Oncogene (2007) 26, 5596–5605 & 2007 Nature Publishing Group All rights reserved 0950-9232/07 $30.00 www.nature.com/onc ORIGINAL ARTICLE Pituitary tumor transforming gene interacts with Sp1 to modulate G1/S cell phase transition Y Tong, Y Tan, C Zhou and S Melmed Department of Medicine, Cedars-Sinai Research Institute, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA Pituitary tumor transforming gene (PTTG1) was isolated et al., 2001, 2003; Chesnokova et al., 2005). Disrupted from rat pituitary tumor cells, and subsequently identified PTTG1 results in insulinopenic diabetes in adult mice, as a securin protein as well as a transcription factor. We with male-selective hyperglycemia (Wang et al., 2003). show here a global transcriptional effect of PTTG1 in Enhanced PTTG1 abundance correlates with tumor human choriocarcinoma JEG-3 cells by simultaneously development and size (McCabe et al., 2003). PTTG1 is assessing 20 000 gene promoters using chromatin immu- induced in the early stages of estrogen-induced rat noprecipitation (ChIP)-on-Chip experiments. Seven hun- prolactinoma pathogenesis (Heaney et al., 1999) and has dred and forty-six gene promoters (Po0.001) were found been suggested as a prognostic marker for differentiated enriched, with functions relating to cell cycle, metabolic thyroid cancer, lymph node invasion and breast cancer controland signaltransduction. Significant interaction recurrence (Solbach et al., 2004), and colon cancer between PTTG1 and Sp1 (Po0.000001) was found by invasiveness and vascularity (Heaney et al., 2000). transcriptionalpattern analysis of ChIP data and further Mechanisms of PTTG1 action include binding to p53, confirmed by immunoprecipitation and pull-down assays. inhibiting its transcriptional activity and interacting PTTG1 acts coordinately with Sp1 to induce cyclin D3 with Ku, the regulatory subunit of DNA-dependent expression Bthreefold, and promotes G1/S-phase transi- protein kinase (Wang and Melmed, 2000; Pei, 2000), tion independently of p21. PTTG1 deletion was also activating c-Myc (Pei, 2001) and basic fibroblast growth protective for anchorage-independent cell colony forma- factor (bFGF) (Chien and Pei, 2000), and acting as a tion. The results show that PTTG1 exhibits properties of securin to inhibit separase, which cleaves sister chroma- a global transcription factor, and specifically modulates tids with fidelity during mitosis (Zou et al., 1999). the G1/S-phase transition by interacting with Sp1. This However, it is unclear how PTTG1 overexpression novelsignaling pathway may be required for PTTG1 mediates tumorigenesis and little is known of the transforming activity. functional mechanisms underlying cellular PTTG1 Oncogene (2007) 26, 5596–5605; doi:10.1038/sj.onc.1210339; action. published online 12 March 2007 Because of widespread PTTG1 functions, we surveyed PTTG1 effects on 20 000 gene promoters. We show here Keywords: PTTG1; pituitary; cell cycle that PTTG1 acts as a global transcription factor, and specifically modulates the G1/S-phase transition by interacting with Sp1, which may underlie its transform- ing activity. Introduction Results Pituitary tumor transforming gene (PTTG1) isolated ChIP-on-Chip experiments and microarray from rat pituitary tumor cells (Pei and Melmed, 1997), image analysis was subsequently identified as a securin protein (Zou The chromatin immunoprecipitation (ChIP)-on-Chip et al., 1999) and a transcription factor (Dominguez assay showed that 746 gene promoters (3.73% of all gene et al., 1998). It also plays a role in organ development promoters studied) were significantly enriched (P 0.001) and metabolism. PTTG1 is required for rat liver o among 20 000 genes by immunoprecipitation with PTTG1 regeneration, human fetal brain development (Boelaert antibody (Supplemental Data 2) and the remaining 13000 et al., 2003) and telencephalic neurogenesis (Tarabykin genes were used as background promoter set. The results et al., 2000). PTTG1-null mice exhibit testicular, splenic, of ChIP-on-Chip assay were confirmed by ChIP poly- pancreatic beta cell and pituitary hypoplasia (Wang merase chain reaction (PCR) as described (Li et al., 2003). Briefly, 19 gene promoters, including ATPAF1, Correspondence: Dr S Melmed, Academic Affairs, Room 2015, 8700 HUMCYT2A, C6orf102, KNTC1, COCH, PP2447, Beverly Blvd, Los Angeles, CA 90048, USA. E-mail: [email protected] CSNK2B, RAD51L3, EIF4B, RASGRF1, FLJ31153, Received 13October 2006; revised 5 January 2007; accepted 7 January RPL32, GABARAPL2, SREBF2, GALNT5, SUCLG2, 2007; published online 12 March 2007 GCDH, TNFRSF13C and GOLGA1, were selected and PTTG1 modulates G1/S phase transition Y Tong et al 5597 conventional ChIP PCR assay used to detect enrichment reported Myc/Max interaction partners TFIIB, Sp1, of the promoters in the immunoprecipitated chromatins. E2F and Ap-2 (Supplemental Data 3). The negative primers used were from the ChIP-IT Among the 746 enriched gene promoters, 618 were kit (Active Motif). The results showed that at