EP0431065A1 - A FULL LENGTH cDNA ENCODING A HUMAN LAMININ BINDING PROTEIN - Google Patents

Abstract

A normal length cDNA sequence containing the human laminin binding protein and derived from human colon carcinoma cells essentially comprises the sequence illustrated in Figure 3. A subsequence essentially corresponds to the sequence that begins with the nucleotide # 162 and which ends with nucleotide # 829, as shown in Figure 3. Another aspect of the invention relates to recombinant clones encoding a human laminin binding protein. In another embodiment, the invention relates to a laminin binding protein encoded by one or the other sequence, and isolated DNA encoding the laminin binding protein forming part of the cDNAs described above. A method and reagent are used to detect the presence of carcinomas in a patient.

Description

TITLE

A FULL LENGTH cDNA ENCODING A

HUMAN LAMININ BINDING PROTEIN

GOVERNMENT INTEREST: ACKNOWLEDGEMENT The invention described herein was made during the course of work under Grant Nos. POl CA44704, PO1 CA22427, and RO1 GM38318 awarded by the National

Institutes of Health.

CROSS REFERENCE TO RELATED APPLICATION

This application is a continuation-in-part of

U.S.S.N. 07/238,955, filed on August 31, 1988.

FIELD OF THE INVENTION

This invention relates generally to human laminin binding protein and, more specifically, to a full length cDNA encoding a human laminin binding protein derived from human colon carcinoma and to the molecular cloning thereof.

BACKGROUND OF THE INVENTION

Colon carcinoma is one of the leading causes for cancer death in humans. Over 140,000 new cases of colon carcinoma are expected in 1988 in the United

States; more than 70,000 of these patients will die of this disease. Current diagnostic techniques find less than half of the patients early enough to ensure

effective control by surgery. In particular, for patients with poorly differentiated colon carcinomas, detection and monitoring is difficult since such

malignancy generally expresses very little

carcinoembryonic antigen (CEA), a marker often useful for well- and moderately-differentiated colon

carcinomas.

Little is known about the transformation of normal colon epithelial cells to cancerous state and the differentiation of colon carcinoma cells at the

molecular level. More than 100 human colon carcinema derived cell lines have been established. Although these cell lines share the common characteristic of human colonic epithelial origin, they may be quite different in their lineage of differentiation, degree of differentiation, tumorigenicity in nude mice, histology of tumor induced, metastatic potential, and transformed phenotype in culture. They are potentially a rich source for unravelling new markers.

Laminin has been implicated in a wide variety of biological processes. Most relevant to colon cancer, laminin has been linked to cell adhesion, motphogenesis, mitogenesis, differentiation and metastasis. Laminin is a major constituent of basement membrane. Laminin is a large glycoprotein consisting of three polypeptide chains, A chain (M_r 400,000), B1 chain (M_r 210,000) and B2 chain (M_r 200,000) in a cross-shaped structure. It mediates the attachment of both normal and neoplastic cells to the basement membrane.

A proteolytic fragment of laminin and a pentapeptide of Tyr-Ile-Gly-Ser-Arg from the B1 chain with cell attachment activity have been identified.

Iwamoto et al. have reported in Science, 238:1132-1134 (1987) that the pentapeptide appears to inhibit the metastasis of melanoma in mice. It has been

demonstrated that much of the effects of laminin are mediated through a 67 kDa laminin receptor, cell surface glycoprotein.

U.S. Patent 4,565,789, issued to Liotta et al. on January 21, 1986, describes the isolation and

characterization of certain aspects of laminin receptor.

Wewer et al., Proc. Nat. Acad. Sci. (USA) 83:7137-7141 (1986) describes the isolation of laminin receptor cDNA as well as the nucleotide and putative amino acid sequence of the 2H5 epitope of human laminin receptor. Two additional laminin receptors have recently been identified with molecular masses of 18 kDa and 110 kDa, (Kleinman et al., Proc. Nat. Acad. Sci.(USA), pages 1282-1286, Vol. 85 (1988), and Smalheiser et al., Proc. Natl. Acad. Sci. (USA), pages 6457-6461 (1987)).

SUMMARY OF THE INVENTION

This invention is directed to a full length human laminin binding protein cDNA sequence derived from human colon carcinoma cells comprising substantially the sequence as set forth in Figure 3. It also concerns a subsequence corresponding substantially to the sequence starting with nucleotide number 162 and ending with nucleotide number 829 as set forth in Figure 3.

In another aspect, this invention also relates to recombinant clones encoding a human laminin binding protein.

In another embodiment, this invention relates to a laminin binding protein encoded by either sequence and to substantially isolated DNA encoding the laminin binding protein of the above-described cDNAs.

