HSP90 and its R2TP/Prefoldin-like cochaperone are involved in the cytoplasmic assembly of RNA polymerase II.

RNA polymerases are key multisubunit cellular enzymes. Microscopy studies indicated that RNA polymerase I assembles near its promoter. However, the mechanism by which RNA polymerase II is assembled from its 12 subunits remains unclear. We show here that RNA polymerase II subunits Rpb1 and Rpb3 accumulate in the cytoplasm when assembly is prevented and that nuclear import of Rpb1 requires the presence of all subunits. Using MS-based quantitative proteomics, we characterized assembly intermediates. These included a cytoplasmic complex containing subunits Rpb1 and Rpb8 associated with the HSP90 cochaperone hSpagh (RPAP3) and the R2TP/Prefoldin-like complex. Remarkably, HSP90 activity stabilized incompletely assembled Rpb1 in the cytoplasm. Our data indicate that RNA polymerase II is built in the cytoplasm and reveal quality-control mechanisms that link HSP90 to the nuclear import of fully assembled enzymes. hSpagh also bound the free RPA194 subunit of RNA polymerase I, suggesting a general role in assembling RNA polymerases.

Drosophila Spag is the homolog of RNA polymerase II-associated protein 3 (RPAP3) and recruits the heat shock proteins 70 and 90 (Hsp70 and Hsp90) during the assembly of cellular machineries.

The R2TP is a recently identified Hsp90 co-chaperone, composed of four proteins as follows: Pih1D1, RPAP3, and the AAA(+)-ATPases RUVBL1 and RUVBL2. In mammals, the R2TP is involved in the biogenesis of cellular machineries such as RNA polymerases, small nucleolar ribonucleoparticles and phosphatidylinositol 3-kinase-related kinases. Here, we characterize the spaghetti (spag) gene of Drosophila, the homolog of human RPAP3. This gene plays an essential function during Drosophila development. We show that Spag protein binds Drosophila orthologs of R2TP components and Hsp90, like its yeast counterpart. Unexpectedly, Spag also interacts and stimulates the chaperone activity of Hsp70. Using null mutants and flies with inducible RNAi, we show that spaghetti is necessary for the stabilization of snoRNP core proteins and target of rapamycin activity and likely the assembly of RNA polymerase II. This work highlights the strong conservation of both the HSP90/R2TP system and its clients and further shows that Spag, unlike Saccharomyces cerevisiae Tah1, performs essential functions in metazoans. Interaction of Spag with both Hsp70 and Hsp90 suggests a model whereby R2TP would accompany clients from Hsp70 to Hsp90 to facilitate their assembly into macromolecular complexes.

CRM1 controls the composition of nucleoplasmic pre-snoRNA complexes to licence them for nucleolar transport.

Transport of C/D snoRNPs to nucleoli involves nuclear export factors. In particular, CRM1 binds nascent snoRNPs, but its precise role remains unknown. We show here that both CRM1 and nucleocytoplasmic trafficking are required to transport snoRNPs to nucleoli, but the snoRNPs do not transit through the cytoplasm. Instead, CRM1 controls the composition of nucleoplasmic pre-snoRNP complexes. We observed that Tgs1 long form (Tgs1 LF), the long isoform of the cap hypermethylase, contains a leucine-rich nuclear export signal, shuttles in a CRM1-dependent manner, and binds to the nucleolar localization signal (NoLS) of the core snoRNP protein Nop58. In vitro data indicate that CRM1 binds Tgs1 LF and promotes its dissociation from Nop58 NoLS, and immunoprecipitation experiments from cells indicate that the association of Tgs1 LF with snoRNPs increases upon CRM1 inhibition. Thus, CRM1 appears to promote nucleolar transport of snoRNPs by removing Tgs1 LF from the Nop58 NoLS. Microarray/IP data show that this occurs on most snoRNPs, from both C/D and H/ACA families, and on the telomerase RNA. Hence, CRM1 provides a general molecular link between nuclear events and nucleocytoplasmic trafficking.

HSP90 and the R2TP co-chaperone complex: building multi-protein machineries essential for cell growth and gene expression.

