The p38 MAPK (mitogen-activated protein kinase)/MK2 [MAPKAP (MAPK-activated protein) kinase-2] signalling pathway is a major regulator of stress- and cytokine-induced gene expression at the transcriptional and post-transcriptional level. Using phosphoproteomics we identified the ER (endoplasmic reticulum)-associated ubiquitin-conjugating enzyme Ube2j1 as a potential substrate of MK2. We demonstrate that Ube2j1 is phosphorylated in a cytokine-, cytosolic stress- and LPS (lipopolysaccharide)-induced manner. The cytosolic stress-induced phosphorylation of Ube2j1 proceeds at Ser184, a site described previously to be phosphorylated in response to ER stress, which is located in a perfect MK2 consensus motif. The cytosolic stress-induced phosphorylation of Ube2j1, but not its ER-stress-induced phosphorylation is sensitive to p38/MK2 inhibitors and abrogated in MK2/MK3-deficient cells. In a pull-down assay we demonstrate the interaction of MK2 with Ube2j1 in HEK (human embryonic kidney)-293T cells. Furthermore, MK2 is able to phosphorylate recombinant Ube2j1, but not the S184A mutant in an in vitro kinase assay. These findings strongly suggest that MK2 directly phosphorylates Ube2j1 at Ser184 upon p38-activating stress in vivo. However, ectopically expressed Ube2j1-S184A mutant displays ubiquitinating activity towards the model substrate ER-synthesized T-cell receptor-α similar to that of the wild-type protein. Interestingly, Ube2j1 is phosphorylated in response to LPS also in macrophages and contributes to MK2-dependent TNFα biosynthesis by a so far unknown mechanism.
p38 MAPK (mitogen-activated protein kinase) is activated by pro-inflammatory cytokines, PAMPs (pathogen-associated molecular patterns), UV radiation, osmotic and chemical stress, and heat shock . The MAPKAP (MAPK-activated protein) kinases MK2 (MAPKAP kinase-2) and MK3 are well-characterized downstream targets of p38 MAPK with similar function, where MK2 plays the dominant physiological role due to its abundance compared with MK3 . MK2 is a major regulator of stress- and cytokine-induced gene expression with well-characterized roles in gene transcription, mRNA stabilization and translation of target transcripts [3–6]. In addition, MK2 has also been shown to phosphorylate the ubiquitin ligase Hdm2 (human double minute 2)/Mdm2 (murine double minute 2) and thereby to regulate p53 protein degradation . HSPB1 [heat-shock protein β-1; also known as HSP27 (heat-shock protein 27 kDa)], a bona fide substrate of MK2, has also been shown to regulate diverse cellular functions including protein folding, cytoskeletal remodelling and protein degradation [8–10]. Except for its role in translational regulation of mRNAs of secreted cytokines, such as TNF (tumour necrosis factor), at the ER (endoplasmic reticulum) [11,12], the role for MK2 in ER function has not been analysed so far, although an evolutionarily conserved role of the p38 pathway in the ER stress response has been established in genetic models [13,14]. Previous studies have identified NogoB, an ER-associated membrane protein, as an MK2 substrate , but the physiological role of this phosphorylation is not yet understood.
Ube2j1 (Ubc6e) and Ube2j2 (Ubc6), which are the mammalian homologues of yeast Ubc6, belong to the non-canonical ubiquitin-conjugating E2 enzyme family  and are ER-membrane-anchored by their C-terminal transmembrane domains. Compared with the yeast protein, Ube2j1 is less conserved than Ube2j2 and exhibits a longer divergent C-terminal segment . The role of Ube2j1 in ERAD (ER-associated protein degradation) has been established in several studies. At first it was demonstrated that Ube2j1 specifically affects proteasomal degradation of two proteins spanning the ER membrane, the subunit of the TCRα (T-cell receptor α) and the mutant CFTR (cystic fibrosis transmembrane conductance regulator)-ΔPhe508 . It was then shown that Ube2j1 is involved in the proteasomal degradation of TRAF2 (TNF-receptor-associated factor 2)  and MHC I  and that it co-operates with the E3 ligases RMA1 and Derlin-1 in degradation of the mutant CFTR . In a previous paper, it was described that ER stress induces the phosphorylation of Ube2j1 in an eIF2α (eukaryotic initiation factor 2α) kinase-dependent manner . In the present study, we show that cytosolic stress, TLR (Toll-like receptor) ligands and cytokines induce phosphorylation of Ube2ji via the p38/MK2 pathway. MK2 directly phosphorylates Ube2j1 at Ser184 and Ube2j1 significantly contributes to the MK2-dependent secretion of TNFα from LPS (lipopolysaccharide)-stimulated macrophages.
