We have previously used a subtractive immunization (SI) approach to generate monoclonal antibodies (mAbs) against proteins preferentially expressed by the highly metastatic human epidermoid carcinoma cell line, M+HEp3. Here we report the immunopurification, identification and characterization of SIMA135/CDCP1 (subtractive immunization M+HEp3 associated 135 kDa protein/CUB domain containing protein 1) using one of these mAbs designated 41-2. Protein expression levels of SIMA135/CDCP1 correlated with the metastatic ability of variant HEp3 cell lines. Protein sequence analysis predicted a cell surface location and type I orientation of SIMA135/CDCP1, which was confirmed directly by immunocytochemistry. Analysis of deglycosylated cell lysates indicated that up to 40 kDa of the apparent molecular weight of SIMA135/CDCP1 is because of N-glycosylation. Western blot analysis using a antiphosphotyrosine antibody demonstrated that SIMA135/CDCP1 from HEp3 cells is tyrosine phosphorylated. Selective inhibitor studies indicated that an Src kinase family member is involved in the tyrosine phosphorylation of the protein. In addition to high expression in M+HEp3 cells, the SIMA135/CDCP1 protein is expressed to varying levels in 13 other human tumor cell lines, manifesting only a weak correlation with the reported metastatic ability of these tumor cell lines. The protein is not detected in normal human fibroblasts and endothelial cells. Northern blot analysis indicated that SIMA135/CDCP1 mRNA has a restricted expression pattern in normal human tissues with highest levels of expression in skeletal muscle and colon. Immunohistochemical analysis indicated apical and basal plasma membrane expression of SIMA135/CDCP1 in epithelial cells in normal colon. In colon tumor, SIMA135/CDCP1 expression appeared dysregulated showing extensive cell surface as well as cytoplasmic expression. Consistent with in vitro shedding experiments on HEp3 cells, SIMA135/CDCP1 was also detected within the lumen of normal and cancerous colon crypts, suggesting that protein shedding may occur in vivo. Thus, specific immunodetection followed by proteomic analysis allows for the identification and partial characterization of a heretofore uncharacterized human cell surface antigen.
Subtractive immunization is an approach to generate monoclonal antibodies (mAbs) against poorly immunogenic or rare antigens (Williams et al., 1992). The technique involves two steps. The first step, known as tolerization, requires that mice are immunized with a specific antigen or set of antigens (the tolerogen). The mice are then ‘tolerized’ to this population of epitopes either by chemical immunosuppression (King and Morrow, 1988) neonatal tolerization (Stocker and Nossal, 1976) or high zone tolerance (Briner et al., 1993). In the second step, mice are immunized with a second antigen or set of antigens (the immunogen) closely related, but not identical to the tolerogen. Manipulation of the mouse immune response in this way results in the generation of antibodies that react preferentially with antigens present in the immunogen, but not in the tolerogen. This approach has been used to generate antibodies that block neurite outgrowth (Matthew and Patterson, 1983), antibodies against embryonic-specific central nervous system antigens (Riggott and Matthew, 1996) and antibodies that can discriminate between proteins sharing high levels of sequence identity (Sleister and Rao, 2002).
We previously have used subtractive immunization, employing highly metastatic (M+) human epidermoid carcinoma HEp3 cells (Ossowski and Reich, 1983) and a low or nonmetastatic (M−) variant of HEp3 (Brooks et al., 1993; Nielsen-Preiss and Quigley, 1993), to produce mAbs that are reactive with antigenic determinants whose expression is elevated on the metastatic human tumor cells (Brooks et al., 1993; Testa et al., 1999). In this approach, mice were immunologically tolerized with M− HEp3 cells. The elicited immune response was then suppressed using cyclophosphamide and immunosuppression was followed with immunization using M+ HEp3 cells. The approach was designed to select for antibodies specific to antigens enriched on highly metastatic HEp3 cells without bias as to the nature or identity of the antigen. By screening the resulting antibodies in vivo for function-blocking ability, a subpopulation of these antibodies were also shown to be antimetastatic (Brooks et al., 1993). In a subsequent study, two of these antibodies (1A5 and 50-6) were used to identify the immunoreactive antigen, PETA-3/CD151 CDCPI, a member of the tetraspanin protein family and to more fully characterize the role of this antigen in metastasis (Testa et al., 1999).
In the present study, we have employed another of the antibodies (designated 41-2) generated by subtractive immunization to purify, identify and partially characterize a new cell surface phosphorylated glycoprotein, SIMA135/CDCP1. The properties of this protein indicate it to be a complex multidomain cell surface antigen, highly expressed by certain human cancer cells, and also normal and cancerous colon.
mAb 41-2, generated by subtractive immunization, recognizes a 135 kDa antigen expressed at elevated levels in highly metastatic human tumor HEp3 cells
We have previously used subtractive immunization to identify antigens more highly expressed by metastatic (M+) human epidermoid carcinoma HEp3 cells compared to a low or nonmetastatic (M−) HEp3 variant cell line (Brooks et al., 1993). The approach involved tolerizing with M− HEp3 cells and immunizing with M+ HEp3 cells. In the present study, we initiated the identification and characterization of the antigen recognized by another of the antibodies generated by this approach, mAb 41-2. As an initial step in determining the significance of the antigen recognized by mAb 41-2, Western blot analysis was performed on lysates prepared from M+ and M− HEp3 cells. As shown in Figure 1 (left panel), mAb 41-2 detected a single band of approximately 135 kDa in both cell types. Consistent with the subtractive immunization approach taken in generating mAb 41-2, the immunoreactive protein was expressed at higher levels in M+ than in M− HEp3 cells. A parallel gel demonstrated that the overall protein pattern and content was indistinguishable for M+ and M− HEp3 cell extracts (Figure 1, right panel), indicating that the observed difference in mAb 41-2 immunoreactivity represents a significant difference in the level of expression of the cognate antigen between the two cell lines.