Also of concern is a method for detecting the presence of carcinoma in a patient and a reagent for detecting the presence of carcinoma in a patient.

BRIEF DESCRIPTION OF DRAWINGS

Figure 1. Blot-hybridization of RNAs from a human colon carcinoma and adjacent normal epithelium obtained from surgery with ³²p-labeled cDNA inserts from three plasmids. Lanes 1,3,5 used total RNA from normal human colonic epithelium; Lanes 2,4,6 from colon

carcinoma. Lanes 1,2 were hybridized with 8-2V; Lanes

3,4 with 3-4W; Lanes 5,6 with 1-4E. Arrows indicated the major species of mRNA hybridized, 1.2 kb, 1.7 kb, and 0.7 kb, respectively.

Figure 2. Blot-hybridization of RNAs from twenty human colon carcinoma cell lines with ³²p-labeled cDNA insert from 8-2V plasmid. Lanes 1 used RNAs from HT29; 2 from CX-1; 3 from RCA; 4 from CCL222; 5 from CCL220.1; 6 from CCL228; 7 from HCT116b; 8 from HTB39; 9 from Clone A; 10 from Clone D; 11 from MIP101; 12 from CCL229; 13 from Gly; 14 from Moser; 15 from CCL233; 16 from CCL224; 17 from CCL237; 18 from CL187, 19 from CCL227, and 20 from CCL234. Arrow indicated the major species of mRNA, 1.2 kb, hybridized.

Figure 3. Nucleotide sequence and deduced amino acid sequence of a full-length cDNA insert from J-9 lambda gt10 phage. All of the sequence of 8-2V insert is located substantially in the sequence shown here (nucleotide #162 to #829). Sequences of interest are underlined.

Figure 4 is a hydropathy plot of the deduced amino acid sequences shown in Figure 3.

Figures 5 and 6 are Northern blots of total RNA isolated from various human carcinoma cell lines which show significant levels of expression of the

1.2 Kb mRNA in human esophageal, cervical, duodenum, lung, prostate carcinoma and human melanoma cell lines.

Figure 7 shows that the in vitro translated products of laminin binding gene from rabbit

reticulocyte.binds to a laminim sepharose column and the molecular weight of laminin binding protein in vitro translated products migrate on SDS PAGE as a 45 kd protein.

STATEMENT OF DEPOSIT

1. The recombinant clone designated J-9 was deposited in the American Type Culture Collection (ATCC) under the Budapest Treaty on August 30, 1988 and was accorded ATCC accession number 40489.

2. The recorr_oir_ant clone designated 8-2V was deposited in the ATCC under the Budapest Treaty on August 30 , 1988 and was accorded ATCC accession number 40490 .

DETAILED DISCUSSION OF THE INVENTION

Unless defined otherwise, all technical or scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the invention belongs. All publications mentioned herein are incorporated herein by reference.

The term "derived from" means that the

sequence encodes the same protein or a functional equivalent with conservative substitutions, additions or deletions.

The term "substantially purified" means synthesized or, if naturally occurring, isolated free of other cellular components with which it is normally associated in a particular genome.

cDNA libraries were constructed from human colon carcinoma cell lines. These libraries were screened for specific clones that hybridized with mRNAs more abundant in colon carcinomas. A 1.2 kb mRNA was identified by its hybridization with the cDNA insert in recombinant clone 8-2V which is discussed below. This mRNA was about 9-fold more abundant in human colon carcinoma than that of adjacent normal colonic

epithelium. The cDNA of this 1.2 kb mRNA appears to be the first full length human laminin binding protein cDNA sequence derived from human colon carcinoma cells comprising substantially the sequence as set forth in Figure 3.

It is understood that more than one nucleotide triplet can code for the same amino acid and, due to the existence of degenerate codes, variations in the

nucleotide sequence can arise. Thus, variations of the nucleotide sequence set forth in Figure 3 can occur and fall within the scope of this invention. Identification of 8-2V Plasmid. To search for cellular mRNAs that are more abundant in human colon carcinoma than that of normal counterparts, a cDNA library was constructed in plasmid pBR322 using

poly (A) +mRNA isolated from a poorly differentiated human colon carcinoma cell line, Clone A. Initial screenings were made by hybridizing 1,500 plasmids with cDNAs generated from mRNAs of Clone A as well as CX-1, a well differentiated human colon carcinoma cell line. Whereas the great majority of plasmids hybridized equally between the two or more with that of CX-1, 10 plasmids were studied because there was greater hybridization with cDNAs from Clone A than that of CX-1, cDNA inserts in these plasmids were then used to compare the levels of mRNA between normal human colonic epithelial tissues and colon carcinoma of the same patient by blot

