|Home | About | Journals | Submit | Contact Us | Français|
Transforming growth factor (TGF)-β-activated kinase 1 (TAK1) is a key serine/threonine protein kinase that mediates signals transduced by pro-inflammatory cytokines such as transforming growth factor-β, tumour necrosis factor (TNF), interleukin-1 (IL-1) and wnt family ligands. TAK1 is found in complex with binding partners TAB1–3, phosphorylation and ubiquitination of which has been found to regulate TAK1 activity. In this study, we show that TAB1 is modified with N-acetylglucosamine (O-GlcNAc) on a single site, Ser395. With the help of a novel O-GlcNAc site-specific antibody, we demonstrate that O-GlcNAcylation of TAB1 is induced by IL-1 and osmotic stress, known inducers of the TAK1 signalling cascade. By reintroducing wild-type or an O-GlcNAc-deficient mutant TAB1 (S395A) into Tab1−/− mouse embryonic fibroblasts, we determined that O-GlcNAcylation of TAB1 is required for full TAK1 activation upon stimulation with IL-1/osmotic stress, for downstream activation of nuclear factor κB and finally production of IL-6 and TNFα. This is one of the first examples of a single O-GlcNAc site on a signalling protein modulating a key innate immunity signalling pathway.
Transforming growth factor (TGF)-β-activated kinase 1 (TAK1), also known as mitogen-activated protein kinase kinase kinase 7 (MAP3K7), is a member of the mitogen-activated protein kinase (MAPK) family (Yamaguchi et al, 1995). TAK1 has a key role in the production of tumour necrosis factor (TNF)α and other inflammatory mediators by activating several MAPKs, such as p38α MAPK, c-Jun N-terminal kinases (JNK1/JNK2), ERK1/2 and the transcription factor nuclear factor κB (NFκB) (Wang et al, 2001; Sato et al, 2005; Shim et al, 2005) via the signalling pathways shown in Figure 1A. TAK1 is essential in several cytokine-mediated innate immunity signal transduction cascades, including the TNFα, interleukin-1 (IL-1) and TGF-β pathways, as well as signalling downstream of Toll-like receptors and NOD1/2 (Shibuya et al, 1996; Ninomiya-Tsuji et al, 1999; Hasegawa et al, 2008) (Figure 1A). In these pathways, various pro-inflammatory cytokines and endotoxins trigger TAK1 activity, leading to its autophosphorylation and subsequent recruitment to the IκB kinase (IKK) complex, ultimately resulting in activation of the transcription factor NFκB (Adhikari et al, 2007) (Figure 1A). The native forms of TAK1 comprise the catalytic kinase subunit in complex with a regulatory subunit TAB1 (TAK1-binding protein 1, a pseudophosphatase; Conner et al, 2006) (Figure 1B), and either of two homologous proteins, TAB2 or TAB3 (Shibuya et al, 1996; Ishitani et al, 2003; Cheung et al, 2004). The activation of TAK1 by lipopolysaccharide (LPS) or IL-1 is triggered by the Lys63-linked poly-ubiquitination of TNF receptor-associated factor 6 (TRAF6), which binds to the C-terminal zinc-finger motifs of TAB2 and TAB3, stimulating autophosphorylation and activation of TAK1 (Wang et al, 2001).
TAK1 activity is also subject to regulation by a feedback loop in which p38α MAPK suppresses the activation of TAK1 by phosphorylation of TAB1 at Ser423 and Thr431 (Cheung et al, 2003). Disruption of the Tab1 gene in mice is embryonic lethal with several developmental phenotypes, including cardiovascular and lung dysmorphogenesis (Komatsu et al, 2002). Studies with Tab1-deficient mouse embryonic fibroblasts (Tab1−/− MEFs) suggest that TAB1 plays several roles in the regulation of the TAK1 complex, namely to recruit p38α MAPK to the TAK1 complex for the phosphorylation of TAB3, to suppress the dephosphorylation of TAB3, and to induce TAK1 catalytic activity (Mendoza et al, 2008). TAB1 is a crucial mediator in TAK1 signalling as Tab1−/− MEFs do not activate TAK1 in response to IL-1 and TNFα (Mendoza et al, 2008). MEKK3 is maintained in an inactive state by interaction with TAK1 in unstimulated cells, preventing basal NFκB signalling. Pro-inflammatory activation of TAK1 leads to disruption of MEKK3–TAK1 complexes via TAB1, allowing both TAK1 and MEKK3 to transduce biochemical signals (Di et al, 2008).
