Sensing soluble uric acid by Naip1-Nlrp3 platform

Uric acid (UA), a product of purine nucleotide degradation able to initiate an immune response, represents a breakpoint in the evolutionary history of humans, when uricase, the enzyme required for UA cleavage, was lost. Despite being inert in human cells, UA in its soluble form (sUA) can increase the level of interleukin-1β (IL-1β) in murine macrophages. We, therefore, hypothesized that the recognition of sUA is achieved by the Naip1-Nlrp3 inflammasome platform. Through structural modelling predictions and transcriptome and functional analyses, we found that murine Naip1 expression in human macrophages induces IL-1β expression, fatty acid production and an inflammation-related response upon sUA stimulation, a process reversed by the pharmacological and genetic inhibition of Nlrp3. Moreover, molecular interaction experiments showed that Naip1 directly recognizes sUA. Accordingly, Naip may be the sUA receptor lost through the human evolutionary process, and a better understanding of its recognition may lead to novel anti-hyperuricaemia therapies.


Introduction
Host responses against harmful signals are basic physiological reactions of all living organisms. Innate immunity pattern-recognition receptors (PRRs) were firstly described as recognizing conserved structural components of microorganisms [1]. The discovery of Toll-like receptor (TLR) [2] led us to understand how the immune system responds to non-self-antigens in the context of an infection [3,4], contrasting to the previous model in which the immune system reacted to all non-self-antigens while being tolerant to selfones [5,6]. Based on Polly Matzinger's studies stating that "the immune system is more concerned with entities that do damage than with those that are foreign" [6], several damage-associated molecular pattern (DAMPs) have been described. Yet, according to this theory, the "foreignness" of a pathogen is not the important feature that triggers a response, and "self-ness" is no guarantee of tolerance. Indeed, receptors for endogenous and exogenous signals may have evolved simultaneously once vertebrates and pathogens have shared eons of evolutionary time and space [6]. Perhaps PRRs have not evolved to bind to pathogens at all; the pathogens, instead, may have evolved to attach to them and enhance their own survival [7], a hypothesis that would explain a puzzling feature of PRRs that each one can attach to many different kinds of molecules.
Among several DAMPs, uric acid (UA), the product of purine catabolism, released mainly from dying cells and ischemic tissues, is considered a major alarmin, especially when it is present at elevated levels and crystalized -also known as monosodium urate (MSU) [8].
In rodents, MSU activates the immune system [9,10], acts as a pro-oxidant molecule, stimulates chemotaxis and also activates NF-κB and MAPK pathways [11]. Moreover, MSU induces the release of IL-1β through the activation of inflammasome-dependent caspases [10,12,13]. The inflammasome is a cytosolic complex mounted upon PAMPs/DAMPs sensing by a nucleotide binding domain (NBD/NACHT) and leucine-rich repeats (LRRs) containing receptor (NLR) [14]. NLRs belong to a superfamily of innate immune proteins with a very conserved structure along the phylogeny, from plants to mammals. NLRs shared NBD/NACHT and LRRs domains. Still they present a sub-family specific N-terminal domain: a pyrin domain (PYD) (NLRP), a caspase-recruitment domain (CARD) (NLRC), a Baculovirus Inhibitor of Apoptosis Protein Repeat (BIR) domain (NLRB), or an Acidic Transactivating Domain (ATD) (NLRA). In distinct phylogenic groups, the number of receptors and of paralogous genes differs, possibly as a consequence of a host/pathogen and/or environmental co-evolution [15]. Up to now, it has been demonstrated that both soluble UA (sUA) [16], as well as MSU, can induce Nlrp3 inflammasome activation in mice. In humans, gout is an inflammatory disease triggered by the deposition of MSU within joints and connective tissues, whereas Nlrp3 inflammasome is activated by UA crystals [17].
In humans, UA crystallization happens when its level reaches 6.8 mg/dL in plasma, while in rodents, the solubility threshold is about 10-fold lower [18]. Great apes have higher levels of UA in the serum (3.02 to 6.72 mg/dL, corresponding to 180 to 400 μΜ), compared to other animals (18 to 40 μΜ). This observation is compatible with the absence of the uricase (or urate oxidase) activity, the enzyme involved in purine catabolism converting UA into allantoin [19,20]. The loss of uricase at the divergence between great apes and other mammals may be related to a survival advantage, as previously hypothesized, due to the UA characteristics as a molecule responsible for saving energy [21]; however, it takes up a tricky question about the role of a mammal's sensor for UA. We hypothesize that along with uricase lost and a consequent elevation of UA serum levels; human has lost the sensors to recognize "high" levels of sUA.
Among NLR receptors able to induce inflammasome activation in mice and humans, the sub-family of NLRB took our attention. In mice, exist 6 paralogous genes, namely Naip 1-6 and 4 functional receptors (Naip-1, 2, 5, and 6), while in humans only one orthologous gene, Naip has been found [22]. Despite the fact that the differences in the aminoacidic content among the Naip proteins, both mice and human receptors have been described to play a role in defense against pathogens [23,24]. Murine (m) Naip1 and Naip2 are respectively responsible for the detection of needle and rod proteins, structural proteins of the bacterial secretion system called type III secretion system (T3SS) [25][26][27]. Naip5 is responsible for the cytosolic recognition of flagellin [28], the major protein component of the bacterial flagellum. Similarly to Naip5, human Naip (hNaip) can bind bacterial flagellin [29] and to activate Nlrc4 inflammasome. In this scenario, Naip acts as a ligand sensor and the Nlrc4 is responsible for inflammasome activation and inflammasomedependent cell death, known as pyroptosis [30]. Considering the differences of NLRB orthologous genes among different species and that until nowadays, any of NLRB was described as having endogenous ligands, in this study we performed transcriptome-and proteome-wide analysis in addition to interaction investigation and structural modeling predictions to study the sensing of sUA by mNaip1. Besides demonstrating that expressing mNaip1 into human cells allow them to be activated upon sUA stimulus, we found that mNaip1 directly recognize sUA. We then hypothesized that Naip could be the lost receptors for UA, and in particular for sUA.

