NEDDylation promotes stress granule assembly

Abstract

Stress granules (SGs) harbour translationally stalled messenger ribonucleoproteins and play important roles in regulating gene expression and cell fate. Here we show that neddylation promotes SG assembly in response to arsenite-induced oxidative stress. Inhibition or depletion of key components of the neddylation machinery concomitantly inhibits stress-induced polysome disassembly and SG assembly. Affinity purification and subsequent mass-spectrometric analysis of Nedd8-conjugated proteins from translationally stalled ribosomal fractions identified ribosomal proteins, translation factors and RNA-binding proteins (RBPs), including SRSF3, a previously known SG regulator. We show that SRSF3 is selectively neddylated at Lys85 in response to arsenite. A non-neddylatable SRSF3 (K85R) mutant do not prevent arsenite-induced polysome disassembly, but fails to support the SG assembly, suggesting that the neddylation pathway plays an important role in SG assembly.

Figure 1. Neddylation pathway regulates SG assembly.

(a) Images from siRNA screen plates with RDG3 stable cells showing knockdown effect of UBE2M on SG assembly. (b) U2OS cells transfected with siCONT or siUBE2M for 72 h were cultured in the absence or presence of arsenite (0.2 mM) for indicated time points, and then immunostained against SG marker eIF3b (green), PB marker RCK (red). Nuclei (Blue) are counterstained with bisbenzamide. (c) Bar graph representing the percentage of cells bearing SGs. Error bars indicate s.e.m. (n=4). **P<0.01, ***P<0.001, Student's t-test. (d) UBE2M knockdown efficiency was assessed using western blot analysis. (e) U2OS cells transfected with siCONT, siNEDD8 or mixture of siNEDD8 and siUBE2M were treated with 0.2 mM arsenite for indicated time points prior to processing for immunofluorescence microscopy using anti-eIF3b and anti-RCK antibodies. (f) Statistical graph showing percentage of cells bearing SGs. Error bars indicate s.e.m. (n=4). **P<0.01, ***P<0.001, Student's t-test. (g) Western analysis for knockdown efficiency of NEDD8 and UBE2M. (h) U2OS cells pretreated with DMSO or 1 μM ML4924 for 18 h were treated with 0.2 mM arsenite for indicated time points and stained against eIF3b and RCK. (i) Statistical data showing percentage of cells bearing SGs. Error bars indicate s.e.m. (n=3). **P<0.01, ***P<0.001, Student's t-test. Scale bar, 10 μm.

Figure 2. UBE2M and NEDD8 are integral components of SGs.

(a–d) UBE2M is a component of SGs. U2OS cells grown on coverslips were either (a) untreated (mock) or treated with (b) 0.5 mM arsenite (c) 20 μM clotrimazole (d) 1 μM thapsigargin for 1 h. Samples were processed and immunostained against eIF3b (green) and UBE2M (red). (e–h) NEDD8 is a component of SGs. U2OS cells either (e) untreated or treated with (f) 0.5 mM arsenite (g) 20 μM clotrimazole (h) 1 μM thapsigargin for 1 h were processed and immunostained against G3BP (green) and NEDD8 (red). Boxed regions are enlarged as both merged and separate channel views. Arrowhead indicates SG. Scale bar, 10 μm.

Figure 3. Identification of neddylated target proteins associated with translation machinery.

(a) Arsenite-induced neddylated proteins are enriched in 80S monosome fractions. Mock or arsenite- (0.5 mM for 1 h) treated U2OS cells were subjected to polysome profiling analysis as described in Methods. A total of 15 fractions collected from sucrose gradient were acetone precipitated, centrifuged and the air-dried pellets were dissolved in SDS sample buffer. The samples were resolved in SDS–PAGE and blotted with anti-NEDD8 antibody. (b) Clear representation of accumulation of neddylated proteins in selected 80S monosome fractions under arsenite stress. (c) Schematic representation for generation of Flag–biopeptide (FB)-tagged NEDD8 stable cell line (see Methods). (d) Identification of arsenite-induced neddylated proteins in FB–NEDD8 stable cells. U2OS cells stably expressing Bir-A or FB–NEDD8 treated with arsenite (0.5 mM) for 1 h were lysed and subjected to polysome fractionation using sucrose gradients (17–50%). A total of 15 fractions were collected, acetone precipitated, air-dried and resolved in SDS–PAGE gel. The membranes were blocked in normal horse serum for 1 h before being blotted directly against streptavidin–HRP to detect NEDD8 conjugates. Bir-A stable cell line was used as control. (e) Brief procedure to isolate NEDD8 conjugates for MS analysis. (f) Isolation and immunopurification of stress-induced neddylated proteins. Stably expressing Bir-A and FB–NEDD8 cells treated with 0.5 mM arsenite for 1 h were subjected to sucrose gradient fractionation, and fraction No. 7 and 8 were collected, acetone precipitated and FB–NEDD8-conjugated proteins were affinity-purified using streptavidin beads. The precipitates were resolved in 12% SDS–PAGE and stained with Coomassie blue. A total of 17 distinct bands were identified and analysed through LC-MS. One representative identified protein from each band is shown.

Figure 4. SRSF3 is a novel target for stress-induced neddylation in vivo.

