WO1995008628A2 - Nucleic acids encoding merosin, merosin fragments and uses thereof - Google Patents

Abstract

This invention provides an isolated nucleic acid molecule encoding a subunit of a protein, the protein having an apparent molecular weight of about 800 kD, designated merosin. Also provided are isolated nucleic acid molecules which encode merosin fragments. Anti-merosin antibodies, vectors for the recombinant production of merosin, and the expression of recombinant proteins by use of a host vector system also are provided. The invention further provides the use of merosin to promote neurite growth and for certain diagnostic applications.

Description

NUCLEIC ACIDS ENCODING MEROSIN. MEROSIN FRAGMENTS AND USES THEREOF

The present invention was supported by grants DK

30051, CA 45546, CA 28896 and Cancer Center Support Grant CA30199 from the National Institute of Health. The United

States Government may have certain rights in this invention.

BACKGROUND OF THE INVENTION

This invention relates generally to basement membranes and specifically to a novel tissue-specific basement membrane-associated protein.

Basement membranes are thin sheets of extracellular matrix separating epithelial cells from underlying tissue stroma. They compartmentalize epithelial and endothelial organs and maintain tissue structures. In some tissues the basement membrane is a product of the interaction of several cell types; for example, the glomerular basement membrane is made by both epithelial and endothelial cells. In skeletal muscle, fibroblasts from the endomysium contribute type IV collagen to the assembly of the basement membrane. The formation of the neural basal lamina requires the interaction of Schwann cells and neurons. Further, basement membranes function in development and tissue repair by promoting attachment, migration and proliferation of cells and by mediating signals for tissue interactions.

All basement membranes contain laminin, type IV collagen, entactin and heparan sulfate proteoglycan. Laminin is a large glycoprotein composed of three polypeptide chains, a 400 kD A chain and two B chains of about 200 kD each. The amino-terminal two thirds of the A chain is homologous to the Bl and B2 chains while the carboxy-terminal third has a distinct structure. Recent studies have revealed that several genetically distinct subunit chains and consequently several laminin isoforms exist. In addition to the EHS laminin chains, A, Bl and B2, merosin (also known as laminin M chain), a homologue of the A chain (Leivo et al. , Proc. Natl. Acad. Sci. USA 85:1544-1548 (1988); Ehrig et al. , Proc. Natl. Acad. Sci. USA 87:3264-3268 (1990)), s-laminin (S chain), a homologue of the Bl chain (Hunter et al . , Nature 338:229- 234 (1989)) and B2t, a truncated homologue of the B2 chain (Kallunki et al. , J. Cell Biol. 119:679-693 (1992)), have been characterized. Recently partial sequence of another Bl chain variant in avian eye was reported (O'Rear et al. , J. Biol. Chem. 267:20555-20557 (1992)) . K-laminin and kalinin are laminin isoforms that are present in epithelial basement membranes. K-laminin contains the Bl and B2 chains and has a third 190 kD chain immunologically distinct from the A chain (Marinkovich et al. , J. Cell Biol. 119:695-703 (1992)) . Kalinin has three subunits of which the largest one is immunologically related to one chain of K-laminin (Rouselle et al. , J. Cell. Biol. 114:567-576 (1991) ; Marinkovich et al. , J. Biol. Chem. 267:17900-17906 (1992)) . For terminology of the laminins, see Engvall, 1993, Kidney International 43:2-6, which is incorporated herein by reference, and Figure 13.

Laminin promotes attachment, spreading, motility and growth of a variety of cell types. One of the most striking features of laminin is its capacity to promote outgrowth of neurites from cultured neuronal cells. A major site of cell adhesion and the neurite-promoting activity appear to reside in the globular domain at the end of the long arm of this molecule.

The metastatic propensity of certain tumor cells may also be influenced by laminin. For example, laminin has been shown to mediate the attachment of malignant carcinoma cells to type IV collagen and to increase the metastatic potential of murine melanoma cells. Other basement membrane proteins and their receptors may be involved in the adhesion of metastasizing tumor cells to basement membranes of blood vessels and other epithelial tissues.

Because of the critical role of basement membranes in development, tissue repair, neurite growth and cancer, there exists a need for the identification of new basement membrane components. The present invention satisfies this need.

SUMMARY OF THE INVENTION

This invention provides an isolated nucleic acid molecule encoding a 380-400 KDa subunit of the protein merosin. Also provided are isolated nucleic acid molecules which encode merosin fragments. The invention further provides antibodies, vectors, and the expression of recombinant proteins by use of a host/vector system. The invention also provides the use of merosin to promote neurite growth.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1 shows the DNA sequence of a partial merosin polypeptide cDNA and the deduced amino acid sequence. Potential N-glycosylation sites are indicated by (A) and cysteines are circled. Sequences obtained by amino acid sequencing are underlined. Conserved motifs of amino acid sequence are boxed.

Figure 2 shows a comparison of the amino acid sequences of merosin fragment and the COOH-terminal portion of the mouse laminin A chain by dot matrix plotting. Sequences were compared using the Micro Genie matrix comparison program. The frame was set at eight amino acids with a minimal match of 40%. Figure 3 shows immunoblotting of placental extract with antiserum. NaDodS0₄ extract of placenta (lanes 1) and the purified fragment of merosin polypeptide from a pepsin digest of placenta (lanes 2) were electrophoresed on a 2- 16% gradient acrylamide gel in the presence of NaDodS0₄ and transferred to nitrocellulose. The blot in (a) was stained with a peptide antiserum raised to a 13-amino acid peptide corresponding to residues 476-488 of the merosin polypeptide in Figure 1. The blot in (b) was stained with monoclonal antibody that recognizes COOH-terminal fragments of merosin polypeptide. For comparison, a blot of mouse laminin was stained with anti-laminin (c) . Arrowhead shows the position of the top of the separating gel and numbers

(KDa) indicated the positions of molecular weight markers.

Figures 4A through C show an analysis of intact merosin from placenta. Figure 4A: NaDodS0₄-polyacrylamide gel electrophoresis of rat laminin (lane 1) and the merosin-containing fraction from human placenta (lane 2) . Positions of molecular weight markers are shown on the left. Figure 4B: Electron microscopy after rotary shadowing of the merosin-containing preparation. Figure 4C: ELISA in microtiter wells coated with the merosin- containing preparation and in wells coated with the large pepsin fragment of laminin. The antibodies were 3E5 (■; anti-Bl), 2E8 (•; anti-B2) , 11D5 (Δ_; anti-A), and 2G9 (A; anti-merosin) .

Figure 5 shows the relative positions of the sequences of cDNA clones for human merosin polypeptide ("laminin M chain"), partial restriction maps and domain structure of the merosin subunit protein. At top, alignment of five overlapping cDNA clones and partial restriction maps of merosin cDNA. ATG indicates the translation initiation signal, and TGA the 3'-end translation stop codon. Restriction enzyme sites EcoRI (E) , Hind III (H) and Pst I (P) are shown. Middle, structure of the protein with domains numbered according to Sasaki et al. , Proc. Natl. Acad. Sci. USA 84:935-939 (1987); Sasaki et al. , J. Biol. Chem. 263:16536-16544 (1988); Sasaki et al. , J. Biol. Chem. 262:17111-17117 (1987) ) , incorporated herein by reference.] Five internal repeats in domain G are indicated by hatched boxes. Domains Ilia, Illb and V consisting of cysteine- rich EGF modules are shown by shaded boxes. Bottom, scale in amino acids (aa) .

Figure 6 shows the complete nucleotide sequence of human merosin cDNA clones and deduced complete amino acid sequence of the entire protein. First line, nucleotide sequence of cDNA clones characterized in this study. Second line, deduced amino acid sequence from the cDNA clones together with the previously determined carboxyl terminal end amino acid sequence (Ehrig et al. , Proc. Natl. Acad. Sci. USA 87:3264-3268 (1990)), incorporated herein by reference. The putative signal peptidase cleavage site is indicated by a triangle. The cysteine residues are circled, and potential attachment sites for asparagine- linked oligosaccharides are boxed.

Figure 7 is an alignment of amino acid sequences of the M (merosin) and A chains of human laminin-type proteins. The upper line shows the amino acid sequence of merosin, and the second line shows the amino acid sequence of the laminin A chain. Both amino acid sequences are numbered from the initiator methionine. All cysteines are circled and N-glycosylation sites are underlined. The structural domains are boxed and indicated by Roman numerals on the right. SP = signal peptide.

Figure 8 shows the chromosomal localization of the merosin encoding sequences. The idiogram of chromosome 6 shows the distribution of signals on that chromosome and assignment of the merosin gene to 6q22->23. Figure 9 shows expression of merosin and laminin A chain mRNA in 17-week-old human fetal tissues. Gene Screen Plus filter containing total RNA (-10 μg) was prepared and hybridized as described Example V below. Ethidium bromide (EtBr) staining of the filter (bottom) is shown to illustrate the relative amounts of RNA in each lane.

