TSP50 Encodes a Testis-Specific Protease and Is Negatively Regulated by p53 Free

Previously, while searching for genes that were overexpressed in human breast cancer tissues, the TSP50 gene was discovered by a hypomethylated DNA fragment isolated via the methylation-sensitive representational difference analysis technique (1). TSP50 is a testis-specific gene encoding a protein that is homologous to serine proteases (enzymes characterized by a triad of histidine, aspartate, and serine residues necessary for catalytic activity). However, the catalytic triad of TSP50 has a substitution of threonine at the serine residue site, which distinguishes it from serine proteases and suggests that it could be a unique type of threonine protease (2). The tissue-specific character (i.e., threonine substitution at a key catalytic site) and cellular locations were also shared by its mouse orthologue mTSP50 (3). These evolutionary conservatives suggested that it might play a crucial role during mammalian spermatogenesis. Most importantly, TSP50 was abnormally activated in many breast cancer biopsies tested, and the gene products were found in malignant epithelial cells. These results imply that it could be involved in breast cancer development. This hypothesis could be reasoned by the fact that spermatogenesis is a complex, multifaceted process by which diploid spermatogonia are committed to further differentiation by dividing mitotically to give rise to haploid spermatocytes that undergo two meiotic divisions to produce haploid round spermatids. Those spermatids experience a dramatic biochemical and morphologic restructure, which leads to the creation of mature spermatozoa (4, 5). Recently, an old theory has regained prominence, claiming that selected features of germ cells undergoing spermatogenesis can be imposed on transformed somatic cells (6). This is because many newly discovered testis-specific genes are expressed in different types of malignant somatic cells (7). These genes are termed cancer testis antigens (CTA; refs. 8–14). Because TSP50 is a testis-specific gene that is re-expressed in human breast cancer cells, it is considered a CTA (1, 2, 14).

We were interested in further investigating the biochemical and cell biological natures of TSP50 as well as the regulatory mechanisms related to its differential expression. To this end, we characterized the enzymatic activity of TSP50, examined its detailed subcellular location, and analyzed its promoter sequence using the Patch Search program. The Search results indicated that there were consensus sequences for several transcription factors, including p53 (15, 16) and Sp1 (17, 18), suggesting that TSP50 gene expression could be regulated by those trans-elements. Because the p53 gene is an important tumor suppressor (15, 19), discovering the regulatory effect of p53 on TSP50 gene expression became our main focus.

Cell lines. MCF7/Adr [p53 del 126-132, multidrug resistant (MDR) cells] and its parental cell line MCF7 (wild-type p53; ref. 20) were grown under the conditions as described (21, 22). Hs578T (p53 V157F), MDA-MB-231 (p53 R280K), MDA-MB-435 (p53 G266E), MDA-MB-436 (mutated p53), and T47D (p53 L194F; refs. 23–25; gifts of Dr. Moll M. Ute at the Department of Pathology and Dr. Cao Jian at the Division of Hematology of State University of New York at Stony Brook) were grown at 37°C with 5% CO₂ in DMEM with 10% FCS, 100 μg/mL streptomycin, and 100 units/mL penicillin. HEK293 (human transformed primary embryonic kidney cells, American Type Culture Collection, Rockville, MD) and HeLa cells were grown at 37°C with 5% CO₂ in RPMI 1640 culture medium with 10% FCS, 100 μg/mL streptomycin, and 100 units/mL penicillin.