Identification of the antigen recognized by mAb 41-2 from metastatic HEp3 cells
Having demonstrated a differential level of expression in M+ and M− HEp3 cells of the antigen recognized by mAb 41-2, the purified antibody was used to immunoprecipitate this antigen from M+ HEp3 cells. Consistent with the Western blot shown in Figure 1, the major protein immunoprecipitated from radiolabeled HEp3 cells had a molecular weight of approximately 135 kDa (Figure 2a). In a parallel experiment using unlabeled HEp3 cells, in which immunoprecipitated proteins were transferred to a PVDF membrane, the 135 kDa protein band was excised, subjected to trypsin digestion and the separated fragments sequenced from the N-terminus. As shown in Figure 2a, three major peptide sequences were obtained. Searches of the GenBank nonredundant protein database indicated that each of the peptide sequences had exact or near exact matches with the sequence of an unpublished protein entry with Accession number BAB15511 (aligned in Figure 2a) translated from unpublished cDNA entry AK026622.
The complete sequence of the identified protein, which we have designated subtractive immunization M+ HEp3 associated 135 kDa protein (SIMA135), is shown in Figure 2b. To confirm that SIMA135 is indeed the same protein specifically recognized by mAb 41-2, Western blot was performed on lysates from HEp3 cells, untransfected HeLa cells and HeLa cells transiently transfected with the SIMA135 cDNA. As shown in Figure 2c, mAb 41-2 reacted with the same 135 kDa protein band that is present in HEp3 cells and in HeLa cells transiently transfected with the SIMA135 cDNA, but is absent in the untransfected HeLa cells. To provide additional confirmation, the protein encoding region of the SIMA135 mRNA was cloned from HEp3 cells by reverse transcription (RT)–PCR. DNA sequence analysis of two clones generated by this approach confirmed that SIMA135 mRNA is indeed expressed by these cells. Four nucleotide differences were identified between GenBank entry AK026622 and SIMA135 sequence obtained from HEp3 cells: nucleotide 1684G→A, 1847T→C, 2236G→A and 2590G→A. The second transition is silent and the others result in amino-acid changes 525Arg→Gln, 709Gly→Asp and 827Ser→Asn, respectively. Following our identification of SIMA135, a report was published on a cDNA encoding a deduced protein designated CUB domain containing protein 1 (CDCP1; Scherl-Mostageer et al., 2001). As SIMA135 and the putative protein encoded by the CDCP1 cDNA are identical, except for amino acid 709 (Gly→Asp), in the remainder of this report we have used the designation SIMA135/CDCP1.
SIMA135/CDCP1 structural features
The SIMA135/CDCP1 protein sequence is comprised of 836 amino acids (Figure 2b) and has a deduced molecular weight of 92.9 kDa. Sequence analysis identified the following structural features. A putative amino terminal signal peptide with cleavage predicted to occur following Ala29. This feature is consistent with the sequence of peptide 1 (Figure 2a) indicating that mature SIMA135/CDCP1, with a predicted molecular weight of 90.1 kDa, starts at Phe30. A potential transmembrane domain, spanning residues 666–686 (boxed in Figure 2b), is predicted (Hartmann et al., 1989) to orient SIMA135/CDCP1 with its carboxy terminus located intracellularly. A total of 12 consensus motifs for N-glycosylation (Figure 2b) were detected. Also indicated in Figure 2b are: a consensus type 1 palmitylation motif (IICCV) (Hansen et al., 1999) at residues 687–691; five PXXP motifs that in other proteins have been shown to mediate binding to Src homology (SH) 3 domains (Pawson, 1995; Mayer, 2001); five tyrosine residues (circled in Figure 2b) that may potentially be phosphorylated; and two closely spaced tyrosine residues (Tyr734 and Tyr743) that are present in consensus motifs (YXXL/I) for SH2 domain binding (Songyang et al., 1993). Although SIMA135/CDCP1 lacked homology with other proteins in the GenBank nonredundant database, two regions of the protein, spanning residues 221–348 and 417–544, were identified with low homology to CUB (complement protein subcomponents C1r/C1s, urchin embryonic growth factor and bone morphogenetic protein 1) domains (Bork and Beckmann, 1993). In other proteins, these domains have reported roles in mediating protein–protein interactions (Sieron et al., 2000; Chen and Wallis, 2001). A third putative CUB domain described by Scherl-Mostageer and co-workers spanning residues 545–660 (Scherl-Mostageer et al., 2001) was below the homology detection threshold of the search algorithms used by us to scan the SIMA135/CDCP1 amino-acid sequence.
Expression pattern of SIMA135/CDCP1 in normal and malignant cells and tissues
The expression pattern of SIMA135/CDCP1 mRNA in 12 normal human tissues was examined by Northern blot analysis hybridizing with a 32P-labeled 2.8 kb SIMA135/CDCP1 cDNA probe. A band of approximately 6.0 kb was detected at highest levels in skeletal muscle and colon with lower levels of expression in kidney, small intestine, placenta and lung (Figure 3a). A barely detectable signal at ∼6.0 kb also was present in peripheral blood leukocytes. In addition, a much weaker signal at approximately 3.3 kb was present in skeletal muscle, colon, placenta and lung. SIMA135/CDCP1 mRNA was not detected in brain, heart, thymus, spleen or liver. Based on alignment of SIMA135/CDCP1 cDNA and genomic sequences (data not shown), it appears most likely that the two SIMA135/CDCP1 transcripts detected by Northern blot analysis result from use of alternate polyadenylation signals within the SIMA-135/CDCP1 3' UTR. The longer, more highly expressed transcript results from use of a more 3' consensus poladenylation signal (at nucleotide 5950 of our SIMA135, cDNA GenBank entry; Accession number AF468010), whereas the shorter, more lowly expressed transcript likely results from use of a variant, less efficient polyadenylation signal located at nucleotide 3186. It is also possible that these variant transcripts result from alternate splicing of the SIMA135/CDCP1 pre-mRNA.
