- Email: [email protected]
Identification of CGI-121, a novel PRPK (p53-related protein kinase)-binding protein
BBRC Biochemical and Biophysical Research Communications 303 (2003) 399–405 www.elsevier.com/locate/ybbrc
Identification of CGI-121, a novel PRPK (p53-related protein kinase)-binding proteinq,qq Akifumi Miyoshi,a,b Katsumi Kito,a,* Takayoshi Aramoto,a Yasuhito Abe,a Nobuaki Kobayashi,b and Norifumi Uedaa a
First Department of Pathology, Ehime University School of Medicine, Shitsukawa, Shigenobucho, Onsengun, Ehime 791-0295, Japan First Department of Surgery, Ehime University School of Medicine, Shitsukawa, Shigenobucho, Onsengun, Ehime 791-0295, Japan
b
Received 20 February 2003
Abstract PRPK (p53-related protein kinase) has been reported as a novel protein kinase which binds to the tumor suppressor protein p53 and induces phosphorylation of p53 at Ser 15. To identify novel binding partners of PRPK, we performed a yeast two-hybrid screening and isolated an expressed sequence tag CGI-121 by which a 20-kDa protein was encoded. We demonstrated the protein– protein interaction of CGI-121 with PRPK in vivo and in vitro. The protein expression of CGI-121 was observed in many cell lines and was immunocytochemically identified in both the nucleus and cytosol. Although PRPK interacted with both CGI-121 and p53, several attempts to demonstrate an association between CGI-121 and p53 were unsuccessful. In addition, coprecipitation of p53 using recombinant PRPK was inhibited by adding recombinant CGI-121 in vitro, suggesting that CGI-121 could act as a potent inhibitor of the binding of PRPK to p53. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: PRPK; CGI-121; p53; Two-hybrid; Protein kinase; Protein interaction
Activation of the tumor suppressor protein p53 in response to cellular damage is accompanied by stabilization of the p53 protein [1]. The p53 protein undergoes extensive post-translational modification to escape degradation and carry out its functions [2]. MDM2 is known as an E3 ubiquitin ligase for p53, and interaction of p53 with MDM2 promotes its rapid degradation through the ubiquitin–proteasome system [3–5]. This interaction between MDM2 and p53 can be impaired by phosphorylation of p53 within the MDM2 binding region [6]. N-terminal phosphorylation of p53 may induce q Abbreviations: DTT; dithiothreitol; FCS, fetal bovine serum; IPTG, isopropyl b-D -1-thiogalactopyranoside; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; EGF, epidermal growth factor; TGF, transforming growth factor; IFN, interferon; PMA, phorbol myristate acetate. qq The nucleotide sequence(s) reported in this paper has been submitted to the GenBank/EBI/DDBJ Data Bank with Accession Nos. AY157986 and AY157987. * Corresponding author. Fax: +81-89-960-5267. E-mail address: [email protected] (K. Kito).
accumulation of p53 by directly inhibiting the N-terminal nuclear export sequence, and by inhibiting MDM2 binding and thereby reducing ubiquitination [7,8]. Specific phosphorylation sites have been identified in the N-terminal region which function primarily in the transcriptional control activity of p53. The amino terminal domain undergoes phosphorylation by several kinases including casein kinase I [9], checkpoint kinases 1 and 2 [10–12], DNA-dependent protein kinase [13], ataxia telangiectasia mutated (ATM) [14,15], Jun kinase [16,17], and mitogen-activated protein kinases [18]. The protein product of Saccharomyces cerevisiae YGR262c gene, piD261, is known to be a putative protein kinase (GenBank Accession No. CAA69084), and the disruption of this gene has been shown to cause severe growth retardation in yeast [19]. More recently, a human homologue of piD261 designated as PRPK (p53related protein kinase; GenBank Accession No. AB017505), which showed 32% identity and 64% similarity to the yeast piD261, has been isolated from an interleukin-2 activated cytotoxic T-cell subtraction
0006-291X/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0006-291X(03)00333-4
400
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
library [20]. We have previously demonstrated that recombinant PRPK protein binds to p53 and preactivated recombinant PRPK phosphorylated not only casein but also p53 protein at Ser 15 [20], however, the biological function of PRPK remains elusive. To identify cellular proteins that are involved in the PRPK-signaling pathway, a yeast two-hybrid screening was performed using PRPK as the bait. As a result, we isolated and characterized a novel PRPK-binding protein, which was referred to as CGI-121.
Materials and methods Yeast two-hybrid assay. For the library screening, an entire coding region of PRPK was inserted into a DNA-binding domain plasmid pGBT9 (Clontech). A yeast strain, HF7c, was transformed with pGBT9/PRPK using a lithium acetate method according to the manufacturerÕs protocol. The HF7c clone carrying pGBT9/PRPK was sequentially transformed with 50 lg of human testis cDNA, which fused to a pACT2 GAL4 DNA-activating domain plasmid (Clontech, Tokyo, Japan). The transformed HF7c cells were cultured for 3 days at 30 °C on SD/-Trp/-Leu/-His (TDO) plates. The positive colonies were replated and assayed for b-galactosidase activity. Escherichia coli KC8 (Clontech) was transformed with the pACT2 plasmid extracted using a Zymoprep kit (Zymo Research, Orange, CA) and positive clones were selected on M9 minimal/-Leu/ ampicillin plates. The recovered plasmids were sequenced by an automated fluorescent sequencing analyzer (Applied Biosystems, Tokyo, Japan). To demonstrate protein–protein interactions, AH109 cells were transformed with both pGBT9 and pGAD424 constructs and were selected on TDO plates. Prediction of structural properties and multiple alignment. The presence of structural features of CGI-121 protein was examined using the following databases: Prosite (http://www.expasy.org/prosite), Pfam (http://www.sanger.ac.uk/Software/Pfam), InterPro (http://www.ebi.ac.uk/interpro), and PEST (http://www.icnet.uk/LRITu/projects/pest). A sequence similarity search was performed using the BLAST 2.0 database (http://www.ncbi.nlm.nih.gov/BLAST). Multiple sequence alignment of homologous proteins was done using MultAlin program [21]. Expression and assay of recombinant proteins. Plasmids for the expression of glutathione S-transferase (GST)-fused and maltose binding protein (MBP)-fused proteins in bacteria, pGEX-6P-2 (Amersham BioSciences, Tokyo, Japan) and pMAL-c2 (New England Biolabs, Beverly, MA), were employed, respectively. The purification of recombinant proteins was performed as described previously [22]. Briefly, fusion proteins were induced in E. coli BL21(DE3) that had been transformed with pGEX-6P-2 or pMAL-c2 constructs by adding 0.1 mM IPTG. Bacteria were lysed in lysis buffer containing completemini protease inhibitors (Roche Diagnostics, Tokyo, Japan) and the extracts were incubated with glutathione–Sepharose 4B (Amersham) or Amylose resin (New England Biolabs) at 4 °C for 3 h. The GST-CGI121 and MBP-PRPK proteins were eluted with 10 mM reduced-form glutathione and 10 mM maltose, respectively. To obtain the CGI-121 moiety, GST-CGI-121 was cleaved with PreScission protease (Amersham) at 5 °C for 16 h. For the in vitro binding assay, CGI-121 was incubated with amylose beads coated with MBP or MBP-PRPK in NETN containing 1 mM DTT and 1 mM ATP at 30 °C for 1 h and then the beads were washed thoroughly. Antibodies and Western blot analysis. Rabbit polyclonal anti-HA (Y-11) and anti-myc (A14) antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) were employed for the immunoprecipitation assay. For the immunoblot, anti-HA (16B12; Covance Research Products, Richmond, CA), anti-myc (9E10; Santa Cruz), anti-p53 (DO-7;