HSP90 (Heat Shock Protein 90) is an essential chaperone involved in the last folding steps of client proteins. It has many clients, and these are often recognized through specific adaptors. Recently, the conserved R2TP complex was identified as a key HSP90 co-chaperone. Current evidences indicate that the HSP90/R2TP system assembles multi-molecular protein complexes. Strikingly, these comprise basic machineries of gene expression: (1) nuclear RNA polymerases; (2) the snoRNPs, essential to produce ribosomes; and (3) mTOR Complex 1 and 2, which control translational activity and cell growth. Another important substrate is the telomerase RNP, required for continuous cell proliferation. We discuss here the assembly of RNA polymerases in bacteria and eukaryotes, the role of HSP90/R2TP in this process and in the assembly of snoRNPs and the PIKK family of TORC1 kinase. Finally, we speculate on the roles of R2TP as a master regulator of cell growth under normal or pathological conditions.

The TRRAP transcription cofactor represses interferon-stimulated genes in colorectal cancer cells.

Transcription is essential for cells to respond to signaling cues and involves factors with multiple distinct activities. One such factor, TRRAP, functions as part of two large complexes, SAGA and TIP60, which have crucial roles during transcription activation. Structurally, TRRAP belongs to the phosphoinositide 3 kinase-related kinases (PIKK) family but is the only member classified as a pseudokinase. Recent studies established that a dedicated HSP90 co-chaperone, the triple T (TTT) complex, is essential for PIKK stabilization and activity. Here, using endogenous auxin-inducible degron alleles, we show that the TTT subunit TELO2 promotes TRRAP assembly into SAGA and TIP60 in human colorectal cancer cells (CRCs). Transcriptomic analysis revealed that TELO2 contributes to TRRAP regulatory roles in CRC cells, most notably of MYC target genes. Surprisingly, TELO2 and TRRAP depletion also induced the expression of type I interferon genes. Using a combination of nascent RNA, antibody-targeted chromatin profiling (CUT&RUN), ChIP, and kinetic analyses, we propose a model by which TRRAP directly represses the transcription of IRF9, which encodes a master regulator of interferon-stimulated genes. We have therefore uncovered an unexpected transcriptional repressor role for TRRAP, which we propose contributes to its tumorigenic activity.

Massive translational repression of gene expression during mouse erythroid differentiation.

We took advantage of a mouse erythroid differentiation system to determine the relative contribution of transcriptional and translational control during this process. Comparison of expression data obtained with total cytoplasmic mRNAs or polysome-bound mRNAs (actively translated mRNAs) on Affymetrix high-density oligonucleotide microarrays revealed different characteristics of the two regulatory mechanisms. Indeed, mRNA expression from a vast majority of genes was affected, albeit most changes were relatively small and occurred at a low pace. Translational control, however, affected a smaller fraction of genes but was effective at earlier time-points. This analysis unravels six clusters of genes showing no significant variation in mRNA expression levels whereas they are submitted to translational regulation. Their involvement in terminal mouse erythropoiesis may prove to be highly relevant. Furthermore, the data from specific and functional categories of genes emphasize that translational control, not only reinforces the transcriptional effect, but allows the cell to increase the complexity in gene expression regulation patterns.

The HSP90/R2TP assembly chaperone promotes cell proliferation in the intestinal epithelium.

The R2TP chaperone cooperates with HSP90 to integrate newly synthesized proteins into multi-subunit complexes, yet its role in tissue homeostasis is unknown. Here, we generated conditional, inducible knock-out mice for Rpap3 to inactivate this core component of R2TP in the intestinal epithelium. In adult mice, Rpap3 invalidation caused destruction of the small intestinal epithelium and death within 10 days. Levels of R2TP substrates decreased, with strong effects on mTOR, ATM and ATR. Proliferative stem cells and progenitors deficient for Rpap3 failed to import RNA polymerase II into the nucleus and they induced p53, cell cycle arrest and apoptosis. Post-mitotic, differentiated cells did not display these alterations, suggesting that R2TP clients are preferentially built in actively proliferating cells. In addition, high RPAP3 levels in colorectal tumors from patients correlate with bad prognosis. Here, we show that, in the intestine, the R2TP chaperone plays essential roles in normal and tumoral proliferation.

Cyclin B synthesis and rapamycin-sensitive regulation of protein synthesis during starfish oocyte meiotic divisions.

