Ribosomal Protein S6 Kinase Activity Controls the Ribosome Biogenesis Transcriptional Program

Home , EEF2K, EIF4B, Ribosomal protein, Ribosomal protein S6, RRP15, Ubiquitin C

ORIGINAL ARTICLE Ribosomal protein S6 kinase activity controls the ribosome biogenesis transcriptional program

C Chauvin1,2, V Koka1,2, A Nouschi1,2, V Mieulet1,2, C Hoareau-Aveilla3, A Dreazen4, N Cagnard2, W Carpentier5, T Kiss3, O Meyuhas4 and M Pende1,2

S6 kinases (S6Ks) are mechanistic target of rapamycin substrates that participate in cell growth control. S6Ks phosphorylate ribosomal protein S6 (rpS6) and additional proteins involved in the translational machinery, although the functional roles of these modiﬁcations remain elusive. Here we analyze the S6K-dependent transcriptional and translational regulation of gene expression by comparing whole-genome microarray of total and polysomal mouse liver RNA after feeding. We show that tissue lacking S6Ks 1 and 2 (S6K1 and S6K2), displays a defect in the ribosome biogenesis (RiBi) transcriptional program after feeding. Over 75% of RiBi factors are controlled by S6K, including Nop56, Nop14, Gar1, Rrp9, Rrp15, Rrp12 and Pwp2 nucleolar proteins. Importantly, the reduced activity of RiBi transcriptional promoters in S6K1;S6K2 À / À cells is also observed in rpS6 knock-in mutants that cannot be phosphorylated. As ribosomal protein synthesis is not affected by these mutations, our data reveal a distinct and speciﬁc aspect of RiBi under the control of rpS6 kinase activity, that is, the RiBi transcriptional program.

Oncogene (2014) 33, 474–483; doi:10.1038/onc.2012.606; published online 14 January 2013 Keywords: signal transduction; ribosome biogenesis; mTOR

INTRODUCTION that fully inhibit mTOR activity, impair mRNA translation to a much 16–18 Biochemical and genetic studies have unequivocally identified S6 greater degree than rapamycin. kinase 1 and 2 (S6K1 and S6K2) as the main in vivo kinases Here we address whether S6K activity controls expression of responsible for ribosomal protein S6 (rpS6) phosphorylation.1–4 specific mRNA classes at the transcriptional and translational levels The kinases sequentially phosphorylate rpS6 on five serine by combining microarray analysis on total and polysomal residues in the C-terminal tail of the protein.5,6 S6K1 and S6K2 fractions. Although we fail to uncover specific mRNAs under the are activated by the mammalian target Of rapamycin kinase translational control by S6Ks, our data reveal that S6K has a broad (mTOR, now officially named mechanistic target Of rapamycin),7 a influence on the transcription of the ribosome biogenesis (RiBi) master controller of growth integrating a large variety of program, providing the cell with nucleolar factors required for environmental cues.8 S6Ks participate in the mTOR-dependent ribosomal RNA (rRNA) synthesis, cleavage, post-transcriptional growth program, as outlined by the growth defect and perinatal modifications, assembly with ribosomal proteins (RPs) and lethality of loss-of-function mutant flies and mice.4,9,10 As rpS6 is a transport. As we demonstrate that the phosphorylation of rpS6 constitutive component of the 40S small ribosomal subunit and as participates in this regulatory event, our findings suggest the rpS6 phosphorylation is triggered by anabolic signals, the favored interesting possibility that the post-translation modification of a hypothesis is that S6K activity regulates growth by affecting RP affects RiBi through the synthesis of RiBi factors. protein synthesis, although direct experimental evidence is lacking. That S6Ks interact with the translational machinery is further RESULTS supported by the findings that additional substrates include Transcriptional versus translational control of gene expression by eukaryotic initiation factor 4B (eIF4B) and eukaryotic elongation S6Ks factor 2 kinase (eEF2K).11,12 Moreover, S6Ks are found in complex To uncover potential translational mRNAs targets of S6Ks, with the eukaryotic initiation factor 3.13 However, global protein ribonucleoprotein complexes were extracted from livers of wild- synthesis is not affected in S6K-deficient cells, as assessed by type and S6K1;S6K2 À / À mice after starvation or 4-h refeeding. The methionine incorporation and polysome profile analysis.14 liver was chosen for this in vivo analysis because the mTOR Consistently, the selective allosteric mTOR inhibitor rapamycin, pathways is readily regulated by nutrient status and the which abrogates S6K activity while having minor effects on the preservation of the ribonucleoprotein complexes is optimal. phosphorylation of other mTOR targets including the eukaryotic Starvation and refeeding affected the phosphorylation of the initiation factor 4E-binding proteins, has minimal effects on cap- mTORC1 substrates S6K1 and 4E-binding protein 1, and the dependent translation.15 In addition, recent data using ribosome downstream S6K substrates rpS6 and eIF4B, as expected profiling and methionine incorporation, demonstrate that (Figure 1a).4,11 After sucrose gradient fractionation, actively Adenosine triphosphate (ATP)-competitive inhibitors of mTOR translating polyribosomes (polysomes) were separated from

