Identification of a New Immunoglobulin Superfamily Protein Expressed in Blood Vessels with a Heparin-Binding Consensus Sequence

[CANCER RESEARCH 55, 2140-2149, May 15, 1995] Identification of a New Immunoglobulin Superfamily Protein Expressed in Blood Vessels with a Heparin-binding Consensus Sequence

Marie E. Beckner,1 Henry C. Krutzsch, Mary L. Stracke, Sonya T. Williams, Jorge A. Gallardo, and Lance A. Liotta

Laboratory of Pathology, National Cancer Institute. N1H, Bethesda, Maryland 20892

ABSTRACT We have identified and sequenced cDNA from a human melanoma cell library that encodes a new protein, called AAMP, that shares A novel immunoglobulin-type protein expressed in blood vessels has characteristics with immunoglobulin superfamily proteins and con been identified. The cDNA for AAMP (angio-associated, migratory cell tains a heparin-binding consensus sequence. Immunoblotting and protein) was first isolated from a human melanoma cell line during a search for motility-associated cell surface proteins. Upon analysis of the immunohistochemistry using affinity-purified antibodies with speci tissue distribution of AAMP, it was found to be expressed strongly in ficity for rAAMP and for a derived peptide, P189, were used to endothelial cells, cytotrophoblasts, and poorly differentiated colon adeno- identify the native protein in human cultured cells and tissues. In carcinoma cells found in lymphatics. The sequence of AAMP predicts a addition, the heparin-binding capacity of peptide 189 was determined protein (M,. 49,000) with distant identity (25%) to known proteins. It in order to evaluate the potential role of this epitope in cell adhesion. contains immunoglobulin-like domains [one with multiple homologies to deleted in colon carcinoma (DCC) protein], the WD40 repeat motif, and a heparin-binding consensus sequence. A 1.6-kilobase niKNA transcript of MATERIALS AND METHODS AAMP is detected in tissue culture cell lines and tissues. Affinity-purified polyclonal antibodies, anti-recombinant AAMP, and anti-peptide 189 cDNA Library Screening. A human melanoma A2058 cDNA expression (AAMP derived) recognize a ,Ur 52,000 protein in human tissue and library (Agtll) was screened with an antibody, 1AA3AA, suspected to inhibit cellular extracts. The protein size is in keeping with the inKNA and cell motility. predicted sequence. The AAMP-derived peptide, P189, contains a hepa 1AA3AA Monoclonal Antibody Preparation. Mouse hybridoma clones rin-binding domain (dissociation constant, 14 pmol) and mediates hepa- were generated with A2058 melanoma cells using the myeloma cell line rin-sensitive cell adhesion. The shared expression of AAMP in endothelial X63.Ag8.653 (a generous gift of R. P. Siraganian, National Institute of Dental cells, trophoblasts, and tumor cells implies a common function in migrat Research, NIH, Bethesda, MD) (20-22). They were screened for inhibition of ing cells. melanoma chemotaxis (23) to autotaxin (24), type IV collagen (Collaborative Biomedicai Products, Bedford, MA), and laminin (a gift of S. Aznavoorian, National Cancer Institute, NIH, Bethesda, MD). INTRODUCTION DNA Sequencing. The Agtl 1 phage insert, AAMP, from the A2058 cDNA Molecules that mediate cell-cell and cell-substrate interactions in library, was subcloned into Bluescript plasmid (Stratagene, La Jolla, CA) for double-stranded DNA sequencing (both strands; Ref. 25). PCR products gen clude members of the immunoglobulin superfamily. These contain erated from the amino terminal region of AAMP cDNA (obtained with human immunoglobulin-like domains that share evolutionary homology and brain mRNA; Clontech, Palo Alto, CA) were sequenced with Taq polymerase function primarily in recognition and binding processes on cell sur (Perkin-Elmer Roche Molecular Systems, Branchburg, NJ) both directly and as faces (1-5). Most are cell surface proteins, a few are intracellular, and subcloned cDNA (26-28). some are secreted (4). Immunoglobulin superfamily proteins and the ALIGN Sequence Comparisons. Immunoglobulin domains (20 residues mechanisms involved in their regulation of migratory cells are of beyond each predicted cysteine bond) of immunoglobulin superfamily mem bers were ALIGN matched with AAMP (29) using standard parameters. Scores special interest. These proteins include those that help mediate endo of 3.1, 4.3, and 5.2 SD indicate probabilities of 10~3, 10~5, and 10~7, thelial cell interactions with lymphocytes, monocytes, neutrophils, and tumor cells that result in adhesion and transendothelial migration respectively, that unrelated domains show similarity by chance (30, 31). (6-9). Activation of T cells (10,11) and the processes of cell adhesion Recombinant AAMP Purification. AAMP insert was subcloned into pGEX-lÃ€T Â£coRI/BAP plasmid (Pharmacia, Uppsala, Sweden) for production and migration in the nervous system also involve immunoglobulin of AAMP protein fused with glutathione 5-transferase in Escherichia coli superfamily members (3). The binding mechanisms of some immu according to the manufacturer's instructions. Bacteria were induced with noglobulin superfamily proteins that participate in the above pro isopropyl-ÃŸ-D-thiogalactopyranoside, sonicated, and centrifuged. Supernates cesses, including PECAM,2 NCAM, CD4, as well as other adhesive were loaded onto glutathione Sepharose columns (Pharmacia), washed, and proteins such as CD44, involve glycosaminoglycans (heparin, chon- thrombin (Sigma Chemical Co., St. Louis, MO) digested for purification of droitan sulfate, and hyaluronan, in these examples) (10,12-18). Some recombinant AAMP from glutathione S-transferase. Digestion ( 1 h) occurred at immunoglobulin superfamily members are multifunctional and par a thrombin-sensitive site in AAMP, aa 14. ticipate in both cell binding and signaling (4, 19). Determining the Peptide Preparation. AAMP-derived peptides and variants were synthe identity and characteristics of new proteins that participate in cellular sized on a Biosearch model 9600 synthesizer (Bedford, MA) using standard Merrifield solid phase synthesis protocols and f-butoxycarbonyl chemistry and interactions involving migrating cells and endothelium should help were analyzed by reverse-phase HPLC. To generate peptides from recombi elucidate mechanisms involved in biological processes such as in nant AAMP, it was reduced and alkylated (AMsopropyliodoacetamide), di flammation, wound healing, and mÃ©tastasesof tumor cells. gested with trypsin, and subjected to HPLC chromatography (32). Peptide Sequencing. Amino acid sequence analysis was performed using Received 9/26/94; accepted 3/2/95. a Porton/Beckman 2090 on-line sequencer (Fullerton, CA) using standard The costs of publication of this article were defrayed in part by the payment of page program no. 1. Phenylthiohydantoin amino acid analysis was carried out on a charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Beckman System Gold system (Fullerton, CA) using a modified sodium 1To whom requests for reprints should be addressed, at Laboratory of Pathology, Bldg. acetate gradient program and a Hewlett-Packard narrow bore C18 column. 10, Rm. 2A-33, National Cancer Institute, NIH, 9000 Rockville Pike, Bethesda, MD Polyclonal Antibody Preparations. Polyclonal antipeptide antibodies spe 20892. 2 The abbreviations used are: PECAM, platelet-endothelial cell adhesion molecule; cific for AAMP (generated in rabbits using peptides conjugated to bovine NCAM, neural cell adhesion molecule; AAMP, angio-associated migratory cell protein; albumin) were affinity purified on columns of peptide covalently attached to rAAMP, recombinant AAMP; aa, amino acid(s); DPBS, Dulbecco's phosphate-buffered Affi-Gel 10 beads (Bio-Rad, Richmond CA). Polyclonal anti-recombinant saline; DMEM, Dulbecco's minimum essential medium; bp, base pair(s); kb, kilobase(s). AAMP antibodies were generated in rabbits against recombinant AAMP. 2140

