Low-iron diet may extend worm lifespan through stress response pathways

From this primary screen, we identified 26 E. coli mutants that significantly reduced FAT-7::GFP levels (Figs. 1B and EV1B; Table EV1). However, no mutants were found that significantly increased FAT-7::GFP expression. For clarity, we will refer to the E. coli mutants that suppressed FAT-7 expression in C. elegans as FAT-7-suppressing diets, while the BW25113 strain will be referred to as the control diet. Notably, worms fed on all the FAT-7-suppressing diets exhibited delayed development compared to those grown on the control diet (Fig. 1C).

Gene ontology (GO) analysis for the molecular function of the identified bacterial mutants revealed enrichment of categories related to electron transport and oxidoreductase activity (Fig. EV1C). This suggested that the E. coli mutants that suppress FAT-7::GFP might have a disrupted redox balance.

FAT-7 converts stearic acid to oleic acid, and elevated oleic acid levels are known to suppress fat-7 expression (Venkatesh et al, 2023). We hypothesized that bacterial diets might reduce FAT-7 expression because they have elevated levels of oleic acid in C. elegans. Previous studies have shown that oleic acid supplementation extends C. elegans lifespan (Han et al, 2017b), suggesting that lifespan extension on FAT-7-suppressing diets might result from increased oleic acid levels. FAT-2 encodes a Δ12 desaturase, and mutants lacking fat-2 cannot convert oleic acid into linoleic acid, leading to elevated oleic acid levels (Watts and Browse, 2002). To examine whether oleic acid accumulation accounts for the extended lifespan observed with FAT-7-suppressing diets, we studied the survival of fat-2(wa17) hypomorphic mutants on the four selected diets. Surprisingly, fat-2(wa17) animals exhibited an increased lifespan on these diets (Fig. 3B), suggesting that the lifespan extension is unlikely to be due to oleic acid accumulation in C. elegans fed the FAT-7-suppressing diets.

To further investigate the role of oleic acid in lifespan extension on FAT-7-suppressing diets, we examined the effects of oleic acid supplementation. As expected, oleic acid supplementation increased the lifespan of worms on the control diet (Fig. EV3A–D). Oleic acid also extended lifespan on the FAT-7-suppressing diets, with variable effects on each of the mutant diets. For example, oleic acid supplementation had a neutral effect on lifespan on tktA and allD mutant diets compared to its effect on the control diet. On the other hand, while oleic acid supplementation had a negative effect on lifespan on the yciA diet compared to its effect on the control diet, it had a positive effect on the pdeI mutant diet (Fig. EV3A–D). Overall, these results suggested that the observed lifespan extension under these conditions is unlikely to be driven by oleic acid.

We next asked whether downregulation of FAT-7 itself was responsible for the extended lifespan on these diets. To test this, we overexpressed FAT-7 to determine whether this manipulation could reverse the lifespan extension seen on the mutant diets. A previous study reported that intestinal overexpression of FAT-7 extends C. elegans lifespan (Han et al, 2017b). Moreover, FAT-7 expression is primarily observed in the intestine in the FAT-7::GFP reporter strain. Therefore, we overexpressed FAT-7 under an intestine-specific promoter. Unexpectedly, we did not observe increased lifespan upon FAT-7 overexpression in worms fed the control diet (Fig. EV3E–H). This discrepancy from Han et al, 2017b may reflect differences in experimental conditions, such as bacterial diets or transgene expression levels. Nonetheless, intestinal FAT-7 overexpression only partially reduced lifespan extension on the FAT-7-suppressing diets (Fig. EV3E–H), indicating that suppression of FAT-7 expression contributes only modestly to the observed phenotype. Collectively, these results suggested that, within our experimental framework, FAT-7 expression likely functions as an indirect proxy for lifespan regulation rather than a direct determinant.

