UAF1 deubiquitinase complexes facilitate NLRP3 inflammasome activation by promoting NLRP3 expression

NOD-like receptor protein 3 (NLRP3) detects microbial infections or endogenous danger signals and activates the NLRP3 inflammasome, which has important functions in host defense and contributes to the pathogenesis of inflammatory diseases, and thereby needs to be tightly controlled. Deubiquitination of NLRP3 is considered a key step in NLRP3 inflammasome activation. However, the mechanisms by which deubiquitination controls NLRP3 inflammasome activation are unclear. Here, we show that the UAF1/USP1 deubiquitinase complex selectively removes K48-linked polyubiquitination of NLRP3 and suppresses its ubiquitination-mediated degradation, enhancing cellular NLRP3 levels, which are indispensable for subsequent NLRP3 inflammasome assembly and activation. In addition, the UAF1/USP12 and UAF1/USP46 complexes promote NF-κB activation, enhance the transcription of NLRP3 and proinflammatory cytokines (including pro-IL-1β, TNF, and IL-6) by inhibiting ubiquitination-mediated degradation of p65. Consequently, Uaf1 deficiency attenuates NLRP3 inflammasome activation and IL-1β secretion both in vitro and in vivo. Our study reveals that the UAF1 deubiquitinase complexes enhance NLRP3 and pro-IL-1β expression by targeting NLRP3 and p65 and licensing NLRP3 inflammasome activation.

Ubiquitination is a key post-translational modification (PTM) that controls NLRP3 inflammasome activation 7,8 . In resting macrophages, NLRP3 is polyubiquitinated with mixed K48 and K63 ubiquitin chains, which is crucial for the maintenance of NLRP3 inactivation. NLRP3 is deubiquitylated upon priming and activation, and that is a key step in the formation and activation of the NLRP3 inflammasome 9,10 . ABRO1 recruits BRCC3 to remove K63-linked ubiquitination of NLRP3, facilitating the assembly and activation of NLRP3 inflammasome 11 . K48-linked ubiquitination mediates protein degradation of NLRP3 and thus limits NLRP3 inflammasome activation. Although several E3 ubiquitin ligases such as TRIM31, March7, ARIH2, and FBXL2 9, [12][13][14] have been reported to attenuate NLRP3 inflammasome activation by mediating NLRP3 protein degradation, the function of K48-linked deubiquitination on NLRP3 inflammasome activity remains largely unclear. Whether any deubiquitinating enzymes exist to specifically remove K48-linked ubiquitination of NLRP3, stabilize its expression, and thus license NLRP3 inflammasome activation, remains to be investigated.
The UAF1/USP12 and UAF1/USP46 complexes interact with p65 and inhibit its ubiquitination and degradation, thus promoting NF-κB activation, resulting in the enhancements of NLRP3 and pro-IL-1β expressions. Consequently, the UAF1 deubiquitinase complexes facilitates NLRP3 inflammasome activation via targeting NLRP3 and p65. Our study thus uncovers mechanisms regulating NLRP3 inflammasome activation and suggests a promising approach for modulating NLRP3-dependent immunopathologies.

Results
UAF1 facilitates NLRP3 inflammasome activation. To investigate the potential role of UAF1 in NLRP3 inflammasome activation, we first examined the effects of Uaf1 deficiency on IL-1β secretion and caspase-1 cleavage. Systemic deletion of Uaf1 causes early embryonic lethality in mice 30 . Therefore, we generated Uaf1 flox/flox mice and then crossed them with Lyz2-Cre mice to specifically knockout Uaf1 in myeloid cells (called 'Uaf1 CKO ' here) ( Supplementary Fig. 1). ATP-stimulated and Nigstimulated IL-1β secretion and caspase-1 cleavage in LPSprimed Uaf1-deficient mouse peritoneal macrophages were markedly attenuated (Fig. 1a-c). In addition, Uaf1 deficiency could inhibit TNF and IL-6 secretions. To confirm the intrinsic role of UAF1, siRNAs targeting mouse Uaf1 were used to suppress endogenous UAF1 expression ( Supplementary Fig. 2a). Uaf1 knockdown substantially suppressed NLRP3 inflammasome activation-dependent IL-1β secretion and caspase-1 cleavage in mouse peritoneal macrophages ( Supplementary Fig. 2b, c). We next investigated the physiological relevance of the effects of UAF1 on NLRP3 inflammasome activation in vivo. Induction of IL-1β by intraperitoneal (i.p.) injection of LPS was shown to be NLRP3 dependent 31,32 . In our study, IL-1β secretion induced by the LPS injection was much lower in the sera of Uaf1 CKO mice than in the sera of WT mice, indicating that Uaf1 deficiency inhibits NLRP3 inflammasome activation in vivo (Fig. 1d). Furthermore, TNF and IL-6 secretions were also reduced in the sera of Uaf1 CKO mice compared with WT mice. Collectively, these results indicate that UAF1 facilitates NLRP3 inflammasome activation, and promotes TNF and IL-6 secretions.