Figures 10A through 10H show in si tu hybridization of merosin mRNA in 17-week-old fetal tissues. In kidney

(Figures 10A and B) signals are seen in mesenchymal cells adjacent to condensing pretubular cells and ureter-derived tubules (t) in the outer cortex. Secretory tubules of the nephron and blood vessels are negative. In heart muscle

(Figures IOC and D) signals can be observed in cardiomyocytes throughout the muscle. In sections of skin (Figures 10E and F) no grains are seen over the epithelial cells of epidermis (e) , while strong signal can be observed in the condensing papillary mesenchymal cells (p) and a developing hair follicles (f) . In lung (Figures 10G and H) signals are present in smooth muscle cells of the peribronchial arterial wall, but alveolar and bronchial cells are negative. Bar A-D 200 μm and E-H 100 μm.

Figure 11 is an alignment of domains VI of the known human A - and B-type laminin chains, the rat S chain, the mouse A chain and the Drosophila A chain. Amino acids that are identical in half of the chains are shaded, and dark shading indicates conserved change Phe (F) <-> Tyr (Y) . Abbreviations: Bl, human Bl chain; S, rat S chain; A, human A chain; mA, mouse A chain; M, human M chain; B2, human B2 chain; dA, Drosophila A chain.

Figure 12 is an alignment of domains V of the known human A- and B-type chains, the rat S chain, the mouse A chain and the Drosophila A chain. Amino acids that are identical in half of the chains are shaded, and dark shading indicates conserved change Phe (F) <-> Tyr (Y) . Abbreviations: Bl, human Bl chain; S, rat S chain; A, human A chain; mA, mouse A chain; M, human M chain; B2, human B2 chain; dA, Drosophila A chain.

Figure 13 is a diagra atic scheme of the structure of the laminins.

DETAILED DESCRIPTION OF THE INVENTION

This invention provides a cDNA molecule encoding the major subunit of human merosin protein which is structurally related to laminin. The merosin protein has an apparent molecular weight of about 800 kd and is composed of four polypeptides having apparent molecular weights of 300, 200, 200 and 80 kD, the 300 kD polypeptide being joined to the 200 kD polypeptides by disulfide bonds, and the 300 kD and 80 kD polypeptides comprising the 380- 400 KDa merosin subunit having substantially the amino acid sequence shown in Figure 6. Merosin is found in placenta, striated muscle, peripheral nerve, trophoblasts and human Schwann cell neoplasms, among other tissues.

Leivo et al. , Proc. Natl. Acad. Sci. USA 85:1544-1548 (1988) , which is incorporated herein by reference, describes the isolation of a 65-KDa and an 80 KDa segment of the basement-membrane-associated polypeptide merosin. These two precursor segments, the full length merosin polypeptide, fragments of the merosin polypeptide, and proteins comprising any of these segments, polypeptide, or fragments have also been termed merosin. Because the 65 KDa and 80 KDa proteins appear to be segments of the 380- 400 KDa merosin polypeptide contained within an 800 KDa protein complex, the term merosin has now also been applied to the 800 KDa protein described herein. The 380-400 KDa subunit is designated merosin polypeptide, merosin subunit, M chain, or laminin M chain. It is understood that limited modifications may be made to the primary sequence of merosin subunit without destroying its biological function, and that only a portion of. the entire primary structure may be required in order to effect activity. One such biological active fragment is a molecule having substantially the sequence shown in Figure 1. In a separate embodiment of the invention, the merosin subunit has an amino acid sequence substantially similar to that shown in Figure 6. Minor modifications of these sequences which do not destroy the activity of the proteins also fall within the definition of merosin and within the definition of the protein claimed as such. Moreover, fragments of the sequences of Figures 1 or 6, but not a fragment consisting solely of the previously described 80 Kd fragment, which retain the function of the entire protein, as determined by the merosin activity assay described in Example II below, and as defined by the protein's ability to elicit merosin-specific antibodies are included within the definition. It is understood that minor modifications of primary amino acid sequence may result in proteins which have substantially equivalent or enhanced function as compared to the sequences set forth in Figures 1 or 6. These modifications may be deliberate, as through site-directed mutagenesis, or synthesis of merosin analogs, or may be accidental such as through mutation in hosts which are merosin producers. All of these modifications are included as long as merosin biological function is retained. The nucleic acid sequences shown in Figures 1 and 6 are useful in the production of recombinant merosin and merosin fragments. Nucleic acid fragments of at least 10 nucleotides are also useful as hybridization probes. The probes are useful to identify tissue (as set forth in more detail below) to isolate the genomic gene encoding merosin, which has now been localized to chromosome 6q22->23, or to identify nucleic acid encoding merosin-like proteins. The isolated nucleic acid fragments also are useful to generate novel peptides. These peptides, in turn, are useful as immunogens for the generation of polyclonal and monoclonal antibodies. Methods of preparing and using the probes and immunogens are well known in the art, and are briefly described below.

Also included within the scope of this invention are nucleic acid molecules that hybridize under stringent conditions to the nucleic acid molecules, the sequences of which are shown in Figures 1 and 6. Such hybridizing nucleic acid molecules or probes, can by prepared, for example, by nick translation of the nucleic acid molecules of Figures 1 or 6, in which case the hybridizing nucleic acid molecules can be random fragments of the molecules, the sequences of which are shown in Figures 1 and 6. For methodology for the preparation of such fragments, see Sambrook et al. , Molecular Cloning: A Laboratory Manual Cold Spring Harbor Press, Cold Spring Harbor, N.Y. (1989) , incorporated herein by reference. As used herein, "nucleic acid" shall mean single and double stranded DNA and RNA.

Further, various molecules can be attached to merosin subunit, for example, other proteins, carbohydrates, or lipids. Such modifications are included within the definition of merosin.

"Purified", when used to describe the state of merosin, denotes the protein free of a portion of the other proteins and molecules normally associated with or occurring with merosin in its native environment. As used herein the term "native" refers to the form of a protein, polypeptide, antibody or a fragment of thereof that is isolated from nature or that which is without an intentional amino acid substitution.

As used herein, the term "antibody" or "immunoglobulin" refers to a protein that is produced in response to immunization with an antigen and specifically reacts with the antigen. This includes polyclonal as well as monoclonal antibodies. Human and mammalian, for example, mouse, rat, rabbit and goat, are intended to be included in this definition. The most predominant human antibody produced is of the IgG isotype, having two light and two heavy chains linked by disulfide bonds, which constitute about 80% of total serum antibodies.

As used herein, "antibody" also encompasses fragments of antibodies. The antibody fragments retain at least some ability to selectively bind with its antigen. Also encompassed by this invention are antibody fragments that have been recombinantly or chemically synthesized that retain the ability to bind the antigen of the corresponding native antibody. The ability to bind with an antigen or hapten is determined by antigen-binding assays known in the art such as antibody capture assays (See, for example, Harlow and Lane, Antibodies, A Laboratory Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, NY (1988) ) . Antibody fragments retaining some binding affinity include, but are not limited to: Fab (the fragment which contains a monovalent antigen-binding fragment of an antibody molecule produced by digestion with the enzyme papain to yield an intact light chain and a portion of one heavy chain) ; Fab' (the fragment of an antibody molecule obtained by treating with pepsin, followed by reduction, to yield an intact light chain and a portion of the heavy chain; two Fab' fragments are obtained per antibody molecule); (Fab')₂, the fragment of the antibody that is obtained by treating with the enzyme pepsin without subsequent reduction; F(ab')₂ is a dimer of two Fab' fragments held together by two disulfide bonds; Fv and single chain antibodies (SCA) .

"Isolated" when used to describe the state of the nucleic acids encoding merosin, denotes the nucleic acids free of at least a portion of the molecules associated with or occurring with nucleic acids in the native environment.

"Recombinant expression vector" includes vectors which are capable of expressing DNA sequences contained therein, where such sequences are operatively linked to other sequences capable of effecting their expression. It is implied, although not always explicitly stated, that these expression vectors must be replicable in the host organisms either as episomes or as an integral part of the chromosomal DNA. In sum, "expression vector" is given a functional definition, and any DNA sequence which is capable of effecting expression of a specified DNA sequence disposed therein is included in this term as it is applied to the specified sequence. In general, expression vectors of utility in recombinant DNA techniques are often in the form of "plasmids" which refer to circular double stranded DNA loops which, in their vector form, are not bound to the chromosome. In the present specification, "plasmid" and "vector" are used interchangeably as the plasmid is the most commonly used form of vector. However, the invention is intended to include such other forms of expression vectors which serve equivalent functions and which become known in the art subsequently hereto.

"Host-vector system" refers to cells which have been transfected with vectors constructed using recombinant DNA techniques. The vectors and methods disclosed herein are suitable for use in host cells over a wide range of procaryotic and eucaryotic organisms.

Reference is made to standard textbooks of molecular biology that contain definitions and methods and means for carrying out basic techniques, encompassed by the present invention. See, for example, Maniatis et al. , Molecular

Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, New York (1982) and Sambrook et al . , Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory, New York (1989) and the various references cited therein. This reference and the cited publications are expressly incorporated by reference into this specification.

