1. ¹Laboratory of Molecular Medicine, Human Genome Center, Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai Minato-ku, Tokyo 108-8639, Japan
2. ²Department of Internal Medicine, Shin Kong Wu Ho-Su Memorial Hospital, 95 Wen Chang Road, Taipei 11160, Taiwan

Correspondence: Y Furukawa, E-mail: furukawa@ims.u-tokyo.ac.jp

Received 19 February 2004; Revised 29 July 2004; Accepted 2 August 2004; Published online 11 October 2004.

Abstract

To disclose mechanisms of colorectal carcinogenesis and identify novel diagnostic markers and drug targets for treatment of these tumors, we previously analysed the expression profiles of 11 colorectal cancers using a genome-wide cDNA microarray containing 23 040 genes. Among the genes commonly transactivated in the cancers, we identified a novel human gene, which we termed CLUAP1 (clusterin-associated protein 1). It encodes a nuclear protein of 413 amino acids containing a coiled-coil domain. To investigate its function, we searched for CLUAP1-interacting proteins using yeast two-hybrid system and identified nuclear Clusterin. Expression of CLUAP1 was gradually increased in the late S to G2/M phases of cell cycle and it returned to the basal level in the G0/G1 phases. Suppression of this gene by short interfering RNAs resulted in growth retardation in the transfected cells. These data provide better understanding of colorectal carcinogenesis, and inactivation of CLUAP1 may conceivably serve in the future as a novel therapeutic intervention for treatment of colon cancer.

Colorectal carcinomas are leading causes of cancer death worldwide. In spite of recent progress in diagnostic and therapeutic strategies, prognosis of patients with advanced cancers remains very poor. Although molecular studies have revealed that alterations of tumor suppressor genes and/or oncogenes are involved in colorectal carcinogenesis, the precise mechanisms remain to be fully elucidated. To disclose mechanisms underlying these tumors from a genome-wide point of view and to discover target molecules for diagnosis and for development of novel therapeutic drugs, we have been analysing their expression profiles by cDNA microarray representing 23 040 genes (Kitahara et al., 2001; Lin et al., 2002). These efforts have pinpointed a number of genes, including ESTs, which appear to be upregulated frequently in the cancer tissues compared with the corresponding noncancerous cells. Since carcinogenesis involves activation of oncogenes and/or inactivation of tumor suppressor genes, enhanced expression of at least some of these upregulated genes may reflect oncogenic properties.

Among the genes with frequently elevated expression in cancer tissues in our microarray data, an EST with an in-house accession number of B9223 corresponding to KIAA0643, Hs. 155995 in UniGene cluster (http://www.ncbi.nlm.nih.gov/UniGene/), was upregulated in the cancer tissues compared to their corresponding noncancerous mucosa in a magnification range between 2.1 and 3.5 in five out of seven cases that passed the cutoff filter (data not shown). To clarify the results of the microarray, we carried out semiquantitative RT–PCR and revealed that expression of B9223 was increased in 14 of additional 20 colon cancers (70%) compared with their corresponding normal mucosae (Figure 1a).

Figure 1.

(a) Semiquantitative RT–PCR showing that B9223 expression was increased in 14 of 20 colon cancers compared to their corresponding normal mucosae. Extraction of RNA, cDNA synthesis, and PCR reaction were carried out as reported previously (Shimokawa et al., 2003). Amplification proceeded for 4 min at 94°C for denaturing, followed by 21 (for GAPDH) and 32 (for CLUAP1), cycles of 94°C for 30 s, 60°C for 30 s, and 72°C for 60 s, in the GeneAmp PCR system 9700 (Perkin-Elmer, Foster City, CA, USA). Primer sequences were as follows: GAPDH, forward 5'-ACAACAGCCTCAAGATCATCAG-3' and reverse 5'-GGTCCACCACTGACACGTTG-3'; CLUAP1, forward 5'-GAGTGGAAGTAACGATGACTC-3' and reverse 5'-GTCATTGTCACTCTCATCCAG-3'. (b) Alignment of CLUAP1 (GenBank Accession number AB089691), a mouse RIKEN cDNA 2610111M03 (BC024910), a zebra fish unknown protein for MGC: 56115 (AAH45921), A. ganbiae ENSANGP00000008052 (XP_310880), and a C. elegans putative protein for 4E162 (NP_500339). Boxed region corresponds to the coiled-coil motif, and shading indicates residues that are conserved in at least three proteins

Full figure and legend (270K)