Naip1 is involved in sUA response
To assess the difference in serum basal level of UA among species, we have initially measured the UA levels of unrelated healthy adult human donors (n=5), C57Bl/6 adult mice (n=5), and adult old-world monkeys (rhesus macaque; n=5). As expected, humans presented an average blood UA concentration of 295 μΜ, 4, and 7 times more elevated UA levels when compared to mice and rhesus macaques, respectively (Fig. 1A). Then, we have stimulated murine LPS-primed BMDM and human and rhesus LPS-primed monocyte-derived macrophages with 200 μΜ of sUA. As observed, human cells did not produce IL1β after sUA stimulus when compared to LPS-primed cells (Fig. 1B). On the other hand, murine BMDM increased the IL1β production after sUA stimulus when compared to LPS-primed ones (Fig. 1C). Surprisingly, despite 200 μΜ being supra physiological level for rhesus macaque, their macrophages did not increase IL1β production after this sUA stimulus (Fig. 1D). Ischemic tissues consistently overproduce UA that trigger immune cell functions [31]. Both sUA stimulus and hypoxia condition of mouse BMDM led to increased Naip1 mRNA expression levels (5 and 15 times, respectively), but not Naip5 (Sup. Fig. 1A and 1B). Additionally, BMDM derived from Naip1 -/-and Naip -/-mice did not increase the IL-1β production upon sUA stimulation when compared to LPS-primed macrophages (Fig. 1E). On the other hand, Naip2 -/-and Naip5 -/-cells behaved as WT macrophages and, despite the variation within the group itself, there were no differences in the IL-1β production in the BMDM from Nlrc4 -/animals after sUA stimulus in comparison with LPS-primed cells from Nlrc4 -/-mice (Fig.   1E). In an attempt to confirm the role of murine Naip (mNaip) platform into sUA response, we virus-transduced human THP1 cells with mNaip1, mNaip5, mNaip6, mNlrc4, and empty backbone vector. Our data demonstrate that PMA-activated and LPSprimed THP1 cells produced IL-1β after sUA stimulus only after mNaip1 transduction, but not mNaip5, mNaip6, mNlrc4 or the control empty vector (Fig. 1F). Such data point to mNaip1 as a target gene involved into sUA response.