In addition, recombinant DNA methods currently used by those skilled in the art include the polymerase chain reaction (PCR) which, combined with the synthesis of oligonucleotides, allows easy reproduction of DNA sequences. A DNA segment of up to approximately 6000 base pairs in length may be amplified exponentially starting from as little as a single gene copy by means of PCR. In this technique a denatured DNA sample is incubated with two oligonucleotide primers that direct the DNA polymerase- dependent synthesis of new complementary strands. Multiple cycles of synthesis each afford an approximate doubling of the amount of target sequence. Each cycle is controlled by varying the temperature to permit denaturation of the DNA strands, annealing the primers, and synthesizing new DNA strands. The use of a thermostable DNA polymerase eliminates the necessity of adding new enzyme for each cycle, thus permitting fully automated DNA amplification. Twenty-five amplification cycles increase the amount of target sequence by approximately 10⁶-fold. The PCR technology is the subject matter of United States Patent Nos. 4,683,195, 4,800,159, 4,754,065, and 4,683,202 all of which are hereby incorporated by reference.

With regard to the present invention, the cDNA shown in Figures 1 or 6, or any portion of them can be reproduced for cloning and expression purposes by amplifying the desired sequence with PCR and cloning it into a suitable vector as is well known in the art. Detection methods for the presence of nucleic acid or protein in cells include hybridization of a nucleic acid probe with the nucleic acid of a cell and cell staining with polyclonal or monoclonal antibodies. Such techniques are accomplished by methods well-known to those skilled in the art. See, for example, Harlow and Lane, Antibodies: A Laboratory Manual, Cold Spring Harbor, 1988, hereby incorporated by reference.

Monoclonal and polyclonal antibodies against merosin were prepared according to procedures well known in the art . The specificity of the antibodies is examined by carrying out enzyme immunoassays and immunoblotting of placental extracts.

For example, monoclonal antibodies are prepared by immunizing an animal with material containing the protein, such as an extract of human placenta tissue, followed by isolating antibody-producing hybridoma cells, as is well known in the art. (See, for example, Harlow and Lane, Antibodies: A Laboratory Manual, supra, and the references cited therein, all which are incorporated by reference into this specification. ) Anti-merosin antibodies are selected by performing immunofluorescence analysis of tissue sections where merosin is localized in the basement membranes of trophoblasts, striated muscle and Schwann cells, and other sites. The identification of antibodies is confirmed by immunoblotting and immunoprecipitation which reveals one or more of the polypeptides described above. The appropriate hybridoma is reactive with purified merosin subunit or merosin fragments. Merosin fragments can be prepared by expressing the merosin cDNA shown in Figure 1, or alternatively, subjecting the cDNA molecules, the sequences of which are shown in Figure 1 and 6, to restriction enzyme digestion and subsequent purification of the restriction enzyme fragments. These methods are well known to those of skill in the art, Sambrook et al. , supra. hereby incorporated by reference. The nucleic acid fragments are then expressed in a procaryotic or eucaryotic expression vector as described above.

Alternatively, anti-merosin antibodies can be prepared by immunizing an animal with synthetic peptides or recombinant protein fragments prepared from molecules having the sequences shown in Figures 1 or 6, or from restriction enzyme fragments, described above. One molecule demonstrated to be suitable for antibody production is the molecule having the sequence shown in Figure 1. A synthetic peptide suitable for antibody production is described in Example I. Selection of anti- merosin antibodies is performed as described above.

The COOH-terminal portion of merosin is structurally related to the COOH-terminus of the laminin A-chain. However, the amino acid sequence of merosin is 61% and 62% different from the homologous portions of mouse and human laminin A chains, respectively. Affinity purified antibodies stain two bands, suggesting that the merosin polypeptide is processed into two fragments of approximately 300 kD and 80 kD respectively.

cDNA clones for merosin A chain were isolated from a human placental lambda gtll cDNA expression library using affinity purified antibodies specific for merosin. Two cDNA clones, designated 271 and 225, with inserts of 3.6 and 1.7 kb respectively were selected for sequencing. The nucleic acid sequence of the cDNA revealed a 3.4 kb open reading frame followed by a 155 bp untranslated 3' region. The cDNA and deduced amino acid sequences are shown in Figure 1. NH₂-terminal amino acid sequences of the fragments isolated from peptic or chymotryptic digests of placenta, and the NH₂-terminal amino acid sequences of a 16 kD fragment generated with thrombin were contained within the deduced sequence, thus defining the clones as merosin cDNA. RNA blot analysis revealed a single transcript of about 10 kb in human placental RNA.

The deduced partial sequence of merosin comprises 1130 amino acids and contains 13 potential sites of N- glycosylation. The sequence includes five repeats of about 190 amino acids. These repeats contain a conserved seven amino acid long sequence, LFVGGLP or variations thereof. This is followed 17-21 and 40-43 residues later by cysteines most of which are preceded by glycines. The average identity among the five repeats is about 25%.

Comparative analysis of the amino acid sequence of merosin with known proteins revealed a striking similarity to the mouse and human laminin A chains. No other significant similarities were found upon search of the data banks. The five repeats of merosin are also present in the COOH-terminal portion of the laminin A chain. The overall identity between the merosin sequence in Figure 1 and the corresponding portion of the mouse laminin A chain is 39%.

The partial cDNA clone, the sequence of which is provided in Figure 1, was used to isolate the full length sequence encoding merosin polypeptide. Several libraries were made from human placental poly(A) RNA and probed with merosin-encoding sequences. Five overlapping cDNA inserts were pieced together to generate the full length sequence, which is shown in Figure 6.

The human M chain is 30 residues longer than the human A chain, which contains 3058 residues. Comparison of the two sequences demonstrates that the domain structure of the M chain is similar to that of the A chain, and these two laminin heavy chains have considerable homology. The overall sequence similarity is 46.6%, and 58.6% when conservative changes are included. Expression of the M and A chain genes was compared by

Northern hybridization; in si tu hybridization was also conducted for the M chain in human fetal tissue. Both procedures confirmed the different tissue expression pattern of these polypeptides.

It has further been discovered that malignant tumors have an insubstantial amount of merosin compared to non- malignant tumors. The precise amount of merosin depends on the specific tumor and can be determined by one skilled in the art given the teaching of this invention.

The following examples are intended to illustrate but not to limit the invention.

EXAMPLE I Purification of Merosin

Screening of cDNA Library

A human placental cDNA library in lambda gtll was screened using affinity purified antibodies to the denatured 65 kD chymotrypsin fragment of merosin as described in Leivo and Engvall, supra. The identity of the isolated cDNA clones was confirmed immunologically following the procedure described by Argraves et al. , J. Cell Biol. 105, 1183-1190 (1987) which is incorporated herein by reference.

Determination and Analysis of cDNA Sequences

Two cDNA clones, designated 271 and 225, with inserts of 3.6 and 1.7 kilobases, respectively, were selected for sequencing. Multiple overlapping fragments were sequenced. Nonoverlapping fragments were sequenced in both directions. Alignment of the fragments that were cloned and sequenced is summarized in Figure 1. cDNA inserts were cleaved with various restriction enzymes, and fragments subcloned into either M13mpl9(+) (Bethesda Research Laboratories, Gaithersburg, MD) or Bluescript SK M13 (+) (Stratagene Cloning Systems, La Jolla, CA) . Nucleic acid sequencing was done by the dideoxy chain termination method of Sanger et al. using deoxyadenosine 5' -a- [³⁵S]thiophosphate (New England Nuclear, Boston, MA) and a kit from USB (Cleveland, OH) . Some areas were sequenced using 15-base oligonucleotide primers synthesized using a DNA synthesizer (Applied Biosystems, Foster City, CA) . Sequence analysis was done using the MicroGenie program (Beckman) . Homology searches were carried out using Bionet with EMBL, Genbank, NBRF/PIR and Swiss-Prot databases.

The nucleic acid sequence of the cDNA revealed a 3.4- kilobase open reading frame followed by a 155-base-pair 3' untranslated region. The deduced amino acid sequence is shown in Figure 1. The NH₂-terminal amino acid sequence of fragments isolated from peptic or chymotryptic digests of placenta and the NH₂-terminal amino acid sequence of a 16- kDa fragment generated with thro bin were contained within the deduced sequence, thus defining the clones as merosin cDNA.

The deduced partial sequence of merosin comprises 1130 amino acids and contains 13 potential sites of N- glycosylation. The sequence includes five repeats of about 190 amino acids. These repeats contain a conserved 7-amino acid sequence, Leu-Phe-Val-Gly-Gly-Leu-Pro, or variations thereof (Figure 1) . This is followed 17-21 and 40-43 residues later by cysteines most of which are preceded by glycines. The average percentage of identity among the five repeats is about 25%. Protein Sequencing

A 55 kD merosin fragment was isolated from a pepsin digest of human placenta using monoclonal antibody affinity chromatography as described in Leivo and Engvall, supra. The pepsin fragment of merosin was digested further with thrombin and a 16 kD fragment was selected for sequence analysis. The merosin fragments were electrophoretically separated on a 10 to 20% gradient polyacrylamide gel in the presence of NaDodS0₄, blotted onto polyvinylidene difluoride membranes (Millipore, Boston, MA) and sequenced on an Applied Biosysterns sequenator as described by Matsudaira, J. Biol. Chem. 262 10035-10038 (1987) incorporated by reference herein.

