Understanding molecular biology of the prostate and the process of prostate carcinogenesis is brought forward by the identification and characterization of new genes specifically expressed or overexpressed in prostate tissue. The encoded proteins may, in addition, provide novel diagnostic and therapeutic tools for prostate carcinoma. Currently, there is limited information regarding the molecular details of normal human prostate function and the genetic events associated with the malignant transformation of prostatic cells. This is largely due to the cellular heterogeneity of prostate and the lack of systematic analysis of prostate tissue gene expression. Few identified highly specific prostate-enriched genes are prostate specific antigen (PSA),1 prostate-specific membrane antigen,2 NKX3.1,3 prostate stem cell antigen4 and PCGEM1.5 These are potential candidates to serve as therapeutic vaccine antigens or as novel cancer markers.
In this study, we describe our work on a novel human protein exhibiting significant homology to prostate cancer associated protein rat PARM-1 (GenBank accession NM_173114) with 98% identity in the C-terminal cytoplasmic tail and the transmembrane domain and 54.5% identity in the extracellular region. The high protein sequence identity suggested that it may be the human ortholog of the rat PARM-1 and was thus named human PARM-1 (hPARM-1). The rat PARM-1 was identified in the ventral prostate and is overexpressed in regressing prostate after androgen withdrawal.6 It has been shown that rat PARM-1 overexpression leads to increased telomerase activity in MAT LyLu cells,7 which is often associated with immortalization of cancer cell lines. On the basis of these results, it was proposed that rat PARM-1 is involved in a survival program enabling certain prostatic cells to resist apoptosis after androgen ablation. The human ortholog of PARM-1 has not yet been characterized; however, the transcript has been identified in several large-scale gene expression analyses of various human tissues and cancer cell lines.8, 9 Moreover, in contrast to rat PARM-1, hPARM-1 has been detected at high levels both in normal prostate and in prostate cancer,10 suggesting that hPARM-1 is present in normal glands and not only during androgen deprivation as shown for rat PARM-1.
In all, this gave us the impetus to characterize this protein for functional, biochemical and cell biological properties and explore its role in human prostate cancer. We found that hPARM-1 is a highly glycosylated protein primarily located at the plasma membrane and in the endocytic pathway. In the prostate cancer cell line, LNCaP, the transcript is upregulated by androgen, which is further corroborated with its expression in the androgen-dependent human prostate cancer xenograft model CWR22. Also, ectopic expression of hPARM-1 in a prostate cancer cell line increases cell proliferation, thereby suggesting its role in the physiobiology of prostate and prostate cancer.
Material and methods
The rabbit polyclonal serum against rat PARM-1 was kindly provided by Dr. J. Closset (Liege, Belgium). The monoclonal antibody 6C4 against lysobisphosphatidic acid (LBPA) was provided by Dr. J. Gruenberg (Geneva, Switzerland). Dr. S. Carlsson (Umeaa, Sweden) provided the rabbit anti-LAMP-1 polyclonal antibody. Mouse anti-EEA1 monoclonal antibody was obtained from Clontech. Goat-anti-mouse and goat anti-rabbit secondary antibodies coupled to Alexa-488/Alexa-594 were obtained from Molecular Probes (The Netherlands). Protein A gold was purchased from Dr. Georg Posthuma (Utrecht, The Netherlands).
Plasmids and xenograft
The hPARM-1-GFP construct was generated as described.11 The pMEP4-hPARM-1 and pcDNA3.1-hPARM-1-GFP constructs were generated by PCR and subcloning. All constructs were verified by sequencing. Total RNA from CWR22 xenograft was used for preparing cDNA, which was further used in real-time RT PCR.
