ?3 domain I fusion
protein (R21) was originally made to evaluate this poten-
tial processing, since epitopes contained within this do-
main should be absent from the processed molecule. In
this regard, the immunohistochemical data is not defini-
tive. Like R16, R21 immunoreactivity is seen at the apical
surface; however, the apical reactivity is distinctly differ-
ent than that for R16. Specifically, R21 immunoreactivity
appears as a plaque-like structure at the cell surface, with
some reactivity within the cells. Thus, it seems possible
that R21 epitopes are entirely intracellular but it is also
possible that some of the R21 epitopes are present at the
apical surface of these epithelial cells. Further experimen-
tation beyond the scope of this report is required to ad-
dress this question. However, laminins containing ?3
chains have been immunoisolated from the medium of
A 204 cells derived from a human rhabdosarcoma (Cham-
pliaud, M.-F., unpublished data), indicating that ?3-con-
taining laminins are capable of being secreted in vitro.
These ?3-containing laminins from A 204 are a mixture
of processed (truncated) and unprocessed (untruncated)
Reports of laminins in tissue locations not identified as
basement membranes are increasingly frequent. In the
brain, laminins have been observed not only within the
basement membrane of capillaries, but also at other sites
not conceptualized as basement membranes (Higuchi et al.,
1991; Jucker et al., 1992, 1996a,b; Mori et al., 1992; Tian
et al., 1996, 1997; Hagg et al., 1997; Y amamoto et al.,
1997). In the eye, the laminin ?2 chain has been identified
in both basement membrane and non-basement mem-
brane locations (Hunter et al., 1992; Libby et al., 1996;
Libby et al., 1997; Toti et al., 1997). Laminins have also
been observed in cartilage (Durr et al., 1996).
Intriguingly, the laminin ?3 chain appears most com-
monly to be associated with non-basement membrane
structures. In the cerebellum, ?3 chains are detected in
the pericellular nets surrounding both neurons and glia
(Brunken, W.J., unpublished observations). Reported else-
where (Libby et al., manuscript submitted for publication,
see above), ?3 is present within the neural retina at two ex-
tracellular sites: between the outer segments of the photo-
receptors, and at the synapses of the photoreceptors with
the bipolar and horizontal cells. In the retina, these ?3-
containing molecules are the products of the Müller glial
cells which, like the tubal epithelium, contain a consider-
able store of intracellular ?3 chain. The laminin ?3 and ?2
chains are also present at these sites, whereas the ?1 and
?2 chains are absent. The functions fulfilled by these lami-
nins are unclear, but possibilities include: stabilization of
neural architecture; induction and stabilization of differen-
tiated neural phenotypes (Hunter et al., 1992b; Libby et
al., 1996; Hunter and Brunken, 1997); and stabilization of
However, the most abundant expression of ?3 as de-
tected by Northern analyses is not within neural tissues,
but rather is in the testis, the placenta, the spleen, the lung,
and the ovary. ?3 immunoreactivity is present at the bases
of the epithelial cilia of the epididymis, the trachea, the
bronchi, and the oviduct. There are no structures resem-
bling basement membrane at these sites. However, ?3
chains may be present within the basement membranes
along the basolateral surfaces of some of these epithelia.
Figure 9. Laminin ?3 is expressed on the surface of rat ciliated
epithelia in testis and lung. V arious tissues from the rat male re-
productive system (A –D) were incubated with either laminin-1
(?1/?1/?1) antiserum (left column) or ?3 antiserum (right col-
umn). (A and B) testis; (C and D) ductus deferens. The laminin-1
immunoreactivity is present in the basement membrane in these
structures (left column); arrows denote interstitial cells in A , and
the apical borders of the epithelium in C; the basal lamina sur-
rounding the smooth muscle is also easily distinguished (C). In
contrast, laminin ?3 expression is distinctly not in the basement
membrane of the testis (B) nor of the ductus deferens (D). ?3 ex-
pression is limited to the seminiferous tubules themselves and
does not surround the interstitial cells (B, asterisk). Within the
seminiferous tubules, ?3 expression appears to change with the
state of maturation of sperm (see text for details). In the ductus
deferens, ?3 expression (D) is limited to the surface of epithelial
cells (basolateral and apical). No ?3 expression is seen in the
smooth muscle of the ductus deferens. In lung (E and F), laminin
?3 is expressed on the apical surface of nonrespiratory airways.