Protein glycosylation with N-acetylglucosamine (O-GlcNAcylation) is an abundant post-translational modification of serines/threonines occurring on nuclear and cytoplasmic proteins (reviewed in Love and Hanover, 2005; Hart et al, 2007). As with phosphorylation, modification by O-GlcNAc is dynamic (Kreppel et al, 1997; Comer and Hart, 2000; Zachara and Hart, 2002), giving rise to functionally distinct protein species and there is evidence to suggest that O-GlcNAc may show interplay with protein phosphorylation (Zeidan and Hart, 2010). O-GlcNAcylation is implicated in virtually all cellular processes examined, for instance gene expression (Comer and Hart, 1999), protein turnover (Hart et al, 2007), and also in regulating cellular responses to insulin (Vosseller et al, 2002; Copeland et al, 2008), cell-cycle control (Slawson et al, 2005), stress protection (Zachara et al, 2004) and calcium cycling (Clark et al, 2003). The enzymes responsible for the attachment (O-GlcNAc transferase, OGT) and removal (O-GlcNAcase, OGA) of this sugar moiety have been found in the nucleus and the cytoplasm of cells. The genes encoding these enzymatic activities have been cloned and characterized (Kreppel et al, 1997; Lubas and Hanover, 2000; Gao et al, 2001). The N-terminus of OGT contains multiple tetratricopeptide repeats thought to mediate protein–protein interactions that are critical for substrate recognition (Kreppel and Hart, 1999; Lubas and Hanover, 2000; Clark et al, 2003). Inactivation of the OGT gene in mouse cells has shown that OGT is required for embryonic stem cell viability and mouse ontogeny (Shafi et al, 2000). In addition, dysfunctional protein O-GlcNAcylation/phosphorylation appears to have a role in the pathology of type II diabetes (Hanover et al, 1999) and Alzheimer's disease (Griffith and Schmitz, 1995; Hanover et al, 1999). Recent evidence supports a central role for O-GlcNAc modification in the regulation of immune cells, particularly in the activation processes of T and B-lymphocytes and possible increased nuclear translocation and activity of nuclear factor of activated T cells and NFκB (Golks and Guerini, 2008).
Here we demonstrate that O-GlcNAcylation of a single residue (Ser395) on TAB1 modulates TAK1 activation in response to IL-1 stimulation or osmotic stress. TAB1 O-GlcNAcylation induces a substantial increase in TAK1 autophosphorylation and activation, phosphorylation of IKK, translocation of NFκB and ultimately cytokine production.
Many cytoplasmic and nuclear proteins such as transcription factors, RNA polymerase II, oncoproteins, nuclear pore proteins, viral proteins, and tumour suppressor proteins have been found to be modified by O-GlcNAc at serine and threonine residues (Hart et al, 2007; Lazarus et al, 2009). During reanalysis of a previously reported phosphosite mapping study of TAB1 (Cheung et al, 2003), a peptide with an increase in mass of 203 was observed, suggesting possible O-GlcNAc modification. To investigate whether TAB1 is an O-GlcNAc modified protein, full-length TAB1 was produced recombinantly in Escherichia coli and was O-GlcNAcylated in vitro using recombinant human O-GlcNAc transferase (hOGT). This resulted in a TAB1 protein species that was recognized by the anti-O-GlcNAc antibody CTD110.6 (Figure 2A). This was further confirmed using an alternative method for O-GlcNAc detection, involving chemoenzymatic labelling of the O-GlcNAc residue. Here, the O-GlcNAc moiety on the protein is labelled with UDP-GalNAz using a mutant galactosyltransferase GalT1 Y289L (mGalT1) with an azide derivative of UDP-GalNAc (UDP-GalNAz) as donor substrate, followed by labelling with biotin alkyne (Khidekel et al, 2003). After in vitro O-GlcNAcylation, TAB1 was subjected to mGalT1 labelling and then detected by probing with streptavidin-conjugated HRP (Figure 2B).
