Role for the splicing factor TCERG1 in Cajal body integrity and snRNP assembly

Nuclear precursor mRNA (pre-mRNA) splicing is a fundamental RNA processing mechanism that occurs as a common means of achieving proteomic cell diversity. Global transcriptome analyses estimate that more than 90% of human multiexon genes undergo alternative splicing (Pan et al., 2008; Wang et al., 2008). Splicing is catalyzed by a dynamic ribonucleoprotein machinery known as the spliceosome, which is composed of RNA–protein complexes called small nuclear ribonucleoproteins (snRNPs) and a group of ∼200 additional proteins (Matera and Wang, 2014; Wahl et al., 2009). The snRNPs are composed of five uridine (U)-rich small nuclear RNAs (snRNAs), U1, U2, U4, U5 and U6, as well as a heptameric ring of Sm proteins and a specific set of proteins. Spliceosomal snRNP formation undergoes an intricate assembly and maturation pathway that involves nuclear and cytoplasmic events. After transcription and initial 3′ end processing, snRNAs are exported to the cytoplasm, where the heptameric ring of Sm proteins is assembled followed by hypermethylation of the 7-methylguanosine (m7G) snRNP cap. Then, the resulting immature snRNPs return to the nucleus, where additional snRNP-specific proteins are added to assemble functional di-snRNP (U4 and U6) and tri-snRNP complexes (U4, U6 and U5) in addition to snRNP recycling, as well as posttranscriptional modifications (Fischer et al., 2011; Matera and Wang, 2014; Staněk, 2016). Nuclear processing of snRNPs is performed in the Cajal bodies (CBs). After the final maturation process, fully assembled snRNPs are released from CBs to the nuclear speckles, where they accumulate prior to functioning in the splicing process.

CBs were first discovered in vertebrate neurons by Spanish neurobiologist Santiago Ramón y Cajal in 1903 (Cajal, 1903). Despite having been described more than a century ago, the major molecular components and functions of CBs related to snRNP assembly and recycling have only begun to be deciphered over the past 20 years (Gall, 2000). It has been demonstrated that a complex termed the little elongation complex (LEC) facilitates efficient synthesis of snRNA by RNAPII (Hu et al., 2013; Smith et al., 2011). In mammalian cells, LEC is composed of ELL, ICE1, and ICE2 proteins, which colocalize with coilin in CBs (Hu et al., 2013; Polak et al., 2003; Smith et al., 2011). Recently, Hutten and collaborators found that the SUMO isopeptidase USPL1, which is an essential component of CBs, interacts and colocalizes with components of the LEC. Importantly, knockdown of USPL1 resulted in disassembly of CBs, reduced RNAPII-mediated snRNP transcription and altered pre-mRNA splicing (Hutten et al., 2014). More recently, the human 7SK snRNP, composed of the 7SK snRNA Larp7 and MePCE, has also been found to specifically associate with the LEC to promote RNAPII-mediated snRNA and small nucleolar RNA (snoRNA) synthesis (Egloff et al., 2017). These data, together with results showing that the human mediator subunit MED26 also localizes to CBs and plays a crucial role in the transcription of snRNA genes through recruitment of the LEC (Takahashi et al., 2015), support the view that CBs have a role in snRNP and snoRNP biogenesis. Further evidence has been found for the role of CBs in snRNA processing. The RNAPII-associated integrator complex (INT) is required for the integrity of CBs and for cleaving the extended 3′-end of snRNAs, which is essential for the biogenesis of spliceosomal snRNAs (Albrecht et al., 2018; Takata et al., 2012). The target of EGR1 protein 1 (TOE1) is a nuclear deadenylase that localizes to CBs and participates in the 3′-end processing of small Cajal body-specific RNAs (scaRNAs), snoRNAs, snRNAs and telomerase RNA component (TERC) (Lardelli et al., 2017; Son et al., 2018). In addition to being involved in snRNP biogenesis, CBs also contain factors involved in other cellular processes, such as transcription and RNA stability, ribosome biogenesis, histone mRNA processing and telomere maintenance (Machyna et al., 2013). Many of these factors are not CB specific and are also present in other types of nuclear bodies, such as gems, the histone locus body or the nucleolus. A comprehensive list of relevant protein components of CBs has recently been published (Sawyer et al., 2019).

