© 2016. Published by The Company of Biologists Ltd | Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

RESEARCH ARTICLE Telomerase activates transcription of cyclin D1 through an interaction with NOL1 Juyeong Hong1, Ji Hoon Lee1,2 and In Kwon Chung1,2,*

ABSTRACT several additional including dyskerin, TCAB1 (also Telomerase is a ribonucleoprotein enzyme that is required for the known as WRAP53), pontin and reptin (Egan and Collins, maintenance of telomere repeats. Although overexpression of 2012; Venteicher et al., 2008, 2009). Telomerase expression is telomerase in normal human somatic cells is sufficient to overcome very low in most human somatic cells but upregulated in many replicative senescence, the ability of telomerase to promote human cancer cells and stem cells, suggesting that activation of tumorigenesis requires additional activities that are independent of telomerase supports the continued cell proliferation (Kim et al., its role in telomere extension. Here, we identify proliferation- 1994). associated nucleolar antigen 120 (NOL1, also known as NOP2) as Although overexpression of telomerase is sufficient to overcome a telomerase RNA component (TERC)-binding that is found in replicative senescence (Bodnar et al., 1998), recent studies have association with catalytically active telomerase. Although NOL1 is suggested that besides its reverse transcriptase activity, telomerase highly expressed in the majority of human tumor cells, the molecular has the noncanonical functions that contribute to cancer mechanism by which NOL1 contributes to tumorigenesis remained development and progression (Stewart et al., 2002; Li and unclear. We show that NOL1 binds to the T-cell factor (TCF)-binding Tergaonkar, 2014). Ectopic expression of telomerase in human element of the cyclin D1 promoter and activates its transcription. mammary epithelial cells results in enhanced expression of Interestingly, telomerase is also recruited to the cyclin D1 promoter in growth-promoting (Smith et al., 2003). Transgenic a TERC-dependent manner through the interaction with NOL1, induction of TERT in mouse skin epithelium has been shown to further enhancing transcription of the cyclin D1 gene. Depletion of cause proliferation of quiescent stem cells (Sarin et al., 2005). This NOL1 suppresses cyclin D1 promoter activity, thereby leading to function for TERT is independent of reverse transcriptase activity induction of growth arrest and altered distributions. (Choi et al., 2008). In addition, TERT has been found to directly Collectively, our findings suggest that NOL1 represents a new route interact with BRG1 (also known as SMARCA4) and activate β by which telomerase activates transcription of cyclin D1 gene, thus transcription of Wnt/ -catenin-dependent genes such as cyclin D1 maintaining cell proliferation capacity. and Myc (Park et al., 2009). However, the proposed noncanonical role of TERT in the Wnt/β-catenin signaling cascade has been KEY WORDS: Telomerase, NOL1, Cyclin D1, Transcriptional controversial. Several studies have reported a lack of physical activation, Tumor cell marker association of TERT with BRG1 or β-catenin (Listerman et al., 2014), as well as no apparent effect of TERT deficiency on INTRODUCTION phenotypes associated with Wnt signaling in TERT-knockout mice Telomeres, the specialized nucleoprotein complexes located at the (Strong et al., 2011). Although TERT appears to regulate the ends of eukaryotic , are essential for maintenance of expression of growth-promoting genes, this event might not be stability and genome integrity (Blackburn, 2001; solely promoted by Wnt signaling. Indeed, TERT has been Smogorzewska and de Lange, 2004). Telomeric DNA is tightly reported to bind to the NF-κB p65 subunit (also known as RelA) associated with the six-subunit protein complex shelterin, which and activate NF-κB-dependent (Ghosh et al., prevents chromosomal ends from being recognized as DNA 2012). damage (Palm and de Lange, 2008; Sfeir and de Lange, 2012). In Given that the large size of human telomerase suggests the the absence of a telomere maintenance pathway, most human existence of additional components, we performed a large-scale somatic cells show a progressive loss of telomeric DNA with each affinity purification to identify proteins that interact with round of cell division due to the end replication problem (Lingner telomerase. Here, we identify proliferation-associated nucleolar et al., 1995; Blasco et al., 1997). The maintenance of telomere antigen p120 (NOL1, also known as NOP2) as a TERC-binding repeats in most eukaryotic organisms requires telomerase, which protein. NOL1 was originally identified as a RNA-binding and adds telomere repeats onto the 3′ ends of linear chromosomes by nucleolar-specific protein that is highly expressed in the majority reverse transcription (Autexier and Lue, 2006; Bianchi and Shore, of human malignant tumor cells but not in normal resting cells 2008). Human telomerase consists of telomerase reverse (Ochs et al., 1988; Jhiang et al., 1990; Fonagy et al., 1992, 1993). transcriptase (hTERT), telomerase RNA component (TERC) and Although NOL1 has been implicated as a tumor cell marker (Gorczyca et al., 1992), the molecular mechanism by which NOL1 contributes to tumorigenesis is poorly understood. We show that 1Department of Integrated Omics for Biomedical Science, College of Life Science and Biotechnology, Yonsei University, Seoul 120-749, Korea. 2Department of NOL1 binds to the T-cell factor (TCF)-binding element (TBE) of Systems Biology, College of Life Science and Biotechnology, Yonsei University, the cyclin D1 promoter and activates its transcription. Telomerase Seoul 120-749, Korea. is also recruited to the cyclin D1 promoter through the interaction *Author for correspondence ([email protected]) with NOL1, further enhancing transcription of cyclin D1 gene. These results suggest a new role for telomerase as a modulator of

Received 25 September 2015; Accepted 15 February 2016 NOL1-dependent transcriptional activation in human cancer cells. Journal of Cell Science

1566 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

RESULTS V5 expression vectors and subjected to immunoprecipitation. Identification of NOL1 as an hTERT-interacting factor NOL1–V5 was specifically bound to Flag–hTERT that was To identify proteins that interact with hTERT, we expressed Flag- immunoprecipitated from HEK293 cells (Fig. 1B). Reciprocal tagged hTERT in HEK293 cells and isolated hTERT complex using immunoprecipitation showed that Flag–hTERT was detected in large-scale affinity purification. Proteins co-purified with Flag- anti-V5 immunoprecipitates, indicating that hTERT associates hTERT were identified by nano-liquid chromatography-tandem with NOL1 in mammalian cells. Interestingly, the interaction mass spectrometry (nano LC-MS/MS). Among the known between Flag–hTERT and NOL1–V5 was disrupted by RNase A telomerase components, TCAB1 and nucleolin were enriched in treatment of the extract, which degrades TERC. Endogenous the hTERT complex (Fig. 1A). In addition, analysis of a band NOL1 was immunoprecipitated by endogenous hTERT, and migrating with an approximate relative molecular mass of 120 kDa this association was also disrupted by RNase A treatment was identified as NOL1, a highly conserved, nucleolar-specific (Fig. 1C), suggesting that NOL1 can associate with hTERT RNA-binding protein (Ochs et al., 1988; Jhiang et al., 1990). Given through TERC binding in intact cells. These findings were that NOL1 has been detected in proliferating tissues but not in further verified by immunoprecipitation experiments with U2OS normal resting cells, it has been implicated as a tumor cell marker. cells, which lack endogenous hTERT and TERC (Jegou et al., Thus, we wanted to investigate the role of NOL1 in telomerase 2009). In this cellular background, Flag–hTERT did not interact function. with NOL1-V5 owing to a lack of TERC (Fig. 1D). Taken To determine whether hTERT and NOL1 associate in vivo, together, these results suggest that NOL1 is a new TERC-binding HEK293 cells were co-transfected with Flag–hTERT and NOL1– protein.

Fig. 1. Identification of NOL1 as an hTERT- interacting protein. (A) Lysates from HEK293 cells expressing Flag–hTERT were immunoprecipitated with anti-Flag antibody and assayed for protein binding by Coomassie staining of a SDS-PAGE gel. Binding proteins were identified by nano-LC- MS/MS. The position of molecular size markers are shown in kDa. Seventeen unique peptides for NOL1 identified from the mass spectrometry are shown. (B) HEK293 cells expressing Flag–hTERT and NOL1–V5 were subjected to immunoprecipitation (IP) with anti-Flag and anti-V5 antibodies, followed by immunoblotting with anti-V5 and anti-Flag antibodies. The indicated extracts were treated with 0.1 mg/ml RNase A during immunoprecipitation to degrade TERC. (C) HEK293 cells were subjected to immunoprecipitation with anti-hTERT antibody, followed by immunoblotting with anti-NOL1 antibody. IgG was used as a negative control. (D) Telomerase-negative U2OS cells expressing Flag–hTERT and NOL1–V5 were subjected to immunoprecipitation with anti-Flag antibody, followed by immunoblotting with anti-V5 antibody. Journal of Cell Science

1567 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

Fig. 2. See next page for legend.

