Dispensability of lipid catabolism in lifespan extension via fasting
Oscillating between catabolic and anabolic states is essential for cellular and organismal adaptation in an ever-changing environment. To model the breakdown of the animal’s main energy reserve, we measured how triglycerides (TAGs) stored in intestinal lipid droplets17 were altered after 24 h of fasting compared to a previously reported 48-h fast18. Thin-layer chromatography (TLC) revealed a 61% depletion in TAG levels at 24 h after fasting versus a 91% depletion after 48 h (Fig. 1a). Despite a more robust decrease in TAG levels at 48 h, the stark depletion of cholesterol raised concerns about the maintenance of membrane integrity and fluidity. Thus, the 24-h fast was selected for further experimentation as this regimen induced significant TAG catabolism and preserved cholesterol levels. Triglycerides are predominantly composed of long-chain fatty acids conjugated to their glycerol backbone19,20,21. Despite no apparent changes in total free fatty acids after 24 h of fasting by TLC, lipidomic profiling revealed selective reduction in specific long-chain fatty acids, including the α- and γ-linolenic acid, as well as oleic acid (Fig. 1b). In summary, fasting C. elegans for 24 h triggers significant catabolism of TAGs and select long-chain fatty acids without altering the total pool of free fatty acids or cholesterol.
Fig. 1: Dispensability of NHR-49 for fasting-induced lifespan extension. The alternative text for this image may have been generated using AI. Full size image a Representative image of a TLC plate showing the lipid profiles of fed worms (control) and worms fasted for 24 and 48 h. Lanes 1-5 are standards in the form of TAGs, FFAs, and Cholesterol. n = 4 for each condition. b Heatmap depicts relative levels of long chain neutral fatty acids from C. elegans. Lipidomic analysis compares fed Day 1 adults to 24 h of fasting and subsequent refeeding for 24 h. n = 4. c Fluorescence micrographs of C. elegans intestinal cells ectopically expressing the lipid droplet localized dehydrogenase, DHS-3::GFP under three conditions: Day 1 fed, Day 2 after 24-h fasting, and Day 3 after 24-h refeeding. Scale bars = 20 µm and 10 µm (zoom). n = 10 per group. d Scatter plot from large-particle flow cytometry of transgenic worms expressing DHS-3::GFP shows relative fluorescence per individual animal. Plot compares Day 1 adults under fed conditions to 24 h of fasting and subsequent refeeding for 24 h. Mean ± 95% CI. n = 1882, 3161, and 1538 from left to right, Kruskal-Wallis test and Dunn’s multiple comparisons test used for statistics p. e Lifespan analysis of control worms (N2) under fed or fasted conditions (24 h of dietary deprivation at Day 1 of adulthood). p < 0.0001 on 3 biological replicates determined by log-rank (Mantel-Cox) test. Additional statistics in Supplementary Table 1. f Heatmap depicts relative transcriptional changes in lipid metabolism annotated genes. Analysis compares Day 1 adults under fed conditions to 24 h of fasting and subsequent refeeding for 24 h. Genes categorized by catabolic or anabolic based on their prominent annotated role in metabolism. g Relative abundance of acs-2 transcription determined by reverse transcription quantitative PCR (RT-qPCR). Analysis compares wild-type (N2) and nhr-49(nr2041) mutant worms at Day 1 of adulthood under fed conditions to 24 h of fasting and subsequent refeeding for 24 h. Mean ± SEM. n = 3, ordinary one-way ANOVA and Tukey’s multiple comparisons test used for statistics. h Scatter plot from large-particle flow cytometry of nhr-49(nr2041) mutant worms expressing DHS-3::GFP shows relative fluorescence per individual animal. Plot compares Day 1 adults under fed conditions to 24 h of fasting and subsequent refeeding for 24 h. Mean ± 95% CI. n = 2436, 3072, and 430 from left to right, Kruskal–Wallis test and Dunn’s multiple comparisons test used for statistics. i Lifespan analysis of nhr-49(nr2041) mutant worms under fed or fasted conditions (24 h of dietary deprivation at Day 1 of adulthood). p < 0.0001 on 3 biological replicates determined by log-rank (Mantel–Cox) test. Additional statistics in Supplementary Table 1.
Next, we evaluated the animal’s capacity to recover after fasting and restore lipid homeostasis upon dietary replenishment (see Supplementary Fig. 1a). Fasted worms were reintroduced to food for an additional 24 h, referred to as refeeding, and this was sufficient to restore free fatty acid profiles to pre-fasted states (Fig. 1b and Supplementary Fig. 1b). Moving forward, we used a fluorescence-based system to monitor, in real time, the dynamics of TAG-enriched lipid droplets in living animals during fasting and refeeding. We selected a well-characterized transgenic C. elegans strain expressing DHS-3::GFP, a lipid droplet-resident short-chain dehydrogenase fused to green fluorescent protein (Fig. 1c)22,23,24,25,26,27. Using the 24-h fasting and refeeding paradigm (Supplementary Fig. 1a), lipid droplet dynamics were monitored in relation to overall DHS-3::GFP fluorescence across a given worm population. Relative to Day 1 fed animals, the average volume and number of DHS-3::GFP positive lipid droplets after a 24-h fast were reduced by 84% and 65%, respectively, as determined by fluorescence microscopy (Fig. 1c and Supplementary Fig. 1c, d). This reduction in average volume and number corresponded to a 71% decrease in DHS-3::GFP fluorescence across the worm population as well as the 61% decrease in TAG levels observed by TLC (Fig. 1a, d). After refeeding for 24 h, both average volume and number of DHS-3::GFP positive lipid droplets recovered (Fig. 1c and Supplementary Fig. 1c, d), which correlated with the restoration of population-wide DHS-3::GFP fluorescence to pre-fasted levels upon refeeding (Fig. 1d). Thus, nutrient replenishment for 24 h after fasting enables restoration of neutral lipid levels.