Coilin is a phosphoprotein that is commonly used to define CBs and is considered an essential structural component of these bodies because loss of coilin results in CB disintegration (Collier et al., 2006; Liu et al., 2009; Strzelecka et al., 2010b; Tucker et al., 2001). Although coilin appears not to be fully essential for viability in Arabidopsis, Drosophila or mice, coilin gene disruption provokes embryonic lethality in zebrafish (Collier et al., 2006; Liu et al., 2009; Tucker et al., 2001; Walker et al., 2009). Coilin depletion in zebrafish leads to defects in snRNP biogenesis that can be rescued upon snRNP injection (Strzelecka et al., 2010b), which strongly suggests that CBs are essential for snRNP assembly. These data, together with other extensive studies that show accumulation of stalled intermediates upon depletion of snRNP-specific proteins or blockage of di- and tri-snRNP formation (Bizarro et al., 2015; Novotný et al., 2015; Schaffert et al., 2004) and the de novo CB formation triggered by inhibition of tri-snRNP (Novotný et al., 2015), suggest a role for CBs in quality control of snRNP assembly and maturation. This putative role for CBs is supported by computational modeling, which predicts that the presence of these organelles increases the local concentration of snRNP components, thereby enhancing snRNP assembly by up to 11-fold (Klingauf et al., 2006). This predicted assembly rate was later experimentally verified by measuring the in vivo kinetics of tri-snRNP assembly using fluorescent recovery after photobleaching (FRAP) (Novotný et al., 2011). The assembly rate may be important when cells are metabolically active (Strzelecka et al., 2010a) and may explain the embryogenesis and/or fertility defects observed in coilin-knockout mice (Tucker et al., 2001; Walker et al., 2009).

TCERG1, previously named CA150 (Suñé et al., 1997), is a nuclear protein involved in splicing regulation and transcriptional elongation. TCERG1 binds to splicing components (Goldstrohm et al., 2001; Lin et al., 2004; Sánchez-Álvarez et al., 2006; Smith et al., 2004), co-purifies with spliceosomal subcomplexes (Deckert et al., 2006; Makarov et al., 2002; Neubauer et al., 1998; Wahl et al., 2009), and affects alternative pre-mRNA splicing of minigene reporters (Cheng et al., 2007; Lin et al., 2004; Montes et al., 2012a,b; Pearson et al., 2008; Sánchez-Hernández et al., 2012) and in putative cellular targets identified by whole-transcriptome analysis upon TCERG1 knockdown (Muñoz-Cobo et al., 2017; Pearson et al., 2008). Consistent with these findings, TCERG1 is enriched at the interface between speckles (the compartments enriched in splicing factors) and nearby transcription sites (Sánchez-Álvarez et al., 2006; Sánchez-Hernández et al., 2016). In addition to this location, TCERG1 is distributed throughout the nucleoplasm, with increased levels in many other granule-like sites besides speckles, suggesting additional functions for TCERG1. Therefore, we undertook a detailed study on the effects of TCERG1 knockdown on nuclear architecture. In this study, we show that, upon knockdown of TCERG1, there are changes in the integrity of CBs, and this phenotype is associated with altered snRNP assembly, thereby suggesting a new role for TCERG1 in CB formation and snRNP biogenesis.

To investigate whether TCERG1 plays a role in the formation or maintenance of nuclear bodies, TCERG1 was depleted in HEK293T and HAP1 cells using CRISPR-Cas9 gene editing technology, and the effect on CBs, speckles, gems, and histone locus body was analyzed by immunofluorescence microscopy. We generated two different HEK293T clones (1AC4 and 2AC2) and one HAP1 clone (c012). In control cells treated with a scramble guide RNA with no homology to any gene (scramble control), coilin displayed the characteristic CB localization (Fig. 1A). Strikingly, no CBs were found in TCERG1-depleted cells. With an extreme overexposure, we found coilin dispersed throughout the nucleoplasm, appearing as tiny dots in the nucleoplasm (Fig. 1A, overexposed panels), which is reminiscent of a phenotype resulting from the depletion of CB components (Girard et al., 2006; Lemm et al., 2006; Li et al., 2014). Quantification analysis of CBs showed a reduction in the number and size of these organelles in the TCERG1-knockdown HEK293T cells (Fig. 1B). Similar results were obtained with the TCERG1-depleted HAP1 (Fig. 1A,B) or SH-SY5Y (Fig. S1) cells. Interestingly, knockdown of TCERG1 did not change the total protein (Fig. 1C) or mRNA (Fig. 1D) levels of coilin. The use of a proximal start site of transcription in HAP1 cells and the difficulty in targeting each allele in the HEK293T polyploidy cell line may account for the residual TCERG1 expression in these cells (Fig. 1C).