NF-κB subunit p65 can be inactivated and degraded by proteasomes 35 . We then examined whether USP12/USP46 could target p65 to remove its polyubiquitin chains. P65 interacted with USP12, USP46 and UAF1 in both resting and LPS-stimulated macrophages, but exhibited no association with USP1 (Fig. 4f). Consistently, the interaction between exogenously expressed p65 and USP12/USP46/UAF1 was observed in HEK293T cells   ( Fig. 4g). We then examined the effects of USP12/USP46/UAF1 on p65 polyubiquitination and observed that all the three could remove its polyubiquitination (Fig. 4h). Under physiological conditions, endogenous p65 was ubiquitinated upon LPS stimulation in macrophages, and Uaf1 deficiency markedly increased endogenous p65 ubiquitination (Fig. 4i). To study the forms of USP12/USP46/UAF1-mediated deubiquitination of p65, ubiquitin mutant vectors K48 and K63, in which all their lysine residues were substituted with arginine except at positions 48 and 63, respectively, were used in the transfection assays. UAF1, USP12, and USP46 markedly inhibited K48-linked ubiquitination of p65 (Fig. 4j), but had no effect on K63-linked ubiquitination of p65 ( Supplementary Fig. 5d). Taken together, these results indicate that the UAF1/USP12 and UAF1/USP46 complexes remove K48-linked ubiquitination of p65 and enhance its expression, thereby promoting NLRP3 transcription. Flag-USP1 Flag-USP12 Flag-USP46 Flag-UAF1   UAF1/USP1 interacts with NLRP3. USP1 enhanced NLRP3 protein expression, with no effect on Nlrp3 mRNA expression, suggesting that the UAF1/USP1 complex regulated posttranslational modification of NLRP3. We first examined the association between NLRP3 and UAF1/USP1. Confocal analysis demonstrated the colocalization between USP1/UAF1 and NLRP3 upon LPS stimulation (Fig. 5a, b). We expressed NLRP3 along with USP1 and UAF1 in HEK293T cells. NLRP3 could be immunoprecipitated with USP1 and UAF1 (Fig. 5c, d). In vitro binding assays demonstrated that NLRP3 could directly interact with USP1 and UAF1 (Fig. 5e, f). Next, to confirm the interaction between NLRP3 and UAF1/USP1 under physiological conditions, we assessed resting and LPS-stimulated macrophages by immunoprecipitation. An association between NLRP3 and USP1/UAF1 was detected in LPS-stimulated mouse peritoneal macrophages (Fig. 5g, h). Taken together, these data suggest that NLRP3 could directly interact with UAF1/USP1. To search for the domains of NLRP3 responsible for the interaction with USP1 and UAF1, a series of Myc-tagged NLRP3 truncated mutants were constructed (Fig. 5i). USP1 and UAF1 were coprecipitated with NLRP3 wild-type (WT), LRR domain deletion mutant (ΔLRR), NACHT domain deletion mutant   Confocal microscopy analysis of colocalization of NLRP3 with USP1 or UAF1. MEFs transfected with Myc-NLRP3, GFP-USP1 or Flag-UAF1 were stimulated with LPS for 2 h, then fixed and incubated with a secondary antibody conjugated to Alexa Fluor 637 or Alexa Fluor 488. Scale bar, 10 µm. c, d Extracts from HEK293T cells transiently transfected with Flag-USP1 and Myc-NLRP3 (c), and Flag-UAF1 and Myc-NLRP3 (d) were subjected to IP with anti-Flag, followed by western blot analysis with anti-Myc and anti-Flag. e PCMV6-USP1 and Myc-tagged NLRP3 were obtained by in vitro transcription and translation. Interaction between USP1 and NLRP3 was analyzed by mixing USP1 and NLRP3 together, followed by IP with USP1 antibody and immunoblot analysis with NLRP3 and USP1 antibody. f HA-tagged UAF1 and Myc-tagged NLRP3 were obtained by in vitro transcription and translation. Interaction between UAF1 and NLRP3 was assayed by mixing UAF1 and NLRP3 together, followed by IP with HA antibody and immunoblot analysis with NLRP3 and HA antibody. g, h Extracts of peritoneal macrophages stimulated with LPS for the indicated time periods were subjected to immunoprecipitation with anti-USP1 (g) and anti-UAF1 (h) followed by western blot analysis with indicated antibody. Proteins from a wholecell lysate were used as positive control. i Schematic diagram of NLRP3 and its truncation mutants. j, k Myc-tagged NLRP3 or its mutants along with Flag-USP1 (j) or Flag-UAF1 (k) were individually transfected into HEK293T cells. The cell lysates were immunoprecipitated with Flag antibody and then immunoblotted with the indicated antibodies. US, unstimulated. Similar results were obtained from three independent experiments.