SIMA135/CDCP1 protein expression in 16 human cell lines was analysed by Western blot analysis in which equal amounts of cell lysate protein (20 μg) were electrophoresed for each cell line. As shown in Figure 3b, SIMA135/CDCP1 was most highly expressed in metastatic HEp3 cells, with the prostate cancer cell line PC3 and the colon cancer cell line DLD-1 also manifesting high levels of expression. Moderate levels of the antigen were detected in the fibrosarcoma cell line HT1080, the gastric cancer cell lines MKN45 and STKM-1, the colon cancer cell line SW480 and the nonmetastatic prostate cancer cell line LNCaP. Low levels of SIMA135/CDCP1 were detected in two liver cancer cell lines, two breast cancer cell lines, the lung cancer cell line A549 and the kidney rhabdoid tumor cell line G401. SIMA135/CDCP1 was not detectable in normal human microvascular endothelial cells and dermal fibroblast cells. Thus, varying levels of SIMA135/CDCP1 protein are expressed in a number of human tumor cell lines, while two normal human cell types do not express the protein.
SIMA135/CDCP1 is a cell surface, phosphorylated glycoprotein
Immunocytochemistry was used to study the cellular location of SIMA135/CDCP1 in HEp3 cells and also in HeLa cells transiently transfected with a SIMA135/CDCP1 expression construct containing a FLAG epitope introduced after the carboxy terminus of the protein. HEp3 cells incubated with mAb 41-2 showed strong staining on the plasma membrane (Figure 4a). HeLa cells transiently transfected with FLAG-tagged SIMA135/CDCP1 also showed similar strong membrane staining when incubated with mAb 41-2 (Figure 4b). In addition, consistent with the predicted intracellular location of the SIMA135/CDCP1 carboxy terminus, transiently transfected HeLa cells permeabilized with Triton X-100 showed strong membrane staining when incubated with an anti-FLAG epitope mAb (Figure 4c), while nonpermeabilized cells exhibited low or near background staining with anti-FLAG mAb (Figure 4d). Untransfected HeLa cells were essentially free of staining when incubated with either mAb 41-2 or an anti-FLAG epitope mAb (data not shown). These data confirmed the predicted cell surface location as well as the type I orientation of SIMA135/CDCP1. Also, the coincidence of staining observed with mAb 41-2 and anti-FLAG mAb in HeLa cells, transiently transfected with the SIMA135/CDCP1-FLAG tag expression construct, provided additional evidence that SIMA135/CDCP1 is indeed the target antigen for mAb 41-2.
The theoretical molecular weight of mature SIMA135/CDCP1 is 90.1 kDa (Figure 2b), whereas the apparent molecular weight of the protein detected by mAb 41-2 is 135 kDa (Figure 1 and Figure 2c). To determine if this difference is because of N-glycosylation, cell lysates from HeLa cells transiently transfected with a SIMA135/CDCP1 FLAG-tag expression construct were treated with N-glycosidase F under conditions optimal for enzyme activity. Proteins were then examined by Western blot analysis using an anti-FLAG epitope mAb. As shown in Figure 5a, N-glycosidase F treatment resulted in the disappearance of the SIMA135/CDCP1 protein band at 135 kDa and replacement with a broad lower molecular weight band of approximately 95–105 kDa. In another series of experiments, lysates of M+ HEp3 cells when immunoprecipitated with mAb 41-2 and treated with N-glycosidase F also manifested a similar diminished molecular weight (data not shown). Therefore, up to 30–40 kDa of the apparent molecular weight of SIMA135/CDCP1 is because of N-glycosylation, consistent with the large number of consensus glycosylation sites in the extracellular region of this protein (Figure 2b).
The intracellular region of SIMA135/CDCP1 contains five tyrosine residues (Figure 2b). To determine whether any of these residues are phosphorylated, Western blot analysis with an antiphosphotyrosine antibody was performed on proteins immunoprecipitated from HEp3 cell lysates with mAb 41-2. As shown in Figure 5b (left panel), the antiphosphotyrosine antibody detected a protein of 135 kDa immunoprecipitated from HEp3 cells with mAb 41-2. The same protein band was detected when the immunoprecipitated proteins were probed with mAb 41-2. As controls, Western blot analysis was also performed on proteins immunoprecipitated from HeLa cell lysates with mAb 41-2, and proteins immunoprecipitated from HEp3 cell lysates with normal mouse IgG. Both immunoprecipitations were free of immunoreactivity when probed with either the anti-phosphotyrosine antibody or mAb 41-2 (Figure 5b, left panel), demonstrating the specificity of the immunoreactions observed with HEp3 cells. The involvement of a Src kinase family member in SIMA135/CDCP1 tyrosine phosphorylation was examined using PP2, an Src family-selective tyrosine kinase inhibitor (Hanke et al., 1996). Western blot analysis with an antiphosphotyrosine antibody of proteins immunoprecipitated from HEp3 cell lysates with mAb 41-2 showed that HEp3 cells treated with PP2 for 30 min had a significant reduction (∼75%) in the level of SIMA135/CDCP1 tyrosine phosphorylation compared to protein from untreated HEp3 cells (Figure 5b, right panel). Western blot analysis, using mAb 41-2 of the same immunoprecipitated proteins, indicated that approximately equal amounts of SIMA135/CDCP1 protein were present in both lanes on the membrane (Figure 5b, right panel). These data indicate that a Src kinase family member is required for tyrosine phosphorylation of SIMA135/CDCP1 in HEp3 cells.
A number of integral cell surface proteins, such as c-met (Wajih et al., 2002) and CD44 (Goebeler et al., 1996), are also produced as soluble molecules. Western blot analysis, probing with mAb 41-2, was employed to examine whether HEp3 cells produce a soluble form of SIMA135/CDCP1. HEp3 cell cultures were washed extensively with PBS then incubated for 20 h with serum-free (SF) medium. The conditioned medium (CM) was harvested and cellular material was removed by centrifugation and the media then concentrated 10-fold. As shown in Figure 5c, mAb 41-2 detected an immunoreactive band of approximately 110 kDa in HEp3 SFCM. As expected, the cell-associated SIMA135/CDCP1 from HEp3 lysates was detected at 135 kDa. In contrast, untransfected HeLa cells, which do not produce SIMA135/CDCP1, yielded no immunoreactive bands in either the lysate or concentrated SFCM. These data indicate that HEp3 cells release a soluble form of SIMA135/CDCP1 and the soluble form presents as a lower molecular weight immunoreactive protein.