Cell culture and transfection
Madin-Darby canine kidney strain II (MDCK) and HEK293 cells were maintained in DMEM containing 10% fetal calf serum, 2 mM L-glutamine, 25 U/ml penicillin, and 25 μg/ml streptomycin. Human umbilical vein endothelial cells (HUVEC) were grown in MCDB 131 Medium as described.12 The LNCaP, PC3 and DU145 human prostate cancer cells were cultured in RPMI 1640 (BioWhittaker) having 10% fetal bovine serum, 2 mM L- glutamine, 50 U/ml penicillin and 50 μg/ml streptomycin. Transient and stable transfections were performed with Lipofectamine 2000 (Invitrogen), according to the supplied protocol. For making stable cell lines, positive clones were identified by immunofluorescence microscopy. Cells transfected with pMEP4-hPARM-1 were induced with 20 μM CdCl2/6 hr to induce protein expression.
Androgen induction of LNCaP cells
For androgen induction, the LNCaP cells were kept in RPMI containing 2% charcoal-treated FBS and glutamine for 48 hr followed by 24 hr incubation in RPMI containing 0.5% charcoal-treated serum. The cells were then induced with 10−7 M R1881(Dupont-NEN), an androgen analogue. At different time points total RNA was isolated using Trizol reagent (Invitrogen) for performing real time RT-PCR.
Metabolic labeling and deglycosylation
For metabolic labeling, MDCK cells transiently transfected with hPARM-1-GFP were starved in cysteine/methionine-free DMEM for 45 min at 37°C and labeled with 100 μCi/ml [35S]-Pro-mix (Amersham Biosciences) for 3 hr. For sulphate labeling, to cells were starved in sulphate-free RPMI 1640 for 45 min, and then 0.3 mCi/ml carrier-free [35SO] (PerkinElmer) was added overnight at 37°C. Cells were washed and lysed in cold lysis buffer (50 mM Tris pH 7.5, 150 mM NaCl, 1% NP-40) containing protease arrest (Gentech, St. Louis, MO) for 20 min. The protein was immunoprecipitated with anti-GFP antibodies overnight, washed in lysis buffer and subjected to SDS-PAGE and blotted onto nitrocellulose membrane (Amersham Biosciences). Immunoprecipitated hPARM-1 was deglycosylated according to the protocol provided in the E-DEGLY deglycosylation kit (Sigma-Aldrich). For digesting glycosaminoglycan chains (GAG), 100 mU Chondroitin ABC lyase in digesting buffer (5 mM Tris-HCl, pH 8.0, 5 mM Sodium acetate buffer, 0.01% BSA) or 50 mU keratanase in 0.05 M Tris-HCl, pH 7.5, overnight at 37°C was used. The HNO2 reagent (0.5 M H2SO4 and 0.5 M Ba(NO2)2 1.1 v/v) was then added to the protein sample (1:1) for 10 min at room temperature and neutralized with 2 M Na2CO3.
Colony formation assay
PC3 cells transfected with 1.0 μg of either empty pcDNA3.1 vector or pcDNA3.1-hPARM-1 construct were seeded on new plates (10,000 cells/100 mm dish) 24 hr after transfection with selection antibiotic (G418, Sigma-Aldrich) and cultured for 2 weeks as described.13 Cells were then methanol-fixed at −20°C for 30 min. The colonies formed were stained with crystal violet (0.1%) and the area covered by the colonies was measured by Image J software (http://www.uhnresearch.ca/facilities/wcif/imagej).
Cells were grown on coverslips, fixed in 3% paraformaldehyde, stained with appropriate primary and secondary antibodies diluted in 1 × PBS/0.02% saponin and mounted with Mowiol (Sigma-Aldrich). Confocal images were acquired with Olympus FW 1000 laser scanning microscope equipped with a 60 x/1.1 oil immersion objective (Olympus, Hamburg, Germany).
Cells grown in 10 cm culture dishes were processed for cryo-electron microscopy (EM) as previously described.14 Cryosections were sequentially incubated with rabbit anti-PARM-1 antibody and protein A-gold particles diluted in PBS/1% BSA for 30 min at room temperature. After contrasting, cryosections were examined using a Philips CM100 transmission electron microscope (FEI, Acht, The Netherlands) and pictures were obtained with a MegaView III 1,000*1,000 pixel digital camera (SIS, Münster, Germany).