The maintenance of cellular energetics is critical for survival during fasting periods28. As with mammalian adipose tissue, the primary energy reserves in C. elegans are stored within intestinal TAG-enriched lipid droplets29. During fasting, hydrolysis of TAGs liberates free fatty acids, which must be distributed to metabolically demanding peripheral tissues to be further catabolized via mitochondrial β-oxidation for adenosine triphosphate (ATP) production30. We sought to understand how animals manage energy homeostasis by redistributing resources provided by lipid catabolism during fasting and refeeding. To investigate this, we first examined transcriptional changes in mitochondrial-annotated genes. Gene pathway analysis after a 24-h fast revealed the transcriptional activation of genes involved in ATP biosynthetic process and a simultaneous repression of genes involved in energetically costly processes like mitochondrial transport and organization (Supplementary Fig. 1f). This coordinated response was fully reversed upon refeeding, demonstrating a rapid and reversible form of metabolic adaptation (Supplementary Fig. 1g). We then focused on the body-wall muscle, arguably the most energetically demanding tissue during fasting. Mitochondria in this tissue exhibit a higher membrane potential than mitochondria from other tissues and are poised to rapidly meet high ATP demands such as the dramatic increase in muscle contractions needed for fasting-induced foraging31. Using the established fluorescent reporter myo-3p::Queen-2m to monitor ATP levels in the body wall muscle32, we detected a mild ~6% decline in ATP levels after 24 h of fasting, a level that recovered upon refeeding (Supplementary Fig. 1h). This ATP maintenance, coupled with sustained foraging activity (Supplementary Fig. 1i), demonstrates that these animals are capable of maintaining energy homeostasis and tissue function during the 24-h fast. Yet, despite their ability to sustain energetics and functionality in critical tissues, animals subjected to this fasting-refeeding cycle displayed lasting physiological changes, including a reduction in body size that persisted well into later life (Supplementary Fig. 1j, k). Our findings suggest that C. elegans exhibits significant metabolic plasticity, in which the modulation of lipid stores effectively accommodates energetic demand during periods of nutrient stress and subsequent recovery.
Dietary restriction is the most evolutionarily conserved method of lifespan extension10. Similar to other reported fasting and refeeding paradigms in C. elegans18,33, we observe that 24 h of fasting in early adulthood was sufficient to extend median lifespan by 40.8% (Fig. 1e) and promote youthfulness, as evidenced by increased motility at older ages (Supplementary Fig. 1i). Analysis of the differentially regulated genes during fasting highlighted fatty-acid degradation as the most affected pathway (Supplementary Fig. 2a, b). Furthermore, fasting-induced fluctuations in lipid metabolism genes were restored upon refeeding. This restoration, similar to that of the mitochondrial-specific transcripts, suggests that the observed changes in lipid droplet availability are mediated, in part, by the transcriptional changes in these key metabolic enzymes (Fig. 1f). Using this fasting and refeeding paradigm, we next investigated the relative importance of lipid mobilization and restoration in fasting-induced longevity.
A substantial body of work establishes that the nuclear hormone receptor, NHR-49, is essential for activating and repressing the expression of genes involved in β-oxidation during starvation and plays a critical role in animal physiology and lifespan determination16. However, NHR-49’s ability to recover after fasting, and its subsequent impact on animal physiology and lifespan extension remained poorly understood. To this end, we first investigated whether NHR-49 activity could recover upon refeeding. Focusing on direct transcriptional targets of NHR-49 with its established roles in lipid metabolism34,35,36,37,38,39,40,41, we confirmed that fasting-induced activation of the mitochondrial medium-chain acyl-CoA ligase (acs-2) and repression of the stearoyl-CoA desaturase (fat-7) were both abrogated in the loss-of-function nhr-49(nr2041) mutant animals (Fig. 1g and Extended Fig. 2c, d)34,42. In the same nhr-49(nr2041) mutants, we still observed a significant reduction in the levels of DHS-3::GFP positive lipid droplets during fasting, accompanied by a mild impairment in lipid droplet restoration upon refeeding (Fig. 1h). Therefore, TAG hydrolysis appears independent of NHR-49, which is further supported by the unaltered expression of critical lipolysis regulators, hormone sensitive lipase, hosl-1, and the adipose triglyceride lipase, atgl-1, in nhr-49(nr2041) mutant animals (Supplementary Fig. 2e). Since transcriptional dynamics dependent on NHR-49 appeared specific for genes involved in β-oxidation, we hypothesized that defective oxidation of the fatty acids would impair ATP maintenance in the nhr-49(nr2041) mutants. Indeed, ATP availability in these mutant animals was reduced by roughly threefold the levels observed in wild-type animals after fasting (Supplementary Fig. 2f). Despite deficits during fasting, ATP levels in mutant animals still recovered upon nutrient replenishment, indicating that other mechanisms may compensate to restore ATP levels upon refeeding. Thus, while transcriptional changes mediated by NHR-49 correspond with the maintenance of energy homeostasis during fasting, animals possessed the capacity to restore their ATP levels upon nutrient replenishment.
Fig. 2: Ligand-independent regulation of NHR-49. The alternative text for this image may have been generated using AI. Full size image a Schematic of NHR-49 construct containing a deletion of amino acids 295–422 (Δ295–422). b Fluorescent micrographs of transgenic worms in L3-L4 larval stages ectopically expressing NHR-49::YFP (full length) or the ligand binding truncation (Δ295-422) in intestinal cells. Scale bar = 25 μm. c Quantification of cytosolic NHR-49::YFP by visual inspection. n = 58 and 51 from left to right over 3 independent trials. Unpaired t-test (two-tailed) used for statistics. d Relative abundance of acs-2 transcription determined by RT-qPCR. Analysis compares nhr-49(nr2041) mutant worms ectopically expressing NHR-49::YFP full-length or Δ295-422 in the intestine. Mean ± SEM. n = 3 independent trials, ordinary one-way ANOVA with Tukey’s multiple comparisons test used for statistics. e Micrographs of stained neutral fatty acids by Oil-Red-O staining in nhr-49(nr2041) mutant worms ectopically expressing intestinal NHR-49::YFP full-length or Δ295-422 at Day 1 of adulthood. f–h Relative abundance of acs-2 transcription determined by RT-qPCR. Analysis compares nhr-49(nr2041) mutant worms ectopically expressing NHR-49::YFP full-length or Δ295-422 in the intestine during f fed, g fasting, and h refeeding conditions. Mean ± SEM. n = 3 independent trials, ordinary one-way ANOVA with Tukey’s multiple comparisons test used for statistics. i Predicted secondary structure of NHR-49 modeled using AlphaFold2. Stick rendition of serine residue 114 (red). j Quantification of cytosolic NHR-49::YFP by visual inspection. n = 31, 35, and 64 from left to right over 2 independent trials. Ordinary one-way ANOVA with Šídák’s multiple comparisons test used for statistics.
