p62/SQSTM1 Enhances NOD2-Mediated Signaling and Cytokine Production through Stabilizing NOD2 Oligomerization

NOD2 is a cytosolic pattern-recognition receptor that senses muramyl dipeptide of peptidoglycan that constitutes the bacterial cell wall, and plays an important role in maintaining immunological homeostasis in the intestine. To date, multiple molecules have shown to be involved in regulating NOD2 signaling cascades. p62 (sequestosome-1; SQSTM1) is a multifaceted scaffolding protein involved in trafficking molecules to autophagy, and regulating signal cascades activated by Toll-like receptors, inflammasomes and several cytokine receptors. Here, we show that p62 positively regulates NOD2-induced NF-κB activation and p38 MAPK, and subsequent production of cytokines IL-1β and TNF-α. p62 associated with the nucleotide binding domain of NOD2 through a bi-directional interaction mediated by either TRAF6-binding or ubiquitin-associated domains. NOD2 formed a large complex with p62 in an electron-dense area of the cytoplasm, which increased its signaling cascade likely through preventing its degradation. This study for the first time demonstrates a novel role of p62 in enhancing NOD2 signaling effects.

Autophagy was originally described as an energy homeostasis process that degrades and recycles damaged molecules and organelles through the formation of double-membrane vesicles. Recent studies have further revealed its essential roles in innate immune responses including entrapment/killing of intracellular microorganisms, antigen presentation, and cytokine production [10,24,25]. NOD2 has been shown to induce autophagy through a RIP2-dependent manner at least in myeloid and epithelial cells [9,24,26]. A portion of NOD2 was also shown to localize at the plasma membrane and recruit the autophagy processing molecule ATG16L1 at the site of bacterial entry, which was a RIP2independent process [10]. Autophagy can also function as a negative feedback process of inflammatory responses, since it was shown to suppress signaling events induced by Toll-like receptors and NLRP3 [25]. A study also suggested that human peripheral blood mononuclear cells with defects in autophagy resulting from a mutation in ATG16L1 produced more inflammatory cytokines at the mRNA level when induced by MDP [27], suggesting that autophagy is also involved in NOD2 signaling.
Of the more than 30 different proteins involved in autophagy, p62 (also known as sequestosome-1), is an adaptor protein which sequesters poly-ubiquitinated proteins [28] and Salmonella-containing vacuoles [29] to autophagy through interacting with microtubule-associated protein 1 light chain 3 (LC3). In addition to these catabolic roles, p62 has also been shown to regulate various signaling events. For example, it up-regulates signaling events initiated by receptors activated by tumor necrosis factor (TNF)-a, IL-1, nerve growth factor, and RANK-L (receptor activator of NF-kB-ligand) through scaffolding for TRAF6 and atypical protein kinase C with these receptors [30]. In contrast, p62 suppresses Toll-like receptor signaling cascades by inducing MyD88-aggregation and down-regulation of MyD88-TRAF6 complex formation [31]. In light of the multifaceted roles of p62 in autophagy and signal transduction, we examined the role of p62 in NOD2 signaling. This study found that p62 interacted with NOD2 and enhanced its signaling response toward NF-kB activation, and TNF-a and IL-1b production, through stabilizing NOD2 signaling complexes.

Cells and Cell culture
The human monocytic cell line THP-1 (ATCCH TIB-202) was maintained in complete RPMI 1640 medium containing 8% heatinactivated fetal bovine serum (FBS, Sigma Aldrich), 1 mM MEM non-essential amino acids solution, 1 mM sodium pyruvate, and antibiotics (mixture of 100 U/mL penicillin G sodium, and 100 mg/mL streptomycin sulfate). HEK293T and RAW 264.7 macrophages (kindly provided by Dr. J. Han, The Scripps Research Inst., La Jolla, CA) were cultured in complete DMEM medium containing 8% FBS, and the same reagents as in complete RPMI 1640. Cells were grown at 37 uC in a humidified atmosphere containing 5% CO 2 .