A human multiple tissue Northern blot (Clontech) containing 1μg of poly(A)+ RNA per lane was hybridized with a full-length hPARM-1-cDNA probe according to the manufacturer's instructions. The cDNA probe was labeled with [α-32P]dCTP (3,000 Ci/mmol) (Amersham Biosciences) by random hexamer primers as described in the protocol. The blot was stripped and reprobed with human β-Actin cDNA probe as control.
Quantitative real time RT-PCR
Total RNA (4 μg) extracted with Trizol reagent (Invitrogen) was used for reverse transcription using the Superscript II system (Invitrogen). The cDNA obtained was subjected to real-time PCR analysis for hPARM-1 expression in the Light Cycler Instrument (Roche) using the Light Cycler—FastStart DNA Master SYBR Green I Kit (Roche). A standard curve made from serial dilutions of cDNA was used to calculate the relative amount of hPARM-1 mRNA in each sample. These values were normalized to the relative amount of the internal standard glyceraldehyde-3-phosphate dehydrogenase (GAPDH) in the same samples, calculated from a standard curve established in the same way. The experiment was repeated 3 times with similar results. The primers used: GAPDH: forward 5′-GGCCTCCAAGGAGTAAGACC-3′, reverse 5′-AG GGGTCTACATGGCAACTG-3′ and hPARM-1: forward 5′-GC TCATAGACATGGAGACCACC-3′, reverse 5′-CCCAGGACCC GTAGTCATGG-3′.
Sequence analysis of hPARM-1
The Gateway™ cloning system was used, as described previously,11 to generate N- and C- terminal green fluorescent protein fusions with open reading frames derived from novel human cDNAs. These expression constructs were transfected into living cells and the subcellular localizations of the encoded proteins were determined.11 Using this approach, one novel human protein (DKFZP564O0823; called hPARM-1) localizing to endocytic structures and showing high similarity to rat PARM-1, which is involved in prostate carcinogenesis was identified (Fig. 1a). BLAST search of the NCBI/EMBL protein database also revealed a 26% identity of hPARM-1 with human mucin, MUC6 (GenBank accession AY184388.1); however, neither similarity to any known protein family nor presence of any conserved functional domains was observed. The hPARM-1 gene is localized at chromosomal locus 4q13.3–q21.3. Its complete mRNA consists of 5,044 bp (GenBank accession NM_015393), which encodes a 310 amino acid polypeptide. The amino acid sequence of hPARM-1 cDNA has characteristics of a type I integral membrane protein (Fig. 1b) with a signal peptide and a single transmembrane domain as revealed by sequence analysis using SMART program. Comparing cDNA and genomic sequences of hPARM-1 showed the boundaries of 4 exons and 3 introns (Fig. 1c). Part of the signal peptide and the extracellular domain are encoded by exon 2, followed by a 78-bp exon 3 encoding the transmembrane domain. Exon 4 encodes the cytoplasmic domain and also contains the 3′ untranslated region (UTR). Protein sequence analysis by ELM resource15 and NetOGlyc 3.1 programs16 revealed 4 motifs for N-glycosylation, 60 sites for mucin-type O-glycosylation and 8 GAG attachment sites in the extracellular part of the protein. Also, serine, threonine and proline residues together constituted >40% of all amino acids in the extracellular part of the protein, a characteristic of mucins.
hPARM-1 is a mucin-like protein
Since the protein sequence analysis of hPARM-1 indicated presence of mucin-type glycosylations and GAGs, which are normally sulphated, we first studied if hPARM-1 is sulphated or not. Transiently transfected MDCK cells with hPARM-1-GFP were labeled with radioactive sulphate followed by immunoprecipitation with an anti-GFP antibody. A single sulphated high molecular weight band with estimated size of 130 kDa was detected on SDS-PAGE (Fig. 2a), which is different from the calculated molecular mass (60.8 kDa) for hPARM-1-GFP. This showed that hPARM-1 is sulphated and may be heavily glycosylated. To verify the presence of O-linked glycans, we treated the immunoprecipitated hPARM-1-GFP protein with a cocktail of enzymes (neuraminidase, O-glycosidase, β(1–4)-galactosidase and β-N-acetyl-glucosaminidase) to remove the glycans, which reduced the molecular size of the protein to 109 kDa. Further addition of N-deglycosidase, PNGaseF, reduced the molecular size of the protein to 100 kDa, confirming the presence of both O- and N-linked glycans (Fig. 2b). The MW of hPARM-1-GFP remained unchanged when treated with cABC, HNO2 and keratanase for removing GAGs (data not shown), indicating that the protein is not a glycosaminoglycan. Thus, hPARM-1 is a highly glycosylated, mucin-like protein, with N- and O-linked glycosylations constituting > 60% of the relative molecular mass. The complicated moiety of mucin-like glycosylation and unavailability of enzymes digesting such linkages most likely explains why we could not obtain the predicted molecular mass of the naked protein even after enzymatic digestion of glycans.
