cDNA cloning and expression of HIP, a novel cell surface heparan sulfate/heparin-binding protein of human uterine epithelial cells and cell lines.

Heparan sulfate proteoglycans and their corresponding binding sites have been suggested to play an important role during the initial attachment of murine blastocysts to uterine epithelium and human trophoblastic cell lines to uterine epithelial cell lines. Previous studies on RL95 cells, a human uterine epithelial cell line, had characterized a single class of cell surface heparin/heparan sulfate (HP/HS)-binding sites. Three major HP/HS-binding peptide fragments were isolated from cell surfaces by tryptic digestion, and partial amino-terminal amino acid sequence for each peptide fragment was obtained (Raboudi, N., Julian, J., Rohde, L. H., and Carson, D. D.(1992) J. Biol. Chem. 267, 11930-11939). In the current study, using approaches of reverse transcription-polymerase chain reaction and cDNA library screening, we have cloned and expressed a novel, cell surface HP/HS-binding protein, named HP/HS interacting protein (HIP), from RL95 cells. The full-length cDNA of HIP encodes a protein of 159 amino acids with a calculated molecular mass of 17,754 Da and pI of 11.75. Transfection of HIP full-length cDNA into NIH-3T3 cells demonstrated cell surface expression and a size similar to that of HIP expressed by human cells. Predicted amino acid sequence indicates that HIP lacks a membrane spanning region and has no consensus sites for glycosylation. Northern blot analysis detected a single transcript of 1.3 kilobases in both total RNA and poly(A) RNA. Examination of human cell lines and normal tissues using both Northern blot and Western blot analyses revealed that HIP is expressed at different levels in a variety of human cell lines and normal tissues but absent in some cell lines and some cell types of normal tissues examined. HIP has relatively high homology (80% both at the levels of nucleotide and protein sequence) to a rodent ribosomal protein L29. Thus, members of the L29 family may be displayed on cell surfaces where they may participate in HP/HS binding events.

Our laboratory has been interested in studying the mechanism of embryo implantation. We have found that HSPGs and their corresponding binding sites on cell surfaces may be important in the initial stage of mouse embryo attachment to uterine epithelium. Upon hatching from the zona pellucida, the embryo initially attaches to the uterus through the adhesion of the apical surfaces of the trophectodermal cells of the blastocyst. HSPGs are expressed by mouse embryos at the two-cell and post-implantation stages (Dziadek et al., 1985). Expression of HSPGs on trophectodermal cell surfaces of mouse blastocysts increases 4 -5-fold at the peri-implantation stage Farach et al., 1987), and HS expressed on mouse embryo surfaces is required for embryo attachment to isolated mouse uterine epithelial cells, fibronectin, and laminin (Farach et al., 1987). Similarly, studies have shown that the initial attachment of JAR cells, a human trophoblastic cell line, to RL95 cells, a human uterine epithelial cell line, is HP/HS-dependent, and enzymatic removal of HS from cell surfaces of both JAR and RL95 cells markedly inhibits JAR-RL95 cell-cell adhesion (Rohde and Carson, 1993). Therefore, HSPGs and their binding proteins also may play an important role in the initial attachment of human trophoblast cells to uterine epithelial cells.
Specific, saturable cell surface HP/HS-binding sites have been identified on mouse uterine epithelial cells and human uterine epithelial cell lines (Wilson et al., 1990;Raboudi et al., 1992). Since mouse uterine epithelial cells have only been available by primary cell culture, there is a practical limitation to isolating HP/HS-binding sites from this source. Effort has been placed on the study of HP/HS-binding proteins expressed on the cell surfaces of a human uterine epithelial cell line, RL95 (Raboudi et al., 1992). Mild tryptic digestion of RL95 cell sur-faces removed most of cell surface HP/HS-binding activity. Three major tryptic peptide fragments, ranging in M r between 6,000 and 14,000, were released from cell surfaces and retained HP/HS-binding activity. Partial amino-terminal amino acid sequences from each of these three peptides were obtained (Raboudi et al., 1992).