Retrovirus production and cell infection were performed as previously described [32]. Briefly, pLNCX-NOD2 recombinant retroviruses were generated in Phoenix Amphotropic producer cells using the calcium phosphate method of transfection. Viruses were produced at 32uC, and virus-containing medium was collected 24 h post-transfection and filtered through a 0.45 mm filter. HEK293T cells were plated in six-well plates at a density of 5610 5 cells/well. One round of retroviral infection was performed by replacing medium with 2 mL of pLNCX-NOD2 virus (containing 4 mg of Polybrene per mL), followed by centrifugation of the six-well plates at 2,000 RPM for 40 min at 32uC. On the third day, culture media were replaced with selection media containing 10 mg/mL of Puromycin (Calbiochem).

Co-immunoprecipitation for poly-ubiquitination and protein-protein interactions
Twenty-four hours after transfection, cells were washed twice in PBS, overlaid with RIPA lysis buffer (50 mM TrisCl, 150 mM NaCl, 1% Igepal CA-630 (NP-40), 0.5% Sodium Deoxycholate, and 0.1% SDS) and harvested by scraping. Cells were homogenized by pipetting on ice, centrifuged at 12,500 g for 15 min at 4uC, and supernatants were transferred to new 1.5 mL tubes. Ubiquitin-conjugated lysates or total lysates were combined with either anti-Myc or anti-GFP antibodies (Clontech) for 1, 2, 4 h or overnight at 4uC. Twenty microliters of Protein G Sepharose TM Fast Flow Beads (Sigma-Aldrich) were added to lysates and incubated for an additional 1 h. Immunoabsorbents were recovered by centrifugation at 10,000 g for 30 sec and washed five times by resuspension and centrifugation in the same lysis buffer. The immunocomplex was eluted with 4X SDS loading buffer (125 mM Tris-HCl [pH 6.8], 4% SDS, 50% Glycerol, 0.08% Bromophenol Blue, and 5% b-mercaptoethanol).

Confocal microscopy
HEK293T cells were seeded at 0.5610 6 cells overnight on coverslips and transfection was performed with GFP-p62, in combination with DsRed-NOD2, DsRed-NBD, DsRed-LRR, or GFP-LC3 and DsRed-NOD2 for 24 h in complete DMEM media at 37uC in a 5% CO 2 incubator.
The cells were fixed with 4.0% paraformaldehyde in PBS (pH 7.4) at RT for 5 min and rinsed twice with PBS at RT for 5 min. Confocal images were obtained using a Zeiss LSM510 META confocal microscope and analyzed with ZEN software.

ELISA and TNF-a bioassay
TNF-a levels in cell culture supernatants were measured by ELISA or bioassay. ELISA was followed by manufacturer's instruction (eBioscience). Bioassay for TNF-a concentrations in cell culture supernatants were measured as previously described [34]. Briefly, murine L929 fibroblasts were seeded to each well of a 96-well tissue culture (7610 4 cells/well) for 4 h. Cell culture media were then replaced with media containing cyclohexamide (0.3 mg/mL) and the supernatants. Known concentrations of recombinant human TNF-a (eBioscience, 1 mg/mL) were prepared by serial dilution and loaded onto the L929 cells for references. The culture plate was incubated at 37 uC overnight, carefully rinsed with 1X PBS, and then exposed to crystal violet (Sigma-Aldrich) solution for 5 min at RT. Cells were rinsed in PBS twice and crystals were solubilized in 50% acetic acid (Caledon) for 30 min. Optical density was measured using an ELISA plate reader (Bio-Rad) at a wavelength of 570 nm. TNF-a in cell culture supernatants was calculated based on the optical density and known TNF-a references.