Expression of hPARM-1 in cell lines and tissues
The rat PARM-1 is expressed in the epithelial cells of prostate gland and also in many other tissues with very high expression in heart. We, therefore, first examined the hPARM-1 mRNA expression by real time RT-PCR in prostate epithelial cell lines (LNCaP, PC3, DU145), as well as in other epithelial cells from kidney (HEK293) and cervix (HeLa) and the endothelial HUVEC cells. Primers were designed to amplify sequence between exon 2 and 4, to avoid amplification of genomic DNA. The results showed a band of expected size (216 bp) in all cell lines, except for PC3, demonstrating the presence of the hPARM-1 mRNA transcript in epithelial and endothelial cells (Fig. 3a).
We next studied the expression of hPARM-1 in various tissues by probing a preprepared Northern blot containing poly A+ RNA from twelve different human tissues with a full-length hPARM-1 cDNA probe. A ∼5,000 bp band corresponding to the reported size of hPARM-1 mRNA (5,044 bp) and a less abundant ∼2,400 bp band were detected in most tissues (Fig. 3b). Equal loading of total mRNA was confirmed by rehybridizing the blot with a probe against actin. High amounts of hPARM-1 mRNA was observed in heart, kidney and placenta; little and no expression of the transcript was seen in liver and peripheral leukocytes, respectively. Thus, hPARM-1 has a wide tissue distribution, which correlates with its presence in various epithelial and endothelial cells.
hPARM-1 expression decreases in regressing human prostate cancer xenograft
The rat PARM-1 is upregulated when normal prostate regresses after androgen ablation. To examine the effect of androgen withdrawal on hPARM-1 expression in human prostate, we used CWR22 xenograft, which recapitulates the human condition as it is dependent on androgen for growth and undergoes rapid regression upon castration. The total RNA from regressing tumors collected at 1, 2 or 4 weeks postcastration was subjected to real time RT-PCR. A marked expression of hPARM-1 in CWR22 xenograft was observed which decreased by 45% by Day 7 postcastration and further by 59% by Day 28 postcastration (Fig. 4a). The positive control, PSA decreased by about 97% by Day 7 after castration. Surprisingly, our results showed that hPARM-1 expression decreases on androgen withdrawal in prostate cancer xenograft.
Androgen stimulation induces hPARM-1 expression in LNCaP
To further scrutinize the relation between androgen and hPARM-1, we assessed the time-dependent effect of androgen stimulation on the hPARM-1 expression by real time RT-PCR using the androgen responsive LNCaP cells. The cells were first incubated in serum- and androgen-deprived medium for 3 days after which a slight but nonsignificant increase in the hPARM-1 transcript level was seen (data not shown). After stimulation with androgen analogue R1881, a ∼7-fold increase in hPARM-1 mRNA was observed at 36 hr and this remained stable for at least 84 hr (Fig. 4b). The expression of PSA (positive control) increased by ∼50-fold after 84 hr of R1881 stimulation (not shown). The specificity of androgen stimulation was confirmed in LNCaP cells using the bicalutamide androgen antagonist Casodex (10−6 M). There was a significant decrease (∼90-fold) in expression of hPARM-1 when Casodex was added along with R1881 to LNCaP cells after 48 hr of treatment (data not shown). Moreover, there was no significant change in the expression of hPARM-1 on androgen stimulation by R1881 in the androgen-receptor negative DU145 cell line (data not shown). These data suggest that expression of hPARM-1 in prostate is regulated by androgens.
