.
F U L L T E X T S O U R C E : Cell Reports
Summary
Biogenesis of the human telomerase RNA (hTR) involves a complex series of posttranscriptional modifications, including hypermethylation of the 5′ mono-methylguanosine cap to a tri-methylguanosine cap (TMG). How the TMG cap affects hTR maturation is unknown. Here, we show that depletion of trimethylguanosine synthase 1 (TGS1), the enzyme responsible for cap hypermethylation, increases levels of hTR and telomerase. Diminished trimethylation increases hTR association with the cap-binding complex (CBC) and with Sm chaperone proteins. Loss of TGS1 causes an increase in accumulation of mature hTR in both the nucleus and the cytoplasm compared with controls. In TGS1 mutant cells, increased hTR assembles with telomerase reverse transcriptase (TERT) protein to yield elevated active telomerase complexes and increased telomerase activity, resulting in telomere elongation in cultured human cells. Our results show that TGS1-mediated hypermethylation of the hTR cap inhibits hTR accumulation, restrains levels of assembled telomerase, and limits telomere elongation.
Introduction
Telomere homeostasis is a major determinant for replicative lifespan, cellular senescence, and tumor progression (Blackburn et al., 2015). Human telomeres consist of arrays of short repetitive sequences at chromosome ends and are shielded from the DNA repair machinery by specialized capping complexes (Palm and de Lange, 2008). Telomere repeats are added by telomerase, an enzyme whose catalytic core is comprised of the telomerase reverse transcriptase (TERT) catalytic subunit and the human telomerase RNA (hTR) template RNA. While hTR is broadly expressed, the expression of TERT is restricted to stem cells and progenitor cells (Wright et al., 1996); telomere elongation occurs only in cells expressing active telomerase (Cristofari and Lingner, 2006). Haploinsufficiency of either TERT or hTR causes pathologic telomere shortening and leads to the stem cell disease dyskeratosis congenita and other telomere-related diseases (Armanios and Blackburn, 2012, Armanios et al., 2005, Batista et al., 2011, Marrone et al., 2004), suggesting that not only the TERT level but also the hTR level is a limiting factor for telomerase activity. Defining the mechanisms that regulate hTR biogenesis and its assembly into telomerase is critically important for our understanding of telomere-related pathologies and telomerase regulation in cancer (Rousseau and Autexier, 2015).
Human hTR is a 451 nt RNA synthesized by RNA polymerase II (Pol II) that acquires a monomethylguanosine (MMG) cap during the early stages of transcription. This MMG cap is further methylated to a N2, 2, 7 trimethylguanosine (TMG) cap, by trimethylguanosine synthase 1 (TGS1), an evolutionarily conserved enzyme that modifies several classes of noncoding RNAs, including small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNas), some viral RNAs, and selenoprotein mRNAs (Mouaikel et al., 2002, Pradet-Balade et al., 2011, Wurth et al., 2014, Yedavalli and Jeang, 2010). Unlike classical Pol II transcripts, hTR lacks a canonical polyadenylation signal and is processed to generate a defined 3′ end.
The 3′ end of hTR contains an H/ACA motif consisting of two hairpins and two single-stranded regions, the hinge and the ACA containing tail (Kiss et al., 2006, Mitchell et al., 1999). The H/ACA motif, which is found also in small Cajal body RNAs (scaRNAs) and in some snoRNAs, is bound cotranscriptionally by the dyskerin (DKC1)-NOP10-NHP2-NAF1 complex that defines the 3′ end of hTR and stabilizes hTR transcripts (Fu and Collins, 2007, MacNeil et al., 2019, Shukla et al., 2016). Mutations in DKC1, NOP10, or NHP2 lead to dyskeratosis congenita (DC), by impairing telomerase and causing telomere shortening (Armanios and Blackburn, 2012).
hTR is initially transcribed as an extended precursor that is trimmed by 3′-5′ RNA exonucleases to generate its mature 451 nt form. hTR transcripts as long as 1,500 nt have been detected, although it is unclear whether these ultra-long transcripts are processed to mature hTR or whether they are aberrantly terminated transcripts removed by nuclear RNA surveillance through the RNA exosome (Nguyen et al., 2015, Tseng et al., 2015, Tseng et al., 2018). Many hTR precursors have 8–10 nt genomically encoded 3′ extensions and are trimmed to generate mature hTR (Goldfarb and Cech, 2013, Roake et al., 2019). These precursors are primarily oligoadenylated by the noncanonical poly(A)polymerase PAPD5 (Moon et al., 2015, Tseng et al., 2015). Oligoadenylated hTR intermediates can either be degraded by the RNA exosome or have their A tails removed by the poly(A)ribonuclease PARN. Patients with biallelic germline mutations in PARN develop DC and idiopathic pulmonary fibrosis (IPF), downstream of telomere shortening (Moon et al., 2015, Stuart et al., 2015, Tummala et al., 2015). In the absence of PARN, oligoadenylated hTR precursors accumulate; the maturation rate of hTR slows, and stalled hTR precursors are degraded causing an overall loss of hTR and telomerase. However, in the absence of both PARN and PAPD5, the maturation of hTR precursors normalizes, indicating that oligoadenylation of hTR precursors governs the maturation rate of hTR (Roake et al., 2019).