Size-exclusion chromatography
Cells were lysed in RIPA buffer containing 1% Igepal CA-630 (NP-40) and the complete protease inhibitor cocktail (Roche Diagnostics). Lysates were spun at 12,500 RPM for 15 min and supernatants were loaded onto Superdex TM 200 10/300 GL (GE Healthcare, Life Science). Superdex column was equilibrated with 50 mM Tris-HCL (pH 8.0) containing 150 mM NaCl and calibrated with standard proteins (Sigma) containing blue Dextran (2,000 kDa), b-amylase (200 kDa) and bovine serum albumin (66 kDa). Fractions (5 ml) were collected after injecting on to the column and each fraction was analyzed by Western blots for NOD2, p62 and p38.

Immunogold staining
HEK293T cells were transfected with HA-tagged p62 and GFP-tagged NOD2 using PolyJet TM (SignaGen Laboratories) following manufacturer's instructions. Twenty-four hours posttransfection, cells were harvested in PBS with 1 mM EDTA, and fixed with 3% paraformaldehyde and 0.025% glutaraldehyde in a 0.1 M cacodylate buffer (CAC; pH 7.4) for 2 h. Following fixing, cells were washed in CAC and incubated overnight in the buffer.
Cells were enrobed in 5% noble agar the following day, and agar pieces containing cells were dehydrated in 50%, 70%, 85%, 90% and 26100% ethanol, respectively, for 10 min each. Dehydrated samples were then infiltrated in a 1:1 mixture of LR White resin: 100% ethanol for 30 min, followed by two infiltrations with pure LR White resin for 90 min and overnight, respectively. Samples were then placed in a gelatine capsule and incubated in an oven at 50C for 24 h to solidify. The hardened capsule was then cut using an ultramicrotome (Ultracut) into 70 nm thin sections, which were placed onto nickel grids (EM Sciences). Grids were blocked in a 0.2 mm-filtered PBS-BSA buffer (10.4 mM Na 2 HPO 4 , 3.2 mM KH 2 PO 4 , 20 mM NaN 3 , 150 mM NaCl, 1% BSA, pH 7.4) overnight. Blocked grids were then incubated with anti-HA (rabbit; Abcam) and anti-GFP (mouse; Clontech) antibodies at 0.01 mg/mL (1:100 dilution) and 0.1 mg/mL (1:10 dilution), respectively, for 2 h at room temperature. Grids were then washed 5 times in PBS-BSA for 5 min each, and probed with anti-rabbit 18 nm colloidal gold (Jackson ImmunoResearch) and anti-mouse 10 nm colloidal gold (Invitrogen) at dilutions of 1:20 for 1 h. Grids were washed 5 additional times in PBS-BSA for 5 min each, followed by an overnight wash in PBS-BSA. The following day, grids were washed 3 times in 0.2 mm-filtered water before being stained in double 0.2 mm-filtered 2% uranyl acetate for 7 min. After staining, grids were washed 3 times for 1 min each in water and left to dry. Grids were viewed using a Phillips CM10 transmission electron microscope at 60 kV.

p62 enhances NOD2 signaling in HEK293T cells
We first examined if p62 is involved in NOD2 signaling using an NF-kB-driven luciferase reporter system in HEK293T cells. Cells were treated with p62 or scrambled small interference RNAs (si-p62; Fig. 1A, upper panel), and transiently transfected with Myctagged NOD2 and Igk luciferase reporter plasmids. As shown previously [35,36], over-expression of NOD2 alone caused about a 5-fold increase in NF-kB reporter activity, which was further increased by the NOD2 ligand N-glycolyl muramyl dipeptide (gMDP) (Fig. 1A, lower panel). However, cells knocked down in p62 failed to respond to NOD2 over-expression alone or gMDP treatments. To further examine the role of p62 in NOD2 signaling, we examined its downstream signaling events: polyubiquitination of RIP2 [8], TRAF6 [37] and p38 mitogenactivated protein kinase phosphorylation [38]. For RIP2 and TRAF6 analysis, cells stably transfected with NOD2 were transiently transfected with HA-tagged ubiquitin and Myc-tagged RIP2 (Myc-RIP2) or TRAF6, together with scrambled siRNAs or si-p62. Cells transfected with scrambled siRNAs showed apparent ubiquitinations of RIP2 (Fig. 1B) and TRAF6 (1C) or ubiquitinated proteins co-immunoprecipitated with RIP2 and TRAF6 in response to gMDP; however, cells treated with si-p62 showed no or much less such ubiquitinated proteins. Similarly, treatments of gMDP in cells stably transfected with NOD2 also gradually induced p38 MAPK tyrosine phosphorylation, which was inhibited in cells knocked down in p62 (Fig. 1D). Collectively, these data suggest that p62 enhances NOD2 signaling cascades.