To investigate the mechanisms underlying the increased lifespan of C. elegans on mutant E. coli diets, we examined transcriptomic changes in worms fed these diets. Wild-type worms were grown on the control diet and four mutant diets (ΔtktA, ΔpdeI, ΔyciA, and ΔallD) until day 1 of adulthood, followed by RNA sequencing. Comparative analysis revealed that 1281 upregulated genes were shared across worms fed all four mutant diets (Fig. 3C; Dataset EV2). Similarly, a significant overlap was observed among downregulated genes (Appendix Fig. S3A; Dataset EV3). These findings suggested a shared molecular mechanism underlying the lifespan-enhancing effects of these diets.

GO analysis of the molecular functions associated with the downregulated genes on all mutant diets revealed enrichment for nucleic acid binding and protein heterodimerization activities (Appendix Fig. S3B). On the other hand, GO analysis of the 1,281 genes upregulated on all mutant diets showed enrichment for molecular functions related to monooxygenase, oxidoreductase, UDP-glucosyltransferase, and iron ion binding activities (Fig. 3D). Notably, these genes are linked to detoxification pathways and are typically upregulated in response to oxidative stress. This suggested that worms feeding on mutant diets may experience elevated reactive oxygen species (ROS) levels compared to those on the control diet. Mutations in the mitochondrial genes nuo-6 and isp-1, which encode subunits of complex I and III of the mitochondrial respiratory chain, respectively, are known to increase superoxide levels (Yang and Hekimi, 2010). A comparison of the genes upregulated on FAT-7-suppressing diets with those induced in nuo-6 and isp-1 partial loss-of-function mutants revealed significant overlap (Fig. 3E,F), supporting the notion that worms fed on FAT-7-suppressing diets experience elevated ROS. Direct measurements confirmed this prediction, showing significantly higher ROS levels in worms fed FAT-7-suppressing diets relative to the control diet (Fig. 3G,H).

Mitochondria are highly sensitive to elevated ROS, which can create a proteotoxic environment and disrupt protein trafficking across the inner mitochondrial membrane (Melber and Haynes, 2018). Such disruptions can impair mitochondrial protein import and activate the UPRmt, a conserved pathway that restores mitochondrial homeostasis (Melber and Haynes, 2018; Shpilka and Haynes, 2018). Activation of the UPRmt has been associated with lifespan extension in C. elegans (Bennett et al, 2014; Shpilka and Haynes, 2018; Xin et al, 2022). Consistently, the mitochondrial mutants nuo-6 and isp-1, which exhibit significant transcriptomic overlap with worms fed FAT-7-suppressing diets, also activate the UPRmt, and their lifespan extension depends on this pathway (Wu et al, 2018).

Next, we investigated whether UPRmt activation is required for the lifespan extension observed on the mutant diets. ATFS-1, a key transcription factor, contains both a mitochondrial localization signal and a weak nuclear localization signal. Under normal conditions, ATFS-1 is imported into mitochondria and degraded. However, during mitochondrial dysfunction, its import is blocked, and ATFS-1 translocates to the nucleus to activate the UPRmt (Nargund et al, 2012). We analyzed the survival of the ATFS-1 loss-of-function mutant atfs-1(gk3094) on the FAT-7-suppressing diets. The lifespan extension observed on these diets was abolished in atfs-1(gk3094) animals (Fig. 4C). Because isp-1(qm150) mutants also activate the UPRmt and display extended lifespan (Wu et al, 2018), we asked whether their longevity pathway overlapped with that induced by FAT-7-suppressing diets. Indeed, lifespan extension was abolished in isp-1(qm150) mutants fed these diets (Fig. 4D). These findings demonstrated that UPRmt activation is essential for the lifespan-enhancing effects of mutant diets in C. elegans.

