WO1990007937A1 - Pregnancy specific proteins applications - Google Patents

Abstract

At least seven genes have been identified which encode proteins generally identified as pregnancy specific protein (SP1), also known as pregnancy specific beta glycoprotein (PSBG). These genes have been found using human placental SP1 labelled cDNA in a number of mammalian species, including human, baboon, cow, sheep, goat, pig and rat, and non-mammalian species, including fish and bird. The genes are classified on the basis of their nucleotide sequence, structure, glycosylation sites, restriction enzyme mapping, and tissue of origin. Genes have been found in cells of human placenta origin (three groups); in intestinal cells; in cells of both testis and placental origin; in tissue of testis origin; and in HeLa cells. One clone encodes a SP1-like protein which is specifically found in placenta and appears to have a hydrophobic C-terminal region, indicating that it is membrane bound. These cDNAs are at least 65% homologous with some members of the immunoglobulin gene superfamily such as CEA. Methods for making and using the DNA sequences, proteins, and antibodies to the proteins are also described, particularly for diagnostic work. Immunosuppressive activity and growth promoting activities have also been demonstrated by suppression of lymphocyte proliferation and inducement of megakaryocytopoiesis, respectively.

Description

PREGNANCY SPECIFIC PROTEINS APPLICATIONS

Background of the Invention

The present invention generally relates to proteins, particularly a group of pregnancy-specific β-1 glycoprotein-like proteins, the genes encoding them, and methods for their use.

This is a continuation-in-part of U.S. Serial No. 07/298,638 entitled "Pregnancy Specific Proteins" filed January 18, 1989 by ai-Yee Chan. Pregnancy-specific β-1 glycoprotein (SP1) is found in the serum of pregnant women and has been isolated in pure form, as reported by Tatarinov and Masyukevich, Bull. Exp. Biol. Med. 69, 66-68 (1970). It is synthesized by the syncytiotrophoblast of the placenta and secreted into maternal serum, as studied by Bohn, Placental Proteins, pp. 71-88, A. Klopper and T. Chard, editors (Springer Verlag, NY 1979) , becoming detectable during the first two to three weeks of pregnancy and increasing as pregnancy progresses to levels of about 200 to 400 μg/ml, as reported by Tatarinov and Masyukevich Bull. Exp. Biol. Med. 69, 66-68 (1970) and Lin, et al. , J. Clin. Invest. 54, 576-582 (1974)_*

Even though the placental origin of human SP1 is well established, low levels of this protein can be detected in serum of the normal male and female by radioi munoassay. The extraplacental source of human SP1 has not been identified. Ectopic production of SPl by cultured human skin fibroblasts and normal brain cells has been reported by Heikinheimo, et al., J.Clin.Endocr.Metab. 51, 1432-1436 (1980) ; Rosen, Pregnancy Proteins, pp. 223-234 (1982) ; and Chou, Oncodevelop. Biol. Med. 4, 319-326 (1983), but the cause is not known. As first reported by Bohn, Blut 24, 292-302

(1972) , SPl isolated from placenta consists of a single polypeptide chain with an N-terminal histidine, having a molecular weight of 90,000, of which 30% is carbohydrate. Several subsequent studies have indicated that this "protein" is actually a heterogeneous mixture of several components. There is a wide discrepancy of reported molecular* weights for human placental SPl. Values determined by gel filtration, ultracentrifugation and SDS-PAGE vary from 110,000 to 42,300, respectively. Carbohydrate varies from 28 to 32%. As reviewed by Bischof, Contri. Gvnecol. Obstet. 12, 6-92 (1984), eleσtrophoretic mobility ranges from α, β, to gamma. Genetic variants of SPl have also been reported in both normal and abnormal pregnancies as well as in tumors. It has been suggested that these variants involve a change in protein size or sequence. cDNA sequences encoding placenta-specific SPl, having slight differences in sequence at the 3* end, were reported by atanabe and Chou in J.Biol. Chem. 263, 2049-2056 in 1988, and Streydio, et al. in Biochem. Biσphys. Res. Comm. 154(1), 130-137, (1S88) . Monoclonal antibodies to human SPl have also been reported to have different and unique specificities and affinities, although it is not known if the antigenic determinants are on the same or different proteins.

SPl determinations have great potential clinical application, even though the function of SPl is still not known. The medical relevance of SPl in a number of situations has been extensively investigated. The most important use of SPl is for monitoring various conditions, both normal and pathological, during pregnancy, reviewed by Bischof (1984) . SPl measurements have also shown promise for the diagnosis and monitoring of trophoblastic and so e nontrophoblastic tumors, as described by Sorensen, Clin. Chim. Acta 121, 199-208 (1984) .

Despite the potential, SPl has not been well utilized in clinical medicine because of the lack of reliable quantitation methods. At present, the majority of the SPl assays depend on antigen-antibody interaction, dependent on the nature of the SPl molecule and on the specificity of the antibody involved. Heterogeneity of the SPl molecules decreases the reliability of these immunoassays. It is therefore an object of the present invention to provide the nucleotide sequences and structure of a group of several SPl proteins. It is a further object of the present invention to provide a method and means for producing extremely pure SPl proteins, for pharmaceutical use.

It is another object of the present invention to provide a methods and means for producing monoclonal antibodies for quantitating the different species of SPl and for improving assays for CEA by eliminating cross-reactions with SPl-like proteins in non-placental tissues.

It is a further object of the present invention to provide methods and means for use in pregnancy assays and in detection and monitoring of trophoblastic and some non-trophoblastic tumors

It is still another object of the present invention to provide reagents for use in preventing fetal rejection, in inducing rejection of the embryo in therapeutical procedures, and in improving the transfer of immunoglobulins into the fetus. Summary of the Invention

At least seven genes have been identified which encode proteins generally identified as pregnancy specific protein (SPl), also known as pregnancy specific beta glycoprotein (PSBG) . These genes have been found using human placental SPl labelled cDNA in a number of mammalian species, including human, baboon, cow, sheep, goat, pig and rat, and non-mammalian species, including fish and bird. These genes can be classified as those specifically found in human placenta (which can be grouped by the presence or absence of specific restriction enzyme sites and hydrophobic regions into at least three groups) ; in intestinal cells; in cells of both testis and placental origin; in tissue of testis origin; and in HeLa cells.

An exemplary clone of the first group of placental specific SPl-like proteins is hPS12. A clone encoding a SPl-like protein which is specifically found in placenta and which appears to have a hydrophobic C-terminal region, indicating that it is membrane bound, is hPS2. Another clone having a sequence found in placenta is hPSll, which is very closely related to clone PSG16 of atanabe and Chou, J. Biol. Chem. 263 (4) , 2049-2056 (1988) , and clones PSBGC and PSBGD of Streydio, et al. , Biocheπ Biophys. Res. Comm. 154(1), 130-137 (1988). Clones isolated from an intestinal library include hISl, hIS2 and hIS3. A clone common to both placenta and testis is hPS3. Clones isolated from a testis cDNA library include hTSl, hTS2 and hTS3. Clones isolated from a HeLa cell library include hHSl, hHS2, hHS8, hHSll, hHS4, hHS3, hHS6, hHS9, hHS12, and hHS14. These cDNAs are at least 65% homologous with some members of the immunoglobulin gene superfamily such as Carcinoembryonic Antigen (CEA) . The proteins encoded by the cDNA of intestinal cell origin appear to be more closely related to the CEA proteins than to the other SPl proteins.

Methods for making and using the DNA sequences, proteins, and antibodies to the proteins are also described, particularly for diagnostic work, as in making and purifying reagents for use in pregnancy assays, in detection and monitoring of trophoblastic and some non-trophoblastic tumors and in purification of reagents used to assay for SPl and for CEA.

Immunosuppressive activity and growth promoting activities have been demonstrated by suppression of lymphocyte proliferation and inducement of megakaryocytopoiesis, respectively.

Brief Description of the Drawings

Figure 1 summarizes the relationship between the SPl genes, CEA and the immunoglobulin gene superfamily.

Figure 2 is the nucleotide and predicted amino acid sequence of hPSll. The amino acid numbers are above and nucleotide numbers below the sequence. Potential glycosylation sites are underlined.

Deduced amino acid sequences identical to that of tryptic fragments of the pure protein are boxed.

(V) Amino acid residue different from those determined by a ino-terminal analysis of pure SPl; (▼) cysteine residues; (***) stop codon. The ends of the different subdomains are indicated by Rln, R2n, and R2c.

Figure 3 is the nucleotide sequence and predicted amino acid sequence of hPS2. Nucleotide numbers are indicated. Potential glycosylation sites are underlined. Open inverted triangles indicated conserved Cys residues. Solid squares indicated polyadenylation sites. * indicates stop codon. Boundaries of the different domains are indicated by: N, N-terminal domain; Rn, n-subdomain of repeat unite; Re, c-subdomain of repeat unit; C, C-terminal domain.