NOD2 interacts with p62 through the NBD domain of NOD2 and UBA or TRAF6 domain of p62
To examine if p62 interacts with NOD2, co-immunoprecipitation analyses were performed in HEK293T cells transfected with Myc-NOD2 and GFP-conjugated p62 (GFP-p62) plasmids. As shown in Fig. 2A-B, immunoprecipitations using anti-GFP or anti-Myc monoclonal antibodies were able to co-immunoprecipitate with Myc-NOD2 or GFP-p62, respectively. Also, cells treated with gMDP (5 mg/mL) further enhanced the Myc-NOD2 and GFP-p62 interaction (Fig. 2C), suggesting that p62 interacted with NOD2 better when NOD2 was activated.
p62 contains at least four distinct motifs (Fig. 3A, right panel): Phox and Bem 1p (PB1), zinc finger (ZZ), TRAF6-binding (TRAF6) and ubiquitin-associated (UBA) domains [43]. The Nterminal PB1 domain is known to accommodate p62 homodimerization as well as hetero-dimerization with various signaling molecules including PKCj/i/l, MEKK3, MEK5 and ERK1. ZZ and TRAF6 domains were shown to be involved in the interaction with RIP1 and TRAF6, respectively. The C-terminal UBA domain preferentially binds to K63-linked poly-ubiquitin chains [44] and the LC3-interacting region (LIR), located between UBA and TRAF6 domains, interacts with LC3. Therefore, p62 is expected to function as an autophagy cargo molecule that targets aggregated proteins, cellular organelles and microbes for degradation [45]. We examined how p62 interacted with NOD2 using a similar co-immunoprecipitation approach with different p62 mutants and LRR-deleted NOD2 (DLRR) to maximize the interaction. Interestingly, both GFP-TRAF6 and GFP-UBA, but not GFP-PB1, domains were co-immunoprecipitated with Myc-DLRR (Fig. 3C). Consistent with these results, DLRR was also coimmunoprecipitated with TRAF6 binding domain-deleted (DTRAF6) or UBA domain-deleted mutants of p62 (Supplemental Fig. S2). Collectively, these results suggest that the NBD domain of NOD2 interacted with either the TRAF6-binding or UBA domain of p62. We found that NOD2 undergoes both K48-and K63-mediated polyubiquitinations (data not shown), which likely contributes interaction between UBA domain of p62 and NOD2. Further detailed experiments are required to elucidate whether p62 binding to NOD2 through UBA domain requires ubiquitination of NOD2, and how p62 interacting through TRAF6-binding domain and UBA domain affect NOD2 signaling.

NOD2 is co-localized with p62 in the cytoplasm as a granulated form
Previous studies demonstrated that NOD2 could be localized in both the plasma membrane and cytosol as speckles [10,15,46]. Indeed, DsRed-NOD2 was localized in both intracellular compartments in punctate form and the plasma membrane in HEK293T cells (Fig. 4A, left panel). Furthermore, cytosolic DsRed-NOD2 positive speckles, but not the plasma membrane associated, were co-localized with p62 (right panel). In line with co-immunoprecipitation results (Fig. 3B), DsRed-conjuated with the NBD of NOD2 also prominently co-localized with GFP-p62 (Fig. 4B, upper lane); whereas, no such co-localization was detected in NOD2 only containing LRR motif (lower lane). To further examine co-localization of these molecules, cells overexpressing both GFP-p62 and HA-tagged full length NOD2 were viewed through EM after immunogold labeling against GFP and HA. As shown in Fig. 4C, aggregated patterns of both GFP-NOD2 (10 nm gold particles) and HA-p62 (18 nm gold particles) were detected in electron-dense areas. However, it could not be determined whether the electron-dense NOD2 and p62-positive areas were autophagosomes, because our immunogold EM staining could not clearly resolve membrane structures. Since p62 associates with autophagosomes through interacting phosphatidylethanolamine conjugated microtubule-associated protein 1 light-chain 3 (LC3) during autophagy [47,48,49], we examined whether LC3 co-localized with the cytosolic NOD2-positive granules. LC3-GFP was detected throughout the cytoplasm and as granular forms close to the nucleus. However, DsRed-NOD2positive granules were localized in distinct locations from those of LC3-GFP (Fig. 4D). These results suggest that cytosolic NOD2-p62 aggregates were not autophagosomes.