NOL1 associates with hTERT in a TERC-dependent manner immunoprecipitating a series of deletion fragments of To determine the domain in hTERT that is responsible for hTERT (Fig. 2A). As shown in Fig. 2B, NOL1–V5 was

NOL1 interaction, we assessed binding of NOL1–V5 by immunoprecipitated only by the hTERT fragments containing Journal of Cell Science

1568 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

Fig. 2. Identification of the domains in NOL1 and hTERT that are required immunoprecipitated NOL1 fragments encompassing amino for their interactions. (A) Schematic representation of the region of hTERT acid residues 380–583 (Fig. 2F), indicating that the rRNA involved in NOL1 binding. The approximate positions of the reverse methyltransferase domain is essential for TERC binding. transcriptase motifs analyzed are indicated. TRBD, TERC-binding domain. (B) Lysates from HEK293 cells expressing the various Flag–hTERT fragments The finding that NOL1 associates with hTERT through TERC was and NOL1-V5 were immunoprecipitated with anti-Flag antibody, followed by further verified by immunoprecipitation experiments. HEK293 cells immunoblotting (IB) with anti-NOL1 antibody. The Flag–hTERT fragments are were transfected with either NOL1–V5 or NOL1-E–V5 (the minimal indicated by arrows. The asterisks mark the positions of nonspecific TERC-binding domain) and subjected to immunoprecipitation, immunoglobulin chains. The position of molecular size markers are shown in followed by semi-quantitative RT-PCR to detect TERC. The results kDa. (C) Schematic representation of the mutant construct of hTERT in which showed that TERC was specifically immunoprecipitated by NOL1- the TRBD was deleted (ΔTRBD). The approximate positions of the reverse transcriptase motifs analyzed are indicated. (D) Removing the TRBD on V5 and NOL1-E-V5 (Fig. 3A). Dyskerin, which has been shown to hTERT abolishes NOL1 association. Lysates from HEK293 cells expressing interact with TERC, was used as a positive-binding control (Lee et al., NOL1–V5 and either Flag–hTERT or Flag–ΔTRBD were immunoprecipitated 2014). To further demonstrate that NOL1 directly interacts with (IP) with anti-Flag antibody, followed by immunoblotting with anti-V5 antibody. TERC, we performed a GST pulldown assay with in vitro transcribed (E) Schematic representation of the region of NOL1 involved in hTERT binding. TERC. GST–NOL1-E, but not the control GST protein, bound The approximate positions of the nuclear localization signal (NLS), coiled-coil efficiently to in vitro transcribed TERC, as did GST–dyskerin domain and putative rRNA methyltransferase (MTase) motif are indicated. (Fig. 3B). (F) Lysates from HEK293 cells expressing the various NOL1–V5 domains and Flag–hTERT were immunoprecipitated with anti-Flag antibody, followed by immunoblotting with the anti-V5 antibody. The NOL1–V5 domains NOL1 associates with catalytically active telomerase but immunoprecipitated with Flag–hTERT are indicated by arrows. The asterisks does not affect telomerase enzymatic activity mark the positions of nonspecific immunoglobulin chains. The position of Given that NOL1 associates with hTERT through TERC binding, molecular size markers are shown in kDa. we determined whether NOL1 is a telomerase holoenzyme subunit. HEK293 cells were co-transfected with NOL1–V5 and Flag– aminoacidresidues1–589. Because this region contains the hTERT, Flag–TCAB1 or Flag–dyskerin, and subjected to TERC-binding domain (TRBD), we examined whether the immunoprecipitation. NOL1–V5 was immunoprecipitated by TRBD is required for NOL1 binding to hTERT. The results Flag–TCAB1 and Flag–dyskerin, as observed for Flag–hTERT showed that removing the TRBD on hTERT abolished (Fig. 4A). We next examined whether NOL1 associates with NOL1 binding (Fig. 2C,D), further supporting the idea that catalytically active telomerase. HEK293 cells expressing Flag– the association of NOL1 with hTERT is dependent on TERC. NOL1 or other Flag-tagged telomerase components were subjected TomaptheregioninNOL1thatisrequiredforTERC to immunoprecipitation with anti-Flag antibody and analyzed for binding, we generated deletion constructs lacking a coiled-coil telomerase activity by a telomeric repeat amplification protocol domain or a putative rRNA methyltransferase (MTase) domain (TRAP) assay. Immunoprecipitates of Flag–NOL1 contained (Fig. 2E) (Koonin, 1994; Gustafson et al., 1998). Flag–hTERT telomerase activity (Fig. 4B), as did those of Flag–hTERT, Flag–

Fig. 3. NOL1 directly interacts with TERC. (A) Lysates from HEK293 cells expressing NOL1–V5 or NOL1-E–V5 or Flag–dyskerin were immunoprecipitated (IP) with anti-V5 or anti-Flag antibodies, followed by immunoblotting (IB) with anti-V5 and anti-Flag antibodies and semi-quantitative RT-PCR to detect TERC. The V5–NOL1 fragments and Flag–dyskerin are indicated by arrows. The asterisks mark the positions of nonspecific immunoglobulin chains. The position of molecular size markers are shown in kDa. (B) GST, GST–NOL1-E or GST–dyskerin were immobilized on glutathione–Sepharose and incubated with in vitro transcribed TERC. Bound TERC was detected by semi-quantitative RT-PCR. The purified GST fusion proteins were visualized by Coomassie Blue staining.

In vitro transcribed TERC was stained with ethidium bromide. Journal of Cell Science

1569 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

Fig. 4. NOL1 associates with catalytically active telomerase. (A) Lysates from HEK293 cells expressing NOL1–V5, together with Flag– hTERT, Flag–TCAB1 or Flag–dyskerin were immunoprecipitated (IP) with anti-Flag antibody, followed by immunoblotting with anti-V5 antibody. (B) Lysates from HEK293 cells expressing Flag– NOL1, Flag–hTERT, Flag–TCAB1 or Flag– dyskerin were immunoprecipitated with anti-Flag antibody and analyzed for telomerase activity by the TRAP assay. (C) HeLa S3 cells expressing control shRNA (shControl) or NOL1 shRNAs (shNOL1-1 and shNOL1-2) were subjected to immunoblotting, to measure the protein levels of telomerase components, and semi-quantitative RT-PCR, to detect the mRNA levels of telomerase components and TERC. (D) HeLa S3 cells expressing shControl or shNOL1 were subjected to immunoblotting to measure the levels of shelterin proteins. (E) HeLa S3 cells expressing shControl or shNOL1 were analyzed for telomerase activity by the TRAP assay. To test RNA-dependent extension, RNase A (0.25 mg/ml) was added to the extracts before the primer extension reaction when indicated. ITAS, internal telomerase assay standard.