We next investigated whether metabolic deficits associated with nhr-49(nr2041) mutants impact fasting-induced longevity. While the mutants exhibited reduced lifespan under ad libitum feeding conditions36, they demonstrated a notable 57.1% extension in lifespan when subjected to a 24-h fast at Day 1 of adulthood (Fig. 1i). These fasted mutants also exhibited physiological changes similar to those observed in wild-type fasted animals, including reduced body size and enhanced motility (Supplementary Fig. 2g–i). These data suggest that activation of β-oxidation is dispensable for lifespan extension via fasting. To further investigate the role of lipid catabolism in fasting-induced longevity, we examined two key regulators: Adipose Triglyceride Lipase-1 (ATGL-1), which initiates lipolysis, and Carnitine Palmitoyl Transferase-2 (CPT-2), which catalyzes carnitine removal and Coenzyme A conjugation to the fatty acid within the mitochondrial matrix43,44. Similar to the nhr-49(nr2041) mutant, fasting-induced lifespan extension was still observed in wild-type animals treated with atgl-1 and cpt-2 RNAi (Supplementary Fig. 3a, b). Moreover, administering atgl-1 RNAi in the nhr-49(nr2041) background did not alter this fasting-induced longevity effect (Supplementary Fig. 3c). Thus, disrupting multiple key steps in the lipid catabolism pathway had little impact on fasting-induced lifespan extension, indicating that the breakdown and utilization of lipids is not needed to confer the longevity benefits associated with fasting.
Stress-induced ligand-independent regulation of NHR-49
Activating lipid catabolism appeared dispensable for fasting-induced lifespan extension. Therefore, we hypothesized that silencing this response upon refeeding plays a more significant role in longevity by restoring and preserving long-term lipid homeostasis. However, little is understood about how NHR-49 and its fasting-induced transcriptional response are attenuated upon dietary replenishment. While the activity of NHR-49 is regulated at multiple levels including cofactor interactions, heterodimerization with other nuclear receptors, and isoform-specific subcellular localization23,42,45,46,47,48, nuclear hormone receptors are classically defined as ligand-regulated transcription factors49,50,51,52,53. Thus, we first examined the role of ligand binding as a means of regulating NHR-49 dynamics during fasting and refeeding. As an orphan receptor, no endogenous ligand has been identified for NHR-49. Previous studies identified an activating mutation within the ligand binding domain of NHR-49 at valine 411 where this hydrophobic valine was mutated to a charged glutamic acid54,55. Molecular modeling indicates that the substitution of a valine to glutamic acid at residue 411 reduces the binding affinity of linolenic acid, a reported exogenous ligand for the NHR-49 ortholog, HNF4α, as well as other potential fatty acid binding partners with comparable structure (Supplementary Fig. 4a, b)56,57. Due to the lipid-based nature of these putative ligands, we reasoned that nutrient replenishment would restore their intracellular abundance, thereby inactivating NHR-49.
While our efforts to define the endogenous ligand for NHR-49 were unsuccessful, mutating the ligand-binding pocket might allow us to determine the role of ligand binding in the transcriptional activity of NHR-49. Based on domain architecture comparison with HNF4α58,59, we engineered an NHR-49 truncation (∆295–422) to disrupt ligand binding by removing a majority of the ligand binding domain, while retaining its analogous N-terminal, DNA-binding, hinge, and C-terminal domains (Fig. 2a). Since the intestine is the primary site of neutral lipid storage in C. elegans60, we generated both a full length and ∆295–422 truncated form of NHR-49::YFP isoform C under the control of the intestinal-specific pept-1 promoter (Supplementary Fig. 4c, k). All experiments utilizing these ectopically overexpressed transgenes were performed in nhr-49(nr2041) mutant backgrounds to rule out contributions from the endogenous NHR-49 gene. Importantly, the full-length NHR-49::YFP fusion protein restored median lifespan of the nhr-49(nr2041) mutant to levels typically observed in wild-type animals, and fasting for 24 h further prolonged lifespan (Supplementary Fig. 4d). Therefore, the full-length isoform C transgene can functionally rescue the nhr-49(nr2041) mutant.
We next investigated the NHR-49::YFP ∆295–422 truncation to better understand how impaired ligand binding impacts its subcellular dynamics and transcriptional activity during fasting and refeeding. Consistent with other reports regarding C-terminal truncations of its ortholog HNF4α61, truncating the ligand-binding domain of NHR-49 altered its nucleocytoplasmic distribution, favoring a more pronounced nuclear localization under ad libitum feed conditions when compared to the full-length receptor (Fig. 2b, c). Despite its increased nuclear occupancy, the ∆295–422 receptor failed to maintain baseline expression of NHR-49 transcriptional targets, acs-2 and fat-7 (Fig. 2d and Supplementary Fig. 4e) and was incapable of phenotypically rescuing nhr-49(nr2041) mutant defects with respect to lipid deposition and fecundity (Fig. 2e and Supplementary Fig. 4f, g). Although atypical, other groups have reported ligand-independent means of nuclear hormone receptor regulation62, and we further investigated whether this might be the case for NHR-49 during fasting conditions. Indeed, ectopic expression of the ∆295–422 receptor displayed similar transcriptional dynamics as the full-length receptor when subjected to fasting and refeeding as evidenced by the activation and subsequent recovery of acs-2 transcription (Fig. 2f–h). Thus, unlike ad libitum fed conditions, ligand binding was dispensable for both the activation and attenuation of NHR-49 upon fasting and refeeding. Furthermore, these NHR-49-mediated transcriptional fluctuations originated within the intestine.