Anti-rAAMP was affinity purified on columns of Affi-Gel 10 beads with human brain mRNA with gene-specific, nested primers using 5'- recombinant AAMP covalently attached. RACE PCR contain sequence that codes for known sequence of Human Tissue and Cell Lysate Preparations for Immunoblotting. AAMP and several additional 5' protein residues consistent with an Human tissues were homogenized in 3% SDS and 4% ÃŸ-mercaptoethanol open reading frame. Four subcloned PCR products code for aa 2-7, buffer at 100Â°C.Whole cell lysates of A2058 melanoma cells (passaged fewer followed by a "G" in the 5' direction. One of these products contains than 20 times) and bovine endothelial cells (passage 6), aortic and corneal a complete AUG codon at this site, consistent with an initiating (gifts of N. H. Guo and J. Kaiser, both from the National Cancer Institute, NIH, methionine. Several other products that also code for aa 2-7 show Bethesda, MD), were prepared in 0.5% NP40 buffer. Immunoblot Preparation. Tissue and cell lysates electrophoresed in HV/c alternative sequence (<40 bp) replacing the AUG codon at this loca SDS polyacrylamide gels were blotted and reacted with either affinity-purified tion that is currently being characterized. The reactivity of the anti- polyclonal rabbit anti-rAAMP or anti-P189 (1-2 fig/ml) and then goat antipeptide antibody generated to P189 (aa 14-25) with the natural rabbit IgG. Blots were developed with diaminobenzidine. Competition studies protein (Fig. 2) indicates that coding message for protein is present in used peptide and recombinant AAMP at 5000X and 2100X their respective the amino terminal region of AAMP. A Kozak sequence (40) has not antibody concentrations on a molar basis. been found. Twenty-two peptides obtained as digestion products from Northern Blots. Total RNA from melanoma cells, Jurkat lymphoma cells, recombinant AAMP were sequenced to confirm that the cDNA se human brain, liver, lung, breast, kidney, and placenta! and rat liver tissues was quencing of the A2058 library clone predicted the protein sequence. extracted (33, 34) and enriched for polyadenylate RNA (35). Total RNA from Recombinant protein was reduced, alkylated, and digested with cy normal and malignant human gastric and breast tissues was isolated (36). All anogen bromide and trypsin or with trypsin alone. These peptide samples were electrophoresed in formaldehyde 1% agarose gels and blotted. DNA probes (AAMP and glyceraldehyde-3-phosphate dehydrogenase) were sequence runs, which averaged eight residues in length, located the labeled with random priming. Hybridization for melanoma cell RNA blots was corresponding peptides with their amino termini at the following according to Church (37). All other blots were hybridized overnight at 42Â°C. residue positions: 15, 121, 134, 139, 153, 178, 199, 204, 221, 228, Immunohistochemistry. Routine deparaffinized sections of human non- 241, 248, 259, 264, 273, 304, 312, 402, 410, 416, 429, and 444. malignant brain, adrenal gland, uterus, and placenta, hyperplastic endome- Polyclonal antibodies (anti-Pi89 and anti-P279), generated against trium, and poorly differentiated colon adenocarcinoma were microwaved for peptides derived from AAMP at positions 14 and 352, reacted with antigen retrieval (38) in 1.5 liters of 10 mM sodium citrate, pH 6.0, at the recombinant protein encoded by AAMP cDNA, also confirming the highest setting for 40 min. Vectastain Elite ABC kits (rabbit and mouse; cDNA sequencing. Excess peptides, P189 and P279 (5000X molar Vector, Burlingame, CA) were used. Affinity-purified rabbit anti-rAAMP and anti-Pi89 (0.25â€”1(ig/ml) were test antibodies. Mouse anti-von Willebrand concentrations), competitively inhibited these reactions appropriately. AAMP Protein Expression. Anti-recombinant AAMP reacts with Factor (1:35; Dako, CarpinterÃa, CA), rabbit anti-human chorionic gonadotro- phin (1:200; Dako), and mouse anti-glial fibrillary acid protein (1:100; Dako) a Mr 52,000 protein in immunoblots of extracts from human cells and were positive controls. Rabbit IgG (1 /xg/ml) and goat anti-rabbit (1:200) were tissues. These include A2058 melanoma cells, bovine endothelial cells negative controls. Sections were counterstained in hematoxylin. (aortic and corneal), activated T cells, and the following tissues: Heparin-binding Assays. Peptides were solubilized (2 mg/ml) by boiling metastatic melanoma, adult liver, skin, kidney, heart, lung, lymph in 50% DMSO/DPBS for 10 min. P189 and P357 (6.25 /ng/ml) were immo node, and skeletal muscle. AAMP was also detected in blots of bovine bilized on CovaLink plates (Nunc, Naperville, IL; Ref. 39). CovaLink Buffer brain tissue. Anti-Pi 89, raised against a synthetic peptide whose without DPBS was used for washes and overnight incubation following pep- sequence was derived from AAMP, also reacted with AAMP in tide linkage. Following buffer removal, tritiated heparin 4 (j.g/well (0.7 mCi/ A2058 melanoma cells, activated T cells, and brain tissue (human and mg; NEN Dupont, Wilmington, DE), alone, and with unlabeled heparin (0.4-40 ^g/well; Sigma) was added in DPBS for 4 h at 27Â°CÂ±2Â°C.The wells bovine). Examples of these blots are shown in Fig. 24. The reactivity of the protein with anti-Pi89 can be competed with 5000X molar were rinsed, separated, and counted. The predominant heparin molecular weight, tritiated and unlabeled, was M, 17,000-20,000. There were 176 and concentration of peptide 189 (Fig. 2B). Reactivity with anti-recom- 143 USP units/mg in unlabeled and tritiated heparin, respectively. binant AAMP is competed also (2100X excess molar concentration). Cell Binding by P189 and Its Variants. Peptides were attached to Co Reactivity of the Mr 52,000 protein in cells and tissues on immuno vaLink plates as described above. For testing the heparin sensitivity of cell blots with both antibodies that have specificity for AAMP (recombi binding by P189 and its scrambled variant, P350, the peptide-coated wells nant protein and a derived peptide) indicate that the Mr 52,000 protein (6.25 /xg/ml) were incubated with varying amounts of heparin sodium must be the native form of AAMP. Its molecular weight is very close (Lyphomed, Deerfield, IL) for l h (27Â°CÂ±2Â°C).Washes removed unbound to that predicted by sequencing (Mt 49,000) and corresponds with the heparin before A2058 melanoma cells (100,000/well) were added, suspended 1.6-Kb message. The original screening antibody, 1AA3AA, reacts in 0.1% BSA/DMEM (diluted 1:2 with DPBS). Following 1-h incubations (37Â°C)and removal of unattached cells, the remaining cells were stained with with two bands, one of which (the smaller) corresponds to AAMP in size (data not shown). Diff Quik (Baxter Healthcare, McGaw Park, IL). Plates were photographed, scanned, and read with densitometry (one densitometer unit represents approx AAMP mRNA Expression. The AAMP mRNA transcript in imately 84 cells). Cell binding by peptides, with and without the heparin- A2058 melanoma cell blots is 1.6 Kb (Fig. 3), consistent with the binding consensus sequence, was performed in 0.1% BSA/DMEM. The Stu AAMP cDNA (1787 bp). Northern blots of other human cells and dent / test was used to test for difference. The group of peptides with the tissues show that the message of AAMP is also expressed in cultured RRXRRX motif, P189 (RRLRRMESESES), P358 (RRLRRMQSOSQS), human breast cancer cells (MCF7 and MDA231), phorbol myristate P359 (RRGRRGESESES), and P360 (RRLRRMEAEAEA) was compared to acetate/phytohemagglutinin-stimulated T cells and monocytes, Jurkat those without it, P350 (MSRERESRERSL), P357 (QQLQQMESESES), and lymphoma cells, fetal brain, lung, liver, and kidney tissues, and adult P369 (RLRRMESESE). heart, lung, brain, liver, breast, pancreas, kidney, placenta!, stomach mucosal, gastric adenocarcinoma, and breast carcinoma tissues. Ex amples of the message in RNA blots of human tissues are shown in RESULTS Fig. 4. AAMP cDNA Sequence. The screening antibody, 1AA3AA, re AAMP Homologies. Since early nucleic acid searches indicated a acted with three cDNA clones in the human melanoma A2058 cell relationship between AAMP and CD4, an immunoglobulin superfam- expression library. They were the same size and cross-hybridized. Â¡lymember, the sequence of AAMP was analyzed for the presence of One selected for sequencing, AAMP, contains a 1335-bp reading immunoglobulin-like domains. The amino acid region 94-303 con frame (aa 8-452; Fig. 1). Amino terminal fragments generated from tains predicted ÃŸsheets and turns consistent with Â¡mmunoglobulin 2141