Oligoadenylated hTR precursors that are not processed to mature hTR are subject to degradation by the exosome, an RNase complex composed of ten core proteins including the 3′-5′ exonuclease Dis3 and the endonuclease Rrp44. The core exosome also interacts with the 3′-5′ exonuclease Rrp6, which functions similarly to Dis3 (Chlebowski et al., 2013). Exosome substrates are recruited by the nuclear exosome targeting complex (NEXT) together with the cap binding complex (CBC), which includes the CBP20, CBP80, and ARS2 subunits (Andersen et al., 2013, Mitchell, 2014). Loss of CBC, exosome subunits, or some specific NEXT components, causes an increase in hTR levels, suggesting that this pathway removes a subset of maturing hTR transcripts (Gable et al., 2019, Shukla et al., 2016, Tseng et al., 2015). In addition, the RNA exosome is involved in degradation of improperly assembled hTR complexes, because exosome knockdown can rescue the hTR loss caused by dyskerin deficiency or mutations (Boyraz et al., 2016, Fok et al., 2019, Shukla et al., 2016).
Beyond its role in hTR biogenesis and stability, the H/ACA domain of hTR also contains a 3-nt sequence called the CAB box that binds additional proteins. TCAB1 binds the CAB box and directs localization of hTR to Cajal bodies (CBs), nuclear structures devoted to RNA modification and assembly. TCAB1 binding to the CAB box is required for full catalytic activity of telomerase (Chen et al., 2018, Freund et al., 2014, Tycowski et al., 2009, Zhong et al., 2011) and for telomerase recruitment at telomeres during the S phase (Cristofari et al., 2007, Tomlinson et al., 2006, Venteicher et al., 2009). The CAB box also associates with a subset of the seven Sm proteins (Fu and Collins, 2006), which form a heteroheptameric ring that encircles and stabilizes several coding and noncoding RNA species, and SmB directly interacts with TGS1 (Mouaikel et al., 2003a, Mouaikel et al., 2003b). In fission yeast, the Sm proteins bind telomerase RNA and contribute to its maturation and stability (Tang et al., 2012).
CBs are also the sites in which hypermethylation of hTR by TGS1 is thought to occur (Fu and Collins, 2006, Girard et al., 2008, Jády et al., 2004). The role of the cap hypermethylation step in telomerase biogenesis and/or trafficking is unknown. Human TGS1 exists as two isoforms, a long (TGS1-LF) and a short (TGS1-SF) isoform (Girard et al., 2008). The TGS1 LF C-terminal portion contains a highly conserved methyltransferase domain and is present both in the cytoplasm and the nuclear CBs (Girard et al., 2008). The short isoform consists only of the C terminus of the protein and it is exclusively enriched in CBs. The two isoforms have different repertoires of RNA targets: TGS1 LF is thought to hypermethylate snRNAs, while TGS1 SF is believed to be specific for snoRNAs. TGS1 LF has been also implicated in the trafficking of these noncoding RNAs through its interaction with the nuclear export factor CRM1 (Boulon et al., 2004, Verheggen and Bertrand, 2012). The TMG cap synthesized by TGS1 LF and the Sm core proteins bound to SMN form a bipartite nuclear targeting signal that directs import of snRNAs and SMN (Hamm et al., 1990, Narayanan et al., 2004). In addition, it has been proposed that CRM1 modulates the interaction between TGS1 LF and hTR (Pradet-Balade et al., 2011).
Here, we investigate the role of TGS1 in regulating hTR biogenesis, stability, and assembly into functional telomerase. By introducing frameshift mutations in the TGS1 gene in human cancer cells using CRISPR/Cas9 genome editing, we create cells depleted of the TGS1 protein. We use these cells to understand how hTR 5′ cap trimethylation controls hTR biogenesis, telomerase levels and telomere length.