To confirm TAB1 as a bona fide OGT substrate with dynamic O-GlcNAc modification in vivo, we studied TAB1 O-GlcNAcylation in human embryonic kidney 293 (HEK293) cells overexpressing the IL-1 receptor (IL-1R cells) treated with/without the potent and specific OGA inhibitor GlcNAcstatin (Dorfmueller et al, 2006). TAB1 was immunoprecipitated from lysates and probed with the anti-O-GlcNAc antibody CTD110.6 (Figure 2C). TAB1 possesses a basal level of O-GlcNAcylation when grown in Dulbecco's modified Eagle's medium (DMEM) media containing 25 mM glucose (high glucose DMEM), increased by treatment with GlcNAcstatin (Figure 2C). To further confirm the O-GlcNAc signal on TAB1, TAB1 immunoprecipitated from IL-1R cells was treated with CpOGA (Rao et al, 2006), a promiscuous bacterial OGA, which removed O-GlcNAc from TAB1 without alteration in protein levels (Figure 2D). The O-GlcNAc signal on TAB1 could also be blocked by competition with free GlcNAc when the CTD110.6 antibody was pre-incubated with 500 mM N-acetylglucosamine (Figure 2D).
To identify the site(s) of TAB1 O-GlcNAcylation, recombinant O-GlcNAcylated TAB1 was analysed by mass spectrometry. In an initial experiment, O-GlcNAcylated TAB1 was trypsin digested and analysed by liquid chromatography-mass spectrometry (LC-MS). A single peptide VYPVSVPYSSAQSTSK (amino acids 387–402) was identified as carrying an additional molecular weight of 203. Two parallel approaches were then employed to identify the O-GlcNAcylation site(s) in this peptide. First, amino acids that could be putative targets for O-GlcNAcylation (part of PVS motif, similarity to target sequences of proline-directed kinases (Hart et al, 1996) were mutated to alanine. A TAB1 triple mutant (S391A/S395A/S396A) or the individual TAB1 mutants S391A, S395A and S396A were expressed as GST-tagged proteins in IL-1R cells treated with GlcNAcstatin. After 48 h, these GST fusion proteins were pulled down and blotted for O-GlcNAc. The triple mutant as well as the single S395A TAB1 mutant showed complete absence of an O-GlcNAc signal (Figure 2E), suggesting that S395 is the single O-GlcNAc site on TAB1. In a parallel approach, the TAB1 O-GlcNAc site was studied by collision-induced dissociation (CID) and electron transfer dissociation (ETD) liquid chromatography (LC) tandem mass spectrometry (MS/MS) experiments. TAB1 was O-GlcNAcylated in vitro, trypsin digested, and analysed by LC-MS/MS. The tryptic peptide VYPVSVPYSSAQSTSK (Mwcalc=1901.9 Da) containing a HexNAc (+203.1 Da) was detected after 25.5 min as [M+HexNAc+2H]2+ m/z 951.904. While the CID spectrum (Figure 2F) did not fully define the site of modification, the ETD spectrum contained the critical fragment ions to unambiguously define S395 as the only O-GlcNAc modified site on TAB1 (Figure 2G).
As shown previously (Figure 2C), levels of O-GlcNAc on TAB1 appear to be increased in the presence of the potent OGA inhibitor GlcNAcstatin. We next investigated stress-induced changes in the O-GlcNAc levels of TAB1 in wild-type (WT) MEFs. To enable these experiments, we attempted to generate a site-specific TAB1 S395 O-GlcNAc antibody. Although several attempts have been made to generate such site-specific O-GlcNAc antibodies recently, most of these still recognize a range of O-GlcNAc proteins (Teo et al, 2010). The only site-specific O-GlcNAc antibodies known to date are gThr58 on c-Myc (Kamemura et al, 2002) and gSer1011 on IRS1, gSer347 on CKII (Teo et al, 2010) and gSer400 on Tau (Yuzwa et al, 2010). Using a classical approach, we synthesized the TAB1-derived glycopeptide CVSVPYS(O-GlcNAc)SAQSTSKTS, exploiting the additional N-terminal cysteine for coupling to KLH. Serum generated from rabbits immunized with this antigen contained antibodies that were capable of recognizing O-GlcNAcylated TAB1, but not the O-GlcNAc-deficient TAB1 S395A mutant or the TAB1 triple mutant (S391A/S395A/S396A) (Figure 3A). Interestingly, while hyperglycaemic conditions resulted in an increase in global O-GlcNAc levels and also increased O-GlcNAcylation of TAB1 in MEFs, stimulation with IL-1 and NaCl, known to specifically activate the TAK1 signalling pathway (Ninomiya-Tsuji et al, 1999; Wang et al, 2001; Cheung et al, 2003; Huangfu et al, 2006), did not raise global O-GlcNAc levels to the same extent but still increased O-GlcNAcylation of WT TAB1 (Figure 3B and C). This suggested that the O-GlcNAcylation levels on TAB1 are modulated by stimuli that are known to activate TAK1 signalling, and could perhaps also affect signalling downstream of TAK1.