Synthetic Peptides, Antibody Production, and Immunoblotting

The length of the open reading frame of the merosin cDNA indicated that the mature merosin polypeptide was much larger than the 80 kD fragment identified originally in placental extracts. The deduced amino acid sequence suggested that the 65 kD fragment and the 80 kD tissue polypeptide are COOH-terminal fragments of merosin. The missing portion of the intact merosin polypeptide was identified after synthesizing two 13-amino acid long peptides from the part of the deduced amino acid presumed to be NH₂-terminal of the 80 kD fragment (residues 475-488 and 457-469 in Figure 1) . Two 13 amino acid long peptides CNNFGLDLKADDKI and CSIVDIDTNQEENI were synthesized based on amino acid sequences deduced from the cDNA sequence. The cysteine at the NH₂-terminus of these peptides was added to facilitate coupling to carrier protein. The peptides were coupled to keyhole limpet hemocyanin using m- maleimidobenzoyl-N-hydroxysuccinimide ester (Pierce

Chemical Co., Rockford, IL) according to 0'Sullivan et al.

(Anal. Biochem 100, 100-108 (1979) incorporated by reference herein. The resulting conjugates were emulsified in Freund's complete adjuvant and injected into rabbits. Boosting immunizations of the conjugate in Freund's incomplete adjuvant were provided one and two months later. The dose of each injection was equivalent to 0.6 mg of peptide. Blood was collected 10 days after the third injection. The antisera obtained were tested against the glutaraldehyde-cross linked peptides in ELISA and against NaDodS0₄ extracts of tissue and isolated proteins in immunoblotting as described in Leivo and Engvall, supra.

Immunization of rabbits with these peptides resulted in antisera which, in immunoblotting, stained a polypeptide of about 300 kD in NaDodS0₄-extracts of placenta. This anti-peptide antisera did not react with the 80 kD or the 65 kD COOH-terminal fragments of merosin. The presence of the 80 kD fragment in the same extract was revealed by a monoclonal antibody (Figure 3B, lane 1) . Antibodies affinity purified from the anti-peptide antiserum on immobilized peptide also stained the 300 kD band. The other peptide antiserum and preimmune sera did not give any staining in immunoblotting. These results suggest that the merosin polypeptide is processed into two fragments of approximately 300 kD and 80 kD, respectively.

Isolation of Intact Merosin from Placenta

Merosin was then isolated using methods previously employed in the isolation of laminin from mouse tissues, Paulsson et al. , Eur. J. Biochem, 166:11-19 (1987) incorporated by reference herein. These methods are based on the selective solubilization of laminin from basement membranes with EDTA-containing buffers. When human placenta was sequentially extracted with a neutral buffer and with the same buffer containing EDTA, merosin antigenic activity was found mainly in the EDTA extract. Merosin could be precipitated from the extract with either 4 M NaCl or 40% saturated ammonium sulphate. Upon gel filtration on Sepharose 6B, merosin antigenic activity eluted in the void volume peak. It bound to DEAE cellulose and was eluted at about 0.2 M NaCl.

Figure 4 shows NaDodS0₄-polyacrylamide gel electrophoresis, electron microscopy after rotary shadowing, and ELISA analysis of the peak merosin- containing fraction from DEAE-cellulose chromatography. The predominant component in this fraction had a molecular weight of about 700 kD, slightly smaller than the 800 kD rat laminin, as determined by gel electrophoresis (Figure 4A) . After reduction with mercaptoethanol, the merosin fraction contained polypeptides of about 600 kD, 300 kD, and 180-200 kD in addition to some minor components of 60- 90 kD (Figure 4A) . The synthetic peptide antiserum bound to the 500-600 kD and 300 kD bands in immunoblotting. Antibodies against the COOH-terminal fragment of merosin bound to an 80 kD band.

Electron microscopy after rotary shadowing was used to further characterize the merosin fraction. Cross-shaped images strongly resembling mouse and rat laminin were the predominant structures seen (Figure 4B) .

Analysis of the fraction by ELISA with merosin- specific and laminin subunit-specific monoclonal antibodies showed that the preparation contained the merosin polypeptide and the laminin Bl and B2 light chains. No reactivity was obtained with laminin heavy chain-specific antibodies (Figure 4C) . The truncated pepsin fragment of laminin, isolated with laminin heavy chain-specific monoclonal antibody, reacted with antibodies specific for the heavy chain as well as with antibodies specific for the Bl and B2 chains. This laminin preparation did not react with merosin antibodies (Figure 4C) . These results show that the high molecular weight, laminin-like molecule isolated from EDTA-extracts of placenta contained no detectable laminin heavy chain but contained laminin light chains associated with the merosin heavy chain.

EXAMPLE II Merosin Activity

Merosin Promotes Cell Attachment

Cell attachment promotion by merosin was determined by methods well known in the art and set forth in Engvall and Ruoslahti, Collagen Rel. Res., 3:359-369 (1983) hereby incorporated by reference. Briefly, polystyrene microtiter plates (Flow Laboratories, Irvine, CA) were coated with various proteins by incubating the wells with 100 μl of different concentrations of the protein in PBS for 3-16 h at room temperature. Nonbound protein was removed by three washes in PBS. In some experiments, the wells with protein solution were air dried at 37°C and then washed. Cells were trypsinized and washed twice with 0.5 mg/ml soy bean trypsin inhibitor in EMEM. A suspension of approximately 250,000 cells per ml EMEM with 10 mM HEPES was prepared and 0.1 ml was added to each well already containing 0.1 ml EMEM. The plate was then incubated at 37°C for 30-90 min in an atmosphere of 10% C0₂ in air. Cell attachment was evaluated by one or more of the following methods: 1) Nonattached cells were removed and counted; 2) attached cells were fixed, stained with toluidine blue, and counted using an Artek cell counter (Dynatech Corporation,

Alexandria, VA) ; or 3) the light absorbed by the fixed and stained cells was measured using an automatic ELISA reader

(Multiscan, Flow Laboratories) . When laminin was tested in solution, it was serially diluted in the plate with a solution of 1 mg/ml BSA in EMEM containing 10 mM HEPES before adding the cells. All assays were done with samples in triplicates. The cell lines in Table 1 have been tested for cell attachment to merosin. Successful attachment is indicated as a "+." The better the attachment the more "+'s."

Table 1 Degree of Attachment Cell Line Merosin Laminin

JAR, Chonocarcinoma - ++

Endothelial Cells - +++

SKLMS, Muscle ++ +++ MG63, Osteosarcoma +++ +++

U251, Glioma +++ +++

JMR 32, Neuroblastoma +++ +++

The results show that merosin promotes attachment by many but not all types of cells.

Merosin Promotes Neurite Outgrowth

Neurite promoting activity by merosin was determined by known methods as set forth in Engvall et al. , J. Cell Biol., 103:2457-2465 (1986) and Manthorpe et al. , A Dissection and Tissue Culture, Manual of the Nervous System, 322-326 (1989), Alan R. Liss, Inc., both of which are hereby incorporated by reference. Briefly, embryonic day 8 chick ciliary ganglion neuronal cultures were used. Polyornithine-coated tissue culture plastic wells (6-mm diameter, 96-well icroplates) were treated with 5 μg/ml of human laminin or merosin in PBS for 2-3 h at 37°C. The wells were washed once with 100 μl PBS containing 1% BSA. 100 μl culture medium (Dulbecco's modified Eagle's basal medium supplemented with 0.5% BSA, 8 X IO^"7 M insulin, 3.3 X IO^"2 M glucose, 2.6 x IO^"2 M NaHC0₃, 2 X IO^-3 M L-glutamine, 100 μm/ml penicillin, and 100 trophic units/ml ciliary neuronotrophic factor) containing 1,000 neurons was added. Cultures were fixed after 3 h by the addition of 200 μl 2% glutaraldehyde for 20 in., washed with water, and stained with 0.1% toluidine blue in water. About 150 neurons were observed microscopically for each culture condition. Neurons were recorded as neurite-bearing if they possessed at least 50 μm of total neurite length.

In addition, surfaces were coated with 100 μg/ml polyoruithine (PORN) for attachment. 25 μg/ml laminin or merosin were then added for neurite outgrowth. Cells were allowed to extend neurites for 72 hours. The degree of promotion is set forth in table 2. Promotion of neurite growth is indicated as a ^■■+.^■■ The greater the promotion, the more "+' s. "

Table 2

No Protein Laminin Merosin

No Porn - - -

Porn + +++ +++

The results show merosin is a promotor of neurite outgrowth and, as such, is as efficient as laminin. This suggests that for certain applications (clinical) merosin would be better than laminin for nerve regeneration because it may not have e.g. angiogenic activity.