TCAB1 and Flag–dyskerin, indicating that NOL1 is a component of NOL1 did not affect the subnuclear localization of catalytically active telomerase. telomerase To examine an involvement of NOL1 in telomerase function, the Telomerase undergoes a highly elaborate, stepwise process expression of endogenous NOL1 was stably depleted in HeLa S3 of assembly and trafficking within the nucleus (Lee et al., cells using short hairpin RNA (shRNA) produced from a retroviral 2014). If NOL1 is required for assembly and trafficking of active vector. NOL1-knockdown cells maintained the reduced levels of telomerase, we would expect NOL1 depletion to impair the NOL1 throughout the duration of the experiments (see below). subnuclear localization of hTERT. To examine this possibility, Depletion of NOL1 did not affect the levels of telomerase we depleted NOL1 in HeLa S3 cells and performed indirect components (Fig. 4C) and shelterin proteins (Fig. 4D). We also immunofluorescence staining to determine the subnuclear found that depletion (Fig. 4E) or overexpression of NOL1 (Fig. S1) localization of endogenous hTERT. Given that telomerase did not affect telomerase activity. Taken together, these results synthesizes telomeres specifically during S phase (Lee et al., indicate that although NOL1 associates with catalytically active 2010; Tomlinson et al., 2006), HeLa S3 cells were synchronized at telomerase, it has no direct regulatory effect on telomerase S phase using a double thymidine block (Lee et al., 2010). In the enzymatic activity. shRNA control cells, the majority of hTERT was found to localize Journal of Cell Science

1570 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040 to nucleoli (Fig. S2A,B). The nucleolar localization of hTERT transcription was induced about four-fold by NOL1 overexpression was not affected by NOL1 depletion. Telomerase has been shown compared to the vector control. The proximal 85-base region, which to accumulate in Cajal bodies prior to telomere elongation contains the TBE, was essential for activation of the cyclin D1 (Venteicher et al., 2009; Venteicher and Artandi, 2009). Thus, promoter, suggesting that the TCF-binding site is a NOL1- we determined whether NOL1 depletion affects colocalization of responsive element. hTERT with Cajal bodies during S phase. Cajal body localization We next examined the effect of NOL1 overexpression on TCF of hTERT was not affected by NOL1 depletion, as indicated by transcriptional activity by transfecting cells with the TCF-sensitive dual staining with a coilin-specific antibody (Fig. S2C,D). These luciferase reporter vector (TOP-Flash) or TCF-insensitive control results demonstrate that NOL1 does not affect the intranuclear vector (FOP-Flash) in HeLa CCL2 cells. The results showed that trafficking of hTERT. ectopic expression of NOL1 increased TOP-Flash activity but not Dysfunctional telomeres are recognized by the canonical DNA FOP-Flash activity (Fig. 5D). Interestingly, ectopic expression of damage signaling pathway, and the resulting telomere-dysfunction- hTERT alone also induced TOP-Flash activity, which was further induced foci (TIFs) represent the foci of DNA damage response increased by co-expression of NOL1. To verify these findings, factors that coincide with telomeres (Takai et al., 2003; D’Adda di telomerase-positive H1299 and MCF7 cells were transfected with Fagagna et al., 2003). To determine the role of NOL1 in telomere- the −964 promoter luciferase reporter vector together with NOL1– damage pathway, the telomeric foci for 53BP1 and the V5 or Flag–hTERT or both. Overexpression of either NOL1 or phosphorylated H2AX marker (γH2AX) were examined in hTERT led to an increase in cyclin D1 promoter activity compared NOL1-knockdown cells. Depletion of NOL1 did not induce more to the vector control (Fig. 5E,F). When both proteins were co- telomere-damage foci in the nucleus compared to the control cells expressed, we observed an additive effect on cyclin D1 promoter (Fig. S3), indicating that NOL1 is not involved in the control of a activity. These results suggest that both NOL1 and hTERT are DNA damage response at telomeres. required for efficient cyclin D1 transcription. Although NOL1 stimulates transcription of the cyclin D1 gene NOL1 activates transcription of the cyclin D1 gene independently of hTERT, it is unclear whether hTERT alone is NOL1 is a proliferation-related nucleolar protein that is highly sufficient to activate cyclin D1 promoter activity without NOL1 expressed in the majority of human malignant tissues (Ochs et al., binding. To test this possibility, dependence of NOL1 was 1988; Jhiang et al., 1990). It is expressed in the early G1 phase and examined in telomerase-negative U2OS cells. As shown in peaks in S phase (Fonagy et al., 1992, 1993; Gorczyca et al., 1992). Fig. 5G, cyclin D1 promoter activity was increased by NOL1 Although the expression of NOL1 has been shown to be induced overexpression but not by hTERT overexpression. Moreover, the rapidly following growth stimulation and produce tumors in the additive effect was not observed upon overexpression of both nude mice (Perlaky et al., 1992), the mechanism by which NOL1 proteins. These results could be due to a lack of TERC in U2OS exerts these effects is poorly understood. To investigate the role of cells. Taken together, these data suggest that telomerase promotes NOL1 in the control of cell cycle and cell proliferation, we transcription of cyclin D1 gene through the interaction with NOL1. examined the effect of NOL1 depletion on the expression of cell- cycle-dependent and proliferation-controlling genes such as cyclin Both NOL1 and hTERT associate with the cyclin D1 promoter D1 and Myc (Musgrove et al., 2011; Sears, 2004). Interestingly, at the TBE depletion of NOL1 caused a clear reduction in the expression of To determine whether both NOL1 and hTERT are recruited to cyclin D1 as shown by immunoblot analysis, as well as in the levels the TBE of cyclin D1 promoter, we carried out chromatin of cyclin D1 mRNA as demonstrated by semi-quantitative RT-PCR immunoprecipitation (ChIP). HeLa S3 cells expressing Flag–NOL1 experiments, suggesting that NOL1 regulates the expression of (or empty vector) and hTERT shRNAs (or control shRNA) were cyclin D1 gene at the transcription level (Fig. 5A). By contrast, the cross-linked with formaldehyde, followed by immunoprecipitation expression of c-Myc was not affected by NOL1 depletion, with anti-Flag antibody. The immunoprecipitated chromatin was suggesting that the effect of NOL1 is specific to the promoter. We used as a template to amplify the TBE in the cyclin D1 promoter. The also found that overexpression of NOL1 increased the expression of results showed that anti-Flag antibody immunoprecipitated the cyclin D1 but did not influence the expression of c-Myc (Fig. S1A). TBE-containing fragments when Flag–NOL1 was overexpressed It has been reported that telomerase modulates Wnt/β-catenin (Fig. 6A). This TBE signal was not altered by depletion of signaling by acting as a transcriptional cofactor at Wnt target genes endogenous hTERT, indicating that NOL1 binds to the TBE (Park et al., 2009). Thus, we examined whether NOL1-mediated independently of hTERT. No amplification was observed when the transcriptional activation of cyclin D1 gene is dependent on immunoprecipitated chromatin was used to amplify the 3′- telomerase. The effect of NOL1 depletion on the cyclin D1 untranslated region (3′-UTR) (Fig. 6A). Because transcriptional expression was examined in telomerase-negative U2OS cells. activation by NOL1 is specific to the promoter, we examined whether Essentially, similar results to those seen in telomerase-positive NOL1 is recruited to the Myc promoter and found that the Flag– HeLa S3 cells were obtained in this cell line (Fig. 5B), indicating NOL1 ChIP signal was not detected at the TBE of the Myc promoter that NOL1-dependent activation of cyclin D1 transcription occurs (Fig. 6B). We next examined the effect of hTERT depletion on the regardless of telomerase expression. cyclin D1 transcription. In HeLa CCL2 cells expressing empty vector, depletion of hTERT reduced cyclin D1 promoter activity Telomerase stimulates transcription of cyclin D1 gene (Fig. 6C). However, in cells expressing Flag–NOL1, cyclin D1 through the interaction with NOL1 promoter activity was not significantly affected by depletion of The cyclin D1 promoter contains several distinct transcription- hTERT, further supporting the idea that NOL1 activates transcription factor-binding sites targeted by different signaling pathways (Pestell of cyclin D1 regardless of hTERT expression. et al., 1999). To determine the NOL1-responsive element, a series of To determine whether the occupancy of the cyclin D1 promoter 5′ cyclin D1 promoter deletion constructs were transfected in the by hTERT is dependent on NOL1, HeLa S3 cells expressing Flag– presence of NOL1 expression. As shown in Fig. 5C, cyclin D1 hTERT (or empty vector) and NOL1 shRNAs (or control shRNA) Journal of Cell Science