Nuclear hormone receptors possess several ligand-independent modalities of regulation63. For instance, these receptors, including NHR-49, can alter their activity through heterodimerization with other nuclear receptors46. However, the inability of the ∆295–422 receptor to rescue baseline transcription (Fig. 2d and Supplementary Fig. 4e) and lipid-related phenotypes (Fig. 2e and Supplementary Fig. 4f, g) under ad libitum fed conditions indicates that it is not functioning through heterodimerization with other nuclear receptors in the intestinal cell. In further support, no nuclear receptor binding partners were detected by mass spectrometry of either the NHR-49::YFP full-length or ∆295–422 immunoprecipitations (see Supplementary Data 1). Alternatively, post-translational modifications have the potential to impact nuclear receptor function62. Starvation-induced phosphorylation of HNF4α by Protein Kinase A decreases DNA binding and reduces transcriptional activity64. While previous studies hint at possible kinase-regulated mechanisms for NHR-49, they lack residue-level specificity16. Utilizing Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS), we performed post-translational modification analysis on NHR-49::GFP immunoprecipitations42 under fed and fasted states and identified a single phosphorylation event with high confidence within the receptor’s hinge region at serine 114 (Fig. 2i and Supplementary Fig. 4h). This S114 phosphorylation was further confirmed by proximity labeling with the biotin ligase tag, TurboID, which was inserted, in independent worm strains, at both the N- and C-terminus of the endogenous NHR-49 gene48. All proteins within an approximate radius of 10 nm65 to either tagged NHR-49 fusion protein were biotinylated, affinity-purified, and subjected to LC-MS/MS48. Following post-translational modification analysis, we again detected phosphorylation at S114, with 15 PSMs in both the N- and C-terminal TurboID strains, ranking it as the 5th most abundant PTM overall. We also identified a phosphorylation site at S131 with a PSM of 2, observed only in the C-terminal strain. An additional low-confidence phosphorylation was detected between residues 110 to 128, though it could not be precisely localized, potentially due to its transient nature (Supplementary Fig. 4i). Thus, the serine residue at position 114 of NHR-49 and potentially others within the hinge-region of NHR-49 were detected across multiple experiments which used both ectopically overexpressed and endogenously tagged forms of NHR-49.
While HNF4α possesses several serine residues within its analogous hinge region, its paralog, HNF4γ, shares more sequence similarity to NHR-49 and exhibits a higher degree of conservation at the respective serine residue and its surrounding motif (Supplementary Fig. 4j). With respect to physiology, phosphorylation and cytosolic localization of HNF4α are linked with liver decompensation and cirrhosis, however evidence is lacking on the effect of phosphorylated HNF4γ66. Based on the supporting evidence in mammals, we sought to understand the significance of this putative phosphorylation event in worms. To this end, we first abrogated phosphate conjugation by mutating the serine residue at the 114 position to an alanine (S114A) within the ectopically expressed NHR-49::YFP transgene. The subcellular distribution of the S114A mutant favored a more pronounced nuclear localization with significantly less fluorescence in the cytosol compared to the full-length transgene (Fig. 2j and Supplementary Fig. 4k, l). In a complementary manner, an activating phosphomimetic mutation was engineered to permanently mimic the phosphate’s bulky negative charge by replacing the serine with an aspartic acid (S114D). This phosphomimetic, S114D mutant, in NHR-49::YFP strongly favored a cytoplasmic localization, which was dramatically different from the wild-type and more so from the phospho-dead S114A mutant (Fig. 2j and Supplementary Fig. 4l). These results are consistent with the reported role of hinge-region serine phosphorylation in the cytoplasmic retention of nuclear hormone receptors67.
NHR-49 phosphorylation by the casein kinase KIN-19
After identifying a putative phosphorylation site at serine 114 and confirming its significance in controlling the subcellular dynamics of NHR-49 through mutational analysis, we next aimed to define the kinase responsible for this post-translational modification. To this end, we employed a multi-pronged strategy, which involved proteomics analysis, genetic screening, in silico binding, sequence motif validation, and in vitro confirmation. Building upon our extensive proteomics data sets, our initial analysis sought to filter kinases that were repeatedly detected in complex with and/or in proximity to NHR-49. We cross-referenced coimmunoprecipitation and proximity labeling data sets, which accounted for three different NHR-49 strains including the overexpressed all-tissue NHR-49::GFP under fed and fasted conditions42 and the endogenous N- and C-terminal TurboID tagged NHR-4948. From these datasets, 96 kinases were detected in immunoprecipitations of overexpressed NHR-49::GFP in all tissues and 41 in the proximity labeling experiments with N- and C-terminal TurboID tagged NHR-49. Of these combined 137 kinases, 22 were mutually detected in both experiments (Supplementary Fig. 5a). Despite the limitations associated with identifying bona fide protein-protein interactions by coupling co-immunoprecipitation or proximity labeling with LC/MS-MS, we were able to confidently generate a shorter, more manageable list of 19 potential interacting kinases, predominantly intestinal, for further study.
Next, we coupled AI-based predictive modeling with RNAi screening to filter kinases with the potential to directly bind NHR-49 and alter its activity. First, we computationally modeled the interactions between NHR-49 and each of the 19 kinases, assuming a one-to-one stoichiometric ratio (see Supplementary Table 2). With an interface predicted templating model (ipTM) score greater than 0.6 as our confidence threshold, none of the modeled kinase/NHR-49 interactions scored over 0.5. While lacking confidence in an interaction does not preclude a genuine association, it highlights that no such interaction between these kinases and similarly structured nuclear receptors has been previously reported. In parallel, we employed a more labor-intensive, candidate-based RNAi screen in vivo to determine whether reduced expression of individual kinases impaired NHR-49 associated phenotypes. Complementary results from phospho-mimetic (S114D) and phospho-dead (S114A) forms of NHR-49::YFP indicate that S114 phosphorylation antagonizes nuclear residency of the receptor (Fig. 2j and Supplementary Fig. 4l). Therefore, we hypothesized that hindering S114 phosphorylation by RNAi depletion of a kinase would enhance the transcriptional activity of NHR-49. To test this, we used transgenic animals expressing the transcriptional rab-11.2p::YFP reporter, which serves as an effective fluorescent diagnostic of NHR-49 activation, though seemingly not a direct transcriptional target23,42. Exhibiting a broad range of fluorescent detection and ranking among the most activated genes during nutrient deprivation, we leveraged this reporter to systematically screen all 19 kinases. Two enzymes were identified whose reduced expression activated the rab-11.2 fasting reporter KIN-2, a cAMP-dependent protein kinase β-subunit (PKA), and to a much greater extent, KIN-19, the ortholog of casein kinase 1 alpha 1 (CSNK1α1) (Fig. 3a and Supplementary Table 3). Through complementary approaches involving the analysis of multiple proteomics datasets and subsequent functional RNAi screening, we identified two potential kinases as candidate regulators. However, these candidates were absent from the initial in silico interaction networks, prompting us to investigate this discrepancy.