Results
TGS1 Is Required for Proper CB Organization and hTR Subcellular Localization
To investigate the role of cap hypermethylation in hTR biosynthesis and telomerase assembly in HeLa cells, we used CRISPR/Cas9 genomic editing to introduce insertions and deletions within both the first and the eighth TGS1 exon, which affect both TGS1 isoforms. We isolated three clones (TGS1 CRISPR M1, M2, and M3) displaying strongly reduced expression of TGS1. Western blotting showed that in these three clones the TGS1 levels are less than 10% of the level observed in parental HeLa cells (Figure 1A). Stable lentiviral expression of FLAG-tagged TGS1 restored TGS1 to levels that approximate those of the endogenous TGS1 protein (M1R, M2R, and M3R correspond to M1, M2 and M3 knockout clones rescued with FLAG-TGS1) (Figure 1A).
Figure 1. TGS1 Loss Affects hTR Localization
(A) Western blot (WB) with an anti-TGS1 antibody on extracts from a TGS1-proficient HeLa cell line, and three independent CRISPR-derived TGS1 mutant clones (M1, M2, and M3) expressing a FLAG-TGS1 rescue construct at a similar level as endogenous TGS1 (endo-TGS1). β-tubulin is a loading control (CTR). See also Table S1.
(B and C) Examples of TGS1 mutant cells (M1) showing reductions in hTR (B) and scaU93 RNA © foci compared to TGS1 M1 cells expressing the FLAG-TGS1 rescue construct (M1R), and accumulation in the nucleoli of hTR and scaU93, which were detected by RNA FISH (red); nucleoli appear as DAPI-dim areas.
(D–G) Average numbers of hTR (D) and U93 (F) foci observed in M (M1, M2, and M3), MR (M1R, M2R, and M3R), and CTR (HeLa parental line) cells. Frequencies of cells with nucleolar accumulations of hTR (E) and U93 (G) in the same cell samples. Data are the means of 3 mutant clones and 3 CTR cell samples. ∗p < 0.05; ∗∗∗∗p < 0.0001, one-way ANOVA.
Because previous studies had shown that TGS1 depletion disrupts CB formation in HeLa cells (Lemm et al., 2006, Roithová et al., 2018), we first tested our TGS1 CRISPR cells for the presence and integrity of CBs using immunofluorescence with antibodies against coilin and TCAB1, which co-localize in CBs (Venteicher et al., 2009) (Figure S1A). While TGS1-proficient HeLa cells showed an average of 2.5 coilin/TCAB1 foci per cell (n = 473), TGS1-deficient cells exhibited only 1.2 coilin/TCAB1 foci/cell (n = 942). The reduction in the CB number was rescued by stable expression of FLAG-TGS1 (2 coilin/TCAB1 foci/cell; n = 1,115) (Figure S1B). In TGS1 mutant cells, the coilin/TCAB1 signals were more diffuse than in controls and formed aggregates of irregular size and shape. These aggregates colocalized with nucleoli in 20% of the nuclei, whereas only 4% of control nuclei displayed nucleolar Coilin/TCAB1 signals (Figures S1A and S1B). These results show that TGS1 is required for CB integrity, likely reflecting an underlying defect in snRNA biogenesis caused by TGS1 loss (Lemm et al., 2006).
We next performed RNA fluorescence in situ hybridization (FISH) to determine the effects of TGS1 depletion on the subcellular localization of hTR, which is normally strongly enriched in the CBs of HeLa cells. Consistent with previous results (Zhu et al., 2004), we found that parental HeLa cells exhibit an average of 2.7 hTR foci/nucleus (n = 142), whereas TGS1 mutant cells (M1 plus M2; n = 263) contain 1.0 hTR focus per nucleus (Figures 1B and 1D). However, TGS1-deficient cells displayed prominent hTR accumulation in nucleoli (Figures 1B, 1E, and S1C). Specifically, while 41% of TGS1 mutant cells showed enrichments of hTR in the nucleoli, only 4% and 13% of control and rescued cells showed similar enrichments, respectively (Figure 1E). TGS1-deficiency also affected localization of scaRNAs. The U93 scaRNA was enriched in CBs of both control cells (2.4 foci/cell, n = 118) and TGS1-rescued cells (2.8 foci/cell n = 344) (Figures 1C and 1F), but showed diminished CB localization (0.7 foci/cell, n = 363) and nucleolar mislocalization in TGS1 M1 and M2 cells (Figures 1F and 1G). Disruption of TGS1 in TGS1 M1 cells led to loss of CB localization of the U2 snRNA (Figure S2A). As expected, normal localization of small nucleolar RNA (snoRNA) in the nucleoli was not affected by TGS1 loss (Figure S2B).