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1995 American Association for Cancer Research. AAMP. A NEW VASCULAR PROTEIN atggactctgggaggcgtttgggcccagagaagtggatccgccgcttgcgccgcatggagtccgaatcgg 70 MDSGRRLGPEKWIRRLRRMESES 23 aaagcggggctgctgctgacacccccccactggagaccctaagcttccatggtgatgaagagattatcga 140 ESGAAADTPPLETLSFHGDEEIIE 47 ggtggtagaacttgatcccggtccgccggacccagatgacctggcccaggagatggaagatgtggacttt 210 VVELDPGPPDPDDLAQEMEDVDF 70 gaggaagaagaggaggaagagggcaacgaagagggctgggttctagaaccccaggaaggggtggtcggca 280 EEEEEEEGNEEGWVLEPQEGVVG 93 gcatggagggccccgacgatagcgaggtcacctttgcattgcactcagcatctgtgttttgtgtgagcct350 SMEGPDDSEVTFALHSASVF V S L 117 ggaccccaagaccaataccttggcagtgaccgggggtgaagatgacaaagccttcgtatggcggctcagc 420 DPKTNTLAVTGGEDDKAFV[WR]LS 140 gatggggagctgctctttgagtgtgcaggccataaagactctgtgacttgtgctggtttcagccatgact 490 D G E L L F E AGHKDSVTCAGFSHD 163 ccactctagtggccacaggggacatgagtggcctcttgaaagtgtggcaggtggacactaaggaggaggt 560 STLVATGDMSGLLKV[WQ]VDTKEEV 187 ctggtcctttgaagcgggagacctggagtggatggagtggcatcctcgggcacctgtcctgttggcgggc 630 WSFEAGDLEWMEWHPRAPVLLAG 210 acagctgacggcaacacctggatgtggaaagtcccgaatggtgactgcaagaccttccagggtcccaact 700 TADGNTWM[WK]VPNGDCKTFQGPN 233 gcccagccacctgtggccgagtcctccctgatgggaagagagctgtggtaggctatgaagatgggaccat 770 PATCGRVLPDGKRAVVGYEDGTI 257 caggatttgggacctgaagcagggaagccctatccatgtactgaaagggactgagggtcaccagggccca 840 RI[WD]LKQGSPIHVLKGTEGHQGP 280 ctcacctgtgttgctgccaaccaggatggcagcttgatcctaactggctctgtggactgccaggccaagc 910 L T VAANQDGSLILTGSVDCQAK 303 tggtcagtgccaccaccggcaaggtggtgggtgtttttagacctgagactgtggcctcccagcccagcct 980 LVSATTGKVVGVFRPETVASQPSL 327 gggagaaggggaggagagtgagtccaactcggtggagtccttgggcttctgcagtgtgatgcccctggca 1050 GEGEESESNSVESLGFCSVMPLA 350 gctgttggctacctggatgggaccttggccatctatgacctggctacgcagactcttaggcatcagtgtc 1120 AVGYLDGTLAI[YD]LATQTLRHQC 373 agcaccagtcgggcatcgtgcagctgctgtgggaggcaggcactgccgtggtatatacctgcagcctgga 1190 QHQSGIVQLLWEAGTAVVYTCSLD 397 tggcatcgtgcgcctctgggacgcccggaccggccgcctgcttactgactaccggggccacacggctgag 1260 GIVRL[WD]ARTGRLLTDYRGHTAE 420 atcctggactttgccctcagcaaagatgcctccctggtggtgaccacgtcaggagaccacaaagcgaaagILDFALSKDASLVVTTSGDHKAK" 1330443

tattttgtgtccaaaggcctgaccgttaatggctgcagcccctgcctgtgtgtctggtgttgaggggacg 1400 VFCVQRPDR- 452

aagggacccctgcccctgtctgccagcagaggcagtagggcacagagggaagaggagggtggggccctgg 1470 atgactttccagcctcttcaactgacttgctcccctctccttttcttctctttagagacccagcccaggg 1540 ccctcccacccttgcccagacctggtgggcccttcagagggaggggtggacctgtttctctttcactttc 1610 atttgctggtgtgagccatggggtgtgtatttgtatgtggggagtaggtgtttgaggttcccgttctttc 1680 ccttcccaagtctctgggggtggaaaggaggaagagatactagttaaagattttaaaaatgta(aataaa 1749 a)tatacttcccagaaaaaaaaaaaaaaaaaaaaaaaaa 1787 Fig. l. The nucleotide sequence of AAMP cDNA with ils predicted amino acid sequence. Base pairs 22-1787 are from a clone isolated from a Agt11 expression library of human A2058 melanoma cells. The first 21 bp are from PCR-generated amino terminal products obtained from human brain tissue. Nucleotide residues are numbered beginning at the 5' end. Amino acid sequence numbering begins with the initiating methionine. The putative heparin binding site, aa 14-18, is underlined with " ". The acidic region, aa 42-102, is underlined with " ". Cysteine pairs, 114 and 148 and 234 and 283. predicted by immunoglobulin domain homology to most likely form disulfide bonds, are marked "< >". The carboxy terminal boundaries of the WD40 repeats are indicated by brackets "[ j" around the last pair of amino acid residues in each repeat. The potential transmembrane region, aa 341-362, is underlined, " ". The polyadenylation site at nucleic acid residues, 1744-1750, is in parentheses.