?3 immunoreactivity is limited to the apical surface of the epithe-
lium (E; compare to the differential interference contrast image
of the same section in F). Scale bars are given for each pair of fig-
ures (A and B, C and D, E and F).
Koch et al. Laminin ?3 Chain
The chain partners for the ?3 chain in these apical lami-
nins are not yet known with certainty. However, in epi-
didymis, the laminin ?2 chain does not colocalize with ?3
at the apical epithelial surface; rather, the ?4 and ?1
chains are present at that location and, thereby, are poten-
tial chain partners. Thus, it seems likely that ?3 will be as
promiscuous as ?1 with respect to partner choice during
The presence of laminins along ciliated epithelial sur-
faces was unexpected and their functions there are un-
known. Perhaps a modified basement membrane con-
taining at least laminin helps organize or stabilize the
specialized cytoskeleton of the cilia. Laminins at these api-
cal surfaces may also participate in the anchorage of mu-
cins to the surface. A lternatively, laminins might stabilize
the outfoldings of the plasma membranes of the cilia. Sim-
ilar functions for laminins have been postulated to stabi-
lize the junctional folds beneath synapses at neuromuscu-
lar junctions (Noakes et al., 1995), and contribute to the
organization of epithelial hemidesmosomes (Langhofer et
al., 1993; Baker et al., 1996). Laminins expressed at the
apical surface of the retina are thought to play a role in
photoreceptor morphogenesis, specifically outer segment
formation and synapse development (Libby et al., 1997,
Consistent with the above speculations regarding an es-
sential function for ?3 containing laminins in neural tis-
sues, the chromosomal locus of LA MC3 at 9q31-q34 is
shared with four diseases having various degrees of neural
dysfunction in common: Fukuyama congenital muscular
dystrophy (FCMD); muscle-eye-brain syndrome; Walker-
Warburg syndrome; and retinitis pigmentosa-21 with deaf-
ness (RP-21). The genetic cause of the latter three of these
conditions is unclear. While ?3 expression may be affected
in FCMD, the LA MC3 cannot be the genetic cause of the
problem as a retrotransposal insertion in a different gene
has recently been identified in 87% of the FCMD alleles
(Kobayashi et al., 1998). However, LA MC3 is an excellent
candidate for mutations underlying one or more of re-
maining syndromes in this cluster of human diseases, par-
We have identified the laminin ?3 chain together with
the ?2 and ?1 chains within laminin 12 from human pla-
centa, but multiple other combinations are possible in
other tissues. It would be of particular interest if ?3 were
to associate with the ?2 chain and well as with the ?2
chain, as both these chains have been reported to show
neural and muscle-associated expression and function.
The ?3-containing laminins are likely to be the subject of
considerable interest as they constitute a novel class of
laminin molecules distributed outside of the traditional
basement membrane. The identification of the function of
this diverse family of laminins remains to be elucidated by
The authors gratefully acknowledge the excellent technical support pro-
vided by Ms. Carol Milbury, and the expert assistance of Dr. Y imin Ge
with confocal microscopy. The authors also thank Dr. Richard Kammerer
for the HisTrx and a pPEP-T vector.
This work was supported by the U.S. Public Health Service grants
A R35689 (R.E. Burgeson), EY 12037 (D.D. Hunter); the E. Matilda Zie-
gler Foundation (W.J. Brunken); and additional support from the Cutane-
ous Biology Research Center, Massachusetts General Hospital.
Received for publication 9 February 1999 and in revised form 23 March
A ebersold, R.H., J. Leavitt, R.A . Saavedra, L.E. Hood, and S.B.H. Kent. 1987.