Recombinant DNA constructs. Detailed information for constructing the pGEX/TSP50 plasmid to generate glutathione S-transferase (GST)/TSP50 fusion proteins was previously described (2, 3). To study the cellular location of the TSP50 protein, TSP50 was amplified by PCR using the sense (TTAAGCTTCCACTGCTGATCC-3′) and antisense primer (5′-CTGAATTCCCACAAGTGAGGGCACAC-3), containing the HindIII and EcoRI sites, respectively. The resulting PCR products were subcloned into pEGFP-C1 (Clontech, Inc., Palo Alto, CA) to generate pEGFP/TSP50, where the enhanced green fluorescent protein (EGFP) was linked to the NH₂ terminus of TSP50. To obtain the pGL/TSP50P plasmid for luciferase assays, a sense (5′-GGGGTACCCTGGCTTGATTTAACTTTCAC-3′) and antisense primer (5′-GAAGATCTTGTGGCAGCCGACTGCGTCTC-3′), containing the KpnI and BglII sites, were designed for PCR amplification to generate the TSP50 promoter (from −521 to +450, the transcriptional initiation site refers as +1) using testis genomic DNA as a template. The resulting PCR fragment was cloned into the luciferase reporter plasmid pGL3 (Promega, Madison, WI). The resulting construct was verified by sequencing (Genewiz, Inc., Brunswick, NJ). pcDNA/P53 and pcDNA/P53/R249S were gifts of Dr. Moll M. Ute.

Site-directed mutagenesis. To analyze the enzyme activity of TSP50, a negative control was preferred. We expected that no enzyme activity would be exhibited when its catalytic threonine residual is replaced by alanine. To obtain this control, base pair A in the ACG sequence that codes the threonine residual was mutated into G using pGEX/TSP50 as a template and the Site Directed Mutagenesis kit (Stratagene, La Jolla, CA) following the manufacturer's instruction. Successful substitution in the resulting construct (pGEX/TSP50/T/A) was verified by DNA sequencing.

Protein purification. GST, GST/TSP50, and GST/TSP50T/A fusion proteins were purified as previously described (3). To purify endogenous TSP50, an AT-50-affinity column was prepared using Sepharose 4B beads conjugated with the TSP50 antibody (AT-50) following the manufacturer's instruction (Amersham, Arlington Heights, IL). In addition, an affinity column conjugated with the serine protein inhibitor benzamidine was purchased from Sigma (St. Louis, MO). TSP50 was purified by passing the cell lysate through the AT-50-affinity column and eluting it with 3.5 mol/L MgCl₂ followed by an immediate desalting process using a desalt spin column. A portion of purified TSP50 was applied onto a benzamidine-affinity column and then eluted with 3.5 mol/L MgCl₂ solution. Alternatively, TSP50 was purified by first passing the cell lysate through the benzamidine-affinity column followed by the eluting/desalting procedure and then applying onto an AT-50-affinity column.

Zymogram gel assay. To test enzymatic activity, equal mole ratios of GST/TSP50, GST/TSP50/T/A, and GST proteins were applied in duplicate to 12% Tris-glycine gels containing β-casein or gelatin (Bethesda Research Laboratory, Bethesda, MD; refs. 26, 27). The positive control for the β-casein gel was homogenous trypsin type IX (Sigma), whereas collagenase type VII (Sigma) was used for the gelatin gel. One set of gels for the zymographies was carried out following the manufacturer's instruction. After the enzymatic reactions, the gels were stained with Coomassie blue R-250 and destained with acetic acid and methanol. To normalize the enzyme-digested band created by GST/TSP50, duplicate gels, without the corresponding substrates, were produced for Western analysis. The experiments were done four times.

Cellular location. pEGFP/TSP50 and the empty vector pEGFP-C1 were transiently transfected, in parallel, into HeLa cells using the LipofectAMINE Plus Transfection kit (Invitrogen, San Diego, CA) following the manufacturer's instructions. The endoplasmic reticulum (ER)–targeting plasmid, pEYFP-ER (Clontech) was transfected into cells in parallel and served as a control for the ER structure. After 24 h, the cells were inspected under a confocal microscope. The experiments were done three times.