DAKO, Kyoto, Japan), anti-phospho-p53 Ser 15 (Cell Signaling Technology, Beverly, MA), and anti-a-tubulin (B-5-1-2; Sigma Chemical, St. Louis, MO) antibodies were used. As secondary antibodies, horseradish peroxidase (HRP)-conjugated antibodies against mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) or rabbit IgG (Santa Cruz) were used. Rabbit anti-CGI-121 antisera were generated by immunization with a GST-fused polypeptide corresponding to amino acids 111–175 of CGI-121. The anti-CGI-121 antibody was separated by Protein-G–Sepharose gels (Amersham) and was further purified using an NHS-coupled CGI-121 column (Amersham). The Western blot analysis was performed according to an ECL kit protocol (Amersham). Cell lines, transfection, and RT-PCR. We purchased the following human cell lines from American Type Culture Collection (Manassas, VA): BurkittÕs lymphoma Daudi, T-cell leukemia Jurkat, chronic myelogenous leukemia K562, promyelocytic leukemia HL60, acute myelogenous leukemia KG-1, epidermoid carcinoma A431, colon adenocarcinoma HT-29, prostatic adenocarcinoma LNCaP, osteosarcoma U2OS, renal cell carcinoma 786-0, and cervical adenocarcinoma HeLa. All cell lines were maintained in DulbeccoÕs modified EagleÕs medium supplemented with 10% FCS and antibiotics. Epitope-tagged CGI-121 or PRPK proteins were expressed in human embryonic kidney (HEK) 293T cells using pcDNA3 constructs. Transfection was performed using LipofectAMINE reagent (Invitrogen) and the cells were harvested after 24 h of transfection. For RT-PCR, cDNA was synthesized with SuperScript reverse transcriptase (Invitrogen) from total RNA, which was extracted from the cell lines using Isogen reagent (Toyobo, Tokyo, Japan). Oligonucleotides of 50 - GTTAGGA TCC ATGCAGTTAA CACATCAGCT GG -30 and 50 -CCGGGAAT TC TCATAAAACATCTTTTGTTG AC -30 were used for CGI-121 primers. Immunoprecipitaion. The 293T cells transfected with HA and/or myc-tagged pcDNA3 constructs were lysed on ice in 1 ml NETN containing complete-mini protease inhibitors (Roche). Extracts were centrifuged at 4 °C for 30 min at 100,000g. Supernatants were incubated with 20 ll of a 50% slurry of protein-G–Sepharose (Amersham) and 0.6 lg of anti-HA or anti-myc polyclonal antibody at 4 °C for 3 h. After incubation, the immunocomplex was collected by centrifugation and was extensively washed in NETN. The immunocomplex was resuspended in 2% SDS sample buffer and detected by immunoblotting using the appropriate antibodies.
Results Isolation of cDNA encoding CGI-121 By the yeast two-hybrid screening using PRPK as the bait, several cDNA fragments that encoded the same protein were isolated. The GenBank database searches yielded an identical sequence that had been deposited as an expressed sequence tag, CGI-121 (GenBank Accession No. AF151879). CGI-121 was deposited as one of the cDNAs that were identified by comparative database searches between human and Caenorhabditis elegans [23]; however, the biological functions of CGI-121 were totally unknown. The longest cDNA fragment included a coding region consisting of 528 bp (Fig. 1A). The putative ORF of the cDNA encodes 175 amino acid residues with a calculated molecular mass of 19.7 kDa. Since there was no stop codon at the site upstream from the first ATG, we repeatedly carried out 50 -RACE. This
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
401
Fig. 1. Structure of CGI-121. (A) The nucleotide and deduced amino acid sequences of CGI-121. (B) Diagram of the predicted CGI-121 gene and the transcripts of CGI-121 and its variants. The BAC clone RP11-434P11 (GenBank Accession No. AC092653) contained the CGI-121 sequence. The two splice regions indicated by a grey shaded box resulted in the CGI-121 variants. (C) Comparison of human CGI-121 with its homologues in other species. Mouse CGI-121, Mus musculus (GenBank Accession No. XM132697); W03F8.4, Caenorhabditis elegans (NM068334); hypothetical protein YML036W, Saccharomyces cerevisiae (NP013676); and hypothetical protein SPCC24B10.12, Schizosaccharomyces pompe (NP588015). Residues invariant or conservative sequence are boxed or printed on grey shaded boxes, respectively.
assay revealed an additional 39 bp upstream of the 50 non-coding sequence. This candidate for the start ATG was not matched to a KozakÕs consensus sequence and no upstream stop codon was identified. Nevertheless, we concluded that this was the authentic ORF of CGI-121, because endogenous protein expression of CGI-121 was confirmed as being consistent with the molecular size calculated from the putative ATG. Further library screening based on PCR using primers designed at both ends of the CGI-121 ORF allowed the isolation of two alternative splicing variants of CGI-121. We designated the isoforms as CGI-121-L1 and CGI121-S1, which consisted of 214 and 142 amino acid residues, respectively (GenBank Accession Nos. AY157986 and AY157987). A few additional splicing variants that were faintly expressed seemed to be present (Fig. 2B); however, they have not yet been isolated. The exon–intron structure of the CGI-121 gene could be hypothesized based on the human genomic database. It is possible that CGI-121 and its variants were generated from the gene that was composed of at least six exons spanning 7.4 kb, tentatively referred to as exons 1–6. According to the structure described in Fig. 1B, the start ATG was positioned in exon 2. The coding region of
CGI-121-L1, which is the longest variant, appeared to consist of five exons (exons 2–6). The coding region of CGI-121 was composed of four exons, with the exception of exon 3 (117 bp). CGI-121-S1, which was the shortest variant, was the result of a transcript in which both exon 3 (117 bp) and 216 bp at 30 -terminal of exon 2 were skipped. We determined the chromosomal location of the CGI-121 gene using fluorescent in situ hybridization. Specific signals of CGI-121 were determined on chromosome 2p12-13 (data not shown). Based on the database searches, CGI-121 appeared to be a conserved protein, from the yeast to human (Fig. 1C). Mouse CGI121 protein shows 85% identity and 93% similarity to human CGI-121. A hypothetical protein in S. cerevisiae, YML036W, which consisted of 189 amino acid residues, showed 34% identity and 53% similarity to human CGI121. However, no significant motifs suggesting biological function were identified by searching the databases listed in the Materials and methods. CGI-121 protein widely expressed in human cell lines Western blot analysis of the total cell extracts revealed that a purified anti-CGI-121 antibody
402
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
Fig. 2. Expression of CGI-121 in various human cell lines. (A) In human cell lines, endogenous CGI-121 protein was widely expressed. However, protein expression corresponding to the splicing variants, including CGI-121-L1 and CGI-121-S1, could not be detected (upper panel). The membrane was reprobed with anti-a-tubulin antibody as an internal control (lower panel). (B) RT-PCR analysis for CGI-121. Several PCR products corresponding to the splicing variants were detected in all of the cell lines described above. The most prominent band at 0.53 kb corresponded to transcripts of CGI-121 (asterisk). The PCR products that migrated around 0.65 and 0.45 kb were identified as the transcripts for CGI-121-L1 (arrow) and CGI-121-S1 (arrowhead), respectively. Other bands have not yet been identified.
recognized a single band at 20 kDa. In 11 human cell lines, protein expression of CGI-121 was widely observed (Fig. 2A). The expression of the CGI-121 at 20 kDa was prominent, whereas no protein expression was observed that corresponded to the variants including CGI-121-L1 and CGI-121-S1, the calculated molecular masses of which were 23.9 and 16.1 kDa, respectively. Absorption of the antibody with the recombinant CGI-121 peptides (111–175 aa) resulted in the disappearance of the 20-kDa band (data not shown), suggesting that the 20-kDa protein was a major product of the CGI-121 transcripts. Conversely, several mRNA transcripts corresponding to the variants were detected by RT-PCR (Fig. 2B). This finding suggested that the isoforms might be translated much less efficiently or degraded much more rapidly than CGI-121. Furthermore, repeated Northern blot analysis failed to detect the messages of both CGI-121 and its variants (data not shown), thus the amount of these transcripts was regarded to be marginal. Fluorescent microscopy revealed that endogenous CGI-121 protein was evenly distributed in both the nucleus and the cytosol (data not shown). The nuclear distribution of CGI-121 seemed to be overlapped with that of PRPK; however, most cytosolic CGI-121 appeared to exist independently of PRPK.