Internal amino acid sequence analysis of proteins separated by one- or two-
dimensional gel electrophoresis after in situ protease digestion on nitrocellu-
lose. Proc. Natl. Acad. Sci. USA. 84:6970–6974.
A ltschul, S.F., W. Gish, W. Miller, E.W. Myers, and D.J. Lipman. 1990. Basic
local alignment search tool. J. Mol. Biol. 215:403–410.
A umailley, M., C. Battaglia, U. Mayer, D. Reinhardt, R. Nischt, R. Timpl, and
J.W. Fox. 1993. Nidogen mediates the formation of ternary complexes of
basement membrane components. Kidney Int. 43:7–12.
Baker, S., S. Hopkinson, M. Fitchmun, G. A ndreason, F. Frasier, G. Plopper,
V . Quaranta, and J. Jones. 1996. Laminin-5 and hemidesmosomes: role of
the alpha 3 chain subunit in hemidesmosome stability and assembly. J. Cell
Battaglia, C., U. Mayer, M. A umailley, and R. Timpl. 1992. Basement-mem-
brane heparan sulfate proteoglycan binds to laminin by its heparan sulfate
chains and to nidogen by sites in the protein core. Eur. J. Biochem. 208:359–366.
Boguski, M.S., T.M. Lowe, and C.M. Tolstoshev. 1993. dbEST-database for
“expressed sequence tags.” Nat. Genet. 4:332–333.
Burgeson, R.E., M. Chiquet, R. Deutzmann, P. Ekblom, J. Engel, H. Kleinman,
G.R. Martin, G. Meneguzzi, M. Paulsson, J. Sanes, et al. 1994. A new no-
menclature for the laminins. Matrix Biol. 14:209–211.
Chait, B.T., and S.B. Kent. 1992. Weighing naked proteins: practical, high-accu-
racy mass measurement of peptides and proteins. Science. 257:1885–1894.
Chen, L., V . Shick, M. Matter, S. Laurie, R. Ogle, and G. Laurie. 1997. Laminin
E8 alveolarization site: heparin sensitivity, cell surface receptors, and role in
cell spreading. Am. J. Physiol. 272:L494–L503.
Cohen, M., C. Jacobson, P. Y urchenco, G. Morris, and S. Carbonetto. 1997.
Laminin-induced clustering of dystroglycan on embryonic muscle cells: com-
parison with agrin-induced clustering. J. Cell Biol. 136:1047–1058.
Day, I.N., L.J. Hinks, and R.J. Thompson. 1990. The structure of the human
gene encoding protein gene product 9.5 (PGP9.5), a neuron-specific ubiq-
uitin C-terminal hydrolase. Biochem. J. 268:521–524.
Deutzmann, R., M. A umailley, H. Wiedemann, W. Pysny, R. Timpl, and D.
Edgar. 1990. Cell adhesion, spreading and neurite stimulation by laminin
fragment E8 depends on maintenance of secondary and tertiary structure in
its rod and globular domain. Eur. J. Biochem. 191:513–522.
Drago, J., V . Nurcombe, and P.F. Bartlett. 1991. Laminin through its long arm
E8 fragment promotes the proliferation and differentiation of murine neu-
roepithelial cells in vitro. Exp. Cell Res. 192:256–265.
Durkin, M., F. Loechel, M. Mattei, B. Gilpin, R. A lbrechtsen, and U. Wewer.
1997. Tissue-specific expression of the human laminin alpha5-chain, and
mapping of the gene to human chromosome 20q13.2-13.3 and to distal
mouse chromosome 2 near the locus for the ragged (Ra) mutation. FEBS
L ett. 411:296–300.
Durr, J., P. Lammi, S.L. Goodman, T. A igner, and K. von der Mark. 1996. Iden-
tification and immunolocalization of laminin in cartilage. Exp. Cell Res. 222:
Ekblom, P., and R. Timpl. 1996. The Laminins. In Cell A dhesion and Commu-
nication. V ol. 2. C. Goridis, editor. Harwood A cademic Publishers, The
Netherlands. 321 pp.