We have employed an approach of reverse transcriptionpolymerase chain reaction (RT-PCR) to identify transcripts encoding cell surface HP/HS-binding peptides. Predicted peptide sequence from one of the RT-PCR products revealed an antigenic sequence that also has features of a HP/HS-binding motif suggested by others (Cardin and Weintraub, 1989). Polyclonal antibodies directed against the synthetic peptide corresponding to this motif recognize a novel HP/HS-binding protein, named HP/HS interacting protein (HIP), expressed on RL95 cell surfaces with an apparent M r of 24,000 determined by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (Rohde et al., 1996). This peptide selectively binds HP/HS, recognizes certain forms of HP and cell surface HS expressed by JAR and RL95 cells, and supports the attachment of human trophoblast cell lines and a variety of other mammalian adherent cell lines. 2 Complete cDNA sequence of HIP has been isolated by screening cDNA libraries using the partial cDNA sequence of RT-PCR product of HIP. HIP cDNA sequence contains a single open reading frame encoding 159 amino acids with a calculated molecular mass of 17,754 Da and a predicted pI of 11.75. This protein is approximately 80% homologous at both the nucleotide and amino acid level to a rodent protein designated as ribosomal protein L29. Transfection of HIP into NIH-3T3 cells results in expression of an M r 24,000 protein that can be detected on the cell surface. Studies on the expression and distribution of HIP revealed that HIP is expressed at varying levels in a variety of human cell lines and normal human tissues.

EXPERIMENTAL PROCEDURES
Materials-Sodium chloride, sodium citrate, Tris base, glycine, bovine serum albumin (BSA), phenylmethylsulfonyl fluoride were purchased from Sigma. Sodium dodecyl sulfate (SDS), ␤-mercaptoethanol, and Tween 20 were purchased from Bio-Rad. Trichloroacetic acid, acetone, paraformaldehyde, calcium chloride were purchased from Fisher. Formamide and restriction enzymes were purchased from Boehringer Mannheim. All chemicals used were reagent grade or better.
Total RNA from RL95 cells was isolated using the method of RNA isolation of Xie and Rothblum (1991). RT-PCR was performed on a DNA Thermal Cycler (Perkin-Elmer) using the protocol described in the GeneAmp RNA PCR kit (Perkin-Elmer) with the concentration of MgCl 2 adjusted to 1 mM. The thermal cycle profile described in the 3Ј-rapid amplification of cDNA ends protocol (Frohman, 1990) was followed. The cDNA pools from RT were amplified by PCR. RT-PCR products were cloned into the pCR II vector using the TA Cloning System kit (Invitrogen, San Diego, CA). Plasmids containing RT-PCR products were isolated either following the method described by Sambrook et al. (1989) or using "Magic Minipreps" DNA Purification System kit (Promega, Madison, WI). All of RT-PCR products were sequenced using either the dideoxy-mediated chain termination method (Sambrook et al., 1989) following the procedure provided in the Sequenase version 2 kit (Amersham Corp.) or automatic sequencing using fluorescently labeled sequencing primers (T7 and SP6) provided in the Applied Biosystems cycle sequencing kit and analyzed on an Applied Biosystem model 373A automated sequencer (Perkin-Elmer).
Northern Blot Analysis-Northern blot analysis of RNA was performed using method described by Sambrook et al. (1989). Total RNA was extracted (Xie and Rothblum, 1991), and poly(A ϩ ) RNA was isolated using Oligotex-dT mRNA kit (Qiagen Inc., Chatsworth, CA). Both total RNA and poly(A ϩ ) RNA were separated on 1% (w/v) agarose gel and transferred to a nylon membrane (Hybond-N, Amersham Corp.). A cDNA probe (Clone 23-1 in Fig. 1A depleted of poly(A ϩ ) tail) was labeled with 32 P by utilizing the random oligonucleotide primer method (Sambrook et al., 1989). Hybridization was performed at 42°C overnight in a solution containing 50% (v/v) formamide, 5 ϫ SSC, 5 ϫ Denhardt's, 50 mM NaH 2 PO 4 , 0.1% (w/v) SDS, and 100 g/ml denatured, sonicated salmon sperm DNA. The blot then was washed using 2 ϫ SSC, 0.1% (w/v) SDS with several changes during 1 h at 25°C and washed once again using 0.5 ϫ SSC and 0.1% (w/v) SDS for 2 h at 42°C before exposure to Kodak XAR film with an intensifying screen at Ϫ70°C.
