Coordinated regulation of Toll-like receptor and NOD2 signaling by K63-linked polyubiquitin chains

Abstract

K63 polyubiquitin chains spatially and temporally link innate immune signaling effectors such that cytokine release can be coordinated. Crohn's disease is a prototypical inflammatory disorder in which this process may be faulty as the major Crohn's disease-associated protein, NOD2 (nucleotide oligomerization domain 2), regulates the formation of K63-linked polyubiquitin chains on the I kappa kinase (IKK) scaffolding protein, NEMO (NF-kappaB essential modifier). In this work, we study these K63-linked ubiquitin networks to begin to understand the biochemical basis for the signaling cross talk between extracellular pathogen Toll-like receptors (TLRs) and intracellular pathogen NOD receptors. This work shows that TLR signaling requires the same ubiquitination event on NEMO to properly signal through NF-kappaB. This ubiquitination is partially accomplished through the E3 ubiquitin ligase TRAF6. TRAF6 is activated by NOD2, and this activation is lost with a major Crohn's disease-associated NOD2 allele, L1007insC. We further show that TRAF6 and NOD2/RIP2 share the same biochemical machinery (transforming growth factor beta-activated kinase 1 [TAK1]/TAB/Ubc13) to activate NF-kappaB, allowing TLR signaling and NOD2 signaling to synergistically augment cytokine release. These findings suggest a biochemical mechanism for the faulty cytokine balance seen in Crohn's disease.

FIG. 1.

NEMO ubiquitination at lysine 285 is required for optimal TLR4 signaling but not for optimal TNF signaling. (A) NEMO-null MEFs were infected with pBABE retrovirus expressing either vector alone, wt NEMO, or K285R NEMO. After 2 weeks of puromycin selection, clones (>1,000) were pooled and selected for an additional week. Each of these reconstituted cell lines was treated with 100 ng/ml of highly purified E. coli LPS (Invivogen) for the indicated time. Lysates were generated, and Western blots were performed using the indicated antibodies. As a control for LPS activity and as control for equivalent loading, phospho-p38 and total p38 Western blots were performed (bottom panels). (B) The vector-only NEMO-null MEFS, wt-NEMO reconstituted NEMO-null MEFs, and the K285R-reconstituted NEMO-null MEFs were treated with TNF for the indicated times. Lysates were generated, and Western blots were performed using the indicated antibodies. Again, as a loading control, a total p38 Western blot was performed (bottom panel). (C) LPS treatment causes K285 to be ubiquitinated. The wt and K285R-reconstituted NEMO-null MEF lines were either left untreated or were treated with LPS (100 ng/ml) for 45 min. NEMO IPs were performed under stringent conditions (RIPA buffer plus 0.1% SDS and 1 M NaCl). Western blots were then performed using either an antiubiquitin antibody (P4D1) or an anti-NEMO (rabbit) antibody. The asterisk indicates a cross-reactive band. No NEMO ubiquitination is present in the K285R-reconstituted cell line despite an approximate twofold increase in the amount of K285R NEMO immunoprecipitated (lower blot).

FIG. 2.

TRAF6 ubiquitinates lysine 285 of NEMO, and this ubiquitination is required for optimal TRAF6-indued NF-κB activation. (A) myc-tagged wt NEMO or myc-tagged mutant NEMOs containing the conservative Lys-Arg mutations at two major NEMO ubiquitination sites (K285R, K399R, or K285/399R) were transfected into the cell with HA-tagged ubiquitin either alone or with TRAF2 or TRAF6. Transfected NEMO was immunoprecipitated and stringently washed. Western blotting was performed on the immunoprecipitate using either an anti-HA antibody (to detect ubiquitinated species) or an anti-NEMO antibody (upper two blots). To control for TRAF2 and TRAF6 expression, total cell lysates were Western blotted using antibodies directed against tagged TRAF2 (FLAG) or tagged TRAF6 (Omni) (lower two blots). (B) NEMO-null MEFs were transiently transfected with wt NEMO, K285R NEMO, K399R NEMO, or K285/399R NEMO as well as an NF-κB reporter construct, a cytomegalovirus (CMV)-driven Renilla luciferase construct (to control for transfection efficiency), and TRAF6 (as indicated). After 2 days, luciferase assays were performed. TRAF6 caused an approximate eightfold activation of the NF-κB reporter in the presence of wt NEMO, and this decreased approximately 40% in the presence of K285R NEMO. The combined K285R/K399R NEMO mutant caused TRAF6-induced NF-κB activity to decrease almost 70%. Each experiment was performed four times with similar results each time. Error bars are standard errors of the mean. (C) NEMO-null MEFs reconstituted with either wt NEMO, K285R NEMO, K399R NEMO, or K285/399R NEMO were transfected with and NF-κB reporter constructs and a CMV-driven Renilla luciferase contruct (to normalize transfection efficiency). Twenty-four hours after transfection, cells were treated with LPS (100 ng/ml), TNF (10 ng/ml), or PC (100 ng/ml) for 6 h. Luciferase assays were then performed. In both the LPS- and the PC-treated cells, diminished NF-κB activity was seen when cells were reconstituted with K285R or K399R NEMO with a greater effect with the double mutant. TNF-stimulated NF-κB activity was greatly diminished in the K399R NEMO-reconstituted cells but not in the K285R reconstituted cells. Error bars are SEM.

