Methods of expressing integrin β6 subunits

Abstract

The present invention provides substantially pure integrins containing a novel β subunit designated as β6. The novel β6 subunit forms heterodimers with αv and αf. Methods of controlling cell adhesion using the β6-containing integrins are also provided.

Description

BACKGROUND ART

This invention relates to receptors for adhesion peptides, and more specifically to a novel receptor subunit having affinity for extracellular matrix molecules.

Multicellular organisms, such as man, have some 10¹⁴ cells which can be divided into a minimum of fifty different types, such as blood cells and nerve cells. During the course of growth and development, cells adhere to other cells, or to extracellular materials, in specific and orderly ways. Such cell adhesion mechanisms appear to be of importance in mediating patterns of cellular growth, migration and differentiation, whereby cells develop specialized characteristics so as to function as, for example, muscle cells or liver cells. Cell adhesion mechanisms are also implicated in dedifferentiation and invasion, notably where cells lose their specialized forms and become metastasizing cancer cells.

The mechanisms underlying the interactions of cells with one another and with extracellular matrices are not fully understood, but it is thought that they are mediated by cell surface receptors which specifically recognize and bind to a cognate ligand on the surface of cells or in the extracellular matrix.

The adhesion of cells to extracellular matrices and their migration on the matrices is mediated in many cases by the binding of a cell surface receptor to an Arg-Gly-Asp containing sequence in the matrix protein, as reviewed in Ruoslahti and Pierschbacher, Science 238:491-497 (1987). The Arg-Gly-Asp sequence is a cell attachment site at least in fibronectin, vitronectin, fibrinogen von Willibrand, thrombopondin, osteopontin, and possibly various collagens, laminin and tenascin. Despite the similarity of their cell attachment sites, these proteins can be recognized individually by their interactions with specific receptors.

The integrins are a large family of cell surface glycoproteins that mediate cell-to-cell and cell-to-matrix adhesion as described, for example, in the Ruoslahti and Pierschbacher article cited above. All known members of this family of adhesion receptors are heterodimers consisting of an α and a β subunit noncovalently bound to each other. When the integrin family was first identified, integrins were grouped into three subfamilies based on the three β subunits that were initially recognized (β₁, β₂ and β₃). Over the past few years, the primary structures of three integrin β subunits from mammalian cells and one from Drosophila have been deduced from cDNA.

Each α subunit was thought to associate uniquely with a single β subunit. Eleven distinct α subunits have thus far been described. As new integrins have been identified, however, it has become clear that this grouping is not entirely satisfactory, since there are clearly more than three β subunits and since some α subunits can associate with more than one β subunit as described, for example, in Sonnenberg et al., J. Biol. Chem. 265:14030-14038 (1988).

Because of the importance of integrins in mediating critical aspects of both normal and abnormal cell processes, a need exists to identify and characterize different integrins. The present invention satisfies this need and provides related advantages as well.

SUMMARY OF THE INVENTION

The present invention relates to a substantially purified β subunit of an integrin cell surface receptor designated as β₆. The amino acid sequence of human β₆ (SEQ ID NO:27) is provided in FIG. 3.

The present invention also relates to amino acid fragments specific to β₆ that have a variety of uses. The invention further relates to vectors having a gene encoding such fragments. Host cells containing such vectors are also provided. The nucleic acids encoding β₆ as well as nucleic acids that specifically hybridize with the nucleic acids encoding β₆ sequences are other aspects of the present invention.

In a further aspect, the present invention relates to a substantially purified integrin comprising β₆ bound to an α subunit, particularly α_v or α_F. Methods of blocking the attachment of the β₆-containing integrins to its ligand and of detecting the binding of such integrins to its ligand are also provided.

The present invention also relates to methods of increasing or decreasing cell adhesion in cells expressing a β₆-containing integrin by overexpressing the integrin or by binding the integrin with a ligand, such as vitronectin.

BRIEF DESCRIPTION OF THE FIGURES

FIG. 1 shows the design of consensus PCR primers (SEQ ID NOS:1-5, 7 and 8)(β₂ human nucleic acids=SEQ ID NOS:10, 14, 18 and 22; corresponding β₂ human amino acids=SEQ ID NOS:50, 51, 54 and 55; βhuman nucleic acids=SEQ ID NOS:11, 15, 19 and 23; corresponding β₃ human amino acids=SEQ ID NOS:52, 53, 56 and 57; β₁ human nucleic acids=SEQ ID NOS:12, 16, 20 and 24; corresponding β₁ human amino acids=SEQ ID NOS:50, 53, 58 and 59; β₁ chicken nucleic acids=SEQ ID NOS:13, 17, 21 and 25; corresponding β₁ chicken amino acids=SEQ ID NOS:50, 53, 60 and 59: β₆ guinea pig sequence from position 219=SEQ ID NO:6; corresponding β₆ guinea pig amino acids=SEQ ID NO:61; β₆ guinea pig sequence from position 1325=SEQ ID NO:8; corresponding 136 guinea pig amino acids=SEQ ID NO:62).