1INSERM, U845, Paris, France; 2Universite´ Paris Descartes, Sorbonne Paris Cite´, Faculte´ de Me´decine, UMRS-845, Paris, France; 3Laboratoire de Biologie Molećulaire Eucaryote, Universite´ de Toulouse-UPS and Centre National de La Recherche Scientifique, Toulouse, France; 4Department of Biochemistry and Molecular Biology, IMRIC, Hebrew University- Hadassah Medical School, Jerusalem, Israel and 5Plateforme Post-Geńomique Pitie´-Salpe´trie`re, Groupe Hospitalier Pitie´-Salpe´trie`re, Universite´ Pierre et Marie Curie, Paris, France. Correspondence: Dr M Pende, U845, Inserm, 156 rue de Vaugirard, Paris, F-75015, France. E-mail: [email protected] Received 23 July 2012; revised 21 September 2012; accepted 4 November 2012; published online 14 January 2013 S6 kinases control RiBi program C Chauvin et al 475 monosomes (80S), large and small ribosomal subunits (60S and These factors may associate with the box C/D class of small 40S), free mRNA and proteins (Figure 1b). After measuring the nucleolar ribonucleoproteins (Nop56 and Rrp9), or with the box relative abundance of peaks corresponding to 40S-60S-80S and H/ACA class (Gar1). As the majority of these mRNAs was also found polysomes, there was an increase in the area under the curve of downregulated in the polysomal fractions (Figure 1c and polysomes after refeeding as compared with starved controls. Supplementary Figure 2), it is likely that S6K deletion also However, we could not detect a significant difference between decreases RiBi expression at the protein level. Immunoblot wild-type and S6K1;S6K2 À / À liver profiles, both in starved and analysis with commercially available antibodies to Nop56 con- refed conditions. These data were consistent with previous firmed a 50% reduction of protein amount, whereas fibrillarin observations in other cell types and experimental conditions,14 levels were not affected (Figure 1e and Supplementary Figure 3). suggesting that S6Ks do not affect global translation. To address Thus, S6Ks regulate the abundance of factors that are implicated whether specific mRNA classes were regulated at the translational in a variety of nucleolar functions. level, whole-genome microarray analysis was performed on To gain further insights on the physiological regulation of the mRNAs associated with polysomes and compared with the RiBi program by S6Ks, mice underwent different nutritional, results from the whole-cell extract. A total of 456 mRNAs were pharmacological and genetic manipulations that affected S6K downregulated in the whole-cell extract from S6K1;S6K2 À / À livers activity (Figure 1a). We focused this analysis on Nop56, Gar1 and after refeeding (Supplementary Table 1). To specifically assess Pwp2 mRNA expression that displayed a high degree of regulation control of gene expression at the translational level, the ratio of by S6Ks. As shown in Figure 2a, 4-h refeeding after overnight signal intensities from the microarray analysis on the whole-cell starvation led to 2.5- to 3-fold increase of gene expression in wild- extracts and polysomal fractions was calculated, as previously type mice. A kinetic study from 2- to 8-h refeeding indicated a performed in distinct experimental settings.19 Strikingly, only similar response throughout the entire time period (Figure 2b). five mRNAs displayed a decreased polysomal/whole-cell ratio Strikingly, S6K deletion abrogated the nutrition-induced gene in S6K1;S6K2 À / À extracts as compared with wild type expression, as Nop56, Gar1 and Pwp2 mRNA levels in S6K1; (Supplementary Table 1). Two of these mRNAs were not S6K2 À / À livers after refeeding were equivalent to the starved annotated. The remaining encoded for myeloid/lymphoid or mice of both S6K1;S6K2 À / À and wild-type genotypes (Figure 2a). mixed-lineage leukemia 5, polycomb group ring finger 5 and A transient inhibition of mTORC1/S6K activities by the rapamycin retinal outer segment membrane protein 1 mRNAs. The signal derivative Temsirolimus in wild-type mice mimicked the effect of intensities for these mRNAs on the microarray chip were close to S6K deletion on RiBi gene expression. Similarly, in primary cultures the low threshold of detection. In addition, when real-time of hepatocytes, serum stimulation increased Nop56 and Pwp2 quantitative PCR (RT–QPCR) analysis was attempted on a larger gene expression in wild-type cells in a rapamycin-sensitive number of mice, we could not confirm a statistical significant manner, whereas basal and serum-stimulated RiBi expression difference in polysomal/whole-cell ratio. To verify whether the was blunted in S6K1;S6K2 À / À cells (Figure 2c). In addition, the polysomal/whole-cell ratio was a reliable read-out of translational defect of Gar1 expression was confirmed in serum-stimulated regulation, the microarray analysis of refed wild-type mice was mutant hepatocytes (Supplementary Figure 4). Taken together, compared with starved animals of the same genotype these data define the essential role of S6Ks in mediating the (Supplementary Table 2). Importantly, the polysomal/whole-cell effects of diet and mTORC1 activity on RiBi gene expression in vivo ratio for RP mRNAs was downregulated after starvation and in a cell autonomous manner. (Supplementary Figure 1). These mRNAs contain a 50 terminal As S6K1 and S6K2 homologs have overlapping yet distinct oligopyrimidine motif that mediates translational repression in functions in mammals,4,30 we addressed their respective cells deprived of growth factors or nutrients in an mTOR- contribution on the RiBi program. As shown in Figure 3a, the dependent and S6K-independent manner.4,20 These findings single deletion of S6K1 or S6K2 was not sufficient to downregulate validated our approach to study translational control of gene RiBi gene expression. Next, we rescued the hepatic expression of expression in livers in vivo. Under these experimental conditions, p70-p85 S6K1 isoforms, or of the S6K2 homolog by adenoviral we failed to identify a positive role for S6Ks in translation of transduction in S6K1;S6K2 À / À mice. Seven days after adenovirus specific mRNA classes. administration, all these kinases were efficiently overexpressed in S6K1;S6K2 À / À livers and re-established the phosphorylation of rpS6 (Figure 3b). Importantly, each S6K isoform alone was able to S6Ks control the RiBi transcriptional program cause a sharp induction of RiBi gene expression (Figure 3c). As When we considered the downregulated mRNAs in whole-cell other tissues were poorly transduced by adenoviral S6Ks, these extracts of S6K1;S6K2 À / À versus wild-type refed livers data accord with the cell autonomous regulation of the RiBi (Supplementary Table 1), it became evident that 23% encoded program, as shown in isolated hepatocytes (Figure 2c). In proteins involved in RNA metabolism. The RiBi transcriptional conclusion, our findings using loss-of-function and gain-of- program was particularly affected by S6K deletion. This program is function approaches indicate that p70/p85 S6K1 and S6K2 are required for the coordinated expression of nucleolar proteins that functionally redundant in the RiBi control. regulate pre-rRNA synthesis, cleavage, post-transcriptional modifications, ribosome assembly and export.21,22 When a heat map of 150 RiBi genes was analyzed, 78% of RiBi mRNAs were rpS6 phosphorylation is involved in the RiBi transcriptional downregulated in S6K-deficient livers (Figure 1c and program Supplementary Figure 2). A gene set enrichment analysis was S6K1 and S6K2 can compensate each other in catalyzing the performed on the set of RiBi genes and showed a significant phosphorylation of rpS6, eIF4B, eEF2K, PCDC4 phosphorylation, enrichment in S6K1;S6K2 À / À versus wild-type livers (Po0.0001). whereas other targets are more selectively controlled by S6K1 and Next, we aimed at confirming the microarray results by RT–QPCR include SKAR, IRS1 and Rictor.31–33 As rpS6 phosphorylation on a larger number of samples. All the tested RiBi genes that were correlated with RiBi expression (Figure 3), the functional role of significantly downregulated in the microarray analysis, were also this post-translational modification was evaluated. This issue is significantly decreased by 25–55% by RT–QPCR, namely Nop56, particularly noteworthy, as rpS6 is an integral component of Nop14, Gar1, Rrp9, Rrp15, Rrp12, Pwp2 and Ddx18 (Figure 1d). The ribosomes. We took advantage of a knock-in mouse, rpS6P À / À in downregulated factors or their yeast orthologs have been which all five serine residues that are phosphorylated by S6K, were implicated in rRNA processing, 60S assembly, 40S assembly, rRNA mutated to alanines.34 Consistently, rpS6 phosphorylation was not pseudouridylation, rRNA methylation and rDNA transcription.23–29 detected in rpS6P À / À livers after refeeding (Figure 4a).

& 2014 Macmillan Publishers Limited Oncogene (2014) 474 – 483 S6 kinases control RiBi program C Chauvin et al 476

WT S6K1;S6K2-/- Starved Refed refed 60S starved refed + rapa starved refed 60S

P-S6K (Thr389) 40S 60S 80S 60S 80S 40S 80S 40S 80S S6K1 40S Polysomes Polysomes Polysomes Polysomes P-S6 (Ser240-244)

P-S6 (Ser235-236) -/- -/- S6 total WT S6K1;S6K2 WT S6K1;S6K2

P-eIF4B (Ser422) Starved Refed 2 4EBP1 1.5 Tubulin

0.5

Polysome to 40S-60S-80S ratio Polysome to 40S-60S-80S 0 -/- -/- WT WT S6K1,2 S6K1,2