EXAMPLE III Merosin Distribution in Human Schwann Cell Neoplasms

The expression of the basement membrane proteins merosin and laminin was studied immunohistochemically in a series of benign and malignant schwannomas and plexiform neurofibromas. Fresh tissue samples were frozen in liquid nitrogen. Monoclonal antibodies to merosin and laminin were applies to frozen sections, and indirect immunoperoxidase or indirect immunofluorescence techniques were used to detect the two proteins in tissues. The results are described in Leivo et al . , Laboratory Investigation, 61:426-432 (1989) . This reference and the references cited therein are hereby incorporated by reference.

Tissue Material

Human neurogenic tumors were obtained fresh without fixation at the Department of Pathology, University of Helsinki. In one instance tissue was derived from the autopsy of a patient with von Recklinghausen' s disease who died of a buccal malignant schwannoma. The tissue samples were frozen in liquid nitrogen and embedded in Tissue-Tek OCT (Miles, Naperville, Illinois) . The frozen sections were air-dried for 1-2 hours and fixed in acetone. Part of each tissue sample was fixed in formalin and embedded in paraffin for conventional histologic evaluation using hematoxylin-eosin.

Antibodies

Monoclonal antibodies raised to the reduced and alkylated 65-kD polypeptide fragment of merosin were used. These antibodies detect denatured human merosin, and they blotted an 80-kD polypeptide band in sodium dodecyl sulfate extracts of human placenta. The following clones of these antibodies giving identical staining results were used: 5H2, 4E10, 2G9, 4H2, 1F6, 2E10, and 2D10. Staining results identical to those obtained with monoclonal antibodies have also been obtained in normal tissues with a polyclonal antiserum to merosin. Monoclonal antibodies to nearly intact human laminin have been described, Engvall et al . supra. The monoclonal antibody 2E8 that blots the 200-kD Bl chain of laminin transferred from sodium dodecyl sulfate-polyacrylamide gels was used.

In immunohistochemical characterization of the Schwann cell tumors, we used a polyclonal rabbit antibody to bovine S-100 protein (Dakopatts, Glostrup, Denmark) at 1:300 dilution and a monoclonal antibody to glial fibrillary acidic protein (Labsystems, Helsinki, Finland) at 1:30 dilution.

Immunohistochemistry

Frozen sections were treated with hybridoma culture medial at 1:2-1:5 dilution. The primary mouse antibodies were applied on sections for 30 minutes or overnight, followed by a 30-minute incubation with biotinylated rabbit antimouse IgG anti-serum (Dako, Copenhagen, Denmark) at 1:500 dilution. Finally, the bound biotin was detected with avidin combined m vitro with biotinylated peroxidase

(AB Complex, Dakopatts), both diluted at 1:160. The color was developed with 3-amino-9-ethylcarbazole (Sigma, St.

Louis, Missouri) supplemented with 0.02% hydrogen peroxide. In some cases, fluorescein isothiocyanate-coupled goat antimouse IgG (Bio-Rad, Richmond, California) was used to detect bound primary antibodies in indirect immunofluorescence.

For controls of specificity for the staining of merosin, normal mouse serum (1:10) or phosphate-buffered saline were used instead of the hybridoma medium. Controls of specificity for the staining of laminin by monoclonal antibodies have been documented. No significant staining was observed in control experiments. The preparations stained with the immunoperoxidase technique were lightly counterstained with Mayer's hemalum (Merck, Darmstadt, West

Germany) to show nuclei. Immunoperoxidase stainings and immunofluorescence preparations were observed and photographed in a Zeiss Axiophot microscope equipped for epi-illumination.

Four human schwannomas, two plexiform neurofibromas, and four malignant schwannomas were examined. Two schwannomas were retroperitoneal; one was mediastinal, and one was from the gastric wall exhibiting the histological features of gastric schwannomas. Histologically, all schwannomas showed a relatively uniform spindle cell morphology with focally palisading arrangement of nuclei. Two cases showed an alternating pattern of cellular and loose areas, representing the so-called Antoni A and Antoni B areas, respectively. Electron microscopic examination performed in three cases disclosed spindle cells rich in rough endoplasmic reticulum exhibiting multiple slender cell processes covered by prominent deposition of basement membrane material. These findings were compatible with the ultrastructural features of schwannomas. In immunohistological studies, all schwannomas were strongly positive for S-100 protein. Glial fibrillary acidic protein (GFAP) was focally seen in three cases.

Prominent staining for laminin was seen in parallel layers of basement membranes in the cellular areas and in the entire thickness of the walls of all blood vessels. The loose, less cellular areas of the tumors and the connective tissue sheaths around vessel walls contained no immunoreactive laminin. The cellular areas including the Verocay bodies contained no or only negligible amounts of merosin. However, distinct staining for merosin was regularly seen at the interface where the cellular areas bordered the loose stromal areas or where the cellular areas bordered vascular septa.

Plexiform Neurofibromas

Two plexiform neurofibromas were from nerve trunks of the subcutis of the back and the mediastinum of patients with von Recklinghausen's disease. These tumors represented enlarged tortuous nerve trunks containing wavy collagen and spindle cells compatible with Schwann cells and fibroblasts. In both tumors, merosin and laminin were colocalized in the form of linear immunoreactivity along basement membranes outlining the tortuous nerve fascicles. Laminin was also found in vessel walls. However, no merosin was seen in this location.

Malignant Schwannomas

These tumors originated from deep nerve trunks of femoral, retroperitoneal, and buccal tissues in patients with von Recklinghausen's disease. Histologically they represented malignant high grade spindle cell sarcomas with pronounced mitotic activity and focal areas of necrosis. The malignant schwannomas showed only minimal focal immunostaining for S-100 protein. No staining with antibody to GFAP was detected.

There was only minor focal staining for laminin in some perivascular tumor cells. All vessel walls were, however, strongly positive for laminin. Three of the four malignant schwannomas showed no immunostaining for merosin in the tumor cells. In contrast to laminin, only the external edges of vessel walls showed some staining. In sections where remnants of the original nerve trunks were microscopically identified, staining for merosin outlined the Schwann cell basement membranes of residual normal axons blending into merosin-negative tumor cell areas. A fibrous capsule surrounding malignant schwannomas was negative for merosin. However, in the adjacent striated muscle tissue, the basement membranes were positive for merosin. In one case, small but definite amounts of merosin were seen as punctate deposits between the tumor cells. In this case, a similar pattern of immunostaining for laminin was seen.

In brief, the distribution of merosin in schwannomas was more restricted than that of laminin, whereas in plexiform neurofibromas both proteins were present in the same location. No significant amounts of either protein were seen in malignant schwannomas.

In schwannomas, a strong staining for laminin was observed in basement membranes of the cellular Antoni A areas. In contrast, these areas were devoid of merosin. Immunoreactive ^'merosin was seen at the border zone between tumor cells and vessel walls. The discordant distribution of the two basement membrane proteins in schwannomas differs from the situation in normal peripheral nerves where both the merosin and laminin are seen in the Schwann cell basement membranes. The reasons for this difference are unknown, but the result may reflect different biological roles for the two basement membrane proteins. Ultrastructurally, no apparent difference seems to exist between the neoplastic basement membranes of schwannomas and the normal basement membranes surrounding Schwann cells.

The presence of merosin only at the boundaries of the schwannoma cells and non-Schwann cell mesenchymal components demonstrates that the expression of merosin may be induced by a contact or an interaction of schwannoma cells with mesenchymal tissues or extracellular matrices and that no expression occurs by isolated schwannoma cells even in relatively well-differentiated tumors. Analogously, Schwann cells in peripheral nerves may require interactions with other cell types of the nerve fascicles such as the neurons, endoneurial fibroblasts, or perineurial cells for synthesis and/or deposition of merosin. It has been shown that the myelination and assembly of Schwann cell basal lamina in the developing nerve in vitro depend on interactions between the Schwann cell and neuron. Likewise, secretion of type IV collagen by cultured Schwann cells is modulated by a contact with neurons. In plexiform neurofibromas, large amounts of both merosin and laminin were seen in an identical location. These neoplasms contain increased numbers of Schwann cells and perineurial cells as well as some residual axons contained within an intact perineurial sheath and enlarge the nerve fascicles. Thus, a relatively well-organized tissue architecture presumably essential for the expression of merosin is maintained. The presence of various cell elements within these nerve fascicles allows for many cellular contacts and interactions, and apparently some of these are essential for the secretion of merosin.

In the malignant schwannomas of this study, both merosin and laminin were absent or only minimally expressed. The concomitant lack of immunohistological markers for Schwann cell differentiation such as S-100 protein and GFAP suggests that these tumors are neurogenous sarcomas at a low level of Schwann cell differentiation.