FIG. 2 shows a map of sequencing strategy.

FIG. 3 shows the nucleotide sequence and amino acid translation for human (H) (SEQ ID NOS:26 and 27) and guinea pig (GP) (SEQ ID NOS:28 and 29) β₆ .

FIG. 4 shows the alignment of human β₆ (SEQ ID NO:27) with, four previously reported integrin β subunits (human β₁ =SEQ ID NO:30; human β₂=SEQ ID) NO:31; human β₃=SEQ ID NO:32; Drosophila β_myo=SEQ ID NO:33).

FIG. 5 shows the alignment of partial nucleotide and amino acid sequences from human (H) and guinea pig (GP) β₁ (human(β_1H)=SEQ ID NOS:34 and 35; guinea pig (β_1GP)=SEQ ID NOS:36 and 37, respectively), β₃ (human(β₃)=SEQ ID NOS:38 and 39; guinea pig (β_6GP)=SEQ ID NOS:40 and 41, respectively), and β₆ (human(β_6H=SEQ ID NOS:42 and 43; guinea pig (β_6GP)=SEQ ID NOS:44 and 45, respectively) for the region just downstream from the B3F primer.

DETAILED DESCRIPTION OF THE INVENTION

The present invention provides a composition of matter relating to a novel, substantially purified integrin β subunit, referred to herein as β₆. The amino acid sequence of 62 ₆ for human (SEQ ID NO:27) and for guinea pig (SEQ ID NO:29) are also provided and are shown in FIG. 3.

By “substantially purified” is meant substantially free of contaminants normally associated with a native or natural environment.

By “β₆” is meant a polypeptide having substantially the same amino acid sequence and binding functions of the polypeptides encoded by the sequences set forth in FIG. 3 for human (SEQ ID NO:26) and guinea pig (SEQ ID NO:28) β₆. Thus, modified amino acid sequences that do not substantially destroy the functions and retain the essential sequence of β₆ are included within the definition of β₆. Amino acid sequences, such as the sequence for β₁ (SEQ ID NO:30), β₂ (SEQ ID NO:31) and β₃ (SEQ ID NO:32), having less than 50% homology with the sequence of 62 ₆, are not substantially the same sequence and, therefore, do not fall within the definition of 62 ₆. Given the amino acid sequences set forth herein, additions, deletions or substitutions can be made and tested to determine their effect on the function of β₆. In addition, one skilled in the art would recognize that certain amino acids, such as the conserved cystines, for example, can be modified to alter a binding function of β₆.

Amino acids are identified herein by the standard one-letter abbreviations, as follows:

Amino Acid    Symbol
Alanine       A
Asparagine    N
Aspartic acid D
Arginine      R
Cysteine      C
Glutamine     Q
Glutamic acid E
Glycine       G
Histidine     H
Isoleucine    I
Leucine       L
Lysine        K
Methionine    M
Phenylalanine F
Proline       P
Serine        S
Threonine     T
Tryptophan    W
Tyrosine      Y
Valine        V

Based on its amino acid sequence, the β subunit of the present invention is clearly different from β₁, β₂, β₃ and other β subunits that have recently been discovered. For example, the 11-amino acid carboxyl-terminal extension on β₆ distinguishes it from β₁, β₂, and β₃. The short cytoplasmic tails of β₁, β₂, and β₃ are thought to be sites of interaction with the cytoskeleton and regions for the transduction of signals initiated by interactions of the large extracellular domains with ligands. These cytoplasmic tails may also be targets for regulation of integrin function. The distinctive 11-amino acid cytoplasmic tail of β₆ indicates that its regulation or pathways for signal transduction may be different from those of β₁, β₂ and β₃.

In addition to β₁, β₂ and β₃. recent studies have suggested the existence of as many as five other integrin β subunits. A β subunit with a molecular weight of approximately 210,000 (β₄) has been found associated with the integrin α subunit “α₆” in colon carcinoma cells and in a variety of other tumor cells of epithelial origin as described, for example, in Kajiji et al., EMBO J., 8:673-680 (1989). On the basis of its high molecular weight, 210,000 compared with the predicted size of 106,000 of the subject novel protein, and on the basis of its clearly different amino-terminal sequence, it is apparent that β₄ is not the same as the subject polypeptide.

Another β subunit, originally called β_x was identified in epithelial-derived tumor cells in association with the integrin α subunit α_v as described, for example, in Cheresh et al., Cell 57:59-69 (1989). This β subunit, having a distinctive amino-terminal sequence, was recently renamed β₅. Based on recent studies of purified preparations, β₅ clearly differs from the β subunit of the present invention. Because the β subunit described in the present report is distinct from each of the five β subunits for which sequence information is available, it has been designated as β₆.