Expression of SIMA135/CDCP1 in normal and cancerous colon
Immunohistochemical analysis was performed to determine the in vivo localization of SIMA135/CDCP1 in normal and cancerous colon. As shown in Figure 6a, in normal colonic mucosa, SIMA135/CDCP1 was expressed exclusively by epithelial cells where it was present uniformly on the luminal (arrowhead) and basal (arrow) surfaces of cells lining the colonic lumen and on the apical surfaces of cells lining the glandular crypts (open arrow). The presence of intense staining in the contents of goblet cells of the crypts and in the mucus in the lumen of glands (Figure 6a asterisk) suggests that SIMA135/CDCP1 is produced in a soluble form by colonic epithelial cells. In colon carcinoma specimens SIMA135/CDCP1 was extensively and heterogeneously expressed (Figure 6b) with some focal accentuation in the mucus within malignant glands (arrow Figure 6b). Some groups of invading cancer cells (Figure 6c) were heavily stained, showing the presence of SIMA135/CDCP1 on the basal, apical and lateral membranes as well as within the glandular mucus. Although there was no conclusive association of intense staining with more malignant, invading glands, there was a definite trend towards this as carcinoma cells deeper in the colonic serosa (Figure 6c) and within draining blood vessels (Figure 6d) were often strongly positive for the SIMA135/CDCP1 antigen. Control sections that were incubated with the secondary, but not the primary, antibody were free of staining (data not shown).
Subtractive immunization is an approach to the generation of antibodies against rare, poorly immunogenic or actively suppressed antigens (Williams et al., 1992). We previously have used this approach of immunosuppression/tolerization treatments to identify antibodies specific to antigens expressed at elevated levels in select metastatic human tumor cells (Brooks et al., 1993). In addition, by use of an in vivo assay, we have identified a tetraspanin plasma membrane protein PETA3/CD151, which is mechanistically involved in the process of tumor dissemination (Testa et al., 1999). Here, we report the use of a monoclonal antibody (mAb 41-2), also generated by subtractive immunization, to purify and characterize a second cell surface protein SIMA135/CDCP1. The efficacy of the subtractive immunization approach was demonstrated by the differential reactivity of mAb 41-2 to human tumor M− and M+ HEp3 cells; thereby also indicating that the target antigen, SIMA135/CDCP1, is more highly expressed by the more metastatic HEp3 cell variant. This differential reactivity again affirmed that subtractive immunization indeed works in generating antibodies that react preferentially with M+ HEp3, and indicates that this approach is likely to also have utility in generating antibodies against antigens selectively upregulated during differentiation/dedifferentiation of other cell lines.
In this report, mAb 41-2 was employed as an immunoprecipitating agent to purify its target antigen from M+ HEp3 cell lysates. Peptide sequence information obtained from the trypsin-digested, immunoprecipitated antigen allowed for database profiling that identified SIMA135/CDCP1. The SIMA135/CDCP1 cDNA was shown to encode a 135 kDa type I transmembrane cell surface protein that specifically immunoreacted with mAb 41-2. Immunocytochemical and biochemical characterization confirmed putative protein modifications and structural features predicted from the SIMA135/CDCP1 protein sequence. Immunopurification and amino-acid sequencing confirmed that the mature protein commences at Phe30 following removal of a 29 amino-acid signal peptide. Immunocytochemical analysis confirmed both the predicted cell surface location of SIMA135/CDCP1 as well as the type I orientation of this protein. In addition, consistent with the presence of 12 potential extracellular glycosylation sites, Western blot analysis of deglycosylated cell lysates indicated that up to 40 kDa of the difference between the apparent (∼135 kDa) and theoretical (∼90 kDa) molecular weight of mature SIMA135/CDCP1 is because of N-linked glycans. Also, Western blot analysis of proteins immunoprecipitated with mAb 41-2 demonstrated that SIMA135/CDCP1 is a phosphotyrosine protein, consistent with the presence of five intracellular tyrosine residues. In addition, using the inhibitor PP2, we demonstrated that a Src kinase family member is required for tyrosine phosphorylation of SIMA135/CDCP1 in HEp3 cells.
Analysis of the SIMA135/CDCP1 amino-acid sequence failed to identify motifs that would indicate it to be a functioning cell surface enzyme, or to belong to a specific class of receptor molecules or cell surface ligands. However, the ability of SIMA135/CDCP1 to be tyrosine phosphorylated, and the presence of potential binding sites for SH2 and SH3 domain containing proteins suggests that it may function in signal transduction. As Src family kinases contain both SH2 and SH3 domains (Martin, 2001), it will be important to determine whether these proteins interact directly with SIMA135/CDCP1 to cause tyrosine phosphorylation. In addition, since myristylation and palmitylation of Src family kinases target these proteins to cell membrane glycolipid-enriched microdomains (Resh, 1994), it will also be important to determine whether the SIMA135/CDCP1 consensus palmitylation signal is capable of such targeting. If this is demonstrated, the SIMA135/CDCP1 consensus palmitylation signal may provide a mechanism for regulation of signal transduction at the cell surface as has been shown for the T-cell integral membrane protein LAK (Zhang et al., 1998).