IL-1β production in human cells is dependent on Nlrp3 and mNaip1
Naip1 carrying plasmid was modified by adding a "self-cleaving" 2A (T2A) sequence between Naip1 and the color-tag sequences. In this sense, the overexpressed Naip1 protein is not color-tag, which could implicate in some misinterpreted data. Initially, it was investigated the ability of PMA-activated, LPS-primed, and mNaip1-expressing THP1 cells to produce IL-1β under the stimulus of UA degradation products, i.e. allantoin, urea and ammonium. None of the investigated products were able to induce IL-1β production, as sUA did (Fig 2A). The role of Nlrp3 for sUA sensing, as previously stated [16], was further investigated. For that, we initially evaluated the IL-1β production upon sUA stimulus in THP1 cells virally transduced with either mNaip1 or empty backbone, both stimulated in the presence or absence of a Nlrp3 inhibitor, CRID3 (1 ) [32]. IL-1β levels are reduced in cells pre-treated with CRID3 (Fig. 2B). IL-1β production was also evaluated in THP1 cells after Nlrp3 gene deletion by Crispr-Cas9. It was confirmed that IL-1β production is dependent on Nlrp3 activation (Fig. 2B-D) once Nlrp3-deleted cells exhibited decreased levels of IL-1β upon sUA stimulus ( Fig. 2C and 2D). To investigate the interaction between Naip1 and Nlrp3, an immunoprecipitation assays with THP1 cell lysates using both Nlrp3 and Naip1 as targets was performed (Sup. Fig. 2). However, the targets were only found in the whole-cell lysates but not in the immunoprecipitants, suggesting that Nlrp3 and Naip1may not directly interact. Altogether, these data indicate that the observed IL-1β production followed by sUA stimulus requires both Naip1 and Nlrp3 inflammasome platforms.

Naip1 triggers enhanced immune responses and altered cellular metabolites content toward sUA
To better define how mNaip1 influences sUA sensing, RNA-seq analysis of mNaip1-and backbone-transduced THP1 cells was performed. The presence of Naip1 affects the gene transcription, which was evident when the gene expression differences are represented as a Volcano plot (Sup. Fig. 3A). Unsupervised hierarchical clustering of 8,000 genes (Sup.  Fig 3B). Detailed inspection of the most highly differentially expressed genes revealed that ccl2, pik3cd, nck2, tab1, and fgfr1 were expressed more strongly in mNaip1-expressing GFP-tagged cells stimulated with sUA, when compared to GFP-transduced (control) cells at the same stimulus (Sup. Fig. 3C). Functional annotation enrichment analysis for KEGG and PANTHER pathways demonstrated that PMAactivated and LPS-primed THP1 cells expressing mNaip1 lead to a shift toward increased inflammation, cancer-and infection-related signaling pathways (Sup. Fig. 3D). Enrichment term analysis using the genes that were upregulated in mNaip1-expressing macrophages, was further visualized as a KEGG pathway enrichment network (Sup. Fig.   3E). Together, these analyses identified that upregulated genes were associated with processes involved in cytoskeleton regulation, adherent junctions, proteoglycans in cancer, as well as bacterial infection and invasion, and immune processes (Sup. Fig. 3E).
These data demonstrate that mNaip1 triggers enhanced immune responses after sensing sUA.
In addition to gene expression profile, we performed a proteomic analysis to identify differentially expressed proteins followed by sUA stimulus in both control (GFPtransduced) and mNaip1-expressing LPS-primed THP1 cells, and we highlighted in orange the proteins only found upon sUA stimulus and in yellow the proteins found only in LPS-primed condition (Sup. Fig. 4A, Sup. Table 02 and Sup. Table 03). Moreover, the two different cells under sUA stimulus were compared (Sup. Fig. 4B). Among 44 proteins only present in mNaip1-expressing cells, but not present in control ones, we highlighted 30 proteins, including some related to immune response such as thymopoietin (A0A024RBH7), CD99 (A8MQT7), stress-associated endoplasmic reticulum protein (Q9Y6X1), and lysosome-associated membrane glycoprotein 2 (H0YCG2) (Sup. Fig 4B).
However, only 2.8% of the expressed proteins correspond to differentially expressed genes (data not shown). We built a STRING network view of proteins only present in mNaip1 expressing cells when compared to GFP + control ones, both under sUA stimulation (Sup. Fig. 4C), which evidenced the metabolism-related up-regulation pathways, especially those related to lipid metabolism. These data demonstrate that additionally to immune response, mNaip1 triggers altered cellular metabolites content in LPS-primed macrophages stimulated with sUA.