The existence of two other integrin β subunits has been inferred from the identification of unique proteins after immunoprecipitation of surface-labeled cell lysates with antibodies to known α subunits. One of these novel proteins, called β_s was found in association with α_v in the human osteosarcoma cell line MG-63, in the fibroblast cell line AF1523, and in human endothelial cells as described, for example, in Freed et al., EMBO J. 8:2955-2965 (1989). This subunit is also different from B₆ since β_s is expressed in MG-63 cells while β₆ is not expressed in these cells as shown in Table 1.

The other novel integrin B subunit identified by co-immunoprecipitation of known α subunits, β_P, is a protein of about M_r 95,000 that is found to be associated with α₄, an α subunit first found as part of the lymphocyte homing receptor VLA-4 as described, for example, in Holzmann et al., Cell 45:37-46 (1989). This subunit is also distinct from β₆ since β_P is expressed in lymphocytes while β₆ is not expressed in lymphocytes as shown in Table 1.

TABLE 1
Distribution of β6
	Type	Results	Source
Cell Lines:
FG-2	Pancreatic	+	Kajiji et al.,
			EMBO J. 3: 673–80
			(1989)
Panc I	Pancreatic	−	Dr. Metzgar,
			Duke U., N.C.
Colo-396	Colon CA	+	Dr. L. Walker,
			Cytel, San
			Diego, CA
UCLA P3	Lung CA	+	Dr. L. Walker,
			Cytel, San
			Diego, CA
HeLa	Cervical	−	ATCC #CCL-2
Jar	Chorio CA	+	ATCC #HTB 36
HT 1080	Fibrosarcoma	−	ATCC #CCL 121
U 937	Monocytoid	−	ATCC #CRL 1593
M 21	Melanoma	−	Dr. R. Reisfeld,
			Scripps Clinic
			& Research
			Foundation, La
			Jolla, CA
B 16	Melanoma	−	Dr. R. Reisfeld
			Scripps Clinic
			& Research
			Foundation, La
			Jolla, CA
MG 63	Osteosarcoma	−	ATCC #CRL 1427
Tissues:
	Cervix	+
	Aortic Endothelium	−
	Leukocytes	−

The invention also provides an integrin comprising β₆ bound to an α subunit. β₆, consistent with recent findings of other β subunits, can associate with a variety of α subunits to form a functional integrin. In one embodiment, β₆ associates with α_v. In another embodiment, β₆ associates with another α subunit referred to herein as α_F. The α_v β₆ integrin, as well as other integrins containing β₆, can bind molecules, for example extracellular matrix molecules. Such molecules are referred to herein as ligands. In a specific embodiment, certain β₆-containing integrins can bind Arg-Gly-Asp-containing polypeptides such as vitronectin or fibronectin. The binding of β₆-containing integrins to various ligands can be determined according to procedures known in the art and as described for example, in Ruoslahti and Pierschbacher, Science 238:491-497 (1987).

The invention also provides an amino acid fragment specific to β₆. Since β₆ is a novel molecule, it contains many fragments which are specific for this β subunit. Fragments specific to β₆ contain sequences having less than 50% homology with sequences of other known integrin β subunit fragments. These fragments are necessarily of sufficient length to be distinguishable from known fragments and, therefore, are “specific for β₆.” The amino acid sequence of such fragments can readily be determined by referring to the figures which identify the β₆ amino acid sequences. These fragments also retain the binding function of the β₆ subunit and can therefore be used, for example, as immunogens to prepare reagents specific for β₆ or as an indicator to detect the novel β₆-containing integrin of the present invention. One skilled in the art would know of other uses for such fragments.

The invention also provides a reagent having specificity for an amino acid sequence specific for β₆. Since β₆ is a novel protein with at least 50% amino acid differences over related β subunits, one skilled in the art could readily make reagents, such as antibodies, which are specifically reactive with amino acid sequences specific for β₆ and thereby immunologically distinguish β₆ from other molecules. Various methods of making such antibodies are well established and are described, for example, in Antibodies, A Laboratory Manual, E. Harlow and D. Lane, Cold Spring Harbor Laboratory 1988, pp. 139-283 and Huse et al., Science 24:1275-1280 (1988).

The invention also provides nucleic acids which encode β₆. Examples of such sequences are set forth in FIG. 3 (SEQ ID NOS:26 and 28). Following standard methods as described, for example, in Maniatis et al., Molecular Cloning, Cold Spring Harbor (1982), nucleic acid sequences can be cloned into the appropriate expression vector. The vector can then be inserted into a host, which will then be capable of expressing recombinant proteins. Thus, the invention also relates to vectors containing nucleic acids encoding such sequences and to hosts containing these vectors.