Engel, J. 1992. Laminins and other strange proteins. Biochemistry. 31:10643–
Engvall, E. 1993. Laminin variants: why, where and when? Kidney Int. 43:2–6.
Engvall, E., D. Earwicker, T. Haaparanta, E. Ruoslahti, and J.R. Sanes. 1990.
Distribution and isolation of four laminin variants; tissue restricted distribu-
tion of heterotrimers assembled from five different subunits. Cell Regul.
Fox, J.W., U. Mayer, R. Nischt, M. A umailley, D. Reinhardt, H. Wiedemann,
K. Mann, R. Timpl, T. Krieg, J. Engel, et al. 1991. Recombinant nidogen
consists of three globular domains and mediates binding of laminin to col-
lagen type IV . EMBO (Eur. Mol. Biol. Organ.) J. 10:3137–3146.
Gerecke, D.R., D.W. Wagman, M.F. Champliaud, and R.E. Burgeson. 1994.
The complete primary structure for a novel laminin chain, the laminin B1k
chain. J. Biol. Chem. 269:11073–11080.
Goodman, S.L. 1992. A lpha 6 beta 1 integrin and laminin E8: an increasingly
complex simple story. Kidney Int. 41:650–656.
Green, T.L., D.D. Hunter, W. Chan, J.P. Merlie, and J.R. Sanes. 1992. Synthe-
sis and assembly of synaptic cleft protein s-laminin by cultured cells. J. Biol.
Harlow, E., and D. Lane. 1988. A ntibodies: A Laboratory Manual. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, NY .
Haaparanta, T., J. Uitto, E. Ruoslahti, and E. Engvall. 1991. Molecular cloning
of the cDNA encoding human laminin A chain. Matrix. 11:151–160.
Hagg, T., C. Portera-Cailliau, M. Jucker, and E. Engvall. 1997. Laminins of the
adult mammalian CNS; laminin-alpha2 (merosin M-) chain immunoreactiv-
ity is associated with neuronal processes. Brain Res. 764:17–27.
Heimann, P., A . Menke, B. Rothkegel, and H. Jockusch. 1996. Overshooting
production of satellite cells in murine skeletal muscle affected by the muta-
tion “muscular dystrophy with myositis” (mdm, Chr 2). Cell Tissue Res. 283:
The Journal of Cell Biology, Volume 145, 1999 Download full-text
Henry, M.D., and K.P. Campbell. 1996. Dystroglycan: an extracellular matrix
receptor linked to the cytoskeleton. Curr. Opin. Cell Biol. 8:625–631.
Higuchi, M., T. Ohnishi, N. A rita, S. Hiraga, H. Iwasaki, S. Mori, and T. Hay-
akawa. 1991. Immunohistochemical localization of fibronectin, laminin and
fibronectin-receptor in human malignant gliomas—in relation to tumor in-
vasion. No to Shinkei. Brain Nerve (Tokyo). 43:17–23.
Hunter, D.D., and W.J. Brunken. 1997. Beta 2 laminins modulate neuronal
phenotype in the rat retina. Mol. Cell. Neurosci. 10:7–15.
Hunter, D.D., R. Llinas, M. A rd, J.P. Merlie, and J.R. Sanes. 1992a. Expression
of s-laminin and laminin in the developing rat central nervous system. J.
Comp. Neurol. 323:238–251.
Hunter, D.D., M.D. Murphy, C.V . Olsson, and W.J. Brunken. 1992b. S-laminin
expression in adult and developing retinae: a potential cue for photorecep-
tor morphogenesis. Neuron. 8:399–413.
Iivanainen, A ., K. Sainio, H. Sariola, and K. Tryggvason. 1995a. Primary struc-
ture and expression of a novel human laminin a4 chain. FEBS L ett. 365:183–188.
Iivanainen, A ., R. V uolteenaho, K. Sainio, R. Eddy, T.B. Shows, H. Sariola,
and K. Tryggvason. 1995b. The human laminin beta 2 chain (S-laminin):
structure, expression in fetal tissues and chromosomal assignment of the
LA MB2 gene. Matrix Biol. 14:489–497.