Previously, p62 was shown to aggregate with several signaling molecules to enhance their signaling effects. For example, p62 aggregates with poly-ubiquitinated caspase-8 that leads to full activation and processing of the enzyme [39]. p62 is also involved in the formation of large structures known as aggresome-like induced structures (ALIS) [50]. Unlike aggresomes, which are rapidly degraded through a proteasomal route, ALIS are devoid of proteasomes and transient in nature, and recruit ubiquitination enzymes including the ubiquitin-activating enzyme E1, the ubiquitin-conjugating enzyme E2 and the ubiquitin ligase E3 [51,52]. ALIS were shown to be induced by Toll-like receptors or various stresses [53] and ubiquitinated proteins associated with ALIS were shown to have a much longer half-life than those present in the cytosol [52]. Several features such as the granular aggregation of NOD2 with p62 in non-autophagic vacuoles may point that the electron-dense organelles are ALIS or ALIS-related structures. However, further studies are required to determine if NOD2 is indeed localized in bona fide ALIS.

p62 stabilizes NOD2 oligomerization
Considering the multiple roles of p62 in protein modifications (through recruiting E1/E2/E3 proteins or other signaling molecules) and stabilization of proteins through forming ALIS, p62 could have enhanced NOD2 signaling through recruiting TRAF6 to nodosomes or stabilizing NOD2 oligomers. However, knocking down p62 had little effects on recruiting TRAF6 to NOD2 (Supplemental Fig. S1). Thus, we examined whether p62 was involved in stabilization of NOD2 at protein levels. To this end, Figure 5. p62 stabilizes gMDP-induced NOD2 oligomers. A. HEK293T cells were stably transfected with pLNCX-NOD2 as described in ''Methods''. These cells were treated with scramble (si-Scramble) or p62 targeting (si-p62) small interference RNAs for 24 h. Cells were then treated with the translation inhibitor cyclohexamide (CHX, 100 mg/ml) and gMDP (5 mg/ml) for the time indicated, and immunoblots against NOD2 were performed. Intensities of NOD2 bands in comparison with p38 bands (loading control) were expressed as 100% for control samples (right panel). The ImageJ (NIH) program was used for densitometry analysis and data were expressed as mean 6 S.D. (n$4). *p,0.05 (Student t-test). B. HEK293T cells were transiently transfected with Myc-NOD2, HA-NOD2, and scramble (si-Scramble) or p62-targeting (si-Scramble) small interference RNAs. Myc-NOD2 was immunoprecipitated with anti-Myc antibodies and immunoblots were performed against HA or Myc. Immunoblots for total lysates against HA and p62 were performed for HA-NOD2 and endogenous p62 inputs (3 rd and 4 th lanes, respectively). C. HEK293T cells were transfected with Myc-NOD2 at 16 h post-transfection with scrambled siRNA or p62-siRNA. After 24 h, cells were further cultured without or with gMDP (5 mg/ml) for 4 h and cell extracts were loaded onto the Superdex TM 200 column. Fractions were analyzed by immunoblot using Myc antibody for estimation of Myc-NOD2 oligomerization (upper panel). Immunoblot for p38 was used as a control. Myc-NOD2 and p38 immunoreactivities were analyzed using NIH ImageJ program (bottom panel). doi:10.1371/journal.pone.0057138.g005 p62 Enhances NOD2 Signaling PLOS ONE | www.plosone.org HEK293T cells stably expressing NOD2 were knocked down in p62 by si-p62 and NOD2 protein levels were examined in the presence of the broad translation inhibitor cyclohexamide (CHX; 100 mg/mL) and gMDP (5 mg/ml). In cells treated with si-Scrambled, both NOD2 and p62, but not p38, gradually degraded over 12 hours (Fig. 5A). However, in cells knocked down in p62, NOD2 degradation was significantly faster. These results are in line with a recent study shown that NOD2 undergoes the 26S proteasome-mediated degradation, which negatively regulates its signaling [17]. Thus, it is possible that p62 leads p62-NOD2 complexes to avoid their degradation by proteasomes. Indeed, the 26S proteasome inhibitor MG132 prevented fast degradation of NOD2 in gMDP-treated p62 knock-down cells (Supplemental Fig.  