FIG. 3.

NOD2 overexpression activates TRAF6 in a manner dependent on RIP2, and this activity is lost with Crohn's disease-associated alleles of NOD2. (A) HEK 293 cells were transfected with HA-tagged ubiquitin, myc-tagged TRAF6 (1 μg) or wt NOD2 (3 μg). TRAF6 was immunopurified under stringent washing conditions (RIPA buffer plus 1 M NaCl). The immunoprecipitate was Western blotted using an HA polyclonal antibody (upper blot) to detect ubiquitin-conjugated species or a myc polyclonal antibody to ensure equivalent IP of transfected TRAF6 (middle blot). To control for NOD2 transfection, total cell lysates were subjected to Western blotting using an Omni antibody (lower blot). (B) 293 cells were transfected with an siRNA targeting RIP2 or a control siRNA 2 days before transfection with HA-ubiquitin, limiting amounts of NOD2 (0.5 μg) and/or myc-TRAF6 (1 μg). TRAF6 was immunopurified under stringent conditions, and Western blotting was performed using an HA-polyclonal antibody (upper blot) or a TRAF6 polyclonal antibody (2nd blot from the top). To control for NOD2 expression and endogenous RIP2 knockdown, total cell lysates were subjected to Western blotting using an Omni or a RIP2 antibody (bottom two blots). (C) myc-TRAF6 (1 μg) was transfected into cells with either Omni-tagged wt NOD2 (3 μg) or Omni-tagged L1007insC NOD2 (3 μg) in the presence of HA-ubiquitin. TRAF6 was immunoprecipitated under stringent washing conditions and was subjected to Western blotting with either an HA polyclonal antibody to detect ubiquitinated species (upper blot) or with a myc polyclonal antibody to ensure equivalent TRAF6 IP (middle blot). To control for equivalent NOD2 and L1007insC expression, total cell lysates were Western blotted using an antibody against the Omni tag. (D) An NF-κB reporter was transfected into 293 cells with a Renilla luciferase reporter to control for transfection efficiency. Cells were transfected with either 75 ng TRAF6, 75 or 150 ng wt NOD2, and/or 75 or 150 ng L1007insC NOD2. Twenty-four hours after transfection, luciferase assays were performed. The experiment was performed four times with similar results each time. Activation (n-fold) of the NF-κB reporter (with standard errors of the mean in parentheses) are as follows: TRAF6 (150 ng), 13.83 (1.24); NOD2 (150 ng), 17.48 (3.39); L1007insC (150 ng), 11.83 (2.08); TRAF6 (150 ng) plus NOD2 (75 ng), 32.3 (6.72); TRAF6 (150 ng) plus NOD2 (150 ng), 35.1 (8.48); TRAF6 (150 ng) plus L1007insC (75 ng), 21.76 (3.23); and TRAF6 (150 ng) plus L1007insC, 19.65 (5.51).

FIG. 4.