Next, we assessed whether other components of CBs, which are shared with other known types of nuclear bodies, are affected upon TCERG1 knockdown. NOLC1 is a nucleolar protein that shuttles between the nucleolus and the cytoplasm, interacts with coilin and accumulates at the CBs, suggesting that NOLC1 functions as a molecular link between the nucleolus and the CBs (Isaac et al., 1998). Immunofluorescence studies showed that NOLC1 localized mainly in the nucleolus and at several foci inside the nucleoplasm, where it colocalized with coilin in the control cells (Fig. 2, scramble control). In the knockdown HEK293T cells, NOLC1 localized in the nucleolus, but similar to coilin, NOLC1 did not accumulate at foci in the nucleoplasm (Fig. 2; Fig. S2). In overexposed images, we did not detect NOLC1 localized in residual coilin foci (Fig. 2, overexposed), which further suggests the loss of CB integrity upon TCERG1 knockdown.

TCERG1-depleted cells were also analyzed for changes in other nuclear structures, such as speckles (SRSF2), gems (SMN) and the histone locus body (LSM11) (Fig. 3). No effects on gems or the histone locus body were observed upon TCERG1 depletion (Fig. 3B,C), demonstrating that TCERG1 is an important factor for CBs but is not essential for other nuclear structures. Depletion of TCERG1, however, affected the morphology of nuclear speckles, which appeared smaller and less diffusely distributed (Fig. 3A). Consistent with these findings, TCERG1 was previously found at the interface between speckles and nearby transcription and processing sites (Sánchez-Álvarez et al., 2006). We also performed quantitative analysis of the number of histone locus bodies and nuclear gems in control and knockdown cells, and we did not observe significant differences among the samples (Fig. S3).

Using the two knockdown HEK293T cell lines, we assessed whether transient TCERG1 overexpression could result in restoration of the phenotype. In fact, overexpression of a limited amount of TCERG1 resulted in significant levels of coilin staining in each of the two cell lines (Fig. 4A). Quantification analysis of CBs showed an increase in the number and size of these organelles in the TCERG1-knockdown HEK293T cells upon TCERG1 re-expression (Fig. 4B). These results further support a role for TCERG1 in the formation of CBs.

To examine the presence of TCERG1 in CBs, we stained HEK293T cells with antibodies against TCERG1 and the CB marker coilin. We observed that TCERG1 is absent at the nucleolar compartment and distributed throughout the nucleoplasm with an increased signal in organized granule-like sites corresponding to the speckles (Fig. 5A, panel a), as previously shown (Sánchez-Álvarez et al., 2006). Dual labeling of cells with antibodies directed against TCERG1 and coilin showed that TCERG1 partially overlapped but did not coincide with the coilin-containing bodies (Fig. 5A, panels c,d), suggesting that TCERG1 is enriched at the CB periphery. The spatial relationship between TCERG1 relative to coilin was identified by quantitatively scanning specific nuclear regions containing CBs (Fig. 5A, panel e). To assess whether the level of colocalization of TCERG1 and coilin changes during the cell cycle, we obtained synchronous populations of cells arrested in G1 and S1 phases by using hydroxyurea (Fig. 5B,C, panel f). We observed an increased level of colocalization during G1 phase (Fig. 5B,C, compare panel d and e) when, presumably, more actively transcribing snRNA genes are recruited to the CB periphery.

Interestingly, immunoprecipitation of endogenous TCERG1 revealed that it associates with coilin (Fig. 5D). Given that the washes of the immunoprecipitation were performed with RIPA buffer (see Materials and Methods), the association of TCERG1 with coilin is fairly strong. TCERG1 is also associated with CRM1, SMN and integrator complex subunit 11 (INTS11) (Fig. 5E), which are factors involved in CB homeostasis and snRNP biogenesis. These results suggest that TCERG1 is present at CBs and is associated with these bodies through its interaction with coilin and/or other CB components.