Figure 4 is the nucleotide sequence and predicted amino acids of hPS12 and the 5* extended sequence derived from hPS90. The nucleotide numbers are indicated. Potential glycosylation sites are underlined. Open inverted triangles indicate conserved Cys residues. Solid squares indicate polyadenylation site. * indicates stop codon.

Boundaries of the different domains are indicated by: N, N-terminal domain; Rn, n-subdomain of repeat unite; Re, c-subdomain of repeat unit; C, C-terminal domain. Figure 5 is the nucleotide sequence and the derived amino acid sequence of hPS3. The nucleotide numbers are shown beneath the sequence. The potential glycosylation sites are underlined. The amino acid sequence determined by protein sequencing is boxed (Peptide C) .

Figure 6 is a comparison of the nucleotide sequence of 3' EcoRI fragment of hTSl (hTSIL) with hPSll.

Figure 7 is the nucleotide sequence of hTS16. Number is the nucleotide base number. Amino acids encoded by exons are shown. Polyadenylation signal sequence is scored by a broken underline. Potential glycosylation sites are underlined and conserved cysteine residues are boxed. Large empty triangle indicates deletion. Boundaries of exon/intron and domain regions are shown. IVS: intervening sequence; N: N-terminal region; Rn: n-subdomain; Re: c-subdomain; C: C-terminal domain.

Figure 8 is the nucleotide sequence and derived amino acid sequence of hHS2. The nucleotide numbers are indicated. Potential glycosylation sites are underlined. Cysteine residues are boxed. The boundaries of the different domains are marked: N- Term (N-terminal domain) , Rln (n-subdomain of repeat 1) , R2n (n-subdomain of repeat 2) .

Figure 9 is the nucleotide sequence and derived amino acid sequence of hISl. Nucleotide numbers are indicated. Potential glycosylation sites are underlined. Cysteine residues are boxed. The boundaries of different domains are marked: N-Term (N-terminal domain) , Rln (n-subdomain of repeat 1) , Rlc (c-subdomain of repeat 1) . (▼) indicates potential signal peptidase cleavage point.

Figure 10 is a comparison of the domain structure of hPSll, hISl and CEA. The following legend is used:

O

I potential glycosylation site

S—S S-S bond

///// signal peptide N-term

n-subdomain

c-subdomain

hydrophobic C-term Hydrophilic C-term Figure 11 is a comparison of the aligned consensus amino acid sequences encoded by human placental SPl cDNAs with that encoded by human CEA and NCA cDNAs^". Potential glycosylation sites are underlined. Solid inverted triangles indicate conserved Cys residues. Amino acids are numbered^' with reference to the beginning of each domain or subdomain. SPl NCon, Consensus sequence of N- terminal domain of SPl cDNAs; SPlRlnCon, consensus sequence of Rln-subdomain of SPl cDNAs; SPlR2cCon, consensus sequence of R2c-subdomain of SPl cDNAs. Domain notation of CEA and NCA is the same as in published references.

Figure 12 is a graph of the immunosuppressive activity of human placental SPl, plotting percent inhibition of mixed lymphocyte reaction versus SPl (micrograms per ml) .

Figure 13. is a graph of the growth promoting activity of human placental SPl, plotting % urine egakaryocyte acetylcholinesterase activity of the control versus SPl (micrograms per ml) .

Detailed Description of the Invention

Several new activities, including growth factor activity and immunosuppressive activity, for SPl proteins have now been determined, based on the structure and sequence of the cDNA clones described in U.S. Serial No. 07/298,638, and confirmed in vitro. Conservation of the genes encoding these proteins has also been demonstrated in a number of mammalian and non-mammalian species, indicating the basic importance of the proteins, and establishing the utility of several animal species as models for further studies. U.S. Serial No. 07,298,638 describes the isolation and characterization of cDNA clones encoding several distinct, but closely related, proteins. At least seven genes are identified which encode proteins specifically found in placenta (which can be grouped by the presence or absence of specific restriction enzyme sites and hydrophobic regions into at least three groups) , genes found in intestinal cells, genes found in cells of both testis and placental origin, genes found only in tissue of testis origin, and genes found in HeLa cells. An exemplary clone of the first group of placental specific SPl-like proteins is hPS12. A clone encoding a SPl-like protein which is specifically expressed in placenta and which appears to have a hydrophobic C-terminal region, indicating that it is membrane bound, is hPS2. Another clone having a sequence expressed in placenta is hPSll, which is very closely related to clone PSG16 of atanabe and Chou, J. Biol. Chem. 263 (4), 2049-2056 (1988), and clones PSBGC and PSBGD of Streydio, et al., Biochem. Biophys. Res. Comm. 154(1), 130-137 (1988). Clones isolated from an intestinal library include hISl, hIS2 and hIS3. A clone common to both placenta and testis is hPS3. Clones isolated from a testis cDNA library include hTSl, hTS2 and hTS3. Clones isolated from a HeLa cell library include hHSl, hHS2, hHS8, hHSll, hHS4, hHS3, hHS6, hHS9, hHS12, and hHS14. These cDNAs are at least 65% homologous with some members of the immunoglobulin gene superfamily such as Carcinoembryonic Antigen (CEA) . Important features of these proteins are the similarities in the occurrence of β-sheet and repeating domain structure, and conserved glycosylation sites and cysteine residues within the repeating domains. indicating that they all evolved from the same primordial gene by gene duplication or exon shuffling. The CEA family differs from the SPl family by having a high degree of glycosylation. The proteins encoded by the cDNA of intestinal cell origin appear to be more closely related to the CEA proteins than to the other SPl proteins.

Several uses of the DNA sequences, proteins, and antibodies to the proteins are also described in 07/298,638, particularly for diagnostic work, as in purifying reagents for use in detecting CEA.

Several cDNA clones for SPl proteins have been isolated and characterized. Some are analogous to previously reported cDNAs for SPl proteins. The SPl genes of the present invention include the cDNA sequences described in the figures and examples, . homologous sequences thereof isolated from any naturally-occurring genome, and any analog thereof in which nucleotides in the sequence are substituted, deleted or added while encoding a protein having at least a portion of the specific biological activity or unique structure of the encoded peptide. The present invention also includes a substantially pure peptide or protein as described in the figures and examples, or as expressed from the described cDNAs, analogues thereof in which amino acids in the sequence are substituted, deleted or added while maintaining at least a portion of the specific biological activity or structure of the peptide, and conjugates of any such peptide or analog.

For purposes of clarification, the following nomenclature is used herein to describe the cDNA clones: cloning vector: lambda, lambda phage; p, plasmid; m, M13 phage lower case letter indicating species of origin of cDNa: h, human; r, rat; b, bovine

Capital letter indicating tissue of origin of cDNa: P, placenta; I, intestine; H, HeLa cells; T, testis

Capital letter indicating protein family, here, S indicates the SPl protein family clone number csubclone number In some cases, clones have previously been referred to in the literature by a different reference number. For example, hPSll has been referred to as hPSPll. hHS2 has been referred to as hHSP2. hPSll is analogous to the clone described by atanabe and Chou, J.Biol.Chem. 263(4), 2049 (1988), PSG16, and that by Streydio, et al., Biochem. Biophvs. Res. Comm. 154(1), 130-137 (1988), PSBGC and D. hPSll is an example of one of the clones of the present invention encoding a placental SPl-like protein that is detected in placenta and in testis. The amino-terminal 143 amino acids of hPSll, besides having a characteristic secondary structure, appears to constitute a specific domain of the protein containing three of the seven glycosylation sites. The rest of the protein is composed of repeating units. There are two internal repeats, Rln and R2n, each of 279 bp, encoding 93 amino acids, that are identical in 73% of their nucleotides and 48% of their amino acids. The glycosylation sites of the repeats are not conserved. The second repeat is more hydrophilic than the first repeat. After the internal repeats there are 90 amino acids before the stop codon TGA, designated R2c. This region contains no glycosylation sites. It is believed that the SPl family was formed by duplication of a primordial gene, which developed into the SPl genes, the Carcinoembryonic Antigen genes, and the immunoglobulin genes. The relationship is depicted in Figure 1. One characteristic of the members of the SPl gene family is the deletion of the c-subdomain of the first repeating unit (Rlc) in several of the cDNAs cloned, indicating that hPSll, hPS12, pSG16 and hHS2 all originate from one ancestor which contains two of the repeating units with one c-subdomain deleted. hISl, even though discovered by hybridization with an SPl probe, is closer to the CEA family than the SPl family, and may therefore be intermediate between the two families of genes.