2142

^ â€” AAMP- AAMP-

Fig. 2. Immunoblots of cultured human cells and tissues probed with polyclonal AAMP-specific antibodies. The immunoreactive Mr 52.000 protein is the putative AAMP protein (see text). A, Lanes I and 2 are nonreduced NP40 whole cell lysates of human A2058 melanoma cells (117 fig/lane) and bovine aortic endothelial cells (41 ^g/lane), respectively, reacted with anti-rAAMP (2 ng/ml). Lanes 4-11 and 13 are human tissue SDS lysates (reduced) from brain, skin, kidney, melanoma, lung, lymph node, liver, heart, and brain, respectively, also reacted with anti-rAAMP (1 /Â¿g/ml).Lane 15 shows the immunoreactivity of AAMP in reduced brain SDS lysate with anti-P189 (1 ng/ml). Lanes 4-11, 13, and /5 were loaded with 40, 26.5, and 26.5 /xg protein, respectively. Anti-rAAMP was generated with affinity-purified rAAMP (see text). Anti-P189 was generated lo a synthetic pcptide derived from the sequence of AAMP. Both antibodies were produced in rabbits and were affinity purified. Lanes 3, ÃŒ2,andÃŒ4arcprotein standards with apparent molecular weights of 215,5(X), 105, KM),69,800, 43,300, and 28,300 in vertical descending order. The molecular weight of AAMP is unaffected by the presence or absence of reducing agents. B. Lanes I and 6 show competition of antibodies' binding to AAMP for anti-P189 (0.5 jJ-gml) and anti-rAAMP (2 ng/ml). respectively. Competition was 5(KK)-fold for anti-P189 and 2100-fold for anti-rAAMP on a molar basis using P189 and rAAMP, respectively. Lanes 3 and 5 were reacted with antibodies only. Brain lysate (40 /ig each) was loaded in Lanes 1 and .?, and A2058 melanoma cell lysate (117 /Â¿geach) was loaded in Lanes 5 and 6. Lanes 2 and 4 contain protein standards (M, 215,(KX), 100.000, 69,(KK),43.0IX). 28,000 and 206,000. 105.(XK),71,000, 44,000, and 28,000, respectively). Gels were 10% polyacrylamide with SDS present.

domain secondary structure (30, 31, 41). Potential AAMP immunoglobulin domains show significant sequence homologies [ALIGN scores > 3 SD (29)] when matched with multiple domains of known family members (Fig. 5). A large portion of the most homologous immunoglobulin-type domain in AAMP is shown aligned with cor responding regions in three domains of deleted in colon carcinoma (DCC) protein (Fig. 6A; Ref. 42). Searches of the databases in the Protein Identification Resource (National BiomÃ©dicalResearch Foundation) using FASTA (43) indi cate that the sequence of AAMP is unique. It has a maximum identity of 25% with a Chlamydomonas G protein ÃŸsubunit-like polypeptide (44), using ALIGN (29) for entire sequence matching. An amino acid motif (approximately 40 residues in length) found as contiguous repeats in transducin homologues is designated as the WD40 repeat motif (frequently ends in tryptophan and aspartic acid) (45-48), and is repeated six times in AAMP with varying degrees of homology (Fig. 6ÃŸ).The first four and the last two are contiguous repeats. A yeast protein, YCR072c (GenPept X59720_170, Saccharomyces cerevisiae, translated from locus SCCHRIII, 239528-241075) (49),

â€” 4.4 Kb 1 23456

-A â€” 2.4 Kb ,â€ž* - * Â» â€” 1.77 Kb â€” 1.52 Kb â€” 1.40 Kb -B

â€” 0.78 Kb Fig. 4. Northern blot demonstrating the widespread tissue distribution of AAMP's Fig. 3. Northern blot of human melanoma A2058 cells probed with AAMP cDNA. mRNA. All RNA was from normal human tissues. Lane I: brain, adult. IÂ¿me2: brain, Lanes I and 2 contain total (41 jig) and polyadenylate-enriched A2058 RNA(2.2 /ig), fetal. Lane 3: liver, adult. Lane 4: liver, fetal. Lane 5: placenta!. Lane 6: breast. Two jig respectively. The 1.6-kb message of AAMP is seen in both lanes. The origin is indicated of polyadenylate-enriched RNA were loaded in each lane. A. AAMP, 1.6-kb message. B, by dotted Unes. Positions of the RNA standards are shown. Glyceratdehyde-3-phosphate dehydrogen ase, standard. 2143

""283 shares a similar arrangement of WD40 repeats, contains a domain AAMP 114 148 234 similar to the immunoglobulin-type domains of AAMP (ALIGN SD scores of 6.58 and 6.79 SD), and its entire sequence shows overall similarity with AAMP (ALIGN score of 10.68 SD; 97 runs). NCAM (I) 2.77 AAMP has one potential transmembrane region, aa 341-362, pre dicted by the method of Rao and Argos (50). A large, negatively NCAM(V) 0.66 charged region, aa 42-102, including a run of seven glutamic acid residues, is present near the amino terminal end (Fig. 1). The isoelec- DCC (II) tric point (calculated from sequence data) is 4.14. AAMP has an DCC (IV) instability index of 35.9 and is thus predicted to be a stable protein (index < 40; Ref. 51). NgCAM (III) Immunohistochemistry. In all human tissues tested (brain, pla centa, periadrenal connective tissue, colon, and uterus), anti-rAAMP NgCAM (IV) 1.92 4.69 and anti-P189 strongly stain endothelial cells in all vessel types seen, compared to controls. Vessels in brain, uterine wall, and colon are NgCAM (V) 2.19 3.54 shown (Fig. 7, A, B, D, G, and H). Anti-rAAMP also strongly stains small groups of poorly differentiated colon adenocarcinoma cells MAG (II) 2.08 3.16 found in lymphatics near the primary tumor and subpopulations of mononuclear inflammatory cells (Fig. 7, G and H). Cytotrophoblasts CD2 (II) 1.78 3.85 stain strongly and syncytiotrophoblasts less strongly (Fig. 7, J and Fig. 5. ALIGN scores for immunoglobulin superfamily members with the immuno- K). Surface staining of glandular cells in the uterine endometrium globulin-like domains of AAMP. Immunoglobulin-like domains (in parentheses) of the is seen with both anti-rAAMP and anti-P189 (Fig. 7, M and N). immunoglobulin superfamily members listed were compared with potential immunoglob- Some central nervous system neurons and their processes show ulin-like AAMP domains (specified above the columns), and the resulting scores in standard deviations (SD) are listed. Scores greater than 3.00 SD (outlined) indicate moderate staining. significant evolutionary immunoglobulin superfamily relationships (30, 31). Immuno Heparin Binding. Peptide 189, derived from the amino terminal globulin superfamily members: NCAM, (58), deleted in colon carcinoma (DCC) protein (42), neural-glial cell adhesion molecule (NgCAM', Ref. 59), myelin-associated glycopro- region of AAMP, contains the sequence RRLRRM, which resembles a consensus sequence for heparan sulfate binding, -XBBXBX- (B's tein (MAC; Ref. 60), and CD2 (61).