S3). Next, we examined whether p62 also enhanced oligomerization of NOD2. Cells were transiently transfected with both HAand Myc-conjugated NOD2 expression vectors together with si-Scramble or si-p62, and oligomerization of NOD2 was examined through co-immunoprecipitation. As shown in Fig. 5B, HA-NOD2 was co-immunoprecipitated by anti-Myc antibody, which was diminished by si-p62, suggesting that dimerization or multimerization of NOD2 is formed through a p62-dependent manner. In addition, size-exclusion gel filtration chromatography was used to examine the degree of NOD2 complex formation. Myc-NOD2 was eluted between fraction number 5 and 7 (Fig. 5C, left panels), suggesting that Myc-NOD2 complex was less than 2000 kDa but higher than 200 kDa in size (Supplemental Fig. S4-A). In gMDPtreated cells, Myc-NOD2 was eluted in earlier fractions (number 3-7), suggesting a higher degree of complex formation with about 2000 kDa in size. However, the gMDP-induced Myc-NOD2 complex was not detected in cells treated with si-p62 (Fig. 5C, right panels). Elution of p38 was used as a control and showed no differences in elution patterns between g-MDP or si-p62 treated or non-treated cells. p62 was detected in fractions between 6 and 8, which partially overlapped with those of Myc-NOD2 (Supplemental Fig. S4-B). In the presence of gMDP, p62 was also formed a higher degree of complex which was eluted in fractions 4-9, Figure 6. p62 enhances pro-IL-1b expression and TNF-a production in macrophages. A. RAW 264.7 cells were transfected with scrambled (si-Scramble) or p62 targeting (si-p62) small interference RNAs using Lipofectamine TM 2000. Twenty four hours post-transfection, cells were treated with a low dose of LPS (50 ng/mL) for 4 h, rinsed with complete media twice, and then incubated with gMDP (5 mg/mL) for another 4 h. Expression of pro-IL-1b was detected using an antibody against IL-1b and p38 was used as a loading control. Densitometric analysis of blots was done using ImageJ (NIH). Data are expressed as mean 6 S.D. (n = 3). N.S., not significant; * p,0.05 (Student t-test). B. RAW 264.7 cells were transfected with scramble siRNA or p62-siRNA as above A, and cells were treated with LPS (20 ng/mL) for 4 h. After two washes with complete media, cells were further incubated with gMDP (5 mg/mL) for another 4 h and TNF-a concentrations in cell culture media were measured by ELISA according to the manufacturer's instructions (eBioscience). Data are expressed as mean 6 S.D. (n = 4); * p,0.05 (Tukey's Multiple Comparison Test). C. THP-1 cells were stably transfected with sh-scramble (sh-Scramble) or p62 targeting (sh-p62) small-hairpin RNA producing constructs as described in ''Methods''. Three cell clones stably knocked down in p62, pooled sh-Scramble control clones and non-treated wild-type cells were treated with gMDP (5 mg/mL) for 4 h and TNF-a production in the cell culture media was measured using TNF-a bioassay as described in ''Methods''. doi:10.1371/journal.pone.0057138.g006 p62 Enhances NOD2 Signaling PLOS ONE | www.plosone.org which fully overlapped with those of Myc-NOD2. Collectively, these results suggest that, in the presence of gMDP, p62 forms a higher degree of complex with NOD2 that may prevent the 26S proteasomal degradation of NOD2.