Besides having immunosuppressive function, CEA is believed to be involved in cell-cell interaction and growth factor like activities. SPl proteins, especially those in the group isolated from an intestine gene library, are expected to have analogous functions.

SPl genes and immunoglobulin genes have similar overall domain structure, conserved disulfide bridges and β-sheet structure and approximately 65% homology in some part of their nucleotide sequence.

This relationship has a very important implication as to the function of SPl proteins. Many immunoglobulin gene family members have receptor or cell recognition functions, suggesting that SPl might also have similar physiological roles. The similarity in structure between the MHC antigen and the membrane bound form of SPl specific to placenta (hPS2) suggests that SPl might compete with the MHC antigens in presentation of the fetal antigens to killer T cells. If so, SPl may prevent fetal allograft rejection by blocking the action of MHC antigens and killer T cells and provide local immunity to the implanted embryo. The similarity in structure between the poly Ig receptor and the cytoplasmic form of SPl (hPSll and hPS12) suggests that SPl might have functions similar to that of poly Ig receptor, i.e., transcellular transfer of immunoglobul-ins from the mother to the fetus. cDNA clones encoding genes for each of the several groups of SPl-like proteins have been isolated and characterized, as described in the following non-limiting examples. It is understood that specific cDNA sequences can be modified by those skilled in the art, for example, by labelling, fusion with regulatory sequences, insertion into expression vectors, and substitution or deletion of nucleotides encoding specific amino acids, without departing from the scope of the nucleotide and amino acid sequences of the present invention, and the methods for their use.

The methods and compositions of the present invention are further illustrated by the following non-limiting examples, using the methods and reagents described below. Vector libraries used in the examples and which are suitable for use in the present invention are sold by Clontech Laboratories, Palo Alto, California, or may be prepared in accordance with known procedures, such as those described in "Construction and Screening cDNA Libraries in lambda gtlO and lambda gtll", A Practical Approach, DNA Cloning (IRL Press, Oxford, England, 1985) Vol. l, pp. 49-79. A lambda gtll human placental expression library obtained from M. D. Anderson Hospital and Tumor Institute was used in examples of the present invention. This library includes double-strand cDNA with over 500 base pairs (bp) , cloned into the EcoRI site of the lambda gtll phage.

Genes isolated from the library can be expressed by plating 2 x 10⁶ phages from a human placental lambda gtll library on E. coli Y1090 at a density of 5 x 10^s plaques per 150 mm L-agar plates and inducing expression by adding IPTG. E- coli Y1090 contains t lac repressor which prevents lacZ- directed gene expression until it is derepressed by the addition of IPTG (Isopropyl-β-D- thiogalactopyranoside) to the medium; a deficiency in the Ion protease which increases the stability of the recombinant fusion protein; and supF to suppress the phage mutation causing defective lysis. To ensure -that fusion proteins toxic_v to the host produced by particular recombinants will not inhibit the growth of particular members of the library, plaque forma¬ tion is initiated without expression from the lacZ gene promoter. After the number of infected cells surrounding the plaques is pin-size, lacZ-directed gene expression is switched on by the addition of IPTG. The production of the protein from the plaque is induced by overnight incubation at 42°C in the presence of IPTG.

It is preferable to plate out only a sufficient number of the sequence-carrying vectors to proportionally represent groups of sequences found in the library. This can be accomplished by following the procedure in Genetic Engineering, Vol.l (Academic Press, New York 1981) .

In order to detect the desired protein, a filter is contacted with the expressed proteins in an agar plate, so that the proteins adhere to the filter. The filter is prehybridized so that the antibodies will not bind non-specifically to the filter. The filter is then incubated in a labeled antibody solution and an autoradiogram made. The DNA encoding the proteins complexing with the antibody are identified by superimposing the marked autoradiogram over the agar plate. Antibody against SPl protein is commercially available from Calbiochem of San Diego, California. This antibody is raised in rabbit, adsorbed with other placental proteins and affinity purified. The antibody (100-150 μg) is labeled with 1 mCi of ^{, 5}I in the presence of iodogen (1,3,4,6-tetrachloro-3o_,6α:-diphenylglycouril) from Sigma Chem. Co. of St. Louis, Missouri, and 1 M Tris- HCl, pH 8.0. The SP, protein antibody can bind to proteins which are selected members or fragments of members of SP, proteins. The antibody-protein complex is identified by autoradiography which detects the radioactivity of the labeled antibody. The autoradiogram then is disposed over the agar dish in the same position as that of the filter which was originally used to adsorb the proteins. In this manner, the phage containing the insert producing the marked protein can be identified. Once the appropriate phages are identified, the E. coli colonies containing those phages are cut out of the agar, and placed in a microfuge tube with 0.5 ml of L-Broth. The suspension is vortexed to break up the agar and release the phage from the ______ coli colony. The supernatant portion of the L-Broth containing the phages is replated on E. coli Y1090 at a density of 5 x 10⁵ plaques per 150 mm L-agar plates. The previous procedure of dispersing the phages, treating the proteins expressed by the DNA contained within the phages with an antibody and identifying the phages carrying the DNA fragments of interest, as previously described, is repeated until all phages in the agar plate contain DNA that express proteins which react with the antibody. A 10 mM solution of Tris-Hcl, pH 7.5, containing 20 mM magnesium chloride is added to the agar plate and the phages allowed to ^' diffuse from the agar into the solution for two hours. The solution containing the phages is then withdrawn from the agar plate. The cDNA screening process is carried out by exposing the autoradiogram to radiolabeled DNA probe, under conditions which promote DNA hybridization. The DNA probe is characterized by a nucleotide sequence that is complementary to at least a portion of the nucleotide sequence of a DNA fragment coding for the selected member of the SPl family of related proteins. Enzymes and Chemicals:

E. coli DNA polymerase 1 and T4 polynucleotide kinase were purchased from Boehringer Mannheim Biochemicals, Indianapolis, IN. DNAse 1 was purchased from Pharmacia, Piscataway, NJ. All restriction enzymes were supplied by either Bethesda Research Laboratories, Gaithersburg, MD, or New England Biolab, Beverly, MA. LambdaSorb phage adsorbent was purchased from Promega Biotec, Madison, WI. The Phage Lambda Mapping Quik-Kit was supplied by Collaborative Research, Inc., Bedford, MA. Nylon filters for library screening and radiochemicals including a and gamma ³²P deoxynucleotides were purchased form Amersham, Arlington Heights, IL. Nitrocellulose filters were supplied by Schleicher & Schuell, Keene, NH. Highly purified human placental genomic DNA was purchased from Sigma, St. Louis, MO. All other chemicals were reagent or molecular biology grade.

Example 1: Isolation and Characterization of cDNA

Encoding a Cytoplasmic Form of Placental SPl. cDNA clones encoding SPl proteins were isolated from a group of fifteen positive clones obtained by screening a human placental cDNA library, initially with SPl protein antibody and then with a partial SPl cDNA probe, described by Chan, et al., in Human Reproduction 3(5), 677-685 (1988). The cDNA insert was released from a lambda gtll vector by complete and partial digestion with EcoRI and subcloned into M13mpl8 and M13mpl9. The DNA sequence was determined by a modified dideoxy chain termination method using Klenow fragment at 50°C or Sequenase from USB, Cleveland, Ohio, at 37°C, as described in L. Johnston-Dow et al., BioTechniques 5,754-765 (1987). M13 universal sequencing primer (Pharmacia, Piscataway, New Jersey) as well as synthetic oligonucleotide primers were used to prime the sequencing reaction. All sequences were determined three or more times as well as from different M13 subclones. Subclones of opposite orientation, identified by the C-test as described by J. Messing, Methods of Enzymology 10lC_r pp. 20-77, R. Wu et al., eds. (Academic Press, New York 1983), were sequenced so that the final sequence was determined from both strands. DNA sequences were analyzed by the Sequence Analysis Software Package of the Genetics Computer Group at the University of Wisconsin as described in H. Devereux et al., Nucl. Acid Res. 12, 387-395 (1984).