The complex process of snRNP biogenesis requires the assembly of seven Sm proteins around the Sm-binding site of snRNAs to form a stable ring-like structure of the core snRNPs in the cytoplasm, which is followed by nuclear import and snRNP maturation in CBs (Matera and Wang, 2014). We have previously shown that TCERG1 interacts with Sm proteins in vitro and in vivo (Sánchez-Álvarez et al., 2006). To investigate whether TCERG1 plays important roles in snRNP formation, we performed immunoprecipitation experiments with anti-Sm (Y12) antibodies using cell lysates from TCERG1-knockdown and scramble control cells, and quantified the amount of associated RNA by real-time quantitative reverse transcription PCR (RT-qPCR). We observed a reduction in the amount of precipitated U1, U2, U4, U5 and U6 snRNAs in the lysates with knockdown of TCERG1 compared with those from control lysates (Fig. 6A). Similar results were obtained with the TCERG1-knockdown HAP1 cells (Fig. 6B). To corroborate these results, we performed a similar experiment after TCERG1 overexpression. When TCERG1 was overexpressed in HEK293T cells, an increased amount of all the precipitated snRNAs was observed (Fig. 6C). To rule out the possibility of a decreased amount of U-rich snRNAs in the depleted samples, we quantified the amount of U1, U2, U4, U5, and U6 RNAs in HEK293T and HAP1 cells by RT-qPCR. We did not observe decreased levels of snRNPs in the cells upon TCERG1 knockdown or increased levels upon TCERG1 overexpression (Fig. 6D–F). In contrast, TCERG1 depletion caused a reproducible, although modest, increase in snRNA levels (Fig. 6D,E). Interestingly, no detectable change of the endogenous Sm proteins was observed under conditions of TCERG1 gene knockdown or overexpression (Fig. 6G), and a similar amount of Sm protein was precipitated with the Y12 antibody from knockdown and scramble control cells (Fig. 6H). Overall, these results indicate that TCERG1 is involved in the recruitment of Sm proteins to spliceosomal snRNAs.

A defect in snRNP assembly upon TCERG1 knockdown would provoke a partially assembled U1 and U2 snRNP accumulation in TCERG1-depleted cells. To test this possibility, we used antibodies against the 5′-end of the snRNA trimethyl cap structure (TMG) to perform co-immunoprecipitations. We observed an increase in TMG-capped U1 snRNA, while the levels of TMG-capped U2 snRNA were similar to control cells upon TCERG1 depletion (Fig. 6I). The results obtained with TMG-capped U1 snRNA are consistent with our previous data suggesting a defect in snRNP assembly in the absence of TCERG1.

To test whether TCERG1 binds directly to pre-snRNAs, we performed immunoprecipitation experiments with TCERG1-specific antibodies, and the resulting pellets were subsequently analyzed in RT-qPCR reactions with primers binding to the 3′-end extension of pre-U1 and pre-U2 snRNA. We did not detect an interaction between TCERG1 and pre-U1 and pre-U2 RNA (data not shown). TCERG1 interacts with mature U2 snRNP (data not shown), which is in agreement with the reported association of TCERG1 with U2AF65 (Sánchez-Álvarez et al., 2006) and SF1 (Goldstrohm et al., 2001) and the presence of TCERG1 in the first spliceosome complex (Neubauer et al., 1998; Zhou et al., 2002). We also investigated the role of TCERG1 in the expression of pre-U1 and pre-U2 snRNAs. We performed RT-qPCR experiments to measure the levels of pre-U1 and pre-U2 snRNA precursors upon TCERG1 knockdown. We observed similar levels of pre-U1 and pre-U2 snRNA precursors in control and TCERG1-depleted cells (Fig. S4). Taken together, these results did not support a direct transcriptional effect of TCERG1 in the expression of these snRNAs.

To further investigate the role of TCERG1 in snRNP production, we analyzed whether nascent snRNP production was affected upon knockdown of TCERG1. We took advantage of a U2OS cell line stably expressing GFP–SmB under a doxycycline (dox)-inducible promoter that was previously generated to monitor nascent snRNP biogenesis from the pool of previously assembled snRNPs (Hutten et al., 2014). As expected (Hutten et al., 2014), immunoblotting of total cell lysates after 20 or 48 h of dox induction revealed high expression of GFP–SmB with no change in TCERG1 or CDK9 (control) expression (Fig. 7A), and GFP–SmB localized correctly in CBs and nuclear speckles (Fig. 7B). Next, we used specific antibodies against the TMG cap to precipitate snRNP and GFP–SmB complexes in control and TCERG1-knockdown cell extracts after 18 h of dox induction, followed by SDS-PAGE and immunoblotting for GFP–SmB detection. Strikingly, we observed a reduced amount of GFP–SmB in the immunoprecipitations from TCERG1-depleted cells (Fig. 7C). These results suggest that the production of nascent snRNA particles is inhibited upon TCERG1 depletion.