(ΔNACHT), PYD domain deletion mutant (ΔPYD), LRR domain mutant (LRR) and NACHT domain mutant (NACHT) (Fig. 5j, k and Supplementary Fig. 6). However, the PYD domain mutant (PYD) lost the ability to bind USP1 and UAF1 (Fig. 5j, k). These results indicate that the LRR and NACHT domains of NLRP3 interact with UAF1/USP1. UAF1/USP1 removes NLRP3 ubiquitination and stabilizes NLRP3. NLRP3 could be ubiquitinated with both K48 and K63 linkage. We identified the deubiquitinase complex UAF1/USP1 as NLRP3-associated proteins (Fig. 5), which prompted us to investigate whether the complex could remove the ubiquitination of NLRP3. Both USP1 and UAF1 considerably inhibited ubiquitination of NLRP3 (Fig. 6a, b). However, the USP1 point mutation (C90S) with substitution of the cysteine residue with serine at position 90, lost the ability to inhibit the polyubiquitination of NLRP3 (Fig. 6a), indicating that USP1 could remove the ubiquitination of NLRP3 through its deubiquitinase activity. Given that NLRP3 possibly forms a complex with other proteins, we performed a two-step immunoprecipitation assay (Re-IP) to exclude the effects of other proteins in the complex. In HEK293T cells, Myc-tagged NLRP3 was cotransfected with ubiquitin and USP1 or UAF1, respectively. The cell lysates were subjected to immunoprecipitation with anti-Myc, and then the immunoprecipitates were denatured, followed by reimmunoprecipitation with Myc antibody. Notably, NLRP3 polyubiquitination was markedly attenuated in the presence of USP1 or UAF1 (Fig. 6c, d). Furthermore, USP1 inhibited NLRP3   h Cell lysates from WT or Uaf1 deficient mouse peritoneal macrophages stimulated with LPS were immunoprecipitated with anti-NLRP3, followed by western blot analysis with anti-Ub. i, j Western blot analysis of NLRP3 and caspase-1 expression in mouse peritoneal macrophages pretreated with DMSO or ML323 for 4 h and then stimulated with LPS for 4 h, followed by treatment with cycloheximide (CHX) for the indicated time periods. NLRP3 expression level was quantitated by measuring band intensities using 'ImageJ' software (j). The values were normalized to actin (mean ± SEM, two-tailed t-test ML323 vs. DMSO, **P < 0.0001, respectively; n = 3 independent experiments). k Western blot analysis of NLRP3 expression in WT and Uaf1 deficient mouse peritoneal macrophages pretreated with LPS for 8 h, followed by treatment with MG132 for 4 h before harvesting the cells. Similar results were obtained from three independent experiments. polyubiquitination with K48-linkage, but not with K63-linkage (Fig. 6e), indicating that USP1 selectively removed K48-linked polyubiquitination of NLRP3. Consistently, UAF1 also inhibited K48-linked polyubiquitination of NLRP3 (Fig. 6f). Under physiological conditions, endogenous NLRP3 was ubiquitinated upon LPS stimulation. Moreover, polyubiquitination of NLRP3 was enhanced in ML323-treated mouse embryonic fibroblasts (MEFs) (Fig. 6g) and Uaf1-deficient macrophages (Fig. 6h). Collectively, these data indicate that UAF1/USP1 directly eliminates K48linked polyubiquitination of NLRP3 through its deubiquitinase activity. K48-linked protein ubiquitination leads to degradation of the corresponding proteins. Both UAF1 and USP1 enhanced LPSinduced NLRP3 expression. We then investigated the effects of UAF1/USP1 complex on NLRP3 protein degradation by cycloheximide (CHX) chase experiment. ML323 treatment significantly promoted NLRP3 protein degradation, with no effect on caspase-1 (Fig. 6i, j). In addition, Uaf1 deficiency resulted in a loss of the inhibitory effects on NLRP3 expression in MG132-treated macrophages (Fig. 6k). Taken together, these data indicate that the USP1/UAF1 deubiquitinase complex targets NLRP3 and inhibits its degradation, thus facilitating NLRP3 inflammasome activation.
