RNF219 regulates CCR4-NOT function in mRNA translation and deadenylation

Abstract

Post-transcriptional regulatory mechanisms play a role in many biological contexts through the control of mRNA degradation, translation and localization. Here, we show that the RING finger protein RNF219 co-purifies with the CCR4-NOT complex, the major mRNA deadenylase in eukaryotes, which mediates translational repression in both a deadenylase activity-dependent and -independent manner. Strikingly, RNF219 both inhibits the deadenylase activity of CCR4-NOT and enhances its capacity to repress translation of a target mRNA. We propose that the interaction of RNF219 with the CCR4-NOT complex directs the translational repressive activity of CCR4-NOT to a deadenylation-independent mechanism.

Figure 1

RNF219 is a RING dependent ubiquitin ligase. (A) Schematic representation of the RNF219 protein and mutants used in this study. (B) Sequence alignment of the human hRNF219 protein and its Mus musculus, Xenopus laevis and Danio rerio, orthologs. Amino acid sequences were aligned with the Prabi CLUSTALW tool. Identical residues are annotated with an asterisk (*) and coloured in red, whereas similar residues are annotated with two (:) and coloured in green and less similar residues are annotated with one dot (.) and coloured in blue. One conserved RING domain of C3HC4 type was predicted in the N-terminus of all the conserved alignments (highlighted in grey); Zinc coordinating residues are marked with a circle. Mutated residues and corresponding mutants are highlighted by a red box. (C) RNF219 but not the RING domain mutant can form poly-ubiquitin chains in an in vitro assay. FHA-RNF219 and FHA-RNF219-CG (cysteine 55 was replaced by a glycine) affinity purified under high salt condition with FLAG antibody and peptide eluted from transfected 293T cells, were incubated with E1 enzyme, UBCH5a, and biotinylated ubiquitin (lanes 1 and 2, respectively). The third lane contains the purification from untransfected cells as negative control. Ubiquitin was detected with HRP conjugated streptavidin (Strep-HRP).

Figure 2

RNF219 binds to the CCR4-NOT complex. (A) (Left) FLAG and HA tandem affinity chromatography was performed on cytoplasmic S100 extracts prepared from HeLa S3 cells stably expressing FHA-RNF219 (S3-FHA-RNF219) or non-transduced HeLa S3 cells. Proteins were resolved by SDS-PAGE and visualized by silver staining. (Right) The identity of FHA-RNF219-associated proteins was determined by tandem mass spectrometry and percent peptide coverage by amino acid count for CCR4-NOT subunits detected are listed. (B) FLAG-immuno-precipitations from S3-FHA-RNF219 cell extracts confirm that FHA-RNF219 is associated with a number of CCR4-NOT subunits as determined by mass spectrometry. The presence of CCR4-NOT subunits CNOT3, CNOT2, CNOT7 in the IPs was analyzed by immunoblotting (IB). (C) Endogenous RNF219 was immuno-purified from HeLa cell extracts using a specific antibody against RNF219 (RNF219-B) or pre-immune IgG control (IgG-B-PI). The presence of the CCR4-NOT subunit CNOT1, CNOT2 and CNOT3 in the IP was analyzed by IB. RNF219 was detected with RNF219-A antibody (long exposure is shown in Fig. S2B). (D) RNF219 interaction with NOT module components CNOT2 and CNOT3 is CNOT1 dependent. FHA-RNF219 expressing HeLa cells were transfected with siRNA against CNOT1 (+) or control siRNA (−). IPs were performed using anti-FLAG agarose beads and were blotted for the indicated proteins. The arrow indicates the expected CNOT1 band. (E) Schematic representation of RNF219-Id, a CCR4-NOT complex interaction defective RNF219 allele. Amino acids 540–549 (ELDSMMSESD) were replaced by VE. (F) RNF219-Id interacts with the CCR4-NOT complex very inefficiently. 293T cells were transfected with the indicated constructs. Whole cell extracts were immuno-purified with anti-FLAG agarose beads and eluates immunoblotted with the indicated proteins. (G) RNF219-Id localizes both in the nucleus and in the cytoplasm similarly to WT RNF219. Immuno-fluorescence was performed with anti-HA antibody on HEK293T cells transfected with the indicated constructs.