cDNA Library Screening-HeLa cell cDNA libraries constructed in the Lambda gt11 vector from Stratagene (La Jolla, CA) and Clontech (Pala Alto, CA) were used for cDNA library screening following the protocol provided by the manufacturer. Briefly, nitrocellulose filters (Schleicher & Schuell) lifted off the plated cDNA library were prehybridized in 0.8 M NaCl, 20 mM Pipes, pH 6.5, 50% formamide, 0.5% (w/v) SDS, and 100 g/ml denatured, sonicated salmon sperm DNA for Ͼ4 h at 42°C and hybridized in a fresh solution of the same composition containing the 32 P-labeled probe (2-4 ϫ 10 6 cpm/ml) for library screening, at 42°C overnight. After hybridization, the blots were briefly washed with 0.1 ϫ SSC, 0.1% (w/v) SDS at room temperature and then washed with same mixture at 65°C for 1 h. The blots were exposed to Kodak XAR films with an intensifying screen at -70°C overnight. Positive plaques were isolated and subjected to subsequent screenings under the same conditions. Southern Blot Analysis and DNA Sequencing-Positive phage DNA from cDNA library screening was purified from the phage lysate (Sambrook et al., 1989) and digested with EcoRI. The cDNA inserts were separated on 1% (w/v) agarose gel, and transferred to a nylon membrane by standard method (Sambrook et al., 1989). Blots were hybridized with 32 P-labeled probe in 6 ϫ SSC, 0.5% (w/v) SDS, 50% (v/v) formamide, and 100 g/ml denatured, sonicated salmon sperm DNA overnight. The blots were washed consecutively with 2 ϫ SSC and 0.5% (w/v) SDS for 5 min at room temperature, 2 ϫ SSC and 0.1% (w/v) SDS for 15 min at room temperature, and 0.1 ϫ SSC and 0.5% (w/v) SDS for 1 h at 37°C. Then the blots were washed with 0.1 ϫ SSC and 0.5% (w/v) SDS for 1 h at 68°C prior to the exposure to Kodak film with an intensifying screen at -70°C. Phage with positive cDNA inserts was digested with EcoRI, separated on a 1% (w/v) agarose gel, purified by phenol/chloroform extraction and ethanol precipitation, and then subcloned into the EcoRI site of pBluescript II SK-(Stratagene). Subcloned inserts were further analyzed using Southern blot analysis as described above. Clones with positive inserts were identified and both strands of the cDNA sequences were determined.
Construction of cDNA Expression Vector and Transfection-The entire HIP cDNA (clone 36 -1 in Fig. 1A) was digested from HIP cDNA-containing pBluescript using NotI, separated on 1% (w/v) agarose gel, purified by phenol/chloroform extraction and ethanol precipitation, and subcloned into the NotI site of mammalian expression vector pOPRSVI (Lac Switch Inducible Mammalian Expression System, Stratagene; La Jolla, CA). The clone that contained the entire cDNA in the correct orientation was determined by PCR using oligonucleotide primers derived from both vector and HIP cDNA sequences and selected for transfection. NIH-3T3 cells (30 -40% confluence in 100-mm cell culture plates) were transfected with 18 g/plate HIP expression vector DNA using the calcium phosphate method (Sambrook et al., 1989) and grown for 48 -72 h. The transfected cells were then harvested and used for Western blot and immunocytochemical analyses. 2 S. Liu and D. D. Carson, submitted for publication.

SDS-Polyacrylamide Gel Electrophoresis and Western
Blotting-Cells on cell culture plates were washed three times with PBS and solubilized in sample extraction buffer: 4 M urea, 1% (w/v) SDS, 50 mM Tris, pH 7.0, 1% (v/v) ␤-mercaptoethanol, and 0.01% (v/v) phenylmethylsulfonyl fluoride. Solubilized samples were concentrated by precipitation with 10% (w/v) trichloroacetic acid at 4°C. Trichloroacetic acid precipitates were centrifuged at 1,200 ϫ g for 10 min at 4°C, washed sequentially with 10% (w/v) trichloroacetic acid and 100% acetone, and air-dried. The pellets were dissolved in equal volumes of sample extraction buffer and sample buffer (Laemmli, 1970), heated for 2 min at 90°C, and subjected to SDS-PAGE. Protein samples were resolved by SDS-PAGE on a 15% (w/v) acrylamide resolving gel as described (Porzio and Pearson, 1977). After a brief rinse in transfer buffer (100 mM Tris base and 100 mM glycine, pH 9.2), the gel was transferred to a nitrocellulose membrane at 4°C for 5 h at 40 V in a Transblot apparatus (Bio-Rad). The transferred blot was blocked with 1% (w/v) BSA in PBS, 0.01% (w/v) sodium azide, and 0.05% (w/v) Tween 20 (P.A.T.) overnight at room temperature. The blot then was incubated with primary antibody diluted in 0.1% (w/v) BSA in P.A.T. for at least 4 -6 h at room temperature, rinsed with P.A.T. three times for 5 min each, and incubated for at least 4 -6 h with 6 Ci of 125 I-protein A (30 Ci/g) in 70 ml of 0.1% (w/v) BSA in P.A.T.