Naip1 activation may be potentiated after the elevation of the cellular content of total lipid
Following the altered cellular metabolites content, UA is also described as increasing the accumulation of triglyceride into hepatic cells [33]. In this sense, we next investigated lipid drops formation in LPS-primed THP1 cells stimulated with sUA. It was observed an increase in cellular content of lipids after sUA stimulus but in a mNaip1-independent way ( Fig. 3A-B). Alteration in metabolites content could also change mitochondrial activity once these plastic organelles sense cellular metabolites, oxygen, and nutrients, and they exert central roles as source of energy and ROS [34]. Changes in mitochondrial membrane potential in live cells upon sUA stimulus were therefore measured ( Fig. 3C-D). Despite no differences in mitochondrial area within mitotracker staining, a reduced mitochondrial membrane potential was observed, as indicated by failure to load the positively charged mitochondrial indicator TMRE in mNaip1 expressing cells stimulated with sUA, when compared to LPS-primed cells ( Fig. 3C-D). We next measured the oxygen consumption ratio (OCR) of THP1 cells virally transduced with empty backbone or mNaip1, both LPSprimed, treated or not with sUA. sUA increased OCR but in a Naip1-independent manner.
The mitochondrial pyruvate carrier inhibitor UK5099 (100 ) was used in order to evaluate the ATP consumption derived from fatty acid indirectly. Again, such OCR increasing can be reverted by UK5099 pre-treatment in both cell types (Sup. Fig. 5A and 5B). Altogether, such data indicate that mNaip1 expression alters the cellular fatty acid content and reduces active mitochondria number upon sUA stimulus.
We next investigated whether the elevation of fatty acid synthesis could trigger IL-1β production. For that, it was measured the levels of IL-1β in the supernatant of LPS-primed cells virally transduced with empty backbone or mNaip1 after 6 hours of incubation with citrate [35] and palmitate. Despite significant production of IL-1β into empty backbonetransduced cells upon both citrate (5 m) or palmitate (100 ) stimulus when compared to LPS-primed cells, mNaip1-expressing cells produce even higher levels of IL-1β when compared to empty backbone transduced cells regardless the stimuli (Fig. 3E).
Moreover, LPS-primed mNaip1-expressing cells were stimulated with sUA or citrate in the presence of the acetyl-CoA carboxylase- inhibitor, TOFA (10 g/mL), or the phosphor-ACLy inhibitor, BMS303141 [36] (25 ), or the fatty acid synthase inhibitors, C75 (50 ) and cerulenin (5 g/mL) [37], as shown in the schematic figure 3F. As observed, all these inhibitors, but TOFA, led to decreased levels of produced IL-1β ( Fig.   3G) upon sUA or citrate stimuli. Altogether, our data suggest that sUA leads to fatty acid synthesis in a way independent of mNaip1. Saturated fatty acids promote Nlrp3 inflammasome activation [38], especially, palmitate [39]. We added that citrate-and palmitate-mediated IL-1β production is potentiated in the presence of Naip1.