Our data indicate that TCERG1 plays an important role in the assembly of Sm–snRNA complexes during the biogenesis of snRNPs and the formation of CBs. By performing TCERG1 knockdown in combination with immunoprecipitation, we detected a striking effect of reduced TCERG1 expression on the association of Sm proteins with snRNAs to form the minimal core snRNP (Fig. 6A,B). A lower Sm core assembly was also observed in vivo using a nascent GFP–SmB fusion protein (Fig. 7C). Given that the endogenous levels of U snRNAs were not decreased in these conditions (Fig. 6D), the effect of TCERG1 on Sm–snRNP formation is unlinked to snRNA transcription. The increased snRNA levels observed upon TCERG1 knockdown are surprising, given that snRNP assembly is inhibited, and the increase could reflect cellular responses to perturbations in the formation of mature snRNP complexes. These snRNAs might be stabilized by other proteins, or their degradation pathways might be affected. More work is clearly necessary to elucidate the reasons for the snRNA increase upon depletion of TCERG1. During the stepwise snRNP biogenesis pathway, snRNAs move from the nucleus to the cytoplasm. In the cytoplasm, the Sm assembly of the snRNPs occurs followed by their import to different nuclear compartments, such as speckles and CBs, to participate in the process of pre-mRNA splicing. Although TCERG1 has been identified as a nuclear protein, it may shuttle between the nucleus and cytoplasm, as suggested in a recent report showing the presence of TCERG1 in the cytoplasm of neuronal cells (Muñoz-Cobo et al., 2017). The observed interaction of TCERG1 with Sm proteins (Sánchez-Álvarez et al., 2006) and coilin (Fig. 5), published data implicating TCERG1 in the functional regulation of the splicing process (Cheng et al., 2007; Lin et al., 2004; Montes et al., 2012a,b; Muñoz-Cobo et al., 2017; Pearson et al., 2008; Sánchez-Hernández et al., 2012), and the implication of TCERG1 in snRNP biogenesis described in this study has led us to hypothesize that TCERG1 takes part in Sm–snRNP formation in the cytoplasm and then facilitates the splicing process in the nucleus. Our results reveal the exciting possibility that the effects of TCERG1 on splicing could be due, at least in part, to modulating Sm–snRNP formation.

Our data also indicate that TCERG1 participates in the formation or integrity of CBs. CBs are conserved nuclear bodies involved in the final biogenesis of snRNPs, which are essential factors for pre-mRNA splicing. Despite being discovered in vertebrates more than 100 years ago, studies on the molecular requirements for the formation and stability of CBs began two decades ago and these details have not been thoroughly elucidated to date. Interestingly, knockdown of TCERG1 disrupts CBs without affecting other nuclear structures (Figs 1–3). The effect of TCERG1 depletion mimics the effect of the loss of coilin, which produces defects in CB formation in many organisms (Collier et al., 2006; Liu et al., 2009; Strzelecka et al., 2010b; Tucker et al., 2001), and the loss of many other proteins involved in snRNP maturation, such as SMN (Girard et al., 2006; Lemm et al., 2006), TGS1 (Lemm et al., 2006), PHAX (Lemm et al., 2006), INTS4 (Takata et al., 2012), WRAP53 (Mahmoudi et al., 2010) and USPL1 (Hutten et al., 2014), that also disrupt CBs.

What is the mechanism by which TCERG1 affects CB integrity? TCERG1 is present at the periphery of CBs (Fig. 5A), suggesting that TCERG1 is neither a constitutive component of CBs nor assists coilin as a scaffold protein in the formation of CBs. One possibility is that TCERG1 affects CB integrity by controlling the expression of an integral component of this organelle. However, we did not find any relevant known protein component of CBs (Machyna et al., 2013) in which transcription or alternative splicing is affected by TCERG1 depletion using data from available transcriptome studies (Muñoz-Cobo et al., 2017; Pearson et al., 2008). Although TCERG1 interacts with SMN (Cheng et al., 2007) (Fig. 5E) and SMN affects Sm core assembly displaying defects in CB formation (Raimer et al., 2017 and references therein), our results show that SMN localization to gems remained unaffected upon TCERG1 knockdown (Fig. 3), thereby making SMN unlikely to be the factor involved in destabilizing these nuclear bodies through altering TCERG1 expression levels. The consequences of the interaction of TCERG1 with other CB factors (Fig. 5E) in the formation of these nuclear structures have not been determined. The formation of CBs also depends on the presence of snRNPs (Lemm et al., 2006; Sleeman et al., 2001). Furthermore, immature snRNPs induce the formation of CBs in cells lacking these nuclear structures, while mature particles fail to generate CBs (Novotný et al., 2015; Roithová et al., 2018). A self-assembly model via coilin, in which different RNA species nucleate residual or sub-CBs that fuse to assemble a mature CB, has been recently proposed (Machyna et al., 2013). TCERG1 may disrupt CBs by interfering with the formation of the snRNP-dependent sub-CB by inhibiting core Sm–snRNP formation. The recent observation that core snRNPs containing the Sm proteins but not naked snRNAs restore the formation of CBs after their depletion and that the Sm ring is sufficient to target snRNAs to CBs (Roithová et al., 2018) is in keeping with this model. Future experiments will determine how the roles of TCERG1 in CB and snRNP formation described in this study are directly linked to the functional roles of this protein in transcription and alternative splicing.