Fig. 3. CGI-121 and PRPK interacted in vivo and in vitro. (A) Immunoprecipitation of transiently expressed HA-tagged and myc-tagged proteins. 293T cells were transfected with (a) myc-PRPK + HA-CGI121; (b) myc-CGI-121 + HA-PRPK; (c) myc-RAD52 + HA-CGI-121; (d) myc-RAD52 + HA-PRPK; and (e) myc-RAD52 + HA-RAD52. The proteins immunoprecipitated with anti-myc were blotted with anti-HA antibody. CGI-121 and PRPK proteins were coprecipitated in vivo (lanes 6 and 7). The blotting of the immunocomplex with anti-myc (lanes 1–5) showed that the immunoprecipitation had been successful. TCL blotted with anti-HA (lanes 11–15) proved that the transfection had been successful. A faint band at 30 kDa (asterisk) was regarded as the IgG light chain used for immunoprecipitation. (B) In vitro binding assay for CGI-121 and PRPK proteins. Recombinant CGI-121 interacted with MBP-PRPK, but not with MBP, as was detected using SDS–PAGE followed by CBB staining. The arrow and arrowhead indicate recombinant PRPK and CGI-121, respectively.
CGI-121 associated with PRPK in vivo and in vitro To support the results of the yeast two-hybrid screening, the interactions between CGI-121 and PRPK in mammalian cells were demonstrated using an immunoprecipitation technique. HA-tagged and myc-tagged proteins transiently expressed in 293T cells were immunoprecipitated with anti-myc antibody and then blotted with anti-HA antibody. As a control, human RAD52 protein, known to make a multimer, was employed. As shown in Fig. 3A, CGI-121 was coprecipitated with PRPK in vivo. The interaction of CGI-121 with PRPK was observed as well when the immunocomplex was precipitated with anti-HA or anti-CGI-121
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
403
antibody (data not shown). Next, to address the question regarding whether CGI-121 directly binds to PRPK, purified recombinant proteins were subjected to an in vitro binding assay. As shown in Fig. 3B, recombinant CGI-121 protein was coprecipitated with MBP-PRPK in vitro. Likewise, GST-PRPK, but not GST, was able to precipitate recombinant CGI-121 (data not shown). These data indicated that CGI-121 and PRPK could physically interact in vitro as well as in vivo. In addition, neither CGI-121 nor PRPK could form a homodimer in vivo or in vitro (data not shown). Since it has been previously noted that PRPK possesses putative kinase domains and that preactivated PRPK actually induces the phosphorylation of both casein and p53 [20], we examined whether CGI-121 underwent phosphorylation by PRPK. Several attempts to demonstrate the phosphorylation of CGI-121 protein in vitro were unsuccessful, regardless of the presence of recombinant PRPK (data not shown). CGI-121 could inhibit p53 coprecipitated with recombinant PRPK, but not with CGI-121 Since the interaction between PRPK and p53 has been demonstrated, it is of interest to know whether CGI-121 also binds to p53. For this binding assay, recombinant proteins were incubated with cell extracts from a human osteosarcoma cell line, U2-OS, in which wild-type p53 protein expression was maintained. As shown in Fig. 4A, endogenous p53 protein was precipitated with recombinant PRPK; moreover, the coprecipitated p53 was in part phosphorylated at Ser 15. In contrast, GST-CGI-121 was unable to precipitate p53 protein in vitro. In addition, an immunoprecipitation assay for demonstrating the interaction of CGI-121 with p53 in vivo was not successful (data not shown). These results suggest the possibility that CGI-121 might compete with p53 for binding to PRPK. To determine whether there was sufficient evidence of this possibility, we preincubated PRPK with an excess amount of CGI121 protein prior to the p53 pull-down assay using U2OS cell lysate. As shown in Fig. 4B, pretreatment with recombinant CGI-121 caused a decrease in the amount of p53 that was coprecipitated with PRPK. This result indicated that the interaction of PRPK with p53 might has been blocked by the excess CGI-121. As expected, the MBP-PRPK captured a large amount of CGI-121 instead of p53 (Fig. 4B, lower panel). These results suggested the possibility that overexpression of CGI-121 in vivo would inhibit the interaction between p53 and PRPK, and might also affect the phosphorylation status of p53. To elucidate the latter possibility, we overexpressed CGI-121 in U2-OS cells using transiently transfected pcDNA3/CGI-121. However, the in vivo overexpression of CGI-121 resulted in no significant alteration in the p53 phosphorylation status at Ser 15
Fig. 4. CGI-121 inhibited the interaction between PRPK and p53. (A) p53 protein precipitated with recombinant proteins was detected with the anti-p53 antibody (DO-7). The recombinant PRPK (lanes 4 and 6), but not CGI-121 (lane 3), precipitated wild-type p53 protein in U2-OS cells (upper panel). Reprobing with anti-phospho-p53 Ser 15 revealed that the p53 protein associated with PRPK was phosphorylated at Ser 15 (lower panel, lanes 4 and 6). (B) CGI-121 was able to inhibit the interaction between PRPK and p53. MBP-PRPK was preincubated with an excess amount of recombinant CGI-121, followed by incubation with U2-OS cell extracts. Western blot analysis revealed that the pretreatment with CGI-121 resulted in a decline in the p53-binding capability of PRPK (upper and middle panels, lane 2). The MBPPRPK following the treatment with recombinant CGI-121 was blotted with anti-CGI-121 antibody, showing that the treatment led to PRPK accompanied by abundant CGI-121 (lower panel, lane 2).
(data not shown). The interpretation of this result was somewhat complicated by the fact that the phosphorylation status of p53 is regulated in vivo not only by PRPK, but also by several kinases and phosphatases.
Discussion It has been reported that PRPK and its yeast orthologue piD261 function as a Ser/Thr protein kinase, despite its small size and lack of some of the conserved features of a protein kinase [19,20]. Although piD261 and PRPK exhibit close structural similarity, their biological functions and intrinsic substrates remain elusive. In this study, we have demonstrated that CGI-121 inhibited the interaction between PRPK and p53; nevertheless, we cannot rule out the possibility that PRPK did not directly phosphorylate p53 due to the fact that preincubation with COS cell lysate was necessary for the in vitro kinase assay [20]. Under stress-free conditions, the p53 level is generally kept low by rapid protein
404
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
degradation resulting from ubiquitination. Once p53 is phosphorylated at specific residues, including Ser 15, p53 might escape from degradation and carry out its transcriptional functions. In this context, CGI-121 appears to induce the inactivation and degradation of p53 by impairing the phosphorylation of p53 via interaction with PRPK. In contrast to the successful demonstration of a marked interaction between PRPK and CGI-121, the in vivo interaction between p53 and PRPK using immunoprecipitation have been detected to be marginal (data not shown). These findings indicate that most endogenous PRPK might already be accompanied by CGI-121 to maintain the inactivation and degradation of p53. If this is indeed the case, then this hypothesis appears to be consistent with the result that transient overexpression of CGI-121 causes no significant alteration of the p53 status. It is of interest to know how CGI-121 protein expression is regulated during diverse cellular events. To address this issue, we treated 293T or HL-60 cells with several biochemical stimulants; however, our tests revealed no significant alterations of CGI-121 protein expression after treatment with the following reagents: EGF; TGF-a and b; IFN-a; b, and c; PMA; retinoic acid; vitamin D3; protease/proteasome inhibitors; and UV irradiation (data not shown). Furthermore, we investigated whether CGI-121 expression was regulated in a cell cycle-dependent manner using synchronized U2OS cells; however, the total abundance of CGI-121 protein varied little throughout the cell cycle (data not shown). A feedback loop for p53 and its related molecules, in which p53 transactivated several target proteins and the activity of p53 was controlled by the induced proteins, has previously been described [24,25]. The loss of p53 function commonly results from deletion and/or dominant negative mutation of the p53 gene [26,27], and it leads to abnormal cell proliferation and immortality. Therefore, we examined the protein expression of CGI121 in several cell lines in which various kinds of mutations and/or deletions were harbored in the p53 gene [28]. As shown in Results, CGI-121 protein was expressed invariably, regardless of the p53 status in those cell lines. At the moment, we have no direct evidence regarding the mechanism by which CGI-121 expression is regulated. In conclusion, our experiments provide insight into the function of CGI-121 as a binding partner of PRPK. One key role of CGI-121 appears to be the inhibition of the interaction of PRPK with p53. The regulation of CGI-121 expression and activity remains of great further interest. Moreover, the existence of abundant CGI121 not only in the nucleus, but also in the cytosol, suggests that CGI-121 may carry out additional physiological tasks besides that of binding to PRPK. The identification of putative CGI-121-associated proteins and additional substrates of PRPK will be necessary in
order to identify and describe this novel signaling pathway.