Antibodies and reagents
Antibodies against p38 MAPK, phospho-p38 MAPK, phospho-HSPB1 (mouse-Ser86/human-Ser82), phospho-Ser/Thr-PKD (protein kinase D) substrate, MK2 and phospho-MK2 (Thr222) were from Cell Signaling Technology. Anti-GAPDH (glyceraldehyde-3-phosphate dehydrogenase) antibody was from Millipore; anti-Hsc70 (heat-shock cognate 70 stress protein) antibody was from Stressgen; anti-His6 antibody was from Stratagene; and anti-HSPB1, anti-Myc (9E10), anti-GFP (B2), anti-Ube2j1 (18-Y), anti-PRAK/MK5 (A-7), anti-TNFα (L-19), anti-EF-2 (C-14) antibodies and other secondary antibodies were from Santa Cruz Biotechnology. Anti-TTP (tristetraprolin) (SAK21B) antibodies were kindly provided by Dr Andrew R. Clark (School of Immunity and Infection, University of Birmingham, Birmingham, U.K.) . Gö 6983, anisomycin and thapsigargin were from Calbiochem; bortezomib was from LC Laboratories; SB202190, BX912 and BIRB-796 were from Axon Ligands; rottlerin was from Millipore; Ro 31-8220 was from Santa Cruz Biotechnology; and tunicamycin was from Enzo. The MK2/MK3/MK5 inhibitor PHA781089 was provided by Dr Robert J. Mourey (Pfizer). IL1α (interleukin 1α) was a gift from Dr Helmut Holtmann (Institute for Physiological Chemistry, Hannover Medical School, Hannover, Germany). Control (Allstars negative control) and mouse Ube2j1 siRNA 1 (Mm_Ube2j1_3: 5′-CAAGACGAUUUACCUACAATT-3′) were obtained from Qiagen. Ube2j1 siRNA 2 (ID-185364: 5′-CCAGUGUUUCUCACUAAAUTT-3′) and Ube2j1 siRNA 3 (ID-185363: 5′-GCUAAUGGACGAUUUGAAGTT-3′) silencer siRNAs were from Ambion, [γ-33P]ATP was from Hartmann Analytic and all other reagents were obtained from Sigma.
Plasmids and cloning
The mouse Ube2j1 coding region was PCR-amplified from total MEF (mouse embryonic fibroblast) cDNA (forward primer, 5′-CACCATGGAGACCCGCTACAACCTG-3′ and reverse primer, 5′-TCACAAACCAGCATTATAACTCAAAGTC-3′) and cloned into the pENTR-D-TOPO directional cloning vector (Invitrogen) following the manufacturer's protocol. Destination expression vectors with GFP, Myc and His tags were generated by LR recombination reactions (Invitrogen). Site-directed mutagenesis was performed using the QuikChange® mutagenesis kit (Stratagene). The GFP–TCRα ERAD reporter plasmid was a gift from Dr Ron Kopito (Department of Biology, Stanford University, CA, U.S.A.) . The GFP–ERK3 (extracellular-signal-regulated kinase 3) construct was obtained as described previously .
Cell culture and treatments
MK2/MK3-deficient immortalized MEFs and macrophages rescued with MK2, MK2-K79R or control vector by retroviral transduction were generated and maintained as reported previously [4,25,26]. HT29 and HEK (human embryonic kidney)-293T cells were maintained, treated and transfected as described previously [25,27]. siRNA transfection for immortalized macrophages using Hiperfect (Qiagen) was performed as described previously . Tail fibroblasts were isolated from 6–8-week-old mice tail tips. Minced tail tips were sequentially digested with collagenase and trypsin at 37°C and plated on to collagen-coated dishes in DMEM (Dulbecco's modified Eagle's medium) supplemented with 20% FBS, minimal essential medium 1× non-essential amino acids (Gibo®) and antibiotics (100 international units/ml penicillin and 100 μg/ml streptomycin). The cells were split in a ratio of 1:4 and maintained in the same growth medium without coated dishes and were treated as described. Animal experimentation was approved to be in agreement with the animal protection law and ethical issues by the Local Government Hannover under file numbers 509.6-42502-04/782 and 33.14-42502-12/0802.