Naip1 directly recognizes sUA
In a study investigating the role of different lipids in macrophage lipidomic, palmitate presented the most pronounced effects [40]. In order to investigate whether mNaip1 directly senses sUA and/or palmitate, we performed a quartz crystal microbalance with dissipation (QCM-D) analysis. After an initial immobilizing step with anti-GFP, we  figure 4B. It is important to observe a significant variation of response (ΔθSPR) obtained after addition of sUA (2 µ), which characterizes the interaction between sUA and mNaip1 protein. In this phase of the study, higher concentrations of sUA (12.5 to 200 µ) were also accompanied by significant responses (data not shown). In turn, the addition of palmitate at the concentration of 2 µ (green curve) and at higher concentrations (12.5 to 200 µ -data not shown) did not trigger a notable response. These results suggest that sUA directly binds to the mNaip1 protein.

hNaip and mNaip1 may respond differently to uric acid
As previously studied, for inflammasomes to be formed, NLR proteins, like Naip, must recognize ligands to be released from their autoinhibited state to finally trigger the oligomerization of NLRCs, and assemble inflammasome complexes [41,42]. After modelling, the structures of mNaip1 and hNaip, in their inactive forms, we observed important differences on their surface electrostatic properties, which may directly interfere with their ability to recognize specific ligands. By performing molecular In association with our experimental results, based on homology modelling and molecular docking, we hypothesize the surface electrostatics of mNaip1 enables this protein to recognize sUA and stop its autoinhibition, tasks that hNaip is unable to perform. With the deletion of the uricase gene on great apes [19], over its evolution, mutations on hNaip surface electrostatics were probably selected to increase its physiological tolerance to high levels of sUA, in such a way to prevent activation of inflammasomes, and allow great apes to benefit from the survival advantages provided by high levels of serum UA [21].