Translation of cyclin mRNAs represents an important event for proper meiotic maturation and post-fertilization mitoses in many species. Translational control of cyclin B mRNA has been described to be achieved through two separate but related mechanisms: translational repression and polyadenylation. In this paper, we evaluated the contribution of global translational regulation by the cap-dependent translation repressor 4E-BP (eukaryotic initiation factor 4E-binding protein) on the cyclin B protein synthesis during meiotic maturation of the starfish oocytes. We used the immunosupressant drug rapamycin, a strong inhibitor of cap-dependent translation, to check for the involvement of this protein synthesis during this physiological process. Rapamycin was found to prevent dissociation of 4E-BP from the initiation factor eIF4E and to suppress correlatively a burst of global protein synthesis occurring at the G2/M transition. The drug had no effect on first meiotic division but defects in meiotic spindle formation prevented second polar body emission, demonstrating that a rapamycin-sensitive pathway is involved in this mechanism. While rapamycin affected the global protein synthesis, the drug altered neither the specific translation of cyclin B mRNA nor the expression of the Mos protein. The expression of these two proteins was correlated with the phosphorylation and the dissociation of the cytoplasmic polyadenylation element-binding protein from eIF4E.

Deep Structural Analysis of RPAP3 and PIH1D1, Two Components of the HSP90 Co-chaperone R2TP Complex.

RPAP3 and PIH1D1 are part of the HSP90 co-chaperone R2TP complex involved in the assembly process of many molecular machines. In this study, we performed a deep structural investigation of the HSP binding abilities of the two TPR domains of RPAP3. We combined 3D NMR, non-denaturing MS, and ITC techniques with Y2H, IP-LUMIER, FRET, and ATPase activity assays and explain the fundamental role played by the second TPR domain of RPAP3 in the specific recruitment of HSP90. We also established the 3D structure of an RPAP3:PIH1D1 sub-complex demonstrating the need for a 34-residue insertion, specific of RPAP3 isoform 1, for the tight binding of PIH1D1. We also confirm the existence of a complex lacking PIH1D1 in human cells (R2T), which shows differential binding to certain clients. These results highlight similarities and differences between the yeast and human R2TP complexes, and document the diversification of this family of co-chaperone complexes in human.

The RPAP3-Cterminal domain identifies R2TP-like quaternary chaperones.

R2TP is an HSP90 co-chaperone that assembles important macro-molecular machineries. It is composed of an RPAP3-PIH1D1 heterodimer, which binds the two essential AAA+ATPases RUVBL1/RUVBL2. Here, we resolve the structure of the conserved C-terminal domain of RPAP3, and we show that it directly binds RUVBL1/RUVBL2 hexamers. The human genome encodes two other proteins bearing RPAP3-C-terminal-like domains and three containing PIH-like domains. Systematic interaction analyses show that one RPAP3-like protein, SPAG1, binds PIH1D2 and RUVBL1/2 to form an R2TP-like complex termed R2SP. This co-chaperone is enriched in testis and among 68 of the potential clients identified, some are expressed in testis and others are ubiquitous. One substrate is liprin-α2, which organizes large signaling complexes. Remarkably, R2SP is required for liprin-α2 expression and for the assembly of liprin-α2 complexes, indicating that R2SP functions in quaternary protein folding. Effects are stronger at 32 °C, suggesting that R2SP could help compensating the lower temperate of testis.

TWE-PRIL; a fusion protein of TWEAK and APRIL.

TWEAK and APRIL are both members of the tumor necrosis factor family, which are involved in respectively angiogenesis and immune regulation. While TWEAK is processed at the cell surface, APRIL is processed inside the cell by a furin-convertase and is solely able to perform its function as a soluble factor. Recently, TWE-PRIL has been identified, which is an endogenous hybrid transcript between TWEAK and APRIL. TWE-PRIL is a transmembrane protein that consists of a TWEAK intracellular, transmembrane and stalk region combined with APRIL as its receptor-binding domain. As such TWE-PRIL is expressed at the cell surface. Although TWE-PRIL, like APRIL, can stimulate T and B cell lines, distinct biological functions that may result from its membrane anchoring cannot be excluded. Understanding the function of this newly identified protein will contribute to the elucidation of the complexity of the tumor necrosis factor family.