The sequences set forth in FIG. 3 (SEQ IS NOS:26 and 28) also provide nucleic acids that can be used as probes for diagnostic purposes. Such nucleic acids can hybridize with a nucleic acid having a nucleotide sequence specific for β₆ but do not hybridize with nucleic acids encoding non-β₆ proteins, particularly other cell surface receptors. These nucleic acids can readily be determined from the sequence of β₆ and synthesized using a standard nucleic acid synthesizer. Nucleic acids are also provided which specifically hybridize to either the coding or non-coding DNA of β₆.

Integrin cell surface receptors bind ligands, such as extracellular matrix molecules. However, the binding of the integrin to the ligand can be blocked by various means. For example, the binding of a β₆-containing integrin can be blocked by a reagent that binds the β₆ subunit or the β₆-containing integrin. Examples of such reagents include, for example, Arg-Gly-Asp-containing peptides and polypeptides, ligand fragments containing the integrin binding site, as well as antibodies specifically reactive with β₆ or a β₆-containing integrin. Alternatively, the blocking can be carried out by binding the ligand or fragment thereof, recognized by a β₆-containing integrin with a reagent specific for the ligand at a site that inhibits the ligand from binding with the integrin. Since the binding of a β₆-containing integrin to its ligand can mediate cell adhesion to an extracellular matrix molecule, preventing this binding can prevent cell adhesion. Alternatively, cell adhesion can be promoted by increasing the expression of β₆-containing integrins by a cell.

Finally, the invention provides a method of detecting ligands which bind a β₆-containing integrin. The method comprises contacting a β₆-containing integrin with a solution containing ligands suspected of binding β₆-containing integrins. The presence of ligands which bind a β₆-containing integrin is then detected.

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

EXAMPLE I

Identification of a Novel β Subunit

A. Generation of cDNA Fragments by Polymerase Chain Reaction

Tracheal epithelial cells, harvested from male Hartley outbred guinea pigs (Charles River Breeding Laboratories, Bar Harbor, Me.) were grown to confluence over 10-14 days on collagen-impregnated microporous filters commercially available from Costar. RNA was harvested from these primary cultures, and mRNA was purified over oligo(dT)-cellulose columns using the Fast Track mRNA isolation kit (Invitrogen, San Diego, Calif.). Two to 5 μg of mRNA was used as a template for cDNA synthesis catalyzed by 200 units of Moloney murine leukemia virus reverse transcriptase (Bethesda Research Laboratories, Gaithersburg, Md.) in a 20-40 μl reaction volume. One to 5 μl of the resultant cDNA was used as a template for polymerase chain reaction (PCR). PCR was carried out in a reaction volume of 25-200 μl. In addition to the template cDNA, each PCR reaction contained 50 mM KCl, 10 mM Tris-HCl (pH 9.0 at 25° C.), 1.5 mM MgCl₂, 0.01% gelatin, 0.1% Triton X-100, 0.2 mM each of dATP, dGTP, dCTP and dTTP, and 0.05 units/μl Taq DNA polymerase (obtained from either United States Biochemical Corporation, Cleveland, Ohio, or from Promega, Madison, Wis.).

For each reaction, two oligonucleotide primers were also added to obtain a final concentration of 1 μM each. The primer pairs are identified below. Each reaction mixture was overlaid with mineral oil, heated to 950° C. for 4 min. in a thermal cycler (Ericomp, San Diego, Calif.), and then subjected to 30 cycles of PCR. Each cycle consisted of 45 seconds at 95° C., 45 seconds at 53° C., and 1 min. at 72° C. Immediately after the last cycle, the sample was maintained at 72° C. for 10 min.

The results of each PCR reaction were analyzed by gel electrophoresis in 1.5% agarose. Reactions that produced fragments of the expected size were electrophoresed in 1.5% low gel temperature agarose (Bio-Rad Laboratories, Richmond, Calif.). The appropriate size band was excised, melted at 68° C., and the DNA was purified by extraction with phenol/chloroform and precipitation in ethanol and ammonium acetate.

B. PCR Primers

To obtain the initial fragment of the novel β subunit cDNA described herein, degenerate mixtures of PCR primers were used. Oligonucleotides were synthesized, trityl-on, by the University of California, San Francisco Biomolecular Resource Center using a DNA synthesizer with standard procedures, and purified over Nen-sorb cartridges (DuPont-New England Nuclear, Boston, Mass.). These consensus primer mixtures were designed to anneal with the nucleotides encoding the highly conserved sequence Asp-Leu-Tyr-Tyr-Leu-Met-Asp-Leu (SEQ ID NO:50) (primer B1F) (SEQ ID NO:1) and Glu-Gly-Gly-Phe-Asp-Ala-Ile-Met-Gln (SEQ ID NO:53) (primer B2R) (SEQ ID NO:2) that flank an approximately 300-nucleotide region beginning approximately 130 amino acids from the amino terminus of each of the integrin β subunits sequenced to date. The sequences of the primers identified herein are depicted in FIG. 1 (SEQ ID NOS:1-8).