Figure 3

RNF219 affects the translation of a targeted mRNA. (A) Schematic representation of the Renilla Luciferase (RL) reporter mRNA, containing the Renilla luciferase gene and five 19-nt BoxB hairpins sequences at its 3′UTR. Recruitment of HA-tagged protein to the 3′UTR is mediated by the fused λN-peptide, which has a high affinity for the BoxB sequence. (B) Recruitment of RNF219 to the 3′UTR mRNA of the RL reporter inhibits its expression. RL activity was determined in the indicated conditions (NHA-LacZ, NHA-RNF219, NHA-CNOT7, NHA-CNOT1-R, FHA-RNF219). (Right) Renilla Luciferase (RL) activity was normalized on Firefly luciferase (FL) activity expressed from a plasmid not containing BoxB sequences, co-transfected with that of Renilla. RL repression levels are relative to that of the control NHA-LacZ set to 1. Error bars represent SD, n = 3. (Left) Protein levels of NHA-fusion proteins in transfected 293T cells was analyzed by immunoblotting. (C) Ubiquitin ligase activity of RNF219 is not necessary for its repressive role. RL activity was determined in the indicated conditions (NHA-LacZ, NHA-RNF219, NHA-RNF219-CG). (Right) Renilla Luciferase (RL) activity was normalized as in (B). Error bars represent SD, n = 3. (Left) Protein levels of NHA-fusion proteins, in HeLa cells transfected with the corresponding plasmids, were analyzed by immunoblotting. (D) Recruitment of RNF219 to the 3′UTR mRNA of the RL reporter leads to decreased reporter mRNA level. Total RNA from the indicated conditions (NHA-LacZ, NHA-RNF219, NHA-CNOT7, NHA-CNOT1-R, FHA-RNF219) was extracted as described in the “Methods”. RT-QPCR was performed using primers specific from both RL and FL reporters to quantify their respective mRNA level (the RT-QPCR primers used are indicated in Table S1). RL mRNA level was normalized on FL mRNA level. RL repression levels are relative to that of the control NHA-LacZ set to 1. Error bars represent SD, n = 3. (E) RNF219 mutant (RNF219-Cd) that resides predominantly in the cytoplasm also represses the RL expression. (Right) RL activity was determined in the indicated conditions (NHA-LacZ, NHA-RNF219, NHA-RNF219-Cd). Renilla Luciferase (RL) activity was normalized as in (B). Error bars represent SD, n = 3. (Left) Protein levels of NHA-fusion proteins in 293T cells transfected with the corresponding plasmids were analyzed by immunoblotting. (F) RL mRNA level relative to FL mRNA level in the monoribosomal (Mono, fraction 4 in Fig. S3B) and polyribosomal fractions (Poly, fraction 9 in Fig. S3B) of extracts from 293T cells transfected with NHA-LacZ or NHA-RNF219. mRNA levels normalized to 18S mRNA levels were quantified by RT-QPCR and shown relative to Input set to 1. Error bars represent SD, n = 2. (G) RNF219 does not affect global translation. 24 h after siRNA treatment (two different siRNA against RNF219: CDS and UTR; Fig. S1B) a 30-min puromycin pulse is analysed by immunoblot using an anti-puromycin antibody. The two siRNA against RNF219 show no difference in global puromycin incorporation compared to the control (SCR).

Figure 4

RNF219 mediated translational repression is CCR4-NOT dependent. (A) CNOT3 interacts with RNF219 tethered to the mRNA reporter. RNA-IPs were performed on extracts of cells transfected with the RL reporter and NHA-LacZ or NHA-RNF219 using CNOT3 antibody. The level of Renilla mRNA in the CNOT3-IP was highly enriched in NHA-RNF219 transfected cells as compared to NHA-LacZ transfected cells, as quantified by RTPCR. Results representative of two replicates are shown. (B) RNF219 interaction with the CCR4-NOT complex is necessary for RFN219 to repress RL reporter translation efficiently. (Top) RL activity was determined in the indicated conditions (NHA-LacZ, NHA-RNF219 and NHA-RNF219-Id). RL activity repression was determined as in Fig. 3B. Error bars represent SD, n = 3. (Bottom) Protein levels of NHA-LacZ, NHA-RNF219, NHA-RNF219-Id in HeLa cells transfected with the corresponding plasmids were analyzed by immunoblotting. (C) mRNA levels of samples in (B) were quantified as in Fig. 3D. Error bars represent SD, n = 3. (D) Knock down of CCR4-NOT scaffold CNOT1 affects NHA-RNF219 mediated repression. RL activity was determined in the indicated conditions (NHA-LacZ, NHA-RNF219 and NHA-RNF219-Id) in control cells (siSCR) or in siCNOT1 depleted cells (siCNOT1A and siCNOT1B are two siRNA targeting different regions of CNOT1 mRNA). (Right) RL activity repression was determined as in Fig. 3B. RL repression levels are relative to that of the control NHA-LacZ in the siSCR cells, set to 1. Error bars represent SD, n = 3. (Left) Protein levels. The arrow indicates the expected CNOT1 band. The lower is a non-specific band as it does not disappear in siCNOT1 treated cells.