Acknowledgment We thank Yuko Hirano for the technical assistance with this study.
References [1] A.J. Levine, p53, the cellular gatekeeper for growth and division, Cell 88 (1997) 323–331. [2] D.B. Woods, K.H. Vousden, Regulation of p53 function, Exp. Cell Res. 264 (2001) 56–66. [3] R. Honda, H. Tanaka, H. Yasuda, Oncoprotein MDM2 is a ubiquitin ligase E3 for tumor suppressor p53, FEBS Lett. 420 (1997) 25–27. [4] Y. Haupt, R. Maya, A. Kazaz, M. Oren, Mdm2 promotes the rapid degradation of p53, Nature 387 (1997) 296–299. [5] M.H. Kubbutat, S.N. Jones, K.H. Vousden, Regulation of p53 stability by Mdm2, Nature 387 (1997) 299–303. [6] S.Y. Shieh, M. Ikeda, Y. Taya, C. Prives, DNA damage-induced phosphorylation of p53 alleviates inhibition by MDM2, Cell 91 (1997) 325–334. [7] S. Nakamura, J.A. Roth, T. Mukhopadhyay, Multiple lysine mutations in the C-terminal domain of p53 interfere with MDM2dependent protein degradation and ubiquitination, Mol. Cell Biol. 20 (2000) 9391–9398. [8] M.S. Rodriguez, J.M. Desterro, S. Lain, D.P. Lane, R.T. Hay, Multiple C-terminal lysine residues target p53 for ubiquitinproteasome-mediated degradation, Mol. Cell Biol. 20 (2000) 8458–8467. [9] K. Sakaguchi, S. Saito, Y. Higashimoto, S. Roy, C.W. Anderson, E. Appella, Damage-mediated phosphorylation of human p53 threonine 18 through a cascade mediated by a casein 1-like kinase. Effect on Mdm2 binding, J. Biol. Chem. 275 (2000) 9278– 9283. [10] A. Hirao, Y.Y. Kong, S. Matsuoka, A. Wakeham, J. Ruland, H. Yoshida, D. Liu, S.J. Elledge, T.W. Mak, DNA damage-induced activation of p53 by the checkpoint kinase Chk2, Science 287 (2000) 1824–1827. [11] S.Y. Shieh, J. Ahn, K. Tamai, Y. Taya, C. Prives, The human homologs of checkpoint kinases Chk1 and Cds1 (Chk2) phosphorylate p53 at multiple DNA damage-inducible sites, Genes Dev. 14 (2000) 289–300. [12] N.H. Chehab, A. Malikzay, M. Appel, T.D. Halazonetis, Chk2/ hCds1 functions as a DNA damage checkpoint in G(1) by stabilizing p53, Genes Dev. 14 (2000) 278–288. [13] S.P. Lees-Miller, K. Sakaguchi, S.J. Ullrich, E. Appella, C.W. Anderson, Human DNA-activated protein kinase phosphorylates serines 15 and 37 in the amino-terminal transactivation domain of human p53, Mol. Cell Biol. 12 (1992) 5041–5049. [14] S. Banin, L. Moyal, S. Shieh, Y. Taya, C.W. Anderson, L. Chessa, N.I. Smorodinsky, C. Prives, Y. Reiss, Y. Shiloh, Y. Ziv, Enhanced phosphorylation of p53 by ATM in response to DNA damage, Science 281 (1998) 1674–1677. [15] C.E. Canman, D.S. Lim, K.A. Cimprich, Y. Taya, K. Tamai, K. Sakaguchi, E. Appella, M.B. Kastan, J.D. Siliciano, Activation of the ATM kinase by ionizing radiation and phosphorylation of p53, Science 281 (1998) 1677–1679. [16] D.M. Milne, L.E. Campbell, D.G. Campbell, D.W. Meek, p53 is phosphorylated in vitro and in vivo by an ultraviolet radiationinduced protein kinase characteristic of the c-Jun kinase, JNK1, J. Biol. Chem. 270 (1995) 5511–5518.
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405 [17] M.C. Hu, W.R. Qiu, Y.P. Wang, JNK1, JNK2 and JNK3 are p53 N-terminal serine 34 kinases, Oncogene 15 (1997) 2277–2287. [18] D.M. Milne, D.G. Campbell, F.B. Caudwell, D.W. Meek, Phosphorylation of the tumor suppressor protein p53 by mitogen-activated protein kinases, J. Biol. Chem. 269 (1994) 9253– 9260. [19] S. Stocchetto, O. Marin, G. Carignani, L.A. Pinna, Biochemical evidence that Saccharomyces cerevisiae YGR262c gene, required for normal growth, encodes a novel Ser/Thr-specific protein kinase, FEBS Lett. 414 (1997) 171–175. [20] Y. Abe, S. Matsumoto, S. Wei, K. Nezu, A. Miyoshi, K. Kito, N. Ueda, K. Shigemoto, Y. Hitsumoto, J. Nikawa, Y. Enomoto, Cloning and characterization of a p53-related protein kinase expressed in interleukin-2-activated cytotoxic T-cells, epithelial tumor cell lines, and the testes, J. Biol. Chem. 276 (2001) 44003– 44011. [21] F. Corpet, Multiple sequence alignment with hierarchical clustering, Nucleic Acids Res. 16 (1988) 10881–10890. [22] H. Wada, K. Kito, L.S. Caskey, E.T. Yeh, T. Kamitani, Cleavage of the C-terminus of NEDD8 by UCH-L3, Biochem. Biophys. Res. Commun. 251 (1998) 688–692.
405
[23] C.H. Lai, C.Y. Chou, L.Y. ChÕang, C.S. Liu, W. Lin, Identification of novel human genes evolutionarily conserved in Caenorhabditis elegans by comparative proteomics, Genome Res. 10 (2000) 703–713. [24] X. Wu, J.H. Bayle, D. Olson, A.J. Levine, The p53-mdm-2 autoregulatory feedback loop, Genes Dev. 7 (1993) 1126–1132. [25] K.H. Vousden, Activation of the p53 tumor suppressor protein, Biochim. Biophys. Acta 1602 (2002) 47–59. [26] B. Vogelstein, E.R. Fearon, S.R. Hamilton, S.E. Kern, A.C. Preisinger, M. Leppert, Y. Nakamura, R. White, A.M. Smits, J.L. Bos, Genetic alterations during colorectal-tumor development, N. Engl. J. Med. 319 (1988) 525–532. [27] S.J. Baker, E.R. Fearon, J.M. Nigro, S.R. Hamilton, A.C. Preisinger, J.M. Jessup, P. vanTuinen, D.H. Ledbetter, D.F. Barker, Y. Nakamura, et al., Chromosome 17 deletions and p53 gene mutations in colorectal carcinomas, Science 244 (1989) 217– 221. [28] M. Hollstein, K. Rice, M.S. Greenblatt, T. Soussi, R. Fuchs, T. Sorlie, E. Hovig, B. Smith-Sorensen, R. Montesano, C.C. Harris, Database of p53 gene somatic mutations in human tumors and cell lines, Nucleic Acids Res. 22 (1994) 3551–3555.