GST pull-down assays
HEK-293T cells were transfected with the indicated expression constructs and at 24 h post-transfection cells were lysed in kinase assay lysis buffer [20 mM Tris, pH 7, 100 μM EDTA, 1 mM EGTA, 1 mM sodium orthovanadate, 10 mM 2-glycerophosphate, 50 mM NaF, 1% Triton X-100, Complete™ protease inhibitor cocktail (Roche) and phosphatase inhibitor cocktail 2 (Sigma)]. Supernatants were incubated with glutathione–Sepharose beads (GE Healthcare) overnight. Beads were washed five times and boiled in gel-loading dye and then analysed by Western blotting.
Western blotting and fluorescence microscopy
Western blotting was performed as described previously using gradient SDS/PAGE . A polyacrylamide gel at 15% was used for detecting TNF. Fluorescent images of live cells expressing GFP-tagged Ube2j1 were acquired using a Leica DM IL LED microscope attached to a Leica EC3 camera.
Protein purification and in vitro kinase assays
His6–Ube2j1 WT (wild-type) and S184A mutant was expressed in Escherichia coli BL21-DE3 cells, solubilized by sonication. His6-tagged membrane proteins were purified using Ni-NTA (Ni2+-nitrilotriacetate)–agarose beads (Qiagen) in the presence of 1% Triton X-100. The beads were directly used for the MK2 kinase assay using radioactive [γ-33P]ATP as described previously .
In vivo ubiquitination assays
In vivo ubiquitination assays were performed in HEK-293T cells as described previously with minor modifications . At 20 h post-transfection of cells with His6–ubiquitin plasmid and other expression constructs, cells were treated for 3 h with 100 nM bortezomib. Cells were scraped out and 10% lysed in gel-loading dye for input controls. Remaining cells were frozen at −80°C and thawed by the addition of room temperature lysis buffer (100 mM sodium phosphate, pH 7.8, 6 M guanidium chloride, 10 mM imidazole, 10 mM 2-mercaptoethanol and 10 mM Tris/HCL). Before pull-down with Ni-NTA beads (Qiagen), samples were sonicated and clarified by centrifugation to reduce viscosity. The ubiquitination signals obtained with anti-GFP antibodies were quantified by analysing the blots using ImageJ (NIH) and normalized to total His6–ubiquitin. His6–ubiquitination of GFP–TCRα in the absence of ectopic expression of conjugating enzyme was taken as 1 (control).
Lysate-kinase assay for HSPB1-kinase activity
MEF cells were lysed in kinase assay lysis buffer and 10 μl of lysate was used in a 25 μl kinase reaction. Non-radioactive kinase assays were performed as described previously with the addition of PKIs (protein kinase inhibitors) HA-1077 and H7 in the reaction mixture .
TNFα quantification by ELISA and real-time PCR
At 24 h post-siRNA transfection, cells were seeded into 96-well plates at a density of 104 cells/well 4 h before stimulation with LPS (1 μg/ml) for an additional 4 h. Supernatants were collected and quantification of murine TNFα by ELISA was performed as described previously  using commercial kits. RNA was isolated from cells and TNF/actin mRNA expression quantified by real-time PCR analysis as described previously [3,12].