Figure 5

RNF219 affects the polyA tail length of a targeted mRNA. (A) RNF219 stabilizes the reporter mRNA poly(A) tail. ePAT assays were performed using HeLa cells expressing the RL reporter and either of the following constructs: NHA-LacZ, NHA-RNF219, NHA-RNF219-CG or NHA-CNOT7. On the left panel, a PAT primer (Luc_3; Table S1) specific to the poly(A) tail of the RL reporter gene is used. On the right panel a PAT primer specific to the poly(A) tail of the GAPDH control gene is used (Table S1). Corresponding image analyzer profiles are shown at the right of each ePAT gel. (B) RNF219 mRNA poly(A) tail stabilization does not depend on its RING Finger domain but depends on its interaction with the CCR4-NOT complex. ePAT assays were performed using HeLa cells expressing the RL reporter and either of the following constructs: NHA-LacZ, NHA-RNF219, NHA-RNF219-[1–480], NHA-RNF219-[121–726]. On the left panel, a PAT primer (Luc_2; Table S1) specific of the poly(A) tail of the RL reporter gene is used. On the right panel a PAT primer specific of the poly(A) tail of the GAPDH control gene is used. Corresponding image analyzer profiles are shown at the right of each ePAT gel. (C) RNF219 mRNA poly(A) tail stabilization depends on its interaction with the CCR4-NOT complex. ePAT assays were performed using HeLa cells expressing the RL reporter and either of the following constructs: NHA-LacZ, NHA-RNF219, NHA-RNF219-Id. On the left panel, a PAT primer (Luc_3; Table S1) specific to the poly(A) tail of the RL reporter gene is used. On the right panel a PAT primer specific to the poly(A) tail of the GAPDH control gene is used. Corresponding image analyzer profiles are shown at the right of each ePAT gel.

Figure 6

Endogenous RNF219 is implicated in cell cycle regulation. (A) Knock-down of RNF219 induces an increase in p27 protein level. HeLa cells were transfected with two siRNA against RNF219 (CDS and UTR) or CNOT3. RNF219, CNOT3 and p27 protein levels were analyzed by immunoblotting using RNF219-A, CNOT3 and p27 antibodies respectively. (B) Knock-down of RNF219 does not substantially affect p27 mRNA level. RNA was extracted in control (SCR) or in RNF219 depleted cells (CDS). RT-QPCR was performed using primers specific to p27/CDKN1b (Table S1) and was normalized on 18S mRNA level. p27/CDKN1b mRNA level was set to 1 in the control (SCR). Error bars represent SD, n = 3. ns P > 0.05 based on unpaired two-tailed Student’s t test. (C) Volcano plot showing differentially expressed (DE) genes of two replicates between siRNF219 CDS versus control SCR. Blue dots represent genes upregulated in RNF219 depleted cells, red represent downregulated genes, black signifying non-DE genes. Differential expression threshold is BH corrected p-value < 0.05 and |LogFC| > 0.5. (D) The top 10 downregulated (red) and upregulated (blue) GO Biological Process terms using gene set enrichment analysis comparing siRNF219 CDS versus SCR. All enrichments are significant with a BH corrected p-value < 0.001.

Figure 7

Schematic representation of RNF219 function. Recruitment of RNF219/CCR4-NOT complexes leads to mRNA translational repression in the absence of poly(A) tail deadenylation.