hybridization of RNA. gels. One of the plasmids, 8-2V, hybridized with 1.2 kb mRNA that was more abundant in carcinoma than that of adjacent normal epithelium (Fig. 1) . Equal amounts of total RNA from carcinoma and normal tissue were compared. Densitometric analysis suggested that the level of this mRNA might be

approximately 9-fold greater in carcinoma than that of normal epithelium. For comparison, the same pair of RNAs were hybridized with cDNA inserts from two other plasmids. Plasmid 1-4E hybridized with a 0.7 kb mRNA far more abundant in normal colonic epithelium than does carcinoma, whereas plasmid 3-4W hybridized with a 1.7 kb mRNA equally abundant between the two. These results suggested that 1.2 kb mRNA which hybridized with the cDNA insert in 8-2V has the ability to encode for a polypeptide which is useful for distinguishing colon carcinoma from normal colonic epithelium.

Blot Hybridization of Total RNA from Various Cell Lines with 8-2V cDNA Insert. To investigate. whether increased levels of 1.2 kb mRNA hybridized with 8-2V in colon carcinoma cells is a general phenomenon, and in particular, its expression by poorly

differentiated, 20 human colon carcinoma cell lines (mentioned above) were examined by blot hybridization of equal amount of total RNAs with ³²P-labeled probe. Fig. 2 shows that essentially all human colon carcinoma cell lines tested express significant levels of this mRNA except CCL234, a mucin-producing, well-differentiated, rectal adenocarcinoma cell line. Correlation between the level of this 1.2 kb mRNA and the degree of

differentiation was not apparent, although all poorly differentiated cell lines (CCL220.1, HCT116b, Clone A, and MIP101 ) have high levels of expression, and those expressing lower levels tend to be the well-differentiated (CCL233 and GLY, both high CEA

producers). On the basis of equal total RNAs, the level of this 1.2 kb mRNA in colon carcinoma cell lines was in the same range as that of surgical tissue of colon carcinoma shown in Fig. 1.

Nucleotide Sequence of 8-2V cDNA Insert. cDNA insert in 8-2V plasmid had 668 bp corresponding

substantially to the sequence in Figure 3 starting with nucleotide number 162 and ending with nucleotide number 829. A computer search of the DNA sequence data base revealed that 8-2V contained a sequence of about 349 nucleotides. This sequence contained the partial cDNA sequence of human laminin receptor reported by Wewer et al. The 2H5 monoclonal antibody is known to bind the laminin receptor on immunoblots, to react with the purified laminin receptor on ELISA, to prevent the binding of laminin to either cells or plasma membrane, and blocks the adhesion of cells to amnion basement membrane enriched in laminin. The epitope recognized by this antibody is present in the deduced amino acid sequence. Furthermore, the predicted polypeptide contains the twenty amino acid sequence (Pro-Thr-Glu-Asp-Trp-Ser-Ala-Gln-Pro-Ala-Thr-Glu-Asp-Trp-Ser-Ala-Ala-Pro-Thr-Ala) recognized by the polyclonal antibodies that inhibit laminin-mediated cell attachment. Taken together, it seems highly likely that the product of the product of 1.2 kB mRNA which hybridiized with 8-2V cDNA also binds laminin.

Molecular Cloning and Sequence of cDNA

Corresponding to 1.2 kb mRNA hybridized with 8-2V cDNA Insert. ³²P-labeled cDNA insert in 8-2V was then used to clone cDNAs with longer inserts from a lambda gt10 cDNA library generated from mRNAs of CX-1, a human colon carcinoma cell line. One of the cDNA clones, J-9, contained an insert approximately 1 kb. The cDNA sequence of this insert is substantially the same as that shown in Fig. 3. 8-2V cDNA sequence from Clone A cell line is completely located within this sequence of cDNA. Thus, the antigenic domain recognized by the 2H5 monoclonal antibody and the C-terminal 20-amino acid recognized by polyclonal antibodies that block laminin binding with laminin receptor has an identical

nucleotide sequence from three different sources, human umbilical vein endothelial cells. Clone A and CX-1.

Computer-assisted analysis revealed an open-reading frame (ORF) starting from the first ATG and terminating at TAA as indicated. This ORF is preceded by an inframe stop codon at position -18, and ANN immediately before the first ATG. Signal sequences for

polyadenylation are also present.

Deduced Amino Acid Sequence. This ORF can encode for a 32,817 Da polypeptide of about 295 amino acid residues. The deduced amino acid sequence

substantially as shown in Fig. 3 revealed some

interesting features. At the N terminus, no leader signal sequence for entry into endoplasmic reticulum (ER) reported for numerous cell surface proteins is found. By hydropathy plots, the N-terminal two-thirds contain some hydrophobic segments that can serve as a signal sequence or provide membrane anchoring. No consensus sequence of Asn-XAA-(Ser/Thr) for N-linked glycosylation is present. There are only two Cys, separated by 14 amino acid residues. No arginine or lysine is found after amino acid residue 225; thus, a complete digestion by trypsin of this polypeptide should lead to a 70-amino acid, C-terminal domain of highly acidic polypeptide (13 asparagine plus glutamic

residues) without positively-charged residue.