Each clone was Southern blot analyzed and sized and placed into one of three groups. The DNA was digested to completion with EcoRI. The digested DNA fragments were separated by electrophoresis on 1% agarose in 1 x TA buffer (40 mM Tris base, 20 mM sodium acetate, 18 mM sodium chloride, 1 mM disodium EDTA, pH 8„0), and transferred to a nitrocellulose filter as described by Smith, in Anal. Biochem. 109, 123-129 (1980) . Blots were analyzed with cDNA inserts previously labeled with ³²P by nick translation as described by Rixon, et al., in Biochemistry 21. 3237-3244 (1983). Hybridization and washing conditions were as described by Kan, Proc. Natl. Acad. Sci. USA 75, 5631-5635 (1978) . Hybridization and washing use high stringency conditions, with 2 x SSC at 65^βC for 15 minutes and^' repeated, 2 X SSC containing 0.1% sodium dodecyl- sulfate at 65°C for 30 minutes and 0.1 x SSC at 65°C for 10 minutes, as recommended by the manufacturer of the filters. Under these conditions only the highly homologous cDNAs are identified. When nylon filters were used instead of nitrocellulose, the procedures were modified as recommended by the manufacturer (Amersham, 1985) . Nylon blots were washed twice with 2 x SSC (0.3 M sodium citrate, 0.3 M sodium chloride) at 65°C for 15 minutes, twice with 0.1 X SSC-O.1% SDS at 65°C for 30 minutes and once with 0.1 x SSC at 65°C for 30 minutes.

For genomic DNA analysis, human placental genomic DNA predigested with EcoRI. BamHI and BamHI/Hindlll was purchased from Oncor, Inc., Gaithersburg, MD. Twelve micrograms of each type of DNA were separated by electrophoresis through 0.8% agarose and Southern blotted.

Clones were divided into three groups. Group 1 clones had cDNA inserts with no internal EcoRI site, and a length of no longer than 1720 bp (base pairs) hPS12 is a representative clone. Group 2 clones had cDNA inserts with one internal EcoRI site, and a length of no longer than 1958 bp with EcoRI fragments of 622 and 1336 bp. hPSll is a representative clone. Group 3 clones had cDNA inserts with two internal EcoRI sites, and a length of no longer than 2215 bp with EcoRI fragments of 450, 645 and 1120 bp. hPS2 is a representative clone. Clones not having internal EcoRI sites because they were less than full length were assigned to the appropriate group based on their restriction enzyme maps or nucleotide sequences. The hybridization of cDNA inserts of all three groups with the probe under highly stringent condition indicated that they were highly homologous to the SPl probe.

The nucleotide sequences of the cDNA insert of the six clones in group 2 were determined by Sanger's dideoxy chain termination method, as described in the Amersham (1985) cloning and sequencing manual. The DNA sequences were analyzed by the computer programs described in Nucleic Acids Res. 12, 605-614 (1984). The nucleotide sequence for hPSll, which is expressed in placental tissue, is set out on the lower series of lines in Figure 2. The peptide sequence of hPSll is shown in the top series of lines in Figure 2, with each amino acid residue positioned directly above the sequence of DNA bases that codes for it. Amino acid sequences identical to those of two tryptic fragments of pure SPl protein, prepared according to Chan, et al., in Human Reproduction 3(5), 677-685 (1988), are present in the coding sequence of hPSll, establishing the identity of this cDNA with purified SPl protein. Purified human SPl protein (1 mg/ml) was incubated with trypsin (L-l- tosylamido-2-phenylethylchloromethyl ketone) TPCK- treated, final concentration of 10 μg/ml in 50 mM ammonium bicarbonate, pH 8, at 37°C for 16 hours. The tryptic fragments were then separated by HPLC using a reverse-phase C18 column, equilibrated with 0o05% trifluoroacetic acid (TFA) . Peptides were eluted from the column by an acetonitrile gradient of 0-40% in 80 minutes at a flow rate of 1 ml/min. Fractions of 1 ml were collected. Peptide peaks were randomly selected and concentrated for further purification on the same column. The procedure for rechromatography was as follows: 0% for 10 minutes, 1 ml/min; 0-20% for 20 minutes, 1 ml/min; 20% constant for 60 minutes, 0.5 ml/min. The amino acid sequences of randomly selected tryptic peptides "B" and "C" were determined by the method described by Edman and Begg, in Eur.- J. Biochem. 1, 80-91 (1967) in a gas phase sequenator as described in Proc. Natl. Acad. Sci. USA 82, 3616-3620 (1985) and using polybrene as a non-protein carrier. About 0.2-0.5 nmol of peptide were applied to the gas phase sequenator. Phenylthiohydantoin (PTH) derivatives from the sequenator were identified by HPLC using a Waters Nova Pak C18 column as described in J. Biol. Chem. 261, 14335-14341 (1986) . Norleucine was added to each sample as an internal standard. All PTH derivatives were monitored at 265 nm and 313 nm for serine and threonine, respectively. These sequences are boxed in Figure 2.

The hPSll cDNA has 1958 bp with a 5¹ non- coding sequence of 73 bp and an open reading frame encoding a ^"protein of 419 amino acids with a calculated molecular mass of 47.2 kDa. Even though no upstream stop codon can be identified, the sequence around the presumed translation initiation codon ACCATGG agrees with the consensus sequence for initiation of translation in vertebrates as suggested by M. Kozack, Nuel. Acid Res. 15: 8125-8148 (1987). Amino-terminal analysis of pure SPl protein demonstrates a sequence of N-X-Thr-Ile-Glu-Ala-Gln- Pro-Pro-Lys-Val-Ser-Glu, corresponding to the predicted amino acids 37-47 of hPSll except Glu-41 and Thr-43. These differences could be due to the majority of the SPl proteins having blocked N- terminals, so that this procedure only gave the N- terminal sequence of a minor component of SPl preparation with unblocked N-terminal. Analysis with the PEPPLOT computer program of the Sequence Analysis Software Package indicated that the N-terminal 34-35 amino acids demonstrate the characteristics of a putative signal peptide having a hydrophobic core, a helix breaker (Pro) and a small uncharged amino acid at the site of cleavage (Ala) , as suggested by G. Blobel et al., J. Cell. Biol. 67, 835-851 (1975) and T. Yama oto et al.. Cell 39, 27-38 (1984).

There are seven possible N-glycosylation sites of the structure Asn-X-Thr/Ser in the predicted amino acid sequence, underlined in Figure 2. All potential glycosylation sites are present within the N-terminal three-fourths portion of the protein. Analysis of the predicted secondary structure with the computer algorithm ,of B. A. Jameson et al., CABIOS 4: 181-186 (1988) shows that the protein is in the form of beta-sheets except for two small areas near the N-terminal (amino acids 35-50 and 160-170) , which are in the form of alpha-helices. With the PEPPLOT computer program, it is possible to determine that the protein is largely hydrophilic, except for the amino-terminal 10 amino acids. The amino-terminal 143 amino acids, besides having a characteristic secondary structure, appears to constitute a specific domain of the protein containing three of the seven glycosylation sites. Analysis with the REPEAT program indicates that the rest of the protein is composed of repeating units. There are two internal repeats, each of 279 bp, encoding 93 amino acids. Rln starts at nucleotide 503 (Leu-144) and runs through 781 (Leu-236) . R2n starts at nucleotide 782 (Pro-237) and runs through 106.0 (Leu-329) . These two repeats are identical in 73% of their nucleotides and 48% of their amino acids. There is only one potential glycosylation site in the first repeat in contrast to the second repeat which has three glycosylation sites. The glycosylation sites of the repeats are not conserved. The two cysteine residues in the repeats are conserved and both are separated by 47 amino acids in their respective domains. The second repeat is shown to be more hydrophilic than the first repeat upon analysis of the predicted amino acid sequence with the PEPPLOT program.

After the internal repeats there are 90 amino acids before the stop codon TGA. This region, designated R2c, begins with nucleotide 1061 (Tyr-330) and continues through 1315 (Ser-414) , and contains no glycosylation sites. There are two cysteine residues 39 amino acids apart. There is 655 bp of 3* non- coding sequence following the stop codon.

Example 2: Isolation and Characterization of cDNA

Encoding an Apparently Membrane Bound Form of Placental Specific SPl.

Clone hPS2, expressed uniquely in placenta and containing two internal EcoRI sites, was sequenced and characterized as described in Example 1. This clone represented partial cDNA with an incomplete 5¹ coding sequence. To obtain more 5' sequence, the most 5' EcoRI-BamHI fragment of hPS2 was used as a probe to rescreen the same cDNA library. No more 5^! sequence was found.

The nucleotide and predicted amino acid sequence of hPS2 is shown in Figure 3. It has 1744 bp with an open reading frame of 1053 bp encoding 351 amino acids, a stop codon TGA, and a 3' non-coding sequence of 659 bp with the polyadenylation signal ATTAAA 14 bp upstream from a 29 bp poly(A) tail.