Jucker, M., P. Bialobok, T. Hagg, and D.K. Ingram. 1992. Laminin immunohis-
tochemistry in brain is dependent on method of tissue fixation. Brain Res.
Jucker, M., M. Tian, and D. Ingram. 1996a. Laminins in the adult and aged
brain. Mol. Chem. Neuropathol. 28:209–218.
Jucker, M., M. Tian, D. Norton, C. Sherman, and J. Kusiak. 1996b. Laminin al-
pha 2 is a component of brain capillary basement membrane: reduced ex-
pression in dystrophic dy mice. Neuroscience. 71:1153–1161.
Kallunki, P., K. Sainio, R. Eddy, M. Byers, T. Kallunki, H. Sariola, K. Beck, H.
Hirvonen, T.B. Shows, and K. Tryggvason. 1992. A truncated laminin chain
homologous to the B2 chain: structure, spatial expression, and chromosomal
assignment. J. Cell Biol. 119:679–693.
Kobayashi, K., Y . Nakahori, M. Miyake, K. Matsumura, E. Kondo-Iida, Y . No-
mura, M. Segawa, M. Y oshioka, K. Saito, et al. 1998. A n ancient retrotrans-
posal insertion causes Fukuyama-type congenital muscular dystrophy. Na-
Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the
head of bacteriophage T4. Nature. 227:680–685.
Langhofer, M., S.B. Hopkinson, and J.C. Jones. 1993. The matrix secreted by
804G cells contains laminin-related components that participate in hemides-
mosome assembly in vitro. J. Cell Sci. 105:753–764.
Libby, R.T., D.D. Hunter, and W.J. Brunken. 1996. Developmental expression
of laminin beta 2 in rat retina. Further support for a role in rod morphogen-
esis. Investig. Ophthalmol. Vis. Sci. 37:1651–1661.
Libby, R.T., Y . X u, L.M. Selfors, W.J. Brunken, and D.D. Hunter. 1997. Identi-
fication and cellular source of laminin ?2 in adult and developing vertebrate
retinae. J. Comp. Neurol. 389:655–667.
Libby, R.T., C. Lavalee, G.W. Balkema, W.J. Brunken, and D.D. Hunter. 1998.
Characterization of laminin ?2 in retinal function. Investig. Ophthalmol. Vis.
Liu, J., S. Swasdison, W. X ie, R.G. Brewton, and R. Mayne. 1998. Primary
structure and expression of a chicken laminin beta chain: evidence for four
beta chains in birds. Matrix Biol. 16:471–481.
Lunstrum, G.P., L.Y . Sakai, D.R. Keene, N.P. Morris, and R.E. Burgeson.
1986. Large complex globular domains of type V II procollagen contribute to
the structure of anchoring fibrils. J. Biol. Chem. 261:9042–9048.
Marinkovich, M.P., G.P. Lunstrum, D.R. Keene, and R.E. Burgeson. 1992. The
dermal-epidermal junction of human skin contains a novel laminin variant.
J. Cell Biol. 119:695–703.
Matter, M.L., and G.W. Laurie. 1994. A novel laminin E8 cell adhesion site re-
quired for lung alveolar formation in vitro. J. Cell Biol. 124:1083–1090.
Mauer, P., and J. Engel. 1996. Structure of laminins and their chain assembly.
In The Laminins. V ol. 2. P. Ekblom and R. Timpl, editor. Harwood A ca-
demic Publishers, The Netherlands. 27–50.
Mayer, U., R. Nischt, E. Poschl, K. Mann, K. Fukuda, M. Gerl, Y . Y amada, and
R. Timpl. 1979. A single EGF-like motif of laminin is responsible for high af-
finity nidogen binding. EMBO (Eur. Mol. Biol. Organ.) J. 12:1879–1885.
Mayer, U., E. Poschl, D.R. Gerecke, D.W. Wagman, R.E. Burgeson, and R.