Identification of CGI-121, a novel PRPK (p53-related protein kinase)-binding proteinq,qq Akifumi Miyoshi,a,b Katsumi Kito,a,* Takayoshi Aramoto,a Yasuhito Abe,a Nobuaki Kobayashi,b and Norifumi Uedaa a
First Department of Pathology, Ehime University School of Medicine, Shitsukawa, Shigenobucho, Onsengun, Ehime 791-0295, Japan First Department of Surgery, Ehime University School of Medicine, Shitsukawa, Shigenobucho, Onsengun, Ehime 791-0295, Japan
b
Received 20 February 2003
Abstract PRPK (p53-related protein kinase) has been reported as a novel protein kinase which binds to the tumor suppressor protein p53 and induces phosphorylation of p53 at Ser 15. To identify novel binding partners of PRPK, we performed a yeast two-hybrid screening and isolated an expressed sequence tag CGI-121 by which a 20-kDa protein was encoded. We demonstrated the protein– protein interaction of CGI-121 with PRPK in vivo and in vitro. The protein expression of CGI-121 was observed in many cell lines and was immunocytochemically identified in both the nucleus and cytosol. Although PRPK interacted with both CGI-121 and p53, several attempts to demonstrate an association between CGI-121 and p53 were unsuccessful. In addition, coprecipitation of p53 using recombinant PRPK was inhibited by adding recombinant CGI-121 in vitro, suggesting that CGI-121 could act as a potent inhibitor of the binding of PRPK to p53. Ó 2003 Elsevier Science (USA). All rights reserved. Keywords: PRPK; CGI-121; p53; Two-hybrid; Protein kinase; Protein interaction
Activation of the tumor suppressor protein p53 in response to cellular damage is accompanied by stabilization of the p53 protein [1]. The p53 protein undergoes extensive post-translational modification to escape degradation and carry out its functions [2]. MDM2 is known as an E3 ubiquitin ligase for p53, and interaction of p53 with MDM2 promotes its rapid degradation through the ubiquitin–proteasome system [3–5]. This interaction between MDM2 and p53 can be impaired by phosphorylation of p53 within the MDM2 binding region [6]. N-terminal phosphorylation of p53 may induce q Abbreviations: DTT; dithiothreitol; FCS, fetal bovine serum; IPTG, isopropyl b-D -1-thiogalactopyranoside; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; EGF, epidermal growth factor; TGF, transforming growth factor; IFN, interferon; PMA, phorbol myristate acetate. qq The nucleotide sequence(s) reported in this paper has been submitted to the GenBank/EBI/DDBJ Data Bank with Accession Nos. AY157986 and AY157987. * Corresponding author. Fax: +81-89-960-5267. E-mail address: [email protected] (K. Kito).
accumulation of p53 by directly inhibiting the N-terminal nuclear export sequence, and by inhibiting MDM2 binding and thereby reducing ubiquitination [7,8]. Specific phosphorylation sites have been identified in the N-terminal region which function primarily in the transcriptional control activity of p53. The amino terminal domain undergoes phosphorylation by several kinases including casein kinase I [9], checkpoint kinases 1 and 2 [10–12], DNA-dependent protein kinase [13], ataxia telangiectasia mutated (ATM) [14,15], Jun kinase [16,17], and mitogen-activated protein kinases [18]. The protein product of Saccharomyces cerevisiae YGR262c gene, piD261, is known to be a putative protein kinase (GenBank Accession No. CAA69084), and the disruption of this gene has been shown to cause severe growth retardation in yeast [19]. More recently, a human homologue of piD261 designated as PRPK (p53related protein kinase; GenBank Accession No. AB017505), which showed 32% identity and 64% similarity to the yeast piD261, has been isolated from an interleukin-2 activated cytotoxic T-cell subtraction
0006-291X/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/S0006-291X(03)00333-4
400
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
library [20]. We have previously demonstrated that recombinant PRPK protein binds to p53 and preactivated recombinant PRPK phosphorylated not only casein but also p53 protein at Ser 15 [20], however, the biological function of PRPK remains elusive. To identify cellular proteins that are involved in the PRPK-signaling pathway, a yeast two-hybrid screening was performed using PRPK as the bait. As a result, we isolated and characterized a novel PRPK-binding protein, which was referred to as CGI-121.
Materials and methods Yeast two-hybrid assay. For the library screening, an entire coding region of PRPK was inserted into a DNA-binding domain plasmid pGBT9 (Clontech). A yeast strain, HF7c, was transformed with pGBT9/PRPK using a lithium acetate method according to the manufacturerÕs protocol. The HF7c clone carrying pGBT9/PRPK was sequentially transformed with 50 lg of human testis cDNA, which fused to a pACT2 GAL4 DNA-activating domain plasmid (Clontech, Tokyo, Japan). The transformed HF7c cells were cultured for 3 days at 30 °C on SD/-Trp/-Leu/-His (TDO) plates. The positive colonies were replated and assayed for b-galactosidase activity. Escherichia coli KC8 (Clontech) was transformed with the pACT2 plasmid extracted using a Zymoprep kit (Zymo Research, Orange, CA) and positive clones were selected on M9 minimal/-Leu/ ampicillin plates. The recovered plasmids were sequenced by an automated fluorescent sequencing analyzer (Applied Biosystems, Tokyo, Japan). To demonstrate protein–protein interactions, AH109 cells were transformed with both pGBT9 and pGAD424 constructs and were selected on TDO plates. Prediction of structural properties and multiple alignment. The presence of structural features of CGI-121 protein was examined using the following databases: Prosite (http://www.expasy.org/prosite), Pfam (http://www.sanger.ac.uk/Software/Pfam), InterPro (http://www.ebi.ac.uk/interpro), and PEST (http://www.icnet.uk/LRITu/projects/pest). A sequence similarity search was performed using the BLAST 2.0 database (http://www.ncbi.nlm.nih.gov/BLAST). Multiple sequence alignment of homologous proteins was done using MultAlin program [21]. Expression and assay of recombinant proteins. Plasmids for the expression of glutathione S-transferase (GST)-fused and maltose binding protein (MBP)-fused proteins in bacteria, pGEX-6P-2 (Amersham BioSciences, Tokyo, Japan) and pMAL-c2 (New England Biolabs, Beverly, MA), were employed, respectively. The purification of recombinant proteins was performed as described previously [22]. Briefly, fusion proteins were induced in E. coli BL21(DE3) that had been transformed with pGEX-6P-2 or pMAL-c2 constructs by adding 0.1 mM IPTG. Bacteria were lysed in lysis buffer containing completemini protease inhibitors (Roche Diagnostics, Tokyo, Japan) and the extracts were incubated with glutathione–Sepharose 4B (Amersham) or Amylose resin (New England Biolabs) at 4 °C for 3 h. The GST-CGI121 and MBP-PRPK proteins were eluted with 10 mM reduced-form glutathione and 10 mM maltose, respectively. To obtain the CGI-121 moiety, GST-CGI-121 was cleaved with PreScission protease (Amersham) at 5 °C for 16 h. For the in vitro binding assay, CGI-121 was incubated with amylose beads coated with MBP or MBP-PRPK in NETN containing 1 mM DTT and 1 mM ATP at 30 °C for 1 h and then the beads were washed thoroughly. Antibodies and Western blot analysis. Rabbit polyclonal anti-HA (Y-11) and anti-myc (A14) antibodies (Santa Cruz Biotechnology, Santa Cruz, CA) were employed for the immunoprecipitation assay. For the immunoblot, anti-HA (16B12; Covance Research Products, Richmond, CA), anti-myc (9E10; Santa Cruz), anti-p53 (DO-7;
DAKO, Kyoto, Japan), anti-phospho-p53 Ser 15 (Cell Signaling Technology, Beverly, MA), and anti-a-tubulin (B-5-1-2; Sigma Chemical, St. Louis, MO) antibodies were used. As secondary antibodies, horseradish peroxidase (HRP)-conjugated antibodies against mouse IgG (Jackson ImmunoResearch Laboratories, West Grove, PA) or rabbit IgG (Santa Cruz) were used. Rabbit anti-CGI-121 antisera were generated by immunization with a GST-fused polypeptide corresponding to amino acids 111–175 of CGI-121. The anti-CGI-121 antibody was separated by Protein-G–Sepharose gels (Amersham) and was further purified using an NHS-coupled CGI-121 column (Amersham). The Western blot analysis was performed according to an ECL kit protocol (Amersham). Cell lines, transfection, and RT-PCR. We purchased the following human cell lines from American Type Culture Collection (Manassas, VA): BurkittÕs lymphoma Daudi, T-cell leukemia Jurkat, chronic myelogenous leukemia K562, promyelocytic leukemia HL60, acute myelogenous leukemia KG-1, epidermoid carcinoma A431, colon adenocarcinoma HT-29, prostatic adenocarcinoma LNCaP, osteosarcoma U2OS, renal cell carcinoma 786-0, and cervical adenocarcinoma HeLa. All cell lines were maintained in DulbeccoÕs modified EagleÕs medium supplemented with 10% FCS and antibiotics. Epitope-tagged CGI-121 or PRPK proteins were expressed in human embryonic kidney (HEK) 293T cells using pcDNA3 constructs. Transfection was performed using LipofectAMINE reagent (Invitrogen) and the cells were harvested after 24 h of transfection. For RT-PCR, cDNA was synthesized with SuperScript reverse transcriptase (Invitrogen) from total RNA, which was extracted from the cell lines using Isogen reagent (Toyobo, Tokyo, Japan). Oligonucleotides of 50 - GTTAGGA TCC ATGCAGTTAA CACATCAGCT GG -30 and 50 -CCGGGAAT TC TCATAAAACATCTTTTGTTG AC -30 were used for CGI-121 primers. Immunoprecipitaion. The 293T cells transfected with HA and/or myc-tagged pcDNA3 constructs were lysed on ice in 1 ml NETN containing complete-mini protease inhibitors (Roche). Extracts were centrifuged at 4 °C for 30 min at 100,000g. Supernatants were incubated with 20 ll of a 50% slurry of protein-G–Sepharose (Amersham) and 0.6 lg of anti-HA or anti-myc polyclonal antibody at 4 °C for 3 h. After incubation, the immunocomplex was collected by centrifugation and was extensively washed in NETN. The immunocomplex was resuspended in 2% SDS sample buffer and detected by immunoblotting using the appropriate antibodies.