ML323 and Uaf1 deficiency ameliorate NLRP3-dependent inflammation. Next, we investigated whether ML323, a selective UAF1/USP1 inhibitor, could suppress NLRP3 inflammasome activation. ML323 treatment attenuated NLRP3-dependent IL-1β secretion and caspase-1 cleavage in macrophages (Fig. 7a, b). ML323 is known to be active in vivo 36 . We then examined the physiological relevance of the UAF1/USP1 complex in inflammation in vivo using ML323. ML323-treated mice produced significantly less IL-1β and TNF in sera after i.p. injection of LPS than control mice, while no difference in IL-6 secretion was observed (Fig. 7c).
Folic acid (FA)-induced acute tubular necrosis (ATN) is tightly associated with NLRP3 inflammasome activation 37 . To evaluate the roles of ML323 in FA-induced ATN, mice were i.p. injected with ML323 followed by an FA injection. ML323 treatment considerably reduced IL-1β secretion in sera (Fig. 7d) and NLRP3 expression in the kidneys (Fig. 7e). Renal inflammation and edema were greatly ameliorated in the kidneys of ML323-treated mice (Fig. 7f, g). Concordantly, ML323-treated mice were found to be more resistant in survival assays upon FA injection (Fig. 7h). Therefore, as a selective UAF1/USP1 complex inhibitor, ML323 inhibits IL-1β secretion and therefore ameliorates NLRP3dependent inflammation both in vitro and in vivo.
We further investigated the physiological relevance of UAF1 on inflammation in FA-induced ATN using Uaf1 CKO mice. Uaf1 deficiency in myeloid cells markedly suppressed IL-1β secretion and pro-IL-1β expression in the kidneys, due to the regulatory roles of UAF1 in NF-κB activation and NLRP3 expression (Fig.7i); less severe renal inflammation and edema were observed in Uaf1 CKO mice (Fig. 7j, k). These data indicate that Uaf1 deficiency in myeloid cells ameliorates FA-induced ATN and suggest UAF1 as a physiological enhancer of inflammation.

Discussion
The NLRP3 inflammasome is activated by a wide variety of stimuli, including PAMPs from bacteria, viruses, and fungi; endogenous DAMPs in sterile inflammation; and exposure to environmental irritants 1,7,38 . Thus, NLRP3 inflammasome activity is strongly associated with a variety of diseases; hence, its activation should be tightly controlled. In this study, we provide several lines of evidence to demonstrate that UAF1 deubiquitinase complexes promote NLRP3 inflammasome activation by enhancing NLRP3 expression. As a cofactor of several deubiquitinases, UAF1 plays different roles by combining with USP1, USP12, and USP46, respectively. On the one hand, UAF1/ USP12 and UAF1/USP46 complexes stabilize p65 expression, thus promoting the NF-κB signaling pathway, which results in the enhancement of NLRP3, pro-IL-1β, TNF, and IL-6 transcription. On the other hand, UAF1/USP1 complex selectively interacts with NLRP3, removes NLRP3 K48-linked ubiquitination and prevents NLRP3 from proteasomal degradation. Therefore, UAF1 facilitates NLRP3 inflammasome activation by enhancing IL-1β and NLRP3 expression at the mRNA and protein levels. Previously, we reported that LPS induced UAF1 expression in macrophages 36 . Thus, UAF1 might be a feedback enhancer of NLRP3 inflammasome activation.