Timpl. 1995. Low nidogen affinity of laminin-5 can be attributed to two
serine residues in EGF-like motif gamma 2III4. FEBS L ett. 365:129–132.
Miner, J.H., B.L. Patton, S.I. Lentz, D.J. Gilbert, W.D. Snider, N.A . Jenkins,
N.G. Copeland, and J.R. Sanes. 1997. The laminin a chains: expression, de-
velopmental transitions, and chromosomal locations of a1-5, identification
of heterotrimeric laminins 8-11 and cloning a novel a3 isoform. J. Cell Biol.
Miner, J.H., R.M. Lewis, and J.R. Sanes. 1995. Molecular cloning of a novel
laminin chain, alpha 5, and widespread expression in adult mouse tissues. J.
Biol. Chem. 270:28523–28526.
Mori, S., N.H. Sternberger, M.M. Herman, and L.A . Sternberger. 1992. V ari-
ability of laminin immunoreactivity in human autopsy brain. Histochemistry.
Nielsen, H., J. Engelbrecht, S. Brunak, and G. von Heijne. 1997. Identification
of prokaryotic and eukaryotic signal peptides and prediction of their cleav-
age sites. Protein Eng. 10:1–6.
Niessen, C.M., F. Hogervorst, L.H. Jaspars, A .A . de Melkar, G.O. Delwel, E.H.
Hulsman, I. Kuikman, and A . Sonnenberg. 1994. The alpha 6 beta 4 integrin
is a receptor for both laminin and kalinin. Exp. Cell Res. 211:360–367.
Noakes, P.G., M. Gautam, J. Mudd, J.R. Sanes, and J.P. Merlie. 1995. A berrant
differentiation of neuromuscular junctions in mice lacking s-laminin/laminin
beta 2. Nature. 374:258–262.
Pall, E., K. Bolton, and J. Ervasti. 1996. Differential heparin inhibition of skele-
tal muscle alpha-dystroglycan binding to laminins. J. Biol. Chem. 271:3817–
Pikkarainen, T., R. Eddy, Y . Fukushima, M. Byers, T. Shows, T. Pihlajaniemi,
M. Saraste, and K. Tryggvason. 1987. Human laminin B1 chain. A multido-
main protein with gene (LA MB1) locus in the q22 region of chromosome 7.
J. Biol. Chem. 262:10454–10462.
Pikkarainen, T., T. Kallunki, and K. Tryggvason. 1988. Human laminin B2
chain. Comparison of the complete amino acid sequence with the B1 chain
reveals variability in sequence homology between different structural do-
mains. J. Biol. Chem. 263:6751–6758.
Poschl, E., J.W. Fox, D. Block, U. Mayer, and R. Timpl. 1994. Two non-contig-
uous regions contribute to nidogen binding to a single EGF-like motif of the
laminin gamma 1 chain. EMBO (Eur. Mol. Biol. Organ.) J. 13:3741–3747.
Poschl, E., U. Mayer, J. Stetefeld, R. Baumgartner, T. Holak, R. Huber, and R.
Timpl. 1996. Site-directed mutagenesis and structural interpretation of the
nidogen binding site of the laminin gamma1 chain. EMBO (Eur. Mol. Biol.
Organ.) J. 15:5154–5159.
Reinhardt, D., K. Mann, R. Nischt, J.W. Fox, M.L. Chu, T. Krieg, and R. Timpl.
1993. Mapping of nidogen binding sites for collagen type IV , heparan sulfate
proteoglycan, and zinc. J. Biol. Chem. 268:10881–10887.
Rousselle, P., R. Golbik, M. van der Rest, and M. A umailley. 1995. Structural
requirement for cell adhesion to kalinin (laminin-5). J. Biol. Chem. 270:
Rousselle, P., D.R. Keene, F. Ruggiero, M.F. Champliaud, M. Rest, and R.E.
Burgeson. 1997. Laminin 5 binds the NC-1 domain of type V II collagen. J.
Cell Biol. 138:719–728.