To investigate the effects of O-GlcNAcylation of TAB1 on activation of the TAK1 kinase and downstream signalling, we reintroduced WT TAB1 and the O-GlcNAc-deficient S395A mutant into Tab1−/− MEFs. In untransfected Tab1−/− MEFs, or the cells transfected with empty plasmid, there was no detectable IL-1α-induced TAK1 activity in agreement with earlier studies (Mendoza et al, 2008), whereas the cells transfected with WT TAB1 showed significant recovery of TAK1 autophosphorylation and activation (Supplementary Figure S1A and B). To establish an optimal reconstituted TAK1–TAB1 system, the Tab1−/− MEFs were transfected with varying amounts of TAB1 plasmid (pEBG GST–TAB1) (Supplementary Figure S1C and D). Reconstituting Tab1−/− MEFs with 5 μg of the TAB1 plasmid resulted in restoration of TAK1 activity similar to that of Tab1+/+ MEFs, with TAB1 protein levels similar to endogenous TAB1 levels (Supplementary Figure S1E and F).
To investigate the effect of TAB1 O-GlcNAcylation on TAK1 activity, reconstituted Tab1−/− MEFs were stimulated with IL-1α for 5 or 15 min, the TAK1 complexes pulled down and analysed for kinase activity and activatory autophosphorylation. When transfected with WT TAB1, IL-1α treatment increased TAB1 O-GlcNAcylation and stimulated TAK1 activity, as evidenced by phosphorylation of T187 in the activation loop of TAK1 (Figure 4A and B). Strikingly, both TAK1 kinase activity and T187 autophosphorylation were reduced with the O-GlcNAc-deficient TAB1 S395A mutant (Figure 4A and B). To rule out the possibility TAK1 is itself O-GlcNAcylated, thus regulating kinase activity, TAK1 O-GlcNAcylation was investigated. GST–TAK1 and GST–TAB1 were co-expressed in IL-1R cells maintained in 1 g/l glucose DMEM (low glucose DMEM). Cells were either treated with GlcNAcstatin or stimulated with IL-1β or NaCl before GST pulldowns and probing for O-GlcNAcylation. As expected, TAB1 was O-GlcNAcylated in the presence of higher glucose conditions or on stimulation with IL-1 or NaCl, whereas TAK1 was not, suggesting that under physiological conditions as well as in stimulated conditions TAK1 is not O-GlcNAc modified (Supplementary Figure S2). Furthermore, the TAK1–TAB1 complex remains intact in these experiments (Supplementary Figure S2).
Once activated, TAK1 translocates from the membrane to the cytosol along with TRAF6 and its binding partners, TAB1 and TAB2/3. TAK1 activation subsequently leads to activation of IKK and c-Jun NH2-terminal kinase (JNK) as well as p38α MAPK. Activated IKK phosphorylates IκB proteins, and phosphorylated IκB proteins are degraded by the ubiquitin-mediated proteasome pathway (Karin and Ben-Neriah, 2000). To investigate the effects of TAB1 O-GlcNAcylation on these downstream events, we investigated Iκβα phosphorylation. In line with reduced TAK1 activation, IL-1α-stimulated phosphorylation of Iκβα Ser32 and Ser36 was reduced by up to 50% in the Tab1−/− MEFs expressing S395A TAB1 as compared with WT TAB1 (Figure 4C and D). Phosphorylation of p38α MAPK or ERK1 was not affected with the S395A mutant (Supplementary Figure S3). Previous work has shown that TAK1 activation can also be robustly induced by osmotic stress, independent of stimulation with cytokines/LPS, leading to downstream activation of JNK1/2 (Inagaki et al, 2008). Indeed, osmotic stress (induced with 0.5 M NaCl) also induces phosphorylation of TAK1 T187, and in parallel O-GlcNAcylation of TAB1 S395, in the Tab1−/− MEFs complemented with WT TAB1 (Figure 4A and E). This effect was reduced with the O-GlcNAc-deficient TAB1 S395A mutant (Figure 4A and E), which also translated into reduced phosphorylation of JNK1/2 on Thr183 and Tyr185 (Figure 4F and G).