Discussion
Evidence is reported that sUA increases Naip1 transcription in murine macrophages. So, Naip1 could be responsible for cellular signalling triggered by sUA. Among all mNaip, Naip1 presents the higher aminoacidic content similarity with the hNaip, about 70%.
Despite some reports pointing to hNaip recognizing the same ligand as mNaip5 [29], some studies indicate that hNaip recognizes the same bacterial components as mNaip1 [26,43]. It is possible that the 30% differences between human and murine proteins are associated with sUA signaling. This is the first study to postulate that Naip recognizes a DAMP. Once our previous work suggests that sUA activates the Nlrp3 inflammasome [16], we also investigated the role of Nlrp3 in mNaip1 expressing cells. LPS-primed human THP1 cells only produce the mature form of IL-1β under sUA stimulus when they express Nlrp3 and mNaip1. Most inflammasomes are believed to include only a single NLR, though other NLR-NLR interactions have been proposed. The interplay between Nlrp3 and Nlrc4 reveals an unexpected overlap between what had been considered distinct inflammasome scaffolds [44]. Besides, it was reported that Nlrc4 can recruit Nlrp3 through its NACHT domain, in the context of S. typhimurium infection [44]. It was also demonstrated that NEK7, a member of the family of mammalian NIMA-related kinases (NEK proteins), is a Nlrp3-binding protein that acts downstream of potassium efflux to regulate Nlrp3 oligomerization and activation [45]. Hence, besides the mechanisms by which sUA activates mNaip1 inflammasome, it remains to be determined if mNaip1 interacts with Nlrp3 after sUA stimulus once we observed no clear evidence of Nlrp3 and mNaip1 protein interaction.
Besides responding to PAMPs and DAMPs, recent data suggest that the immune system act as a signal integrator able to detect disturbances in cytoplasmic cells related to metabolites. This monitored disruption is termed "homeostasis-altering molecular processes" (HAMPs) [46], and it provides powerful flexibility in the ability of the innate immune system to detect infections and chronic inflammatory diseases. It has been shown that the Nlrp3 inflammasome complex activation and the posterior caspase-1 and IL-1 production occurs following the saturated fatty acid palmitate triggering even in humans' cells [47,48]. Both QCM and SPR immunosensor based on the immobilization of Naip1 GFP /anti-GFP/Au for specific sUA detection were successfully performed. The results indicate that the three-dimensional structure of mNaip1 provided a great accessible area for interaction with sUA, and no accessible area for interaction with palmitate. Homology modelling and molecular docking analysis indicate that the surface electrostatics of mNaip1 enables it to recognize sUA through abolishing its autoinhibition state, tasks that hNaip is unable to perform. Despite we have demonstrated that sUA, but not palmitate, is responsible by directly binding to mNaip1, our data pointed to an increased IL-1 production in the context of Naip1 expression in human's cells followed by palmitate stimulus. Therefore, further analyses are necessary in order to better investigate the mechanisms by which palmitate could lead to Naip1 permissiveness to sUA.
Additionally, the transcriptome and the cellular metabolites content was changed in cells upon sUA stimulus, being the presence of mNaip1 altered some metabolism-related enzymes besides favoured the increase in the immune responses toward sUA. Based on the central dogma, it was generally assumed that there exists a direct correlation between mRNA transcripts and protein expressions, however, recent studies have been showing that this correlation can be low due to various factors such as post transcription modifications [49] or even the time mRNA and protein are expressed [50]. Moreover, we observed it reduced active mitochondria in mNaip1 expressing cells stimulated with sUA.
This result corroborates a study demonstrating that the percentage of TMRE + cells was significantly lower in LPS-primed macrophages stimulated with ATP, compared to the control ones [51]. In this mentioned study, the TMRE dye was not trapped in the mitochondrial membrane due to depolarization consequent to calcium release.
Uricase activity missing through the evolutionary process gave UA a puzzling character in the evolutionary history of Humans. Great apes have, in the basal state, high physiological levels of sUA and humans' macrophages do not respond to 200 μΜ sUA, an inflammatory condition for murine cells. Uricase inhibition therapy and the consequent elevation in the serum UA is responsible for the triggering metabolic syndrome comorbidities in murine models [52][53][54]. We have demonstrated, on the other hand, that rhesus macaque, a primate with uricase enzyme activity evolutionarily maintained, has reduced levels of IL-1β production by their monocyte-derived macrophage following sUA stimulation. It is possible that rhesus' macrophages require a higher priming activation in order to induce IL-1β transcription since LPS alone could not increase IL-1β production in these experiments. Moreover, several cytokines, including IL-1β, circulate at very low levels in both affected and unaffected rhesus macaques, in a different model of diseases [55][56][57]. It is also speculated that rhesus Naip protein was selected to tolerate elevated levels of serum UA.
In recent years, an understanding of additional adverse effects of high levels of serum UA has been advanced [58]. Early scientific literature suggested an association between uric acid concentration and incidence of cardiovascular disease, specifically, the development of hypertension [59], metabolic syndrome [60], endothelial dysfunction [61], and microalbuminuria [62]. Lifestyle and socioeconomic changes that occurred over time have resulted in a marked reduction of physical activity as well as in profound dietary changes. These changes correlate to increased rates of metabolic diseases triggered by overly active innate immune functions [63], being the chronic inflammation termed 'metaflammation' [64,65]. Furthermore, multiple genetic and non-genetic risk and protective factors are also thought to contribute to the pathogenesis of metabolic diseases, specifically those related to hyperuricemic conditions. Different states of tolerance to sUA sensing by hNaip could predict innate immune activation state in the context of hyperuricemic-related diseases. In this sense, besides understanding the humans evolutionary process, investigating which mechanisms mediate the immune modulatory function of sUA is also essential to better design rational novel antiinflammatory therapies.

Soluble uric acid preparation
Media was pre-warmed (37°C), uric acid (Ultrapure, Sigma; 200 μΜ) was added and then sterilized through 0.20 μm filters. Crystals were not detectable under these conditions (polarizing microscopy), nor did they develop during cell incubation.