Proteomic and 3D structure analyses highlight the C/D box snoRNP assembly mechanism and its control.

In vitro, assembly of box C/D small nucleolar ribonucleoproteins (snoRNPs) involves the sequential recruitment of core proteins to snoRNAs. In vivo, however, assembly factors are required (NUFIP, BCD1, and the HSP90-R2TP complex), and it is unknown whether a similar sequential scheme applies. In this paper, we describe systematic quantitative stable isotope labeling by amino acids in cell culture proteomic experiments and the crystal structure of the core protein Snu13p/15.5K bound to a fragment of the assembly factor Rsa1p/NUFIP. This revealed several unexpected features: (a) the existence of a protein-only pre-snoRNP complex containing five assembly factors and two core proteins, 15.5K and Nop58; (b) the characterization of ZNHIT3, which is present in the protein-only complex but gets released upon binding to C/D snoRNAs; (c) the dynamics of the R2TP complex, which appears to load/unload RuvBL AAA(+) adenosine triphosphatase from pre-snoRNPs; and (d) a potential mechanism for preventing premature activation of snoRNP catalytic activity. These data provide a framework for understanding the assembly of box C/D snoRNPs.

The Hsp90 chaperone controls the biogenesis of L7Ae RNPs through conserved machinery.

RNA-binding proteins of the L7Ae family are at the heart of many essential ribonucleoproteins (RNPs), including box C/D and H/ACA small nucleolar RNPs, U4 small nuclear RNP, telomerase, and messenger RNPs coding for selenoproteins. In this study, we show that Nufip and its yeast homologue Rsa1 are key components of the machinery that assembles these RNPs. We observed that Rsa1 and Nufip bind several L7Ae proteins and tether them to other core proteins in the immature particles. Surprisingly, Rsa1 and Nufip also link assembling RNPs with the AAA + adenosine triphosphatases hRvb1 and hRvb2 and with the Hsp90 chaperone through two conserved adaptors, Tah1/hSpagh and Pih1. Inhibition of Hsp90 in human cells prevents the accumulation of U3, U4, and telomerase RNAs and decreases the levels of newly synthesized hNop58, hNHP2, 15.5K, and SBP2. Thus, Hsp90 may control the folding of these proteins during the formation of new RNPs. This suggests that Hsp90 functions as a master regulator of cell proliferation by allowing simultaneous control of cell signaling and cell growth.

Nuclear envelope breakdown may deliver an inhibitor of protein phosphatase 1 which triggers cyclin B translation in starfish oocytes.

In vertebrates, enhanced translation of mRNAs in oocytes and early embryos entering M-phase is thought to occur through polyadenylation, involving binding, hyperphosphorylation and proteolytic degradation of Aurora-activated CPEB. In starfish, an unknown component of the oocyte nucleus is required for cyclin B synthesis following the release of G2/prophase block by hormonal stimulation. We have found that CPEB cannot be hyperphosphorylated following hormonal stimulation in starfish oocytes from which the nucleus has been removed. Activation of Aurora kinase, known to interact with protein phosphatase 1 and its specific inhibitor Inh-2, is also prevented. The microinjection of Inh-2 restores Aurora activation, CPEB hyperphosphorylation and cyclin B translation in enucleated oocytes. Nevertheless, we provide evidence that CPEB is in fact hyperphosphorylated by cdc2, without apparent involvement of Aurora or MAP kinase, and that cyclin B synthesis can be stimulated without previous degradation of phosphorylated CPEB. Thus, the regulation of cyclin B synthesis necessary for progression through meiosis can be explained by an equilibrium between CPEB phosphorylation and dephosphorylation, and both aspects of this control may rely on the sole activation of Cdc2 and subsequent nuclear breakdown.

Regulation of mRNA expression in macrophages after Yersinia enterocolitica infection. Role of different Yop effectors.