Biosynthesis of laminin, type IV collagen, heparan sulfate proteoglycan, and entactin has been repeatedly shown in Schwann cell and schwannoma cell cultures. Moreover, in solid choriocarcinomas merosin was expressed by cells of the intermediate trophoblast type. No merosin could be detected in cultured choriocarcinoma cell lines, although these cell lines synthesized laminin. Apparently, cultured and neoplastic Schwann cells and other cells lose the capacity to secrete merosin but retain some other matrix proteins characteristic of the corresponding mature cells. EXAMPLE IV

ISOLATION OF cDNA ENCODING FULL LENGTH HUMAN MEROSIN

Generation and Characterization of cDNA Clones

cDNA libraries were made from human placental poly(A)

RNA. First, RNA was primed with primer ML-1 (nucleotide residues 6917-6942, Figure 6) made according to the M chain

(merosin) sequence in Figure 1 and Ehrig et al. , Proc.

Natl. Acad. Sci. USA 87:3264-3268 (1990), incorporated herein by reference. The cDNA was prepared with a cDNA synthesis kit according to the manufacturer's instructions

(Amersham International) , purified and cloned into a λgtlO vector (Promega) using EcoRI/NotI adaptors (Pharmacia) and packaged using the Packagene extract system (Promega) . Two other primer extension libraries were prepared similarly using primers M-10 (nucleotide residues 4153-4167, Figure 6) and ML-5 (nucleotide residues 1028-1050, Figure 6) . The first library was screened using the previously characterized merosin cDNA (Example I and Ehrig et al. , supra, incorporated herein by reference) as probe. The 5' end 1.4 kb EcoRI fragment of clone Ml-1 was used to screen the second extension library, and the third cDNA library was screened using a 1.3 kb Notl/AccI fragment of clone M10-22. To obtain clones for the 5' end of merosin, ML-6 (nucleotides 706-731, Figure 6) primed cDNA was synthesized and EcoRI adaptors (Promega) were ligated to the cDNA. An EcoRI adapter primer and specific primers were used to amplify the 5' end of the cDNA by PCR. Purified cDNA clones and PCR products were subcloned into Bluescript II (Stratagene) and sequenced from both strands using dideoxy sequencing (Sanger et al. , Proc. Natl. Acad. Sci. USA 84:935-939 (1977) hereby incorporated by reference). Northern Analysis

Total RNA from 18-19 week-old human fetal tissues was isolated and samples containing 10 μg of each RNA were electrophoresed, transferred to a GeneScreen Plus filter and hybridized with human laminin A chain and merosin cDNA probes.

In si tu Hybridization

To obtain sense and antisense probes for in si tu hybridization a 260 bp Notl-Sall fragment from laminin A chain cDNA clone C2-12 and a 350 bp Xhol-Clal fragment from merosin cDNA clone Ml-1 were cloned into the Bluescript II vector. Probes were labeled with ³⁵S-UTP (Amersham) using T3 and T7 polymerases. Human fetal tissues from the 17th gestational week were used. In si tu hybridization was performed according to Cox et al., Devi. Biol. 101:485-502 (1984) and Wilkinson et al., In: Postimplantation mammalian embryos: a practical approach (A.J. Copp and D.L. Cockrof, eds. ) IRL Press, Oxford 155-171 (1990) , each of which is hereby incorporated by reference.

Characterization of cDNA Clones and Amino Acid Sequence of the Merosin Chain

A cDNA clone providing 1130 amino acid residues from the carboxyl terminal end of human merosin is described in Example 1, and by Ehrig et al. , Proc. Natl. Acad. Sci. USA 87:3264-3268 (1990). This cDNA clone (MER 3') and its sequence were used for priming and screening of the first primer extension library. The longest positive 2.9 kb clone Ml-1 (Figure 5) was further characterized and its 5' end sequence was used to prime and screen the second primer extension library yielding clone M10-22 (3.2 kb) . The 5' end of clone M10-22 was similarly used for screening of the third primer extension library resulting in the isolation of clone M5-1 (0.8 kb) . Several libraries were made in order to obtain clones spanning the entire 5' end sequence. However, all clones obtained through those efforts were either of similar lengths or shorter than M5-1. Genomic clones that were characterized (data not shown) contained the putative exon 2, but not the coding region for the signal peptide and 5' untranslated region. The 5' end sequences were finally obtained by PCR amplification. The primer ML-6 was used to make cDNA to which EcoRI adaptors were ligated. An EcoRI adaptor primer and two specific primers were then used in PCR to amplify a 300 bp 5' end fragment, Mg-16 (Figure 5) , containing sequences for the 5' end untranslated region of the mRNA, the signal peptide and the amino-terminal end of merosin.

The nucleotide sequence of the overlapping cDNA clones and the deduced amino acid sequence are shown in Figure 6. The C-terminal end amino acid sequence described in Example I and in Ehrig et al., Proc. Natl. Acad. Sci. USA 87:3264- 3268 (1990) is included in that sequence. The clones generated and characterized in this study covered a total of 6942 bp, consisting of a 49 bp 5' end untranslated region and 6893 bp of an open reading frame. The 5' end sequence has an open reading frame but the sequence ACUACGAUGC around the initiator methionine is in agreement with the Kozak consensus sequence for translation initiation (Kozak, M. , J. Cell. Biol. 115:887-903 (1991) hereby incorporated by reference) . The putative signal peptide contains 22 amino acids starting with the initiator methionine followed by a hydrophobic leucine-rich sequence. Computer program analysis predicting the signal peptidase cleavage site, based on the method of von Heijne (1986) , incorporated here by reference, suggested a cleavage site after Ala22, whereby mature merosin would start with a glutamine residue as do most laminin chains. Altogether, merosin contains 3088 amino acid residues after cleavage of the tentative 22-residue signal peptide. Domain Structure of Merosin Comparison with the Laminin A Chain

Mature human merosin is 30 residues larger than the human laminin A chain which contains 3058 residues (Nissinen et al. , Biochem. J. 276:369-379 (1991); Haaparanta et al. , Matrix 11:151-160 (1991)). The amino acid sequences of both chains are aligned in Figure 7. Similarly to all laminin chains, the merosin protein has distinct domains which are predicted to have globular regions, cysteine-rich rod-like regions and helical structures. Additionally, merosin, like the laminin A chain, has a large globular domain at the carboxy-terminal end (Ehrig et al. , Proc. Natl. Acad. Sci. USA 87:3264-3268 (1990) ) . Comparison of the two sequences demonstrates that the domain structure of merosin is similar to that of the laminin A chain, and that these two laminin heavy chains have considerable homology.

The amino-terminal end domains VI (residues 23-286) , IVb (residues 528-723) and IVa (residues 1176-1379) of merosin are predicted to form globular structures. Domains

V (residues 287-527), Illb (residues 724-1175) and Ilia

(residues 1380-1573) contain cysteine-rich EGF-like repeats and are predicted to have rigid rod-like structures. The number of EGF-like repeats is identical in merosin and the laimin A chain. Domain V has four and one-half repeats, domain Illb has ten and one-half, and domain Ilia has four repeats. Beck et al. , FASEB J. 4:148-160 (1990) and Beck et al. , In: W. Taylor and P. Argos, (eds) Springer series in biophysics, Springer-Verlag, Berlin 7:231-256 (1992) count the half repeats as one, and according to that both chains contain 17 cysteine-rich repeats. Domains I+II (residues 1574-2153) , a part of which has previously been reported (Ehrig et al. , Proc. Natl. Acad. Sci. USA 87:3264- 3268 (1990) ) , participate together with two B-type chains in the formation of a triple coiled-coil structure that forms the long arm of the laminin molecule. In addition, merosin contains one cysteine residue in this region which has no counterpart in the laminin A chain or any of the B- type chains characterized thus far. The large carboxy- terminal G domain (residues 2154-3110) forms the large globule at the end of the long arm of the laminin molecule.

The amino-terminal domain VI in the M chain has 12, domain IVa two, domain Illb one, domain I+II has 10 and domain G has seven amino acid residues more than the A chain. Domain V in the A chain has two residues more than the corresponding domain in the M chain. Comparison of the amino acid sequences of the human merosin and laminin A chain shows that the overall sequence similarity is 46.6% (Table 3) and 58.6% when conservative changes are included (Figure 7) . The sequence similarity is highest in the globular domains VI, or 73.9%, although this domain in merosin contains 12 residues more at the amino-terminus than the laminin A chain. If the additional glutamine rich amino-terminal sequence is excluded, the homology is 77.4%. All six cysteine sites in this domain are conserved. The amino acid sequence identities of the cysteine-rich domains V, Illb and Ilia between merosin and laminin A chain are 60.1%, 54.9% and 50.2%, respectively. All cysteine residues in these domains are conserved and the length of domains are about the same. The globular domains IVb and IVa of the two chains also have approximately the same number of amino acids, although the sequence similarity is lower, or 42%. The sequence similarity is lowest between domains I+II where it is only 32.3%. There is also an extra cysteine residue (residue 1970) in domain I+II in merosin that has no counterpart in the laminin A chain. The sequence identity between domains G is 41.8%. There are 28 putative N-glycosylation sites in merosin and 34 in the laminin A chain, ten of these sites are conserved between the two chains. Most putative glycosylation sites are in domains G and I+II. Chromosomal Assignment of Human Merosin Gene

The human merosin gene was mapped to chromosome 6 by hybridization of labeled cDNA clone M10-22 to DNA from a panel of 39 somatic cell hybrids. Hybridization of the merosin cDNA clone correlated with the distribution of chromosome 6. In si tu hybridization of the cDNA to metaphase chromosomes confirmed the localization of the merosin gene to chromosome 6, and more precisely to bands 6q22->q23 (Figure 8) .