Phosphorylation-induced band-shift of Ube2j1 correlates with p38 MAPK activity
We compared protein phosphorylation in MK2/MK3-deficient anisomycin-stimulated MEFs rescued by MK2 or GFP expression  by a phosphoproteomic approach using Glu-C protease cleavage, IP (immunoprecipitation) by the phospho-PKD-consensus motif antibody (anti-LXRXXpSer/pThr) and subsequent MS of the immunoprecipitated peptides (Cell Signaling Technology Phosphoscan) to identify novel substrates of MK2. This approach was adopted due to the fact that the phosphorylation consensus motif for MK2 is almost identical with that of PKD  and could be used to specifically detect or enrich peptides or proteins phosphorylated at similar motifs. In this screen, the Ube2j1 peptide L179ARQISFKAE188 displayed 7.7-fold higher abundance in IP from MK2-rescued cells when compared with IP of MK2/MK3-deficient control transduced cells, indicating that Ube2j1 could be a substrate of MK2. Since human Ube2j1 phosphorylation was reported previously to induce a marked shift in the mobility of the protein band in SDS/PAGE , we first utilized this property to verify the stress-dependent phosphorylation of Ube2j1 by the p38/MK2 pathway. Treatment of HT29 and HEK-293T cells with various inducers of cellular stress, including UV, anisomycin [protein synthesis inhibitor and p38/JNK (c-Jun N-terminal kinase) activator] and sorbitol (osmotic stress stimuli) resulted in strong p38 phosphorylation and activation as monitored by p38-pThr181pTyr183 (pp38), the phosphorylation-induced band-shift of the downstream component MK2 and the phosphorylation of the MK2 substrate HSPB1 (HSP27) at Ser82. Interestingly, the pattern of activation of the p38/MK2 pathway completely matches the phosphorylation-induced band-shift of Ube2j1. Additionally, the selective p38 MAPK inhibitor BIRB-796 was able to efficiently inhibit both the p38/MK2 pathway and the Ube2j1 phospho-shift (Figure 1). This observation strongly suggests that oxidative-, ribotoxic- and osmotic-type cytosolic stresses induce phosphorylation of Ube2j1 via the p38/MK2 pathway. DTT, a reducing agent interfering with disulfide bond formation, and thapsigargin, an inhibitor of the ER calcium import, are both strong inducers of the unfolded protein response and ER stress. To our surprise, both of these originally described inducers of ER-stress-mediated Ube2j1 phosphorylation  were only weak inducers of phospho-Ube2j1 in these cell lines.
p38 MAPK activity and Ube2j1 phosphorylation correlate in stimulated cells
p38 MAPK-dependent phosphorylation of Ube2j1 at Ser184
The PhosphoSite database (http://www.phosphosite.org) documents the existence of four conserved serine/threonine phosphorylation sites in murine Ube2j1. Of these, Ser184 is located in the peptide immunoprecipitated by the anti-PKD phospho-motif antibody and situated in the LXRXXS motif (Figure 2A) identical with the phosphorylation consensus motif of MK2/MK3 . To confirm the results from phosphoproteomics, we analysed lysates of Myc–Ube2j1-overexpressing HEK-293T cells with the PKD substrate motif antibody and detected a p38-dependent phosphorylation of this motif (Figure 2B). To reconfirm that Ser184 is the anisomycin-induced p38-dependent phosphorylation site in vivo, we mutated this site to alanine (S184A) or aspartate (S184D). When Myc-tagged WT and mutant constructs were expressed in HEK-293T cells, a clear p38 activity-dependent band-shift, which was inhibited by BIRB-796 treatment, was detected only for the WT protein (Figure 2C). The mutant S184A co-migrates with the non-phosphorylated band of the WT protein and cannot be shifted, indicating the loss of the only phosphorylation site for the p38/MK2 pathway. The phospho-mimicking mutant S184D migrates similar to the phosphorylated isoform. These findings support the notion that Ser184 is the only stress-dependent phosphorylation site in Ube2j1. To analyse whether Ser184 phosphorylation affects the subcellular localization of Ube2j1, we monitored the expression of GFP–Ube2j1-WT, GFP–Ube2j1-S184A and GFP–Ube2j1-S184D in HEK-293T cells (Figure 2D). Consistent with the previous findings , there was no significant alteration in the localization of the mutant proteins (Figure 2D).