AAMP aa#231-299 G P N LP D G[Kj - [RjA[y]v DG DCC (IV) " 349-416 EFE VP TVNWM-KN DV- Y- DCC (I) " 58-133 L L D VP v i [K]W[K]- K D l H L AL R|K - - Q Q DCC (III) " 259-327 V L E YPPP S F T W - - (TlR E[TI- R|S K KpTIs

AAMP cont. -H VL K G|T E|- - G H Q DCC (IV) -R l L GVVK--SDE DCC (l) QN l L HSRHHKPDE N vfT~Dl- - D D S S L G [D] S G Sll l S DCC (III) -L l FS T Y K N E N| l |s A|S

V T l S

B AAMP aa# 96-138 EGPDDSEV T A S F c[V]s L DK T N T IT A VT G G E Ã– D K A F VWV 139-180 LSDGELLFEC A G HK D S V T C A G F S H D[Ã•Ã•]T L V A - T G|D M S G L ITI K WM 181-220 - -VDTKEEVWSFEAGDljjE W[Mj E W H P R A P V|L L WRQKl 221-261 V P N G D - C K T[?]Q[G]P NCPATCGRv[T P|D[G]KR A|V WDI 322-363 ASQPSLGEGEESESNSfv]ES[L|GFC -|_S VMP YDL 364-404 -LATQTLRHQC oFTTla S G l V Q L L W E A G T A- WD

WD40 motif l G H 1LVM1LVM1LV GA DY L SAG NF V (PMGSS N)1LVM1LVMST(SG M T)DN1LVMC1LVMW F Fig. 6. A. homology of the immunoglobulin-like domain of AAMP (defined by cysteines 234 and 283) with the immunoglobulin-like domains (I. III. and IV) of deleted in colon carcinoma (DCC; Ref. 42). Identical and/or similar amino acid residues are boxed together. Similar amino acids are grouped as follows: S.T; R,K: Q,N; E,D; F,W,Y; and L,M,I,V. Note that seven identities and four additional similarities are conserved in all domains. Gaps introduced to optimize alignments are shown as dashes. The significant ALIGN score listed in Fig. 5 [the domain of AAMP compared with DCC (IV) j indicates that there is an evolutionary relationship. B. the WD40 repeat motif found in AAMP is shown. This motif, indicated at the bottom, is found in ÃŸtransducin and other proteins (45-49, 54). Residues in parentheses are interchangeable. The identities of AAMP with the motif are boxed. The first four repeats are contiguous. The fifth and sixth repeats are also contiguous. Gaps are as described above. The functional significance of the motif is not known, but there is speculation that it may be involved in protein-protein recognition (54). 2144

D-

.^ i s â€¢.>Â£h'^ â€¢-\ .â€¢â€¢' VW> >/Ã¨ 0*'?T"'Â»L '>mi\'Â¿*Â¿. 1 ,Â»..-.-'<â€¢V

â€¢â€¢*Â»i i -. : . i* _ .Â«. *^

M N >^mfiP. v..-

/ ^

Fig. 7. Immunoperoxidase staining of human tissues with anti-recombinant AAMP and anti-PI89. Anli-rAAMP (1 /xg/ml) stains endothelial cells (veins and capillaries) as seen in sectionsof brain (A and ÃŸ).Italso stains endothelial cells in the uterine wall (D) and in the colon (G and H). In the colon, venules are shown (with red cells), and lymphalics (containing tumor and mononuclear inflammatory cells and lined with discontinuous endothelial cells) are shown in (G) and (H). Arterioles are seen in the upper portion of (G). Endothelial cells stained with anti-von Willebrand Factor (positive control) are seen in (Â£).The small groups of poorly differentiated colon adenocarcinoma tumor cells present in the lymphalics stain strongly positive with anti-rAAMP (1 fig/ml). Cytotrophoblasts stain strongly, and syncytiocytotrophoblasts stain moderately in sections of plÃ¡cenla(J and K). The concentration of anti-rAAMP is 0.25 /xg/ml in (J) and 1 jitg/ml in (K). Anti-rAAMP (1 /xg/ml) and anti-PI89 (1 ^ig/ml) stain endometrial glandular cell surfaces in (M) and (N), respectively. Staining with rabbit IgG (1 /ig/ml) used as a negative control is seen in (C), (/â€¢"),(/),(L), and (O). Paraffin-embedded tissues were prepared routinely, deparaffinized, immersed in buffer, and heated in a microwave oven for "antigen retrieval" (38) prior to staining. Anti-rAAMP and anti-P189 were generated as polyclonal antibodies in rabbits and were affinity purified using pure antigenic recombinant protein and peptide. Goal anti-rabbit secondary antibody conjugated with horseradish peroxidase was reacted with diaminobenzidinc and peroxidase. Sections were counterstained with hematoxylin. The bars shown on (C), (F), (/). (/.), and (O) indicale Ihe magnifications (100 (im) for their rows. 2145

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1995 American Association for Cancer Research. AAMP, A NEW VASCULAR PROTEIN are basic and X's are hydropathic residues; Ref. 52). Solubilized P189, attached to CovaLink plates (0.3 /ng/well) binds heparin strongly with a dissociation constant, Kd, of 14 pmol (Fig. 8A) according to Scatchard analysis (53) and is saturable (Fig. 8fl). A peptide variant of P189 that lacked the heparin-binding consensus sequence, P357 (sequence in "Materials and Methods"), showed no consistent heparin binding. This was also seen with plates prepared with no peptide. Heparin binding for P189 was at background levels (no peptide present on the plates) when competing unlabeled heparin (1-10X) was added. The presence of a heparin-binding epitope in recombinant fusion AAMP protein was confirmed qualitatively by comparing its binding of tritiated heparin to that of a control protein, BSA. Recombinant fusion AAMP protein (approximately 15 fig), gel purified away from bacterial proteins, and transblotted to Immo- bilon-P bound tritiated heparin (10.6 /xCi/ml) 2.2 Â±0.4 times more (six assays) and was competed (77% reduced) with an equal amount of unlabeled heparin (three assays). Cell Binding by P189 and Its Variants. P189 and its variants that contain the RRXRRX sequence also bind A2058 melanoma cells (CovaLink assays). Cell binding to P189-coated plates was pro nounced and heparin sensitive (Fig. 9). In contrast, plates prepared with P350, the scrambled version of P189 with dispersed basic resi dues, showed only slight cell binding with no competition by heparin at most concentrations (Fig. 9A). The cell binding of P189 was 2.95 times that of P350 with no heparin present. The morphology of cells in culture bound to P189 appears similar to that of the few cells that bound to P350 (Fig. 10). Peptides were solubilized and immobilized on linker arms 2 nm above the plate surfaces. Some cells were able to spread slightly with both peptides, and this is shown for P189 (Fig. IOC). The pictures in Fig. 10 are higher magnification photographs of the bottom wells (no heparin present) in Fig. 9/4. In additional assays, B = 1400

0.64 0.79 0.95 Bound (ng) B 0.95T 1 10 100 Heparin (units/ml)