Results Isolation of cDNA encoding CGI-121 By the yeast two-hybrid screening using PRPK as the bait, several cDNA fragments that encoded the same protein were isolated. The GenBank database searches yielded an identical sequence that had been deposited as an expressed sequence tag, CGI-121 (GenBank Accession No. AF151879). CGI-121 was deposited as one of the cDNAs that were identified by comparative database searches between human and Caenorhabditis elegans [23]; however, the biological functions of CGI-121 were totally unknown. The longest cDNA fragment included a coding region consisting of 528 bp (Fig. 1A). The putative ORF of the cDNA encodes 175 amino acid residues with a calculated molecular mass of 19.7 kDa. Since there was no stop codon at the site upstream from the first ATG, we repeatedly carried out 50 -RACE. This
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
401
Fig. 1. Structure of CGI-121. (A) The nucleotide and deduced amino acid sequences of CGI-121. (B) Diagram of the predicted CGI-121 gene and the transcripts of CGI-121 and its variants. The BAC clone RP11-434P11 (GenBank Accession No. AC092653) contained the CGI-121 sequence. The two splice regions indicated by a grey shaded box resulted in the CGI-121 variants. (C) Comparison of human CGI-121 with its homologues in other species. Mouse CGI-121, Mus musculus (GenBank Accession No. XM132697); W03F8.4, Caenorhabditis elegans (NM068334); hypothetical protein YML036W, Saccharomyces cerevisiae (NP013676); and hypothetical protein SPCC24B10.12, Schizosaccharomyces pompe (NP588015). Residues invariant or conservative sequence are boxed or printed on grey shaded boxes, respectively.
assay revealed an additional 39 bp upstream of the 50 non-coding sequence. This candidate for the start ATG was not matched to a KozakÕs consensus sequence and no upstream stop codon was identified. Nevertheless, we concluded that this was the authentic ORF of CGI-121, because endogenous protein expression of CGI-121 was confirmed as being consistent with the molecular size calculated from the putative ATG. Further library screening based on PCR using primers designed at both ends of the CGI-121 ORF allowed the isolation of two alternative splicing variants of CGI-121. We designated the isoforms as CGI-121-L1 and CGI121-S1, which consisted of 214 and 142 amino acid residues, respectively (GenBank Accession Nos. AY157986 and AY157987). A few additional splicing variants that were faintly expressed seemed to be present (Fig. 2B); however, they have not yet been isolated. The exon–intron structure of the CGI-121 gene could be hypothesized based on the human genomic database. It is possible that CGI-121 and its variants were generated from the gene that was composed of at least six exons spanning 7.4 kb, tentatively referred to as exons 1–6. According to the structure described in Fig. 1B, the start ATG was positioned in exon 2. The coding region of
CGI-121-L1, which is the longest variant, appeared to consist of five exons (exons 2–6). The coding region of CGI-121 was composed of four exons, with the exception of exon 3 (117 bp). CGI-121-S1, which was the shortest variant, was the result of a transcript in which both exon 3 (117 bp) and 216 bp at 30 -terminal of exon 2 were skipped. We determined the chromosomal location of the CGI-121 gene using fluorescent in situ hybridization. Specific signals of CGI-121 were determined on chromosome 2p12-13 (data not shown). Based on the database searches, CGI-121 appeared to be a conserved protein, from the yeast to human (Fig. 1C). Mouse CGI121 protein shows 85% identity and 93% similarity to human CGI-121. A hypothetical protein in S. cerevisiae, YML036W, which consisted of 189 amino acid residues, showed 34% identity and 53% similarity to human CGI121. However, no significant motifs suggesting biological function were identified by searching the databases listed in the Materials and methods. CGI-121 protein widely expressed in human cell lines Western blot analysis of the total cell extracts revealed that a purified anti-CGI-121 antibody
402
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
Fig. 2. Expression of CGI-121 in various human cell lines. (A) In human cell lines, endogenous CGI-121 protein was widely expressed. However, protein expression corresponding to the splicing variants, including CGI-121-L1 and CGI-121-S1, could not be detected (upper panel). The membrane was reprobed with anti-a-tubulin antibody as an internal control (lower panel). (B) RT-PCR analysis for CGI-121. Several PCR products corresponding to the splicing variants were detected in all of the cell lines described above. The most prominent band at 0.53 kb corresponded to transcripts of CGI-121 (asterisk). The PCR products that migrated around 0.65 and 0.45 kb were identified as the transcripts for CGI-121-L1 (arrow) and CGI-121-S1 (arrowhead), respectively. Other bands have not yet been identified.
recognized a single band at 20 kDa. In 11 human cell lines, protein expression of CGI-121 was widely observed (Fig. 2A). The expression of the CGI-121 at 20 kDa was prominent, whereas no protein expression was observed that corresponded to the variants including CGI-121-L1 and CGI-121-S1, the calculated molecular masses of which were 23.9 and 16.1 kDa, respectively. Absorption of the antibody with the recombinant CGI-121 peptides (111–175 aa) resulted in the disappearance of the 20-kDa band (data not shown), suggesting that the 20-kDa protein was a major product of the CGI-121 transcripts. Conversely, several mRNA transcripts corresponding to the variants were detected by RT-PCR (Fig. 2B). This finding suggested that the isoforms might be translated much less efficiently or degraded much more rapidly than CGI-121. Furthermore, repeated Northern blot analysis failed to detect the messages of both CGI-121 and its variants (data not shown), thus the amount of these transcripts was regarded to be marginal. Fluorescent microscopy revealed that endogenous CGI-121 protein was evenly distributed in both the nucleus and the cytosol (data not shown). The nuclear distribution of CGI-121 seemed to be overlapped with that of PRPK; however, most cytosolic CGI-121 appeared to exist independently of PRPK.