Homology searches with the sequence of B9223 in public databases using BLAST program in National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/BLAST/) identified a gene that had been annotated as KIAA0643 protein, clone MGC:9638 (GenBank Accession number BC017070), and a genomic sequence with GenBank Accession number of NT_010552.13 assigned to chromosomal band 16p13. To determine the coding sequence of the gene, we predicted candidate–exon sequences in the genomic sequence using GENSCAN (http://genes.mit.edu/GENSCAN.h
tml) and Gene Recognition and Assembly Internet Link (GRAIL, http://compbio.ornl.gov/Grail-1.3/) program, and performed exon-connection experiments. As a result, we obtained an assembled sequence of 1681 nucleotides (GenBank Accession number AB089691) containing an open-reading frame of 1239 nucleotides encoding a putative 413-amino-acid protein, and termed the gene CLUAP1 (clusterin-associated protein 1) because of the reason described below. The first ATG was flanked by a sequence (AGCGTTATGT) that agreed with the consensus sequence for the initiation of translation in eukaryotes, with an in-frame stop codon upstream. Comparison of CLUAP1 cDNA and the genomic sequence disclosed that this gene consisted of 11 exons. The amino-acid sequence of the predicted CLUAP1 protein showed 89, 62, 38, 38, and 31% identities to a mouse RIKEN cDNA 2610111M03 (NP_084014), a zebra fish unknown protein for MGC:56115 (AAH45921), Anopheles ganbiae ENSANGP00000008052 (XP_310880), a Caenorhabditis elegans hypothetical protein (BAD 23996), and Drosophila melanogaster CGI7599-PA (NP_608470), respectively. A search for protein motifs with the Simple Modular Architecture Research Tool (SMART, http://smart.embl-heidelberg.de) revealed that the predicted protein contained a single coiled-coil region (codons 195–267), which was conserved in the homologues. In addition to the coiled-coil motif, these homologues also shared similarities in amino acids between codons 1 and 184 of CLUAP1 (Figure 1b).

Multiple-tissue Northern blotting with CLUAP1 cDNA as a probe showed that a 1.6-kb transcript was expressed abundantly in human testis, thyroid and trachea, and moderately in spinal cord and adrenal gland (Figure 2a). To examine the expression and explore the function of CLUAP1, we prepared polyclonal antibody against CLUAP1. Western blot analysis using whole extracts of colon cancer cells, including HCT116, SNU-C4, and SW480, showed a 53 kDa band that corresponded to CLUAP1 (Figure 2b). As a negative control for the Western blot analysis, we used extracts from TE8 esophageal cancer cells that showed low level of CLUAP1 expression by semiquantitative RT–PCR (Figure 2b).

Figure 2.

(a) Human multiple-tissue blots (Clontech) were hybridized with a ³²P-labeled PCR product of CLUAP1. Prehybridization, hybridization, and washing were performed according to the supplier's recommendations. A 1.6-kb transcript was expressed abundantly in human testis, thyroid, and trachea. Expression of -actin served as an internal control. (b) Western blot analysis of CLUAP1 in cancer cell lines. HCT116, SW480, and SNU-C4 were established from colon cancer tissues, while TE8 was from esophageal cancer tissue. The band indicated with an arrow corresponded to CLUAP1. Expression of -actin served as an internal control. (c) Immunocytochemical staining using anti-CLUAP1 antibody in HCT116 cells showing the nuclear localization of endogenous CLUAP1. Recombinant His-tagged CLUAP1 protein was produced in Escherichia coli and inoculated for the immunization of rabbits. The polyclonal antibody to CLUAP1 was purified from the sera as described elsewhere (Shimokawa et al., 2003). Immunohistochemical staining of HCT116 cells with rabbit anti-CLUAP1 antibody was performed as reported previously (Okabe et al., 2003). The reaction was visualized after incubation with fluorescein-conjugated anti-rabbit second antibody (Leinco and ICN). Nuclei were counter-stained with 4',6'-diamidine-2'-phenylindole dihydrochloride (DAPI). (d) Immunohistochemical staining of CLUAP1 in paraffin-embedded cancer tissues using anti-CLUAP1 antibody. We analysed a total of 24 clinical samples including 14 cancer and 10 adenoma tissues, and assessed the reactivity semiquantitatively using the following grading system; -, no staining; +, slight staining; 3+, strong staining; and 2+, staining between + and 3+. For positive controls (3+), sections of normal testis were utilized. If the score of the cancer cells was greater than that of corresponding epithelial cells in noncancerous mucosae in the same tissue section, we defined that CLUAP expression was elevated in the cancer by immunohistochemistry. Two pathologists independently reviewed the sections. (e) Western blot analysis of CLUAP1 using three colon cancer tissues (T) and their corresponding noncancerous mucosa (N). The band indicated with an arrow corresponded to CLUAP1