The domain structure of SIMA135/CDCP1 indicates that it will also likely interact with extracellular proteins such as soluble ligands, other cell surface proteins and/or matrix components; potentially via putative CUB domains present within its amino terminal region. Although the functions of CUB domains in other proteins have not been clearly defined, reports indicate that these structures mediate binding to a variety of protein ligands. For example, homodimerization of the MASP serine proteases acting within the lectin branch of the complement cascade is stabilized through interactions involving CUB domains (Chen and Wallis, 2001). Also, a number of the type II transmembrane serine proteases contain CUB domains thought to mediate enzyme–substrate interactions (Hooper et al., 2001). In addition, CUB domains of cubilin mediate binding to both the intrinsic factor-cobalamin as well as albumin (Yammani et al., 2001). As SIMA135/CDCP1 is heavily glycosylated within its extracellular domain, it is probable that ligand binding will be, at least partially, dependent on carbohydrate moieties as has been demonstrated for various isoforms of the cell surface glycoprotein CD44 (Bajorath, 2000). Glycosylation may also contribute to SIMA135/CDCP1 protein folding, and trafficking to and maintenance at the cell surface (Gorelik et al., 2001; Grogan et al., 2002). Identification of putative interacting cytoplasmic, membrane-associated and extracellular partner molecules as well as elucidation of the possible position of SIMA135/CDCP1 in specific signaling pathways are the subject of ongoing studies.
An intriguing aspect to the characterization of SIMA135/CDCP1 was the observation of differences in protein sequence encoded by the cDNAs obtained by us from HEp3 cells and by others from signet ring carcinoma (GenBank entry AK026622), and the nonsmall lung cell carcinoma cell line Calu 6 (GenBank entry AY026461) (Scherl-Mostageer et al., 2001). It is not yet known what effect these amino-acid differences would have on protein function. However, it appears likely, based on the position of the affected residues, that the ability of SIMA135/CDCP1 to interact with other molecules would be altered. The first amino-acid change, 525Arg→Gln, occurs within an extracellular potential ligand-binding domain; the second of the potential CUB domains of SIMA135/CDCP1. The second amino-acid change, 709Gly→Asp, is located two residues after a tyrosine residue. This change from a nonpolar amino acid to a charged residue could be expected to have a significant impact on the ability of the proximal tyrosine to be phosphorylated, and therefore may have an impact on the capacity of SIMA135/CDCP1 to bind to, for example, SH2 domains. The last change, 827Ser→Asn, is located four residues from a PXXP motif. Accordingly, this change may also impact on the ability of SIMA135/CDCP1 to interact with other proteins; in this case, SH3 domain containing proteins.
Scherl-Mostageer et al. (2001) have recently reported that SIMA135/CDCP1 mRNA is overexpressed in human colon and lung cancers. It was shown that SIMA135/CDCP1 mRNA was expressed at much higher levels in microdissected colon cancer samples from four patients than in two nonmatched normal colon samples. Our Northern blot analysis indicated that of 12 normal human tissues examined, SIMA135/CDCP1 mRNA is most abundantly expressed in colon and skeletal muscle with lower expression levels in lung as well as kidney, small intestine and placenta. Interestingly, immunohistochemical analysis using mAb 41-2 of colon tissue specimens from three patients did not detect major differences in the overall level of SIMA135/CDCP1 protein expression between normal and adjacent regions of cancer tissue. Possibly a larger number of matched normal/cancer colon tissue samples should be analysed to determine whether protein expression levels increase in the same way as Scherl-Mostageer and co-workers observed for SIMA135/CDCP1 mRNA levels.
In normal colon tissue, we observed SIMA135/CDCP1 protein on basal and apical surfaces of epithelial cells lining the colon lumen and on the apical surface of crypt epithelial cells. In contrast to its distinct localization in normal colon, SIMA135/CDCP1 distribution in colon tumor tissue was disarrayed and heterogeneous, appearing dysregulated with both plasma membrane and cytoplasmic staining. However, although only a small number of patient samples were examined, it appeared that expression of SIMA135/CDCP1 was more intense in invading glands deeper in the colonic serosa and within draining blood vessels. These data may indicate that increased SIMA135/CDCP1 protein expression is associated more with later stages of carcinogenesis such as local invasion and metastasis. This proposal is partly supported by Western blot analysis of pairs of human tumor cell lines originating from the same tissue. For example, SIMA-135/CDCP-1 levels were much higher in highly metastatic M+ HEp3 cells compared to the congenic and low metastatic variant, M− HEp3. In addition, the noncongenic prostate cancer cell lines PC-3 and LNCaP showed a similar trend; the former, a metastatic cell line, showing much higher levels of SIMA135/CDCP1 compared to the latter, a low metastatic cell type (Soos et al., 1997). For the noncongenic breast cancer cell pair, MDA-MB-231 and MCF-7, the trend was not as pronounced, as only low levels of the 135 kDa protein were detected in metastatic MDA-MB-231 cells. Moderate to high levels of SIMA135/CDCP1 were present in the metastatic gastric carcinoma cell lines MKN45 and STKM-1, but a comparison of the relative metastatic capabilities of these cell lines have not been reported. Relatively high levels of SIMA135/CDCP1 were detected in the colon cancer cell line, DLD-1, but in our hands this human tumorigenic cell line is nonmetastatic in SCID mice and chick embryos (data not shown). Thus, the level of expression of SIMA135/CDCP1 by a variety of human tumor cell lines and in tumor tissue sections was linked only weakly to the metastatic status of these cells. Additional patient samples and more congenically matched pairs of high and low metastatic tumor cell lines will have to be examined to establish whether a clear link exists between SIMA135/CDCP1 expression and the extent of malignant progression.
Our observation of apparently free SIMA135/CDCP1 in glandular mucus of both normal and malignant glands (Figure 6) is consistent with the observation that a 110 kDa soluble form of this protein is released in vitro by HEp3 cells. The distinct loss of glandular tissue ultrastructure that is apparent during tumorigenesis may permit the release of the soluble form of SIMA135/CDCP1 into the fluid and vascular system of the colon cancer patient. Accordingly, although there may not be a major difference in SIMA135/CDCP1 protein expression level in normal and cancerous colon, this protein might have utility as a serum or tissue fluid marker as has been proposed for the transmembrane proteins MUC1 (Rye and McGuckin, 2001), CD44 (Adham et al., 1990) and ICAM-1 (Maruo et al., 2002). Additional experiments are necessary to determine whether SIMA135/CDCP1 has a functional role in malignant progression and whether a released form of this protein is present at elevated levels in cancer patient fluids.