least well characterized from the promoter sequence database 95% of the observed enriched genes are true positives (http://biowulf.bu.edu/zlab/PromoSer/promoser.html) (Figure 1a). and the corresponding sequences extracted for tran- scriptional motif analysis as stated above. Analysis of PTTG1 ChIP-on-Chip data and the background data Transcriptional pattern analysis sets revealed a significant interaction between PTTG1 The transcriptional pattern analysis as described in and Sp1 (Po0.000001) (Figure 1b). To further identify Materials and methods was then performed to unravel the binding location of SP1, the promoter sequence of transcriptional patterns yielded by the ChIP-on-Chip each gene in both the enriched and background results. This approach allowed us to test functional promoter set was evenly divided into 10 intervals, each relationships of PTTG1 with other transcription factors comprising 200 bp (from À1800 bp upstream to 200 bp and to identify transcriptional regulation modules downstream of transcription starting site (TSS)) and the controlling gene expression. The same transcriptional frequency of motif occurrence in each interval was pattern analysis was also applied to published c-Myc calculated and plotted in Figure 1c. The results show ChIP-on-Chip data (Li et al., 2003) and the results that the greatest difference of Sp1 binding location showed that the method appropriately identified between enriched and background promoter sets occurs a None Beads Anti-PTTGInput None Beads Anti-PTTGInput ATPAF1 HUMCYT2A C6orf102 KNTC1 COCH PP2447 CSNK2B RAD51L3 EIF4B RASGRF1 FLJ31153 RPL32 GABARAPL2 SREBF2 GALNT5 SUCLG2 GCDH TNFRSF13C GOLGA1 Negative primer b 45 c 40 0.4 Sp1 35 Sp1 0.35 30 0.3 PTTG1 enriched Background 25 0.25 20 0.2 15 Percentage 0.15 10 0.1 Number of promoter set 5 0.05 TSS 0 0 0.58 0.6 0.62 0.64 0.66 0.68 0.7 12345 678910 Frequency Interval Figure 1 ChIP-on-Chip and transcription pattern analysis. (a) Nineteen genes were randomly selected from the ChIP-on-Chip results. JEG-3chromatin was immunoprecipitated by PTTG1 antibody and regular PCR assay carried out using ChIP-IT kits (Active Motif) to detect and confirm promoter enrichment. All 19 gene promoters were confirmed as positive. (b) Transcription pattern analysis: frequency distribution of occurrence of the Sp1 motif in the 1000 randomly selected background promoter data sets. Solid red line represents frequency of occurrence of the Sp1 motif in the enriched promoter data set. Frequency of Sp1 in the PTTG1-enriched set was significantly (Po0.000001) higher than expected by chance. (c) Frequency distribution of occurrence of the Sp1 motif along the promoter sequence (10 intervals, each 200 bp long, ranging from À1800 bp to 200 bp) for both the enriched and background promoter set (containing the remaining non-enriched promoters). The largest difference appeared to be À200 bp from the TSS. Oncogene PTTG1 modulates G1/S phase transition Y Tong et al 5598 at À200 to 0 bp, suggesting that the Sp1 binding site may PTTG1 regulates Sp1 binding to its motif and colocalizes approximate the TSS position (Figure 1c). with Sp1 on cyclin D3 promoter We further examined Sp1 binding to its motif in the presence of PTTG1 and PTTG1 fragments. In PTTG1 interacts with Sp1 Figure 3a1, we show that PTTG1 and the consensus Sp1 We used agarose-conjugated Sp1 antibody to immuno- binding sequence in the absence of Sp1 protein did not precipitate the Sp1 complex, and as shown in Figure 2a, cause a shift, suggesting that PTTG1 does not bind the PTTG1 associated with Sp1 in both JEG-3and HCT116 Sp1 binding sequence. Incubation of Sp1 with the Sp1 cells, suggesting that they interact in vivo. We confirmed binding sequence is shown to shift the band, suggesting PTTG1 and Sp1 interaction by His-tag pulldown assay that Sp1 binds to this sequence as expected. Incubation and further mapped their interaction domains using of Sp1, PTTG1 and the consensus Sp1 binding sequence poly-His-tagged PTTG1, Sp1 and their respective resulted in both a shifted and a super-shifted band. The fragments (Figure 2b and c). The PTTG1 N-terminus shifted band likely comprises Sp1 and consensus Sp1 (aa 1–120) and Sp1 C-terminus (aa 401–785) bind to the binding sequence complex. As only two proteins are full-length protein, respectively, suggesting the location present in this system, the super-shifted bands appear of protein domains required for their physical interac- to reflect a formation of PTTG1-Sp1-consensus Sp1 tions (Figure 2b, c and d). binding sequence complexes. abSp1 and its fragments Binding to PTTG1 1 400 Sp1.F1 - 301 500 - Sp1.F2 Input Neg IgG Sp1 Ab 401 785 + Sp1.F3 JEG-3 1 785 Sp1 + PTTG1 HCT116 Control Sp1.F1 Sp1.F2 Sp1.F3 Sp1 Input PTTG1 c PTTG1 and its fragments Binding to Sp1 1 120 + PTTG1.F1 101 202 PTTG1.F2 - 1 202 PTTG1 + Control PTTG1-F1PTTG1-F2PTTG1 Input Sp1 d Sp1 N C Interactio PTTG1 N C Sp box Transactivating DNA binding Zinc finger Ken box SH-3 binding D box Figure 2 PTTG1 interacts with Sp1.