Transient transfection and luciferase assays. The LipofectAMINE Plus Transfection kit (Invitrogen) was employed for the transient transfection assays. Host cell lines used to test the influence of p53 on TSP50 promoter activity were HeLa, HEK293, and paired MCF7 cells. In brief, 0.2 μg pcDNA/P53, pcDNA/P53R249S, or pcDNA3.1, and pGL/TSP50P or pGL3 along with 0.2 μg pCMV/β-galactosidase (transfection efficiency control) were cotransfected into different host cells when they (seeded onto 24-well plates) reached 50% to 70% confluence. After 24 h of transfection, luciferase activity was measured using the Luciferase Assay System (Promega) according to the manufacturer's instructions. β-Galactosidase activity was evaluated using the Beta-Glo Assay System (Promega). The luciferase activity of transfectants was compared after normalizing their β-galactosidase activity and protein concentrations (28, 29). The experiments were done four times. To test if the TSP50 gene promoter was differentially regulated in paired MCF7 cells, pGL3 or pGL/TSP50P were transfected into those cell lines in parallel to perform luciferase assays. In addition, MCF7/Adr cell line was used as host to test if overexpression of the p53 transgene could reduce endogenous TSP50 gene expression.

Semiquantitative reverse transcription-PCR. Total RNA was prepared from cell lines using the High Pure RNA Isolation kit (Roche, Indianapolis, IN) following the manufacturer's instructions. Normal testis and breast tissue RNAs were purchased from Clontech and served as positive and negative controls, respectively. To evaluate the TSP50 gene expression levels in different cell lines, semiquantitative reverse transcription-PCR (RT-PCR) was done using the Titan One tube RT-PCR system (Roche) as previously described (26, 29). The sense and antisense primers for TSP50 and GAPDH (internal control) were also previously described (1, 22, 29). The lengths of the TSP50 and GAPDH PCR products were about 700 and 300 bp, respectively. The PCR and quantification of PCR products were done as described (2, 22, 28, 29). The experiments were done three times.

Electrophoretic mobility shift assay and super electrophoretic mobility shift assay. Electrophoretic mobility shift assay (EMSA) was done using two probes, which were located in the 5′-end untranslated region, and the nuclear extracts were prepared from MCF7/Adr and MCF7 cells. Probe 1, which did not include p53- and Sp1-binding motifs, represented the region from +89 to +138 in the TSP50 promoter, whereas probe 2 contained p53- and Sp1-binding motifs and covered the region from +210 to +259. Both probes were synthesized and their forward sequences were 5′-TGAGCCAGTCCAGGGCAATGGGCTGCGAGGGTGAGGAGGTCTGAGAATCG-3′ and 5′-G[AGGGAAGTCA]GCCG[GGGCGTACCC]AGCCCGGGAGGTATAAGGGAGGGCT-3′, respectively. Patch Research indicated p53-response elements in probe 2 were bracketed, whereas the indicated Sp1-binding site was overlapped with the second half of the p53 motif. The forward and reverse sequences were annealed. Probe was then end-labeled using T4 polynucleotide kinase in the presence of [γ-³²P]dATP. Nuclear extracts were generated according to the method of Dignam et al. (30). EMSA and super EMSA experiments were done as previously described (28, 29). The antibodies for p53 and Sp1 (Santa Cruz Biotechnology, Santa Cruz, CA) were used to carry out super EMSA. The experiments were done three times.

Purification of GST, GST/TSP50, and GST/TSP50T/A fusion proteins. To analyze TSP50 enzyme activity, GST, GST/TSP50, and GST/TSP50T/A were expressed in Escherichia coli and purified via GST-affinity columns. Purification of the fusion proteins was confirmed on a SDS-PAGE gel (Fig. 1A). In addition, a duplicate gel was transferred onto a nylon membrane to perform Western blot analysis where the TSP50 antibody AT-50 was used (Fig. 1B). The successful purification was further verified using the GST antibody (Fig. 1B). In addition, endogenous TSP50 was obtained from the MDA-MB-231 breast cancer cell line by passing the cell lysate solely through an AT-50-affinity column, or through both AT-50- and benzamidine-affinity columns in either order. The purified TSP50 proteins were directly analyzed by zymogram gels and Western blot assays (see below).