Reagents
Ultrapure LPS was obtained from InvivoGen, nigericin was obtained from Invitrogen.

Cytokine profile
Cells lysates were maintained at RIPA buffer with protease inhibitors, at -80°C until dosage. IL-1β protein was measured using IL-1β (R&D Systems, Minneapolis, MN, USA), according to the manufacturer's instructions.

RNA extraction, library construction and sequencing
Total RNA was extracted from GFP-sorted THP1 cells containing lentiviral vectors NAIP1

RNA-Seq data analysis
Before read mapping, clean reads were selected after preprocessing with Trimmomatic [68] removing adapter and poly-N sequences. After cleaning, the quality of reads was checked by FastQC tool then aligned to the human genome (GRCh38/hg38) using HISAT2 aligner (V2-2.0.0) [69] considering strandness. Overall mapping quality and uniformity of read coverage on exons were checked by RSeQC tool to ensure good RNA integrity and reproducible RNA sequencing. Stringtie (v.1.3.4) [70] and Ballgown [71] algorithms were applied to identify significantly differentially expressed genes (q-value < 0.05), based on the "new Tuxedo" package [72]. Gene set enrichment analyses were performed using Enrichr tool and GAGE package using up-regulated (FC > 1 and q-value < 0.05) and downregulated genes (FC < 1 and q-value < 0.05). software (Thermo, USA). The search criteria were as follows: full tryptic specificity was required, two missed cleavage was allowed, carbamidomethylation (C) was set as the fixed modification, and the oxidation (M) was set as the variable modification, precursor ion mass tolerances were set at 10 ppm for all MS acquired in an orbitrap mass analyzer, and the fragment ion mass tolerance was set at 0,6 Da for all MS2 spectra acquired. All covariates were log-transformed before statistical analysis. All the analyses were performed using STRING software and UniProt for protein-protein interactions, identification, and statistics. P ≤0.05 was considered significant.

Oxygen consumption rates
An hour before oxygen consumption measurements, cell media was replaced by assay

Surface plasmon resonance (SPR)-based immunosensor development
An Autolab Sprit instrument (Eco Chemie B. V., The Netherlands), which presents the phenomenon of attenuated total internal reflection (Kretschmann configuration) as operation mode [73] was employed in the SPR analysis. This SPR technique is equipped with a glass prism (BK7), a planar gold SPR sensor chip, and two measurement channels protein, an aqueous solution of sUA was added to evaluate its interaction with Naip1 protein. As control assay, the interaction of palmitate with Naip1 protein was also evaluated.

Protein structure analysis
The homology models were obtained using MODELLER v.9.18 [75] with the structure 4KXF as a template [76].To fix residues with bad torsion angles, the target proteins were repaired using RepairPDB [77], and Chimera [78] was used to add hydrogen atoms and charges where adequate. Surface electrostatic potentials were calculated using the AMBER force field implemented on APBS [79], taking as input files converted on PDB2PQR [80]. For each NAIP structure, a blind docking was performed on SwissDock [81], having the uric acid on its ionized form (Urate) as a ligand (ZINC AC: 2041003).
During the docking, the surfaces of both proteins were scanned for putative binding pockets in more than 250 iterations. Several low energy ligand clusters with similar binding modes (poses) were found. All poses showing ∆G < -6 kcal/mol were considered in further analyses, and those showing the lowest energies were selected as the best representation of the binding between urate and both human and murine Naip.

Statistics
Experiments were performed in duplicate or triplicate and at least two independent tests were performed for each assay. The data were described in terms of the mean and S.E.M.
unless specified in the figure legend. Differences between groups were compared using ANOVA (with Tukey's post-test) and Student's t-test. Significant differences were regarded as p<0.05, p<0.01 or p<0.001, according to the figure. All statistical analyses were performed using GraphPad PRISM 6.01 (La Jolla, CA, USA).