Full figure and legend (133K)

To reveal its subcellular localization, fluorescent immunocytochemical staining of endogenous CLUAP1 was carried out in HCT116 cells. Cells were stained with anti-CLUAP1 and visualized by fluorescein-conjugated secondary antibody. Signals were observed mainly in the nuclei (Figure 2c). To compare the expression levels of CLUAP1 protein between noncancerous epithelial cells and tumor cells, formalin-fixed paraffin-embedded clinical tissues including 14 advanced cancers and 10 adenomas of the colon were subjected to immunohistochemical staining. Cancerous cells were more strongly stained with anti-CLUAP1 antibody than noncancerous epithelial cells in six tumors out of the 14 cancers and two out of the 10 adenomas (Figure 2d). Further Western blot analysis using three clinical tissues demonstrated that its expression was augmented in one out of the three cancer tissues compared to the noncancerous mucosa (Figure 2a).

To clarify further the function(s) of CLUAP1, we searched for CLUAP1-interacting proteins using yeast two-hybrid screening system. Among a total of 70 positive clones that were picked and sequenced, 48 sequences matched a known protein called Clusterin (CLU). Subsequent simultaneous transformation of pAS2-CLUAP1 with pACT2-CLU corroborated an interaction between CLUAP1 and CLU in the yeast cells. Apart from the secreted form of CLU, another form is synthesized from a second AUG and expressed in the cytoplasm and nucleus (Reddy et al., 1996; Leskov et al., 2003). Since subcellular localization of CLUAP1 was nucleus, we focused on the nuclear form of CLU (nCLU). The positive CLU clones contained codons between 252 and 449, indicating that the region responsible for the interaction was in nCLU. To prove the association between CLUAP1 and nCLU in vivo, we transfected COS7 cells with plasmids expressing myc-tagged CLUAP1 (pcDNA-myc-CLUAP1) with or without plasmids expressing FLAG-tagged C-term nCLU (p3XFLAG-Clusterin) and carried out co-immunoprecipitation assay. Immunoprecipitation with anti-FLAG antibody and Western blot using anti-myc antibody showed a single band corresponding to CLUAP1, and immunoprecipitation with anti-myc antibody and Western blot using anti-FLAG showed a band corresponding to nCLU, suggesting that CLUAP1 associates with nCLU in vivo (Figure 3a).

Figure 3.

(a) Interaction between CLUAP1 and nCLU in vivo. Yeast two-hybrid assays were performed with the MATCHMAKER GAL4 Two-Hybrid System according to the manufacturer's protocols (BD Biosciences). A total of 1 10⁶ clones from a human testis MATCHMAKER cDNA library were screened using pAS2-CLUAP1 expressing the entire coding region of CLUAP1 as bait (BD Biosciences). The C-terminal region of nuclear Clusterin from the isolated positive clones was subcloned into p3XFLAG-CMV vector (p3XFLAG-Clusterin). The coding region of CLUAP1 was amplified by RT–PCR using gene specific primer sets 5'-GAGGAATTCCGACCCTGGGCTCCTGGGGAC-3' and 5'-AAGCTCGAGAAGTCATTGTCACTCTCATCCAG-3', and cloned into appropriate cloning site of pcDNA3.1myc/His vector (pcDNA-myc-CLUAP1). COS7 cells were transfected with pcDNA-myc-CLUAP1 expressing myc-tagged CLUAP1 and/or p3XFLAG-Clusterin as indicated, and their lysates were subsequently immunoprecipitated with anti-myc (BD Biosciences) or anti-FLAG M2 (SIGMA) antibody. The precipitated protein was separated by SDS–PAGE and immunoblot analysis was carried out using with anti-myc antibody (upper panel) or rabbit anti-FLAG antibody (lower panel). (b) Colocalization of CLUAP1 and nCLU protein in cells. COS7 cells were transfected with pcDNA-myc-CLUAP1 and p3XFLAG-Clusterin, and stained with mouse anti-myc antibody and visualized with anti-mouse antibody IgG conjugated with FITC. (c) The cells were stained with rabbit anti-FLAG antibody and visualized with anti-rabbit antibody IgG conjugated with Rhodamine. (d) Nuclei were counter-stained by DAPI. (e) Merged image of (b–d)

Full figure and legend (84K)

To confirm the nuclear localization of these proteins, we cotransfected COS7 cells with pcDNA-myc-CLUAP1 and p3XFLAG-Clusterin, and examined their subcellular localization by immunocytochemical staining. Staining with anti-myc antibody revealed that the tagged CLUAP1 protein localized in the nucleus, while that with anti-FLAG antibody demonstrated that the tagged nCLU was also in the nucleus (Figure 3b–d). Cotransfection with both pcDNA-myc-CLUAP1 and p3XFLAG-Clusterin and double staining with the antibodies revealed colocalization of these proteins in the nucleus (Figure 3e), supporting the view that CLUAP1 and nCLU interact in the cells.