Unusual sequences include two repeats of Thr-Glu-Asp-Trp-Ser-Ala-X-Pro (264-271 and 273-280(*)

Figure 3), noted previously (Wewer, U. M., Liotta,

L. A., Jaye, M., Ricca, G. A., Drohan, W. N., Claysmith, A. P., Rao, C. N., Wirth, P., Coligan, J. E.,

Albrachtsen, R., Mudryj, M. & Sobel, M. E. (1986) Proc. Nat. Acad. Sci. 83, 7137-7141); two repeats of Ala-Ala-Ala-X-X-Ala (91-96 and 216-221 (^) Figure 3); three repeats of Lys-Glu-Glu (11-13, 212-214 and 224-226), with two of them in Lys-Glu-Glu-Gln-X-X-X-Glu (212-219) and Lys-Glu-Glu-X-Gln-X-Glu (224-230); two repeats of Asp (Glu) -X-X-X-Glu (Asp) -X-Tyr-X-Tyr-Lys (Arg)-X-X-X- Asp (Glu) (31-44 and 196-209(:) Figure 3); one symmetrical sequence of Leu-Met-Trp-Trp-Met-Leu (173-178 (0)

Figure 3).

Increased levels of the 1.2 kb mRNA which hybridized with cDNA insert in the recombinant clone designated as 8-2V were detected in both poorly-differentiated and well-differentiated human colon carcinoma cell lines. At present, markers for poorly differentiated human colon carcinomas are needed since these cells express little carcinoembryonic antigen. The 1.2 kb mRNA which encodes for a laminin binding

protein, can distinguish colon carcinomas of both poorly and well differentiated from normal colon epithelial

cells based on their interactions with laminin.

Figures 5 and 6 show there is a significant level of expression of the 1.2 kB mRNA in human

esophageal, cervical, duodenum, lung, prostate carcinoma and human melanoma cell lines.

Figure 5 is a Northern blot of total RNA isolated from various human carcinoma cell lines. Total RNA

(15 μg/lane) was separated on a 1.0% agarose/formaldehyde gel, transferred to nitrocellulose filters, and probed with ³²P-labeled cDNA insert from clone 8-2V which encodes a laminin binding protein. Lanes 1-6 contained RNA isolated from the following carcinoma cell lines: 1, CE81T; 2,

CE48T; 3, HID (1, 2 & 3, human esophageal carcinoma);

4, HTB34MS751, 5, HTB35SiHa; 6, HeLa (4, 5 & 6, human cervical carcinoma). Arrowhead indicates the hybridized 1.2 Kb RNA.

Figure 6 is also a Northern blot of total RNA isolated from various human carcinoma cell lines. Total RNA (15 μg/lane) was separated on a 1.0% agarose/formaldehyde gel, transferred to nitrocellulose filters, and probed with ³²P-iabeled cDNA insert from clone 8-2V which encodes a laminin binding protein. Lanes 1-6 contained RNA isolated from the following carcinoma cell lines: 1, T24; 2, EJ (1 & 2, human bladder adenocarcinoma); 3, CX-1; 4, CCL218WiDr (3 & 4, human colon adenocarcinoma); 5, HTB40HuTu80, human duodenum adenocarcinoma; 6, CCL185A549; 7, 0AT4; 8, Calu (6, 7 & 8, human lung carcinoma); 9, A2058 (human melanoma);

10, CRL1435PC-3 (human prostate adenocarcinoma). Arrowhead indicates the hybridized species of 1.2 Kb RNA.

At the protein level, Figure 7 shows that the in vitro translated proαucts of laminin binding protein gene appeared as a 45 kd protein on SDS PAGE which bound to laminin sepharose column. In Figure 7, the in vitro

translated products of laminin binding protein gene obtained from rabbit reticulocyte lysate. The molecular weight of laminin binding protein in vitro translated products

migrated on SDS PAGE as a 45 Kd protein as shown in lane 2. Lane 1 is the background control of in vitro translation of this particular batch rabbit reticulocyte lysate. The

in vitro translated product was passed through laminin sepharose affinity column. After extensive wash, a 45 Kd band protein was eluted with 0.5 M NaCl buffered with 50 mM Tris pH 8.0 (lane 3).