Immunocytochemical Analysis-Transiently transfected and parental NIH-3T3 cells were grown on coverslips for 24 h in Dulbecco's modified Eagle's medium/Ham's F12 supplemented with 10% (v/v) heatinactivated fetal bovine serum. The cells were briefly rinsed with PBS and fixed with 2.5% (w/v) paraformaldehyde in PBS for 15 min at room temperature. Cells were rinsed with PBS and soaked in PBS for 10 min. Aldehyde groups were blocked by incubation with 2 ml of 50 mM ammonium chloride in PBS for 15 min at room temperature. Immunostaining with primary and secondary antibodies and mounting of coverslips were performed as described before (Julian et al., 1994). Briefly, coverslips were then incubated with primary antibody in PBS for 1 h at 37°C. After three rinses of 5 min each with PBS, the cells were incubated with the second antibody, fluorescein-conjugated donkey antirabbit IgG (Amersham Corp.), for 40 min at 37°C. Unbound antibody was removed by several 5-min rinses at room temperature with PBS prior to mounting for fluorescence microscopy.
Computer Analysis-Nucleotide and protein sequence analyses were carried out using GCG and MicroGenie programs and the data bases from GeneBank (release 90.0), EMBL (release 43), and SWISS-PROT (release 31).

RESULTS
Characterization of RT-PCR Products-RT-PCR products obtained from each pair of primers were isolated, subcloned, and sequenced. Sequences of all of the RT-PCR products were determined. Among all of the RT-PCR products, one product (RT-PCR 224 in Fig. 1A) displayed several interesting features as follows. 1) There was a single open reading frame. 2) The amino-terminal peptide sequence used to design the primer was contained in the predicted peptide sequence. 3) A polyadenylation signal was found adjacent to the poly(A ϩ ) tail. 4) The predicted polypeptide sequence of the cDNA contained an antigenic peptide sequence, CRPKAKAKAKAKDQTK, with features associated with HP/HS-binding motifs (Cardin and Weintraub, 1989). Computer analysis predicted that this motif was ␣-helical and hydrophilic and likely to be exposed on the external surface of the protein where it may bind HP/HS. The sequence of this 270-bp RT-PCR product was compared with available sequences in GenBank. The highest similarities found by this analysis were 64% (in 265-bp overlap) similarity to Ratus norvegicus (rat) mRNA for ribosomal protein L29 and 67% (in 188-bp overlap) similarity to a rat mRNA for a ribosomal protein related to yeast ribosomal protein YL43. These features and homology analysis suggested that the parental mRNA corresponding to this RT-PCR product encodes a novel protein related to L29.
Isolation and Characterization of a cDNA Encoding HIP-A HeLa cell cDNA library was screened using the cDNA of HIP RT-PCR product (RT-PCR 224 depleted of poly(A ϩ ) tail in Fig.  1A) as a probe. As a result, 20 positive clones with identical inserts to that of 23-1 (or 42-1) (Fig. 1A) were obtained from approximately 5 ϫ 10 5 plaque-forming units. The cDNA sequence of inserts of 23-1 and 42-1 was determined by primer walking. The determined cDNA sequence contains an incomplete open reading frame encoding 117 amino acid residues. Re-screening of a 5Ј-stretched HeLa cDNA library (Clontech) using 23-1 cDNA as a probe resulted in the isolation of 12 positive clones from 3.0 ϫ 10 5 plaque-forming units. The inserts ranged between 400 and 650 bp. Restriction analysis revealed that they had the same restriction enzyme maps, indicating that they were likely to be derived from the same mRNA transcript. Sequencing of the two largest inserts, 35-2 and 36 -1, revealed identical sequences except that 1) 36 -1 contains 26 more bp of 5Ј-untranslated region than 35-2; 2) 35-2 contains an extra 6 bp of nucleotides (marked with brackets in Fig. 1B) at the 3Ј end of 3Ј-untranslated region; and 3) there is a nucleotide replacement (C replaced with T, marked with [ ] in Fig. 1B) at nucleotide position 145. This nucleotide replacement does not affect the coding of the amino acid, leucine. The nucleotide sequence of HIP cDNA including differences between different clones and predicted amino acid sequence of HIP are shown in Fig. 1B (accession number,  U49083).