1571 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

Fig. 5. hTERT stimulates transcription of cyclin D1 gene through the interaction with NOL1. (A) HeLa S3 and (B) U2OS cells expressing control shRNA (shControl) or NOL1 shRNAs (shNOL1-1 and shNOL1-2) were subjected to immunoblotting, to measure the protein levels of cyclin D1 and c-Myc, and semi- quantitative RT-PCR to detect the mRNA levels. (C) The structures of the promoter luciferase (luc) constructs containing various lengths of upstream fragments from the cyclin D1 gene are shown on the left. The binding sites of known transcription factors are indicated. The results of the luciferase assay are shown on the right. HeLa CCL2 cells were co-transfected with the promoter luciferase constructs together with NOL1–V5 or empty vector. The firefly luciferase activity was normalized against the Renilla luciferase activity. Results are mean±s.d. of three independent experiments. (D) The effects of NOL1 and hTERT overexpression on TCF transcriptional activity. HeLa CCL2 cells were transfected with TOP-Flash or FOP-Flash luciferase reporter vectors together with NOL1–V5, Flag–hTERT or both. The firefly luciferase activity was normalized against the Renilla luciferase activity. Results are mean±s.d. of three independent experiments. (E–G) The luciferase assay in H1299 (E), MCF7 (F) and U2OS cells (G) co-transfected with the −964 promoter luciferase reporter vector together with NOL1–V5, Flag– ’ t hTERT or both. Results are mean±s.d. of three independent experiments. Statistical analyses were performed using a two-tailed Student s -test. Journal of Cell Science

1572 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

Fig. 6. NOL1 and hTERT associate with the cyclin D1 promoter at the TBE. (A,B) HeLa S3 cells were transfected with Flag–NOL1 (or empty vector) together with hTERT shRNA (shhTERT-1 or shhTERT2) or control shRNA, and ChIP analyses were performed using anti-Flag antibody. The recruitment of NOL1 to the cyclin D1 TBE (A) or Myc TBE (B) was quantified by performing a gel-based semi-quantitative RT-PCR assay. The 3′-untranslated region (3′-UTR) was used as a negative control. (C) A luciferase assay of cyclin D1 transcription in HeLa CCL2 cells transfected with Flag–NOL1 (or empty vector) together with hTERT shRNA (or control shRNA). The firefly luciferase activity was normalized against the Renilla luciferase activity. Results are mean±s.d. of three independent experiments. (D,E) HeLa S3 cells were transfected with Flag–hTERT (or empty vector) together with NOL1 shRNA (or control shRNA), and ChIP analyses were performed using anti-Flag antibody. The recruitment of NOL1 to the cyclin D1 TBE (D) or Myc TBE (E) was quantified by performing a gel-based semi-quantitative RT-PCR assay. The 3′-UTR was used as a negative control. (F) A luciferase assay of cyclin D1 transcription in HeLa CCL2 cells transfected with Flag–hTERT (or empty vector) together with NOL1 shRNA (or control shRNA). The firefly luciferase activity was normalized against the Renilla luciferase activity. Results are mean±s.d. of three independent experiments. were subjected to ChIP. Flag–hTERT associated with the TBE- depletion (Fig. 6F), suggesting that hTERT activates cyclin D1 containing fragment of the cyclin D1 promoter in cells expressing transcription in a NOL1-dependent manner. control shRNA (Fig. 6D). When endogenous NOL1 was depleted, To further validate the co-occupancy of NOL1 and hTERT in the the ability of Flag–hTERT to bind to the TBE fragment was cyclin D1 promoter, we performed a ChIP-re-ChIP assay (Qiu et al., abrogated. These results suggest that hTERT can be recruited to the 2013). HeLa S3 cells expressing Flag–hTERT and NOL1–V5 were cyclin D1 promoter through the interaction with NOL1. subjected to ChIP with anti-Flag antibody, followed by re-ChIP Intriguingly, the faint but consistent ChIP signal of Flag–hTERT using anti-V5 antibody. Flag–hTERT alone bound to the cyclin D1 was detected at the TBE fragment of the Myc promoter in a NOL1- TBE (Fig. 7A). Interestingly, this binding was increased by co- independent manner (Fig. 6E). We also examined the effect of expression of NOL1–V5. These results are consistent with the NOL1 depletion on the cyclin D1 transcription. As expected, cyclin previous findings that co-expression of both proteins led to an D1 promoter activity was reduced by NOL1 depletion in HeLa additive effect on cyclin D1 promoter activity (see Fig. 5E,F). When CCL2 cells (Fig. 6F). Even in the presence of overexpression of a sequential ChIP (re-ChIP) assay was performed, we observed that hTERT, cyclin D1 promoter activity was also reduced by NOL1 Flag–hTERT associates with the cyclin D1 TBE only in the Journal of Cell Science

1573 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040 presence of NOL1–V5 overexpression (Fig. 7A), further confirming effect of NOL1 depletion on . HeLa S3 cells were that hTERT is recruited to the cyclin D1 TBE through the transduced with the retrovirus particles expressing NOL1 shRNAs interaction with NOL1. In contrast, although Flag–hTERT or control shRNA, and stable cell lines were established from associates with the Myc TBE in a NOL1-independent manner, the separate transductions to monitor the population doubling levels at re-ChIP results demonstrated that NOL1–V5 does not associate with regular intervals. HeLa S3 cells expressing the control shRNA grew the Myc promoter (Fig. 7B). normally and continued to divide throughout the duration of the Transcriptional activation of cyclin D1 gene is regulated by many experiments (Fig. 8A). The growth rates of two independent NOL1- other factors including c-Myc, NF-κB and β-catenin during the cell knockdown cells (shNOL1-1 and shNOL1-2) gradually slowed cycle (Pestell et al., 1999). β-catenin has been shown to occupy the down and stopped dividing after ∼50 population doublings and TBE to activate transcription of cyclin D1 gene in the presence of ∼30 population doublings, respectively. To determine whether the Wnt signaling (Park et al., 2009), whereas c-Myc and NF-κB growth arrest correlates with an altered cell cycle distribution, associate with other binding sites on the cyclin D1 promoter. To NOL1-knockdown cells were subjected to flow cytometric analysis examine whether NOL1 depletion has any effect on β-catenin by propidium iodide staining. During the first few population recruitment to the cyclin D1 TBE, HeLa S3 cells expressing Flag–β- doublings, NOL1-knockdown cells did not exhibit a substantial catenin (or empty vector) and NOL1 shRNA (or control shRNA) change in the cell cycle distribution compared with the control cells were subjected to ChIP. Flag–β-catenin became associated with the (Fig. 8B). By contrast, NOL1-knockdown cells exhibited a TBE at the basal level in a NOL1-independent manner (Fig. 7C). substantial increase in sub-G1 DNA content at later population This association was markedly increased upon stabilization of doublings, a characteristic of apoptosis (Fig. 8C). Thus, the growth β-catenin by lithium chloride (LiCl) treatment but was not affected arrest of NOL1-knockdown cells appeared to be due to an increased by NOL1 depletion (Fig. 7C). rate of cell death. In addition, NOL1-knockdown cells showed increased levels of Bax and cleaved caspase-3, and decreased levels Depletion of NOL1 induces cell growth arrest of anti-apoptotic Bcl-2 protein (Fig. 8D). Taken together, these data Because cyclin D1 plays an important role in the cell cycle support the idea that a growth-arrest phenotype associated with progression through G1 phase (Yu et al., 2001), we determined the NOL1 depletion results from suppression of cyclin D1 transcription.

Fig. 7. hTERT is recruited to the cyclin D1 TBE through the interaction with NOL1. (A,B) HeLa S3 cells expressing Flag– hTERT and NOL1–V5 were subjected to ChIP with anti-Flag antibody, followed by re- ChIP using anti-V5 antibody. The recruitment of Flag–hTERT and NOL1–V5 to the cyclin D1 TBE (A) or c-Myc TBE (B) was quantified by performing a gel-based semi-quantitative RT-PCR assay. The 3′- untranslated region (3′-UTR) was used as a negative control. (C) HeLa S3 cells expressing Flag–β-catenin (or empty vector) and control shRNA (shControl) or NOL1 shRNAs (shNOL1-1 and shNOL1-2) were subjected to ChIP with anti-Flag antibody in the absence or presence of LiCl treatment. The recruitment of Flag–β-catenin to the cyclin D1 TBE was quantified by performing a gel-based semi-quantitative RT-PCR assay. The 3′-UTR was used as a negative control. Journal of Cell Science