MDP causes the activation of endogenous TRAF6. (A) THP-1 cells were stimulated with either MDP (10 μg/ml) or PC (1 μg/ml) or were left untreated. Cells were lysed in RIPA buffer and were immunopurified using antiubiquitin-Sepharose beads (Pierce Biotech.). These antiubiquitin beads bind to polyubiquitin chains containing more than four ubiquitin molecules. The immunoprecipitated polyubiquitin proteins were subjected to SDS-PAGE, and Western blotting was performed using an anti-TRAF6 antibody (upper blot). The total cell lysates was also probed for TRAF6 expression as a control for equivalent starting amounts of TRAF6 (middle blot) and for equivalent amounts of precipitated ubiquitinated proteins (lower blot). (B) MDP stimulates the autoubiquitination of endogenous TRAF6. The mouse macrophage cell line RAW264.7 was left untreated or stimulated with MDP (10 μg/ml) or purified LPS (10 ng/ml) for the time periods indicated. Cells were lysed in RIPA buffer and were immunoprecipitated using anti-TRAF6 antibody (Santa Cruz). The immunoprecipitated polyubiquitinated proteins were subjected to SDS-PAGE, and Western blotting was performed using an antiubiquitin antibody (clone P4D1). The immunoprecipitates were also probed with an anti-TRAF6 antibody to ensure that equivalent amounts of TRAF6 were immunoprecipitated. (C) To determine if NOD2 caused TRAF6 activation indirectly through a paracrine or autocrine loop, NOD2 and the Crohn's disease-associated allele, L1007insC, were transfected into HEK293 cells. One millilter of the medium from these transfectants was subjected to cytokine array analysis (Ray Biotech; bottom panel) showing high levels of IL-8, MCP-1, angiopoietin, and VEGF. The relative positions of the cytokines are shown in Fig. S3 in the supplemental material. Three milliliters of this medium was then added to cells transfected with mouse myc-tagged TRAF6 and rabbit (rbt) HA-tagged ubiquitin. After 30 min, the cells were lysed and TRAF6 was immunoprecipitated. As a positive and negative control, TRAF6 was also cotransfected with wt NOD2 and L1007insC NOD2. Western blotting again showed a strong TRAF6 ubiquitination when wt NOD2 was cotransfected; however, a much weaker TRAF6 ubiquitination was seen when the cells were exposed to medium from cells transfected with wt NOD2. The overexposed blot is shown to highlight the difference. In addition, no change in TRAF6 ubiquitination was identified between cells exposed to media from cells transfected with wt NOD2 or L1007insC NOD2. sup., supernatant.

FIG. 5.

TRAF6 is not the only E3 ubiquitin ligase responsible for NOD2/RIP2-induced NEMO ubiquitination. (A) TRAF6 expression was inhibited by siRNA (second blot from bottom) using two siRNAs designed to be unique to TRAF6, and RIP2-induced NEMO ubiquitination assays were performed as described previously (1). (B) TRAF6 or RIP2 expression was inhibited by siRNA, and NOD2-induced NF-κB activation was assayed (300 ng NOD2, 200 ng NF-κB reporter, 200 ng Renilla luciferase).

FIG. 6.

RIP2 and TRAF6 utilize the TAK1/TAB complex of proteins to induce NEMO ubiquitination and IKK activation. (A) myc-tagged wt NEMO and Omni-tagged TRAF6 were transfected into cells. wt NEMO was immunopurified via four washes with modified RIPA buffer, four washes with modified RIPA buffer containing 1 M NaCl, four washes with phosphate-buffered saline, and four washes with modified RIPA buffer. The immunoprecipitate was eluted from the beads using 1% SDS at 60°C. SDS-PAGE was performed followed by Coomassie staining (left panel). Stained bands were excised and subjected to mass spectrometry analysis. In addition to modified forms of NEMO, the TAK1/TAB complex was present in the NEMO binding complexes. MW, molecular mass markers. (B) myc-tagged wt NEMO and Omni-tagged RIP2 were transfected into cells. The experiment was performed as described for panel A. Again, in addition to modified forms of NEMO, the TAK1/TAB complex was present in the NEMO binding complex. (C) To determine TAK1's role in RIP2-induced IKK activation, HA-tagged IKKβ was transfected into 293 cells with Omni-tagged RIP2, FLAG-tagged kinase-dead TAK1 (K63A), or, as a control, Omni-tagged CYLD. IKK was immunoprecipitated, and Western blotting was performed using the indicated antibodies. Kinase-dead TAK1 reduced RIP2-induced IKK activation to essentially undetectable levels.

FIG. 7.