NLRP3 protein expression is considered to be a rate-limiting step and causes subsequent NLRP3 inflammasome activation. In resting macrophages, NLRP3 expression is relatively low; hence, NLRP3 inflammasome assembly is hardly induced. However, following stimulation with exogenous and endogenous factors, including TLR agonists and proinflammatory cytokines, NLRP3 expression is dramatically induced to establish quick responses to NLRP3 activators. At the transcriptional level, NF-κB is critical for Nlrp3 mRNA expression 34 . Zinc finger protein A20 deubiquitinates NF-κB, thereby signaling molecules to suppress NF-κB activation and NLRP3 transcription 39 . Aryl hydrocarbon receptor (AhR) binds to the xenobiotic response elements (XREs) located within two NF-κB binding sites in the NLRP3 promoter and attenuates NLRP3 transcription and NLRP3 inflammasome activation 40 . In this study, we demonstrate that UAF1/USP12 and UAF1/USP46 complexes target p65 and promote NF-κB activation, resulting in the enhancement of NLRP3 transcription and NLRP3 inflammasome activation. In addition, the UAF1/USP12 and UAF1/USP46 complexes also enhance the transcription of proinflammatory cytokines, including pro-IL-1β, TNF, and IL-6, which further promote NLRP3 inflammasome activation. At posttranslational level, multiple E3 ubiquitin ligases promote NLRP3 ubiquitination and protein degradation, which limit NLRP3 inflammasome activity. In this study, we observed that the UAF1/USP1 deubiquitinase complex inhibits proteasomal degradation of NLRP3 and enhances NLRP3 inflammasome activation. Deubiquitination of NLRP3 is required for optimal NLRP3 inflammasome activation. Our results indicate that NLRP3 deubiquitination is important for the maintenance of NLRP3 protein level, thus providing a mechanism to elucidate the indispensable roles of ubiquitination in NLRP3 inflammasome activation.
UAF1, as a cofactor of USP1, USP12, and USP46, could enhance their deubiquitinase activity by forming stable UAF1/ USP protein complexes 41 . UAF1/USP1, UAF1/USP12, and UAF1/USP46 complexes play vital roles in DNA repair processes and tumor pathogenesis. Here, we demonstrated that UAF1 works together with USP12/USP46 and USP1 to facilitate NLRP3 and pro-IL-1β expression, and license NLRP3 inflammasome activation. As NLRP3 inflammasome needs two signals (priming signal and activation signal) to be activated, it acts just like a precise lock. Under certain circumstances, UAF1 and its partners (USP1/USP12/USP46) work as multiple keys to unlock and tightly control the activation of NLRP3 inflammasome at different levels ( Supplementary Fig. 7).
Inappropriate NLRP3 inflammasome activation has been implicated in a variety of diseases, which makes NLRP3 an attractive drug target. Many inhibitors of the NLRP3 pathway have been developed and extensively investigated in multiple NLRP3 inflammasome-related disease models. For example, MCC950, a best studied NLRP3 inhibitor, displays efficacy in a wide range of NLRP3-dependent murine disease models 32 . CY-09, another selective NLRP3 inhibitor, is also found to be efficacious in a mouse model of cryopyrin-associated periodic syndromes (CAPS) 42 . In this study, we identified ML323 as a novel NLRP3 inflammasome inhibitor. ML323 is a specific inhibitor of the UAF1/USP1 complex, with no inhibitory effects on both UAF1/USP12 and UAF1/USP46 complexes 33 . ML323 inhibits NLRP3 protein expression, suppresses NLRP3 inflammasome activation, thereby ameliorating NLRP3-dependent inflammation both in vitro and in vivo. Our results suggest that ML323 might be a promising candidate for the treatment of inflammatory diseases caused by aberrant NLRP3 activity.
In conclusion, by identifying UAF1 as a critical enhancer of NLRP3 and pro-IL-1β expression, this study provides insights into the mechanism of NLRP3 inflammasome activation. UAF1 enhances NLRP3 expression at both the mRNA and protein levels, as well as pro-IL-1β transcription through recruiting USP1/12/46. The UAF1/USP1 deubiquitinase complex directly targets NLRP3 and inhibits its ubiquitous degradation. The UAF1/USP12 and UAF1/USP46 complexes suppress the ubiquitous degradation of p65, promote NF-κB activation, and enhance NLRP3 and pro-IL-1β transcription. These results indicate that the UAF1 deubiquitinase complexes enhance NLRP3 and pro-IL-1β expressions via targeting NLRP3 and p65, and thus facilitates NLRP3 inflammasome activation.