Total Polysome

1.4

1.2

a 0.8 a a a a a 0.6 a a

0.4

Relative mRNA expression level 0.2

0 1.00 0.67 0.33 0.00 -0.33 -0.67 -1.00 WT S6K1;S6K2-/-

NOP56 1.2

-/- WT S6K1;S6K2 0.8 a NOP56 0.6 Tubulin 0.4

Expression ratio Expression to tubulin 0.2

0 -/- WT S6K1;S6K2

Oncogene (2014) 474 – 483 & 2014 Macmillan Publishers Limited S6 kinases control RiBi program C Chauvin et al 477 Importantly, S6K activation by mTORC1, as assessed by Thr389- The activity of a RiBi promoter depends on rpS6 kinase activity phosphorylation, and the phosphorylation of other S6K substrates To dissect the level of regulation of RiBi expression by S6K activity, P À / À including eIF4B was not affected in rpS6 livers. In addition, mice were hydrodynamically injected with a linearized plasmid the phosphorylation of the mTORC1 substrate 4EBP was not encoding a luciferase reporter under the control of the RiBi Nop56 decreased, but actually slightly increased, as assessed by the promoter spanning 2-kb upstream of the transcription start abundance of the slower migrating g band corresponding to the site. A sixfold stimulation of Nop56 promoter activity was hyperphosphorylated protein (40% of g band over the total 4EBP observed in wild-type liver tissue after refeeding of mice P À / À amount in rpS6 livers, versus 32% in wild type). Thus, (Figure 6a). However, this nutrient-stimulated promoter activity P À / À mTORC1/S6K are fully functional in rpS6 livers, whereas rpS6 was blunted in S6K1;S6K2 À / À livers. Similarly, serum stimulation phosphorylation is completely absent. The proportion of poly- of primary hepatocytes increased reporter gene transcription P À / À some-associated ribosome was slightly increased in rpS6 driven by the Nop56 promoter in wild-type cells in a rapamycin- livers after refeeding, as compared with wild type (Figure 4b). sensitive manner (Figure 6b). Importantly, Nop56 promoter P À / À These data are consistent with previous studies with rpS6 activity was low in cells from both S6K1;S6K2 À / À and rpS6P À / À 34 mouse embryonic fibroblasts, suggesting a direct negative role livers (Figures 6b and c). Our findings define a transcriptional of rpS6 phosphorylation on protein synthesis and/or the presence response under the control of rpS6 kinase activity that drives the of feedback mechanisms in these mutants (for example, increased RiBi program. 4EBP phosphorylation). Thus, both S6K À / À and rpS6P À / À mutants do not support a positive role of S6K activity in translation initiation. Importantly, livers of refed rpS6P À / À mice displayed a 25–35% reduction in Nop56, Nop14, Gar1, Rrp9, Rrp15, Rrp12, DISCUSSION Pwp2 and Ddx18 mRNA levels, as compared with refed wild-type RiBi is an extremely complex and energy-demanding process, as in mice (Figure 4c). The percentage of inhibition of RiBi genes was actively growing cells 470% of the total gene transcription is quantitatively more pronounced in S6K1;S6K2 À / À than in rpS6P À / dedicated to ribosome production. Eukaryotic RiBi involves À livers (Figures 1d and 4c), suggesting that additional S6K targets transcription, covalent modification and nucleolytic processing could be involved in this regulation. In conclusion, the phosphor- of pre-rRNAs, packaging of the maturing rRNAs with nearly 80 RPs ylation of rpS6, among other substrates, participates in the control as well as continuous nucleo-cytoplasmic trafficking of RPs and of the RiBi program by S6Ks. nascent ribosomal subunits. Therefore, eukaryotic cells developed sophisticated mechanisms: (i) to coordinate RNA Pol I-, Pol II- and Pol III-dependent transcription required for rRNA and RP mRNA Defects of RiBi in S6K mutants production; (ii) to control rRNA processing and ribosome assembly To address the functional consequence of changes in RiBi in the nucleolus, and rp mRNA translation in the cytosol. The RiBi expression, rRNA processing was evaluated by northern blot program is essential to this regulation, as it produces hundreds of analysis with a probe recognizing the internal transcribed spacer specialized protein and ribonucleoprotein factors that assist rDNA sequence (Figure 5a). As shown in Figure 5b, the 45S rRNA transcription, pre-rRNA processing and ribosome assembly. Here, precursor, and 32S and 12S intermediates of rRNA processing we define a pathway that is triggered by nutrient availability and were detected in starved and refed livers of wild-type and S6K- that, through the activity of S6K, has a broad influence on deficient genotypes. However, there was a significant defect of transcription of RiBi factors. 32S and 12S levels in S6K-deficient livers after refeeding. Next, we À / À The defect in RiBi expression and RiBi in S6K mutant tissues evaluated the number of ribosomes per cell in S6K1;S6K2 as does not correlate with an impairment of translation initiation, as compared with wild-type livers. As 85% of total cellular RNA is assessed by the analysis of polysome profiles and by methionine rRNA, changes in the overall amount of RNA mainly reflect incorporation (Figures 1 and 4; Mieulet et al.14 and Ruvinsky changes in the amount of rRNA and thus in ribosome content. et al.34). This is consistent with the analysis of mouse mutants for Importantly, the measurement of the RNA to DNA ratio showed a À / À RiBi factors, such as the pseudouridine synthase dyskerin. Loss-of- 40% decrease in S6K1;S6K2 as compared with wild-type refed function mutations in mouse dyskerin inhibits rRNA processing livers (Figure 5c). The RNA to DNA ratio was similarly decreased in À / À and pseudouridylation without affecting global translation serum-stimulated S6K1;S6K2 MEF cells as compared with wild initiation.36,37 S6K1;S6K2 À / À and rpS6P À / À cells may have type (data not shown). Our data suggest a defect in pre-rRNA actually upregulated mechanisms that sustain protein synthesis. synthesis and/or maturation because of S6K inhibition, although at The slight increase in 4E-binding protein 1 phosphorylation that this stage the two processes cannot be easily taken apart. These we observe in the mutant cells may be one of the compensatory findings are consistent with published literature implying that the mechanisms promoting translation initiation (Figure 4; Espeillac multiple RiBi factors under the control of S6K activity may be et al.38). Such mechanisms may explain the increased polysome involved in a variety of different steps, such as the regulation of P À / À 21,35 amount in rpS6 cells, and partly compensate the defect in rDNA transcription, 40S or 60S maturation. the protein anabolic responses because of the alteration of the

Figure 1. S6K deletion affects RiBi transcriptional program but not global translation initiation. (a) Immunoblot analysis of protein extracts from liver of 2-month-old mice of the indicated genotypes. Mice were starved overnight or 4-h refed after overnight starvation. When indicated (rapa), mice were injected intraperitoneally with 5 mg/kg of the rapamycin derivate temsirolimus 1 h before refeeding. Proteins were revealed using the indicated antibodies. Each well corresponds to a different mouse. (b) Liver polysomal proﬁle of the indicated genotypes in 2-month-old mice starved overnight or 4-h refed after overnight starvation. The area under the curve was calculated and indicated in blue for 40S, 60S and 80S peaks and in yellow for polysomal peaks. The ratio of polysome/40S-60S-80S is shown. Data are mean±s.e.m. (n ¼ 5). aPo0.05 versus wild-type (WT) mice. (c) Heat map of the expression levels of 150 genes involved in RiBi at the transcriptional (total) or polysomal (polysome) level in 4-h refed livers of S6K1;S6K2 À / À versus wild-type 2-month-old mice. Each column corresponds to a different mouse. The brightness of green and red represents the degree to which expression is respectively lower or higher in the S6K1;S6K2 À / À versus wild-type livers. The RiBi gene list was established using the Gene Ontology database. An enlarged heat map is shown in Supplementary Figure 3 with the name of the RiBi genes. (d) RT–QPCR analysis of relative transcript levels for RiBi genes normalized to H2AZ gene in 4-h refed livers of 2-month-old mice of the indicated genotypes. Data are mean±s.e.m. (nX5). aPo0.05 versus WT mice. (e) Immunoblot analysis of protein extracts from 4-h refed livers of 2-month-old mice of the indicated genotypes. Proteins were revealed using the indicated antibodies. Each well corresponds to a different mouse. The ratio of the densitometric assay over tubulin is shown. Data are mean±s.e.m. aPo0.05 versus WT mice.

& 2014 Macmillan Publishers Limited Oncogene (2014) 474 – 483 S6 kinases control RiBi program C Chauvin et al 478 RiBi program. At this stage, we do not exclude that one difference between wild-type and mutant cells during the pioneer physiological role of S6Ks may be the regulation of translation round of translation may not be appreciated using our initiation and elongation, as suggested by their ability to experimental approach. phosphorylate eIF4B and eEF2K. However, using the whole- Whole-genome screens in yeast uncovered several nucleolar genome microarray analysis of the polysomal fractions we were RiBi factors involved in the regulation of cell size during the G1/S not able to uncover this role. It is possible that the observed phase transition of the cell cycle.39,40 Similarly, studies in compensatory mechanisms in S6K mutants may mask the Drosophila cells also pointed to an enrichment for RiBi factors in defective eIF4B and eEF2K regulation. Alternatively, the knockdown experiments causing a decrease in cell size.41 RiBi technique is not adequate to detect subtle changes in mutants have more prominent effects on cell size, as compared translation initiation or elongation. For instance, a quantitative with rp mutants, suggesting that the nucleolar RiBi, rather than the cytoplasmic protein synthesis, is setting the cell size limits through an unknown mechanism. In yeast, loss-of-function NOP56 GAR1 PWP2 mutations of the protein kinase Sch9 lead to small cell size and 4 impairment of RiBi because of a defect in rDNA transcription, rp 35,39 3.5 a expression and RiBi synthesis. The closest mammalian 39 a orthologs of Sch9 are Akt and S6Ks. As Sch9 is a downstream 3 a target of TORC1, it has been proposed that S6Ks are the functional 2.5 orthologs of Sch9.42 Our studies suggest that during the evolution b b 2 to metazoans, S6Ks have maintained the Sch9 role in the b b regulation of the RiBi program, whereas loosing the control of 1.5 b b rp expression that in metazoans is mainly regulated at the 1 translational level. As S6K mutants in metazoans also have a 9,10 0.5 specific defect in cell size, it is tempting to speculate that S6K Relative mRNA expression level controls cell size via the regulation of the RiBi program. However, 0 in mice cell atrophy correlates with the deletion of S6K1,30 refed refed refed whereas the reduction of rpS6 phosphorylation and RiBi -/- -/- -/- starved starved starved -/- -/- -/- WT refed WT refed WT refed expression are caused by the combined deficiency of both S6K1 WT starved WT starved WT starved and S6K2. Therefore, it is possible that the defect of RiBi on S6K WT refed + rapa WT refed + rapa WT refed + rapa deletion contributes to the deregulation of cell growth, although S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 additional events linked to the single deletion of S6K1 are dominant. NOP56 GAR1 PWP2 mTOR is an integral component of the signal transduction 4 a mechanisms matching RiBi with nutrient availability and a 20,35,43–46 3.5 a growth factor signals. When our manuscript was in a preparation, mTOR has been proposed to regulate rp mRNA 3 a translation through the relief of 4E-binding protein-mediated a a a 2.5 repression.17,18 Our studies demonstrate a distinct and specific 2 branch of mTOR signaling regulating RiBi at the transcriptional