In the canonical NFκB activation pathway, upon phosphorylation by IKKβ, Iκβα and IκBβ, the cytoplasmic inhibitors of NFκB, are marked for ubiquitination and subsequent proteosomal degradation (DiDonato et al, 1996; Ghosh and Karin, 2002). The degradation of IκBα and IκBβ allows translocation of NFκB to the nucleus, leading to transcription of a plethora of genes including those encoding various cytokines (Li and Verma, 2002; Hayden and Ghosh, 2008). Using a reporter assay, we studied the role of O-GlcNAc on TAK1/TAB1-mediated activation of NFκB in response to IL-1α stimulation in Tab1−/− MEFs overexpressing WT or S395A TAB1 (Figure 5). While a robust level of NFκB activation was observed with WT TAB1, S395A TAB1 showed a significant reduction in NFκB activation (Figure 5A), in line with the observed reduction in phosphorylation of IκBα (Figure 4C). To evaluate whether the observed reduction in NFκB activity of the TAB1 S395A mutant resulted in effects on cytokine production, levels of IL-6 and TNFα cytokines were measured in the cell culture medium at different time points after IL-1α stimulation. IL-6 secretion was reduced by 50% after 8 h and 40% after 24 h in samples from the cells transfected with the S395A TAB1 mutant, as compared with the WT TAB1 control (Figure 5B). Similarly, TNFα production, 8 and 16 h after IL-1α stimulation, was significantly reduced with the O-GlcNAc-deficient TAB1 mutant (Figure 5C).
TAB1 possesses three characterized phosphorylation sites (S423, T431 and S438) that are modified under various stimuli and are involved in TAK1 activation (Mendoza et al, 2008) (Figure 1A and B). p38α MAPK phosphorylates S423 and T431 whereas ERK1/2 and JNK1/2 phosphorylate S438 (Mendoza et al, 2008), although the specific roles of the individual phosphorylation sites have not yet been defined. The regulatory TAB1 O-GlcNAc site described here does not correspond to a known phosphorylation site; however, it is in proximity to the phosphorylation sites and the C-terminal region of TAB1 that is required for interaction with TAK1 (Ono et al, 2001) (Figure 1B). To investigate possible effects of O-GlcNAcylation of TAB1 on phosphorylation at S423, T431 and S438, Tab1−/− MEFs transiently transfected with TAB1 (WT and S395A) were stimulated with sodium chloride in the presence or absence of GlcNAcstatin. When stimulated with NaCl, phosphorylation of S438 was observed to increase in both WT and S395A TAB1 mutant when compared with control. However, in the presence of both NaCl and GlcNAstatin, there is a further increase in phosphorylation of S438 in WT, but this further increase was not seen in the S395A mutant. This suggests that O-GlcNAc at S395 may regulate TAB1 phosphorylation at S438.
WT TAB1 showed increased phosphorylation of S438 in the presence of both GlcNAcstatin and NaCl as compared with the S395A mutant, while there was no change on S423 or T431 phosphorylation (Figure 6). S438 is phosphorylated by ERK1/2 and JNK1/2, and TAB1 has a role in stress response and activation of JNK1/2 pathway (Inagaki et al, 2008). The reduced phosphorylation at S438 on S395A TAB1 as compared with that of WT TAB1 could be related to the reduced activation of JNK1/2 as observed earlier, although its not yet clear how TAB1 O-GlcNAcylation may regulate JNK1/2 activity towards the TAB1 S438 phosphorylation site.
TAK1 is a key regulator of NFκB activity and cytokine production in response to stimulation with LPS and cytokines. Over the past decade, a large number of studies have shown that TAK1 activity is extensively regulated by phosphorylation, ubiquitination, and binding to the regulatory binding partners TAB1–3 (Sakurai et al, 2000; Wang et al, 2001; Ishitani et al, 2003; Ea et al, 2004; Singhirunnusorn et al, 2005). The data presented here show that TAB1 possesses a single, inducible O-GlcNAc site, Ser395 that is responsive to IL-1 and NaCl, known activators of the TAK1 pathway. By reintroducing WT TAB1 or an O-GlcNAc-deficient mutant in Tab1−/− MEFs, we were able to delineate the role of O-GlcNAcylation of TAB1. With IL-1 stimulation, TAB1 O-GlcNAcylation enhances Iκβα phosphorylation, which in turn regulates NFκB activation. We show here that O-GlcNAcylation of TAB1 is required for full activation of TAK1 on Thr187 and that there is a direct correlation between O-GlcNAcylation of TAB1, autophosphorylation of TAK1 Thr187, activation of NFκB and, ultimately, production of IL-6 and TNFα.