p62 is required for cytokine production mediated by NOD2 in macrophages
To confirm the role of p62 in physiologically relevant cell types, we used two macrophage cell lines of murine and human origin: RAW 264.7 (mouse) or THP-1 (human). RAW 264.7 cells express low levels of NOD2 which is rapidly induced by LPS [54]. Consistently, gMDP alone did not induce expression of pro-IL-1b in si-Scramble RNA-transfected RAW 264.7 cells (Fig. 6A). LPS induced pro-IL-1b expression at low levels. In cells pretreated with LPS for 4 h, gMDP significantly enhanced pro-IL-1b expression, as similarly demonstrated before [36,55]. However, in RAW 264.7 cells transfected with si-p62, no such enhancing effect was detected. Knocking down p62 had no effects on LPS-induced pro-IL-1b expression. In addition, production of TNF-a in response to gMDP was measured in LPS-primed RAW 264.7 cells with or without si-p62. LPS alone induced TNF-a production which was further increased by gMDP (Fig. 6B). However, si-p62 significantly prevented gMDP-induced TNF-a in LPS-primed cells. To further examine the role of p62 in human macrophages, THP-1 cells were knocked down in p62 using small hairpin RNAs (shRNA-p62). Three THP-1 cell clones stably knocked down in p62, pooled clones stably transfected with scrambled sh-RNAs (sh-Scramble), and non-infected wild-type cells were treated with gMDP (Fig. 6B). THP-1 cells responded to gMDP without priming with LPS and induced high levels of TNF-a in wild-type and sh-Scramble transfected clones. However, all three clones knocked down in p62 failed to respond to gMDP. Collectively, these results suggest that p62 was indeed required for optimal IL-1b and TNF production in response to NOD2 in mouse and human macrophages, respectively. p62 traffics ubiquitinated molecules to autophagosomes through interacting with LC3 [48]. Therefore, we examined the involvement of LC3 in p62-mediated regulation of NOD2. RAW 264.7 cells were knocked down in LC3 using si-RNAs (si-LC3) and pro-IL-1b expression in response to gMDP, LPS and LPS+gMDP were examined. However, knocking down LC3 had no effects on IL-1b production induced by LPS or LPS+gMDP (Supplemental Fig. S5). These results, together with data shown in Fig. 4D suggest that the enhancing effects of p62 in NOD2 stabilization and signaling are not mediated through autophagy.