1.5 level through the S6K-mediated control of the RiBi program that affects rRNA synthesis and maturation. It is likely that S6Ks 1 have additional mechanisms to affect rRNA production, including 0.5

Relative mRNA expression level a direct effect on rDNA transcription factors as previously 43,47 0 suggested. Upstream-binding factor is a putative mTOR/S6K target involved in Pol-I activity,43,47 although its identiﬁcation as a direct S6K substrate requires a more thorough analysis using tools that do not titrate mTOR activity. In yeast, Sch9 WT starved WT starved WT starved

WT refed -2h WT refed -4h WT refed -8h WT refed -2h WT refed -4h WT refed -8h WT refed -2h WT refed -4h WT refed -8h directly phosphorylates the transcription repressors Dot6 and Tod6, thus affecting the RPD3L histone deacetylase recruitment at NOP56 PWP2 1.6 a 1.4 a Figure 2. RiBi gene regulation by S6Ks. (a) RT–QPCR analysis of 1.2 relative transcript levels for RiBi genes normalized to H2AZ gene in b liver of 2-month-old mice of the indicated genotypes. Mice were 1 b b b starved overnight or 4-h refed after overnight starvation. When 0.8 c indicated (rapa), mice were injected intraperitoneally with 5 mg/kg of 0.6 the rapamycin derivate temsirolimus 1 h before the refeeding. Data are mean±s.e.m. (nX3). aPo0.05 versus starved wild-type (WT) 0.4 mice; bPo0.05 versus refed WT mice. (b) RT–QPCR analysis of relative 0.2 transcript levels for RiBi genes normalized to H2AZ gene in liver of Relative mRNA expression level 2-month-old mice of the indicated genotypes. Mice were starved 0 overnight or 2-, 4- or 8-h refed after overnight starvation. Data are mean±s.e.m. (nX3). aPo0.05 versus starved WT mice. (c) RT–QPCR serum serum

starved starved analysis of relative transcript levels for RiBi genes normalized to -/- -/- -/- -/-

WT serum WT serum H2AZ gene in cultured hepatocytes of the indicated genotype. Cells WT starved WT starved

serum + rapa serum + rapa were starved overnight and serum stimulated for 4 h with or without -/- -/-

WT serum + rapa WT serum + rapa a 15-min pre-treatment with rapamycin. Data are mean of three a S6K1;S6K2 S6K1;S6K2

S6K1;S6K2 S6K1;S6K2 independent experiments±s.e.m. Po0.05 versus starved WT cells; bPo0.05 versus serum-stimulated WT cells; cPo0.05 versus serum- À / À S6K1;S6K2 S6K1;S6K2 stimulated S6K1;S6K2 cells.

Oncogene (2014) 474 – 483 & 2014 Macmillan Publishers Limited S6 kinases control RiBi program C Chauvin et al 479 the RiBi promoters.35,39 Mammalian orthologs of Dot6 and own expression by binding the pre-mRNA and inhibiting Tod6 have not revealed yet. Surprisingly, we find that splicing.49 RpL11 interacts with and inhibits the trans- rpS6 mutants that cannot be phosphorylated by S6Ks also cription factor Myc that drives transcription of 5S rRNA and present a defect in RiBi expression, albeit to a lesser degree transfer RNA genes as well as of RP genes.50 In yeast, the binding than the S6K mutants. These data open the question how rpS6 of RPs to the gene promoters is enriched at the transfer RNA phosphorylation may affect RiBi transcription. Other RPs have genes.51 Future studies should define whether rpS6 directly binds been shown to participate in the control of gene expression the RiBi promoters and how rpS6 phosphorylation could modify at the level of transcription and splicing.48 In particular, the transcriptional networks. functions of the target genes may contribute to set the overall protein synthetic capacity. For instance, rpS13 downregulates its MATERIALS AND METHODS Materials NOP56 GAR1 PWP2 1.8 pNOP56-luc2CP plasmid containing firefly luciferase complementary 1.6 DNA under the control of NOP56 promoter was constructed. The 2031- bp Nop56 promoter fragment was generated by PCR using custom- 1.4 designed primers with additional restriction sites: sense primer, 50-ACGG 1.2 GGTACCACATGTCTACTCTCTGTTGG-30; antisense primer, 50-CCCAAGCTTG 1 CGGAGACCACACCTTCTCCG-30. The digested fragment was inserted in the 0.8 KpnI–HindIII sites of the expression vector pGL4.12[luc2CP] (Promega, Charbonnieres, France), which contains the firefly luciferase gene. 0.6 0.4 0.2 Animals

Relative mRNA expression level 0 Generation of S6K1;S6K2-deficient mice (C56BL/6-129/Ola), S6K1- and -/- -/- -/- -/- -/- -/- S6K2-deficient mice (C57BL/6) and rpS6P À / À knock-in mice has been WT WT WT previously described.4,34 Each genotype was compared with wild-type S6K1 S6K1 S6K1 S6K2 S6K2 S6K2 mice of the same genetic background. Mice were maintained at 22 1C with a 12-h dark–12-h light cycle and had free access to food. All animal studies WT S6K1;S6K2-/- were approved by the Direction De´partementale des Services Ve´te´rinaires, Myc- Myc- Pre´fecture de Police, Paris, France (authorization number 75–1313). All AdV - GFP p70 S6K1 S6K2 p85S6K1 studies were done in male animals. When indicated, mice were infected with 5.109 of infectious particles of green fluorescent protein, Myc-tagged Low exposure p70 S6K1, p85 S6K1 or Myc-tagged S6K2 adenovirus by retro-orbital intravenous injection. p85 was not tagged to avoid masking of a putative P-S6K (Thr389) nuclear localization sequence in the N terminal end. In the p85 construct, High exposure the second ATG corresponding to p70 was mutated. One week after infection, mice were refed for 4 h after overnight starvation and then p85 euthanized. When indicated, mice were hydrodynamically injected into the S6K1 * p70 tail vein in 5 s with 2 ml of saline solution containing 36 mg of pNop56- luc2CP plasmid and 4 mg of a plasmid encoding renilla luciferase under a Myc constitutive promoter (pRL-TK). Four days after injection, mice were refed for 4 h after overnight starvation and then euthanized. A piece of frozen P-S6 (Ser240-244) liver tissue was lysed in 200 ml of 1X passive lysis buffer (Promega). Luciferase activity was measured using the Dual-Luciferase Reporter Assay P-S6 (Ser235-236) System (Promega).