Recent studies have shown effects of hyperglycaemia on the transcriptional activity of NFκB and also on IKκβ, key regulators of innate immunity pathways. Both IKKβ and NFκB are also O-GlcNAc modified proteins (Yang et al, 2008; Kawauchi et al, 2009). Taken together with the data presented here, this suggests that the innate immune response is not only regulated by phosphorylation and ubiquitination but also by O-GlcNAcylation. Further work will be required to understand how O-GlcNAcylation of TAB1 is induced/regulated, and also to unravel the molecular details of O-GlcNAc-dependent TAK1 activation by TAB1.
GlcNAcstatin was obtained from GlycoBioChem (Dundee, UK). Mouse IL-6 and TNFα Elisa kits were purchased from Peprotech, UK. Human IL-1β was from Sigma, murine IL-1α was from Peprotech and glutathione-sepharose was from GE Healthcare. Luciferase reporter assay kit was from Promega. Click-iT™ O-GlcNAc Enzymatic Labeling System and Click-iT Biotin Glycoprotein Detection Kit were from Invitrogen. Dynabeads® Protein G was from Invitrogen.
Full-length TAB1 was cloned and inserted into pGEX6P1 for recombinant protein production in E. coli and pEBG6P for transfection in mammalian cells, as described previously (Cheung et al, 2004). Mutations for putative O-GlcNAc sites on TAB1 (S391A, S395A and S396A) were created following the Quick Change method (Stratagene) using KOD Hot start Polymerase (Novagen). Recombinant OGT was produced as described previously (Clarke et al, 2008). GST–TAK1 was obtained from the Division of Signal Transduction Therapy DSTT in Dundee. CpOGA was produced as described previously (Rao et al, 2006). For NFκB reporter assays, DNA encoding a ConA basal promoter incorporating three copies of the NFκB DNA response element (termed ‘3 × κB luciferase reporter construct') was provided by Professor Ron Hay, College of Life Sciences, University of Dundee (Rodriguez et al, 1999). The pRL-TK vector driving Renilla luciferase expression was from Promega.
The antibodies that recognize TAK1 phosphorylated at Thr187, total TAB1, TAB1 phosphorylated at Ser423 and TAB1 phosphorylated at Ser438 were used as described previously (Cheung et al, 2003). Antibodies recognizing the active phosphorylated forms of ERK1/2, JNK1/2, p38α MAPK and total ERK1/2, JNK1/2 were from Cell Signalling Technologies. ExtrAvidin®-Peroxidase was from Sigma. For immunoblotting with the phospho-specific antibodies for TAK1 and TAB1, the antibodies were incubated at 3 μg/ml in the presence of 30 μg/ml of the unphosphorylated peptide immunogen to neutralize any antibodies that recognize the unphosphorylated form of the protein. The anti-O-GlcNAc antibody CTD110.6 was purchased from Abcam. Secondary antibodies conjugated to horseradish peroxidase were from Pierce.
The O-GlcNAc peptide CVSVPYS(O-GlcNAc)SAQSTSKTS, corresponding to residues 389–403 of TAB1, was synthesized on a Liberty microwave-assisted peptide synthesizer (CEM) using MBHA Rink-amide low load resin (Novabiochem) with standard protocols of Fmoc SPS chemistry. The QS dipeptide was introduced as pseudoproline to suppress formation of truncated sequences detected in pilot experiments. 4Ac-GlcNAcSerFmoc was synthesized in-house following a published procedure (Saha and Schmidt, 1997). After high peformance liquid chromatograpy (HPLC) purification of the peptide, it was conjugated to keyhole limpet haemocyanin and injected into rabbits. Antibodies from the serum were first precipitated with ammonium sulphate followed by a one-step purification by passing the resuspended antibody over a non-GlcNAcylated peptide column. The flowthrough from the column was collected and used for immunoblotting.
In vitro o-GlcNAcylation of TAB1 (1 μM) was performed in 20 μl assay volumes containing 100 nM of OGT, reaction buffer (50 mM Tris–HCl, pH 7.5, 1 mM DTT, 12.5 mM MgCl2), and 1 mM UDP-GlcNAc. The reactions were incubated for 90 min at 37°C, stopped by adding 4 × SDS–PAGE sample buffer, resolved on SDS–PAGE, transferred to PVDF and probed with appropriate antibodies.