1574 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

As a control, we also examined the effect of cyclin D1 depletion on 2002; Li and Tergaonkar, 2014). Here, we identify NOL1 as a new the cell cycle status (Fig. S4A). Depletion of cyclin D1 led to an TERC-binding protein that is found in association with catalytically increase in the proportion of cells in sub-G1 phase compared to the active telomerase. Given that NOL1 has no direct regulatory effect control cells (Fig. S4B,C). on the assembly and trafficking of telomerase or its enzymatic activity, it is likely that NOL1 is involved in a non-telomeric DISCUSSION function of telomerase. We show that NOL1 activates transcription Given that normal human somatic cells express very low levels of of cyclin D1 gene by binding to the TBE. Telomerase is also telomerase, they have a limited proliferative lifespan and ultimately recruited to the cyclin D1 promoter through the interaction with enter a non-dividing state of replicative senescence (Bodnar et al., NOL1, further enhancing transcription of cyclin D1 gene. These 1998). Although ectopic expression of telomerase is sufficient to data suggest that NOL1 represents a new pathway by which extend lifespan, recent studies have suggested that the ability of telomerase activates transcription of the cyclin D1 gene. telomerase to promote tumorigenesis requires additional activities Besides its primary role in telomere extension, telomerase has that are independent of its role in telomere extension (Stewart et al., been demonstrated to have non-canonical functions in signaling

Fig. 8. Depletion of NOL1 induces cell growth arrest. (A) Cell growth curves of HeLa S3 cells stably expressing control shRNA (shControl) or NOL1 shRNAs (shNOL1-1 and shNOL1-2). HeLa S3 cells were infected with the retrovirus particles to establish stable cell lines. Stable cells were replated every 3–4 days to maintain log-phase growth and calculate the growth rate, with day 0 representing the first day after puromycin selection. (B) Flow cytometric analysis of HeLa S3 cells stably expressing control shRNA or NOL1 shRNAs. Cells were stained with propidium iodide after a few or and many population doublings (PD), followed by FACS analysis. (C) The percentage of total cells in sub-G1 phase of the cell cycle is shown. Results are mean±s.d. of three independent experiments. (D) HeLa S3 cells stably expressing control shRNA or NOL1 shRNAs were analyzed by immunoblotting to measure the protein levels of Bcl2, Bax and cleaved caspase-3. (E) Proposed model for the two fates of telomerase during its assembly, telomere extension and transcriptional activation. Journal of Cell Science

1575 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040 pathways that influence human tumorigenesis (Choi et al., 2008; the interaction with TERC, these two proteins might compete for Park et al., 2009; Ghosh et al., 2012). Telomerase has been shown to binding to telomerase in the nucleolus. The outcome of this bind the NF-κB p65 subunit and localize to the promoters of a competition likely determines which of the two fates of telomerase subset of NF-κB target genes (Ghosh et al., 2012). Inhibition of is favored. Although we cannot rule out the possibility that both telomerase reduces the expression of NF-κB-dependent genes, proteins exist in the same telomerase RNP, it is yet unclear what suggesting that telomerase acts as a transcriptional modulator of the fractions of the telomerase RNP contain NOL1 or TCAB1. Recently, NF-κB signaling cascade in cancer cells. Telomerase has also been it has been reported that there are several hundred copies of telomerase found to act as a transcriptional modulator of Wnt/β-catenin RNP in a human cancer cell (Xi and Cech, 2014; Akıncılar et al., signaling pathway and enhance the expression of Wnt target genes 2015). Furthermore, the two telomerase components, hTERT and (Park et al., 2009). Moreover, overexpression of alternatively TERC, appear to be in excess of telomerase RNPs, suggesting the spliced variants that lack telomerase activity stimulate cell existence of unassembled telomerase components. Thus, it will be proliferation by activating Wnt signaling (Hrdlicková et al., interesting to investigate how many molecules of NOL1 exist in a cell 2012). Given that Wnt signaling target genes are also regulated by and how many of these are associated with TERC. other signaling pathways (Guo and Wang, 2009), the mechanism by NOL1 has been shown to be expressed early in the G1 phase and which telomerase enhances the expression of growth-promoting peaks during the S phase (Fonagy et al., 1992, 1993; Gorczyca et al., genes cannot be solely dependent on Wnt signaling. Recently, it 1992). Thus, NOL1-dependent transcriptional activation of cyclin has been reported that telomerase regulates Myc-dependent D1 gene might occur in a cell-cycle-dependent manner. When oncogenesis by stabilizing Myc levels on chromatin (Koh et al., telomerase is upregulated in cancer cells, telomerase could interact 2016). Taken together, these findings suggest that telomerase with NOL1 and occupy the TBE of the cyclin D1 promoter to contributes to activation of growth-promoting genes through further enhance gene expression. Thus, NOL1 plays an important multiple signaling pathways in cancer. In this work, we show that role in cell cycle progression through G1 phase and is implicated as telomerase interacts with NOL1 and promotes transcription of the a tumor cell marker. By contrast, repressing cyclin D1 expression by cyclin D1 gene in the absence of Wnt or NF-κB signaling. Whereas NOL1 depletion prevents the tumor cells exiting from G1 phase, NOL1 alone is sufficient to bind the cyclin D1 promoter and reversing tumor characteristics. Consistent with this idea, depletion promote its transcription, hTERT is recruited to the cyclin of NOL1 induced a growth arrest in telomerase-positive HeLa S3 D1 promoter through its interaction with NOL1, suggesting that cells. This growth arrest was accompanied by several features telomerase activates cyclin D1 transcription in a NOL1-dependent consistent with the induction of apoptosis, including a substantial manner. Whereas TERC is not required for telomerase-dependent increase in sub-G1 DNA content, an increase in the levels of Bax transcriptional activation of Wnt/β-catenin signaling, it is essential and cleaved caspase-3, and a decrease in the levels of anti-apoptotic for transcriptional activation by NOL1 and telomerase. Moreover, Bcl2. These findings suggest that NOL1 plays a key role in the domain mapping analysis has revealed that the TERC-binding control of cell cycle progression through transcriptional activation domain in hTERT is required for NOL1 binding, further supporting of cyclin D1 gene. Overall, our results provide an insight into the the idea that the association between NOL1 and hTERT is new function of NOL1 as an important regulator of cell cycle and dependent on TERC. Thus, a functional interplay between NOL1 cell proliferation, as well as the non-canonical mechanism by which and telomerase modulates the prolonged expression of the cyclin telomerase promotes cell proliferation in cancer cells. D1 gene that is crucial for the maintenance of cell proliferation. Interestingly, it has been recently reported that non-canonical NF- MATERIALS AND METHODS κB signaling can upregulate mutant hTERT promoter activity along Cell culture and plasmids with ETS transcription factors (Li et al., 2015). Reactivated hTERT Human cervical carcinoma HeLa cells and human embryonic kidney can induce the binding of NOL1 to the cyclin D1 promoter by HEK293 cells were grown in Dulbecco’s modified Eagle’s medium setting a potential feed-forward loop in cell proliferation. containing 10% fetal bovine serum with 100 units/ml penicillin and 100 μg/ml streptomycin in 5% CO at 37°C. Human osteosarcoma U2OS NOL1 is a proliferation-related nucleolar protein that is highly 2 cells were grown in McCoy’s modified medium with 10% fetal bovine expressed in most malignant tumor cells but not in normal resting cells μ serum, 100 units/ml penicillin and 100 g/ml streptomycin in 5% CO2 at (Ochs et al., 1988; Jhiang et al., 1990). Given that telomerase is 37°C. Theexpressionvectorsfor NOL1–V5and Flag–NOL1wereconstructed initially assembled in the nucleolus (Lee et al., 2014), the finding that by inserting the full-length NOL1 cDNA into pcDNA 3.1/V5-His (Invitrogen) NOL1 is a new component of catalytically active telomerase suggests and p3xFlag-CMV 7.1 plasmid (Sigma-Aldrich), respectively. The Flag– that the nucleolus could be the site where NOL1 associates with the hTERT expression vector was constructed by cloning the full-length hTERT telomerase holoenzyme. Based on data presented in this work, we cDNA into a pCMV-Tag2 vector (Stratagene). The human cyclin D1 promoter propose a model for the two fates of telomerase during its initial luciferase plasmids were constructed by inserting the various promoter assembly, telomere extension and transcriptional activation (Fig. 8E). fragments into pGL4.20[luc2/Puro] vector (Promega). The expression vectors The assembly of the active telomerase holoenzyme occurs in a highly were transiently transfected using Lipofectamine-PLUS reagent according to the manufacturer’s protocol (Invitrogen). elaborate, stepwise fashion (Lee et al., 2014). After transcription, a TERC molecule assembles with a preformed dyskerin complex, and the subsequent assembly of a TERC–dyskerin ribonucleoprotein Peptide identification using LC-MS/MS (RNP) with hTERT occurs specifically during the S phase in the Nano LC-MS/MS analysis was performed with a nano HPLC system (Agilent) as previously described (Her and Chung, 2013). nucleolus. For telomere extension, telomerase associates with TCAB1 and is transported to Cajal bodies (Venteicher et al., 2009; Venteicher Immunoprecipitation and immunoblotting and Artandi, 2009). Telomerase-containing Cajal bodies are loaded on Immunoprecipitation and immunoblot analyses were performed as telomeric chromatin to elongate telomere repeats. By contrast, when described previously (Lee et al., 2004). Briefly, cells were lysed with lysis NOL1 is recruited to the telomerase RNP in the nucleolus, the NOL1– buffer (0.5% NP-40, 0.5% Triton X-100, 150 mM NaCl, 2 mM EDTA, pH telomerase complex binds to the cyclin D1 promoter at the TBE. 8.0, 50 mM Tris-HCl, pH 7.5 and 10% glycerol) supplemented with a