Fig. 3: NHR-49 phosphorylation by casein kinase, KIN-19. The alternative text for this image may have been generated using AI. Full size image a Scatter plots from large-particle flow cytometry of transgenic worms expressing rab-11.2p::YFP show relative fluorescence per individual animal. Plots compare Day 1 adults under empty vector control conditions to the respective RNAi conditions. Mean ± 95% CI. n = 1067, 664, 1017, 755, 563, 408, 333, 1033, 131, 536, 799, 1023, 555, 631, 1036, 853, 566, 633, 521 and 886 from left to right, ordinary one-way ANOVA with Dunnett’s multiple comparisons test used for statistics (see Supplementary Table 3). b Predicted secondary structure and intermolecular interface of KIN-19 and NHR-49, modeled using AlphaFold3, an advanced AI-driven deep learning algorithm for protein structure and interaction prediction. Surface density for KIN-19 (orange) and NHR-49 (teal) with stick rendering of regulatory serine residues. c Quantification of γ-phosphate transferred from [32P] ATP to either wild-type (blue) or pS114 (teal) primed NHR-49 polypeptides by recombinant CSNK1A1 over 10 min, taken over three independent repeats. A non-linear fit to each line. Error bars represent SEM. d Relative abundance of acs-2 transcripts determined by RNA-sequencing. Analysis compares Day 1 adults on an empty vector control and kin-19 RNAi. Mean ± SEM. n = 4, statistics represent Differential Expression Analysis in Two Groups with FDR correction (Qiagen CLC Workbench v9.5). e Venn diagram highlighting the overlap between transcriptional changes observed in animals treated with kin-19 RNAi and those fasting for 24 h. f Heatmap depicts relative transcriptional changes in genes previously reported to be partially regulated by NHR-49. Analysis compares Day 1 adults on an empty vector control and kin-19 RNAi conditions.
Protein kinases exhibit distinct and highly specific consensus binding motifs that dictate their enzymatic phosphorylation activity. To investigate the phosphorylation of NHR-49’s S114 residue, we compared the consensus motifs of the highly conserved human orthologs of KIN-2 and KIN-19. Although neither kinase was predicted to phosphorylate S114 directly, a highly conserved sequence neighboring S114 was strongly predicted to be CSNK1α1 (KIN-19) phosphorylation sites (Supplementary Fig. 5b). Casein kinases are a broad family of serine/threonine kinases involved in diverse cellular processes, including signal transduction, circadian rhythm regulation, and metabolic pathways68,69. Substrates of casein kinase typically require a priming phosphorylation event to catalyze phosphate addition70,71,72. Sequence motif analysis indicated that KIN-19 would preferentially phosphorylate the conserved serine residues adjacent to S114 at positions S117 and S120 (Supplementary Fig. 5b)73. Thus, phosphorylation at S114 may act to prime NHR-49 for further phosphate addition by KIN-19. In support, AI-based structural algorithms74 now predicted a significant interaction between KIN-19 and S114-phosphorylated NHR-49 with an ipTM of 0.69 (Fig. 3b, Supplementary Fig. 5c, d and Supplementary Table 2). To determine whether this was specific for KIN-19, we tested the other 18 kinases that were previously screened against the primed S114 and found only KIN-19 as a significant predicted interaction. In summary, a comprehensive analysis of 19 kinase candidates, which integrated LC-MS/MS data from several co-immunoprecipitations, in silico modeling, RNAi screening, and consensus sequence validation, identified KIN-19 as the most promising kinase candidate for further investigation.
Casein kinases are reported to act on a wide range of molecular targets, and we sought to determine whether KIN-19 could physically modify NHR-49. Using the C. elegans KIN-19 ortholog, CSNK1A1, we conducted [32P] kinase-substrate in vitro assays. The human ortholog was chosen for its high sequence identity (87%, 274/315) and conserved secondary structure (RSMD = 0.220), ensuring functional equivalence to KIN-19 (Supplementary Fig. 5e). This assay utilized recombinant human CSNK1A1 and peptides corresponding to the hinge region of NHR-49 as the substrate to test for direct phosphorylation. Consistent with in silico modeling, CSNK1A1 exhibited basal phosphorylation activity on the unprimed NHR-49 peptide (k m = 106.6; V max = 10.98). However, pre-phosphorylation of the NHR-49 peptide at the corresponding S114 (pS114) position markedly enhanced transfer of the γ-phosphate from [32P] ATP onto the primed peptide (k m = 285.7; V max = 68.63) (Fig. 3c and Supplementary Fig. 5f). Pre-phosphorylation of the serine residue resulted in a 2.68-fold increase in the Michaelis constant (K m ) and a 6.25-fold increase in the maximum reaction rate (V max ) when compared to the unprimed site. This observation is consistent with previously reported mechanisms of the casein kinase 1 family, which are capable of phosphorylating non-primed substrates but do so far more efficiently when an upstream priming phosphorylation is present75,76,77. Specifically, casein kinase 1 family members are reported to act on substrates via sequential phosphorylation of serine/threonine residues following a canonical pSxx(S/T) motif. This priming-dependent enhanced phosphorylation aligns with our in vivo post-translational modification analysis: S114 phosphorylation is consistently detectable, whereas downstream residues S117 and S120 remain elusive, likely due to the transient nature of their phosphorylation following S114 priming.
To identify the kinase responsible for phosphorylating S114, we revisited our analytical pipeline. The second kinase that activated our fasting reporter, KIN-2, is the ortholog of PKA, which has previously been shown to phosphorylate HNF4α during fasting64. However, we did not detect KIN-2 as a conserved interactor in our co-immunoprecipitation assays, and its consensus phosphorylation motif does not align with residues surrounding S114. Conversely, while MPK-1’s consensus motif closely matches the residues surrounding S114, it failed to activate our fasting reporter (Fig. 3a). Despite KIN-19 being the only kinase to satisfy our initial criteria, we performed in vitro kinase assays using [32P] ATP with the human orthologs of both KIN-2 (PKA) and MPK-1 (ERK1). Both kinases yielded non-significant phosphorylation results for the peptide corresponding to the hinge region of NHR-49 (Supplementary Fig. 5g–j). These findings validated our initial selection, leading us to focus on the further characterization of the KIN-19/NHR-49 phosphorylation event.