Overexpression of hPARM-1 increases colony formation
To investigate any biological effect of hPARM-1 expression on the regulation of cell growth, colony formation assays were performed with PC3 cells that do not express hPARM-1. PC3 cellstransfected with either an empty pcDNA3.1 vector or the pcDNA3.1-hPARM-1 construct were cultured in the selection medium containing G418. There was significantly higher number of colonies in the hPARM-1-transfected cells (Figs. 5a and 5b), suggesting a probable role of hPARM-1 in cell proliferation. The ectopic expression of hPARM-1 in PC3 cells was confirmed by immunofluorescence confocal microscopy using anti-PARM-1 polyclonal antibody (result not shown).
hPARM-1 is sorted to early endocytic pathway
The cellular localization of hPARM-1 was examined by immunofluorescence confocal microscopy and EM of MDCK cells stably transfected with either hPARM-1 or hPARM-1-GFP. The hPARM-1-GFP colocalized completely with anti-PARM-1 polyclonal antibody demonstrating a specific immunoreaction with the antibody and nomal trafficking of the hPARM-1-GFP construct (data not shown). Cells were immunolabeled with antibodies against specific endosomal markers. The hPARM-1 protein colocalized with early endosomal marker, EEA1 (Fig. 6a), late endosomal marker, LBPA (Fig. 6b), but not with lysosomal marker, LAMP-1 (Fig. 6c). The EM studies showed presence of hPARM-1 in the Golgi, on the internal vesicles of both early and late endosomes (Figs. 7a–7c) and at the plasma membrane, both on microvilli and the flat membrane (Fig. 7d) but not in lysosomes (not shown). Noteworthy, the protein was present in many small vesicular structures throughout the cytoplasm, indicating high transport activity of the protein.
In this study we have characterized hPARM-1, the human ortholog of rat PARM-1. The hPARM-1 gene encodes a 310-amino acid protein containing a single transmembrane domain with a cytoplasmic C-terminal end and an extracellular N-terminal end having mucin-like glycosylations. We found abundant expression of hPARM-1 in heart, placenta, kidney and lung tissues, suggesting that it may play an important role in many different tissues. This is in accordance with the microarray analysis of the human transcriptome that identified the transcript in several human carcinomas (prostate, colon, kidney and lung).9 Interestingly, there is a high expression of hPARM-1 in leukemia and in the lymphoma cell line K422.10 Our tissue blot results showed absence of hPARM-1 mRNA in peripheral leukocytes, indicating a possible upregulation of the transcript in leukemic cells. The tissue distribution of hPARM-1 is similar to that reported for rat PARM-1.6
Inspite of hPARM-1 sharing high sequence similarity with the rat ortholog, our data indicates that PARM-1 gene is differently regulated in human and rat in many ways. First, there is an elevation in hPARM-1 mRNA level in the LNCaP human prostate cancer cells on androgen stimulation. Also, there is a significant expression of hPARM-1 transcript in the androgen-dependent human CWR22 prostate xenograft, which decreased on androgen deprivation. This is in contrast to rat PARM-1, which is upregulated in normal prostate gland after androgen withdrawal.6 However, it is important to mention that we studied expression of hPARM-1in a cancer xenograft tumor, whereas rat PARM1 expression was studied in normal prostate tissue. Second, it is interesting to note that hPARM-1 transcript is highly expressed in normal human prostate tissue,10 while rat PARM-1 is only expressed in prostate upon androgen withdrawal. These observations indicate a difference in the PARM-1 gene regulation in rat and human. This is not uncommon as probasin, an androgen-regulated prostate specific gene is differently regulated on postcastration in mouse and rat. The expression level of rat probasin increases after postcastration whereas that of mouse probasin decreases.17 Third, in Northern blot, two variant forms of hPARM-1 mRNA were detected in contrast to the tissue blot analysis of rat PARM-1 that gave only one product. The presence of an additional shorter transcript of 2,400 bp in hPARM-1 could be due to a second polyadenylation site predicted by GENSCAN web server at 2,331–2,336 bp, which is absent in rat PARM-1. This further indicates a complex regulation and expression of the human PARM-1 gene. Endolyn/CD164, a sialomucin fundamentally involved in cell proliferation, is an example of a gene whose expression depends on differential usage of polyadenylation sites within a single 3′UTR.18 The 3′UTR regulates the stability and translatability of the mRNA transcript thereby affecting the protein expression and thus could be of functional relevance.19 Further studies are required to shed light on this area.