In conclusion, we have used a monoclonal antibody, 41-2, generated by subtractive immunization, and with differential reactivity to M− and M+ HEp3 cells, as an efficient tool and probe for immunopurification, permitting the detection, isolation and identification of the cell surface phosphorylated glycoprotein, SIMA135/CDCP1. Expression of this protein is dysregulated in colorectal carcinoma and also may be selectively upregulated during malignant progression. The next phase in the characterization of SIMA135/CDCP1 will be to determine whether it actively contributes to the metastatic ability of HEp3 cells, and also whether it has utility as a diagnostic marker for specific cancers. The utilization of subtractive immunization to generate specific antibodies with the ability to select out the cognate protein antigens provides a form of functional proteomics that is valuable in directly linking specific protein expression to cellular phenotype.
Materials and methods
Cell lines and hybridomas
Human cervical adenocarcinoma HeLa, fibrosarcoma HT1080, colon adenocarcinoma DLD-1 and SW480, breast adenocarcinoma MCF7, prostate adenocarcinoma PC-3, prostate carcinoma lymph node metastasis LNCaP, lung carcinoma A549 and kidney rhabdoid tumor G401 cells were obtained from the American Type Culture Collection (Rockville, MD, USA). Human liver cancer HuH7 and HLE, and gastric cancer MKN45 and STKM-1 cells were provided by Dr. Peter Vogt (The Scripps Research Institute, La Jolla, CA, USA). Breast adenocarcinoma MDA-MB-231 cells were provided by Dr Liliana Ossowski (Mount Sinai School of Medicine, NY, USA). Human epidermoid carcinoma HEp3 cells, were obtained from solid tumors serially passaged on the chorioallantoic membrane (CAM) of chicken embryos (Testa, 1992; Brooks et al., 1993). The metastatic variant of HEp3 cells, M+ HEp3, was cultured for less than 20 days before use. The low metastatic variant, M− HEp3, was maintained in culture for at least 80 days before use. Human microvascular endothelial cells (HEC) and dermal fibroblasts (HDF) were obtained from Clonetics (San Diego, CA, USA) and maintained in EGM-2 MV and FGM-2 media (Clonetics), respectively. Cancer cell lines were maintained as monolayer cultures in DMEM (Invitrogen, Carlsbad, CA, USA) supplemented with 10% FBS (HyClone, Logan, UT, USA), sodium pyruvate, penicillin/streptomycin and nonessential amino acids (Invitrogen), and grown in a humidified 5% CO2 atmosphere at 37°. Hybridomas producing mAb 41-2 were generated by a previously described subtractive immunization approach (Brooks et al., 1993). Hybridoma culturing and purification of mAbs were performed by the Protein and Nucleic Acids Core Facility of The Scripps Research Institute using standard procedures.
Protease inhibitors, normal mouse IgG, anti-FLAG M2 mAb, DAB reagent and Gill hematoxylin were purchased from Sigma (St Louis, MO, USA). Reverse transcription and PCR reagents, and the pCR-II Topo vector were from Invitrogen. PP2 was obtained from Calbiochem (La Jolla, CA, USA).
Protein purification, peptide sequencing and protein analysis
Immunoprecipitations were performed on lysates from either unlabeled or 35S-labeled HEp3 cells (5 × 107). Metabolic labeling was performed overnight in methionine/cysteine-free DMEM containing Tran35S-label (100 μCi/ml; ICN, Costa Mesa, CA, USA). Cells were washed thoroughly with PBS, then lysed in a buffer containing 0.1 M Tris (pH 8.0), 0.1% Triton X-100, 150 mM NaCl, 5 mM EDTA, 10 μ M trans-epoxysuccinyl-L-leucylamido (4-guanidino) butane, 20 μg/ml soybean trypsin inhibitor and 25 μg/ml aprotinin. Lysates were precleared against protein G-Sepharose (Pharmacia Biotech, Piscataway, NJ, USA) at 4° for 30 min, then incubated overnight at 4° with 20 μg of either mAb 41-2 or, as control, nmIgG. Immunocomplexes was precipitated using protein G-Sepharose and complexes were denatured by boiling in reducing SDS loading buffer before analysis by polyacrylamide gel electrophoresis (PAGE). For 35S-labeled proteins, the gel was dried and exposed to film at −80°. Otherwise proteins were transferred to polyvinylidine difluoride (PVDF) membranes (Millipore, Bedford, MA, USA). The predominant coomassie-stained band, at 135 kDa, was excised and then digested with trypsin. The resulting peptides were separated by high-pressure liquid chromatography and sequenced on a Procise 494 protein sequencer (Applied Biosystems, Inc., Foster City, CA, USA). Trypsin digestion and peptide sequencing were performed by the Protein and Nucleic Acids Core Facility of The Scripps Research Institute. Peptide sequences were used to search the GenBank database using algorithms available at the National Center for Biotechnology Information (NCBI) website. The complete SIMA135/CDCP1 protein sequence was analysed for structural domains, cellular processing signals and consensus post-translational modification motifs using the Prosite database (Falquet et al., 2002), the SMART algorithm (Schultz et al., 1998), the PSORT algorithm (Nakai and Kanehisa, 1992) and the NetPhos 2.0 algorithm (Blom et al., 1999).