Given the essentiality of KIN-19, we focused our investigation on its functional role. Repeated attempts to engineer an enzymatically dead kinase via a D135A mutation were unsuccessful, strongly suggesting that this kinase and its enzymatic activity are vital for animal viability. Therefore, we proceeded with RNAi to knockdown KIN-19 expression. We confirmed the knockdown efficiency with a 79% reduction in transcript levels and approximately a 95% reduction in steady-state protein levels of endogenous KIN-19, as determined by RNA-sequencing and whole-worm proteomics, respectively (Supplementary Fig. 5k, l). Animals treated with kin-19 RNAi exhibited a gene expression profile characteristic of fasting-induced NHR-49 activation, including the differential regulation of established NHR-49 targets, acs-2 and fat-7 (Fig. 3d and Supplementary Fig. 5m). Gene Ontology (GO) analysis of differentially regulated genes revealed that the most significantly altered categories were related to lipid transport and metabolism (Supplementary Fig. 5n). Cross-referencing the transcriptional changes induced by kin-19 RNAi with those induced by fasting showed a significant overlap in genes differentially regulated between the two conditions, including 20 partially dependent NHR-49 targets beyond acs-2 and fat-7 (Fig. 3e, f). Based on these transcriptional signatures, impairing KIN-19 expression appeared to mimic a chronically-fasted state in which NHR-49 is hyperactivated. While the precise nature of the priming event at S114 remains unclear, our findings demonstrate that the casein kinase ortholog, KIN-19, is capable of phosphorylating adjacent serine residues in the hinge region of NHR-49. This ultimately reveals that impairing KIN-19 function via RNAi mimics a perpetually fasted response.
Silencing lipid catabolism is required for fasting induced longevity
Next, we investigated the impact of KIN-19 gene silencing on NHR-49 regulation and its subsequent effects on energy homeostasis, animal physiology, and age determination. Previous research links the worm casein kinase 1A1 to oocyte development, asymmetric cell division, and embryogenesis78,79; while its function in adulthood and more particularly in metabolism and aging, remains largely unexplored. Though some studies report that KIN-19 overexpression can lead to extensive protein aggregation in older worms80,81,82, our analysis of Day 1 adult worms revealed no evidence of endogenous KIN-19 aggregates (Supplementary Fig. 6a). Our study found that kin-19 RNAi in adult worms resulted in several physiological changes consistent with chronic fasting. We observed a significant reduction in lipid deposition and body size, traits typical of fasted animals (Fig. 4a–c and Supplementary Fig. 6b). Additional transcriptomic analysis of lipid metabolism or mitochondrial-associated genes revealed similar transcriptional profiles between kin-19 RNAi treated and fasted animals (Supplementary Fig. 6c, d). These findings demonstrate that silencing KIN-19 triggers the physiological hallmarks of fasting even when animals are under ad libitum conditions, indicating that KIN-19 plays a critical role in mediating the fasting-induced response impacting both metabolic and physiological processes.
Fig. 4: Silencing lipid catabolism is required for fasting induced longevity. The alternative text for this image may have been generated using AI. Full size image a Scatter plots from large-particle flow cytometry of transgenic worms expressing DHS-3::GFP show relative fluorescence per individual animal. Plots compare Day 1 adult animals on empty vector control or kin-19 RNAi under fed conditions or after 24 h of fasting. Mean ± 95% CI. n = 1855, 953, 1774, and 797 from left to right, ordinary one-way ANOVA with Šídák’s multiple comparisons test used for statistics. b Micrographs of stained neutral fatty acids by Oil-Red-O staining in wild-type, Day 1 adult worms on empty vector control or kin-19 RNAi. n = 22 for empty vector control, n = 19 for kin-19 RNAi. c Fluorescence micrographs of C. elegans intestinal cells ectopically expressing the lipid droplet localized dehydrogenase, DHS-3::GFP. Micrographs compare kin-19 RNAi treated worms under fed or fasting conditions (24 h). Scale bar = 50 µm. d Recovery of DHS-3::GFP fluorescence upon refeeding. Graph compares transgenic worms treated with empty vector control or kin-19 RNAi. Mean ± 95% CI. n = 736 and 927 from left to right, unpaired t-test (two-tailed) was used for statistics. e–g Lipidomic analysis compares the relative abundance of different long chain neutral fatty acids as a percentage of the total TAG content from worms on empty vector control and kin-19 RNAi under fasting and refeeding conditions. e palmitic acid, f oleic acid, and g γ-linolenic acid. Mean ± SEM. n = 4. Šídák’s multiple comparisons test used for statistics. h Oxygen consumption rate of Day 3 N2 worms on control or kin-19 RNAi after a 24-h fast, followed by a 24-h refeed. Mean ± SEM. n = 10 per trial over three independent repeats, unpaired t-test (two-tailed) used for statistics. i Scatter plots from large-particle flow cytometry of transgenic worms expressing DHS-3::GFP show relative fluorescence per individual animal. Plots compare Day 1 adult wild-type or nhr-49(nr2041) mutants on an empty vector control or kin-19 RNAi. Mean ± 95% CI. n = 6748, 6248, 2452, and 2446 from left to right, ordinary one-way ANOVA with Šídák’s multiple comparisons test used for statistics. j Relative abundance of acs-2 transcription determined by RT-qPCR. Analysis compares Day 1 adult wild-type or nhr-49(nr2041) mutant worms on empty vector control or kin-19 RNAi. n = 3, ordinary one-way ANOVA with Šídák’s multiple comparisons test used for statistics. k Lifespan analysis of wild-type (N2) worms on empty vector or kin-19 RNAi under fed or fasted conditions (24 h of dietary deprivation at Day 1 of adulthood). Based on 3 biological replicates, log-rank (Mantel–Cox) test used for statistics (see Supplementary Table 1).
Upon analysis of the kin-19 RNAi-dependent metabolic response, treated animals lacked the metabolic plasticity that was observed in wild-type or nhr-49(nr2041) animals. Specifically, while fasting typically reduces body size, it had no effect on kin-19 RNAi-treated animals on Day 3 of adulthood, with even a slight increase in body size by Day 7 (Supplementary Fig. 6e, f). Despite starting with significantly lower lipid reserves, kin-19 RNAi treated animals displayed no additional decrease in lipid availability after 24 h of fasting (Fig. 4a, c). This was in sharp contrast to the pronounced lipid depletion observed in both wild-type and nhr-49(nr2041) animals (Fig. 1d, h). Moreover, these worms lack the ability to fully restore lipid droplet levels after refeeding, recovering only 58.27% of their pre-fasted levels (Fig. 4d). Lipidomic profiling exasperated this lack of metabolic plasticity, revealing that several long-chain neutral fatty acids, namely palmitic, oleic, and γ-linolenic acids exhibited defective transitions during fasting in kin-19 RNAi treated animals (Fig. 4e–g). Collectively, this evidence indicates that KIN-19 is essential for metabolic flexibility, with its absence mimics a permanent fasting state that severely impairs the animal’s ability to adapt to nutrient changes.