S6 Western blot

GAPDH A piece of frozen tissue was ground to powder under liquid N2 and lysed in extraction buffer (20 mM Tris-HCl (pH 8,0), 5% glycerol, 138 mM NaCl, 2,7 mM KCl, 1% Nonidet P-40, 20 mM NaF, 5 mM EDTA, 1 mM dithiothreitol) NOP56 GAR1 PWP2 3 b b 2.5 b b b b Figure 3. S6K1 and S6K2 are functionally redundant in the RiBi b 2 b b control. (a) RT–QPCR analysis of relative transcript levels for RiBi genes normalized to H2AZ gene in 4-h refed liver of 2-month-old 1.5 mice of the indicated genotypes. Data are mean±s.e.m. (nX3). (b) Immunoblot analysis of protein extracts from 4-h refed after 1 overnight starvation liver of 2-month-old mice of the indicated a a a genotypes. S6K1;S6K2 À / À mice were injected with either control 0.5 (green ﬂuorescent protein (GFP)), Myc-p70 S6K1, p85 S6K1 or Myc- S6K2 adenovirus 1 week before killing. Proteins were revealed using Relative mRNA expression level 0 the indicated antibodies. Star indicates a cross-reacting band. Each WT WT WT well corresponds to a different mouse. (c) RT–QPCR analysis of relative transcript levels for RiBi genes normalized to ubiquitin C

AdV GFP AdV GFP AdV GFP (Ubc) gene in 4-h refed liver of 2-month-old mice of the indicated AdV S6K2 AdV S6K2 AdV S6K2

-/- -/- -/- À / À -/- -/- -/- genotypes. S6K1;S6K2 mice were injected with either control

AdV p70 S6K1 AdV p85 S6K1 AdV p70 S6K1 AdV p85 S6K1 AdV p70 S6K1 AdV p85 S6K1 (GFP), Myc-p70 S6K1, p85 S6K1 or Myc-S6K2 adenovirus 1 week -/- -/- -/- -/- -/- -/- before killing. p85 S6K1 was not tagged to avoid masking of a putative nuclear localization signal. Data are mean±s.e.m. (n ¼ 2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2

S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 for wild-type (WT) mice and nX3 for adenovirus-injected S6K1;S6K2 À / À mice). aPo0.05 versus WT mice; bPo0.05 versus S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 GFP-injected S6K1;S6K2 À / À mice.

& 2014 Macmillan Publishers Limited Oncogene (2014) 474 – 483 S6 kinases control RiBi program C Chauvin et al 480 WT WT P-/- anti-tubulin (Sigma, Saint Quentin Fallavier, France) and anti-glyceralde- starved refed S6 refed hyde 3-phosphate dehydrogenase (Abcam, Paris, France). P-S6K (Thr389) Polysome fractionation S6K1 Sucrose density gradient centrifugation was used to separate the P-S6 (Ser240-244) subpolysomal from the polysomal ribosome fractions. A piece of frozen liver was ground to powder under liquid N2 and lysed in 50 mM Tris-HCl P-S6 (Ser235-236) (pH 7.8), 10 mM MgCl2, 240 mM KCl, 250 mM sucrose, 2% Triton X-100, 5 mM dithiothreitol, 100 mg/ml cycloheximide and 100 U/ml RNase inhibitor. To 1 S6 remove cell debris, homogenates were spun at 8000 Â g for 5 min at 4 C. An aliquot of the supernatant was removed to measure protein concen- P-eIF4B (Ser422) tration. Heparin was added to the supernatants at a final concentration of γ 1 mg/ml. The extracts were rapidly frozen and stored at À 80 1C. In all, 4EBP1 β α 1.2 mg of protein were layered on a 0.5–1.5 M linear sucrose gradient (20 mM Tris-HCl, pH 7.5, 80 mM NaCl, 5 mM MgCl2, 1 mM dithiothreitol) and Tubulin centrifuged in a SW41 rotor at 36 000 r.p.m. for 2 h at 4 1C. Following centrifugation, the gradient was displaced upward through a flow cell 60S 2 recording absorbance at 260 nm with the use of the density gradient 60S a 1.8 fractionation system (Isco, Lincoln, NE, USA) and fractionated in 12 80S 80S 1.6 fractions. To calculate the ratio between polysome and monosome-free 40S 1.4 subunits, the area underlying the corresponding peaks on the profile was 40S Polysomes 1.2 measured using the Image J software (rsbweb.nih.gov/ij/). Polysomes 1 0.8 RNA extraction 0.6 For northern blot analysis, RNA isolation after polysome fractionation was 0.4 3 -/- performed as described previously. Northern blots were hybridized WT S6P 0.2

Polysome to 40S-60S-80S ratio 0 0 with speciﬁc probes against ITS (5 -ACCCACCGCAGCGGGTGACGCGA 0 0 WT S6P-/- TTGATCG-3 ) and eEF1A (5 -GCCGGAATCTACGTGTCCGATTACGACGATGTT GATGTGAGTCTTTTCCTTTCCCAT-30). Quantiﬁcation was made using Storm ImageQuant software (GE Healthcare Life Science, Munich, Germany). For microarray analysis, a piece of frozen liver was 1.2 homogenized in 3 ml of the lysis solution (4 M guanidinium thiocyanate, 25 mM sodium citrate, 0.5% sarcosyl,100 mM 2-mercaptoethanol, pH 7) and 52 1 total RNA was extracted as described. For polysomal RNA isolation, polysomal fractions were pooled and RNA was extracted with guanidinium 52 a a a thiocyanate as described. RNA integrity of total and polysomal RNA was 0.8 a a a a a checked on an Agilent 2100 bioanalyser (Agilent Technologies, Massy, 0.6 France).

0.4 RNA and DNA quantitation in liver 0.2 A piece of frozen liver (o50 mg) was homogenized in 1 ml of TRIzol

Relative mRNA expression level reagent. RNA extraction was performed from aqueous phase and DNA was 0 precipitated from the phenol phase and interphase with 300 ml of ethanol. NOP56 NOP14 GAR1 RRP9 RRP15 RRP12 PWP2 Ddx18 After a series of washes with 0.1 M trisodium citrate in 10% ethanol to WT remove residual phenol, the DNA pellet was solubilized in a mild alkaline S6P-/- solution (300 ml NaOH 8 mM) and the pH was adjusted to 7 with 1 M Hepes. Figure 4. rpS6 phosphorylation participates in the control of the RiBi RNA and DNA content were determined using specific fluorescent dyes program. (a) Immunoblot analysis of protein extracts from liver of selective for RNA (Qubit RNA BR Assay Kit, Invitrogen, Life Technologies, 3-month-old mice of the indicated genotypes. Mice were starved Saint Aubain, France) or double-stranded DNA (Qubit dsDNA BR Assay Kit, overnight or 4-h refed after overnight starvation. Proteins were Invitrogen), respectively. Measurements were realized using Qubit 2.0 revealed using the indicated antibodies. Each well corresponds to a fluorometer (Invitrogen). different mouse. (b) Polysomal profile of 4-h refed liver in 3-month- old mice of the indicated genotypes. The area under the curve was Gene expression profiling and bioinformatic analysis calculated and indicated in blue for 40S, 60S and 80S peaks and in Total and polysomal RNA were used for microarrays. Complementary RNA yellow for polysomal peaks. The ratio of polysome/40S-60S-80S is was synthesized, amplified and purified using the Illumina TotalPrep RNA ± X a shown. Data are mean s.e.m. (n 3). Po0.05 versus wild-type Amplification Kit (Ambion, Foster City, CA, USA) following the manufac- (WT) mice. (c) RT–QPCR analysis of relative transcript levels for RiBi turer’s recommendations. Briefly, 200 ng of RNA was reverse transcribed. genes normalized to H2AZ gene in 4-h refed livers of 3-month old After second-strand synthesis, the complementary DNA was transcribed mice of the indicated genotypes. Data are mean±s.e.m. (nX3). a in vitro and complementary RNA labeled with biotin-16-UTP. Labeled Po0.05 versus WT mice. A full colour version of this figure is probe hybridization to Illumina BeadChips Mouse WG-6 v2 (Illumina, San available at the Oncogene journal online. Diego, CA, USA) was carried out using Illumina’s protocol. The Beadchips were scanned on the Illumina IScan using Illumina IScan image data acquisition software. Illumina GenomeStudio software (Illumina) was used complemented with complete protease inhibitors (Roche, Boulogne- for preliminary data analysis. The Illumina data were then normalized using Billancourt, France). To remove cell debris, homogenates were spun at the ‘normalize quantiles’ function in the GenomeStudio Software 8000 Â g for 10 min at 4 1C. Protein extract from liver was resolved by (Illumina). One S6K1;S6K2 À / À mice did not pass the signal quality control sodium dodecyl sulfate–polyacrylamide gel electrophoresis before transfer and was discarded from further analysis. We used a threshold at 0.01 to onto polyvinylidene difluoride membrane and incubation with anti-S6K1 convert ‘Detection pval’ into flags: flag ¼ 0 if pval 40.01 and flag ¼ 1if (Santa Cruz, Santa Cruz, CA, USA), anti-phospho S6K1 (Thr389), anti-4EBP1, pvalo ¼ 0.01. Furthermore, all the intensity values X15 000 were flagged anti-phospho rpS6 (Ser 235/236), anti-phospho rpS6 (Ser 240/244), 0. For changes at the transcriptional level, total RNAs from livers of two anti-phospho-eIF4B (Ser 422), anti-rpS6, anti-c-Myc (Cell Signalling, different genotypes or conditions were compared. For translational Danvers, MA, USA), anti-NOP56 (Novus Biologicals, Littleton, CO, USA), analysis, the ratio of polysomal RNA to total RNA signals was calculated