Given that both TCAB1 and NOL1 associate with telomerase through proteinase inhibitor cocktail (Roche) for 15 min at 4°C, followed by Journal of Cell Science

1576 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040 centrifugation (16,000 g) for 15 min at 4°C. Lysates were pre-cleared with buffer (low-salt buffer containing 500 mM NaCl), LiCl buffer (0.25 M protein-A–Sepharose beads (GE Healthcare) for 30 min at 4°C. After LiCl, 0.5% NP-40, 0.5% Na-deoxycholate, 1 mM EDTA, 10 mM Tris-HCl, centrifugation, the supernatants were incubated with primary antibodies at pH 8.0), and twice with TE buffer. For the re-ChIP assay, immune 4°C overnight, followed by incubation with protein-A–Sepharose beads for complexes were eluted from the agarose beads using 20 mM dithiothreitol 1 h at 4°C. After binding, the beads were washed extensively with lysis and used for a secondary immunoprecipitation. DNA and proteins were buffer and subjected to immunoblot analysis. Immunoprecipitation and recovered after incubation in 1% SDS and 0.1 M NaHCO3 at 65°C, for 2 h. immunoblotting were performed using anti-Flag (1:10,000, cat. no. F-1804, Then, NaCl was added to a 10 mM final concentration and protein–DNA Sigma-Aldrich), anti-V5 (1:5000, cat. no. 46-0575, Invitrogen), anti-NOL1 cross-links were reversed by incubating samples at 65°C overnight. Finally, (1:2000, cat. no. NBP1-92192, Novus Biologicals), anti-cyclin D1 (1:2000, samples were digested with proteinase K at 45°C for 1 h. The isolated DNAs cat. no. ab16663, Abcam), anti-c-Myc (1:1000, cat. no. sc-764, Santa Cruz were analyzed by semi-quantitative RT-PCR. The primers for amplification Biotechnology), anti-TCAB1 (1:3000, cat. no. ab99376, Abcam), anti- of the TBE and 3′-UTR regions of each gene promoter were as follows: dyskerin (1:3000, cat. no. sc-48794, Santa Cruz Biotechnology), anti- cyclin D1 TBE (forward, 5′-CGCTCCCATTCTCTGCCGGG-3′ and reverse, hTERT (1:1000, cat. no. 600-401-252, Rockland), anti-tubulin (1:1000, cat. 5′-CCGCGCTCCCTCGCGCTCTT-3′), cyclin D1 3′-UTR (forward, 5′-C- no. sc-8035, Santa Cruz Biotechnology), anti-TRF1 (1:1000, cat. no. sc- AAGAGAAGATTACCGCCCGAG-3′ and reverse, 5′-TCCCCAGCCTTT- 1977, Santa Cruz Biotechnology), anti-TRF2 (1:1000, cat. no. D1Y5D, Cell TTGACACC-3′), c-Myc TBE (forward, 5′-CGTCTAGCACCTTTGATTT- Signaling), anti-RAP1 (1:3000, cat. no. A300-306A, Bethyl Laboratory), CTCCC-3′ and reverse, 5′-CTCTGCCAGTCTGTACCCCACCGT-3′), and anti-POT1 (1:3000, cat. no. ab21382, Abcam), anti-TPP1 (1:1000, cat. no. c-Myc 3′-UTR (forward, 5′-CTAATGTATCACAAAGTCCTTTA-3′ and ab54685, Abcam), anti-TIN2 (1:1000, cat. no. ab136997, Abcam), anti- reverse, 5′-GTGATCTGCTCTGTTAGCTTTTGA-3′). Bcl2 (1:1000, cat. no. sc-492, Santa Cruz Biotechnology), anti-Bax (1:1000, cat. no. sc-493, Santa Cruz Biotechnology), anti-cleaved Dual luciferase reporter assay caspase-3 (1:1000, cat. no. 9661, Cell Signaling) antibodies as specified. HeLa cells were transiently co-transfected with cyclin D1 promoter All the immunoblots are representatives of at least three experiments, which luciferase reporter constructs, together with empty vector or NOL1-V5. demonstrated similar results. At 24 h after transfection, cells were lysed, and firefly and Renilla luciferase activities were measured using the dual-luciferase reporter assay system Semi-quantitative RT-PCR (Promega). The firefly luciferase activity for each sample was normalized Total RNA was isolated using Easy BLUE (Intron). The reverse based on transfection efficiency as measured by Renilla luciferase activity. transcription reaction was performed with 1 μg of total RNA using the M- The results are expressed with the standard deviation from the mean of three MLV reverse transcriptase (Promega), and cDNA was used for semi- independent experiments. quantitative RT-PCR. The following primers were used: hTERT (forward, 5′-CGGAAGAGTGTCTGGAGCAA-3′ and reverse, 5′-GGATGAAGC- Establishment of stable cell lines GGAGTCTGGA-3′), TERC (forward, 5′-TCTAACCCTAACTGAGAA- The retrovirus vectors were constructed by cloning the shRNA-expressing GGGCGTAG-3′ and reverse, 5′-GTTTGCTCTAGA ATGAACGGTGG- oligonucleotides targeting NOL1 (5′-GATCCCCGCGTTGCTGCCCAT- AAG-3′), dykerin (forward, 5′-ACAGGGTGAAGAGTTCTGGCACAT-3′ TGAAATTTTCAAGAGAAATTTCAATGGGCAGCAACGCTTTTTA-3′ and reverse, 5′-TGAAGGTGAGGCTTCCCAACTCAA-3′), cyclin D1 for shNOL1-1; 5′-GATCCCCGGACGATGCTGATACGGTATTTTCAA- (forward, 5′-CACACGGACTACAGGGGAGT-3′ and reverse, 5′-CACA- GAGAAATACCGTATCAGCATCGTCCTTTTTA-3′ for shNOL1-2) into GGAGCTGGTGTTCCAT-3′), c-Myc (forward, 5′-AATGAAAAGGCC- pSUPER.retro.puro vector (OligoEngine). The retrovirus vectors expressing CCCAAGGTAGTTATCC-3′ and reverse, 5′-GTCGTTTCCGCAACAA- NOL1 shRNAs were co-transfected with pGP (for gag-pol expression) and GTCCTCTTC-3′), and GAPDH (forward, 5′-CTCAGACACCATGGGG- pE-ampho (for env expression) into HEK293T packaging cells according to AAGGTGA-3′ and reverse, 5′-ATGATCTTGAGGCTGTTGTCATA-3′). the manufacturer’s instructions (Takara). After 48 h, the culture super- natants were harvested and filtered through a 0.45 μm filter. To generate GST pulldown assay with in vitro transcribed TERC stable cell lines, HeLa S3 cells were transduced with the viral supernatants TERC transcripts were prepared by in vitro transcription using a MAXIscript containing 4 μg/ml polybrene (Sigma-Aldrich). After selection with 1 μg/ml T7/T3 Kit and cleared by MEGAclear Kit according to the manufacturer’s puromycin (Gibco) for 2 weeks, multiple independent single clones were recommendation (Ambion). For GST pulldown assays, GST fusion proteins isolated and checked for NOL1 expression. (2 μg) were incubated with in vitro transcribed TERC (200 ng) in reaction buffer (10 mM HEPES, 3 mM KCl, 150 mM NaCl, 1 mM MgCl2) Fluorescence-activated cell sorting analysis supplemented with a proteinase inhibitor cocktail (Roche) and RNasin HeLa S3 cells were washed with PBS and fixed for 30 min in ice-cold 70% (20 U/ml, Promega). Complexes between GST fusion proteins and TERC ethanol. The fixed cells were resuspended in PBS containing RNase A were isolated with glutathione–Sepharose-4B (GE Healthcare) and used for (200 μg/ml) and propidium iodide (50 μg/ml), and incubated in the dark for semi-quantitative RT-PCR to detect TERC. 30 min at room temperature. Cell cycle progression was monitored by flow cytometry using a FACScan flow cytometer (BD Biosciences). Telomerase assay The telomeric repeat amplification protocol (TRAP) was used as previously Immunofluorescence and telomere fluorescence in situ described (Kim et al., 2003). hybridization Cells grown on glass coverslips were fixed, permeabilized and blocked as Chromatin immunoprecipitation assay described previously (Abreu et al., 2011). Cells were incubated with rabbit HeLa S3 cells transfected with either Flag–NOL1 or Flag–hTERT were anti-TERT (500 ng/ml, cat. no. 600-401-252, Rockland), mouse anti-coilin cross-linked with 1% formaldehyde in phosphate-buffered saline (PBS) for (2 μg/ml, cat. no. ab11822, Abcam), and mouse anti-nucleolin (200 ng/ml, 30 min at room temperature and neutralized with addition of 125 mM cat. no. sc-17826, Santa Cruz Biotechnology) for 16 h at 4°C. After glycine. Cells were lysed in lysis buffer (50 mM Tris-HCl, pH 8.0, 1% SDS, thorough washing with PBS, cells were incubated with Alexa-Fluor-488- 10 mM EDTA) supplemented with a proteinase inhibitor cocktail (Roche). conjugated anti-rabbit immunoglobulin (green) (cat. no. A-11034, Thermo Chromatin was sonicated to obtain chromatin fragments with an average size Fisher) and Alexa-Fluor-568-conjugated anti-mouse immunoglobulin (red) of 600 bp, as assessed by electrophoresis on a 1% agarose gel. Sonicated (cat. no. A-11004, Thermo Fisher) for 1 h in the dark. The coverslips were samples were used for ChIP immunoprecipitation with Flag M2 agarose mounted on microscope slides using Vectashield mounting medium with beads (Sigma-Aldrich) overnight at 4°C and pulled down by centrifugation DAPI (Vector Laboratories). Immunofluorescence images were captured (16,000 g) for 60 s. Chromatin–Flag-M2-agarose complexes were using a confocal laser-scanning microscope LSM 510 (Carl Zeiss). sequentially washed once with low-salt buffer (0.1% SDS, 1% Triton X- Telomere FISH staining was performed with Cy3-(CCCTAA)3 PNA