Expression of Merosin and Laminin A Chain Genes in Human Tissues

Expression of merosin and laminin A chain genes was compared by Northern hybridization using RNA from several 18-19-week-old human fetal tissues (Figure 9) . As previously reported (Nissinen et al. , Biochem. J. 276:369- 379 (1991) ) , the laminin A chain gene has highly restricted expression in human adult tissues. Signals for the laminin A chain were observed only in brain, neuroretina, kidney and testis, while no signals were obtained with RNA from skin, colon, pancreas, adrenal glands, cardiac muscle, lung, thymus, spleen, liver or calvarial bone, even after long exposures. The signal was by far the strongest in the neuroretina and in brain tissues the laminin A chain gene is expressed in the meninges, the intermediate zone, cerebellum, olfactory bulb and weak expression was observed also in choroid plexus and the ependymal zone.

The merosin protein has a different expression pattern, signals being observed with RNA from most tissues studied except thymus, liver, calvarial bone and ependymal and intermediate zones of brain. The strongest expression of the merosin gene was seen in cardiac muscle, pancreas, choroid plexus and meninges. In si tu Hybridization

The location of merosin mRNA was analyzed by in si tu hybridization in 17-week-old human fetal tissues. A cell- type-specific expression pattern for merosin mRNA was obvious in kidney, heart, skin and lung. In embryonic kidney, the transcripts for merosin were predominantly found in the undifferentiated nephrogenic mesenchyme of the outermost cortex (Figures 10A and 10B) , whereas the nephric tubules and renal blood vessels remained negative. In heart muscle expression was observed in myocytes throughout the tissue (Figures IOC and 10D) . The epidermal cells of the skin did not express merosin mRNA which, however, was abundant in the condensing mesenchyme around the tip of the developing hair follicles (Figures 10E and 10F) . In the lung (Figures 10G and 10H) label was found in the smooth muscle cells of the pulmonary arteria, while the alveolar and bronchiolar cells were negative. Thus, the epithelial and endothelial cells were negative for merosin mRNA and the transcripts were found only in various mesenchymal cells. No cell specific signals were observed with the laminin A chain specific hybridizations in the tissues studied.

The present results, together with the 3' end sequence described in Example I and by Ehrig et al. , Proc. Natl. Acad. Sci. USA 87:3264-3268 (1990), incorporated herein by reference, provide the complete primary structure for the human merosin. The merosin and laminin A chains were shown to be very similar. The overall sequence similarity between the two human chains (46.6%) is about the same as that between the homologous Bl and S chains. The human merosin and laminin chain genes have been localized to different chromosomes, with the exception of the genes for the closely related B2 and B2t which are located in the q25->q31 region of chromosome 1 (Fukushima et al. , Cvtogen. Cell Genet. 48:137-141 (1988) . In this study, the merosin gene was assigned to 6q22->q23 while the related laminin A chain gene has been localized to chromosome 18pll.3 (Nagayoshi et al . , Genomics 5:932-935 (1989)) .

Domain Structure

The domain structure of merosin contains several features similar to other laminin chains and it is practically identical to that of the laminin A chain. The amino terminal globular domains VI share the highest homology, although merosin has additional 12 amino acids at the amino terminus. In fact, domain VI of all known human laminin chains, the mouse A chain, the rat S chain and the Drosophila A chain can be aligned so that the cysteine residues, some glycine, serine, proline and arginine residues, and short amino acid sequences RP, TCG and WWQS match in all chains (Figure 11) . A conserved sequence,

Y(Y/F)Yxhxdhxh(G/R)G (h: hydrophobic residue, d: D, E or N)

(according to Beck et al., In: W. Taylor and P. Argos,

(eds) Springer series in Biophysics, Springer-Verlag,

Berlin 7:231-256 (1992)) , incorporated herein by reference, at the carboxyl terminus of domain VI also is found in merosin. The function of these conserved sequences is not known; but, while not wanting to be bound by any theory, the conserved regions can have significance for the role of this domain in laminin self-assembly which is apparently mediated by the amino terminal globular domains.

Domains V, Illb and Ilia contain EGF-like repeats with eight cysteine residues at regular positions. The number of residues between the eighth to the second and the fifth to the seventh cysteine is the same in all laminin repeats and the order of repeats is specific. The number of repeats in merosin is 20 according to Sasaki et al . , J. Biol. Chem. 263:16536-16544 (1988) , incorporated herein by reference, or 17 according to Beck et al., FASEB J. 4:148-160 (1990) ; Beck et al. , In: W. Taylor and P. Argos, (eds) Springer series in biophysics. Springer-Verlag, Berlin 7:231-256 (1992)), each incorporated herein by reference. The order of the repeats is conserved in human merosin and, generally, the repeats are very similar to the repeats present in human and mouse laminin A chains. The repeats in domain V of the human A, M (merosin) , Bl, B2 and B2t chains, the rat S chain, the murine A chain and the Drosophila A chain can be aligned in order (see Figure 12) . The human B2t chain lacks the first EGF-like repeat, but the rest of the repeats match with repeats of the other chains, except that after the second repeat there is an insertion making the distance between the eighth to the second cysteine longer than in the other chains. The alignment of domain V includes in addition to cysteine and glycine residues also other conserved sequences like HNT in first repeat between cysteines five and six. In contrast to other laminin chains, the Drosophila A chain contains 10 and a half EGF-like repeats in domain V. The two first cysteine-rich repeats in the Drosophila A chain can be aligned with repeats in the other chains but the rest of domain V differs more, although some similarities are found between repeats 3, 4, 5 and 6 in the Drosophila A chain and repeats 3, 4 and 5 in the other chains. All EGF-like repeats of known A-type chains can be aligned but this alignment is based mainly on conserved cysteines and glycines and the number of residues between them.

Globular domains IV of the A- and B2-type chains have been suggested to have evolved by an insertion between the third and fourth cysteines in one EGF-like repeat, and to be duplicated in A chains to form domains IVb and IVa. These domains are present in merosin and are, thus, well conserved in the laminin A-type chains, except for the Drosophila laminin A chain which contains only one domain IV. It also has another domain IV" that consists of duplicated sequences that are more similar to the Drosophila Bl chain domain IV. Domains I+II form the long arm helical region. The EHS laminin chains have been shown to contain heptad repeats and similar repeats can be found also in the human laminin A chain and merosin. Proline residues are known to interrupt helices. There are four conserved proline residues in domain I+II in the mouse laminin A chain and the human laminin A chain and merosin. The cysteine pair that is suggested to form interchain disulfide bonds is conserved in merosin.

Domain G of merosin consists of five internal repeats that contain 107 to 178 amino acid residues (Ehrig et al., Proc. Natl. Acad. Sci. USA 87:3264-3268 (1990)), incorporated herein by reference. These repeats share 30 to 50% homology when compared with the human or mouse laminin A chain. The Drosophila laminin A chain also has five repeats in the G domain, but there is a large spacer sequence rich in threonine residues between subdomains G3 and G4 (Kusche-Gullberg et al., EMBO J. 11:4519-4527 (1992) ) . Several proteins are known to be homologous to the G domain in the laminin A chain and merosin. For example, one domain of the HSPG (heparin sulfate proteoglycan) core protein, perlecan, has 33% homology with the domain G of the human laminin A chain and merosin. Other homologous proteins are sex hormone binding globulin (Beck et al. , In: W. Taylor and P. Argos, (eds) Springer series in biophysics. Springer-Verlag, Berlin 7:231-256 (1992)) , androgen binding protein (Joseph et al., FASEB J. 6:2477-2481 (1992)) _. and neurexins (Ushkaryov et al. , Science 257:50-56 (1992)) , each of which is hereby incorporated by reference. Also Drosophila proteins fat, slit and crumbs share similarities with domain G of merosin and laminin A chain (Patthy, L., FEBS Lett. 298:182-184 (1992) ) . Table I. Similarity of amino acid sequences of the human laminin Am and A chains as aligne in Fig. 3.