TSP50 exhibited protease activity. To determine if TSP50 was a protease, zymogram gel assays were done using purified GST/TSP50 and GST/TSP50/T/A, where GST served as a negative control. Briefly, three proteins were applied onto a 12% Tris-glycine gel containing β-casein or gelatin, where homogenous trypsin type IX and collagenase type VII served as positive controls, respectively. Zones of protein lyses due to enzyme digestion appeared as clear bands with a blue background. A duplicate gel without substrate was also carried out for Western analysis to normalize the location of the enzymatic bands. The results showed that an identically sized digested band appeared in all three products, which was considered background. However, only the GST/TSP50 protein, which was confirmed by Western analysis, exhibited a better cleavage efficiency for β-casein (Fig. 1C and D) compared with the gelatin gel (data not shown). No enzyme activity was observed for GST/TSP50T/A and GST. These findings suggested that GST/TSP50 contained enzyme activity, and the threonine catalytic site was crucial for such activity. To confirm these results, endogenous TSP50 was purified from the human breast cancer cell line MDA-MB-231 employing different strategies: (a) via the AT-50-affinity column alone; (b) via AT-50- followed by the benzamidine-affinity column; and (c) via benzamidine- and then the AT-50-affinity column. Zymogram gels and Western analyses suggested that three batches of TSP50 proteins also expressed a better cleavage efficiency for β-casein (Fig. 1E) than the gelatin gel (data not shown). Therefore, our studies indicated that TSP50 was a new threonine enzyme.

TSP50 was located in the ER and cytoplasm membrane. To determine the subcellular location of TSP50, pEGFP/TSP50 was transfected into HeLa cells, where pEGFP-C1 served as the control. The EGFP/TSP50 fusion protein was monitored by confocal microscopy. The results showed that the fusion protein was located in the ER and the cytoplasm membrane (Fig. 2A–D). The ER location of TSP50 was compared with an image generated by an ER indicator (pEYFP-ER), which was also introduced into HeLa cells. Consequently, the ER structure was clearly outlined, and this pattern was similar to the one created by EGFP/TSP50 (Fig. 2E and F).

The TSP50 promoter was negatively regulated by p53. A typical TATA box and several TATA- and CAT-like boxes were located in the 5′-end genomic and untranslated region of the TSP50 gene, which featured the gene promoter. Due to the reactivation of TSP50 in many breast cancer patients, we were interested in elucidating the mechanisms involved in its differential expression. We first searched for potential transcription factor–binding motifs in the promoter region by the Patch Search Program. Thus, several motifs were indicated, which included motifs for p53 and Sp1 transcription factors. Because p53 is an important tumor suppressor gene involved in human malignancies, we wanted to test if it regulated TSP50 expression. To do so, the TSP50 promoter was constructed into pGL3 to generate pGL/TSP50P; this was then cotransfected with pcDNA/P53 or pcDNA3.1 along with pCMV/β-galactosidase into different hosts, which included HeLa, HEK293, MCF7/Adr, and MCF7 cells. Luciferase assay results showed that the p53 transgene reduced the promoter function about three to five times in all the hosts compared with the controls containing the empty vector (Fig. 3). These findings suggested that p53 negatively regulated the TSP50 promoter. We also cotransfected the mutated p53 gene construct pcDNA/P53R249S and pGL/TSP50P into those cells. We found that the mutated gene had no influence on TSP50 promoter activity in HeLa and MCF7/Adr, which either contained the mutated or attenuated p53 gene (20, 31); however, it increased the promoter activity in HEK293 and MCF7 cells, both of which had the wild-type p53 gene.

The p53 transgene reduced endogenous TSP50 gene expression. To further confirm that p53 down-regulated the TSP50 gene, MCF7/Adr cells, which expressed a very high level of TSP50, were transfected with pcDNA/P53 or pcDNA3.1 and pCMV/β-galactosidase to see if the p53 transgene could inhibit endogenous TSP50 gene expression. After 24 h, RNA was isolated from the transfectants and converted into cDNA. Semiquantitative RT-PCR was then done to evaluate the TSP50 transcript. As expected, cells expressing the p53 transgene produced three times less TSP50 RNA than the control cells containing the pcDNA3.1 vector (Fig. 4A and B). This result provided additional evidence that p53 negatively regulated the TSP50 gene expression. Accordingly, we measured TSP50 expression levels in a breast cancer cell line panel whose p53 gene status was either wild type or mutated/defected.