Given that the transcriptional profiles of worms fed FAT-7-suppressing diets significantly overlapped with nuo-6 and isp-1 loss-of-function mutants, we next asked whether these mitochondrial mutants also showed reduced fat-7 expression. Indeed, transcriptomic data from multiple studies showed that fat-7 is consistently downregulated in nuo-6 and isp-1 mutants (Park et al, 2020; Senchuk et al, 2018; Wu et al, 2018; Yee et al, 2014). This led us to hypothesize that mitochondrial stress more broadly downregulates fat-7. Supporting this, reanalysis of published datasets revealed reduced fat-7 expression in several mitochondrial mutants with activated UPRmt, including clk-1, cco-1, and hsp-6 (Fischer et al, 2014; Mao et al, 2019; Matilainen et al, 2017; Tian et al, 2016; Zhu et al, 2020). To confirm whether mitochondrial stress results in the downregulation of fat-7, we exposed the fat-7p::fat-7::GFP reporter strain to paraquat (PQ). While PQ treatment resulted in the upregulation of hsp-6p::GFP, it led to the downregulation of FAT-7::GFP levels (Fig. EV4A–D). Similarly, knockdown of tomm-22, which elicits UPRmt, also led to downregulation of FAT-7::GFP (Fig. EV4E–H). Together, these findings suggested that mitochondrial stress suppresses fat-7 expression and that the FAT-7 reporter may have functioned as an indirect indicator of mitochondrial stress in our Keio library screen.

A previous study by Zhang et al identified E. coli Keio mutants that delayed C. elegans development (Zhang et al, 2019). Interestingly, most of these mutants also upregulated hsp-6p::GFP expression in C. elegans that were rescued by NAC supplementation. Similarly, our study observed delayed development and increased hsp-6 expression in worms fed on mutant diets. A comparison between the two studies revealed a nearly complete overlap, with 23 out of 26 mutants from our screen matching those identified by Zhang et al (Appendix Fig. S4B). Zhang et al attributed the observed phenotypes to elevated ROS levels and reduced bioavailable iron in E. coli mutants.

To explore whether the lifespan extension observed under iron chelation and on mutant diets involved overlapping mechanisms, we studied the survival of worms fed on mutant diets upon supplementation with bipyridyl. While supplementation of bipyridyl enhanced C. elegans lifespan on the control diet (Fig. 7F), it did not further extend the lifespan on mutant diets (Fig. 7G–K). Taken together, these results suggested that the lifespan extension induced by mutant diets arises from a low-iron environment, similar to that created by bipyridyl supplementation.

We next investigated the mechanisms underlying lifespan extension in worms fed FAT-7-suppressing or iron-depleted diets. Because changes in food intake can influence lifespan, we first tested whether the mutant diets affected feeding behavior. Worms fed FAT-7-suppressing diets showed a significant reduction in pharyngeal pumping (Fig. EV5A). To determine whether reduced food intake accounted for the observed phenotypes, we examined eat-2(ad465) mutants, which display markedly reduced pharyngeal pumping (Avery, 1993). If decreased pumping were causal, eat-2 mutants should exhibit reduced fat-7 expression and elevated hsp-6 expression. However, eat-2 mutants showed neither phenotype (Fig. EV5B–E). Thus, reduced pharyngeal pumping was not the cause of the observed phenotypes but was more likely a consequence of elevated oxidative stress. Supporting this idea, mitochondrial mutants with increased oxidative stress are known to show reduced pumping (Jafari et al, 2015; Yee et al, 2014).

We next investigated the host genetic pathways involved in the lifespan extension observed in worms fed FAT-7-suppressing or iron-depleted diets. Since our screen utilized FAT-7 expression, we tested whether the nuclear hormone receptor NHR-49, which regulates FAT-7 and lifespan (Naim et al, 2021; Ratnappan et al, 2014), was required for lifespan extension. The nhr-49 loss-of-function mutant exhibited extended lifespan on FAT-7-suppressing diets, suggesting that NHR-49 is not essential for lifespan extension (Appendix Fig. S5A). Because worms experienced oxidative stress on the mutant diets (Fig. 3G,H), we asked whether oxidative stress response pathways are required for the increased lifespan. The hypoxia-inducible factor (HIF-1) is activated by ROS and is required for lifespan extension mediated by ROS (Hwang et al, 2014; Lee et al, 2010; Ravi and Singh, 2025). We examined whether HIF-1 was required for increased lifespan on the FAT-7-suppressing diets. The hif-1 loss-of-function mutant exhibited an enhanced lifespan on the mutant diets, indicating that HIF-1 was not required for the increased lifespan (Appendix Fig. S5B).