Murine Ube2j1 phosphorylation at Ser184 is responsible for its stress-induced band-shift in SDS/PAGE
Stress-induced Ube2j1 phosphorylation is inhibited by PHA781089
p38 MAPK is a proline-directed kinase and hence could not be involved directly in phosphorylation of Ube2j1 at Ser184. MK2 and MK3 are well-characterized basophilic kinases activated downstream of p38α/β MAPKs. Since stress-induced Ube2j1 Ser184 phosphorylation was shown to be p38-dependent and is situated within a perfect consensus motif for MK2/MK3 phosphorylation, we further analysed the role of MK2/MK3 in p38-mediated Ube2j1 phosphorylation. To this end, we monitored the effect of the MK2/MK3/MK5 inhibitor PHA781089  on anisomycin-induced Ube2j1 phosphorylation in HEK-293T and HT29 cells. In both, PHA781089 strongly inhibited the anisomycin-induced Ube2j1 phosphorylation, whereas phosphorylation of MK2 monitored by band-shift was unaltered (Figure 3). Efficient MK2/MK3 inhibition was confirmed by the effect on phosphorylation of its bona fide substrate HSPB1 at Ser82. As expected, the p38 MAPK inhibitors SB202190 and BIRB-796 efficiently down-regulated MK2, HSP27 and Ube2j1 phosphorylation. These data suggest a role for the p38-MK2/MK3 axis in the regulation of Ube2j1 by direct MK2/MK3 phosphorylation at Ser184.
p38 MAPK inhibitors and the MK2 inhibitor PHA781089 block anisomycin-induced Ube2j1 phosphorylation
MK2 interacts with mouse Ube2j1 and directly phosphorylates it at Ser184
MK2 is a nuclear–cytoplasmic shuttling protein and Ube2j1 is an ER-membrane-anchored enzyme. The predicted topology for Ube2j1 suggests a single-pass type IV transmembrane structure  with the N-terminal catalytic domain and the intermediate region with the Ser184 site exposed to the cytosolic side of the ER, making interaction with and phosphorylation by MK2 possible (Figure 2A). To characterize further the role of MK2 in direct regulation of Ube2j1, we analysed the interaction of both proteins in GST pull-down assays. When co-expressed in HEK-293T cells, GST–MK2 specifically interacts with Myc-tagged murine Ube2ji (Figure 4A). To prove that Ube2j1 is a direct substrate of MK2, we performed in vitro kinase assays using bacterially purified murine Ube2j1. Western blot analysis confirmed similar amounts of Ube2j1-WT and Ube2j1-S184A protein (Figure 4B). Since recombinant GST–MK2 displays only a very low basal kinase activity , we used GST–p38 for in vitro activation of MK2 . The results clearly demonstrate phosphorylation of His–Ube2j1-WT by active MK2, whereas p38 alone had no effect (Figure 4D). An MK2-mediated phosphorylation was not observed for His–Ube2j1-S184A, confirming the presence of a unique MK2 phosphorylation site in murine Ube2j1. Phosphorylation of HSPB1, a bona fide substrate of MK2, is shown as a control. Autophosphorylation of p38 and p38-mediated MK2 phosphorylation is also detected (Figure 4D, lanes 3, 6 and 9). The identity of the additional phospho-protein bands in the kinase assay autoradiogram was reconfirmed by identical non-radioactive kinase assays followed by Western blotting (Supplementary Figure S1 at http://www.biochemj.org/bj/456/bj4560163add.htm).
MK2 interacts with Ube2j1 and phosphorylates Ube2j1 at Ser184
Cytosolic stress-induced, but not ER stress-induced, phosphorylation of Ube2j1 is abrogated in MK2/MK3-deficient MEFs
To conclusively prove the involvement of MK2/MK3 in stress-induced Ube2j1 phosphorylation, we used MEFs from MK2/MK3-deficient mice. MK2/MK3-deficient cell lines transduced with MK2, catalytically inactive MK2 (MK2-K79R) or empty vector (GFP) were stimulated with anisomycin, UV and sorbitol. We observed almost complete abrogation of stress-induced band-shift of Ube2j1 in MK2/MK3-DKO (double knockout) cells and only a small portion of phosphorylated Ube2j1 in MK2-K79R transduced cells (Figure 5A). This residual phosphorylation was inhibited by a p38 MAPK inhibitor and correlated with the residual kinase activity of the MK2-K79R mutant against HSPB1 (Supplementary Figure S2 at http://www.biochemj.org/bj/456/bj4560163add.htm). Interestingly, the ER stress inducers thapsigargin and tunicamycin led to weak phosphorylation of Ube2j1 independently of MK2 (Figure 5B). Thapsigargin induces an ER stress response due to altered ER calcium homoeostasis and tunicamycin induces protein misfolding by inhibition of N-linked glycosylation. We also tested whether the related kinase MK5 can act as the ER-stress-induced Ube2j1 kinase. However, anisomycin- and thapsigargin-induced phosphorylation of Ube2j1 was not significantly affected in MK5-deficient cells (Figure 5C). This supports the notion of an ER-stress-independent pathway for Ube2j1 modification activated by cytosolic stressors via activation of the p38/MK2/MK3 pathway. In addition, inhibitor studies performed in MK2/MK3-double-deficient MEFs suggest the involvement of PKC/PKD isoforms in ER-stress-induced Ube2j1 phosphorylation (Supplementary Figure S3 at http://www.biochemj.org/bj/456/bj4560163add.htm). However, compared with the MK2-dependent phosphorylation these additional phosphorylations seem minor.