0.79- Fig. 9. The inhibition by heparin of cell binding by P189. Increasing concentrations of heparin were added in wash solutions to CovaLink wells coated with P189 and P350 O (scrambled version of P189). These peptides had been previously solubilized and attached to the plates. Unbound heparin was removed prior to addition of A2058 melanoma cell m suspensions. Following a 1-h 37Â°Cincubation, unbound cells were removed, and the cells remaining on the plates were stained. A, the stained cells bound by P189 are seen in the 0.64 2.50 3.55 4.60 first two lanes labeled J and 7. Those bound by P350 are in Lanes 6 and 8. The bottom wells were never exposed to heparin, and the others in ascending order were exposed to Log Free (ng) 0.5, 1, 5, 10, 25, 50, and 100 units of heparin in a wash prior to the addition of cells. B, densities of stained cells bound by P189 (O) and P350 (â€¢)(computer scanned) Fig. 8. Heparin binding mediated by the derived peptide of AAMP, P189. Solubilized were quantitated by image analysis. Each densilometry unit represents approximately P189 (0.3 (ig/well) was attached to CovaLink plates and incubated with tritiated heparin. 84 cells. 4 fig/well, alone, and with fresh unlabeled heparin (0.4-40 /ig/well) of the same molecular weight. Following incubation and washing, the wells were counted to determine their Iritiated heparin contenls. cpm from multiple wells (3-5) for each amount of competing unlabeled heparin were averaged. Calculations using amounts of tritiated heparin displaced provided the amounts of unlabeled heparin bound/well as the concen peptide variants containing the RRXRRX motif, P189, P358, P359, tration of unlabeled heparin increased. A, a Scatchard plot was generated (54) showing and P360, showed comparable cell binding (no heparin), with results strong heparin binding with a dissociation constant, Kj, of 14 pmol. B, the heparin binding of P189 is saturable as shown. ranging from 66,000 Â±6,000 to 72,000 Â±11,000 cells/well. Peptides 2146

teines 234 and 283 identify AAMP as an immunoglobulin superfamily member. The immunoglobulin superfamily relatives (Fig. 5) include proteins that either are known or suspected to play a role in cell adhesion (NCAM, NgCAM, MAG, CD2, and DCC; Refs. 3, 4, and 42). Most are in the immunoglobulin superfamily C2 subset. The eleven identities/similarities that three domains in DCC share with one domain in AAMP may constitute a previously unrecognized motif (Fig. 6/1). The AAMP domain defined by cysteines 114 and 148 is possibly a second immunoglobulin domain. Other potential domains in AAMP defined by the cysteine pairs, 157 and 226 and 300 and 344, show weaker immunoglobulin domain homology. The WD40 repeat motif found in ÃŸtransducin and other proteins is speculated to represent a general protein-protein recognition/binding site (54). Several of these, such as ÃŸtransducin, PRP4, and Tupi, are parts of large protein complexes (47, 48). This motif is present as six repeats or half repeats in AAMP. Many of the variations in the GH and WD amino acid positions seen in some of the repeats of AAMP can also be found in those of known members (48). The first four WD40 repeats in AAMP are contiguous and involve a large portion of the region containing its immunoglobulin-type domains. The fifth and sixth repeats are in the carboxy terminal region of AAMP. The fifth repeat involves the potential transmembrane region. Of all the repeats present in AAMP, the sixth one shows the strongest homology with the WD40 motif. It is notable that AAMP contains a region that shares homology with both immunoglobulin-type domains and WD40 re peats. These motifs are both thought to contain ÃŸstrands (30, 31, 46). The potential heparin-binding domain in the amino terminal region of AAMP and the large, negatively charged region (61 amino acids in length) that follows it, precede the region of immunoglobulin-type domains and WD40 repeats. The tissue and cellular distributions of AAMP seen on RNA blots (1.6-kb message; Fig. 4) and on immunoblots (Mr 52,000 band iden tified by both anti-recombinant AAMP and anti-AAMP derived pep tide; Fig. 2) were compared. Correspondence of the widespread dis tribution seen only for the M, 52,000 protein with the broad mRNA distribution also supports its identity as the AAMP protein. The A/r 52,000 band is also the only prominent band that reacts with both anti-rAAMP and anti-Pi89 on immunoblots (Fig. 2). Also, it is almost the same size as predicted by the sequence of its cDNA clone (Mr 49,000). Although the broad distribution of AAMP in tissues could be attributed to its presence in endothelial cells, it also appears in other cell types, including many with migratory potential in human tissues (A2058 melanoma cells from a brain metastasis, metastatic melanoma tissue, and activated T cells on immunoblots, and in the following cells with immunohistochemistry: cytotrophoblasts, invasive poorly differentiated colon adenocarcinoma cells in lymphatics, mononuclear inflammatory cells, and some central nervous system neurons). RNA Fig. 10. MagnifÃ¬edviews of the wells with no heparin from Fig. 9A are shown to isolated from several types of human cells in tissue culture (melano illustrate cell morphology with exposure to immobilized P189 and its scrambled version. ma, breast cancer, lymphoma, monocytes, and T cells) contained P350. A. peptide 350. B. peptide 189. C, closer view of P18Â°-exposed cells. P189 is oriented on linker arms (2 nm above the plate surfaces) so that its heparin-binding site is AAMP mRNA as shown on Northern blots. available. More cells bind to P189 compared to the control peptide. P350. but the The reactivity for AAMP in endothelium may provide clues in morphology of the bound cells does not appear to differ significantly for the two peptides. determining its function. Endothelial cells express several immuno globulin superfamily proteins such as PECAM, Cell-Cam 105, lacking this motif, P350, P357, and P369, exhibited significantly LFA-3, ICAM-1, ICAM-2, VCAM, and HT7 (4, 10, 55, 56). Endo lower cell binding (P < 0.001) as a group. thelial cells also interact with immunoglobulin-type proteins on mi gratory cells that adhere to the endothelium (lymphocytes, neutrophils, and tumor cells). In general, immunoglobulin-type proteins are DISCUSSION known to be part of the complex collection of binding proteins that AAMP is a newly identified protein found in normal and malignant mediate selective and nonselective adhesion and transendothelial mi human tissues that shares sequence homology with two types of gration of various nonmalignant and malignant circulating cells (6, proteins, immunoglobulin superfamily members and proteins that 55). Studies have shown that endothelial adhesion proteins vary in contain the WD40 repeat motif. The homologies of the immunoglob their expression (constitutive or induced), location (cell surface or ulin superfamily members with the AAMP domain defined by cys- intracellular granules that relocate to the surface when induced), type 2147