Finally, we tested whether these oxidative stress response pathways were also required for lifespan extension under iron-depleted conditions. To this end, we studied the lifespan of skn-1(zj15), sek-1(km4), and hlh-30(tm1978) animals upon iron chelation. Supplementation of bipyridyl did not increase the lifespan of skn-1(zj15), sek-1(km4), and hlh-30(tm1978) animals (Fig. 8D–F). Together, these findings demonstrated that oxidative stress response pathways, including those involving SKN-1, SEK-1, and HLH-30, are critical for lifespan extension on both mutant diets and iron-depleted conditions.

The bacterial strains used in this study include Escherichia coli OP50, E. coli HT115(DE3), E. coli BW25113, and mutants from the E. coli Keio collection (Baba et al, 2006). E. coli OP50 and E. coli BW25113 cultures were grown in Luria-Bertani (LB) broth at 37 °C. The Keio collection mutants were grown in LB broth supplemented with 25 µg/mL kanamycin at 37 °C during the FAT-7::GFP screen, while for other experiments, they were grown in LB without the antibiotics.

C. elegans hermaphrodites were maintained on NGM plates seeded with E. coli OP50 at 20 °C unless otherwise specified. The Bristol N2 strain was used as the wild-type control unless indicated otherwise. The following strains were used in this study: DMS303 nIs590 [fat-7p::fat-7::GFP + lin15(+)], BX26 fat-2(wa17), SJ4100 zcIs13 [hsp-6p::GFP + lin-15( + )], VC3201 atfs-1(gk3094), QV225 skn-1(zj15), KU4 sek-1(km4), JIN1375 hlh-30(tm1978), STE68 nhr-49(nr2041), ZG31 hif-1(ia4), MQ887 isp-1(qm150), eat-2(ad465);nIs590 [fat-7p::fat-7::GFP + lin15(+)], and eat-2(ad465);zcIs13 [hsp-6p::GFP + lin-15(+)]. Some of the strains were obtained from the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MN). The eat-2(ad465);nIs590 [fat-7p::fat-7::GFP + lin15(+)] and eat-2(ad465);zcIs13 [hsp-6p::GFP + lin-15(+)] strains were generated using standard genetic crosses. For all the experiments, worms were synchronized by bleach treatment to obtain the same-stage L1 larvae.

For the overexpression of fat-7, the fat-7 gene was amplified using the cDNA of N2 worms. The gene, including its stop codon, was cloned into the pPD95_77 plasmid using the restriction sites SalI and KpnI. The promoter region of the intestine-specific gene vha-6 (1248 bp upstream) was cloned upstream of fat-7 using the restriction sites HindIII and SalI. N2 worms were microinjected with vha-6p::fat-7 plasmid along with pCFJ104 (myo-3p::mCherry) as a coinjection marker to generate the overexpression strain, jsnEx4 [vha-6p::fat-7 + myo-3p::mCherry]. The vha-6p::fat-7 plasmid was used at a concentration of 50 ng/µL, while the coinjection marker was used at 25 ng/µL.

The following supplements were obtained from HiMedia BioSciences: ferric chloride (#TC583), N-acetylcysteine (NAC) (#RM3142), and 2,2'-bipyridyl (#GRM791). Paraquat dichloride (PQ) (#856177) and sodium oleate (# O7501) were purchased from Sigma. Stock solutions were prepared as follows: 1 M ferric chloride, 0.5 M PQ, and 0.5 M NAC in water, and 100 mM 2,2'-bipyridyl in dimethyl sulfoxide (DMSO). All stock solutions were stored at −20 °C and diluted to their final concentrations in NGM before pouring the plates. For sodium oleate supplementation, Nonidet P-40 was added to a final concentration of 0.001% in liquid NGM before autoclaving, in both supplemented and control plates. Sodium oleate was weighed and added directly to the NGM before pouring plates. For experiments with 2,2'-bipyridyl, control plates were supplemented with an equivalent amount of DMSO. Worms were grown from the synchronized L1 stage on all supplements except PQ. For PQ treatment, worms were exposed at the late L4 stage and incubated for 24 h prior to fluorescence imaging.