TGS1 Regulates the Abundance of hTR Molecules
To determine whether TGS1 loss affects the abundance of hTR molecules, we analyzed hTR levels by northern blotting using total RNA isolated from either TGS1-proficient cells (parental HeLa cells or TGS1 M1R cells) or TGS1 mutants cells (TGS1 M1, M2). These analyses revealed that the levels of hTR in TGS1 mutant cells increase by 2-fold compared to control or rescued cells (Figure 2A). Similar results were obtained by analyzing hTR levels using quantitative RT-PCR (qRT-PCR). hTR levels were increased on average by 1.8-fold in TGS1 mutant clones (M1, M2, and M3) expressing FLAG-GFP compared with their rescued counterparts stably expressing FLAG-TGS1 (Figure 2B).
Figure 2. Mutations in TGS1 Increase the hTR Levels
(A) Northern blots (NBs) of total RNA from TGS1-proficient (CTR), TGS1 mutant (M1, M2), and TGS1 M1-rescued (M1R) probed against hTR or U1 snRNA. Ethidium bromide (EtBr)-stained rDNA is a loading CTR. Bottom panel: quantification of the hTR levels normalized to U1snRNA. The hTR doublet represents different folding states of mature hTR (Mitchell et al., 1999). Error bars represent standard deviation derived from three independent NBs.
(B) hTR levels determined by qRT-PCR on total RNA prepared from TGS1 mutant (M1, M2, M3) cells expressing either FLAG-GFP or FLAG-TGS1. Bars represent means from 3 independent experiments, are relative to parental HeLa cells (set to 1) and are normalized to GAPDH (∗∗p < 0.01; ∗∗∗p < 0.001; one-way ANOVA).
© WB showing the TGS1 abundance in UMUC3 and BJ-HELT cells treated with non-targeting (CTR) or TGS1 siRNA for 6 days. Tubulin is a loading CTR.
(D) Quantification of hTR and TGS1 transcripts by qRT-PCR performed on total RNA prepared from UMUC3 or BJ-HELT cells at the indicated days following siRNA treatment; data are relative to cells treated with CTR siRNA (set to 1); hTR levels are normalized to the GAPDH transcript. (∗∗p < 0.01; ∗∗∗p < 0.001, ns, not significant; one-way ANOVA, left panel; Student’s t test, right panel.)
We also determined whether loss of TGS1 upregulates hTR in other human cell lines and whether the degree of hTR upregulation depends on the basal hTR level in the cell line before TGS1 inactivation. We performed RNAi-mediated TGS1 knockdown in the UMUC3 bladder cancer line and in BJ-HELT cells (BJ fibroblasts stably expressing TERT and SV40 large T antigen) (Hahn et al., 1999, Xu and Blackburn, 2007). TGS1 RNAi efficiently reduced the levels of both TGS1 mRNA and TGS1 protein, resulting in increased hTR levels in each cellular context (Figures 2C and 2D). Together, these results indicate that loss of TGS1 increases hTR levels in human cell lines.
TGS1 Catalyzes Formation of the TMG Cap in hTR Molecules Associated with Telomerase
We next asked whether TGS1 catalyzes formation of the hTR TMG cap in the hTR molecules that associate with telomerase and whether loss of this cap affects hTR incorporation into functional telomerase complexes. To address these questions, we used a two-step immunopurification procedure (Figure 3A). We first precipitated endogenous telomerase complexes from nuclear extracts of either parental cells or TGS1 mutant clones (M1 and M2) using an anti-TERT antibody (Venteicher et al., 2009). We then purified hTR from hTR-TERT complexes using an antibody that specifically recognizes the trimethylated guanosine cap (TMG) (Bringmann et al., 1983). Northern blotting analysis of TERT-IP eluates confirmed that hTR is more abundant in TGS1 mutant cells than in TGS1-proficient control cells (Figure 3B), indicating that loss of TGS1 increases the abundance of both hTR and hTR-TERT complexes in HeLa cell nuclei. qRT-PCR of RNA precipitated by the anti-TMG antibody further showed that hTR molecules are more abundant in control cell precipitates than in precipitates from TGS1 mutant cells (Figure 3C), suggesting that in TGS1-deficient cells most hTR molecules do not contain a TMG cap. These results clearly show that TGS1 catalyzes formation of the hTR TMG cap and that hTR molecules devoid of this cap can efficiently associate with TERT.
.
Edited by Engadin, 05 February 2020 - 03:50 PM.