Laminin receptor has been suggested to be involved in the initial interactions of tumor cells via laminin with the vascular basement membrane to

facilitate invasion and subsequent promotion of

metastasis. Increased levels of this mRNA in colon

carcinomas would be consistent with previous

observations that carcinoma cells tend to increase

expression of laminin receptor when compared with normal epithelial cells. Highly malignant, poorly

differentiated human carcinomas including that of colon have been shown to express high levels of cytoplasmic laminin receptor by immunoperoxidase stainings than that of well-differentiated. Moreover, inhibition of

metastasis of melanoma in animal model systems has been demonstrated by laminin-derived pentapeptide recognized by laminin receptor. It is possible that intervening interactions between laminin and its receptor prevent the metastasis of colon carcinomas.

The size of this mRNA is significantly smaller than the 1.7 kb mRNA previously described by Wewer et al. for laminin receptor. The ORF in cDNA described

here will only suffice for a 32,817 Da of polypeptide far smaller than the 67 kDa protein described for

laminin receptor. However, because all partial cDNA sequences reported for laminin receptor are included in the ORF described here, the product of 1.2 kb mRNA, if not the 67 kDa laminin receptor, is probably a laminin binding protein. That the product of this 1.2 kb mRNA might be a laminin binding protein is strongly supported by the fact that the coding region for an epitope in 67 kDa laminin receptor recognized by the monoclonal antibody, 2H5, which blocks the binding of laminin with its receptor is present entirely in the sequence

reported here. Moreover, the C-terminal 20-amino acid synthetic peptide recognized by polyclonal antibodies that interfere with laminin-mediated cell attachment is also present at the C-terminus of this sequence.

Since three laminin receptors with different molecular mass have already been reported (Kleinman et al., Proc. Natl. Acad. Sci. (USA), pages 1282-1286, Vol. 85 (1988) and Smalheiser et al., Proc. Natl. Acad. Sci. (USA), pages 6457-6461, Vol. 84 (1987)), it is of interest to consider whether all receptors share common sequences. While most of the blot-hybridizations of RNA gels shown here revealed a predominant band of 1.2 kb, there are minor bands of larger sizes. Whether these bands are related to other laminin receptors requires further investigation. Furthermore, a 5.5 kb mRNA has been identified which hybridized strongly with 8-2V cDNA in human esophageal carcinomas.

The deduced amino acid sequence of Figure 3 revealed some interesting features. At the N terminus no leader signal sequence for entry into the endoplasmic reticulum reported for numerous cell surface proteins was found. There is no obvious transmembrane regions similar to other membrane proteins, or amphipathic alpha helices. If this amino acid sequence is one of the laminin receptors it might be associated with membrane by other pathways. By hydropathy plots, the N-terminal two-thirds of the molecule contained some hydrophobic segments that might serve as a signal sequence or provide membrane anchoring. For example, there is a sequence of highly hydrophobic Leu-Leu-Leu-Ala-Ala-Arg-Ala-Ile-Val-Ala-Ile as a part of 19-amino acid segment which might provide association with the membrane. The other segment for potential membrane anchoring is the N-terminal 15-30 residues which contains many hydrophobic residues. However, it. is also possible that laminin receptor might use myristate or palmitate for membrane anchoring. No consensus sequence of Asn-X-Ser (Thr) for N-linked glycosylation was present. There are 24 Thr and 14 Ser for potential O-linked glycosylations.

Asp(Glu)-X-X-X-Glu(Asp)-X-Tyr-X-Tyr-Lys(Arg)-X-X-X-Asp(Glu) in positions 31-44 and 196-209 was a notable repeat. If the 140 amino acid stretch flanked by the two cysteines is considered a linking domain, the two flanking domains would each have one of this repeat sequence. There is another repeat, Ala-Ala-Ala-X-X-Ala-X(-)-Thr, one each in these two flanking domains.

The C-terminal domain after residue 225 contains a 70-amino acid segment with 13 aspartic plus glutamic residues and without any lysine or arginine residues. This trypsin-resistant, highly acidic domain might mark laminin binding proteins. Within this 70-amino acid domain, there are five repeats of Asp(Glu)-Trp-Ser(Thr), with two of them in nearly tandem repeats of Thr-Glu-Asp-Try-Ser-Ala-X-Pro.

In another aspect, this invention concerns a reagent and method for detecting carcinoma in a patient which comprises providing a labelled cDNA probe having one of the cDNA sequences mentioned above, exposing the probe to a sample of tissue or body fluid from the patient, and monitoring the reaction for hybridization. Hybridization selection techniques are described in Maniatis et al., Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory (1982).

Useful labels include any moiety detectable subsequent to hybridization such as radionuclides, e.g., ³²P, ³H, or ¹⁴C, enzymes, fluorophores, chemiluminescent compounds, chromogens and chromophores. These labels can be attached directly or indirectly to the probe using techniques commonly known to those skilled in the art. It is also within the skill of the art to identify the probe using specific labelled antibodies.