The Yop virulon, which comprises a complete type III secretion system and secreted proteins, allows bacteria from the genus Yersinia to resist the nonspecific immune response of the host. This virulon, which is encoded by a plasmid called pYV in Yersinia enterocolitica, enables extracellular bacteria to inject six Yop effectors (YopE, -H, -T, -O, -P, -M) into the host cell. To investigate the role of YopP, YopM, and the other pYV-encoded factors on the expression of the host cell genes, we characterized the transcriptome alterations in infected mouse macrophages using the microarray technique. PU5-1.8 macrophages were infected either with an avirulent (pYV(-)), a wild type (pYV(+)), or two knockout (yopP(-) and yopM(-)) mutants of Y. enterocolitica. Expression alterations in response to Y. enterocolitica infection were monitored for 6657 genes. Among those, 857 genes were affected, 339 of which were specifically regulated by the action of the Yop virulon. Further analysis of those 339 genes allowed identification of specific targets of YopP, YopM, or the other pYV-encoded factors. According to these results, the main action of the Yop virulon is to counteract the host cell pro-inflammatory response to the infection. YopP participates to this inhibition, whereas another pYV-encoded factor appears to also be involved in this down-regulation. Besides, YopM was found to induce the regulation of genes involved in cell cycle and cell growth, revealing for the first time an in vitro effect for YopM. In addition to YopM, other pYV factors distinct from YopP affected the expression of genes involved in cycling. In conclusion, these results provide new insight into the mechanisms of Yersinia pathogenicity by identifying the changes in host genes expression after infection and highlight the concerted actions of the different Yop effectors.

Global and specific translational control by rapamycin in T cells uncovered by microarrays and proteomics.

Rapamycin has been shown to affect translation. We have utilized two complementary approaches to identify genes that are predominantly affected by rapamycin in Jurkat T cells. One was to compare levels of polysome-bound and total RNA using oligonucleotide microarrays complementary to 6,300 human genes. Another was to determine protein synthesis levels using two-dimensional PAGE. Analysis of expression changes at the polysome-bound RNA levels showed that translation of most of the expressed genes was partially reduced following rapamycin treatment. However, translation of 136 genes (6% of the expressed genes) was totally inhibited. This group included genes encoding RNA-binding proteins and several proteasome subunit members. Translation of a set of 159 genes (7%) was largely unaffected by rapamycin treatment. These genes included transcription factors, kinases, phosphatases, and members of the RAS superfamily. Analysis of [(35)S]methionine-labeled proteins from the same cell populations using two-dimensional PAGE showed that the integrated intensity of 111 of 830 protein spots changed in rapamycin-treated cells by at least 3-fold (70 increased, 41 decreased). We identified 22 affected protein spots representing protein products of 16 genes. The combined microarray and proteomic approach has uncovered novel genes affected by rapamycin that may be involved in its immunosuppressive effect and other genes that are not affected at the level of translation in a context of general inhibition of cap-dependent translation.

T cell activation in vivo targets diacylglycerol kinase alpha to the membrane: a novel mechanism for Ras attenuation.

Diacylglycerol kinase (DGK) phosphorylates diacylglycerol to produce phosphatidic acid, leading to decreased and increased levels, respectively, of these two lipid messengers that play a central role in T cell activation. Nine DGK isoforms, grouped into five subtypes, are found in higher organisms; all contain a conserved C-terminal domain and at least two cysteine-rich motifs of unknown function. In this study, we have researched in vivo the regulation of DGK alpha, using a transgenic mouse model in which injection of an antigenic peptide activates the majority of peripheral T cells. We demonstrate that DGK alpha, highly expressed in resting T lymphocytes, is subject to complex control at the mRNA and protein levels during in vivo T cell activation. Subcellular fractionation of T lymphocytes shortly after in vivo engagement of the TCR shows rapid translocation of cytosolic DGK alpha to the membrane fraction. At early time points, DGK alpha translocation to the membrane correlates with rapid translocation of Ras guanyl nucleotide-releasing protein (RasGRP), a nucleotide exchange activator for Ras that associates to the membrane through a diacylglycerol-binding domain. To demonstrate a causal relationship between DGK alpha activity and RasGRP relocation to the membrane, we determined RasGRP translocation kinetics in a T cell line transiently transfected with constitutive active and dominant-negative DGK alpha mutants. We show that membrane localization of DGK alpha is associated with a negative regulatory signal for Ras activation by reversing RasGRP translocation. This study is the first demonstration of in vivo regulation of DGK alpha, and provides new insight into the functional role of a member of this family of lipid kinases in the regulation of the immune response.