(a) HEK293T cells were transiently co-transfected with indicated plasmids. After 36 h, cells were treated with either mock or arsenite, lysed under denaturing condition and affinity-purified using Ni²⁺-NTA agarose beads (see Methods). The precipitates were then blotted against anti-Flag antibody. (b) HEK293T cells transfected with empty vector or Flag–SRSF3 were treated with arsenite, lysed under denaturing condition (1% SDS) and immunoprecipitated using Flag beads (see Methods). (c) In vivo neddylation assay was carried out for cells treated with arsenite in a dose-dependent manner (0.15 to 1 mM) for 1 h. (d) In vivo neddylation assay was performed for cells treated with 0.2 mM arsenite at different time points (0–60 min) as indicated. (e) HEK293T cells transfected with indicated plasmids were treated with 0.2 mM arsenite in lane 3 and 4 for 30 min and 60 min, respectively. Cells in lanes 5 and 6 were treated with 0.2 mM arsenite for 60 min and then allowed to recover from stress by changing with fresh medium for 90 min (lane 5) and 180 min (lane 6) and subjected to neddylation assay. (f) NAE inhibitor MLN4924 significantly attenuates SRSF3 neddylation. HEK293T cells pretreated with DMSO, MLN4924 (1 μM) for 18 h or MLN4924 (3 μM) for 1 h prior to arsenite treatment were subjected to neddylation assay. (g) SiRNA mediated knockdown of UBE2M reduces SRSF3 neddylation. In vivo neddylation assay was performed for HEK293T cells transfected with siCONT or siUBE2M under arsenite stress. (h) Ectopic expression of UBE2M-C111S displays dominant negative effect on SRSF3 neddylation. HEK293T cells co-transfected with indicated plasmids were treated with arsenite and subjected to neddylation assay. (i) UBE2M interacts with SRSF3 under arsenite stress. Empty vector or Flag–SRSF3 transfected cells were treated with arsenite, lysed in IP buffer and immunoprecipitated using Flag agarose beads. The precipitates were subjected to western analysis using anti-UBE2M antibody. (j) NEDP1 overexpression inhibits SRSF3 neddylation. HEK293T cells transfected with indicated plasmids were treated with arsenite and affinity purified using Ni²⁺-NTA agarose beads.

Figure 5. SRSF3 is neddylated at Lys85 under arsenite stress.

(a) Schematic representation of SRSF3 protein domains and positions of lysine residue. SRSF3 contains 2 domains, one RRM and one RS domain and has 5 lysine residues positioned at 11, 23, 85, 146 and 164 (highlighted by underline). (b) SRSF3 is neddylated at Lys85 under arsenite stress. His-NEDD8 was co-transfected with empty vector or Flag–SRSF3 WT and series of single lysine mutant (11, 23, 85, 146, 164), double lysine mutant (11/23, 23/85, 146/164) or null mutant (all lysine mutant) as indicated and treated with 0.5 mM arsenite for 30 min. Cells were then lysed under denaturing condition, affinity-purified with Ni-NTA agarose beads and the precipitates were subjected to western analysis using anti-Flag antibody.

Figure 6. Flag–SRSF3-WT but not K85R mutant can rescue the SG assembly defect in SRSF3-depleted cells.

(a) Stably expressing empty vector, Flag–SRSF3-WT or K85R mutant cells transfected with siCONT or siSRSF3-3′UTR were treated with 0.2 mM arsenite before processing for immunofluorescence microscopy using anti-G3BP and anti-Flag antibody. Percentage of cells bearing SGs in (b) siCONT-transfected cells and (c) siSRSF3-3′UTR-transfected cells are shown as bar graph. Error bars indicate s.e.m. (n=4). **P<0.01, Student's t-test. (d) Knockdown efficiency of endogenous SRSF3 and the expression Flag constructs were assessed by western blot analysis. Scale bar, 10 μm.

Figure 7. SRSF3–K85R displays defects in association with SG components.

(a) Empty vector, Flag–SRSF3-WT or K85R mutant transfected cells were treated with arsenite for 30 min, lysed in IP buffer and immunoprecipitated using Flag agarose beads. The precipitates were subjected to western analysis using antibodies against SG proteins. (b) SRSF3–K85R has a defect in associating with TIA-1 under arsenite stress. Immunoprecipitation was carried out as described in a and the precipitates were blotted against TIA-1 and G3BP. (c) U2OS cells transiently transfected with Flag–SRSF3-WT or K85R mutant were treated with either mock or 0.2 mM arsenite and subjected to immunofluorescence microscopy. Recruitment of TIA-1 into SGs was assessed by co-staining with anti-TIA-1 and anti-Flag antibodies. Arrowhead indicates SG. (d) Model diagram of SG assembly mediated by neddylation pathway. During stressful condition, stress kinases (HRI, PKR, PERK, and GCN2) are activated and phosphorylate eIF2α to initiate abortive translation initiation and polysome disassembly. Subsequently, neddylation pathway possibly targets ribosomal proteins to promote polysome disassembly. Finally, neddylation of SRSF3 promotes SG aggregation through interaction with initiation factors such as eIF4G, eIF3s and SG assembly factor TIA-1. Scale bar, 10 μm.