The mammalian expression vectors pEFBOST7-TCERG1 and pEFBOS/GFP/T7-TCERG1[1-1098] have been described previously (Sánchez-Álvarez et al., 2006; Sánchez-Hernández et al., 2016; Suñé and Garcia-Blanco, 1999).

For immunoblotting, the following primary antibodies were used at the indicated dilution: anti-coilin (sc-55594; Santa Cruz Biotechnology, Dallas, TX) at 1:1000; anti-CDK9 (sc-484; Santa Cruz Biotechnology) at 1:2000; anti-Sm (Y12) (MS-450; Thermo Fisher Scientific, Waltham, MA) at 1:1000; anti-SMN (sc-32313; Santa Cruz Biotechnology) at 1:500; anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (sc-365062; Santa Cruz Biotechnology) at 1:2000; and anti-CRM1 and anti-INTS11 at 1:1000 (generously provided by Edouard Bertrand, IGMM-CNRS, Montpellier, France). The TCERG1 antibodies used in the western blot analysis in this study were generated in guinea pigs using two truncated glutathione S transferase (GST) fusion TCERG1 proteins spanning amino acids 234 to 662 and 631 to 1098, and were used at a dilution of 1:10,000. For immunofluorescence studies, we used the following antibodies: anti-TCERG1 (Suñé et al., 1997) at 1:2000; anti-coilin at 1:500; anti-T7 (A190-116A; Bethyl Laboratories, Montgomery, TX) at 1:1000; anti-NOLC1 (ab106324; Abcam, Cambridge, UK) at 1:500; anti-SRSF2 (S4045; Sigma, St Louis, MO) at 1:4000; anti-Sm at 1:50, anti-SMN at 1:200 and anti-LSM11 (generously provided by Joseph G. Gall, Carnegie Institution for Science, Baltimore, MD) at 1:200. For western blot analysis, primary antibodies were detected using HRP-conjugated secondary antibodies to rabbit-IgG and mouse-IgG (PerkinElmer Life Sciences; Waltham, Massachusetts) at a dilution of 1:5000. For immunofluorescence, we used Alexa Fluor 488-conjugated goat anti-rabbit-IgG, Alexa Fluor 550-conjugated goat anti-rabbit-IgG, and Alexa Fluor 647-conjugated goat anti-mouse-IgG antibodies from Molecular Probes (Eugene, Oregon) at a dilution of 1:500.

HEK293T cells were originally obtained from the American Type Culture Collection (ATCC, Manassas, VA) and were grown and maintained as described previously (Sánchez-Álvarez et al., 2010). Transfection assays were carried out using protocols described previously (Sánchez-Álvarez et al., 2006, 2010). Cells were transfected using calcium phosphate or Lipofectamine 2000 reagent (Invitrogen, Carlsbad, CA) according to the protocols of the manufacturer. Clonal U2OS GFP–SmB-expressing cells were generously provided by Angus I. Lamond (University of Dundee, UK) and were grown and maintained in Dulbecco's modified Eagle medium (Invitrogen) supplemented with 10% fetal bovine serum, penicillin–streptomycin 100 U and 100 μg/ml, respectively, 150 mg/ml hygromycin B, and 15 mg/ml blasticidin-HCl (Hutten et al., 2014). Expression of GFP–SmB was induced with 10 ng/ml doxycycline (Sigma) for 18 h. The source of SH-SY5Y cells was reported elsewhere (Muñoz-Cobo et al., 2017) and were grown and maintained as described previously (Muñoz-Cobo et al., 2017). The scramble, 1AC4 and 2AC2 cell lines were obtained using a TCERG1 human gene knockdown kit in HEK293T cells according to the manufacturer's instructions (Origene, Rockville, MD). Cells were cultured in Dulbecco's modified Eagle medium (Invitrogen) supplemented with 10% fetal bovine serum, penicillin–streptomycin (100 U and 100 μg/ml), and 1.5 μg/ml puromycin. TCERG1-knockdown HAP1 cells (Horizon Discovery, Waterbeach, Cambridge, UK) were maintained in Iscove's modified Dulbecco's medium (IMDM) (Gibco, Thermo Fisher Scientific) supplemented with 10% FBS (Gibco) and penicillin-streptomycin 100 U and 100 μg/ml, respectively. All cell lines were tested regularly for contamination.