Expression of TSP50 was significantly higher in cells whose p53 gene was mutated. Semiquantitative RT-PCR was done to measure TSP50 gene expression in seven breast cancer cell lines. Combined RNAs isolated from normal breast and testis tissues served as negative and positive controls, whereas GAPDH served as quantitative control. We found that, with the exception of Hs578T, the TSP50 transcripts in cell lines with mutated p53, such as MCF7/Adr, MDA-MB-231, MDA-MB-435, MDA-MB-436, and T47D, were much higher than that in MCF7 cells, which contained the wild-type p53 gene. Although, as expected, none and high TSP50 expression levels were detected in normal breast and testis, respectively (Fig. 5A and B). Because there was a significant disparity concerning TSP50 promoter activity in paired MCF7 cells, we were interested in testing if the promoter was targeted by different transcription factors by performing EMSA and super EMSA assays.

Diverse nuclear proteins in paired MCF7 cells targeted the TSP50 promoter. To perform EMSA assays, we designed two probes: probe 1 that did not contain p53- and Sp1-binding sites and probe 2 that included those sites. The probes were then interacted with the nuclear extracts prepared from paired MCF7 cells. The results showed that both probes displayed diverse binding patterns because they incubated with different nuclear protein sources (Fig. 6). To see if p53 and Sp1 proteins were part of the DNA/protein complexes, super EMSA was done using p53 and Sp1 antibodies. We found that the Sp1 antibody was able to shift probe 2-protein but not probe 1-protein complexes generated by both extracts to a higher position (Fig. 6). In addition, the quantity of Sp1 in MCF7 was notably greater than that in MCF7/Adr cells. These findings suggested that Sp1 could be involved in the differential regulation of TSP50 via sequence of probe 2 in those cells. Because MCF7 cells contained the wild-type p53 gene, we expected that the p53 antibody could shift probe 2-MCF7-protein complexes. However, it did not. This failure could be attributed to low p53 gene expression in the nonstressed MCF7 cells. Therefore, we employed another strategy to perform super EMSA, where probe 2 was incubated with the purified recombinant p53 protein. The results were also negative (data not shown). In addition, chromatin immunoprecipitation assay was done using the p53 antibody and MCF7 cells that were treated with doxorubicin to induce p53 expression. Again, we were unable to generate a PCR product that represented the TSP50 promoter sequence using the antibody-trapped DNA as the template.

Previously, we reported that the TSP50 protein had a similar catalytic structure to many serine proteases; however, a substitution at a key catalytic site (serine) with threonine made it a unique protease candidate (1–3). To prove that TSP50 encoded an enzyme, zymogram gel assays were done using the purified GST-fused TSP50 proteins, whose NH₂-terminal consensus sequences signaling for enzymatic activation were eliminated, as well as endogenous TSP50 that was purified from MDA-MB-231 cells. The result showed that both forms TSP50 proteins were capable of digesting β-casein and gelatin. However, replacing the catalytic threonine with an alanine residual in the fusion protein diminished this activity. This finding verified that TSP50 was a protease possessing a rare catalytic site. Therefore, one could consider TSP50 a new type of threonine protease. However, we noticed that TSP50 displayed relatively weak enzymatic activity for both substrates compared with the positive controls. Two possibilities could explain this phenomenon: (a) the key catalytic site of TSP50 (threonine) could be less active than the serine residual, and/or (b) neither substrate used in our study is the prime substrate of TSP50. If this is so, it will be important to identify its native target, which is critical in illuminating its biological function in spermatogenesis and tumor development.

Although earlier results showed that TSP50 was normally expressed in spermatocytes and abnormally re-expressed in breast cancer epithelial cells, detailed information regarding its subcellular location was missing. By using the pEGFP/TSP50 construct and confocal microscope technique, we found that the TSP50 protein was located in the ER-like structures as well as the cytoplasm membrane. The membrane location was consistent with the fact that TSP50 contains a COOH-terminal hydrophobic membrane–anchoring region, which is also present in a few serine proteases (32–34). Several reports suggest that these regions could be trans-membrane domains that remain on the membrane after protein shedding (31, 35).