Ryan, M.C., R. Tizard, D.R. V anDevanter, and W.G. Carter. 1994. Cloning of
the LamA 3 gene encoding the alpha 3 chain of the adhesive ligand epiligrin.
Expression in wound repair. J. Biol. Chem. 269:22779–22787.
Sugiyama, S., A . Utani, S. Y amada, C.A . Kozak, and Y . Y amada. 1995. Cloning
and expression of the mouse laminin gamma 2 (B2t) chain, a subunit of epi-
thelial cell laminin. Eur. J. Biochem. 228:120–128.
Tian, M., T. Hagg, N. Denisova, B. Knusel, E. Engvall, and M. Jucker. 1997.
Laminin-alpha2 chain-like antigens in CNS dendritic spines. Brain Res. 764:
Tian, M., C. Jacobson, S.H. Gee, K.P. Campbell, S. Carbonetto, and M. Jucker.
1996. Dystroglycan in the cerebellum is a laminin alpha 2-chain binding pro-
tein at the glial-vascular interface and is expressed in Purkinje cells. Eur. J.
Toti, P., C. De Felice, A . Malandrini, T. Megha, C. Cardone, and M. V illanova.
1997. Localization of laminin chains in the human retina: possible implica-
tions for congenital muscular dystrophy associated with alpha 2-chain of
laminin deficiency. Neuromuscular Disorders. 7:21–25.
V errando, P., B.L. Hsi, C.J. Y eh, A . Pisani, N. Serieys, and J.P. Ortonne. 1987.
Monoclonal antibody GB3, a new probe for the study of human basement
membranes and hemidesmosomes. Exp. Cell Res. 170:116–128.
V uolteenaho, R., M. Nissinen, K. Sainio, M. Byers, R. Eddy, H. Hirvonen, T.B.
Shows, H. Sariola, E. Engvall, and K. Tryggvason. 1994. Human laminin M
chain (merosin): complete primary structure, chromosomal assignment, and
expression of the M and A chain in human fetal tissues. J. Cell Biol. 124:
Wewer, U., and E. Engvall. 1996. Merosin/laminin-2 and muscular dystrophy.
Neuromuscular Disorders. 6:409–418.
Wewer, U.M., D.R. Gerecke, M.E. Durkin, K.S. Kurtz, M.G. Mattei, M.F.
Champliaud, R.E. Burgeson, and R. A lbrechtsen. 1994. Human beta 2 chain
of laminin (formerly S chain): cDNA cloning, chromosomal localization, and
expression in carcinomas. Genomics. 24:243–252.
X u, H., X .R. Wu, U.M. Wewer, and E. Engvall. 1994. Murine muscular dystro-
phy caused by a mutation in the laminin alpha 2 (Lama2) gene. Nat. Genet.
Y amamoto, T., N. Shibata, M. Kanazawa, M. Kobayashi, T. Komori, K. Ikeya,
E. Kondo, K. Saito, and M. Osawa. 1997. Localization of laminin subunits in
the central nervous system in Fukuyama congenital muscular dystrophy: an
immunohistochemical investigation. Acta Neuropathol. 94:173–179.
Y bot-Gonzalez, P., S. Runswick, N. Smyth, and D. Edgar. 1995. Regulated ex-
pression of a novel laminin beta subunit during the development of the chick
embryo. Differentiation. 59:215–223.
Y urchenco, P.D., and Y .S. Cheng. 1993. Self-assembly and calcium-binding
sites in laminin. A three-arm interaction model. J. Biol. Chem. 268:17286–
Y urchenco, P.D., and Y .S. Cheng. 1994. Laminin self-assembly: a three-arm in-
teraction hypothesis for the formation of a network in basement mem-
branes. Contrib. Nephrol. 107:47–56.
Y urchenco, P.D., Y .S. Cheng, and H. Colognato. 1992. Laminin forms an inde-
pendent network in basement membranes [published erratum appears in J.
Cell Biol. 118:493]. J. Cell Biol. 117:1119–1133.