The Examples below exemplify the invention.

EXAMPLE 1

Tissues and Cell Lines. Surgical specimens of primary human colon carcinomas and adjacent normal colon tissues were obtained from Dept. of Surgery at New

England Deaconess Hospital. Immediately after surgery, normal colonic epithelium was dissociated from muscle and connective tissues under dissection microscope and placed in liquid nitrogen (Iversen, P. L., Mata, J. E. & Hines, R. N. (1987) BioTechnique 5, 521,523). Human colon carcinoma cell lines, Clone A and Clone D,

subclones of DLD-1, were from D. Dexter (Du Pont Co.), MIP101 from N. Zamcheck (Mallory Institute), CX-1 from S. Bernal (Dana-Farber Cancer Institute), RCA, HCT116b, Gly and Moser from M. Brattain (Baylor College of

Medicine), HT29, CL187, CCL220.1, CCL222, CCL224,

CCL227, CCL228, CCL229, CCL233, CCL234, CCL237 and HTB39 from American Type Culture Collection. All these cell lines have been well characterized and documented, and are of human colon origin. All cells were grown in 50%

Dulbecco's modified Eagles' medium (GIBCO) and 50% RPMI 1640 MEDIUM (GIBCO) supplemented with 10% fetal calf serum (MA Bioproducts) and 1% Nutridoma NS (Boehringer) at 37°C with 5% CO₂ and 100% humidity. Preparation of RNA. Total cellular RNAs from cultured cells were isolated by the methods of Cox (Cox, R. A. (1986) Methods Enzym. 12, 120-129) and Strohman et al. (Strohman, R. C, Moss, P. S., Eastwood, J. M. & Spector, D. (1977) Cell 10, 265-273). Total cellular RNAs from surgical specimens were prepared according to Chirgwin et al . (Chirgwin, J . M . , Pryzbyla, A . E . ,

MacDonald, R. J. & Rutter, W. J. (1979) Biochemistry 18, 5294-5299) . Normal and tumor tissues were ground into sand-like particles in the presence of liquid nitrogen before extraction with guanidine isothiocyanate.

Construction of pBR322 Clone A cDNA Library. Poly (A)⁺RNAs from Clone A cells were purified by a poly(U)-Sephadex column twice as described by Haff et al. (Haff, L. A. & Bogorad, L. (1976) Biochem. 15, 4110-4115). The first strands of cDNAs were synthesized by AMV reverse transcriptase (Molecular Genetics Resources) using oligo-dT as a primer, the second strands with DNA polymefase I according to Wickens et al. (Wickens, M. P., Buell, G. N. & Schimke, R. T. (1978) J. Biol. Chem. 263, 2483-2495). cDNAs were treated with S1 nuclease, tailed with dCTP, and fractionated by Bio-Gel A-50 column. Only dC-tailed cDNAs larger than 400 bp were pooled and annealed to Pst I-digested poly(G)-tailed pBR322 plasmids. Transformation in E. coli HB101 was carried out according to Dagert et al. (Dagert, M. & Ehrlich, S. D. (1979) Gene 6, 23-28). The complexity of this library was approximately 5X10^ clones/μg cDNA.

Screening the Plasmid cDNA Library. The individual transformed E. coli colonies were transferred onto a master agar plate and two other filter plates containing tetracycline (Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, CSH, NY)). The colonies were arranged in a grid pattern. Each colony was streaked in an identical position on all three plates. The colonies on the nitrocellulose filters were lysed with alkali and fixed. 32p-iabeled first strand cDNAs generated from Poly(A)⁺RNAs of Clone A or CX-1 were used for colony hybridization. A total of 1,500 colonies were screened.

RNA Gels Blot-Hybridization. Total RNAs (15 μg) extracted from tissues and cultured cells were electrophoresed in agarose gels containing formaldehyde, and transferred to GeneScreenPlus (Du Pont) according to Seed (Seed, B. (1982) Genet. Eng. 4, 91-102).

Recombinant plasmids were prepared by alkaline lysis and CsCl gradient centrifugation (Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A

Laboratory Manual (Cold Spring Harbor Laboratory, CSH, NY)). Plasmid DNA was purified and digested by Pst I. After electrophoresis, cDNA insert was eluted with LID/X filter syringes (Xydex), extracted with

phenol/chloroform, concentrated by ethanol

precipitation, and suspended in TE buffer (Maniatis, T., Fritsch, E. F. & Sambrock, J. (1982) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, CSH, NY)). Nick translated, ³²P-labeled cDNA insert was used to hybridize blots on GeneScreenPlus. Specific

activities of the DNA probes were approximately 2-5X10⁸ cpm/μg of DNA. Densitometry readings were made with a LKB UltroScan XL Laser Densitometer (LKB).