Fig. 3. CGI-121 and PRPK interacted in vivo and in vitro. (A) Immunoprecipitation of transiently expressed HA-tagged and myc-tagged proteins. 293T cells were transfected with (a) myc-PRPK + HA-CGI121; (b) myc-CGI-121 + HA-PRPK; (c) myc-RAD52 + HA-CGI-121; (d) myc-RAD52 + HA-PRPK; and (e) myc-RAD52 + HA-RAD52. The proteins immunoprecipitated with anti-myc were blotted with anti-HA antibody. CGI-121 and PRPK proteins were coprecipitated in vivo (lanes 6 and 7). The blotting of the immunocomplex with anti-myc (lanes 1–5) showed that the immunoprecipitation had been successful. TCL blotted with anti-HA (lanes 11–15) proved that the transfection had been successful. A faint band at 30 kDa (asterisk) was regarded as the IgG light chain used for immunoprecipitation. (B) In vitro binding assay for CGI-121 and PRPK proteins. Recombinant CGI-121 interacted with MBP-PRPK, but not with MBP, as was detected using SDS–PAGE followed by CBB staining. The arrow and arrowhead indicate recombinant PRPK and CGI-121, respectively.
CGI-121 associated with PRPK in vivo and in vitro To support the results of the yeast two-hybrid screening, the interactions between CGI-121 and PRPK in mammalian cells were demonstrated using an immunoprecipitation technique. HA-tagged and myc-tagged proteins transiently expressed in 293T cells were immunoprecipitated with anti-myc antibody and then blotted with anti-HA antibody. As a control, human RAD52 protein, known to make a multimer, was employed. As shown in Fig. 3A, CGI-121 was coprecipitated with PRPK in vivo. The interaction of CGI-121 with PRPK was observed as well when the immunocomplex was precipitated with anti-HA or anti-CGI-121
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
403
antibody (data not shown). Next, to address the question regarding whether CGI-121 directly binds to PRPK, purified recombinant proteins were subjected to an in vitro binding assay. As shown in Fig. 3B, recombinant CGI-121 protein was coprecipitated with MBP-PRPK in vitro. Likewise, GST-PRPK, but not GST, was able to precipitate recombinant CGI-121 (data not shown). These data indicated that CGI-121 and PRPK could physically interact in vitro as well as in vivo. In addition, neither CGI-121 nor PRPK could form a homodimer in vivo or in vitro (data not shown). Since it has been previously noted that PRPK possesses putative kinase domains and that preactivated PRPK actually induces the phosphorylation of both casein and p53 [20], we examined whether CGI-121 underwent phosphorylation by PRPK. Several attempts to demonstrate the phosphorylation of CGI-121 protein in vitro were unsuccessful, regardless of the presence of recombinant PRPK (data not shown). CGI-121 could inhibit p53 coprecipitated with recombinant PRPK, but not with CGI-121 Since the interaction between PRPK and p53 has been demonstrated, it is of interest to know whether CGI-121 also binds to p53. For this binding assay, recombinant proteins were incubated with cell extracts from a human osteosarcoma cell line, U2-OS, in which wild-type p53 protein expression was maintained. As shown in Fig. 4A, endogenous p53 protein was precipitated with recombinant PRPK; moreover, the coprecipitated p53 was in part phosphorylated at Ser 15. In contrast, GST-CGI-121 was unable to precipitate p53 protein in vitro. In addition, an immunoprecipitation assay for demonstrating the interaction of CGI-121 with p53 in vivo was not successful (data not shown). These results suggest the possibility that CGI-121 might compete with p53 for binding to PRPK. To determine whether there was sufficient evidence of this possibility, we preincubated PRPK with an excess amount of CGI121 protein prior to the p53 pull-down assay using U2OS cell lysate. As shown in Fig. 4B, pretreatment with recombinant CGI-121 caused a decrease in the amount of p53 that was coprecipitated with PRPK. This result indicated that the interaction of PRPK with p53 might has been blocked by the excess CGI-121. As expected, the MBP-PRPK captured a large amount of CGI-121 instead of p53 (Fig. 4B, lower panel). These results suggested the possibility that overexpression of CGI-121 in vivo would inhibit the interaction between p53 and PRPK, and might also affect the phosphorylation status of p53. To elucidate the latter possibility, we overexpressed CGI-121 in U2-OS cells using transiently transfected pcDNA3/CGI-121. However, the in vivo overexpression of CGI-121 resulted in no significant alteration in the p53 phosphorylation status at Ser 15
Fig. 4. CGI-121 inhibited the interaction between PRPK and p53. (A) p53 protein precipitated with recombinant proteins was detected with the anti-p53 antibody (DO-7). The recombinant PRPK (lanes 4 and 6), but not CGI-121 (lane 3), precipitated wild-type p53 protein in U2-OS cells (upper panel). Reprobing with anti-phospho-p53 Ser 15 revealed that the p53 protein associated with PRPK was phosphorylated at Ser 15 (lower panel, lanes 4 and 6). (B) CGI-121 was able to inhibit the interaction between PRPK and p53. MBP-PRPK was preincubated with an excess amount of recombinant CGI-121, followed by incubation with U2-OS cell extracts. Western blot analysis revealed that the pretreatment with CGI-121 resulted in a decline in the p53-binding capability of PRPK (upper and middle panels, lane 2). The MBPPRPK following the treatment with recombinant CGI-121 was blotted with anti-CGI-121 antibody, showing that the treatment led to PRPK accompanied by abundant CGI-121 (lower panel, lane 2).
(data not shown). The interpretation of this result was somewhat complicated by the fact that the phosphorylation status of p53 is regulated in vivo not only by PRPK, but also by several kinases and phosphatases.
Discussion It has been reported that PRPK and its yeast orthologue piD261 function as a Ser/Thr protein kinase, despite its small size and lack of some of the conserved features of a protein kinase [19,20]. Although piD261 and PRPK exhibit close structural similarity, their biological functions and intrinsic substrates remain elusive. In this study, we have demonstrated that CGI-121 inhibited the interaction between PRPK and p53; nevertheless, we cannot rule out the possibility that PRPK did not directly phosphorylate p53 due to the fact that preincubation with COS cell lysate was necessary for the in vitro kinase assay [20]. Under stress-free conditions, the p53 level is generally kept low by rapid protein
404
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405
degradation resulting from ubiquitination. Once p53 is phosphorylated at specific residues, including Ser 15, p53 might escape from degradation and carry out its transcriptional functions. In this context, CGI-121 appears to induce the inactivation and degradation of p53 by impairing the phosphorylation of p53 via interaction with PRPK. In contrast to the successful demonstration of a marked interaction between PRPK and CGI-121, the in vivo interaction between p53 and PRPK using immunoprecipitation have been detected to be marginal (data not shown). These findings indicate that most endogenous PRPK might already be accompanied by CGI-121 to maintain the inactivation and degradation of p53. If this is indeed the case, then this hypothesis appears to be consistent with the result that transient overexpression of CGI-121 causes no significant alteration of the p53 status. It is of interest to know how CGI-121 protein expression is regulated during diverse cellular events. To address this issue, we treated 293T or HL-60 cells with several biochemical stimulants; however, our tests revealed no significant alterations of CGI-121 protein expression after treatment with the following reagents: EGF; TGF-a and b; IFN-a; b, and c; PMA; retinoic acid; vitamin D3; protease/proteasome inhibitors; and UV irradiation (data not shown). Furthermore, we investigated whether CGI-121 expression was regulated in a cell cycle-dependent manner using synchronized U2OS cells; however, the total abundance of CGI-121 protein varied little throughout the cell cycle (data not shown). A feedback loop for p53 and its related molecules, in which p53 transactivated several target proteins and the activity of p53 was controlled by the induced proteins, has previously been described [24,25]. The loss of p53 function commonly results from deletion and/or dominant negative mutation of the p53 gene [26,27], and it leads to abnormal cell proliferation and immortality. Therefore, we examined the protein expression of CGI121 in several cell lines in which various kinds of mutations and/or deletions were harbored in the p53 gene [28]. As shown in Results, CGI-121 protein was expressed invariably, regardless of the p53 status in those cell lines. At the moment, we have no direct evidence regarding the mechanism by which CGI-121 expression is regulated. In conclusion, our experiments provide insight into the function of CGI-121 as a binding partner of PRPK. One key role of CGI-121 appears to be the inhibition of the interaction of PRPK with p53. The regulation of CGI-121 expression and activity remains of great further interest. Moreover, the existence of abundant CGI121 not only in the nucleus, but also in the cytosol, suggests that CGI-121 may carry out additional physiological tasks besides that of binding to PRPK. The identification of putative CGI-121-associated proteins and additional substrates of PRPK will be necessary in
order to identify and describe this novel signaling pathway.