Oncogene (2014) 474 – 483 & 2014 Macmillan Publishers Limited S6 kinases control RiBi program C Chauvin et al 481 18S 5.8S 28S 12 Probe 47S 5’ETS ITS 3’ETS 45S

Cleavage at site 1 first Cleavage at site 2 first

41S 30S 32S 21S 32S 21S 17S 18SE 12S 28S 18SE 12S 28S 18S 5.8S 18S 5.8S

WT S6K1;S6K2–/– 45S 32S 12S 35 Starved RefedStarved Refed a 30 45S c 25 ITS probe 32S 20 a 12S 15 b eEF1A probe 10 a b 5

Expression ratio to eEF1A mRNA 0 EtBr refed refed refed starved starved starved –/– –/– –/– WT refed WT refed WT refed –/– –/– –/– WT starved WT starved WT starved S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 S6K1;S6K2 40 S6K1;S6K2 S6K1;S6K2 35 30 a 25 20 15 10

RNA to DNA ratio (µg/µg) 5 0 -/- WT S6K1;S6K2 Figure 5. rpS6 kinase activity has functional consequences on RiBi. (a) Scheme of pre-rRNA processing in mouse liver showing the target sequence of the internal transcribed spacer (ITS) probe. (b) Northern blot of RNA from liver of 2-month-old mice of the indicated genotypes. Mice were starved overnight or 4-h refed after overnight starvation. Each well corresponds to a different mouse. The membrane was probed successively with ITS and eEF1A probes as indicated. As a loading control, the ethidium bromide stained gel is shown. The data of the quantiﬁcation of the autoradiography are shown. Data are mean±s.e.m. (n ¼ 3). aPo0.05 versus starved WT mice; bPo0.05 versus refed WT mice; cPo0.05 versus starved S6K1;S6K2 À / À mice. (c) RNA to DNA ratio in 4-h refed livers of 2-month-old mice of the indicated genotypes. The assay was performed by measuring RNA and DNA contents and calculating the ratio of RNA to DNA. Data are mean±s.e.m. (nX7). aPo0.05 versus wild-type (WT) mice.

for each genotype or condition. Then changes in the polysomal/total Real-time quantitative PCR between livers of each genotype or condition were identified. The group Total and polysomal RNAs prepared from liver for microarray analysis were comparisons were done using Student’s t-test on probes flagged as ‘1’ for also used for RT–QPCR. In all, 1 mg of RNA first treated for 15 min at room at least half of the samples involved in the comparison. We filtered the temperature with RNase-free DNase (Invitrogen) according to the resulting P-values at 5% and fold 1.5. All the resulting genes further manufacturer’s instructions, was used to prepare complementary DNAs discussed were validated by RT–QPCR on independent samples. Heat map with random hexamer primers and SuperScript II (Invitrogen). Total RNA representation of log2 normalized expression data were generated using R from hepatocytes was isolated using an RNeasy Mini Kit (Qiagen, Hilden, (http://www.r-project.org/foundation/) and Treeview software (http://tax- Germany), according to the manufacturer’s instructions. Single-strand onomy.zoology.gla.ac.uk/rod/treeview.html) using the Spearman correla- complementary DNA was synthesized from 1 mg of RNA with random tion similarity measure and average linkage algorithm. Enrichment tests hexamer primers and SuperScript II (Invitrogen). RT–QPCR was performed were performed with Gene Set Enrichment Analysis software v2.07 (www. using a Taqman instrument (Applied Biosystem, Life Technologies, Saint broadinstitute.org/gsea/index.jsp) using a phenotype-based (1000) permu- Aubain, France) according to the manufacturer’s instructions using a SYBR tation test procedure. Microarray data have been deposited into the Array Green PCR Master Mix (Applied Biosystem). We determined the relative Express repository (http://www.ebi.ac.uk/arrayexpress/) under the acces- amounts of the mRNAs studied by means of the 2 À DDCT method, with sion no. E-MEXP-3518 for S6K1;S6K2 À / À versus wild-type liver analysis and H2AZ or ubiquitin C as reference genes pinin and wild-type samples as the E-MEXP-3522 for starved versus refed wild-type liver analysis. invariant control. The murine primer sequences used were as follows:

& 2014 Macmillan Publishers Limited Oncogene (2014) 474 – 483 S6 kinases control RiBi program C Chauvin et al 482 WT S6K1;S6K2-/- WT S6P-/- 0.9 2.5 1.8 a a a 0.8 1.6 2 0.7 1.4

0.6 1.2 b 1.5 0.5 b 1 b 0.4 b 0.8 1 0.3 b 0.6 0.2 0.5 0.4 (fold induction over WT starved) Firefly to renilla luciferase activity Firefly to renilla luciferase activity (fold induction over WT starved) 0.1 Firefly to renilla luciferase activity 0.2

0 0 0 WT WT S6K1;S6K2-/- starved refed refed serum serum serum serum starved starved starved starved serum + rapa serum + rapa serum + rapa serum + rapa Figure 6. rpS6 kinase activity regulates RiBi promoters. (a) Luciferase activity in mice of the indicated genotypes after hydrodynamic injection of plasmids encoding firefly luciferase under the 2031-bp Nop56 promoter in combination with a plasmid encoding renilla luciferase under a constitutive promoter (pRL-TK). Three days after injection, mice were starved overnight or 4-h refed after overnight starvation. Hepatic lysates were prepared and the ratio between firefly and renilla luciferase activity was measured. Data are mean±s.e.m. (nX7). aPo0.05 versus starved wild-type (WT) mice; bPo0.05 versus refed WT mice. (b, c) Luciferase activity in cultured hepatocytes of the indicated genotypes after transfection of plasmids encoding firefly luciferase under the 2031-bp Nop56 promoter in combination with a plasmid encoding renilla luciferase under a constitutive promoter (pRL-TK). Six hours after transfection, cells were starved overnight and serum stimulated for 4 h with or without a 15-min pre-treatment with rapamycin. Hepatic lysates were prepared and the ratio between firefly and renilla luciferase activity was measured. Data are expressed as fold increase over the starved WT cell luciferase activity. Data are mean of three independent experiments±s.e.m. aPo0.05 versus starved WT cells; bPo0.05 versus serum-stimulated WT cells.