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1995 American Association for Cancer Research. AAMP. A NEW VASCULAR PROTEIN of binding (homophilic, heterophilic, and some use bridging mole 18. DeLisser, H. M., Yan, H. C., Newman, P. J., Muller, W. A.. Buck, C. A., and Albelda, S. M. Platelet/endothelial cell adhesion molecule-1 (CD3l)-mediated cellular aggre- cules), and dependency on calcium (3, 6, 7, 57). Diffuse cellular galion involves cell surface glycosaminoglycans. J. Biol. Chem., 268: 16037-16046, staining of some cells for AAMP (Fig. 7, J and K) does not rule out 1993. its possible participation in these processes. 19. Doherty, P., Ashton, S. V., Moore, S. E., and Walsh, F. S. Morphoregulatory activities of NCAM and N-cadherin can be accounted for by G protein-dependent aclivation of The domain in AAMP which has the strongest homology with L- and N-type neuronal Ca24 channels. Cell, 67: 21-33, 1991. Spitz, M. "Single-shot" intrasplenic immunization for the production of monoclonal known immunoglobulin domains contains a large region (aa 231-299) 20. that shares local homology with immunoglobulin domains in two antibodies. Methods Enzymol., 121: 33-41, 1986. 21. Hook, W. A., Berenstein, E. H., Zinsser, F. U., Fischler, C., and Siraganian, R. P. endothelial adhesive proteins, HT7 (ALIGN score, 3.43 SD) and Monoclonal antibodies to the leukocyte common antigen (CD45) inhibit IgE- PECAM (ALIGN score, 3.63 SD). The cellular distribution of AAMP mediated histamine release from human basophils. J. Immunol., 147: 2670-2676, is similar to that of PECAM (endothelial cells and migratory cells 1991. 22. Fox, P. C., Berenstein, E. H., and Siraganian, R. P. Enhancing the frequency of including monocytes, T cells, and tumor cells). PECAM mediates antigen-specific hybridomas. Eur. J. Immunol.. //: 431-434, 1981. cellular aggregation that is inhibited by heparin and chondroitan 23. Stracke, M. L., Guirguis, R., Liotta, L. A., and Schiffmann, E. Pertussis toxin inhibits stimulated molility independently of the adenylate cyclase pathway in human mela sulfate (18). Similarly, the derived peptide of AAMP, P189, also binds noma cells. Biochem. Biophys. Res. Commun., 146: 339-345, 1987. cells in a heparin-sensitive manner (Fig. 9). In addition to its normal 24. Stracke, M. L., Krutzsch, H. C., Unsworth, E. J., Ã‚restad, A., Cioce, V., Schiffmann. physiological functions, PECAM participates in intercellular adhesive E., and Liotta, L. A. Identification, purification, and partial sequence analysis of processes involved in tumor metastasis, such as tumor cell-tumor cell autotaxin. a novel motility-slimulation protein. J. Biol. Chem., 267: 2524-2529, and tumor cell-platelet-endothelial interactions (8). AAMP's sequence 1992. 25. Sanger, F.. Nicklen, S.. and Coulson, A. DNA sequencing with chain-terminating homologies, distribution among endothelial and migratory cells, and inhibitors. Proc. Nati. Acad. Sci. USA, 74: 5463-5467, 1977. 26. Apte, A. N., and Siebert. P. D. Anchor-ligated cDNA libraries: a technique for potential for heparin-sensitive adhesive functions indicate that it may generating a cDNA library for the immediate cloning of the 5' ends of mRNAs. also participate in adhesion/migration activities. Biotechniques, 75: 890-893, 1993. 27. Innis, M. A., Myambo, K. B., Gelfand, D. H., and Brow, M. A. D. DNA sequencing with Thermus aquaticus DNA polymerase and direct sequencing of polymerase chain reaction-amplified DNA. Proc. Nati. Acad. Sci. USA, 85: 9436-9440, 1988. ACKNOWLEDGMENTS 28. Lee, J. Laboratory methods: alternative dideoxy sequencing of double-stranded DNA by cyclic reactions using Taq polymerase. DNA Cell Biol., 10: 67-73, 1991. We thank Reuben P. Siraganian, Cynthia Fischler, Susan M. Mackem, 29. Dayhoff, M. O., Barker, W. C., and Hunt, L. T. Establishing homologies in protein Douglas E. Moul, Mark E. Sobel, and Edward A. Unsworth for advice and sequences. Methods Enzymol., 91: 524-545, 1983. technical assistance. William G. Stetler-Stevenson, Anna T. Levy, and Jean E. 30. Williams, A. F., and Barclay, A. N. The immunoglobulin superfamily-domains for cell surface recognition. Annu. Rev. Immunol., 6: 381-405, 1988. Maguire allowed us to reprobe previously prepared Northern blots (data not 31. Barclay, A. N., Birkeland, M. L., Brown, M. H., Beyers, A. D., Davis, S. J., Somoza, shown). Mariana Doval (Washington Hospital Center, Washington, DC), C., and Williams, A. F. Protein superfamilies and cell surface molecules. In: The Stephanie Barr (George Washington University, Washington, DC), and Jean Leukocyte Antigen Facts Books, pp. 38-62. London: Academic Press, 1993. Colandrea (St. Agnes Hospital, Baltimore, MD) provided tissue samples. 32. Krutzsch, H. C.. and Inman. J. K. N-isopropyliodoacetamide in the reduction and alkylation of proteins: use in microsequence analysis. Anal. Biochem., 209: 109-116, 1993. 33. Gough, N. M. Rapid and quantitative preparation of cytoplasmic RNA from small REFERENCES numbers of cells. Anal. Biochem., 173: 93-95, 1988. 1. Bourgois. A. Evidence for an ancestral immunoglobulin gene coding for half a 34. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J.. and Rutter, W. J. Isolation of domain. Immunochemistry, 12: 873-876, 1975. biologically active ribonucleic acid from sources enriched with ribonuclease. 2. McLachlan, A. D. Early evolution of the antibody domain. In: H. Peelers (ed.). Biochemistry. 18: 5294-5299, 1979. Proteins and Related Subjects: Protides of the Biological Fluids, pp. 29-32. New 35. Aviv, H., and Leder, P. Purification of biologically active globin messenger RNA by York: Pergamon Press, 1980. chromatography on oligothymidylic acid-cellulose. Proc. Nati. Acad. Sci. USA, 69: 3. Edelman, G. M., and Crossin, K. L. Cell adhesion molecules: implications for a 1408-1412. 1972. molecular histology. Annu. Rev. Biochem., 60: 155-190, 1991. 36. Chomcznski. P., and Sacchi, N. Single-step melhod of RNA isolation by acid 4. Pigoli. R., and Power, C. The immunoglobulin superfamily. In: The Adhesion guanidium thiocyanate-phenol-chloroform extraction. Anal. Biochem., /62: Molecule Facts Book. pp. 13-16, 52-53. San Diego: Academic Press, 1993. 156-159, 1987. 5. Edelman, G. M. A golden age for adhesion. Cell Adhesion Commun., /: 1-7, 1993. 37. Church, G. M., and Gilbert, W. Genomic sequencing. Proc. Nati. Acad. Sci. USA, 81: 6. Honn, K. V., and Tang, D. G. Adhesion molecules and tumor cell interaction with 1991-1995. 1984. endothelium and subcndothelial matrix. Cancer Metastasis Rev., //: 353-375. 1992. 38. Gown, A. M., de Wever, N., and Battifora, H. Microwave-based antigcnic unmask 7. Muller, W. A., Weigl, S. A., Deng, X., and Phillips. D. M. PECAM-1 is required for ing: a revolutionary new technique for routine immunohistochemistry. Appi. Immu- Iransendothelial migration of leukocytes. J. Exp. Med., 178: 449-460, 1993. nohistochem., /: 256-266, 1993. 8. Tang, D. G.. Chen. Y. Q.. Newman, P. J., Shi, L., Gao, X., Diglio, C. A., and Honn, 39. Sondergard-Andersen, J., Lauritzen, E., Lind, K.., and Holm, A. Covalently K. V. Identification of PECAM-1 in solid tumor cells and its potential involvement linked peptides for enzyme-linked immunosorbent assay. J. Immunol. Methods. in tumor cell adhesion to endothelium. J. Biol. Chem., 268: 22883-22894, 1993. 131: 99-104, 1990. 9. Vaporciyan, A. A., DeLisser, H. M.. Yan, H., Mendiguren, I. I., Thorn. S. R.. Jones, 40. Kozak, M. Compilation and analysis of sequences upstream from the translational M. L., Ward, P. A., and Albelda, S. M. Involvement of platelet-endothelial cell start site in eukaryotic mRNAs. Nucleic Acids Res., 12: 857-872, 1984. adhesion molecule-1 in neutrophil recruilment I'Mi'iVo. Science (Washington DC), 41. Chou, P. Y.. and Pasman, G. D. Predictions of the secondary structure of proteins 262: 1580-1582, 1993. from their amino acid sequence. Adv. Enzymol., 47: 45-148, 1978. 10. Buck, C. A. Immunoglobulin superfamily: structure, function and relationship to 42. Fearon, E. R.. Cho, K. R., Nigro, J. M., Kern, S. E., Simons, J. W., Ruppert, J. M., other receptor molecules. Semin. Cell Biol., 3: 179-188, 1992. Hamilton, S. R., Preisinger, A. C., Thomas, G., Kinzler, K. W., and Vogelstein, B. 11. Shevach, E. M. Intercellular interactions in the immune response. In: S.J. Oppenheim and Identification of a chromosome 18q gene that is altered in colorectal cancers. Science E. M. Shevach (eds.). Immunophysiology: The Role of Cells and Cytokines in Immunity (Washington DC), 247: 49-56, 1990. and Inflammation, pp. 104-128. New York: Oxford University Press, 1990. 43. Pearson, W. R., and Lipman, D. J. Improved tools for biological sequence compar 12. Cole, G. J., Loewy, A., and Glaser, L. Neuronal cell-cell adhesion depends on ison. Proc. Nati. Acad. Sci. USA, 85: 2444-2448, 1988. interactions of N-CAM with heparin-like molecules. Nature (Lond.), 320: 445-447, 44. Schloss, J. A. A Chlam\domtmas gene encodes a G protein ÃŸsubunit-like polypep- 1986. tide. Mol. Gen. Genet., 22/: 443-452, 1990. 13. Cole, G. J., and Akeson, R. Identification of a heparin binding domain of the neural 45. Fong, H. K. W., Amatruda, T. T., Ill, Birren, B. W., and Simon, M. I. Distinct forms cell adhesion molecule N-CAM using synthetic peplides. Neuron, 2: 1157-1165, of the ÃŸsubunit of GTP-binding regulatory proteins identified by molecular cloning. 1989. Proc. Nati. Acad. Sci. USA, 84: 3792-3796, 1987. 14. Reyes, A. A., Akeson, R., Brezina, L., and Cole, G. J. Structural requirements for 46. Duronio, R. J., Gordon, J. I., and Boguski, M. S. Comparative analysis of the ÃŸ neural cell adhesion molecule-heparin interaction. Cell Regul., /: 567-576, 1990. transducin family with identification of several new members including PWP1, a 15. Aruffo, A., Stamenkovic, I., Melnick, M., Underhill, C. B., and Seed, B. CD44 is the nonessential gene of Saccharomycea cerevisiae that is divergently transcribed from principal cell surface receptor for hyaluronate. Cell, 61: 1303-1313, 1990. NMT1. Proteins Struct. Funct. Genet., 13: 41-56, 1992. 16. Miyake, K., Underhill, C. B., Lesley, J., and Kincade, P. W. Hyaluronate can function 47. Neer, E. J., Schmidt, C. J., and Smith, T.: LIS is more. Nat. Genet., 5: 3-4, as a cell adhesion molecule and CD44 participates in hyaluronate recognition. J. Exp. 1993. Med., Â¡72:69-75, 1990. 48. van der Voorn. L., and Ploegh, H. L. The WD-40 repeat. FEBS Lett., 307: 131-134, 17. Lederman, S., Gulick, R., and Chess. L. Dexlran sulfate and heparin interact with 1992. CD4 molecules to inhibit the binding of coat protein (gpl20) of HIV. J. Immunol., 49. Slonimski, P. P., and Brouille!, S. A data-base of chromosome III of Saccharomyces 143: 1149-1154, 1989. cerevisiae. Yeast, 9: 941-1029, 1993. 2148