On the basis of the initial sequence obtained, a specific forward primer was designed to anneal with the sequence encoding the amino acids Pro-Leu-Thr-Asn-Asp-Ala-Glu-Arg (SEQ ID NO:61) (primer BTE2F) (SEQ ID NO:7) ending approximately 49 nucleitides from the 3′ end of the region that had been sequenced. An additional forward primer (B3F) (SEQ ID NO:3) and two reverse primers (B3R and B4R) (SEQ ID NOS:4 and 5) were also designed to recognize highly conserved consensus regions encoding the sequences Gly-Glu-Cys-Val-Cys-Gly-Gin-Cys (SEQ ID NO:58) (B3 region) (SEQ ID NOS:3 and 4) and Ile-Gly-Leu-Ala-Leu-Leu-Leu-Trp-Lys (SEQ ID NO:59) (B3 region) (SEQ ID NO:5). The alignment of these primers with previously published sequences of human β₁, β₂, and β₃ and chicken β₁ is shown in FIG. 1. PCR as described above was performed with cDNA from guinea pig trachel epithelial cells and the primer pairs BTE2F/B3R (SEQ ID NOS:7 and 4) and B3F/B4R (SEQ ID NOS:3 and 5).

The primer pair BTE2F/B3R (SEQ ID NOS:7 and 4) yielded 1095 additional base pairs of new sequence. Based on this sequence another specific primer (BTE3F) (SEQ ID NO:8) was designed to recognize the sequence Val-Ser-Glu-Asp-Gly-Val (SEQ ID NO:9) near the 3′ end of this sequence, and PCR was performed with this primer in combination with primer B4R (SEQ ID NO:5).

FIG. 1 shows the design of the PCR primers. β subunit consensus primer mixtures were designed on the basis of alignment of published sequences of human β₁, β₂, β₃ and chicken β₁. For forward primers (B1F and B3F) (SEQ ID NOS:1 and 3), the primer sequences included a single nucleotide whenever possible for each of the first two nucleotides of each codon and were usually either degenerate or included deoxyinosine for the third base in codons for amino acids other than methionine. Reverse primers (B2R, B3R, and B4R) (SEQ ID NOS:2, 4 and 5) were designed in the same manner for the complementary DNA strand. Two specific forward primers were designed to recognize β₆. The first (BTE2F) (SEQ ID NO:7) was designed to work across species and was thus degenerate or included deoxyinosine in the third codon position. The second, BTE3F (SEQ ID NO:8), was not degenerate and was designed to only recognize guinea pig β₆.

C. Cloning of Fragments Obtained by PCR

Individual fragments were cloned in pBluescript (Stratagene, San Diego, Calif.) as follows. Purified fragments were resuspended in distilled water containing deoxynucleotides and treated with 2.5 units of DNA polymerase I, large fragment (Promega) to fill in any 3′ recessed ends left after the last cycle of PCR. The 5′ ends were phosphorylated with 5 units of T4 polynucleotide kinase (New England Biolabs, Beverly, Mass.). An aliquot of the above reaction mixture containing approximately 100-200 ng of DNA, was ligated into pBluescript that had been cut with EcoRV (Promega) and dephosphorylated with calf intestinal alkaline phosphatase (Boehringer Mannheim, Indianapolis, Ind.). Ligations were performed at 22° C. for 1 hour with T4 DNA ligase (Bethesda Research Laboratories). The ligation mixture was used to transform competent Escherichia coli (JM109, Clontech, San Francisco, Calif.). Plasmids containing inserts were purified using the Pharmacia miniprep lysis kit (Pharmacia LKB Biotechnology, Inc., Piscataway, N.J.) denatured in 0.3 H NaOH, further purified over spin columns containing Sephacryl S-400 (Pharmacia), and then sequenced using the Sequenase™ version 2.0 sequencing kit (United States Biochemical Corp., Cleveland, Ohio) and [³⁵S]dATP (Amersham Corp., Arlington Heights, Ill.).

D. Library Screening

PCR fragments generated with the primer pairs B 1 F/B2R (SEQ ID NOS:1 and 2) and BTE3F/B4R (SEQ ID NOS:8 and 5) were uniformly labeled with alpha-[³²P]dCTP and used as probes to screen a random-primed cDNA library and an oligo-dT-primed cDNA library. Both libraries were constructed in the plasmid pTZl 8R-BstXI obtained from Invitrogen (San Diego, Calif.) from mRNA obtained from the human pancreatic carcinoma cell line FG-2. Plasmid was purified from clones found to hybridize with either region and inserts were sequenced. A portion of insert DNA from one clone was in turn labeled and used to screen the same libraries. Fourteen independent overlapping clones were sequenced from both ends using primers that recognize regions of the pTZ polylinker. The regions flanking the 3′ end of the putative translated region of the new β subunit were sequenced in both directions from three clones using primers constructed to recognize sequences close to the 3′ end. On the basis of the initial sequences thus obtained, an additional internal sequence was obtained from clones T10, T11, T12 and T14 (FIG. 2) after digestion with specific restriction endonucleases and religation. Three internal fragments thus generated were subcloned into pBluescript and were also sequenced in both directions. Approximately 90% of the new sequence reported was obtained from both strands of DNA, and 97% was obtained from two or more overlapping clones (FIG. 2).