For the experiments with cells synchronized for the cell cycle, cells were seeded on sterile glass coverslips and synchronized at G1 phase by incubation with 0.5 mM hydroxyurea (Thermo Fisher Scientific) for 18 h. For synchronization at S phase, cells were then incubated in normal DMEM for 3 additional hours before being fixed. Synchronized cells were also analyzed using a FACSAria III cell sorter flow cytometer (Becton–Dickinson). Data were evaluated using FlowJo cell analysis software.

For RNA interference (siRNA) knockdown experiments, HEK293T and U2OS cells were seeded to ∼50% confluence in 35 mm plates. The cells were transfected using the Lipofectamine 2000 reagent (Invitrogen) for HEK293T and Lipofectamine RNAiMaxx (Invitrogen) for U2OS cells according to the manufacturer's protocol with 60 nM (final concentration) of either of the following small interfering RNA (siRNA) duplexes: siEGFP, 5′-CUACAACAGCCACAACGCU-3′; siControl, 5′-CAGUCGCGUUUGCGACUGG-3′; siTCERG1, 5′-GGAGUUGCACAAGAUAGUU-3′.

Immunofluorescence staining in HEK293T and HAP1 cells was performed as previously described (Sánchez-Hernández et al., 2017). SH-SY5Y cells were grown on coverslips and fixed with 4% paraformaldehyde and 4% sucrose in PBS (pH 7.4) for 20 min at 37°C and permeabilized with 90% chilled methanol for 5 min. The cells were blocked in PBS containing 10% BSA and 0.5% Triton X-100 for 1 h at room temperature. The cells were incubated with primary antibody in PBS containing 10% BSA for 2 h at room temperature (humidity chamber). The cells were subsequently washed three times with 0.1% BSA in PBS. This step was followed by labeling with secondary antibodies for 2 h at room temperature (humidity chamber). The cells were subsequently washed three times with 0.1% BSA in PBS and three times with PBS. Coverslips were mounted onto glass slides using the ProLong Gold Antifade reagent with DAPI (Life Technologies, Carlsbad, California). Images were acquired using a Leica SP5 spectral laser confocal microscope and processed using the LAS AF software v2.3.6. Quantification was performed using ImageJ software. The images were digitally processed for presentation using the Adobe Photoshop CS3 extended v10.0 software.

Cells were lysed in T7 buffer [20 mM HEPES pH 7.9, 150 mM NaCl, 5 mM EDTA, 1% NP-40, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride (PMSF), and protease inhibitor mixture (Complete Roche)], RIPA buffer (50 mM Tris-HCl pH 7.5, 1% NP-40, 0.05% SDS, 1 mM EDTA, 150 mM NaCl, 0.5% sodium deoxycholate, 1 mM dithiothreitol, 1 mM PMSF, and protease inhibitor mixture) for 30 min at 4°C. The cell extracts were centrifuged at maximum speed (16,100 g) for 5 min. The proteins were separated by SDS-PAGE, transferred to a nitrocellulose membrane (Amersham Biosciences; Little Chalfont, UK) and then incubated with specific antibodies. After washing, the membrane was incubated with peroxidase-conjugated secondary antibodies, and the bound antibodies were detected by enhanced chemiluminescence (PerkinElmer Life Science).