Expression constructs and transient transfections
SIMA135/CDCP1 cDNA in the eukaryotic expression vector pME18S-FL3 (GenBank Accession number AK 026622) was generated as part of the Japanese NEDO human cDNA sequencing project and kindly provided by Dr Hiroko Hata (Department of Virology, Institute of Medical Science, University of Tokyo, Japan). The SIMA135FLAGin construct was generated by PCR placing sequences encoding the FLAG epitope (DYKDDDDK) immediately before the stop codon of the parent construct. Both constructs were sequenced. HeLa cells (4 × 105) were transiently transfected with either the SIMA135/CDCP1 or SIMA135FLAGin expression constructs using Superfect reagent (Qiagen, Valencia, CA, USA) as described by the manufacturer. Cells were lysed in ice-cold buffer containing 10 mM Tris (pH 8.0), 150 mM NaCl, 1% Triton X-100, 5 mM EDTA and 1 × Complete mini EDTA-free protease inhibitor cocktail (Roche, Indianapolis, IN, USA). Insoluble material was removed by centrifugation at 14 000 r.p.m. for 10 min.
Cloning of the SIMA135/CDCP1 cDNA from HEp3 cells
Total RNA was isolated using an RNeasy kit (Qiagen) and 2 μg served as template in a reverse transcription reaction using Superscript II reverse transcriptase. PCR was performed on 1 μl of the resulting cDNA using primers IndexTermTCCCCACC- GTCGTTTTCC and IndexTermGGTTAGGAACACGGACGGGTG (designed based upon GenBank Accession number AK 026622), and the proof reading enzyme Platinum Pfx DNA polymerase. PCR cycling conditions were 94° for 3 min, 30 cycles of 94° for 30 s, 55° for 30 s and 72° for 150 s, followed by a final 72° extension for 10 min. PCR products were gel purified (Qiagen) adenosine tailed using Platinum Taq DNA polymerase, then cloned in the pCR-II Topo vector and sequenced.
HeLa cells transiently transfected with the SIMA135FLAGin expression construct and HEp3 cells were plated on coverslips. After incubation for 48 h at 37°, cells were washed with PBS then fixed in 2% formaldehyde. HeLa cells to be incubated with anti-FLAG mAb were either not permeabilized or permeabilized by incubating in 0.5% Triton X-100 in PBS for 5 min at room temperature. Both cell types were blocked in 5% BSA in PBS. Following overnight incubation at 4° with either mAb 41-2 (5 μg/ml) or anti-FLAG M2 mAb (4 μg/ml) in blocking buffer, cells were washed with PBS then incubated with Alexa Fluor 546 conjugated goat anti-mouse IgG (2 μg/ml) (Molecular Probes). Labeled cells were visualized and photographed using a BioRad 1024 MRC2 scanning confocal imaging system.
Northern blot analysis
A human 12-lane multiple tissue Northern blot (Clontech) was hybridized with [α-32P]dCTP labeled (Ambion) EcoRI/HincII DNA insert fragments of the SIMA135/CDCP1 cDNA overnight in UltraHyb solution (Ambion) at 68°. The blot was washed to a final stringency of 0.1 × SSC, 0.1% SDS at 68° and then exposed to film at −80°. Blots were reprobed with ß-actin cDNA to determine consistency of RNA loading in each lane.
Western blot analysis
Cell lysates, serum-free conditioned media and immunoprecipitated proteins were separated by electrophoresis through 8% SDS–PAGE and then transferred to nitrocellulose membranes (Millipore). Membranes were blocked in 5% nonfat skim milk powder in PBS, then incubated overnight at 4° with either mAb 41-2 (2 μg/ml), anti-FLAG M2 mAb (0.8 μg/ml) or anti-phosphotyrosine mAb (1 μg/ml; Upstate Biotechnology, Lake Placid, NY, USA). Following extensive washing, membranes were incubated for 2 h at room temperature with goat anti-mouse IgG (0.16 μg/ml, Pierce, Rockford, Il, USA) and immunoreactive bands detected by enhanced chemiluminescence (Pierce).
Biochemical characterization procedures
For removal of N-linked glycans, lysates (50 μl) from M+ HEp3 cells and HeLa cells transiently transfected with the SIMA135FLAGin expression construct were denatured and reduced in 0.5% SDS, 1% β-mercaptoethanol for 10 min at 100°, then incubated with PNGase F (New England Biolabs, Beverly, MA, USA) at 37° for 45 min. For analysis of the basal level of tyrosine phosphorylation of SIMA135/CDCP1, subconfluent cultures of HEp3 and HeLa (as negative control) cells were incubated at 37° for 30 min with serum-free DMEM containing 50 mM NaF and 1 mM Na3VO4 and then washed with ice-cold PBS. For inhibition of Src kinase family phosphorylation, HEp3 cells were cultured in serum-free DMEM without NaF and Na3VO4 for 30 min at 37° with PP2 (50 μ M). Cells were then lysed in ice cold buffer containing 50 mM Tris (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 25 μg/ml aprotinin, 25 μg/ml leupeptin, 50 mM NaF and 1 mM Na3VO4. Insoluble material was removed by centrifugation at 14 000 r.p.m. for 10 min. Immunoprecipitation was performed as described above on 300 μg of cell lysates using 1 μg of either mAb 41-2 or nmIgG (as negative control). For assays, for the presence of soluble SIMA135/CDCP1, HEp3 cells approaching confluence were washed three times with PBS then incubated in serum-free conditioned media for 24 h. The media was collected and centrifuged at 4° and 10 000 g, then concentrated 10-fold using micron centrifugal filters with a molecular weight cutoff of 30 000 kDa (Millipore). Cells lysates were collected as described above.
Cryostat sections (6 μm) from archival human adenocarcinoma colon tissue samples from three patients were fixed in zinc–formalin for 15 min, rinsed briefly with PBS, then nonspecific binding sites were blocked by incubating in PBS containing 3% BSA. mAb 41-2 (5 μg/ml) was applied at 4° overnight. Specific antibody binding was detected by the addition of biotin-conjugated anti-mouse antibodies (Pierce) followed by peroxidase-conjugated neutravidin (Pierce), which was visualized using DAB reagent. Sections were counterstained using Gill hematoxylin.
Adham IM, Klemm U, Maier WM and Engel W . (1990). Hum. Genet., 84, 125–128.
Bajorath J . (2000). Proteins, 39, 103–111.
Blom N, Gammeltoft S and Brunak S . (1999). J. Mol. Biol., 294, 1351–1362.