To understand how these changes in lipid dynamics affect cellular energy homeostasis, we directly measured mitochondrial respiration by quantifying the oxygen consumption rates (OCR) in live animals under fasting and refeeding. Worms treated with kin-19 RNAi consumed 50.4% less oxygen than control animals (average 61.3 pmol O 2 /min vs. 121.7 pmol O 2 /min, respectively), confirming a significant impairment in their bioenergetic capacity (Fig. 4h). To further investigate, we assessed mitochondrial morphology using an endogenously tagged COX-4::GFP83 strain and a transgenic muscle-specific myo-3p::GFP reporter84. To validate our findings, we used atp-3 RNAi, a positive control, which is known to cause mitochondrial fragmentation due to ATP synthase dysfunction85,86. In both intestinal and muscle tissues, we observed that kin-19 RNAi-treated animals display a similar fragmented mitochondrial network, mimicking the phenotype observed in the atp-3 control animals (Supplementary Fig. 6g, h). Lastly, we measured ATP levels and found that unlike wild-type and nhr-49(nr2041) animals, kin-19 RNAi treated worms failed to restore ATP levels upon refeeding (Supplementary Figs. 6i, 1h, and 2f). This defective recovery of both cellular energetics and intracellular lipid availability highlights a critical role for KIN-19 in the restoration of energy homeostasis after fasting.
Given that KIN-19 reduction modulates NHR-49-dependent gene expression, we sought to determine whether this response was acting through NHR-49 to impact lipid metabolism. To this end, we utilized nhr-49(nr2041) mutants to investigate whether the loss of NHR-49 function would abrogate physiological and transcriptional changes observed with kin-19 RNAi. Loss of lipid deposition caused by kin-19 RNAi was not observed in the nhr-49(nr2041) mutant animals as monitored by DHS-3::GFP fluorescence (Fig. 4i). Furthermore, the altered expression of fasting-induced genes by kin-19 RNAi was also absent in the nhr-49(nr2041) mutants (Fig. 4j and Supplementary Fig. 6j). This suggests that the chronic fasting state induced by kin-19 RNAi under ad libitum fed conditions is, at least in part, dependent on NHR-49 activity. To further examine this, we introduced the transgenic ATP sensor into nhr-49(nr2041) mutant animals. Unlike wild-type animals treated with kin-19 RNAi (Supplementary Fig. 6i), these mutants efficiently restored their ATP levels upon refeeding (Supplementary Fig. 6k). These findings underscore a critical role for KIN-19 in facilitating the animal’s recovery from fasting-induced β-oxidation. When KIN-19 is disrupted, it impairs the restoration of energy, ultimately compromising the metabolic plasticity needed for a successful transition during nutrient replenishment.
While impairing the activation of lipid catabolism was dispensable for fasting-induced longevity, the importance of silencing this catabolic response remained unclear. To test the necessity for NHR-49 attenuation in this process, we used kin-19 RNAi to prevent its phosphorylation and subsequent inactivation. This manipulation dramatically reduced the benefits of fasting, leading to an average lifespan increase of only 6% and negating the positive effects on longevity (Fig. 4k). This result demonstrates that the ability to properly silence lipid catabolism via KIN-19 is essential for animals to achieve the full, age-related benefits of fasting.
Silencing of NHR-49 by KIN-19 mediated protein turnover
Although KIN-19-dependent phosphorylation is required to silence NHR-49 activity, the downstream effects of this modification remain unclear. We found that phosphorylation of the S114 priming residue altered the nucleocytoplasmic distribution of NHR-49 and was necessary for subsequent phosphorylation of adjacent residues by recombinant KIN-19 (Figs. 2j and 3c). To examine how the loss of KIN-19 impacts NHR-49 protein dynamics, we immunoprecipitated NHR-49::GFP from transgenic animals and observed a threefold increase in the phosphorylated S114 peptide upon kin-19 RNAi (Fig. 5a), indicating that reduced phosphorylation at S117 and S120 leads to accumulation of the S114 mark. Given that incomplete phosphorylation of S117 and S120 could stabilize NHR-49, we next quantified total protein levels. Casein kinases from the CK1 family are known to promote proteasomal degradation of substrates involved in circadian rhythm, cancer immunotherapy, and hepatic lipogenesis87,88,89,90,91. Consistent with this, kin-19 RNAi significantly increased steady-state levels of NHR-49::YFP (Fig. 5b and Supplementary Fig. 7a). Despite the fact that peptides phosphorylated at S117 and S120 were not detected, these transient modifications may facilitate NHR-49 inactivation through nuclear export or turnover. Since mutation of the S114 priming site altered the subcellular distribution of NHR-49 (Fig. 2j), we next asked whether additional cofactors contribute to its shuttling or turnover.
Fig. 5: Silencing NHR-49 via primed phosphorylation. The alternative text for this image may have been generated using AI. Full size image a Relative abundance of phosphorylated S114 NHR-49 peptides detected by mass spectrometry in empty vector control and kin-19 RNAi conditions. Mean ± SEM. n = 4 biological replicates. A paired t-test was used for statistical analysis. b Western blot of NHR-49::YFP in nhr-49(nr2041) background under empty vector control and kin-19 RNAi conditions at Day 1 of adulthood. c TurboID proximity labeling of N- or C-terminally tagged NHR-49 to XPO-1 and RAN-1. Normalized to the C-terminally tagged strain. n = 3 (C-terminal), 4 (N-terminal) biological replicates. Box boundaries = 25% (lower), 75% (upper) percentile. Whiskers = Min/Max. Centre line = Median. Ordinary one-way ANOVA with Šídák’s multiple comparisons test used for statistics. d Scatter plots from large-particle flow cytometry of transgenic worms expressing rab-11.2p::YFP show relative fluorescence per individual animal. Plots compare Day 1 adults under empty vector control conditions to the respective nucleocytoplasmic RNAi conditions. Mean ± 95% CI. n = 80, 70, 245, 118, 254, 215, 188, 235, 125, 75, 225, 141, 233, 450, 345, 224, 24, 92, 81, 222, 278, 1259, 909, 1243, 1311, 94, 94, 74, 104 and 203 from left to right, ordinary one-way ANOVA with Dunnett’s multiple comparisons test used for statistics (see Supplementary Table 4). e Western blot of NHR-49::YFP in nhr-49(nr2041) background under empty vector control and xpo-1 RNAi conditions at Day 1 of adulthood. f Predicted secondary structure and intermolecular interface of XPO-1, RAN-1 and NHR-49, modeled using AlphaFold3, an advanced AI-driven deep learning algorithm for protein structure and interaction prediction. Surface density for XPO-1 (purple) and RAN-1 (black) and a cartoon rendering of NHR-49 (ligand binding domain in blue and DNA binding domain in yellow). g Immunoblot analysis of XPO1 binding to an NHR-49 polypeptide array. A total of 123 overlapping 15mer peptides (spanning the full length of NHR-49 isoform c and overlapping by 11 amino acids) were immobilized on cellulose and incubated with recombinant XPO1 and the RAN-1 ortholog, GSP1, prior to XPO1 immunoblotting. See Supplementary Fig. 7f for schematic. h Western blot of T7 epitope tagged N-terminal NHR-49 (1-155). Utilizing recombinant enzymes and substrates, either HECD-1 or WWP-1 were incubated with NHR-49 in the presence of ubiquitin and the E1 and E2 ubiquitin ligases.