Cells were harvested into 500 ml of RIPA buffer for 30 min at 4°C, and the cell extracts were centrifuged at 16,100 g for 5 min at 4°C. 10% fractions of whole cell extract (WCE) were boiled in SDS-PAGE loading buffer and saved for immunoblotting analysis. The supernatants were incubated with 3 μl of control mouse, rabbit IgG or Y12 (0.6 μg), anti-TMG (0.75 μg; MABE302, Millipore) or anti-TCERG1 antibodies (3 μl) overnight at 4°C and then incubated with Protein A/G agarose beads (Millipore, Burlington, MA) with end-over-end rotation for 2 h at 4°C. The anti-TMG antibodies (K121) have high affinity for trimethylguanosine and cross-react weakly with 7-methylguanosine at very high concentration (Krainer, 1988). Because the anti-TMG antibodies preferentially bind to the trimethylguanosine, they have been routinely used to detect specifically the trimethyl cap versus the m7G cap. After six washes with RIPA buffer, the proteins bound to the antibody resin were eluted by boiling the samples with SDS-PAGE loading buffer (0.2% Bromophenol Blue, 100 mM DTT, SDS 4%, 20% glycerol and Tris-HCl 0.1 M, pH 6.8), separated on 10% SDS-PAGE gels, and analyzed by western blotting. RNA immunoprecipitation was performed as previously described (Niranjanakumari et al., 2002) with minor modifications. Briefly, cells were fixed with 1% formaldehyde for 10 min at room temperature with slow mixing; cross-linking was arrested by adding glycine at 0.125 M for 5 min at room temperature. Cells were lysed in RIPA buffer and sonicated 3 times for 20 s on ice. Supernatants were precleared with protein A/G-agarose fast-flow slurry (Millipore) and tRNA at 100 µg/µl with end-over-end rotation for 1 h at 4°C. Beads were removed by centrifugation, and the supernatant was incubated with nonspecific mouse IgG and specific antibodies overnight. The RNA–antibody complexes were collected with protein A/G–agarose beads for 2 h at 4°C. Subsequently, the beads were collected and washed six times with RIPA buffer and resuspended in a buffer containing 100 µl of 50 mM Tris-HCl pH 7.0, 5 mM EDTA, 10 mM dithiothreitol and 1% SDS with a final incubation at 70°C for 45 min.

Total RNA was extracted from transfected cells with peqGOLD TriFast (peQlab, Fareham, UK) according to the manufacturer's protocol. Approximately 1 μg of RNA was digested with 10 units of RNase-free DNase (Ambion, Foster City, CA), and RNA was reverse transcribed using the qScript cDNA Supermix (Quanta Biosciences, Gaithersburg, MD) following the manufacturer's protocol. Quantification of endogenous transcripts by real-time PCR was performed using the iQ SYBR Green Supermix (Bio-Rad, Hercules, CA) and the iCycler thermal cycler station (Bio-Rad) with the following oligonucleotides: Coilin-Fw, 5′-CTTGAGAGAACCTGGGAAATTTG-3′ and Coilin-Rv, 5′-GTCTTGGGTCAATCAACTCTTTCC-3′; U1-Fw, 5′-GATACCATGATCACGAAGGTGGTT-3′ and U1-Rv, 5′-CACAAATTATGCAGTCGAGTTTCC-3′; U2-Fw, 5′-ATCGCTTCTCGGCCTTTTGG-3′ and U2-Rv, 5′-GGTGCACCGTTCCTGGAGG-3′; U4-Fw, 5′-GCGCGATTATTGCTAATTGAAA-3′ and U4-Rv, 5′-AAAAATTGCCAATGCCGACTA-3′; U5-Fw, 5′-GGTTTCTCTTCAGATCGCATAAATC-3′ and U5-Rv, 5′-CTCAAAAAATTGGGTTAAGACTCAGA-3′; U6-Fw, 5′-GCTTCGGCAGCACATATACTAAAAT-3′ and U6-Rv, 5′-ACGAATTTGCGTGTCATCCTT-3′; preU1-Fw, 5′-ACTGCGTTCGCGCTTTCCC-3′; preU1-Rv, 5′-GCAGGCGACATGTTACTTCC-3′; preU2-Fw, 5′-AACATAGGTACACGTGTGCCACGG-3′; preU2-Rv, 5′-ACAAATAGCCAACGCATGCGGGGC-3′. GAPDH was used as an internal control gene and was amplified with the oligonucleotides GAPDH-Fw (5′-ATGGGGAAGGTGAAGGTCG-3′) and GAPDH-Rv (5′-GGG TCATTGATGGCAACAATATC-3′). Statistical analysis was performed using Microsoft Excel 2010 and GraphPad Prism 5.0 software.

All experiments were repeated at least three times, and statistical analysis was performed using Prism 5.0 software (GraphPad). Two-tailed Student's tests (unpaired t-tests) were used to compare the samples and their respective controls. The P-values are represented by asterisks (*P≤0.05, **P≤0.01 and ***P≤0.001). The absence of an asterisk indicates that the change relative to the control is not statistically significant.

We thank Joseph Gall for the anti-LSM11 antibody, Edouard Bertrand for the anti-CRM1 and anti-INTS11 antibodies, and Angus I. Lamond for the U2OS GFP-SmB cells. We are grateful to David Staněk for experimental conversations and advice. The technical assistance of Laura Montosa during the confocal microscopy studies is deeply appreciated.