The peptide encoded by hPS2, as compared with hPSll and hPS12, contains part of the N-terminal domain including 92 amino acids that are 91.2% homologous at the nucleotide level and 82.2% homologous at the amino acid level. Two of the three glycosylation sites are also conserved (a total of five potential glycosylation sites in the encoded protein can be identified) . The Rln subdomain containing 93 amino acids is 91.3% homologous at the nucleotide level and 82.3% homologous at the amino acid level. When compared to Rln of hPSll, both the glycosylation site and the positions of the cysteine residues are conserved. The c-subdomain has 81 amino acids. As in the n-subdomain, there is 91.3% homology at the nucleotide level and 92.3% homology at the amino acid level. The cysteine residues but not the glycosylation sites are conserved. However, no homology is observed beyond the c-subdomain. One important difference between the two cDNAs, hPSll and hPS2, is in the hydrophobicity of their C-terminal regions. The secondary structure of the encoded protein of hPSll, as predicted with the PEPPLOT program, shows the protein to be largely hydrophilic, indicating that it is not a membrane protein. The only hydrophobic part is at the N- ter inal 34-35 amino acids which corresponds to the signal peptide. The hydrophobic C-terminal region of hPS2 is comparable with the membrane anchor region of a number of membrane bound proteins including that of

CEA and some of the immunoglobulins. Prediction of secondary structure using programs available in the

Sequence Analysis Software Package of the Genetics

Computer Group also show that the protein is mainly in the form of beta-sheets.

Example 3: Isolation and Characterization of cDNA Encoding a third type of Placental Specific SPl.,

Figure 4 is the nucleotide sequence and the encoded amino acid sequences of hPS12, the clone described in Example 1 as representative of "group 1" placenta-specific SPl-like proteins characterized by no internal EcoRI sites. The open reading frame encodes 395 amino acids^', a stop codon and 3 * non- coding sequence of 252 bp. hPS12 has only been detected in the placenta.

This clone also represented only partial cDNA, with an incomplete 5* coding sequence. To obtain more 5¹ sequence, the most 5' EcoRI-BamHI fragment of hPS12 was used as a probe to rescreen the cDNA library. Two clones, hPS89 and hPS90, with more 5' sequence than hPS12, were identified. The composite cDNA has 1573 bp with a 5¹ non-coding sequence of 45 bp, an open reading frame of 1272 bp encoding 424 amino acids with a calculated molecular mass of 47.5 kD, a stop codon of TAA and a 3* non- coding sequence of 253 bp. Even though no upstream stop codon can be identified, the sequence around the presumed translation initiation codon ACCATGG agrees with the consensus sequence for initiation of translation in vertebrates as suggested by Kdzak, Nucleic Acid Res. 15, 8125-8148 (1987). No poly(A) tail was found in any of the hPS12 clones sequenced. Eight potential glycosylation sites, two of the form Asn-X-Ser and six of the form Asn-X-Thr are present. The encoded protein contains an N-terminal domain, two n-subdomains and one c-domain each containing two conserved Cys residues and a C-terminal domain, as is characteristic of all SPl proteins reported other than hPS2. Prediction of secondary structure with the

Sequence Analysis Software Package of the Genetics Computer Group show that the protein encoded by hPS12 is mainly in the form of beta sheets. The PEPPLOT program (Goldman, et al., Ann. Rev. Biophys. Chem. 15, 321-353 (1986)) and the PLOTSTRUCTURE program

(Chou, et al., Adv. Enzvmol. 47, 145-147 (1978)) both indicated that hPS12 is largely hydrophilic except for the N-terminal amino acids. The N-terminal 34-35 amino acids have the characteristics of a putative signal peptide as observed for hPSll. The N-terminal domain of hPS12, as in hPS2, is hydrophobic.

Comparison of the domain structure of hPS12, hPSll and hPS2 shows that hPS12 is very similar to hPSll. There is 90.6% homology at the amino acid level and 93.7% homology at nucleotide level between hPS12 and hPSll. The cysteine residues and majority of the glycosylation sites are conserved. Both hPS12 and hPSll differ from hPS2 by having an extra n- subdomain and but not a hydrophobic c-terminal domain.

Examples 1-3 demonstrate that there are multiple species of highly homologous SPl-like proteins present in the human placenta, including at least one membrane bound and one cytoplasmic or secretory protein. Example 4: Isolation and Characterization of cDNA Encoding SPl-like Proteins expressed in Cells of Testis and Placental Origin.

The presence of SPl in human testis was shown

5 by Western blot analysis of testicular homogenate with anti-placental SPl antibodies. Instead of two major hybridization bands corresponding to 72 kDa and

61.5 kDa, as observed with human placental extract, testicular homogenate shows only one weakly

10 hybridizing band corresponding to 47 kDa. The presence of SPl in testis was further confirmed by Northern blot analysis with labelled placental SPl CNA as the probe. Screening of human testis cDNA library with placental SPl cDNA as the probe yielded

15 17 positive clones which can be divided into four groups according to their restriction enzyme maps.

The procedure was as follows. A partial cDNA clone of SPl, lambda hPS3, was isolated from a human placental library. ^' The nucleotide sequence is shown

20 in Figure 5. Lambda hPS3 cDNA includes the 5' non- translating sequence and the N-terminal 48 amino acids of human SPl. This cDNA insert subcloned into M13mpl8 was released from the recombinant phage with EcoRI, fractionated by electrophoresis in 3.5% - 25 polyaery1amide gel, and recovered by electroelution according to the method of T. Maniatis et al. , Molecular Cloning: A Laboratory Manual pp. 167, 173- 177, 184 (Cold Spring Harbor Laboratory, Cold Spring Harbor, NY 1982) . SPl cDNA was labeled by nick

30 translation with DNase 1 and DNA polymerase 1 in the presence of 15 μCi of each of the (α-³²P) deoxynucleotides according to the methods of M. W. Rixon et al., Biochemistry 22, 3237-3244 (1983). The nick translated probe was desalted through an AcA 54

35 column and denatured before use. Group 1, contains two clones, lambda hTS2 and lambda hTS3. Estimation of cDNA insert size by polyacryla ide gel electrophoresis and nucleotide sequence analysis shows that both clones are about 550 bp, and are identical to nucleotide base 433 to 910 of placental SPl cDNA hPSll, shown in Figure 2. These clones do not have internal EcoRI sites.

Group 2 consists of three clones, ^"hTSl, hTS7 and hTS13. cDNA inserts of this group have an internal EcoRI site. Sequence analysis of the 3' 750 bp EcoRI fragment of the longest cDNA, hTSl, shows that it has 95.9% homology with nucleotide base 1332 to 1930 of hPSll. The nucleotide sequence of 3' EcoRI fragment of hTSl (hTSIL) , as compared with the nucleotide sequence of hPSll, is shown in figure 6. The 5' EcoRI fragment appears to indicate that the group 1 and group 2 clones are allelic gene products. Group 3 consists of three clones, hTS5, hTS6, and hTS9. Sequence analysis of these clones shows that they are identical to pSPl-i, one of the placental SPl cDNA reported by Rooney, et al., Gene 71, 439-449 (1988).

Group 4 consists of nine clones. Five cDNAs of this group were sequenced. Their nucleotide sequences are identical. The nucleotide sequence of the longest clone, hTS16, is shown in figure 7, along with the potential glycosylation sites, conserved cysteine residues, and the boundaries of the exon/intron and domain regions. It has 1957 bp and is apparently an incompletely spliced product of the SPl gene. The sequences marked as IVS1 and IVS2 have no homology with any SPl cDNA sequence known and are apparently unspliced introns. There is also a deletion of 161 bp within the coding sequence (between nucleotide bases 1314 and 1315) . The presumed introns and the deletion of hTS16 occur between boundaries of the different domains. This is similar to what was observed regarding the positions of introns in a different SPl gene by Oikawa, et al., Biochem. Biophvs. Res. Comm. 156, 68-77 (1988) The segments of sequence marked N, Rln, R2n, R2c and C and the sequence after the stop codon are identical with that of the group 3 clones and encode a protein identical to pSPl-i.

Example 5: Isolation and Characterization of cDNA

Encoding SPl-like Proteins expressed in HeLa Cells and He atopoietic Cells.

Screening of a HeLa cell library using hPS3 probe yields 10 positive clones with cDNA inserts varying from 630 to 950 bp in size, as indicated by polyacrylamide gel electrophoresis. Southern blot analysis shows that the HeLa cell clones are highly homologous to placental SPl cDNA. The nucleotide sequence of one of the HeLa cell cDNA clones, hHS2, was determined.