To identify factors regulating NHR-49 turnover, we utilized our previously performed TurboID proximity labeling data set that tagged endogenous NHR-49 at either the N- or C-terminus48. Because this phosphorylation cluster resides within the first 120 amino acids, we anticipated that N-terminal tagging would preferentially biotinylate key regulators coordinating with the hinge region. After filtering for common proteins between the N- and C-terminal data sets, 149 candidates were enriched at least twofold (p < 0.05) in the N-terminal dataset compared to the C-terminal one. Among these, XPO-1, highly conserved exportin, ranked 14th with a 9.59-fold enrichment (p = 0.0044), and its binding partner RAN-1 showed a 3.56-fold enrichment (p = 0.0599) (Fig. 5c and Supplementary Fig. 7b). We next performed RNAi knockdown of nucleocytoplasmic trafficking candidates identified both here and in the fasted NHR-49::GFP immunoprecipitation datasets. Depletion of XPO-1 significantly activated the rab-11.2p::YFP fasting reporter (Fig. 5d) and increased fluorescence of the acs-2p::GFP reporter by 68% (Supplementary Fig. 7c). Unlike kin-19 knockdown, however, loss of xpo-1 did not cause NHR-49 accumulation (Fig. 5e and Supplementary Fig. 7d), suggesting an alternative role for this interaction, potentially through nuclear-based degradation or functions independent of canonical nuclear export.
Sequence analysis of NHR-49 did not identify a canonical nuclear export signal (NES) to explain its interaction with XPO-1; however, XPO-1 has recently been reported to mediate transcription factor–dependent chromatin docking at the nuclear pore complex independently of a canonical NES92. Consistent with this, AI-based structural prediction74 indicated a direct interaction between NHR-49 and XPO-1 that required its small GTPase cofactor RAN-1 (ipTM = 0.66 versus 0.54 without RAN-1) (Fig. 5f and Supplementary Fig. 7e). The predicted binding interface was located at the N-terminus of NHR-49, consistent with our proximity labeling results. To further validate this interaction, we employed an immobilized peptide array spanning the length of NHR-49 and probed with recombinant human XPO1 together with the yeast RAN-1 ortholog, GSP1. XPO1 specifically bound to NHR-49 peptides corresponding to the AI-predicted N-terminal interfaces (Fig. 5g and Supplementary Fig. 7f). Together, these results support an atypical interaction between XPO-1 and NHR-49 that may facilitate translocation of NHR-49–bound DNA to the nuclear pore.
Building on our observation that NHR-49 localizes in proximity to nuclear pore components48, we revisited our nucleocytoplasmic trafficking RNAi screen. Knockdown of several nuclear pore proteins, most notably the inner ring component NPP-8, blocked activation of the rab-11.2 fasting reporter (Fig. 5d). We previously reported that NPP-8 was 2.42-fold enriched at the N-terminus of NHR-49, suggesting a potential cooperative interaction with XPO-148. These findings support a model in which XPO-1 binding to the N-terminus of NHR-49 serves as a critical upstream event that promotes NHR-49’s association with the nuclear pore complex, potentially coordinating its nuclear engagement under conditions of stress. The observation that XPO-1 knockdown does not cause NHR-49 accumulation, unlike kin-19 knockdown, further suggests that XPO-1 regulates NHR-49 through a distinct, non-canonical mechanism. Whereas KIN-19 directly influences protein stability, XPO-1 appears to act earlier in the trafficking pathway.
While KIN-19 appears to influence the steady-state levels of NHR-49, we sought to investigate the mechanisms underlying NHR-49 degradation. A common pathway for nuclear hormone receptor degradation is the ubiquitin-proteasome system, wherein specific E3 ubiquitin ligases mediate receptor ubiquitination and degradation93,94. Consistent with this, phosphorylation of other nuclear receptors, such as the androgen and progesterone receptors, has been shown to enhance their recognition by E3 ligases95. Based on proximity labeling and co-immunoprecipitation, seven E3 ubiquitin ligases were identified as potential interactors of NHR-49, with the top candidates being the HECT domain-containing E3 ligase, HECD-1, with 30 PSMs detected via proximity labeling, and the WW domain-containing E3 ligase, WWP-1, with 110 PSMs identified through co-immunoprecipitation (Supplementary Fig. 7g). When testing their ability to conjugate ubiquitin to NHR-49 in vitro, WWP-1 successfully transferred ubiquitin, resulting in detectable polyubiquitinated species, whereas HECD-1 failed to do so (Fig. 5h). Using both recombinant proteins, we confirmed that WWP-1 directly binds to NHR-49 with a 1-to-1 stoichiometry in vitro (Supplementary Fig. 7h). While WWP-1 has the capacity to ubiquitinate NHR-49 and has previously been implicated as an aging regulator during dietary restriction96,97, wwp-1 RNAi did not reproducibly stabilize steady-state NHR-49 levels within the worm as determined by western blot (Supplementary Fig. 7i, j). While this discrepancy could be due to systemic redundancy, the action of deubiquitinating enzymes, or a requirement for adapter proteins or specific post-translational modifications, WWP-1 remains a promising candidate capable of ubiquitinating and modulating the activity of NHR-49 during metabolic stress. and modulating its activity.