Construction of CX-1 cDNA lambda gt10 phage library.

Total RNAs were isolated by extraction of cultured human colon carcinoma cell line, CX-1, with guanidinium thiocyanate. Poly(A)+RNAs were selected on oligo(dT)-cellulose affinity columns.

(a) First strand synthesis: Synthesis of first strand cDNA was carried out in a reaction (200 μl volume) containing 50 mM Tris-HCl, pH 8.3, 50 mM KCl, 6 mM MgCl₂, 10 mM DTT, C.5 mM dNTPs, 50 μg/ml of

oligo(dt) 12-18, 100 μ/ml of RNAs in 10 μg of poly(A)+RNA, and 1000 μ/mi of avian reverse transcriptase for 1.5 h at 42°C. To minimize frequent nicking of cDNA labelled with radioactive nucleotides, the main reaction for the first strand synthesis was not radiolabelled. Labelling was done only, in a separate tube, to monitor the incorporation reaction containing the same

components as above with an addition of 10 uCi of

[³²P]dCTP (specific activity: 3000 Ci/umol). The reactions were terminated by adding SDS and EDTA to the final concentration of 0.4% and 10 mM, respectively. The products were extracted with phenol and

chloroform/isoamylalcohol (24/1) and chromatographed on a 5 ml column of G-50 Sephadex to remove unincorporated nucleotide triphosphates. cDNA fractions in the void volume were pooled and concentrated by ethanol

precipitation. Based on the radioactivity incorporated in ethanol precipitated cDNA, approximately 0.1 μg of the first strand cDNA per μg of added poly(A)+ RNA was produced.

(b) Second Strand Synthesis: The cDNA pellet from first strand reaction was dissolved in water and the second strand synthesis conducted in a reaction (400 μl volume) containing 0.5-1.0 μg of first strand cDNA, 20 mM Tris-HCl, pH 7.5, 100 mM KCl, 5 mM MgCl₂, 10 mM DTT, 10 mM (NH⁴)₂SO₄, 50 μg/ml BSA (nuclease-free), 150 μM B-NAD μ/ml RNAse H, 50 μ/ml E. Coli DNA ligase, 40 μM dNTPs, 20 uCi [³²P]dCTP (specific activity: 3000

Ci/umol), and 250 μ/ml T4 DNA polymerase. Incubation was at 15°C overnight and terminated with 0.4% SDS and 10 mM EDTA. Purification of double-stranded cDNA from unincorporated nucleotide triphosphates was achieved by Sephadex G-50 column chromatography as described above. The approximate yield of ds-cDNA from poly(A)+ RNA was 15-20%.

(c) Fill-in and Methylation Reaction: In cloning full-length cDNA inserts into lambda gt10 vectors at the EcoRI site, it was necessary to generate blunt-ended cDNA for the adaptation of blunt-ended EcoRI linkers and to protect the internal EcoRI sites of cDNA by

methylation. The fill-in reaction was performed in a reaction (80 μl volume) consisting of 30 mM

Tris·Acetate, pH 7.8, 60 mM K-acetate, 10 mM MgCl₂, 0.2 mM DTT, 0.1 mg/ml BSA, 0.1 μg/ml RNAse A, 100 μ/ml RNAse H, 50 μM B-NAD, 250 μ/ml EcoRI DNA ligase, 300 μM dNTPs and 500 μ/ml T4 DNA polymerase at 37°C for 30 min. The cDNA was recovered by EtOH precipitation and

subsequently methylated in a buffer containing 50 mM Tris-HCl, pH 8.0, 5 mM EDTA, 5 mM DTT, 0.1 mM Naacetate, 50 μM S-adenosyl-methionine, and 1000 μ/ml EcoRI methylase at 37°C for 30 min. The cDNA was recovered by EtOH precipitation and subsequently

methylated in a buffer containing 50 mM Tris-HCl, pH 8.0, 5 mM EDTA, 5 mM DTT, 0.1 mM Na-acetate, 50 μM S-adenosyl-methionine, and 1000 μ/ml EcoRI methylase at 37°C for 1 h.

(d) Ligation of EcoRI Linkers to cDNA: To

generate external EcoRI sites on cDNAs, blunt-ended ligation of 5'-phosphorylated EcoRI linkers (0.5 μg/ml) to cDNAs was carried out in the presence of 1 mM ATP, 200 units of T4 DNA ligase, in a standard ligation buffer (50 mM Tris-HCl, pH 7.6, 10 mM MgCl2, 20 mM DTT at 15°C overnight.