We performed cell-cycle analysis and immunoblot analysis of CLUAL1 using HCT116 cells synchronized at the G₁ phase by aphidicolin treatment. After withdrawal of aphidicolin, the cells entered S phase around 2–4 h, G₂/M at 6–8 h and completed their cycle in 12 h. Western blot showed the expression of CLUAP1 increased from G₁ phase, peaked at S phase and gradually decreased after G₂M  phase (Figure 4a).

Figure 4.

(a) Cell-cycle-dependent expression of CLUAP1. To examine whether CLUAP1 has a role in cell-cycle progression, HCT116 cells were growth arrested in G₁ phase by incubation with 5 g/ml aphidicolin (Sigma Chemical Co.) for 24 h and released from G₁ by washing five times with PBS. Cells were collected by trypsinization at 2-hour intervals for a total of 12 h and fixed in 70% cold ethanol. Cells treated with RNase and propidium iodide (50 g/ml) in PBS were analysed by FACS analysis (Becton Dickinson, San Jose, CA, USA). Western blot was also carried out at the same time points to see if the expression levels of CLUAP1 altered during the cell cycle. Time after the withdrawal of aphidicolin is indicated at the top. (b) Effect of psiH1BX-CLUAP1 on the expression of CLUAP1 in SNU-C4 cells. We constructed plasmid vector, psiH1BX, expressing siRNA as described elsewhere (Shimokawa et al., 2003). Plasmids expressing CLUAP1-siRNAs were prepared by the cloning of double-stranded oligonucleotides into psiH1BX vector. The oligonucleotides used for CLUAP1-siRNA (psiH1BX-CLUAP1) were 5'-TCCCGACCATCATAGGATGGAGCTTCAAGAGAGCTCCATC
CTATGATGGTC-3' and 5'-TTTTGACCATCATAGGATGGAGCTCTCTTGAAGCTCCATC
CTATGATGGTC-3'. psiH1BX-CLUAP1, psiH1BX-EGFP, or psiH1BX-mock plasmids were transfected into SNU-C4 cells, using FuGENE6 reagent (Roche) or Nucleofector reagent (Amaxa) according to the supplier's recommendations. Total RNA was extracted from the cells 48 h after the transfection. Semiquantitative RT–PCR was carried out using RNA extracted from cells transfected with psiH1BX-mock, psiH1BX-EGFP, or psiH1BX-CLUAP1. (c) MTT assay showing the effect of psiH1BX-CLUAP1 on the viability of SNU-C4 cells. Cells were cultured in the presence of 400–800 g/ml geneticin (G418) for 14 days before the assay was carried out. The data were subjected to analysis of variance (ANOVA) and the Scheffé's F-test

Full figure and legend (93K)

To investigate the function of CLUAP1 in cancer cells, we constructed plasmids expressing CLUAP1-short interfering RNA (siRNAs) and examined their effect on CLUAP1 expression. Transfection of SNU-C4 or HCT116 colon cancer cells with psiH1BX-CLUAP1, psiH1BX-EGFP or psiH1BX-mock revealed that psiH1BX-CLUAP1 significantly suppressed the expression of CLUAP1 in the cells compared to psiH1BX-EGFP or psiH1BX-mock (Figure 4b, and data not shown for HCT116). To test whether the suppression of CLUAP1 may affect the growth of colon cancer cells, we transfected these cells with psiH1BX-CLUAP1 or psiH1BX-mock. Viable cells transfected with psiH1BX-CLUAP1 were markedly reduced compared to those transfected with psiH1BX-mock, suggesting that decreased expression of CLUAP1 suppressed growth of SNU-C4 cells as well as that of HCT116 cells (Figure 4c, and data not shown for HCT116).

In this paper, we described the isolation and characterization of a novel gene, CLUAP1, whose expression was significantly increased in colon cancer tissues compared to the corresponding noncancerous mucosa. The high identities of the putative CLUAP1 protein with its homologues in other species suggest that CLUAP1 may play a crucial role in the cell. In addition to the coiled-coil domain between amino acids 195 and 267, it contained a conserved N-terminal region, which may be essential for the function of this protein.