FIG. 2 shows a map of the sequencing strategy. Shown are the location of clones used to obtain the partial cDNA sequence of guinea pig β₆ (clones 1F, 3L, 3N and 3Y, top) and the complete sequence of human β₆ (clones T1-T19 bottom). Also shown is the location of the translated region (Protein). The location of the transmembrane domain is shown by the letters TM. Clones shown often represent one of several identical clones. Internal sequence of clones with long inserts was obtained by restriction endonuclease digestion and relegation and by ligation of internal fragments into pBluescript. Specific restriction sites employed are shown (Hind, HindIII; Hinc, HincII; Kpn, KpnI; Pst, PstI). The direction and extent of sequencing are shown by arrows. 1109 and 1110 are the sites recognized by oligonucleotide sequencing primers. T18 and T19 each terminated in a poly(A) tail. The regions recognized by the degenerate PCR primers B1F (B1), B2R (B2), B3R/F (B3), and B4R (B4) and the β₆ primers BTE2F (BTE2) and BTE3F (BTE3) are noted above the guinea pig cDNA map, kb, kilobases.

E. Nucleotide Sequence of a Novel Guinea Pig Integrin β Subunit

PCR using cDNA from guinea pig airway epithelial cells and the consensus primer mixtures B1F and B2R (FIG. 1) amplified DNA fragments with the expected size of approximately 350 nucleotides. When the fragment DNA was sequenced after cloning into pBluescript, recombinant clones each contained inserts with one of two distinct sequences. One sequence encoded a stretch of 98 amino acids that was 97% identical to the expected region of human β₁ and was therefore presumed to be guinea pig B₁. The other sequence encoded 98 amino acids that were only 53% identical to human β₁, 45% identical to human β₂, and 57% identical to human β₃ (FIG. 2, clone 1F). Both of the guinea pig sequences included the integrin β subunit consensus sequences Ser-X-Ser-Met-X-Asp-Asp-Leu (SEQ ID NO:46) and Gly-Phe-Gly-Ser-Phe-Val (SEQ ID NO:47), and both contained the 2 cysteine residues found in this region in all known integrin β subunits. These data suggest that one of the two sequences we obtained encoded a new member of the integrin β subunit family.

This novel sequence was extended by further PCR steps utilizing primers specific for the novel sequence (BTE2F, BTE3F) (SEQ ID NOS:7 and 8) in combination with two additional degenerate primers (B3R and B4R, see FIGS. 1, 2 and 4). With the primer pair BTE2F/B3R (SEQ ID NOS7 and 4) two different cDNA products were obtained (3L and 3N in FIG. 2) due to an unexpected hybridization of the B3R primer with a site 220 nucleotides further downstream (B3′ in FIG. 2). The 1732-nucleotide sequence determined from these clones is shown in FIG. 3.

FIG. 3 shows the nucleotide sequences and corresponding amino acid sequences for human (H) β₆ (SEQ ID NOS:26 and 27) and guinea pig (GP) β₆ (SEQ ID NOS:28 and 29). The amino acid translation is denoted by the single letter code beneath the second nucleotide of each codon from the translated region of human β₆. For the guinea pig sequence, only amino acids that differ from the human sequence are shown. The numbers along the right-hand margin denote the nucleotide or amino acid number of the last entry on each line. The numbering system used starts with the first nucleotide or amino acid available for each sequence shown. The nine potential sites for N-glycosylation in the putative extracellular domain of human β₆ are underlined.

F. Nucleotide Sequence of Human B₆

Screening of cDNA libraries constructed from the human pancreatic carcinoma cell line FG-2 with guinea pig cDNA probes 1F and 3Y (see FIG. 2) and subsequent screening with a probe constructed from a portion of clone T10 (FIG. 2) produced 14 independent positive clones. The two longest clones (T18 and T19) extended to the poly(A) tail. A map of these clones, constructed on the basis of sequence information and of the mobility of inserts cut out of these clones in agarose gels is shown in FIG. 2. This map-predicts an mRNA of approximately 5 kilobases including at least a 226-nucleotide untranslated region at the 5′ end and, a 2364-nucleotide open reading frame, and a 3′ untranslated region of approximately 2.5 kilobases. This molecule has been termed integrin β₆.