(e) Digestion of EcoRI-linked cDNA with EcoRI Endonuclease: The above ligation products were digested with EcoRI endonuclease to expose the external EcoRI sites on cDNAs. The digestion reaction contained 50-100 units of EcoRI endonuclease per 0.1-0.5 μg cDNA in a standard EcoRI buffer (ICO mM Tris-HCl, pH 7.5, 100 mM NaCl, 5 mM MgCl₂). The incubation was 4-8 h at 37°C after which EcoRI enzyme was inactived by heating at 68°C for 15 min.

(f) Purification and Size-Fractionation of EcoRI- digested cDNA by CL-4B column chromatography: EcoRI- digested cDNAs were purified and selected for desirable size for cDNA cloning using CL-4B-Sepharose columns. The purpose of this column purification was (i) to isolate and recover cDNA from excess polymer linkers and/or digested linkers (usually less than 200 bp in size) which interfere with effective ligation of the cDNA to vector DNA and (ii) to fractionate cDNAs by size prior to cloning into specific vector. CL-4B-200

Sepharose resin was prepared in 5 ml pipet and

equilibrated with column buffer (10 mM Tris·HCl, pH 8.0, 600 mM NaCl, 1mM EDTA, and 0.1% Sarkosyl). EcoRI-digested cDNA samples were loaded in 0.3 ml column buffer, and twenty of 100 μl fractions collected. Each fraction was monitored for radioactivity and analyzed for size on 1% agarose gel. For our purpose of

constructing total cell cDNA libraries in lambda gt10 vectors, cDNAs longer than 500 bp were selected and pooled. The recovered cDNAs were then concentrated by ethanol precipitation and dissolved in a small volume of

10 mM Tris-HCl, pH 7.5/1 mM EDTA buffer.

(g) Ligation of Selected cDNA Inserts to lambda gt10 Vectors: Cloning of cDNA libraries into lambda gt10 vector was accomplished by ligation of EcoRI-digested, column-purified and size-selected cDNA inserts to lambda gt10 arms previously EcoRI-cleaved and 5'-dephosphorylated. The ligation reaction (10 μl)

contained vector DNA, 0.1-0.5 μg insert cDNA, 1 mM ATP, 1-10 units of T4 DNA ligase in a standard ligation buffer. Incubation was carrier out at 15°C overnight. Screening of Phage cDNA Library. ³²P-labeled cDNA insert from plasmid 8-2V was used to screen CX-1 lambda gt10 library (Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Laboratory, CSH, NY)).

Approximately 10,000 plaques were screened.

DNA Sequencing. M13 dideoxyoligonucleotide chain termination method was used (Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Nat. Acad. Sci. 74, 5463-5467). Both inserts, 8-2V and J-9, were sequenced from both directions with the M13 universal primer or with four specifically synthesized 17-base

oligonucleotides which had the following sequences:

#1 ²¹⁵CTGATGTCAGTGTTATA²³¹

#2 ²⁵¹TGGCCAGTATTCCTGGA²³⁵ (complementary strand) #3 ⁶⁴⁷CTGCTGCTGAGAAGGCA⁶⁶³

#4 ⁵⁹⁷GAGTCCATTCACCCTGA⁶⁸¹ (complementary strand) M13 was from New England Biolab and Sequenase sequencing kit from United States Biochemical. Synthetic

oligonucleotides were kindly provided by P . H. Sayre in the laboratory of E. Reinherz (Dana-Farber Cancer

Institute).

Computer-assisted Sequence Analysis was performed at the Molecular Biology Computer Research Resource/Harvard School of Public Health at Dana-Farber Cancer Institute.

Claims

WHAT IS CLAIMED IS:

1. A full length human laminin binding protein cDNA sequence derived from human colon carcinoma cells comprising substantially the sequence as set forth in Figure 3.

2. A subsequence of the cDNA sequence of claim 1 corresponding substantially to the sequence starting with nucleotide number 162 and ending with nucleotide number 829 as set forth in Figure 3.

3. A subsequence of the cDNA sequence of claim 1 corresponding substantially to the sequence starting with nucleotide number 1 and ending with nucleotide 43 as set forth in Figure 3.

4. A recombinant clone designated J-9 having ATCC Accession No. 40489.

5. A recombinant clone designated 8-2V having ATCC Accession No. 40490.

6. A laminin binding protein encoded by the sequence of claim 1, 2 or 3.

7. A substantially purified DNA sequence encoding the laminin binding protein of claim 6.

8. A method for detecting the presence of carcinoma in a patient comprising:

a. providing a labelled cDNA probe having the

cDNA sequence of claim 1, 2 or 3;

b. exposing the probe to a sample of tissue or body fluid from the patient; and c. monitoring the reaction or step b. for hybridization.

9. A method according to claim 8 wherein the carcinoma is colon carcinoma.

10. A reagent for detecting the presence of carcinoma in a patient comprising a labelled cDNA probe having the cDNA sequence of claim 1, 2 or 3.