Figure 8 shows the nucleotide sequence and the encoded amino acids of hHS2. hHS2 has 756 bp with an open reading frame encoding 248 amino acids. Alignments of hPSll and hHS2 according to the different domains at the nucleotide level and at the protein level demonstrate that hHS2 cDNA contains the N-terminal domain, the first internal repeat and part of the second repeat of hPSll. Percent homology with the corresponding domains in hPSll is 92.5, 95.0 and 92.7, respectively, at the nucleotide level, and 87.1, 88.2 and 94 at the amino acid level. The positions of the cysteine residues and the potential glycosylation site are conserved in hPSll and hHS2. SPl mRNA has also been identified in different hematopoietic cells. Expression of different SPl genes appears to be lineage specific. Culture cells were stimulated for four hours with 10 ng/ml of phorbol 12-myristate 13-acetate (PMA) (P) , 5 μg/ l lipopolysaccharide (LPS) (L) , 100 U/ml interferon-gamma (IFN-gamma) (gamma) , or the combination of LPS and IFN-gamma (L.gamma) . Northern blot analysis revealed the presence of hybridizing mRNA in unstimulated cells (C) with a 2.4 k band in the KG-1 line (myelomonocytic) and a 3.3 kb band in the HEL line (erythroid) . Stimulation with LPS and IFN-gamma have little effect on SPl mRNA expression in KG-1 over a four hour period. However, LPS plus IFN-gamma increases SPl mRNA expression in HEL several fold. The results demonstrate for the first time both expression of the SPl genes in hematopoietic cells and differential expression of the SPl genes.

Example 6: Isolation and Characterization of cDNA Encoding SPl-like Proteins expressed in Cells of Intestinal Origin.

A human intestine cDNA library was screened with hPS3 using procedures described in Example 5.

21 positive clones were identified. Analysis with restriction enzymes showed that these clones can be divided into four main groups. Four of the clones have one internal EcoRI site. Polyacrylamide gel electrophoresis analysis showed that the longest clone of this group, hISl, gave two insert fragments of l.l kb and 1.3 kb upon digestion with EcoRI. 17 clones have no internal EcoRI sites but include an interna Xbal site. The longest clone in this group is hIS3 with a cDNA insert of 1.5 kb. The other three groups do not have internal Xbal sites and have similar restriction enzyme maps as those of the three groups of placental SPl cDNA. The smaller EcoRI fragments of clones hISl and hIS2 of the first group have been sequenced. cDNA inserts of these two clones are identical.

Figure 9 shows the nucleotide sequence and the encoded amino acids of hISl. hISl has 1024 bp with a 5" non-translated sequence of 61 bp, an open reading frame of 963 bp encoding a potential signal peptide of 34 amino acids, an N-terminal domain of 107 amino acids and a complete repeating unit consisting of an n-subdomain of 93 amino acids and a c-subdomain of 85 amino acids. Comparison of the corresponding domains of hISl and hPSll shows that the percentage of homology between the two cDNAs is approximately 74.5, with the highest homology between the n-subdomains (79.3%) and the lowest homology between the c-subdomains (68.3%). The percentage homology of the encoded amino acid sequence of the two cDNAs is approximately 57.4, and is also higher between the n-subdomains and lower between the c- subdomains.

Example Ϊ Comparison of Members of the Pregnancy- specific Bl glycoprotein (SPl) .

A human genomic library was screened, clones identified and restriction mapped, as described in

Examples 1-3. Partial restriction maps were constructed which demonstrated the presence of at least seven groups of unique SPl genomic clones and suggested that multiple genes code for SPl. The multigene nature of SPl was confirmed by hybridization of the SPl cDNA probe to multiple bands on Southern blots of human genomic DNA. Further analysis with chromosomal DNA dot blot demonstrated the presence of homologous sequences on the X chromosome and autosomal chromosome 6.

The methods and materials used to localize the SPl genes using a mixture of labeled hPS2 and hPS3 cDNAs to screen a human genomic library was reported by Chan, et al., in Am.J.Hu .Genet. 43,152- 159 (1988).

Comparison of the encoded amino acid sequence of all SPl cDNAs was used to derive a consensus sequence, shown in Figure 11. During derivation of the consensus sequence for the amino coding region hPSll, PSG16, PSBGC and PSBGD were considered as one unit because of their near identical coding sequence and the probability that they are products of the same gene. An amino acid or nucleotide is considered to be consensus if it occurs at least three times among the clones hPS12, hPS2, PSBGE, pSOl-i and hPSll/PSG16/PSBGC-D. All cDNAs showed a high degree of homology in the regions compared. Percent of homology with the consensus amino acid* sequence ranged from 92.5 to 100% in the N-terminal domain and the n-subdomains except for the N-terminal domain of hPS2 which shows only 85.9% homology. Levels of homology among the c-subdomains are slightly lower than the other domains and range from 85.9 to 97.6%. Nucleotide sequence homology of the N-terminal and n- and c-subdomains are similar to that of amino acid sequences and range from 90 to 99%. In contrast, the amino acid and nucleotide sequences of the C-terminal domain of these cDNAs share very little sequence homology with two exceptions: the C-terminal domain of hPSll and PSBGD are identical and the 3' non- coding sequence of PSG16 is identical with that of hPSll other than for the deletion of 86 bp near the 3¹ end of the coding region. Positions of the cysteine residues in the n- and c- subdomains of all cDNAs are conserved. Positions of the potential glycosylation sites in the N-terminal domain and Rln- subdo ain are also conserved. Comparison of the aligned amino acid sequences encoded by human placental SPl cDNAs shows that hPS12 differs from the other SPl cDNAs in the deletion of three nucleotides encoding Ile-90 in the N-terminal domain. The comparison also makes clear that hPS12 and pSPl-i are different from each other as well as the other SPl cDNAs. Again, hPSll and PSBGD are identical. PSG16 is identical to hPSll except at four positions in the coding region resulting in the change of three amino acids, a G to C mutation in the 3' non-coding sequence and deletion of 86 bp near the 3' end of the coding region. PSBG is also identical to hPSll and PSBGD except for the C-terminal domain. Both hPS2 and PSBGE have one less n-subdomain when compared to the other SPl cDNAs. The n- subdomain of hPS2 and PSBGE are more comparable with the Rln-subdomain than the-R2n-subdomain. The percentage of homology with the Rln-subdomain consensus sequence is 93.5 and 92.5 and that of the R2n-subdomain is 52.7 and 48.4 for hPS2 and PSBGE respectively. The c-subdomain of hPS2 and PSBGE are less homologous to the consensus sequence than that of the other SPl cDNAs, being 87.1% and 85.9% respectively, while all other cDNAs show greater than 90% homology. The most significant difference between hPS2 and the other SPls is the presence of the 81 amino acid hydrophobic C-terminus while all other SPls have only relatively short (14 amino acids or less) hydrophilic C-termini.

Comparison of the 5' non-coding sequence of several SPl cDNAs shows over 91% similarity among all of the SPl sequences. Comparison of the 3' non- coding sequences produces more variation. The 3 * non-coding sequence of hPS2 has very little homology with any of the SPl cDNAs reported. On the other hand, the 3* non-coding sequence of hPS12 is very homologous to that of PSBGE and pSPl-i. Aside from 27 bp, 30 bp and 39 bp at the 5' end of this region in hPS12, PSBGE and pSPl-i, respectively, there is 95% homology among the three cDNAs. The 3' non- coding sequences of hPSll, PSG16 and PSBGD are almost identical except for the deletion of 70 bp after the stop codon in PSG16 and one mismatch in both PSG16 and PSBGD when compared with hPSll.

All the SPl cDNAs reported demonstrate characteristics required for inclusion in the Ig gene superfamily, namely sequence homology, characteristic domain structure, conserved disulfide bond within domains and beta sheet structure.