Ube2j1 phosphorylation in MK-deficient cells
Cytokines and TLR induce MK2/MK3-dependent phosphorylation of Ube2j1
Since p38 MAPK is also activated by TLR ligands and pro-inflammatory cytokines, we analysed whether MK2-dependent Ube2j1 phosphorylation is downstream of these physiological stimuli. When immortalized BMDMs (bone-marrow-derived macrophages) were stimulated with the TLR4 ligand LPS, a strong phosphorylation of Ube2j1 was observed, which was completely abrogated in MK2/MK3-deficient cells (Figure 5E). Similarly, IL1α also induced phosphorylation of Ube2j1, which was completely abrogated in MK2/MK3-deficient cells (Figure 5D). Since TLR/cytokine-induced inflammatory gene expression and cytokine secretion is a major function of the p38/MK2 cascade, strong MK2-dependent phosphorylation of Ube2j1 in response to these stimuli suggests a possible role for this ER enzyme in innate immunity.
Analysis of the role of Ube2j1-Ser184 phosphorylation in ER-associated ubiquitination
Ube2j1 has been shown to play an evolutionarily conserved role in ER protein quality control . To analyse its role in ER-associated ubiquitination, we applied ubiquitination assays in HEK-293T cells [26,29]. Whereas the yeast homologue of Ube2j1 is an unstable protein undergoing stimulus-dependent ubiquitination and self-destruction , mammalian Ube2j1 is stable and not subject to ubiquitination in the experimental system chosen (Figure 6A). We analysed the role of Ube2j1 in ubiquitination of different unstable proteins known to be targets of proteasomal degradation in the cytosol or at the ER membrane. The targets chosen were the TCRα and the atypical MAPK ERK3. GFP–TCRα is a reporter for ERAD , whereas ERK3 is a soluble unstable cytosolic protein . Although expression of His6–ubiquitin induced the expected strong ubiquitination of all proteasomal targets, additional co-expression of Myc–Ube2j1 enhanced the specific ubiquitination of GFP–TCRα, but not of the cytosolic protein ERK3 (Figure 6B). To analyse the role of Ube2j1 phosphorylation on its ubiquitination activity, we then compared the effect of WT Ube2j1 and the phosphorylation mutant Ube2j1-S184A in the TCRα ubiquitination assays. Both co-expression of Ube2j1 and of Ube2j1-S184A consistently enhanced the ubiquitination of the ERAD reporter (Figures 6C and 6D). The double band for WT Myc–Ube2j1, which is not seen for Myc–Ube2j1-S184A (Figure 6C), indicates that prominent phosphorylation takes place under these experimental conditions and that it is different for WT and mutant, but has no functional consequences towards the model substrate TCRα.