p62 was first found to be required for NF-kB activation induced by IL-1 [56] or NGF [41,57] and its sustained activation in RANK (Receptor Activator of Nuclear Factor k B)-activated osteoclasts [40]. Consistently, p62-deficient mice have defects in sustaining activation of NF-kB in T cells [58]. It was shown that p62 interacts with TRAF6, protein kinase C, MAP kinase kinases and PDK1, which enhances NF-kB and Akt activation [59,60]. However, p62 was also found to be involved in both positive and negative regulation of NF-kB by interacting with a deubiquitinating enzyme, CYLD [61,62]. In macrophages, p62 is involved in both TLR-and NLRP3-mediated signaling events. It plays a suppressive role in interferon c and CpG DNA (TLR9 ligand)-induced cytokine production [63]; whereas, only a partial effect has been detected on TLR4-induced signaling events, namely, activation of the p38 and c-Jun N-terminal kinase but not NF-kB, and production of IL-6 but not TNF [31]. p62 also plays a critical role in NLRP3 inflammasome activation induced by Mycobacterium abscessus [64] but at the same time limits NLRP3 inflammasome activation through targeting inflammasomes to autophagy-medi-ated destruction as a feedback mechanism [65]. Therefore, p62 is involved in multiple signaling cascades with different roles. Here, we demonstrated that p62 plays a positive role in NOD2-mediated signaling cascades probably through forming aggregation of NOD2/p62 oligomers that prevents their degradation. Further studies are required to delineate the mechanism of p62 in preventing degradation of NOD2 and its physiological significance in NOD2 innate immune function. Figure S1 p62 has no effects on TRAF6 and NOD2 interaction. HEK293T cells were transiently transfected with pCMV-HA-NOD2 and pCMV-Myc-TRAF6 with or without si-p62 using a PolyJetTM (SignaGen Laboratories). Myc-TRAF6 was immunoprecipitated with anti-Myc antibody (the second lane). HA-NOD2 was co-precipitated with Myc-TRAF6 regardless of the presence of small interference RNAs against p62 (si-p62). Endogenous p62 was knocked down by si-p62 (bottom lane). (TIF) Figure S2 NOD2 interacts with TRAF6 or UBA domain deletion mutants. A. The schematic structure of p62 and mutant constructs are shown. B. HEK293T cells were transiently transfected with expression vectors encoding Myc-tagged LRR region deleted NOD2 (Myc-DLRR) and/or different mutants of deletion mutant of p62. Twenty four h post-transfection, total cell lysates were subjected to immunoprecipitation using anti-GFP antibodies and the immune complexes were resolved by SDS-PAGE, followed by immunoblotting against anti-Myc antibodies. Both TRAF6-interacting domain or UBA domain deletion mutants co-immunoprecipitated with NOD2. Data shown are representative images of 3 independent experiments. (TIF) Figure S3 Degradation of NOD2 in p62 knocked down cells was prevented by the 26S proteasome inhibitor MG132. HEK293T cells stably transfected with pLNCX-NOD2 were treated with si-p62 using a PolyJetTM (SignaGen Laboratories) for 24 h. Cells were then treated with the translation inhibitor cyclohexamide (CHX, 100 mg/ml) and gMDP (5 mg/ml) with or without the 26S proteasome inhibitor MG132 (25 mM) for the time indicated. Stability of NOD2 was analyzed by Western blots using anti-NOD2 (4A11). Western blots for p38 were used as loading controls. (TIF) Figure S4 Size exclusion gel filtration analysis and formation of a higher form p62 complex formaiton by gMDP. A. A mixture of 2000 kDa (Blue dextran; Peak I), 200 kDa (b-amylase; Peak II) and 66 kDa (Bovine serum albumin; Peak III) proteins were eluated through Superdex TM 200 gel filtration column. Elution of standard protiens were detected by UV light. B. HEK293T cells were transiently transfected with pCMV-Myc-NOD2 using a PolyJetTM (SignaGen Laboratories) for 24 h. Cells were then treated with gMDP (5 mg/mL) for 4 h and cell extracts were loaded onto the gel filtration column. Elution of p62 was analyzed using Western blots against p62 on each fraction (left panel) and intensities of immuno-reacted bands were ploted (right panel, n = 2). In non-treated cells, p62 complexes were eluted between 2000 kDa-200 kDa fractions; whereas, in gMDP-treated cells, p62 was eluted in $ 2000 kDa fractions. These results indicate that gMDP caused a higher degree of p62 complex formation. (TIF) p62 Enhances NOD2 Signaling Figure S5 Knocking down LC3 has no effects on p62mediated NOD2 signaling regulation. RAW264.7 cells were treated with scrambled-or map1lc3a (LC3)-specific siRNA for 24 hr. Cells were treated with LPS (50 ng/mL) for 4 hr and were rinsed twice with fresh media, followed by a subsequent treatment with gMDP (5 mg/mL) for an additional 4 hr. Total cell lysates were resolved by 14% SDS-PAGE, transferred onto PVDF and blotted with anti-LC3 and anti-IL-1b antibodies. (TIF)