The TSP50 gene was considered a testis antigen (14) because it was reactivated in breast cancer cells. We were interested in understanding the mechanisms involved in its transcription regulation, which would help us comprehend its cancer-related differential expression. Therefore, we did the Patch Search, which led to the discovery of a putative p53- and Sp1-binding motif. To see whether TSP50 gene expression was controlled by p53, a “landmark” tumor suppressor and whose mutations are associated with a broad range of human cancers (28, 30, 36–39), luciferase assays were done. We found that the p53 transgene significantly reduced promoter activity in all the tested host cells of diverse origins, which suggested that TSP50 was negatively governed by p53. We also noticed that the mutated p53 gene exerted different effects on the TSP50 gene promoter in cells that had the wild-type or abnormal endogenous p53. These disparities could be caused by competition between p53R249S and p53 for an unknown mediator(s) in HEK293 and MCF7, which was necessary for p53 to regulate TSP50 gene expression. However, no such competition occurred in HeLa and MCF7/Adr cells because they contained either the mutated or attenuated p53 gene, respectively (20, 40). Because EMSA and chromatin immunoprecipitation assays did not indicate a physical relationship between the p53 protein and TSP50 promoter region tested, which suggested that the negative effect of p53 on that region could be achieved indirectly via an unknown mediator(s).

The negative effect of p53 on TSP50 expression was further confirmed by the observation that the p53 transgene was able to considerably reduce endogenous TSP50 gene expression in MCF7/Adr cells. In addition, by measuring the TSP50 transcript levels in seven breast cancer cell lines, we found that five of the six lines containing mutated p53 produced much higher transcripts than the MCF7 cell line, which expressed the wild-type p53 gene. Furthermore, the significant disparity of TSP50 gene expression in paired MCF7 cells (both are estrogen receptor positive) suggested that it was estrogen receptor independent. It is also possible that TSP50 gene expression might be connected to the degree of malignancy among breast cancer cell lines because the tumorigenic cells, such as MDA-MB-231, MDA-MB-435, MDA-MB-436, and T47D (23, 41), as well as the advanced cancer cell line MCF7/Adr generated substantially higher TSP50 transcripts compared with the nontumorigenic lines Hs578T and MCF7 (42–44). This scenario remains to be further tested, although many enzymes involved in human cancers, such as matrix metalloproteinases, are associated with tumor metastasis (45, 46).

We also discovered that higher amounts of the Sp1 transcription factor bound to the TSP50 gene promoter in MCF7 compared with MCF7/Adr cells, which indicated that Sp1 could behave as a suppressor of gene expression. Although the primary role of Sp1 is to bind to GC boxes present in a variety of cellular and viral promoters to stimulate their functions (17, 18), it was also reported that it is capable of inhibiting gene expression (47). In addition, we wondered if TSP50 was involved in MDR development because of its differential expression in paired MCF7 cells. This speculation was unlikely because integrating TSP50 into the genome of HEK293 cells did not considerably increase their IC₅₀ to doxorubicin compared with the control cells incorporated with the empty vector (data not shown).

In summary, we have shown that the TSP50 gene encodes a unique protease and is located in the ER and cytoplasm membrane. These features make it a perfect testis antigen for immunotherapy and a biomarker for breast cancer diagnosis. Most importantly, its down-regulation by the p53 gene suggests that it could be an oncogene participating in the onset and progression of breast cancer. Therefore, we are interested in testing the tumorigenic and metastatic potentials of the TSP50 gene in the near future.

Grant support: Department of Defense Breast Cancer Research Award DAMD17-97-7070, Friends for an Earlier Breast Cancer Test Foundation, and Waldbaum Foundation for Breast Cancer Research.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Drs. U.M. Moll and J. Cao for the cell lines and constructs and J.C. Duffy for article preparation.