The cell proliferation/colony formation-promoting effects of hPARM-1 overexpression in PC3 prostate cancer cells suggest that it may have a functional role on cell growth in prostate cancer. This is supported by the data reporting presence of hPARM-1 in many malignant tumors (kidney, colon, prostate), which display dysregulated growth. Many androgen-regulated genes are known to either stimulate or inhibit cellular growth, thereby playing a critical role in the growth and differentiation of prostate tissue and in prostate carcinogenesis.20 Interestingly, in a yeast two-hybrid screen using cytoplasmic tail of hPARM-1 (residues 280–310) as “bait,” TRIP-1 (TGF-β receptor interacting protein 1) was identified as a potential interacting partner (data not shown). TRIP-1 interacts with TGF-β receptor, a multifunctional cytokine involved in the regulation of proliferation, differentiation, migration and survival of many endothelial cells.21 However, additional work is required to establish the significance of this potential interaction.
The hPARM-1 protein is localized to Golgi, plasma membrane and early and late endosomes, but not in lysosomes. No colocalization was observed with the lysosomal marker LAMP1 even after treatment with lysosomal protease inhibitors (Leupeptin and E-64d) and an inhibitor of V-ATPase (concanamycin A) (unpublished results). Thus, we suggest that hPARM-1 functions at the plasma membrane and in the early endocytic pathway only. We have also demonstrated that the tyrosine-sorting signal, 287YGRL290 in the cytoplasmic tail of hPARM-1 is necessary for mediating its proper intracellular trafficking and localization as the mutated protein (287YGRL290 to 287AGRA290) was found mainly at the plasma membrane (data not shown). It is important to note that hPARM-1 is not unique among the prostate-specific proteins to be localized in early endocytic pathway and contribute to cell proliferation. STAMP2, for instance, an androgen-regulated protein with high expression in prostate, is localized to the Golgi, plasma membrane and early endosomes and plays a role in prostate cell proliferation.22
This is the first evidence showing hPARM-1 to be a mucin-like molecule. A high content of sulphated N- and O-linked sugars and presence of proline, threonine and serine residues as the predominant amino acids concentrated in certain regions of the polypeptide termed “PTS regions” suggest hPARM-1 to be a mucin-like molecule. Also, interesting to note is the similarity of hPARM-1 with MUC6, which is involved in intestinal and breast carcinomas.23 Mucins are expressed in secretory epithelial surfaces of specialized organs (like liver, pancreas, kidney and eye) and contribute to the regulation of differentiation and proliferation of tumor cells. Mucins like MUC1 and MUC4 are aberrantly expressed in prostate and pancreatic adenocarcinomas, respectively.24
In conclusion, we have cloned and characterized the human ortholog of PARM-1 and further demonstrated that it is a novel androgen-regulated gene highly expressed in androgen-dependent cancer xenograft. We have also showed that forced expression of hPARM-1 in hPARM-1 nonexpressing human prostate cancer cells increases their growth. Also, we propose that the protein encoded by hPARM-1 gene is a mucin-like molecule. However, further characterization of this gene will reveal its associated proteins and precise cellular functions thereby contributing to understanding of its role in prostate and prostate cancer.
We thank Dr. Kristian Prydz and Dr. Trond Berg for valuable discussions and comments on the manuscript. Dr. Heidi Tveit is acknowledged for technical help on deglycosylation studies. The help from Miss Ling Wang is highly appreciated in conducting colony formation assay. The authors sincerely thank Mr. Torstein Lindstand for providing the control DNA samples for androgen stimulation studies. The EM department at Oslo University is acknowledged for the use of equipments.