Further analysis of the sequence showed that HIP cDNA contains a single open reading frame of 477 bp, starting with an ATG codon at position 28 with characteristic purines at positions -3 and ϩ4 relative to the start ATG codon (Kozak, 1987), and ending with a stop codon TAG at position 506 (* in Fig. 1B). This open reading frame encodes a protein of 159 amino acid residues with a calculated molecular mass of 17,754 Da. The HIP peptide sequence used for antibody production is indicated in the predicted protein sequence (shaded letters in Fig. 1B). Following the open reading frame, there are 121 bp of a 3Јuntranslated region that contains a polyadenylation signal (AATAAA) at nucleotide position 613. The predicted protein sequence has high content of positively charged amino acid residues (K ϩ R ϭ 29.6%) and a predicted pI of 11.75.
A comprehensive search of the GenBank, EMBL, and SWISS-PROT data bases revealed that the nucleotide sequence of HIP has 80.5% identity in 549-bp overlap to a rat mRNA for ribosomal protein related to yeast ribosomal protein YL43, 77.6% identity in 603-bp overlap to R. norvegicus (rat) mRNA for ribosomal protein L29, and 76.6% identity in 640-bp overlap to Mus musculus (murine) large ribosomal subunit protein mRNA. A BLAST homology search using GenBank revealed two human nucleotide sequences, designated as a putative human ribosomal protein L29, in GenBank (accession number U10248 and Z49148) showing the same nucleotide sequence as that of HIP cDNA; however, these sequences are not published and no further information is available. The predicted amino acid sequence of HIP has 80.3% identity in 157-amino acid residue overlap to a rat 60 S ribosomal protein, L29; however, the region encoding the peptide sequence (HIP peptide) used for antibody production and HP/HS-binding activity studies is not conserved among human and rat or mouse (Fig. 1C). Consequently, the antibodies are specific to human and do not cross-react with L29 of rat or mouse.
Northern Blot Analysis of RNA from RL95 Cells-The expression of HIP mRNA was examined by Northern blot analysis using a 32 P-labeled clone 23-1. A single predominant transcript of 1.3 kb was detected either using total RNA or poly(A ϩ ) RNA from RL95 cells (Fig. 2). Further studies using a variety of human cell lines (see below) indicated that the 1.3-kb transcript is the major HIP mRNA found in most cases.
Transfection of HIP cDNA into NIH-3T3 Cells and Expression of HIP-To demonstrate that the cloned cDNA sequence encodes the protein recognized by antibodies generated against the HIP peptide and to verify cell surface expression of HIP, clone 36 -1 was subcloned into a Rous sarcoma virus-based mammalian expression vector and used to transfect NIH-3T3 cells. Transiently transfected NIH-3T3 cells were fixed with paraformaldehyde, and expression of HIP was detected using anti-HIP. Fig. 3 shows the immunostaining of a representative field of paraformaldehyde-fixed intact NIH-3T3 cells transiently transfected with HIP cDNA. Staining demonstrated cell surface expression of transfected HIP protein in a portion of transfected cells, whereas other cells that presumably were not transfected during the transient assay were negative. Negative staining also was observed using paraformaldehyde-fixed parental NIH-3T3 cells (data not shown). Similar controls as that presented in the accompanying paper (Rohde et al.,1996;Fig. 7) using antibodies directed against cytokeratins were done for the cell surface staining of the transiently HIP-transfected NIH-3T3 cells, and no reactivity was observed under the fixing conditions performed (data not shown). Therefore, the immunostaining of the transfected cells was cell surface staining. Western blot analysis of total protein extracted from transiently transfected NIH-3T3 cells using anti-HIP-peptide de-tected a newly expressed protein with an apparent M r of 24,000, i.e. the same molecular weight as that of HIP detected in RL95 cells. In contrast, the M r 24,000 component was not detectable in the parental NIH-3T3 cells (Fig. 4). Collectively, these data demonstrate that the isolated cDNA sequence encodes the protein recognized by the anti-HIP peptide antibody and that this protein can be expressed on cell surfaces.