NOP56, sense 50-AGGTGGAGGAATGTGTGCTT-30; rapamycin (Calbiochem). Luciferase activity was measured using the Dual- antisense 50-TCAGACACAGCATTGGCATT-30; Luciferase Reporter Assay System (Promega). NOP14, sense 50-CGGAGGCAGGTCCACAAAG-30; antisense 50-GCCAAGGATCTGGAACTTCTG-30; CONFLICT OF INTEREST GAR1, sense 50-CCTCCAGAACGTGTCGTCTT-30; The authors declare no conflict of interest. antisense 50-CCTTGTTCTCCTCGGTGGTA-30; Rrp9, sense 50-CGGCGAAAGGTAGATTCTGCT-30; antisense 50-TTCCTGTGCGGTCTCTTCCA-30; ACKNOWLEDGEMENTS 0 0 Rrp15, sense 5 -CAGCGTCGAAGCTAGAAGATG-3 ; We thank the Novartis Foundation and George Thomas laboratory for the use of S6K 0 0 antisense 5 -AGGATTCAACATTGTCCTCACTG-3 ; mutant mice. We are grateful to the members of INSERM-U845 for support, and to 0 0 Rrp12, sense 5 -TGTGAAGTTGCATAATGAGCTGC-3 ; Stefano Fumagalli and Olivier Jean-Jean for helpful discussions and sharing reagents. 0 0 antisense 5 -TCGAATCACCTCTGTAACAGCA-3 ; We thank Sophie Berissi, Ana Diaz and Sylvie Fabrega for excellent technical support. 0 0 PWP2, sense 5 -GGTTCTCGAATTTGTTGGGTACG-3 ; We thank Pfizer for the generous gift of Temsirolimus. This work was supported by 0 0 antisense 5 -ACAGTGACTCTATTTCCCACGG-3 ; grants to MP from the European Research Council, from Fondation de la Recherche 0 0 Ddx18, sense 5 -GAAGCGGAATGCCAAGCTG-3 ; Medicale, from Fondation Schlumberger pour l’Education et la Recherche and from 0 0 antisense 5 -TCCTGTTTCTTTAGGCACATCTC-3 ; ANR and by grants to OM from the US-Israel Binational Science Foundation 0 0 H2AZ, sense 5 -AAGCGTATCACCCCTCGTCA-3 ; (2009054), Israel Cancer Research Fund and Ministry of Health. CC received a 0 0 antisense 5 -AGCGATTTGTGGATGTGTGG-3 ; fellowship from the Coddim Ile de France. ubiquitin C, sense 50-GGACGTCGAGCCCAGTGTTA-30; antisense 50-CCCATCACACCCAAGAACAAG-30. REFERENCES 1 Kozma SC, Ferrari S, Bassand P, Siegmann M, Totty N, Thomas G. Cloning of the Hepatocyte culture and luciferase assay mitogen-activated S6 kinase from rat liver reveals an enzyme of the second Primary hepatocytes were isolated from mice by collagenase perfusion messenger subfamily. Proc Natl Acad Sci USA 1990; 87: 7365–7369. method as described previously.38 In all, 4 Â 105 cells were plated in six- 2 Grove JR, Banerjee P, Balasubramanyam A, Coffer PJ, Price DJ, Avruch J et al. well plates in William’s medium (Invitrogen) supplemented with 10% fetal Cloning and expression of two human p70 S6 kinase polypeptides differing only bovine serum (Invitrogen), insulin (4 mg/ml), bovine serum albumin (1 mg/ at their amino termini. Mol Cell Biol 1991; 11: 5541–5550. ml), 10 units/ml penicillin, 10 g/ml streptomycin and 25 nM dexamethasone 3 Shima H, Pende M, Chen Y, Fumagalli S, Thomas G, Kozma SC. Disruption of the (Sigma). For RNA analysis, 24 h after plating, the cells were serum-starved p70(s6k)/p85(s6k) gene reveals a small mouse phenotype and a new functional overnight in William’s medium and then in Earle’s Balanced Salt Solution S6 kinase. EMBO J 1998; 17: 6649–6659. (EBSS) (Invitrogen) for 2 h. Cells were then stimulated with 10% fetal 4 Pende M, Um SH, Mieulet V, Sticker M, Goss VL, Mestan J et al. S6K1( À / À )/ / bovine serum for 4 h with or without a 15-min pre-treatment with 20 nM S6K2( À À ) mice exhibit perinatal lethality and rapamycin-sensitive 50-terminal rapamycin (Calbiochem, SanDiego, CA, USA). For luciferase assay, oligopyrimidine mRNA translation and reveal a mitogen-activated protein kinase- transfection of luciferase plasmids was performed 20 h after plating in dependent S6 kinase pathway. Mol Cell Biol 2004; 24: 3112–3124. the presence of Lipofectamine LTX with PLUS Reagent (Invitrogen). In total, 5 Thomas G, Siegmann M, Kubler AM, Gordon J, Jimenez dA. Regulation of 40S 3.6 mg pNop56-luc2CP plasmid, 0.4 mg pRL-TK plasmid and 5 ml ribosomal protein S6 phosphorylation in Swiss mouse 3T3 cells. Cell 1980; 19: Lipofectamine per well were applied in a final volume of 2 ml Opti-MEM. 1015–1023. After 6 h, the cells were serum-starved overnight in William’s medium and 6 Krieg J, Hofsteenge J, Thomas G. Identification of the 40 S ribosomal protein S6 then in EBSS (Invitrogen) for 2 h. Cells were then stimulated with 10% fetal phosphorylation sites induced by cycloheximide. J Biol Chem 1988; 263: bovine serum for 4 h with or without a 15-min pre-treatment with 20 nM 11473–11477.