100, 2 mM EDTA, 150 mM NaCl, 20 mM Tris-HCl, pH 8.0), high-salt probe (Panagene) as previously described (van Steensel et al., 1998). Journal of Cell Science

1577 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

Acknowledgements Jegou, T., Chung, I., Heuvelman, G., Wachsmuth, M., Görisch, S. M., Greulich- We are grateful to Sun Ah Jeong, Prabhat Khadka, Joonyoung Her and Yu Young Bode, K. M., Boukamp, P., Lichter, P. and Rippe, K. (2009). Dynamics of Jeong for technical assistance and helpful comments on the manuscript. telomeres and promyelocytic leukemia nuclear bodies in a telomerase-negative human cell line. Mol. Biol. Cell 20, 2070-2082. Competing interests Jhiang, S. M., Yaneva, M. and Busch, H. (1990). Expression of human proliferation-associated nucleolar antigen p120. Cell Growth Differ. 1, The authors declare no competing or financial interests. 319-324. Kim, N. W., Piatyszek, M. A., Prowse, K. R., Harley, C. B., West, M. D., Ho, P. L., Author contributions Coviello, G. M., Wright, W. E., Weinrich, S. L. and Shay, J. W. (1994). Specific J.H. and J.H.L. designed and performed experiments, analyzed data and assisted in association of human telomerase activity with immortal cells and cancer. Science manuscript preparation. I.K.C. conceived and designed the experiments, analyzed 266, 2011-2015. data and wrote the manuscript. Kim,J.H.,Kim,J.H.,Lee,G.E.,Lee,J.E.andChung,I.K.(2003). Potent inhibition of human telomerase by nitrostyrene derivatives. Mol. Pharmacol. 63, Funding 1117-1124. This work was supported in part by a grant from the National Research Foundation of Koh, C. M., Khattar, E., Leow, S. C., Liu, C. Y., Muller, J., Ang, W. X., Li, Y., Korea [grant number NRF-2013M3A9B6076413 to I.K.C.]; and by a Yonsei Franzoso, G., Li, S., Guccione, E. et al. (2016). Telomerase regulates MYC- University internal grant [grant number 2014-22-0096 to I.K.C.]. driven oncogenesis independent of its reverse transcriptase activity. J. Clin. Invest. 125, 2109-2122. Koonin,E.V.(1994). Prediction of an rRNA methyltransferase domain in Supplementary information human tumor-specific nucleolar protein p120. Nucleic Acids Res. 22, Supplementary information available online at 2476-2478. http://jcs.biologists.org/lookup/suppl/doi:10.1242/jcs.181040/-/DC1 Lee, G. E., Yu, E. Y., Cho, C. H., Lee, J., Muller, M. T. and Chung, I. K. (2004). DNA-protein kinase catalytic subunit-interacting protein KIP binds telomerase by References interacting with human telomerase reverse transcriptase. J. Biol. Chem. 279, Abreu, E., Terns, R. M. and Terns, M. P. (2011). Visualization of human telomerase 34750-34755. localization by fluorescence microscopy techniques. Methods Mol. Biol. 735, Lee, J. H., Khadka, P., Baek, S. H. and Chung, I. K. (2010). CHIP promotes 125-137. human telomerase reverse transcriptase degradation and negatively regulates Akıncılar, S. C., Low, K. C., Liu, C. Y., Yan, T. D., Oji, A., Ikawa, M., Li, S. and telomerase activity. J. Biol. Chem. 285, 42033-42045. Tergaonkar, V. (2015). Quantitative assessment of telomerase components in Lee, J. H., Lee, Y. S., Jeong, S. A., Khadka, P., Roth, J. and Chung, I. K. cancer cell lines. FEBS Lett. 589, 974-984. (2014). Catalytically active telomerase holoenzyme is assembled in the dense Autexier, C. and Lue, N. F. (2006). The structure and function of telomerase reverse fibrillar component of the nucleolus during S phase. Histochem. Cell Biol. 141, transcriptase. Annu. Rev. Biochem. 75, 493-517. 137-152. Bianchi, A. and Shore, D. (2008). How telomerase reaches its end: mechanism Li, Y. and Tergaonkar, V. (2014). Noncanonical functions of telomerase: implications of telomerase regulation by the telomeric complex. Mol. Cell 31, 153-165. in telomerase-targeted cancer therapies. Cancer Res. 74, 1639-1644. Blackburn, E. H. (2001). Switching and signaling at the telomere. Cell 106, Li, Y., Zhou, Q.-L., Sun, W., Chandrasekharan, P., Cheng, H. S., Ying, Z., 661-673. Lakshmanan, M., Raju, A., Tenen, D. G., Cheng, S. Y. et al. (2015). Non- Blasco, M. A., Lee, H.-W., Hande, M. P., Samper, E., Lansdorp, P. M., DePinho, canonical NF-κB signalling and ETS1/2 cooperatively drive C250T mutant TERT R. A. and Greider, C. W. (1997). Telomere shortening and tumor formation by promoter activation. Nat. Cell Biol. 17, 1327-1338. mouse cells lacking telomerase RNA. Cell 91, 25-34. Lingner, J., Cooper, J. P. and Cech, T. R. (1995). Telomerase and DNA end Bodnar, A. G., Ouellette, M., Frolkis, M., Holt, S. E., Chiu, C.-P., Morin, G. B., replication: no longer a lagging strand problem? Science 269, 1533-1534. Harley, C. B., Shay, J. W., Lichtsteiner, S. and Wright, W. E. (1998). Extension of Listerman, I., Gazzaniga, F. S. and Blackburn, E. H. (2014). An investigation of the life-span by introduction of telomerase into normal human cells. Science 279, effects of the core protein telomerase reverse transcriptase on Wnt signaling in 349-352. breast cancer cells. Mol. Cell. Biol. 34, 280-289. Choi, J., Southworth, L. K., Sarin, K. Y., Venteicher, A. S., Ma, W., Chang, W., Musgrove, E. A., Caldon, C. E., Barraclough, J., Stone, A. and Sutherland, Cheung, P., Jun, S., Artandi, M. K., Shah, N. et al. (2008). TERT promotes R. L. (2011). Cyclin D as a therapeutic target in cancer. Nat. Rev. Cancer 11, epithelial proliferation through transcriptional control of a Myc- and Wnt-related 558-572. developmental program. PLoS Genet. 4, e10. Ochs, R. L., Reilly, M. T., Freeman, J. W. and Busch, H. (1988). Intranucleolar Cancer Res. D’Adda di Fagagna, F., Reaper, P. M., Clay-Farrace, L., Fiegler, H., Carr, P., Von localization of human proliferating cell nucleolar antigen p120. 