Expression and Distribution of HIP-Expression and distribution of HIP in different human cell lines and normal tissues were examined using both Northern blot analysis and Western blot analysis. Northern blot analysis using the HIP cDNA probe detected a single transcript of 1.3 kb in most human cell lines tested. HIP is expressed highly in most human epithelial cell lines including RL95, JAR, HeLa, HEC, and Ishikawa, as well as AFb-11, human fibroblastic cells, moderately in HU-VEC (endothelial cells), and relatively low in HL60, a human leukemic cell line. HIP mRNA was not detectable in MDA-231, a human breast cancer cell line, and NCI-H69, a human lung epithelial cell line, or mouse NIT-3T3 cells (Fig. 5). Western blot analyses on several human cell lines also revealed a similar distribution pattern of HIP as that shown in Northern blot Deduced amino acid sequence is indicated below the nucleotide sequence. Nucleotides are numbered from the beginning of the cDNA sequence, and the deduced amino acid sequence is numbered from the beginning of the open reading frame. An in-frame stop codon is indicated with *, and the consensus polyadenylation signal, AATAAA, is underlined. HIP peptide sequence used for antibody production and HP/HS-binding activity study is shaded. Differences in nucleotide sequence among different clones are indicated with [ ]. C, comparison of amino acid sequence of HIP with mouse and rat L29. The deduced amino acid sequences of HIP and mouse and rat L29 are indicated. The HIP peptide sequence used for antibody production and HP/HS-binding studies is underlined. Asterisks denote the identical amino acid residues. Deletions or insertions are indicated by dashes. analyses ( Fig. 6 and Table I). Again, consistent with Northern blot analysis, HIP was not detectable in cell lines of MDA-231 (Table I) and NCI-H69 (Fig. 6). Table I provides a summary of results showing differential HIP expression among a panel of human cell lines examined. Collectively, these results indicate that HIP mRNA and protein are expressed differentially in human cell lines.

DISCUSSION
In the present study, we have isolated and sequenced a full-length cDNA encoding HIP, a novel human HP/HS-binding protein expressed on cell surfaces. This cDNA encodes a protein of 159 amino acids with high content of basic amino acids. There is no a potential transmembrane domain present in the predicted amino acid sequence of HIP. HIP is associated with the cell surface (Rohde et al., 1996 and present article). Therefore, it is likely that HIP is a peripheral membrane protein, perhaps bound to other proteins, lipids, or polysaccharides. The predicted amino acid sequence of HIP does not contain a classical hydrophobic amino-terminal signal peptide (Blobel and Dobberstein, 1975); however, there are multiple reports of the lack of a signal peptide in the sequences of membrane or secreted proteins (Kikutani et al., 1986;Bettler et al., 1989;Brown et al., 1987). The predicted protein sequence of HIP predicts a molecular mass of 17,754 Da. This is significantly less than expected for M r 24,000 protein recognized by anti-HIP-peptide on SDS-PAGE. The anomalous molecular mass may be due to the highly basic character of HIP (predicted pI ϭ 11.75). Other highly basic proteins, e.g. histones, migrate rel-atively slowly on SDS-PAGE (von Holt et al., 1989;Weber and Osborn, 1975), apparently due to an inordinately high amount of SDS binding. Alternatively, post-translational modifications may increase the size of HIP. No consensus sites for glycosylation are evident in this sequence, but other modifications are possible. Transfection of full-length cDNA of HIP into NIH-3T3 cells resulted in the expression of a protein with M r of 24,000 determined by SDS-PAGE, further demonstrating that the cloned cDNA sequence encodes the same protein recognized by the antibody and contains the predicted peptide sequence. Northern blot and Western blot analyses revealed that both the 1.3-kb mRNA and M r 24,000 protein are expressed coordinately in a variety of human cell lines. Differential expression of HIP protein also has been observed in normal human tissues examined (Rohde et al., 1996). 3 FIG. 2. Northern analyses of RNA from RL95 cells. Approximately 3.5 g of poly(A ϩ ) RNA (lane 1) and 20 g of total RNA (lane 2) were isolated from RL95 cells and subjected to Northern blot analysis using 32 P-labeled cDNA of clone 23-1 (Fig. 1)  Sequence comparison with available data bases revealed that HIP has a relative high similarity (80%) to rodent L29, a ribosomal protein, at both the nucleotide and protein sequence level. It is possible that HIP is the human homologue of rodent L29. It is noteworthy that in the mouse, L29 is a member of 15-18 genes or pseudogenes (Rudert et al., 1993). It is not clear what functions, if any, these sequences serve in rodents. Several lines of evidence indicate that HIP does not function simply as a ribosomal protein. First, HIP can be detected on cell surfaces of cells transfected with HIP cDNA or RL95 cells (Rohde et al., 1996). Second, HIP is expressed in a nonconsti-tutive fashion in different human cell lines and normal tissues. Constituent ribosomal proteins would be expected to be expressed at stoichiometric levels in different cells and with respect to the cellular content of rRNA species. While there is precedent for limited modulation of ribosomal proteins in some cases (Nomura et al., 1982;Rudert et al., 1993), these proteins are never essentially absent as in the case for both HIP mRNA and protein in cells like MDA-231 and NCI-H69. Collectively, these data strongly argue that HIP is not critical to ribosomal function.