RIP2 utilizes the same E2 as TRAF6 to cause NEMO ubiquitination. (A) RIP2 was transfected into 293 cells with the dominant-negative E2 ligase Ubc13 with myc-tagged K399R NEMO and HA-tagged ubiquitin. NEMO was immunoprecipitated, and Western blotting was performed using the indicated antibodies. (B) MDP-induced NF-κB activation is dependent on the expression of the E2-conjugating enzyme Ubc13. The mouse macrophage cell line RAW 264.7 was infected with lentiviruses containing two separate shRNA sequences specific for Ubc13 (Open Biosystems) or green fluorescent protein (GFP) as a control, and stable lines were selected. These lines were left untreated or stimulated with MDP (10 μg/ml) for the time periods indicated, and NF-κB activation was measured by probing with a phospho-IκBα antibody (Cell Signaling). Ubc13 and β-actin levels were determined by immunoblotting with anti-Ubc13 or anti-β-actin (Sigma) antibodies. (C) To determine whether NEMO ubiquitination was also decreased in these Ubc13-knockdown cell lines, either the control shRNA line or the two Ubc13 cell lines were treated with MDP for 0, 30, 45, or 60 min. NEMO was immunoprecipitated under stringent washing conditions, and Western blotting was performed using either an anti-NEMO antibody or an antiubiquitin antibody. NEMO ubiquitination was significantly decreased in the Ubc13-knockdown cell lines. rbt, rabbit.

FIG. 8.

MDP, a NOD2 agonist, and PC, a TLR2 agonist, synergize to increase cytokine release, and this correlates with a prolonged NF-κB activation and with prolonged NEMO ubiquitination. (A) MDP and PC synergize to increase specific cytokine expression. THP-1 cells were either left untreated or were treated with 10 μg/ml MDP, 1 μg/ml PC, or 10 μg/ml MDP plus 1 μg/ml PC overnight. The medium was collected and subjected to cytokine array analysis (Ray Biotech, Inc.). The position of each of the 40 cytokines on the array is indicated in Fig. S3 in the supplemental material. In this analysis, there was limited cytokine release in either the untreated or the MDP-treated THP-1 cells (upper two panels). PC caused strong upregulation of IL-8 and RANTES and weaker upregulation of GRO (macrophage inhibitory protein β/γ [MIPβ/γ]) (lower left panel). Cells treated with MDP plus PC showed strong upregulation of IL-8, RANTES, MCP-1, IL-6, and GRO (MIPβ/γ) and weaker upregulation of TNF-α (lower right panel). (B) IL-6 ELISA performed under identical conditions shows an additive effect of PC and MDP. (C and D) Quantitative TaqMan RT-PCR was performed on THP-1 cells 2 h after stimulation with 500 ng/ml PC, 1 μg/ml MDP, or 500 ng/ml PC plus 1 μg/ml MDP. Total RNA from each sample was extracted and equalized. Within the RT-PCR cycle, each sample was then internally standardized to the 18S RNA present in that sample. Relative expression levels (with standard errors of the mean) are presented. (E) Macrophages were treated with PC (500 ng/ml), 10 μg/ml MDP, or both PC and MDP for 15, 30, or 60 min. Lysates were generated, and Western blots were performed. Phospho-IκB is shown in the lower blot. This blot was stripped and reprobed for total IκB (upper blot). (F) To correlate the synergy seen at the cytokine level with NEMO ubiquitination, macrophages were treated with 10 μg/ml MDP, 200 ng/ml PC, or both for 15, 30, or 60 min. In addition, one plate of cells was left untreated. At the indicated time, NEMO was immunopurified under stringent washing conditions and Western blotting was performed. MDP caused a slower and more prolonged NEMO ubiquitination, while PC caused a stronger, more acute NEMO ubiquitination (upper blot). When both MDP and PC were added, the NEMO ubiquitination was both more pronounced and more prolonged.

FIG. 9.

Model of the molecular synergy of TLR and NOD2 signaling converging on NEMO ubiquitination. In this model, TLRs are activated by extracellular bacteria to activate TRAF6. TRAF6 then causes ubiquitination of NEMO, which then allows nucleation of the TAB/TAK1 complex to ubiquitinated NEMO. This induced proximity of TAK1 to the IKKs allows TAK1 to phosphorylate the activation loop of the IKKs. Separately, the poorly controlled infection moves intracellularly to activate the NOD2/RIP2 complex to further activate TRAF6 as well as other, redundant E3 ligases, which activate to enhance NEMO ubiquitination and increase both the amount and duration of TAK1/TAB loading on NEMO. PGN, peptidoglycan.