Role of Ube2j1 in ERAD
Novel role for Ube2j1 in LPS-induced TNFα production
Since we did not observe significant effects of Ube2j1 phosphorylation in its function in ERAD using the model substrate TCRα, we next analysed whether Ube2j1 could be involved in one of the well-defined functions of the p38/MK2 pathway. The most prominent phenotype of MK2 and MK2/MK3 deficiency is compromised inflammatory gene expression in response to LPS challenge in vitro and in vivo [5,33]. MK2 is involved in the regulation of IL6 and TNFα production by different post-transcriptional control mechanisms . Our recent studies show specific involvement of p38/MK2 signalling in the LPS-induced translational activation of TNFα mRNA at ER-membrane-associated ribosomes . This suggested the possibility for a regulatory role of ER-associated MK2 substrates in TNFα production. For analysing this possibility, we performed siRNA-mediated down-regulation of Ube2j1 in immortalized macrophages. This approach led to almost complete inhibition of Ube2j1 expression in specific siRNA-treated cells, whereas scrambled siRNA control did not suppress Ube2j1 expression (Figure 7A). We further analysed the effect of Ube2j1 depletion on MK2-dependent LPS-induced TNFα secretion of macrophages. Consistent with the role of MK2/MK3 signalling in TNFα production, strong differences in TNFα secretion could be detected between GFP-transduced and MK2-rescued MK2/MK3-DKO cells in response to 4 h LPS treatment when transfected with control siRNA (Figure 7B). Interestingly, treatment with the three different Ube2j1 siRNAs (1–3) blocked TNF secretion in the MK2-rescued cells, and in the two cases with the best protein knockdown (1 and 3), even down to the levels of GFP-transduced DKO cells (Figure 7B). Immunoblot analysis showed an MK2-dependent increase in pro- and mature TNFα levels and, more interestingly, also a reduction in these levels in the MK2-rescued DKO cells as after knockdown of Ube2j1 (Figure 7C). In addition, down-regulation of Ube2j1 led to some increase in p38 activation in response to LPS and a slight reduction in the induction of the translational repressor of TNF, the mRNA-binding protein TTP (Figure 7D). An important feature of MK2/MK3 deficiency is the absence of significant reduction in TNFα mRNA levels while strongly reducing protein synthesis. Analysis of the kinetics of LPS-induced TNFα mRNA expression in immortalized macrophages showed only marginal effects of MK2 or Ube2j1 depletion (Figure 7E). This demonstrates the involvement of Ube2j1 in MK2-dependent translational control of TNFα synthesis.
Role for Ube2j1 in LPS-induced TNFα production
In the present study, we have demonstrated the involvement of the p38/MK2 signalling pathway in cytoplasmic stress-, LPS- and cytokine-induced phosphorylation of the ubiquitin-conjugating enzyme Ube2j1. Interestingly, the same phosphorylation site which is described in the present study (Ser184) has been identified for the ER-stress-induced PERK [PKR (dsRNA-dependent protein kinase)-like endoplasmic reticulum kinase]-dependent phosphorylation of human Ube2j1 . Hence two pathways converge in phosphorylation of Ube2j1 following cytoplasmic and ER stress. This idea is especially supported by the observations that Ube2j1 phosphorylation upon ER stress elucidated by thapsigargin, tunicamycin or DTT treatment is not inhibited by the p38 inhibitor BIRB-796 (Figure 1) and not reduced in MK2/MK3-knockout cells (Figure 5B). In addition, stimulation of phosphorylation of Ube2j1 by the ER stress is significantly weaker than that by the p38-dependent pathway, indicating different efficiencies of both pathways at least in the cell lines analysed here. It was described that the ER-stress-induced phosphorylation of human Ube2j1 did not alter the stability, subcellular localization or interaction with the E3 ubiquitin-protein isopeptide ligase parkin . This is in agreement with our observation that subcellular localization of the phosphorylation site mutants is not different from the WT (Figure 2D) and that the expression levels of mutants and WT Ube2j1 are comparable (Figures 2C and 6C). Similar to human Ube2j1, the mouse protein seems to be stable and we did not observe any ubiquitination of murine Ube2j1 in ubiquitination assays. However, a conserved lysine residue (Lys186) in close proximity to the phosphorylation site has been reported to be ubiquitinated in human cells . Hence it is still possible that specific signalling events and ubiquitin ligases induce non-degradative ubiquitination of Ube2j1. The Ser184 phospho-mimicking mutants of human Ube2j1 display decreased capacity to bind ubiquitin in vitro , indicating an inhibitory role of this phosphorylation. It is not clear how this effect will be translated into protein function in vivo. Our in vivo ubiquitination assays did not reveal a significant effect of S184A mutation on ubiquitination of the ERAD reporter TCRα (Figure 6C). A previous study could show negative influence of ectopic expression of catalytically active as well as inactive Ube2j1 on the degradation of ERAD reporters . By flow cytometry analysis we obtained similar results for GFP–TCRα degradation (results not shown), which supports the hypothesis that the levels of Ube2j1 is critical in maintaining a multi-protein ubiquitination complex, dissociation of which by overexpression could lead to substrate stabilization. It is still unclear whether phosphorylation of Ube2j1 at Ser184 may modulate its interaction with specific E3 enzymes or even with specific substrates, which might be of importance in maintaining active multi-protein complexes in the ER surface.