FIG. 3 shows the partial nucleotide and complete amino acid sequences for human β₆ (SEQ ID NOS:26 and 27) (excluding most of the 3′-untranslated region) and the alignment of the 1732 nucleotides of sequence obtained from PCR of guinea pig airway epithelial cell cDNA. Of the 577 amino acids deduced from the region sequenced in both species only 36 residues differ; The amino acid sequences are 94% identical. Furthermore, of the 1732 nucleotides sequenced in both species, 91% are identical. Nine potential glycosylation sites present in the putative extracellular domain of human β₆ are shown by underlining. All seven of these sites that lie within the 577 amino acids obtained for guinea pig β₆ are also present in the guinea pig protein.If all of the potential glycosylation sites are occupied with oligosaccharides having an average molecular weight of 2,500, the predicted molecular weight of human β₆ would be 106,000.

Comparison of the 788-amino acid sequence deduced from the open reading frame to the three previously sequenced human β subunits (SEQ ID NOS:30-32) and the myospheroid protein of Drosophila (SEQ ID NO:33) is shown in FIG. 4.

FIG. 4 shows the alignment of β₆ with four previously reported integrin β subunits. Previously published sequences for human β₁ (SEQ ID NO:30), human β₂ (SEQ ID NO:31), human β₃ (SEQ ID NO:32), the myospheroid gene product (βmyo) of Drosophila (SEQ ID NO:33), and the novel sequence described as β₆ (SEQ ID NO:27) are shown using the single letter amino acid code. The 56 conserved cysteines are noted by * and the 120 other invariant amino acids by=above each line. The transmembrane domain is underlined. The regions used for constructing the consensus β subunit primers B1F (1) (SEQ ID NO:2), B2R (B2) (SEQ ID NO:2), B3F/R (B3) (SEQ ID NOS:3 and 4), and B4R (B4) (SEQ ID NO:5) are labeled below the alignment in bold type. The numbers along the right-hand margin denote the number of the last amino acid in each line beginning from the first amino acid of each putative signal sequence.

There are 179 amino acid residues that are identical, in each of the other β subunits and in β₆ including 56 conserved cysteine residues. The overall percentage of identical amino acids between β₆ and the other human β subunits is 47% for β₃, 42% for β₁ and 38% for β₂. Human β₆ is also 39% identical to the Drosophila β subunit. Human β₁, β₂ and β₃ and the Drosophila β subunit all have cytoplasmic regions consisting of 41 amino acids (beginning after the putative transmembrane domain shown by the underline in FIG. 4). Although β₆ contains each of the 10 conserved amino acid residues in this cytoplasmic region it also contains an 11-amino acid extension at the carboxyl terminus. β₆ also contains two Arg-Gly-Asp sequences, one at amino acids 514-516 and the other at 594-596. These regions could serve as recognition sites for other ligands of the integrin family.

PCR using the primer pair B3F/B4R (SEQ ID NOS:3 and 5) (see FIG. 1) amplified fragments of the expected size of approximately 750 nucleotides. Cloning and sequencing of the fragments did not result in any additional clones containing the novel β subunit sequence but did result in several clones with inserts encoding an amino acid sequence that was 97% identical to the corresponding region of human β₃ and several others encoding an amino acid sequence that was 93% identical to human β₁ (SEQ ID NO:35) (FIG. 5). These are presumably the guinea pig homologues of β₁ and β₃, respectively (SEQ ID NOS:37 and 41). The nucleotide sequences of guinea pig (SEQ ID NO:36) and human β₁ (SEQ ID NO:34) are 80% identical, and those of guinea pig (SEQ ID NO:40) and human β₃ (SEQ ID NO:38) are 91% identical.

FIG. 5 shows the alignment of partial nucleotide and amino acid sequences from human (H) and guinea pig (GP) β₁ (SEQ ID NOS:34-37), β₃ (SEQ ID NOS:38-41), and β₆ (SEQ ID NOS:42-45) for the region just downstream from the B3F primer. Amino acid translations denoted by the one-letter code are shown below the second nucleotide of each codon. For the guinea pig sequences, only amino acids that differ from the human sequences are shown. The numbers shown along the right-hand margin denote the nucleotide number for human β₆. The sequences for human β₁ and β₃ are from previously published reports.

EXAMPLE II

β₆ Associates with α_v And α_F Subunits

To determine that the novel β subunit of the present invention is associated with an α chain similar to other known integrins, antisera against peptides from the cytoplasmic domain sequence of β₆ were prepared. The following amino acid peptides from the cytoplasmic sequence of β₆ were prepared and used to immunize rabbits: RGSTSTFKNVTYKHR (SEQ ID NO:48) (residues 763-777) and YKHREKQKVDLSTDC (SEQ ID NO:49) (residues 774-788). The antisera were raised in rabbits according to standard procedures known in the art. Briefly, peptides were chemically coupled to keyhole limpet hemocyanin, and were injected in rabbits in either complete (first injection only) or incomplete Freund's adjuvant as described, for example, in Antibodies: A Laboratory Manual, E. Harlow and D. Lowe, eds., Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 11724. Antisera were termed 6830 (to peptides corresponding to residues 763-777) and 6341 (to peptides corresponding to residues 774-788).