Although cytogenetic analysis of colorectal tumors using comparative genomic hybridization have so far identified a set of amplified chromosomal regions including 3q, 5p, 6p, 7p, 7q, 8q, 12p, 13q, 17q, 19p, 19q, 20q, and X (Nakao et al., 1998; De Angelis et al., 1999; Korn et al., 1999; He et al., 2003), amplification of chromosome 16p has not been reported to our knowledge. On the contrary, frequent LOHs were shown in the chromosomal region of 16p13. Therefore, its elevation may not be a result from genetic alterations, but from other mechanisms such as transcriptional activation or enhanced RNA stability. Future studies concerning the mechanism(s) of its up-regulation including promoter assay are necessary.

The identification of nCLU as a CLUAP1-interacting protein suggested a possible involvement of CLUAP1 in signal transduction of growth and/or survival. CLU was originally identified as SP-40,40, a secreted glycoprotein in ram rete testis enhancing cell aggregation in vitro (Murphy et al., 1988). It is also known as complement lysis inhibitor, testosterone-repressed prostate message-2, sulfated glycoprotein 2 or apolipoprotein J. CLU is reportedly implicated in diverse physiological processes such as sperm maturation, lipid transportation, complement inhibition, tissue remodeling, membrane recycling, cell adhesion, stress responses, senescence, proliferation and apoptosis (Trougakos and Gonos, 2002). However, its precise role and regulation remain poorly understood. Nuclear CLU was first identified as a Ku70 associating protein and it was shown to act as a prodeath signal inhibiting cell growth and survival (Yang et al., 1999, 2000). Only nCLU proteins with both the C-terminal coiled-coil domain and a functional NLS could kill MCF7 breast cancer cells, and alterations in, or deletions of, the NLS or coiled-coil domain was reported to abrogate cell death (Leskov et al., 2003). Although we tested whether transduction of CLUAP1 interfered with nCLU-induced apoptosis in MCF7 cells or not, FACS analysis detected unchanged sub-G1 population between cells transfected with nCLU alone and those transfected with nCLU and CLUAP1, and similar results were observed by colony formation assay (data not shown). Hence, elevated expression of CLUAP1 may not affect the apoptotic activity, but modulate other functions of nCLU. Notably, since codons between 256 and 448 in nCLU are responsible for its interaction with Ku70, CLUAP1 may affect the association and result in altered DNA repairing capacity in cancer cells. Alternatively, sharing structural and functional similarities with heat-shock proteins, CLU may affect the function of CLUAP1 through formation of aggresomes. In agreement with this notion, a recent report has demonstrated that nCLU can indeed initiate aggresome formation and this led to a severe disruption of the mitochondrial distribution pattern (Debure et al., 2003). Expression of CLU is reportedly enhanced in sex hormone-induced prostate carcinogenesis (Ouyang et al., 2001), ovarian cancer (Hough et al., 2001), lymphomas (Wellmann et al., 2000), and breast tumor (Redondo et al., 2000). Another recent report has revealed that CLU was also upregulated in murine intestinal neoplasias and human colorectal tumors (Chen et al., 2003). Therefore, elevated expression of CLUAP1 and CLU may additively or synergistically enhance their function and play a significant role in colon tumors.

To clarify the role of CLUAP1 in cancer cells, we constructed siRNA designed to suppress the expression of CLUAP1. Consequently, the suppression of CLUAP1 resulted in the decrease in colony number of cells, suggesting that decreased levels of CLUAP1 lead to growth suppression of the cancer cell lines tested. Consistently, expression of CLUAP1 was induced in S phase of cell-cycle progression, suggesting that its elevated expression was relevant to cellular proliferation. Since its expression was greatly enhanced in more than half of the colon cancers we examined, our results raised a hypothesis that activated CLUAP1 may lead to neoplasms by constitutively accelerating proliferation of colonic epithelial cells. This notion prompted us to examine oncogenic activity of CLUAP1 by colony formation assay, which resulted in unaltered colony-forming activity. Hence, CLUAP1 may exert an essential role for proliferation in collaboration with unidentified factor(s) or for maintenance of proliferating cells. Although the precise molecular mechanism by which enhanced expression of CLUAP1 is involved in cell growth needs to be clarified, CLUAP1 may be a good candidate for the development of effective drugs to treat colorectal tumors, because downregulated expression of CLUAP1 suppressed the growth of colon cancer cells in our experiments.