The evidence shown in Examples 1 to 6 suggests the presence of multiple species of highly homologous SPl proteins in human placenta and that the SPl proteins are encoded by multiple genes. The results shown in Examples 1 to 3 support the contention that SPl protein in human placenta consists of products of three or more genes. Two of the reported placental SPl cDNAs, hPSll and PSBGD are identical. PSG16 and another partially sequenced cDNA PSG93 differ from these two cDNAs at only four bases which could correspond to polymorphisms. Based on their almost identical protein coding sequences, hPSll, PSG16, PSBG93,PSBGC and PSBGD, are likely to be the products of the same gene with differentially spliced exons encoding the C-terminus and 3 * non- coding sequence. Three other cDNAs hPS12, PSBGE and pSPl-i have 221 bp in their 3¹ non-coding sequence which are highly homologous (95%) . Considering that this is observed in the non-coding region and that the three clones are derived from two different libraries, it is conceivable that the few differences are individual polymorphisms. These three cDNAs are products of splicing of different amino acid coding exons to one common exon which contains the 3* non- coding sequence. The exon encoding the N-terminal domain of hPS12 is unique in that it has a 3 bp deletion when compared to the other cDNAs. PSBGE differs from hPS12 and pSPl-i by having one less n- subdomain which could be the result of difference in splicing. The gene encoding hPS12/PSBGE/pSPl-i could be different from that encoding hPSll/PSG16/PSG93/PSBGC-D. The other SPl CDNA, hPS2, is unique. It has no significant homology with any of the SPl cDNAs reported. The presence of three different species of mRNA in. human placenta was confirmed by the Northern blot analysis. The 3' EcoRI fragment of hPSll contains the entire 3*. non-coding sequence and is specific for that cDNA. The Ncol-EcoRI fragment of hPS12 contains the C-terminal 24 bp of the coding sequence and the 3' non-coding sequence and is unique for cDNAs of this group. The probe for hPS2 is the fragment encoding the unique hydrophobic C-terminal domain of the molecule. These three probes are cDNA specific. Hybridization of the Northern blot with these probes therefore indicates the presence of the specific mRNA species. The hPS2 specific probe hybridized to both the 1.65 Kb and 2.25 Kb mRNA bands, suggesting that each mRNA band contains more than one species of mRNA. Up to the present time only one placental SPl cDNA with sequence encoding a hydrophobic C-terminal has been found. The Northern blot results suggest that there may be more than one species of membrane-bound SPl present in human placenta. Multiple mRNAs have also been reported for the CEA gene family. The SPl genes, like the CEA family of genes, are probably localized in clusters on chromosome 6 and the X chromosome, although there is also the possibility that all SPl cDNAs so far reported are derived from one very large gene which gives rise to the above three groups of products by differential splicing.

Both hPS2 and PSBGE, unlike the other SPl cDNAs, contains only one n-subdomain. These cDNAs could be formed by having one complete repeating unit containing an n-subdomain and c-subdomain (i.e. n- subdomain in hPS2 is R2n-subdomain) or by splicing the n-subdomain of one repeating unit with the c- subdomain of another repeating unit (i.e. n-subdomain in hPS2 is Rln-subdomain) . By comparisons with the consensus sequence of Rln- and R2n-subdomains of the SPl cDNAs, it is that the n-subdomain of hPS2 is likely to be derived from the Rl repeating unit instead of the R2 repeating unit. A similar conclusion is drawn for PSBGE.

The distinctive feature of hPS2, the presence of an 81 amino acids hydrophobic C-terminus, is very similar to that observed in CEA, TM-CEA and NCA. Hydropathy plot analysis showed that this C-terminal domain of hPS2 is very hydrophobic and supports the inference that it represents the membrane-anchor region of the molecule. These results support the theory that there are two types of SPl proteins in human placenta, the cytoplasmic or soluble SPl and the membrane-bound SPl. Analogous phenomenon has been reported for the CEA family of proteins.

Even though the SPl proteins and CEAs appear to share a lot of common properties, there are sufficient features unique to the SPl proteins to qualify them as a separate subfamily of the Ig gene superfamily instead of being members of the CEA gene subfamily. For example, even though the SPls are very similar to CEA and NCA, the homology among the different SPl family members are even higher, being consistently greater than 90% at the nucleotide level and greater than 85% at the amino acid level. The SPl proteins are also less glycosylated. The number of potential glycosylation site ranges from four in pSPl-i to eight in PSBGC with the majority having six or seven. Most of these sites are also conserved. Both CEA and NCA are more heavily glycosylated, with twenty-seven potential sites in NCA and twelve potential sites in NCA. A number of these sites are conserved between CEA and NCA but not between CEAs and SPls.

The strong conservation of both nucleotides and amino acids among the internal repeats of SPl and CEA genes suggests that both gene families evolved recently by the duplication of a primordial gene. The percentage homology at the amino acid level between the Rln-subdomain and R2n-subdomain of SPl is 52.7, which is significantly lower than that between the Rln-subdomain of SPl and the n-subdomains of CEA (62.4 with IA, 57.0 with IIA, and 59.1 with IIIA) and NCA (62.4). It is comparable with the percentage similarity between the R2n-subdomain and the n- subdomains of CEA and NCA. Comparison of nucleotide homology gives the same results, suggesting that the duplication of Rl repeat unit to R2 repeat unit occurred before the divergence of SPl and CEA and that the divergence of the two genes involved the duplication of the Rl repeat unit only. Example 8: Comparison of CEA and SPl proteins and Removal of Cross-reactive proteins from CEA assays.

As described above, cDNA clones for seven apparently different, but closely related, genes for

SPl proteins have been isolated. One group of clones appears to encode placenta-specific SPl proteins.

Clones in this group are characterized by not having an internal EcoRI site. An example is hPS12. A second group of clones, reported by Chan, et al., DNA 7(8), 545-555 (1988), includes hPSll, which has one internal EcoRI site, and has been isolated from placenta and non-placental sources, including testis, although it is not known whether hPSll is secreted from normal testis into the blood stream. While similar, hPSll differs from the BSG16 reported by Watanabe and Chou, J. Biol. Chem. 263(4), 2049-2054 (1988), at amino acids 41, 43 and 319, and 86 nucleotides near the 3* end of the coding regions. A third group of clones has been isolated only from placenta and has two internal EcoRI sites. An example of the third group is hPS2.

Clones have also been isolated from a HeLa cell cDNA library, a testis cDNA library and an intestinal cDNA library, which are highly homologous to placental SPl cDNA. hISl shows higher homology to Normal Crossreacting Antigen (NCA) , a member of the CEA family, than to the placental^' sPl cDNAs. Figure 10 compares the domain structures among hPSll, hISl and CEA. hISl is 93% homologous at the nucleotide level, and 67% homologous at the amino acid level, with CEA. The N-terminal domains of both hISl and CEA are two amino acids shorter than that of hPSll. The domain structure, as well as the number of potential glycosylation sites, are more similar between hISl and CEA than hPSll.

All cloned SPl cDNAs show significant structural homology with carcinoembryonic antigen (CEA) . CEA is a large membrane glycoprotein present in high quantities in adenocarcinomas of endodermally derived digestive tract epithelia and fetal colon. There are a number of molecules closely related immunologically to CEA. These molecules have very similar but distinct amino-terminal amino acid sequences and have been postulated to be encoded by different genes by Shively and Beatty (1985) . As shown in the foregoing examples, the SPl genes encode proteins which are structurally and chemically similar to CEA and are expressed, at least in part, in tissue where CEA is also expressed. Among the placental SPl cDNAs, hPS2 and PSBGE have the domain structure closest to that of CEA and NCA. Both hPS2 and PSBGE, similar to NCA, have only one complete repeat unit, i.e, one n-subdomain and one c- subdomain, while CEA has three complete repeating units. All other SPl cDNAs have one complete repeating unit and an additional n-subdomain. Comparison of the consensus amino acid sequences of the different domains of the SPl cDNAs with the corresponding domains in CEA and NCA shows that the percentage homology of both CEA and NCA to SPl is quite similar in all domains compared, ranging from 53.8% to 62.4% at the amino acid level. Nucleotide sequence comparison of the different domains of SPl, CEA and NCA showed similar results as that of amino acid comparison. It is therefore possible to use the nucleotide and protein sequences for the SPl-like proteins to screen reagents used in detecting CEA, to thereby decrease the number of false positives arising due to detection of the SPl-like proteins rather than the CEA, in various clinical circumstances including but not limited to monitoring and detection of tumors of the gastrointestinal tract.

Example 8s Use of Purified SPl Protein as an Immunosuppressive Agent.

Immunological studies show that pure human placental SPl (containing all species of SPl protein present in the placenta capable of reacting with the same antibodies) at a concentration of as low as 10 μg/ml inhibits lymphocyte proliferation significantly (50%) in a mixed lymphocyte assay. Inhibition is 85 at 30 μg/ml and 100% at 100 μg/ml. The results of two independent studies are shown in Figure 12. This immunosuppressive effect of SPl appears to be specific since phytohemagglutinin (PHA) stimulated lymphocytes are not affected by comparable dosage of SPl. The results confirm the structural evidence for an immunological role for the placental SPl proteins.

Example 9: Use of Purified SPl Protein as a Growth Factor.

Growth promoting activity of placental SPl was tested by addition of varying amounts of pure human placental SPl preparation to serumless murine bone marrow cultures. The effect of SPl was followed by assessing the number and size of megakaryocytes as well as measuring acetylcholinesterase (AchE) activities. Addition of SPl to murine bone marrow cultures induced a significant increase in the diameter of megakaryocytes. The potency of SPl in inducing megakaryocytopoiesis is comparable to that of interleukin 6, an established growth factor stimulator for megakaryocytes. The effect of SPl on murine megakaryocyte AchE is shown in Figure 13. SPl at a concentration of as low as 10 ng/ml caused a 45% increase in AchE activity of megakaryocytes. These results demonstrate that SPl has growth promoting activities.