The resulting polyclonal antibodies were used to immunoprecipitate detergent lysates from the pancreatic carcinoma cell line FG-2 that had been surface radioiodinated according to procedures well known in the art such as described, for example, in Kajiji et al., EMBO J. 3:673-680 (1989). A complex of two bands was precipitated of respectively 150 kilodaltons (Kd) and 97 Kd in SDS-PAGE under non-reducing conditions. Under reducing conditions, the two bands migrated as a diffused band, extending from 130 Kd to 116 Kd. These bands were specific since pre-immune serum did not precipitate any of them and they were not present when the immunoprecipitation was carried out in the presence of the corresponding immunogenic peptide. Furthermore, the same complex of two bands was precipitated by both the 6830 and 6841 antibodies, which were raised against independent peptides from the cytoplasmic sequence deduced from β₆ cDNA clones.

To determine which of the two precipitated bands corresponds to β₆, a SDS-heat denaturated lysate from surface-radioiodinated FG-2 cells was immunoprecipitated with the 6841 antibody. Only the 97 Kd band was detectable (non-reducing conditions), identifying it as the β₆ band. Under reducing conditions, the apparent molecular weight of this band increased to 116 Kd suggesting the presence of many intra-chain disulfide bonds, which is consistent with the primary structure of β₆ and of other integrin B chains.

The other band, of 150 Kd or 130 Kd under non-reducing or reducing conditions, respectively, is likely to be an α subunit since it dissociates after SDS-heat denaturation of the lysate, indicating that it is non-covalently associated with the B₆ polypeptide. Furthermore, similar to certain other integrin α chains, its molecular weight decreases under reducing conditions by about 20 Kd (130 Kd versus 150 Kd under non-reducing conditions) probably due to a disulfide linked small peptide that dissociates upon reduction.

To identify which α chain is associated with β₆, the αβ₆ integrin complex was purified by immuno-affinity chromatography on a 6841-protein A sepharose matrix according to procedures well known in the art such as described, for example, in Kajiji et al., EMBO J. 3:673-680 (1989). The eluted material was immunoprecipitated with antibodies specific for α₁, α₂, α₃, α₅, α₆ and α_v, which are known to be expressed in FG-2 cells. Only the anti-α_v monoclonal antibody 142.19, obtained from Dr. David Cheresh, The Scripps Research Institution, La Jolla, Calif., reacted with the purified material, which indicates that the α_v is associated with β₆ in this pancreatic carcinoma cell line.

To confirm this data, immunodepletion experiments on surface-radioiodinated FG-2 lysates were performed according to methods well known in the art such as described in Kajiji et al., EMBO J. 3:673-680 (1989). The cell lysate was depleted with the 6841 anti-β₆ antibody or, in parallel, with a control antiserum, and then immunoprecipitated with the 142.19 anti-α_v antibody. A smaller amount of α_v was present in the immunoprecipitation on the β₆ depleted lysate and no 97 Kd β₆ band was visible. Instead, a smaller band of about 90 Kd was present. It is hypothesized that this smaller band represents the β₅ chain also associated with α_v in these cells. In the control lysate depleted with normal rabbit serum, all three bands, 150 Kd (α_v), 97 Kd (β₆) and 90 Kd (β₅) were present after immunoprecipitation with the anti-α_v 142.19 antibody.

Another immunodepletion was carried out using 142.19 antibody as the depleting antibody, or in parallel a mouse monoclonal as a control antibody. Immunoprecipitations of α_v-depleted lysate with anti-α_v 142.19 antibodies did not. show the presence of any band, indicating that all α_v-containing integrins had been removed. However, the 6841 anti-β₆ antibody still precipitated a complex of two bands, one corresponding to β₆, the other with a molecular weight close to that of α_v. This α chain, however, must differ from α_v since it is unreactive with anti-α_v monoclonal antibodies and is referred to herein as α_F. In the control depleted lysates, the 6841 anti-β₆ antibody precipitates much stronger bands, consistent with the possibility that, in FG-2 cells, two β₆ integrins exist, α_vβ₆ and α_Fβ₆.

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 claims.

Claims

1. A method of blocking binding of αvβ6 to a ligand, the method comprising, contacting a cell in vitro with an antibody that binds a β6 subunit of αvβ6, wherein the β6 subunit comprises SEQ ID NO:27, and wherein the cell is selected from the group consisting of a pancreatic cancer cell, a colon cancer cell, a lung cancer cell, and a chorio cancer cell, and wherein the ligand is fibronectin.

2. The method of claim 1, wherein the cell is a pancreatic cell.

3. The method of claim 1, wherein the cell is a colon cancer cell.

4. The method of claim 1, wherein the cell is a lung cancer cell.

5. The method of claim 1, wherein the cell is a chorio cancer cell.