Cell surface localization (Rohde et al., 1996) and HP/HSbinding activity 2 suggest that HIP may play a role in HP/HSinvolved cell-cell or cell-matrix interactions or have other functions yet to be determined. In the studies of rodent L29, the identification and localization of this protein was based on sequence homology analysis and standard procedures of ribosomal protein isolation (Ostvold et al., 1992;Svoboda et al., 1992;Rudert et al., 1993). The distribution of the protein was only examined in one study by Northern blot analysis and in situ hybridization (Rudert et al., 1993). No studies of the expression of the L29 protein are reported. Thus, it is of interest to re-examine the expression of rodent L29 considering the possibility that it may not be a "housekeeping" protein. Considering the existence of the high number of sequences closely related to L29 (Rudert et al., 1993), it will be important to use probes specific for each gene in such studies.
HIP may be expressed both at cell surfaces and intracellularly. Several reports indicated that some proteins are present both at cell surfaces or secreted as well as inside the cell (Terada et al., 1995). These examples include certain growth factors (Abraham et al., 1986;Jaye et al., 1986), cytokines (Matsushima et al., 1986), and lectins (Cooper and Barondes, 1990). Why these proteins are expressed in both locales is unclear. In the present case, it is not known if intracellular HIP is contained within vesicles or organelles or in the cytoplasm. The following paper (Rohde et al., 1996) demonstrates that almost all of the cell-associated HIP is formed in a 100,000 ϫ g sedimentable fraction and, therefore, is not present in a freely soluble form. Several mechanisms for sorting of cytoplasmic and secreted proteins have been postulated, including that cell lysis, death, or leakage might be responsible for the release of these proteins (D' Amore, 1990) or that the release might be induced by plasma membrane evaginations (Cooper and Barondes, 1990).
Previous studies in our laboratory have suggested that HSPGs and their corresponding binding sites may play an important role in the initial attachment of mouse embryo to uterine epithelium. In the present study, a novel cell surface HP/HS-binding protein from a human uterine epithelial cell line has been cloned and expressed. The accompanying article (Rohde et al., 1996) describes expression of this protein in normal human lumenal epithelium, a location where HIP could participate in embryo attachment. Rigorous examination of a role for HIP in human embryo attachment is not possible; however, such studies can readily be performed in rodents. Therefore, it should be possible to identify the murine functional homologue of HIP and study the expression and physiological functions of this protein in the mouse in order to test its potential role in embryo implantation.

TABLE I HIP Expression in Human Cell Lines
Total RNA (20 g) and total protein (50 -100 g) from different human cell lines were extracted and subjected to Northern blot analysis using clone 23-1 cDNA probe and Western blot analysis using anti-HIP-peptide antibody, respectively, as described under "Experimental Procedures." The relative expression level of HIP mRNA in different cells is designated semi-quantitatively (see Fig. 5) as following examples: JAR or Ishikawa, ϩϩϩϩ; HeLa or RL95, ϩϩϩ; AFb-11, ϩϩ; HL60, ϩ; and MDA-231 or NCI-H69, Ϫ. Similarly, the level of HIP expression is designated as in the following examples (see Fig. 6): JAR or Ter9113, ϩϩϩ; RKO or NeuroB1, ϩϩ; HeLa, ϩ; and NCI-H69, Ϫ. N.D., not determined.