48, Zglinicki, T., Saretzki, G., Carter, N. P. and Jackson, S. P. (2003). A DNA 6523-6529. damage checkpoint response in telomere-initiated senescence. Nature 426, Palm, W. and de Lange, T. (2008). How shelterin protects mammalian telomeres. Annu. Rev. Genet. 194-198. 42, 301-334. Egan, E. D. and Collins, K. (2012). Biogenesis of telomerase ribonucleoproteins. Park, J.-I., Venteicher, A. S., Hong, J. Y., Choi, J., Jun, S., Shkreli, M., Chang, W., Meng, Z., Cheung, P., Ji, H. et al. (2009). Telomerase modulates Wnt signalling RNA 18, 1747-1759. Nature Fonagy, A., Swiderski, C., Dunn, M. and Freeman, J. W. (1992). Antisense- by association with target gene chromatin. 460, 66-72. Perlaky, L., Valdez, B. C., Busch, R. K., Larson, R. G., Jhiang, S. M., Zhang, mediated specific inhibition of p120 protein expression prevents G1- to S-phase W. W., Brattain, M. and Busch, H. (1992). Increased growth of NIH/3T3 cells by transition. Cancer Res. 52, 5250-5256. transfection with human p120 complementary DNA and inhibition by a p120 Fonagy, A., Swiderski, C., Wilson, A., Bolton, W., Kenyon, N. and Freeman, antisense construct. Cancer Res. 52, 428-436. J. W. (1993). Cell cycle regulated expression of nucleolar antigen p120 in normal Pestell, R. G., Albanese, C., Reutens, A. T., Segall, J. E., Lee, R. J. and and transformed human fibroblasts. J. Cell Physiol. 154, 16-27. Arnold, A. (1999). The cyclins and cyclin-dependent kinase inhibitors in Ghosh, A., Saginc, G., Leow, S. C., Khattar, E., Shin, E. M., Yan, T. D., Wong, M., hormonal regulation of proliferation and differentiation. Endocr. Rev. 20, Zhang, Z., Li, G., Sung, W.-K. et al. (2012). Telomerase directly regulates NF-kB- 501-534. Nat. Cell Biol. dependent transcription. 14, 1270-1281. Qiu, R., Yang, Y., Zhao, H., Li, J., Xin, Q., Shan, S., Liu, Y., Dang, J., Yu, X., Gong, Gorczyca, W., Bruno, S., Melamed, M. R. and Darzynkiewicz, Z. (1992). Cell Y. et al. (2013). Signal transducer and activator of transcription 6 directly regulates cycle-related expression of p120 nucleolar antigen in normal human lymphocytes human ORMDL3 expression. FEBS J. 280, 2014-2026. and in cells of HL-60 and MOLT-4 leukemic lines: effects of methotrexate, Sarin, K. Y., Cheung, P., Gilison, D., Lee, E., Tennen, R. I., Wang, E., Artandi, Cancer Res. camptothecin, and teniposide. 52, 3491-3494. M. K., Oro, A. E. and Artandi, S. E. (2005). Conditional telomerase induction Guo, X. and Wang, X.-F. (2009). Signaling cross-talk between TGF-beta/BMP and causes proliferation of hair follicle stem cells. Nature 436, 1048-1052. Cell Res. other pathways. 19, 71-88. Sears, R. C. (2004). The life cycle of C-myc: from synthesis to degradation. Cell Gustafson, W. C., Taylor, C. W., Valdez, B. C., Henning, D., Phippard, A., Cycle 3, 1131-1135. Ren, Y., Busch, H. and Durban, E. (1998). Nucleolar protein p120 contains Sfeir, A. and de Lange, T. (2012). Removal of shelterin reveals the telomere end- an arginine-rich domain that binds to ribosomal RNA. Biochem. J. 331, protection problem. Science 336, 593-597. 387-393. Smith, L. L., Coller, H. A. and Roberts, J. M. (2003). Telomerase modulates Her, Y. R. and Chung, I. K. (2013). p300-mediated acetylation of TRF2 is expression of growth-controlling genes and enhances cell proliferation. Nat. Cell required for maintaining functional telomeres. Nucleic Acids Res. 41, Biol. 5, 474-479. 2267-2283. Smogorzewska, A. and de Lange, T. (2004). Regulation of telomerase by Hrdlicková, R., Nehyba, J. and Bose, H. R. (2012). Alternatively spliced telomeric proteins. Annu. Rev. Biochem. 73, 177-208. telomerase reverse transcriptase variants lacking telomerase activity stimulate Stewart, S. A., Hahn, W. C., O’Connor, B. F., Banner, E. N., Lundberg, A. S.,

cell proliferation. Mol. Cell. Biol. 32, 4283-4296. Modha, P., Mizuno, H., Brooks, M. W., Fleming, M., Zimonjic, D. B. et al. Journal of Cell Science

1578 RESEARCH ARTICLE Journal of Cell Science (2016) 129, 1566-1579 doi:10.1242/jcs.181040

(2002). Telomerase contributes to tumorigenesis by a telomere length- Venteicher, A. S. and Artandi, S. E. (2009). TCAB1: driving telomerase to Cajal independent mechanism. Proc. Natl. Acad. Sci. USA 99, 12606-12611. bodies. Cell Cycle 8, 1329-1331. Strong, M. A., Vidal-Cardenas, S. L., Karim, B., Yu, H., Guo, N. and Greider, Venteicher, A. S., Meng, Z., Mason, P. J., Veenstra, T. D. and Artandi, S. E. − − − C. W. (2011). Phenotypes in mTERT+/ and mTERT / mice are due to short (2008). Identification of ATPases pontin and reptin as telomerase components Cell telomeres, not telomere-independent functions of telomerase reverse essential for holoenzyme assembly. 132, 945-957. Venteicher, A. S., Abreu, E. B., Meng, Z., McCann, K. E., Terns, R. M., Veenstra, transcriptase. Mol. Cell. Biol. 31, 2369-2379. T. D., Terns, M. P. and Artandi, S. E. (2009). A human telomerase holoenzyme Takai, H., Smogorzewska, A. and de Lange, T. (2003). DNA damage foci at protein required for Cajal body localization and telomere synthesis. Science 323, Curr. Biol. dysfunctional telomeres. 13, 1549-1556. 644-648. Tomlinson, R. L., Ziegler, T. D., Supakorndej, T., Terns, R. M. and Terns, M. P. Xi, L. and Cech, T. R. (2014). Inventory of telomerase components in human cells Mol. (2006). Cell cycle-regulated trafficking of human telomerase to telomeres. reveals multiple subpopulations of hTR and hTERT. Nucleic Acids Res. 42, Biol. Cell 17, 955-965. 8565-8577. van Steensel, B., Smogorzewska, A. and de Lange, T. (1998). TRF2 protects Yu, Q., Geng, Y. and Sicinski, P. (2001). Specific protection against breast cancers human telomeres from end-to-end fusions. Cell 92, 401-413. by cyclin D1 ablation. Nature 411, 1017-1021. Journal of Cell Science

1579