GST–TAB1 was bound to glutathione-sepharose beads and was O-GlcNAcylated in vitro. The beads were washed with 10 mM HEPES (pH 7.9) and resuspended in reaction buffer (1% SDS, 20 mM HEPES) then 2 μl of GalT1 Y289L (Invitrogen) and 2 μl of 0.5 mM UDP-GalNAz (Invitrogen) were added in a final reaction volume of 50 μl. The reaction was performed overnight at 4°C. The beads were washed twice with reaction buffer to remove excess UDP-GalNAz. The samples were then reacted with biotin alkyne (Invitrogen) according to the manufacturer's instructions. The proteins were resolved by SDS–PAGE and transferred to PVDF. The blot was incubated with ExtrAvidin-Peroxidase and biotinylated TAB1 was visualized by ECL reaction.
For site mapping analysis of digested TAB1 protein, an Ultimate 3000RSLC nano-HPLC system (Dionex) with a 3D high capacity ion trap mass spectrometer with ETD capability (amaZon ETD; Bruker Daltonics) were used to perform HPLC electrospray MS/MS (ESI-MS/MS). Digested TAB1 samples were reconstituted in 0.1% formic acid, injected and concentrated onto a Dionex PepMap C18 nano-trap column, after a wash step with (2% acetonitrile, 0.1% formic acid (v/v)) peptides were resolved by a Dionex Acclaim PepMap C18 reverse phase column (75 μm i.d. × 15 cm × 2 μm) over a 25-min linear gradient from 4% to 50% buffer B (80% acetonitrile and 0.08% formic acid in Milli-Q water (v/v)), followed by another 2 min to 90% buffer B. The column was then washed by holding the 90% buffer B for 10 min before returning to initial conditions of 96% buffer A (0.1% formic acid in Milli-Q water (v/v)). A typical tandem mass spectrometric (MS/MS) cycle (Alternating CID/ETD) in amaZon ETD happens in the following order: (A) 1 MS full-range scan and precursor ion selection. (B) Accumulation of precursor ion (~10 ms) and fragmentation by CID. (C) (CID-MS/MS spectrum recorded) accumulation of the same precursor ion (~5 ms), which is allowed to mix and react with fluoranthene (50–100 ms; variable) for ETD fragmentation, and ETD-MS/MS spectrum acquired. Steps (B) and (C) are repeated automatically for each precursor ion. In this study, precursor ion selection was set up to five ions per cycle, excluding singly charged ions, with a dynamic exclusion time of 0.5 min for both CID and ETD. Helium gas was used as collision gas (60–200%), collision energy sweep with amplitude 1.0 parameters was used for CID fragmentation. For ETD experiments, the maximum output of ETD reagent ion (202 m/z) was achieved with the following nCI source tuned parameters: reagent ion charge control target (ICC) 500 000, maximum emission current 4 μA and ionization energy of 75 eV, reactant remove cutoff 210 m/z without supplemental energy activation.
Raw MS data were processed using software packages BioTool 3.2 SR1 and DataAnalysis 4.0 (Bruker Daltonic GMBH). In parallel, two database searches were performed for CID and ETD using Mascot v2.3 (Matrix Science Ltd), database used IPI-human 20110731 (91 522 sequences; 36 630 302 residues) with the following Mascot MS/MS ion search parameters: peptide charges considered 2+, 3+ and 4+ ions, peptide tolerance and fragment tolerance was set to ±0.5 Da, # of 13C=1, two missed cleavages allowed, trypsin as proteolysis enzyme, ESI-TRAP (for CID) and ETD-TRAP (for ETD) for instrument type. Carbamidomethyl (C) was used as fixed modification, where Deamidated (NQ), Oxidation (M), Phospho (ST) and HexNAc (ST) (+203.0794 Da) were set as variable modifications.
IL-1R cells HEK293 (cells stably expressing the IL-1 receptor) and immortalized Tab1-deficient MEFs (Tab1−/− MEFs) were provided by Professor Philip Cohen, MRC Unit, University of Dundee. Cells were cultured in DMEM with 1 or 4.5 mg/ml glucose, containing 10% (v/v) heat inactivated fetal calf serum (FCS), and 2 mM L-glutamine. Prior to stimulation with human IL-1β in IL-1R cells or murine IL-1α in mouse cells, the medium was removed and replaced with DMEM from which FCS had been omitted. IL-1R cells were serum deprived for 16 h and MEFs for 6 h. GlcNAcstatin (1 μM) was added to cells during serum starvation, if required. For osmotic shock, cells were treated with either 0.25 or 0.5 M of NaCl by adding it into DMEM for 15 min before harvesting the cells.