Although the role of stress-induced Ube2j1 phosphorylation in the ERAD pathway seems elusive, we could demonstrate the involvement of Ube2j1 in the MK2-dependent endotoxin response. This finding establishes a novel link between the ER–ubiquitin conjugation system and endotoxin-induced cytokine production. Hence we propose a non-canonical role of this ER-anchored enzyme in MK2-dependent inflammatory response and ER-associated translational activation. Our preliminary data indicate a role for Ube2j1 in the activation of TNFα translation in the microsome fractions (results not shown). We demonstrated recently the involvement of MK2 signalling in ER-associated translation of TNFα mRNA . The important MK2 downstream-signalling mediators characterized in this pathway are TTP and HuR, both antagonizing RNA-binding proteins exhibiting predominant non-ER subcellular localization. Besides Nogo-B , Ube2j1 is the only other known ER-associated substrate of MK2 and may function in providing a specific docking platform for the translational activation of TNFα mRNA at the ER. It may also be that Ube2j1 regulates the expression of another signalling intermediate in the pathway via its canonical ubiquitin-conjugating function. Since TTP is a well-characterized target of ubiquitination and proteasomal decay, we also analysed whether Ube2j1 could enhance TTP ubiquitination. Ube2j1 co-expression had no effect on TTP ubiquitination (results not shown) and TTP levels were only marginally altered on Ube2j1 knockdown (Figure 7A). The molecular mechanism, by which MK2-mediated phosphorylation of Ube2j1 and its catalytic ubiquitin-conjugating activity contribute to specific cytokine production, needs further investigation. The generation of conditional Ube2j1-deficient mice could strongly contribute to understanding the importance of these mechanisms.
Unbalanced ERAD and ER stress have been associated with inflammatory pathologies such as inflammatory bowel disease [35,36]. Interestingly, RA (rheumatoid arthritis) was also described as a ‘hyper-ERAD disease’, where the ER-resident E3 ubiquitin ligase synoviolin/HRD1 acts as a pathogenic factor, which co-operates with the E2 enzyme Ube2j1 in ERAD and triggers synovial cell overgrowth. Increased expression levels of synoviolin/HRD1 are associated with RA, whereas HRD1 deficiency leads to an enhanced resistance of mice in the collagen-induced arthritis model [37,38]. Interestingly MK2-deficient mice also show enhanced resistance to collagen-induced arthritis attributed to reduced cytokine production . Hence the pro-inflammatory p38/MK2 signalling-mediated activation of Ube2j1 could constitute a missing link between inflammation and the pathological consequences of ‘hyper-ERAD’ in RA.
cystic fibrosis transmembrane conductance regulator
ER-associated protein degradation
extracellular-signal-regulated kinase 3
human embryonic kidney
heat-shock protein 27 kDa
heat-shock protein β-1
mitogen-activated protein kinase
mouse embryonic fibroblast
T-cell receptor α
tumour necrosis factor
Manoj Manon, Christopher Tiedje and Natalia Ronkina designed and performed the experiments and analysed the data. Juri Lafera and Timo Konen performed the experiments. Alexey Kotlyarov and Matthias Gaestel discussed the project, contributed to the design of the experiments and discussed the data. Manoj Menon and Matthias Gaestel wrote the paper.
We thank Dr Ron Kopito (Stanford University) for providing the GFP–TCRα construct, Dr Robert J. Mourey (Pfizer) for providing PHA781089, Dr Andrew R. Clark (University of Birmingham) for providing anti-TTP antibodies and Dr Helmut Holtmann (Hannover Medical School) for providing IL1α.
This work was supported by Deutsche Forschungsgemeinschaft [grant numbers GA 453/10-2 and GA 453/13-1].