Oncogene (2014) 474 – 483 & 2014 Macmillan Publishers Limited S6 kinases control RiBi program C Chauvin et al 483 7 Isotani S, Hara K, Tokunaga C, Inoue H, Avruch J, Yonezawa K. Immunopurified 30 Ohanna M, Sobering AK, Lapointe T, Lorenzo L, Praud C, Petroulakis E et al. mammalian target of rapamycin phosphorylates and activates p70 S6 kinase Atrophy of S6K1( À / À ) skeletal muscle cells reveals distinct mTOR effectors for cell alpha in vitro. J Biol Chem 1999; 274: 34493–34498. cycle and size control. Nat Cell Biol 2005; 7: 286–294. 8 Laplante M, Sabatini DM. mTOR signaling. Cold Spring Harb Perspect Biol 2011; 4: 31 Ma XM, Yoon SO, Richardson CJ, Julich K, Blenis J. SKAR links pre-mRNA splicing pii a011593. to mTOR/S6K1-mediated enhanced translation efficiency of spliced mRNAs. 9 Montagne J, Stewart MJ, Stocker H, Hafen E, Kozma SC, Thomas G. Drosophila S6 Cell 2008; 133: 303–313. kinase: a regulator of cell size. Science 1999; 285: 2126–2129. 32 Dibble CC, Asara JM, Manning BD. Characterization of Rictor phosphorylation sites 10 Pende M, Kozma SC, Jaquet M, Oorschot V, Burcelin R, Marchand-Brustel Y et al. reveals direct regulation of mTOR complex 2 by S6K1. Mol Cell Biol 2009; 29: Hypoinsulinaemia, glucose intolerance and diminished beta-cell size in S6K1- 5657–5670. deficient mice. Nature 2000; 408: 994–997. 33 Tremblay F, Brule S, Hee US, Li Y, Masuda K, Roden M et al. Identification of IRS-1 11 Shahbazian D, Roux PP, Mieulet V, Cohen MS, Raught B, Taunton J et al. The Ser-1101 as a target of S6K1 in nutrient- and obesity-induced insulin resistance. mTOR/PI3K and MAPK pathways converge on eIF4B to control its phosphorylation Proc Natl Acad Sci USA 2007; 104: 14056–14061. and activity. EMBO J 2006; 25: 2781–2791. 34 Ruvinsky I, Sharon N, Lerer T, Cohen H, Stolovich-Rain M, Nir T et al. Ribosomal 12 Wang X, Li W, Williams M, Terada N, Alessi DR, Proud CG. Regulation of elongation protein S6 phosphorylation is a determinant of cell size and glucose homeostasis. factor 2 kinase by p90(RSK1) and p70 S6 kinase. EMBO J 2001; 20: 4370–4379. Genes Dev 2005; 19: 2199–2211. 13 Holz MK, Ballif BA, Gygi SP, Blenis J. mTOR and S6K1 mediate assembly of the 35 Huber A, French SL, Tekotte H, Yerlikaya S, Stahl M, Perepelkina MP et al. Sch9 translation preinitiation complex through dynamic protein interchange and regulates ribosome biogenesis via Stb3, Dot6 and Tod6 and the histone deace- ordered phosphorylation events. Cell 2005; 123: 569–580. tylase complex RPD3L. EMBO J 2011; 30: 3052–3064. 14 Mieulet V, Roceri M, Espeillac C, Sotiropoulos A, Ohanna M, Oorschot V et al. S6 36 Yoon A, Peng G, Brandenburger Y, Zollo O, Xu W, Rego E et al. Impaired control of kinase inactivation impairs growth and translational target phosphorylation in IRES-mediated translation in X-linked dyskeratosis congenita. Science 2006; 312: muscle cells maintaining proper regulation of protein turnover. Am J Physiol Cell 902–906. Physiol 2007; 293: C712–C722. 37 Ge J, Rudnick DA, He J, Crimmins DL, Ladenson JH, Bessler M et al. Dyskerin 15 Choo AY, Yoon SO, Kim SG, Roux PP, Blenis J. Rapamycin differentially inhibits ablation in mouse liver inhibits rRNA processing and cell division. Mol Cell Biol S6Ks and 4E-BP1 to mediate cell-type-specific repression of mRNA translation. 2010; 30: 413–422. Proc Natl Acad Sci USA 2008; 105: 17414–17419. 38 Espeillac C, Mitchell C, Celton-Morizur S, Chauvin C, Koka V, Gillet C et al. S6 kinase 16 Thoreen CC, Kang SA, Chang JW, Liu Q, Zhang J, Gao Y et al. An ATP-competitive 1 is required for rapamycin-sensitive liver proliferation after mouse hepatectomy. mTOR inhibitor reveals rapamycin-insensitive functions of mTORC1. J Biol Chem J Clin Invest 2011; 121: 2821–2832. 2009; 284: 8023–8032. 39 Jorgensen P, Nishikawa JL, Breitkreutz BJ, Tyers M. Systematic identification of 17 Thoreen CC, Chantranupong L, Keys HR, Wang T, Gray NS, Sabatini DM. A unifying pathways that couple cell growth and division in yeast. Science 2002; 297: model for mTORC1-mediated regulation of mRNA translation. Nature 2012; 485: 395–400. 109–113. 40 Jorgensen P, Rupes I, Sharom JR, Schneper L, Broach JR, Tyers M. A dynamic 18 Hsieh AC, Liu Y, Edlind MP, Ingolia NT, Janes MR, Sher A et al. The translational transcriptional network communicates growth potential to ribosome synthesis landscape of mTOR signalling steers cancer initiation and metastasis. Nature 2012; and critical cell size. Genes Dev 2004; 18: 2491–2505. 485: 55–61. 41 Guertin DA, Guntur KV, Bell GW, Thoreen CC, Sabatini DM. Functional genomics 19 Powley IR, Kondrashov A, Young LA, Dobbyn HC, Hill K, Cannell IG et al. Trans- identifies TOR-regulated genes that control growth and division. Curr Biol 2006; lational reprogramming following UVB irradiation is mediated by DNA-PKcs and 16: 958–970. allows selective recruitment to the polysomes of mRNAs encoding DNA repair 42 Urban J, Soulard A, Huber A, Lippman S, Mukhopadhyay D, Deloche O et al. Sch9 enzymes. Genes Dev 2009; 23: 1207–1220. is a major target of TORC1 in Saccharomyces cerevisiae. Mol Cell 2007; 26: 20 Patursky-Polischuk I, Stolovich-Rain M, Hausner-Hanochi M, Kasir J, Cybulski N, 663–674. Avruch J et al. The TSC-mTOR pathway mediates translational activation of TOP 43 Hannan KM, Brandenburger Y, Jenkins A, Sharkey K, Cavanaugh A, Rothblum L et mRNAs by insulin largely in a raptor- or rictor-independent manner. Mol Cell Biol al. mTOR-dependent regulation of ribosomal gene transcription requires S6K1 2008; 29: 640–649. and is mediated by phosphorylation of the carboxy-terminal activation domain of 21 Lempiainen H, Shore D. Growth control and ribosome biogenesis. Curr Opin Cell the nucleolar transcription factor UBF. Mol Cell Biol 2003; 23: 8862–8877. Biol 2009; 21: 855–863. 44 Mayer C, Zhao J, Yuan X, Grummt I. mTOR-dependent activation of the tran- 22 Kiss T, Fayet-Lebaron E, Jady BE. Box H/ACA small ribonucleoproteins. Mol Cell scription factor TIF-IA links rRNA synthesis to nutrient availability. Genes Dev 2004; 2010; 37: 597–606. 18: 423–434. 23 Girard JP, Lehtonen H, Caizergues-Ferrer M, Amalric F, Tollervey D, Lapeyre B. 45 Lempiainen H, Uotila A, Urban J, Dohnal I, Ammerer G, Loewith R et al. Sfp1 GAR1 is an essential small nucleolar RNP protein required for pre-rRNA processing interaction with TORC1 and Mrs6 reveals feedback regulation on TOR signaling. in yeast. EMBO J 1992; 11: 673–682. Mol Cell 2009; 33: 704–716. 24 Liu PC, Thiele DJ. Novel stress-responsive genes EMG1 and NOP14 encode con- 46 Iadevaia V, Zhang Z, Jan E, Proud CG. mTOR signaling regulates the processing of served, interacting proteins required for 40S ribosome biogenesis. Mol Biol Cell pre-rRNA in human cells. Nucleic Acids Res 2011; 40: 2527–2539. 2001; 12: 3644–3657. 47 Mariappan MM, D’Silva K, Lee MJ, Sataranatarajan K, Barnes JL, Choudhury GG et 25 Venema J, Vos HR, Faber AW, van Venrooij WJ, Raue HA. Yeast Rrp9p is an al. Ribosomal biogenesis induction by high glucose requires activation of evolutionarily conserved U3 snoRNP protein essential for early pre-rRNA upstream binding factor in kidney glomerular epithelial cells. Am J Physiol Renal processing cleavages and requires box C for its association. RNA 2000; 6: Physiol 2011; 300: F219–F230. 1660–1671. 48 De S, Brogna S. Are ribosomal proteins present at transcription sites on or off 26 Bernstein KA, Granneman S, Lee AV, Manickam S, Baserga SJ. Comprehensive ribosomal subunits? Biochem Soc Trans 2010; 38: 1543–1547. mutational analysis of yeast DEXD/H box RNA helicases involved in large ribo- 49 Malygin AA, Parakhnevitch NM, Ivanov AV, Eperon IC, Karpova GG. Human ribosomal subunit biogenesis. Mol Cell Biol 2006; 26: 1195–1208. somal protein S13 regulates expression of its own gene at the splicing step by a 27 Dosil M, Bustelo XR. Functional characterization of Pwp2, a WD family protein feedback mechanism. Nucleic Acids Res 2007; 35: 6414–6423. essential for the assembly of the 90 S pre-ribosomal particle. J Biol Chem 2004; 50 Dai MS, Sun XX, Lu H. Ribosomal protein L11 associates with c-Myc at 5 S rRNA 279: 37385–37397. and tRNA genes and regulates their expression. J Biol Chem 2010; 285: 28 Gautier T, Berges T, Tollervey D, Hurt E. Nucleolar KKE/D repeat proteins Nop56p 12587–12594. and Nop58p interact with Nop1p and are required for ribosome biogenesis. Mol 51 De S, Varsally W, Falciani F, Brogna S. Ribosomal proteins’ association with tran- Cell Biol 1997; 17: 7088–7098. scription sites peaks at tRNA genes in Schizosaccharomyces pombe. RNA 2011; 29 De Marchis ML, Giorgi A, Schinina ME, Bozzoni I, Fatica A. Rrp15p a novel com- 17: 1713–1726. ponent of pre-ribosomal particles required for 60S ribosome subunit maturation. 52 Chomczynski P, Sacchi N. Single-step method of RNA isolation by acid guanidinium RNA 2005; 11: 495–502. thiocyanate-phenol-chloroform extraction. Anal Biochem 1987; 162:156–159.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

& 2014 Macmillan Publishers Limited Oncogene (2014) 474 – 483