Cells were lysed in ice-cold lysis buffer (50 mM Tris–HCl pH 7.5, 1 mM sodium orthovanadate 1 mM EDTA, 10 mM sodium β-glycerophosphate, 1 mM EGTA, 50 mM sodium fluoride, 5 mM sodium pyrophosphate, 1% (v/v) Triton X-100, 0.27 M sucrose, 0.1% (v/v) 2-mercaptoethanol, 0.1 mM PMSF, 1 mM benzamidine and 5 μM leupeptin). Lysates were centrifuged at 13 000 g for 15 min at 4°C and the supernatants were used immediately or snap frozen in liquid nitrogen and stored in aliquots at −80°C until use. Protein concentrations were determined using the Bradford assay.
To immunoprecipitate endogenous TAB1, 1 mg of cell lysate was incubated for 2 h at 4°C with 10 μg of anti-TAB1 antibody coupled with 10 μl of Dynabeads Protein G. The immunoprecipitates were washed twice with 1 ml of lysis buffer containing 0.25 M NaCl, followed by two washes with 1 ml of 50 mM Tris/HCl, pH 7.5, 50 mM NaCl and 0.1 % (v/v) 2-mercaptoethanol and subjected to SDS–PAGE followed by western blotting.
For OGA treatment on immunoprecipitated O-GlcNAcylated TAB1, 3 mg of lysate was used for immunoprecipitation. The samples were divided in three equal volumes. One set of samples was treated with CpOGA (1 μM) for 30 min at room temperature, while the other two sets were left untreated at room temperature. The samples were subjected to SDS–PAGE and western blotting with CTD110.6 antibody. The third set of samples was incubated with CTD110.6 antibody, which was pre-incubated with 500 mM N-acetylglucosamine for 1 h.
IL-1R cells were transfected at 40–50% confluence using poly- ethyleneimine using DNA encoding GST (glutathione transferase)–TAB1, whereas MEFs were transfected at a density of 3 × 106 cells, with the Amaxa MEF2 kit according to the manufacturer's instructions.
For the measurement of NFκB-dependent luciferase gene expression, Tab1−/− MEFs (3 × 106) were co-transfected with either 3 μg of WT TAB1, S395A TAB1 or empty pEBG6 plasmid; 0.5 μg of DNA encoding the 3 × κB luciferase reporter construct; and 0.5 μg pRL-TK and plated 3 × 105 cells per well. After 24 h, the cells were stimulated with 10 ng/ml of IL-1α for 24 h and then the cells were lysed in Passive Lysis Buffer (Promega). The luciferase activity was then measured using a Dual-Luciferase Reporter Assay System (Promega) as per the manufacturer's instructions. Firefly luciferase activity was normalized by Renilla luciferase activity for each transfection.
TAK1 present in TAB1 immunoprecipitates was assayed by its ability to activate MKK6, as judged by the activation of SAPK2α/p38α. The active SAPK2α/p38α generated in this first stage of assay was then quantitated in a second assay by measuring phosphorylation of myelin basic protein (Cheung et al, 2003). The TAK1 complexes were pulled down from 1 mg of cell lysate obtained from TAB1 reconstituted MEFs. The cell lysates were incubated for 2 h at 4°C with 20 μl of glutathione-sepharose beads per sample. The beads were washed twice with 1 ml of lysis buffer containing 0.25 M NaCl, followed by two washes with high salt wash buffer (1 ml of 50 mM Tris–HCl pH 7.5, 0.5 M NaCl and 0.1% (v/v) 2-mercaptoethanol). In all, 25% of the TAK1 complex bound to the beads was used to measure TAK1 activity. The remaining 75% of the TAK1 complex was taken for immunoblotting using appropriate antibodies as described earlier.
For measuring TNFα and IL-6 secretion into the medium, 3 × 106 MEFs were transfected with the WT or S395A TAB1 and were seeded in 24-well plates at density of 3 × 105 cells/well. At 24 h after transfection, the cells were stimulated with IL-1α for different lengths of time. The media were collected and after brief centrifugation, 100 μl of clear media was used for cytokine ELISA as per the protocol from Peprotech.
We thank Sharon Shepherd and Ramon Hurtado-Guerrero for the supply of protein reagents. This work was funded by a Wellcome Trust Senior Research Fellowship to DVA.
Author contributions: DMFA and SP designed the experiments. SP performed the experiments and analysed the data. VSB synthesized the peptides. OA and DGC performed mass spectrometry. AI cloned the constructs. SP and DMFA wrote the manuscript.
The authors declare that they have no conflict of interest.