Example 10: Use of pregnancy-specific proteins and nucleic acid sequences as pharmaceutical agents.

There are at least three proteins, including the protein encoded by hPSll, hPS2 and hPS12, which appear to be expressed in placental tissue. hPS2 and hPS12 have only been found in placental tissue. Since placental tissue exists only in individuals who are pregnant, the levels of hPSll, hPS2 and hPS12 proteins can function as an index of whether the female, either a human or non-human mammal such as a rat, cow or pig, is pregnant. The present invention therefore includes a method and reagents for testing for pregnancy wherein a biological sample from the subject is assayed for the level of one of the placenta specific peptides, and correlating an above- normal level, if any, with pregnancy, and correlating a normal or below-normal level, if any, with the absence or abnormalities of pregnancy. In one embodiment of the pregnancy testing method of the present invention, a biological sample from the female is obtained and is treated with monoclonal antibodies against the protein described in Figure 2. The level of antibody-protein reaction is measured using methods known to those skilled in the art (i.e., radiolabelling, ELISA, etc.) to determine the level of protein in the sample.

The present invention also includes a method and compositions for enhancing the fertility of a female deficient in one of the placenta specific proteins such as the protein described in Figure 2 wherein an effective amount of the protein or peptide is administered to the female. .

The peptide is preferably administered in combination with a pharmaceutically acceptable carrier, as described below. The preferred dosage of the peptide is based on the normal levels of the protein and is preferably administered as determined by those skilled in the art. In general, this will be between 0.1 and 50 mg/kg body weight. Any pharmaceutically acceptable carrier may be selected which will maintain the biological activity of the protein, such as 5% dextrose in sterile water and sterile normal saline. The present invention further includes a method and compositions for enhancing the viability of a fetus wherein an effective amount of a pregnancy specific peptide or protein such as the one shown in Figure 2 is administered to either the mother or the fetus. If the mother has an inadequate level of SPl- like proteins, a spontaneous abortion could occur. Accordingly, the administration of the composition can raise the level of SPl proteins necessary to maintain a viable fetus. The peptide is preferably administered in combination with a pharmaceutically acceptable carrier, as described above.

The SPl proteins described here are present in cows, pigs, sheep, horses, d gs, cats, rats, and primates, including chimpanzee, cynomolgus, monkeys, and baboons, as well as humans. On this basis, it is assumed that the proteins are present in all placental mammals, and that the methods described above can therefore be utilized in all placental mammals. Modifications and variations of the present invention, proteins expressed from nucleotide sequences encoding members of the SPl protein family, antibodies and hybridization probes to the proteins and sequences, and methods for use thereof, will be obvious to those skilled in the art from the foregoing detailed description of the invention. Such modifications and variations are intended to come within the scope of the appended claims. We claim:

Claims

1. A genetically engineered nucleic acid sequence encoding a pregnancy-specific β-1 glycoprotein-like protein (SPl-like protein) having at least 70% homology to Carcinoembryonic Antigen (CEA) .

2o The nucleic acid sequence of claim 1 wherein the sequence is specifically expressed in placenta and the protein is encoded at least in part by the nucleotide sequence of clone hPS12.

3. The nucleic acid sequence of claim 1 wherein the sequence is expressed in placenta and the protein is encoded at least in part by the sequence of clone hPSll.

4. The nucleic acid sequence of claim 1 wherein the sequence is specifically expressed in placenta, and the protein contains a hydrophobic C- terminal region and is encoded at least in part by the sequence of clone hPS2.

5. The nucleic acid sequence of claim 1 wherein the sequence is expressed in intestinal tissue and the protein is encoded at least in part by the DNA from an intestinal gene library which hybridizes with a nucleic acid sequence encoding a SPl-like protein.

6. The nucleic acid sequence of claim 5 wherein the sequence selected from the group consisting of hISl and hIS3.

7. The nucleic acid sequence of claim 1 wherein the sequence is expressed in both testis and placenta tissue and the protein is encoded at least in part by a sequence selected from the group consisting of clones hPS3 and hPSll.

8. The nucleic acid sequence of claim 1 wherein the sequence is expressed in testis and the protein is encoded at least in part by a sequence selected from the group consisting of clones hTS2 and hTS3.

9. The nucleic acid sequence of claim 1 wherein the sequence is expressed in HeLa cells and the protein is encoded at least in part by the sequence of a clone selected from the group consisting of hHS2.

10. A substantially pure homogeneous protein sequence forming at least a portion of a pregnancy- specific β-1 glycoprotein-like protein (SPl-like protein) having at least 70% homology to Carcinoembryonic Antigen (CEA) expressed from a genetically engineered nucleic acid sequence selected from the group consisting of clones hPS12, hPSll, hPS2, hISl, hIS3, hPS3, hTSl, hTS2, hTS3, hTS16, hHS2, and portions thereof.

11. An antibody to a pregnancy-specific β-1 glycoprotein-like protein (SPl-like protein) having at least 70% homology to Carcinoembryonic Antigen (CEA) expressed from a nucleic acid sequence selected from the group consisting of clones hPS2, hISl, hIS3, hPS3, hTSl, hTS2, hTS3, hTS16, hHS2, and portions thereof, wherein said antibody does not recognize CEA.

12. A substantially pure membrane bound SPl- like protein.

13. The protein of claim 12 expressed in placenta.

14. The protein of claim 13 encoded at least in part by hPS2.

15. An antibody to a membrane bound SPl-like protein.

16. A method of testing for pregnancy in a female comprising: assaying a biological sample from the female for the level of a peptide encoded by a nucleic acid sequence expressed in placenta homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll; and correlating the level with pregnancy.

17. A method of enhancing the fertility of a female deficient in SPl-like proteins comprising: administering to the female an effective amount of a peptide encoded by a nucleic acid sequence expressed in placenta homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll.

18. A method of enhancing the viability of a fetus carried by a female subject deficient in SPl- like proteins comprising: administering to the subject an effective amount of a peptide encoded by a nucleic acid sequence expressed in placenta homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll.

19. A method of inducing the rejection of a fetus comprising: administering to the mother of the fetus an effective amount of an antibody or an inhibitor of a peptide encoded by a nucleic acid sequence expressed in placenta homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll.

20. A composition for enhancing the fertility of a female comprising: a pharmaceutically acceptable carrier and a peptide encoded by a nucleic acid sequence expressed in placenta and homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll.

21. A composition for enhancing the viability of a fetus carried by a female deficient in SPl-like proteins or inducing rejection of a fetus comprising: a pharmaceutically acceptable carrier; and a peptide encoded by a nucleic acid sequence expressed in placenta and homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll.

22. A method for screening for CEA comprising: providing an antibody reactive with CEA but not with a pregnancy-specific β-1 glycoprotein-like protein (SPl-like protein) having at least 70% homology to Carcinoembryonic Antigen (CEA) expressed from a nucleic acid sequence selected from the group consisting of clones hPS2, hPSll, hPS12, hISl, hIS3, hPS3, hTS2, hTS3, and hHS2.

23. An antibody reactive with CEA but not with a pregnancy-specific β-1 glycoprotein-like protein (SPl-like protein) having at least 70% homology to Carcinoembryonic Antigen (CEA) expressed from a nucleic acid sequence selected from the group consisting of clones hPS2, hPSll, hPS12, hISl, hIS3, hPS3, hTS2, hTS3, and hHS2.

24. The assay of claim 22 wherein the SPl- like proteins are expressed from sequence present in the X chromosome and chromosome 6.

25. The assay of claim 22 wherein the SPl- like proteins are expressed from sequence transcribed in extraplacental tissue.

26. A method for enhancing transfer of immunoglobulins and immunoglobulin-like molecules to a fetus comprising administering a SPl-like protein encoded by a nucleic acid sequence expressed in placenta and homologous to a clone selected from the group consisting of hPS2, hPS12, and hPSll.

27. A method of inducing the rej ction of a fetus comprising: administering to the subject an effect amount of an antibody to a pregnancy-specific βl- glycoprotein (SPl-like protein) .

28. A method^* for inhibiting lymphocyte proliferation comprising: administering an effective amount of a pregnancy-specific β-1 glycoprotein-like protein (SPl-like protein) .

29. The method of claim 28 wherein the protein is of placental origin.

30. The method of claim 28 wherein the protein is isolated by binding to antibodies to SPl proteins.

31. A method for stimulating growth and proliferation of cells comprising: administering an effective amount of a pregnancy-specific β-1 glycoprotein-like protein (SPl-like protein) to the cells.

32. The method of claim 31 wherein the protein is of placental origin.

33. The method of claim 32 wherein the protein is isolated by binding to antibodies to SPl proteins.