50. Rao, J. K. M., and Argos, P. A conformational preference parameter to predict helices 57. Palei, K. D., Zimmerman, G. A.. FrescoÂ».S. M., McEver, R. P.. and Mclntyre, T. M. in intÃ©gralmembrane proteins. Biochim. Biophys. Acta, 869: 197-214, 1986. Oxygen radicals induce human endothclial cells to express GMP-140 and hind 51. Guruprasad, K., Reddy, B. V. B., and Pandit, M. W. Correlation between stability of a neutrophils. J. Cell Biol.. 112: 749-759, 1991. protein and its dipeptide composition: a novel approach for predicting in vim stability of 53. Cunningham, B. A.. Hemperly, J. J., Murray, B. A., Prediger, E. A., Brackenbury, R., a protein from its primary sequence. Protein Engineering, 4: 155-161, 1990. and Edelman, G. M. Neural cell adhesion molecule: structure, immunoglobulin-likc 52. Cardin, A. D., and Weintraub, H. J. R. Molecular modeling of protein-glycosamin- domains, cell surface modulation, and alternative RNA splicing. Science (Washington oglycan interactions. Arteriosclerosis, 9; 21-32, 1989. T)C\ 236- 799 806 1987 53. Scatchard, G. The attractions of proteins for small molecules and ions. Ann. NY 5C BurÂ¿00 'M. p Grumeli 'M Mauro. v Eddman, G M., and Cunningharn, B. A. Acad. Sci., 51: 660-672, 1949. _*,.,. . . â€ž . .. â€ž.,, . . , 54. Dalrymple, M. A., Petersen-Bjorn, S., Friesen, J. D., and Beggs, J. D. The product of S.lruct"re Â°f'he chl=kc" "^""f'' Â«Â»â€¢*â€¢*"mokcule Ng-CAM: origin of the PRP4 gene of S. cererai shows homology to ÃŸsubunitfof G proteins Cell, 58: the polypept.deÂ« and rela.ion to the Ig superfam.ly. J. Cell B.ol., 112: 1017-1029, 811-812, 1989. "'â€¢ 55. Beynon, H. L. C, and Haskard, D. O. Lymphocyte-endothclial cell interactions. In. M- Salzer- J- L- Holmes. W. P., and Colman, D. R. The ammo acid sequences of the S. D. Brain (ed.), Immunopharmacology of the Microcirculation, pp. 113-114. San myelm-associated glycoprotems: homology to the .mmunoglobul.n gene superfamily. Diego: Academic Press, 1994. J- Cell Biol., 104: 957-965, 1987. 56. Seulberger, H., Lottspeich, F., and Risau, W. The inducible blood-brain barrier 61. Sewell, W. A., Brown, M. H., Dunne, J., Owen, M. J., and Crumpton, M. J. Molecular specific molecule HT7 is a novel immunoglobulin-like cell surface glycoprotein. cloning of the human T-lymphocyle surface CD2 (Til) antigen. Proc. Nati. Acad. EMBO J., 9: 2151-2158, 1990. Sci. USA, S3: 8718-8722, 1986.

2149

Downloaded from cancerres.aacrjournals.org on September 28, 2021. © 1995 American Association for Cancer Research. Identification of a New Immunoglobulin Superfamily Protein Expressed in Blood Vessels with a Heparin-binding Consensus Sequence

Marie E. Beckner, Henry C. Krutzsch, Mary L. Stracke, et al.

Cancer Res 1995;55:2140-2149.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/55/10/2140

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/55/10/2140. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.