Length of Matches aligned (conservative

Domain sequence Matches substitutions) Unmatches Matches Length

%

VI 264 195 25 12 73.9 (83.3)

V 243 146 23 2 60.1 (69.5)

IV-b 199 85 28 6 42.7 (56.8)

Illb 452 248 38 1 54.9 (63.3)

IVa 207 88 33 8 42.5 (58.5)

Ilia 195 98 14 2 50.2 (57.4)

I+II 591 191 79 32 32.3 (45.7)

G 987 413 136 67 41.8 (55.6)

Total 3138 1464 376 130 46.6 (58.6)

Expression of Merosin and Laminin A Chain in Human Fetal Tissues

Expression of the merosin gene was observed in many tissues known to contain the respective protein from immunohistological studies. However, the strong level of expression at an early embryonic stage contrasts previous immunostaining studies wherein merosin was not detected in the mouse embryo (Leivo et al. , Proc. Natl. Acad. Sci. USA 85:1544-1548 (1988)) . The reason for this discrepancy is obscure; but, while not intending to be bound by any theory, it could be due to some unknown limitation in the antibodies or the transcripts may not be efficiently translated into proteins. Merosin has been reported to appear in mouse muscle tissues first after birth (Leivo et al., Proc. Natl. Acad. Sci. USA 85:1544-1548 (1988)) and at adult stages also in some other tissues in several mammalian species (Sanes et al. , J. Cell Biol. 111:1685- 1699 (1990) ) . The data presented here on 17-week-old human fetal tissues revealed strong expression of the merosin gene in cardiac muscle, pancreas, choroid plexus and meninges, significant expression also being observed in testis, skin, adrenal glands, kidney, lung, spleen, neuroretina, olfactory bulbs and cerebellum. Practically no signals were observed in thymus, liver, bone or some brain tissues such as the intermediate and ependymal zones or cortical plates. The in si tu hybridization analyses localized the expression of the merosin gene to myocytes of heart muscle, which agrees with several previous studies

(Leivo et al., Proc. Natl. Acad. Sci. USA 85:1544-1548 (1988) ; Paulsson et al. , J. Biol. Chem. 264:18726-18732 (1989); Klein et al. , Development 110:823-837 (1990); Engvall et al., Cell Regul . 1:731-740 (1990) ; Paulsson et al., J. Biol. Chem. 266:17545-17551 (1991)) . However, expression also was seen in stromal cells close to condensing mesenchyme in kidney and skin. Merosin has been localized by a monoclonal antibody to a narrow region located between the stromal cells and pretubular condensates in the outer cortex. A good concordance between merosin mRNA and protein expression also is seen in other embryonic tissues. The strong expression observed in mesenchymal cells located immediately beneath cells at the tip of the developing hair follicle and sebaceous glands indicates the potential role of merosin in exocrine gland development. Expression of merosin was not found in epithelial or endothelial cells of any of the tissues analyzed. Consequently, it can be concluded that during embryogenesis expression of merosin is primarily, if not only, the property of cells of mesenchyme origin.

Expression of the laminin A chain gene was shown to be considerably more restricted in human fetal tissues than that of the merosin gene. As previously reported for newborn human tissues (Nissinen et al. , Biochem. J. 276:369-379 (1991)) Northern analysis revealed expression of the laminin A chain gene in kidney. The present studies did not locate the expression at this stage of kidney development to specific cells by in si tu hybridizations. The laminin A chain has been localized in the kidney to tubular and glomerular basement membranes of adult tissues

(Sanes et al. , J. Cell Biol. 111:1685-1699 (1990)) and in polarized kidney epithelial cells (Holm et al. , Cell

Differ. 24:223-238 (1988); Klein et al. , Cell 55:331-341 (1988); Ekblom et al. , Cell 60:337-346 (1990). Klein et al. , Development 110:823-837 (1990) reported the detection of laminin A chain mRNA in embryonic heart, liver, lung and intestine, and laminin containing the A chain has been isolated from skeletal and heart muscle, lung, liver, kidney and intestine (Paulsson et al. , J. Biol. Chem. 264:18726-18732 (1989)) . However, in this study on tissues from a 17-week-old human fetus, no signal for the A chain mRNA was observed in lung, heart or liver, even after long exposures. This discrepancy could be due to the differences in temporal expression during development. The intense expression of the laminin A chain gene in neuroretina, olfactory bulbs and cerebellum, is interesting and indicates its role in brain and nerve development. Detailed immunohistological and in si tu hybridization analyses on developing brain tissues have been initiated to further analyze the temporal and spatial expression during brain development.

Several studies including the present study have demonstrated variability in both spatial and temporal expression of laminin subunit chains in vivo . This, in part, implicates tissue-specific functions of different laminin isoforms. With regard to merosin and the laminin A chain, Engvall et al. , Cell Regul. 1:731-740 (1990) and Sanes et al. , J. Cell Biol. 111:1685-1699 (1990), each incorporated herein by reference, showed that they are often mutually exclusive in a distinct type of basement membranes, suggesting that the laminin molecules contain either an M (merosin) or an A chain as a heavy chain. The present Northern blot and in si tu hybridization analyses carried out on RNA from human fetal tissues supports the different tissue distribution of the M (merosin) and A chains. In particular, the results showed that the merosin gene is expressed in several tissues during embryonic development and possibly only by mesenchymal cells. However, the results also demonstrated that some laminin producing cells and tissues, such as skin and lung epithelia as well as vascular endothelia did not express either gene, or its expression was very weak in these tissues. This suggests that there exist laminin isoforms containing some, as yet, unidentified heavy A-type chains. Such isoforms may include kalinin or K-laminin. Although the invention has been described with reference to the presently-preferred embodiment, it should be understood that various modifications can be made without departing from the spirit of the invention. Accordingly, the invention is limited only by the following claims.

Claims

What is claimed is :

1. An isolated nucleic acid molecule which encodes human merosin.

2. The nucleic acid molecule of claim 1, wherein the nucleic acid is DNA.

3. The nucleic acid molecule of claim 1, wherein the nucleic acid is cDNA.

4. The nucleic acid molecule of claim 1, wherein the nucleic acid is mRNA.

5. An isolated nucleic acid molecule having substantially the sequence shown in Figure 1 or Figure 6.

6. A method of making anti-merosin antibodies comprising the steps of:

a) immunizing an animal with a polypeptide corresponding to CNNFGLDLKADDKI or CSIVDIDTNQEENI; and

b) obtaining antibodies reactive with merosin or collecting antibody-producing hybridoma cells and selecting antibodies reactive with merosin.

7. Antibodies made by the method of claim 6.

8. The antibodies of claim 7, wherein the antibodies are polyclonal antibodies.

9. The antibodies of claim 7, wherein the antibodies are monoclonal antibodies.

10. A biologically active polypeptide having substantially the amino acid sequence as the 380-400 KDa polypeptide of merosin as shown in Figure 6.

11. Antibodies reactive with the 380 KDa polypeptide of claim 10.

12. The antibodies of claim 11, wherein the antibodies are polyclonal.

13. The antibodies of claim 11, wherein the antibodies are monoclonal antibodies.

14. A recombinant expression vector comprising the nucleic acid molecule of claim 1.

15. A recombinant expression vector comprising the nucleic acid molecule having a sequence substantially the same as the sequence shown in Figure 1 or Figure 6.

16. A host-vector system comprising the recombinant expression vector of claim 14 and 15.

17. A host-vector system of claim 16, wherein the host is a procaryotic cell.

18. A host-vector system of claim 16, wherein the host is a eucaryotic cell.

19. The procaryote of claim 17, wherein the procaryote is an Escherichia coli cell.

20. A cell culture comprising the cells of claim 17, and a suitable medium.

21. A nucleic acid probe comprising at least ten nucleotides encoding a protein having a sequence shown in Figures 1 or 6 and the probe being sufficiently complementary to portions of a nucleic acid molecule having the sequence of Figures 1 or 6, to allow hybridization.

22. A method of detecting the presence of a nucleic acid encoding merosin in a sample comprising the steps of:

a) contacting the probe of claim 21 to the sample under conditions favoring hybridization of the probe to complementary nucleic acid;

b) detecting the presence of any probe hybridized to the m vitro and in si tu or extended nucleic acid, the presence of hybridized probe indicating the presence of nucleic acid encoding merosin in the sample.

23. The method of claim 22, wherein the probe is labeled with a detectable marker selected from radionucleotides, enzymes, biotin, and fluorescent markers.

24. A method of detecting the presence of merosin in a sample, comprising the steps of:

a) contacting the antibody of claim 11 with the sample under conditions favoring the formation of antibody- merosin complexes;

b) detecting the presence of any complex which formed, the presence of complex indicating the presence of merosin in the sample.

25. The method of claim 24, wherein the antibody is labeled with detectable marker.

26. A nucleic acid molecule encoding a fragment of merosin, wherein the fragment is produced by restriction enzyme digestion of nucleic acid molecules shown in Figures 1 or 6.

27. A method of determining the malignancy of a tumor comprising detecting merosin produced by the tumor, the substantial absence of merosin indicating a malignant tumor.

28. A method of inducing the production of merosin comprising contacting a merosin producing cell with a cell or matrices which induce the production of merosin in the merosin producing cell.

29. The method of claim 28, wherein the cell which induces the production of merosin is a mesenchymal.

30. A method of promoting neurite growth comprising contacting a neuron with the protein produced by the nucleic acid molecule of claim 1.

31. A method of inhibiting neurite growth comprising contacting merosin with an inhibitor of the neurite promoting activity.

32. A method of promoting neurite growth comprising preventing the binding of an inhibitor of the neurite promoting activity of merosin, to merosin.

33. A method of promoting cell attachment of a cell which binds to merosin to a substrate, comprising contacting the cell with substrate containing merosin fragment or a cell attachment promoting portion thereof.