Acknowledgment We thank Yuko Hirano for the technical assistance with this study.
References [1] A.J. Levine, p53, the cellular gatekeeper for growth and division, Cell 88 (1997) 323–331. [2] D.B. Woods, K.H. Vousden, Regulation of p53 function, Exp. Cell Res. 264 (2001) 56–66. [3] R. Honda, H. Tanaka, H. Yasuda, Oncoprotein MDM2 is a ubiquitin ligase E3 for tumor suppressor p53, FEBS Lett. 420 (1997) 25–27. [4] Y. Haupt, R. Maya, A. Kazaz, M. Oren, Mdm2 promotes the rapid degradation of p53, Nature 387 (1997) 296–299. [5] M.H. Kubbutat, S.N. Jones, K.H. Vousden, Regulation of p53 stability by Mdm2, Nature 387 (1997) 299–303. [6] S.Y. Shieh, M. Ikeda, Y. Taya, C. Prives, DNA damage-induced phosphorylation of p53 alleviates inhibition by MDM2, Cell 91 (1997) 325–334. [7] S. Nakamura, J.A. Roth, T. Mukhopadhyay, Multiple lysine mutations in the C-terminal domain of p53 interfere with MDM2dependent protein degradation and ubiquitination, Mol. Cell Biol. 20 (2000) 9391–9398. [8] M.S. Rodriguez, J.M. Desterro, S. Lain, D.P. Lane, R.T. Hay, Multiple C-terminal lysine residues target p53 for ubiquitinproteasome-mediated degradation, Mol. Cell Biol. 20 (2000) 8458–8467. [9] K. Sakaguchi, S. Saito, Y. Higashimoto, S. Roy, C.W. Anderson, E. Appella, Damage-mediated phosphorylation of human p53 threonine 18 through a cascade mediated by a casein 1-like kinase. Effect on Mdm2 binding, J. Biol. Chem. 275 (2000) 9278– 9283. [10] A. Hirao, Y.Y. Kong, S. Matsuoka, A. Wakeham, J. Ruland, H. Yoshida, D. Liu, S.J. Elledge, T.W. Mak, DNA damage-induced activation of p53 by the checkpoint kinase Chk2, Science 287 (2000) 1824–1827. [11] S.Y. Shieh, J. Ahn, K. Tamai, Y. Taya, C. Prives, The human homologs of checkpoint kinases Chk1 and Cds1 (Chk2) phosphorylate p53 at multiple DNA damage-inducible sites, Genes Dev. 14 (2000) 289–300. [12] N.H. Chehab, A. Malikzay, M. Appel, T.D. Halazonetis, Chk2/ hCds1 functions as a DNA damage checkpoint in G(1) by stabilizing p53, Genes Dev. 14 (2000) 278–288. [13] S.P. Lees-Miller, K. Sakaguchi, S.J. Ullrich, E. Appella, C.W. Anderson, Human DNA-activated protein kinase phosphorylates serines 15 and 37 in the amino-terminal transactivation domain of human p53, Mol. Cell Biol. 12 (1992) 5041–5049. [14] S. Banin, L. Moyal, S. Shieh, Y. Taya, C.W. Anderson, L. Chessa, N.I. Smorodinsky, C. Prives, Y. Reiss, Y. Shiloh, Y. Ziv, Enhanced phosphorylation of p53 by ATM in response to DNA damage, Science 281 (1998) 1674–1677. [15] C.E. Canman, D.S. Lim, K.A. Cimprich, Y. Taya, K. Tamai, K. Sakaguchi, E. Appella, M.B. Kastan, J.D. Siliciano, Activation of the ATM kinase by ionizing radiation and phosphorylation of p53, Science 281 (1998) 1677–1679. [16] D.M. Milne, L.E. Campbell, D.G. Campbell, D.W. Meek, p53 is phosphorylated in vitro and in vivo by an ultraviolet radiationinduced protein kinase characteristic of the c-Jun kinase, JNK1, J. Biol. Chem. 270 (1995) 5511–5518.
A. Miyoshi et al. / Biochemical and Biophysical Research Communications 303 (2003) 399–405 [17] M.C. Hu, W.R. Qiu, Y.P. Wang, JNK1, JNK2 and JNK3 are p53 N-terminal serine 34 kinases, Oncogene 15 (1997) 2277–2287. [18] D.M. Milne, D.G. Campbell, F.B. Caudwell, D.W. Meek, Phosphorylation of the tumor suppressor protein p53 by mitogen-activated protein kinases, J. Biol. Chem. 269 (1994) 9253– 9260. [19] S. Stocchetto, O. Marin, G. Carignani, L.A. Pinna, Biochemical evidence that Saccharomyces cerevisiae YGR262c gene, required for normal growth, encodes a novel Ser/Thr-specific protein kinase, FEBS Lett. 414 (1997) 171–175. [20] Y. Abe, S. Matsumoto, S. Wei, K. Nezu, A. Miyoshi, K. Kito, N. Ueda, K. Shigemoto, Y. Hitsumoto, J. Nikawa, Y. Enomoto, Cloning and characterization of a p53-related protein kinase expressed in interleukin-2-activated cytotoxic T-cells, epithelial tumor cell lines, and the testes, J. Biol. Chem. 276 (2001) 44003– 44011. [21] F. Corpet, Multiple sequence alignment with hierarchical clustering, Nucleic Acids Res. 16 (1988) 10881–10890. [22] H. Wada, K. Kito, L.S. Caskey, E.T. Yeh, T. Kamitani, Cleavage of the C-terminus of NEDD8 by UCH-L3, Biochem. Biophys. Res. Commun. 251 (1998) 688–692.
405
[23] C.H. Lai, C.Y. Chou, L.Y. ChÕang, C.S. Liu, W. Lin, Identification of novel human genes evolutionarily conserved in Caenorhabditis elegans by comparative proteomics, Genome Res. 10 (2000) 703–713. [24] X. Wu, J.H. Bayle, D. Olson, A.J. Levine, The p53-mdm-2 autoregulatory feedback loop, Genes Dev. 7 (1993) 1126–1132. [25] K.H. Vousden, Activation of the p53 tumor suppressor protein, Biochim. Biophys. Acta 1602 (2002) 47–59. [26] B. Vogelstein, E.R. Fearon, S.R. Hamilton, S.E. Kern, A.C. Preisinger, M. Leppert, Y. Nakamura, R. White, A.M. Smits, J.L. Bos, Genetic alterations during colorectal-tumor development, N. Engl. J. Med. 319 (1988) 525–532. [27] S.J. Baker, E.R. Fearon, J.M. Nigro, S.R. Hamilton, A.C. Preisinger, J.M. Jessup, P. vanTuinen, D.H. Ledbetter, D.F. Barker, Y. Nakamura, et al., Chromosome 17 deletions and p53 gene mutations in colorectal carcinomas, Science 244 (1989) 217– 221. [28] M. Hollstein, K. Rice, M.S. Greenblatt, T. Soussi, R. Fuchs, T. Sorlie, E. Hovig, B. Smith-Sorensen, R. Montesano, C.C. Harris, Database of p53 gene somatic mutations in human tumors and cell lines, Nucleic Acids Res. 22 (1994) 3551–3555.