Bork P and Beckmann G . (1993). J. Mol. Biol., 231, 539–545.
Briner TJ, Kuo MC, Keating KM, Rogers BL and Greenstein JL . (1993). Proc. Natl. Acad. Sci. USA, 90, 7608–7612.
Brooks PC, Lin JM, French DL and Quigley JP . (1993). J. Cell Biol., 122, 1351–1359.
Chen CB and Wallis R . (2001). J. Biol. Chem., 276, 25894–25902.
Falquet L, Pagni M, Bucher P, Hulo N, Sigrist CJ, Hofmann K and Bairoch A . (2002). Nucleic Acids Res., 30, 235–238.
Goebeler M, Kaufmann D, Brocker EB and Klein CE . (1996). J. Cell Sci., 109 (Part 7), 1957–1964.
Gorelik E, Galili U and Raz A . (2001). Cancer Metast. Rev., 20, 245–277.
Grogan MJ, Pratt MR, Marcaurelle LA and Bertozzi CR . (2002). Annu. Rev. Biochem., 71, 593–634.
Hanke JH, Gardner JP, Dow RL, Changelian PS, Brissette WH, Weringer EJ, Pollok BA and Connelly PA . (1996). J. Biol. Chem., 271, 695–701.
Hansen SG, Grosenbach DW and Hruby DE . (1999). Virology, 254, 124–137.
Hartmann E, Rapoport TA and Lodish HF . (1989). Proc. Natl. Acad. Sci. USA, 86, 5786–5790.
Hooper JD, Clements JA, Quigley JP and Antalis TM . (2001). J. Biol. Chem., 276, 857–860.
King SW and Morrow Jr KJ, (1988). Biotechniques, 6, 856–861.
Martin GS . (2001). Nat. Rev. Mol. Cell Biol., 2, 467–475.
Maruo Y, Gochi A, Kaihara A, Shimamura H, Yamada T, Tanaka N and Orita K . (2002). Int. J. Cancer, 100, 486–490.
Matthew WD and Patterson PH . (1983). Cold Spring Harb. Symp. Quant. Biol., 48 (Part 2), 625–631.
Mayer BJ . (2001). J. Cell Sci., 114, 1253–1263.
Nakai K and Kanehisa M . (1992). Genomics, 14, 897–911.
Nielsen-Preiss SM and Quigley JP . (1993). J. Cell Biochem., 51, 219–235.
Ossowski L and Reich E . (1983). Cell, 33, 323–333.
Pawson T . (1995). Nature, 373, 573–580.
Resh MD . (1994) Cell, 76, 411–413.
Riggott MJ and Matthew WD . (1996). J. Neurosci. Methods, 68, 235–245.
Rye PD and McGuckin MA . (2001). Tumour Biol., 22, 269–272.
Scherl-Mostageer M, Sommergruber W, Abseher R, Hauptmann R, Ambros P and Schweifer N . (2001). Oncogene, 20, 4402–4408.
Schultz J, Milpetz F, Bork P and Ponting CP . (1998). Proc. Natl. Acad. Sci. USA, 95, 5857–5864.
Sieron AL, Tretiakova A, Jameson BA, Segall ML, Lund-Katz S, Khan MT, Li S and Stocker W . (2000). Biochemistry, 39, 3231–3239.
Sleister HM and Rao AG . (2002). J. Immunol. Methods, 261, 213–220.
Songyang Z, Shoelson SE, Chaudhuri M, Gish G, Pawson T, Haser WG, King F, Roberts T, Ratnofsky S, Lechleider RJ, Neel BG, Birge RB, Fajardo JE, Chou MM, Hanafusa H, Schaffhausen B and Cantley LC . (1993). Cell, 72, 767–778.
Soos G, Jones RF, Haas GP and Wang CY . (1997). Anticancer Res., 17, 4253–4258.
Stocker JW and Nossal GJ . (1976). Contemp. Top. Immunobiol., 5, 191–210.
Testa JE . (1992). Cancer Res., 52, 5597–5603.
Testa JE, Brooks PC, Lin JM and Quigley JP . (1999). Cancer Res., 59, 3812–3820.
Wajih N, Walter J and Sane DC . (2002). Circ. Res., 90, 46–52.
Williams CV, Stechmann CL and McLoon SC . (1992). Biotechniques, 12, 842–847.
Yammani RR, Seetharam S and Seetharam B . (2001). J. Biol. Chem., 276, 44777–44784.
Zhang W, Trible RP and Samelson LE . (1998). Immunity, 9, 239–246.
We thank Rebecca Mellor, Giano Panzarella, Sarah Lustig and Jeanine Kleeman for expert technical assistance and Dr Linda Wasserman (Department of Pathology, University of California, San Diego, CA, USA) for archival colon tissue samples. This work was supported by National Institutes of Health Grants CA65660, HL31950 (JPQ), and Training Grants T32 HL07695 (AZ) and T32 HL07195 (GFC), and a National Health and Medical Research Council of Australia CJ Martin/RG Menzies Fellowship (JDH).
About this article
Cite this article
Hooper, J., Zijlstra, A., Aimes, R. et al. Subtractive immunization using highly metastatic human tumor cells identifies SIMA135/CDCP1, a 135 kDa cell surface phosphorylated glycoprotein antigen. Oncogene 22, 1783–1794 (2003). https://doi.org/10.1038/sj.onc.1206220
- subtractive immunization
- colon cancer
Expression and functional characterization of the CUB domain-containing protein from the triangle sail mussel (Hyriopsis cumingii) in response to pathogenic infection
Comparative Biochemistry and Physiology Part B: Biochemistry and Molecular Biology (2021)
Journal of Proteome Research (2020)
Annals of Hematology (2020)
An oncogenic viral interferon regulatory factor upregulates CUB domain-containing protein 1 to promote angiogenesis by hijacking transcription factor lymphoid enhancer-binding factor 1 and metastasis suppressor CD82
Cell Death & Differentiation (2020)
Journal of Clinical Investigation (2020)