Bacterial mutants were grown overnight at 37 °C in LB broth supplemented with 25 μg/mL kanamycin in 96-well plates. Subsequently, 30 μL of the overnight cultures were seeded onto 24-well plates containing NGM agar supplemented with 25 μg/mL kanamycin. The plates were incubated at room temperature for at least 2 days to allow bacterial growth before use in experiments. For screening, embryos of the fat-7p::fat-7::GFP strain were harvested from gravid adults using an alkaline bleach solution and incubated in M9 buffer at room temperature for 22 h to obtain synchronized L1 larvae. Approximately 30–40 synchronized L1 larvae were transferred to each well of 24-well NGM agar plates seeded with individual E. coli single-gene deletion mutants and incubated at 20 °C. Since FAT-7 levels vary across developmental stages, screening was performed at the day-1-adult stage to identify mutants that modulate FAT-7 expression. GFP fluorescence was monitored in each well to identify E. coli mutants that either enhanced or suppressed FAT-7 levels compared to the E. coli BW25113 control diet.

Bacterial mutants identified as hits in the initial screen were retested three times on individual NGM plates seeded with the corresponding mutants. E. coli mutants that consistently reproduced the phenotype across three independent trials were considered as primary hits. Gene ontology analysis was performed using the DAVID Bioinformatics Database (https://david.ncifcrf.gov/tools.jsp↗).

Primary cultures of each bacterial strain were grown in LB broth without antibiotics at 37 °C with shaking for 12 h. These cultures were then diluted to an initial optical density (OD₆₀₀) of 0.01 in 15 mL of fresh LB broth in 50-mL centrifuge tubes. The diluted cultures were incubated at 37 °C with continuous shaking, and OD₆₀₀ measurements were taken hourly to monitor bacterial growth.

RNAi was performed to generate loss-of-function phenotypes by feeding nematodes the E. coli strain HT115(DE3) expressing double-stranded RNA homologous to tomm-22. RNAi was carried out as described previously (Das et al, 2024; Rao et al, 2025). Briefly, E. coli strains were cultured overnight in LB containing ampicillin (100 µg/mL) at 37 °C. They were then concentrated 20 times and plated on an NGM plate containing 3 mM isopropyl β-D-thiogalactoside and ampicillin (100 µg/mL) (RNAi plate). The plated bacteria were allowed to grow overnight at 37 °C before use. For worm synchronization, gravid adults were bleached, and embryos were allowed to hatch in M9 buffer for 22 h at room temperature to obtain L1 larvae. These synchronized L1s were transferred to RNAi plates and incubated at 20 °C till the day-1-adult stage. The tomm-22 RNAi clone was obtained from the Ahringer RNAi library.

Lifespan assays were conducted as described earlier (Das et al, 2024). Briefly, gravid adults were lysed using an alkaline bleach solution to obtain embryos, which were then incubated in M9 buffer for 20–24 h to synchronize them at the L1 larval stage. Synchronized L1 larvae were transferred to NGM plates seeded with either wild-type E. coli BW25113 or bacterial mutants identified from the FAT-7::GFP screen. For assays involving ferric chloride, NAC, sodium oleate, and 2,2'-bipyridyl supplementation, synchronized L1 larvae were transferred to NGM plates containing these supplements. At the late L4 larval stage, the animals were transferred to corresponding bacterial diet plates or supplement plates containing 50 µg/mL FUdR and incubated at 20 °C. Worms were monitored daily or every other day and scored as alive or dead. Animals that failed to exhibit touch-provoked movement were classified as dead, while those that crawled off the plates were censored from the analysis. For the lifespan analysis, young adult animals were designated as day 0. Three independent experiments were performed for each condition.

Gravid N2 hermaphrodites were lysed with an alkaline bleach solution to isolate eggs, which were then incubated in M9 buffer at room temperature for 22 h. Approximately 50–100 synchronized L1 larvae were transferred onto NGM plates seeded with either the E. coli BW25113 control or mutant diets and incubated at 20 °C for 60 h. The assays were similarly carried out for NAC and ferric chloride-supplemented diets. Animals at various developmental stages (L1/L2, L3, L4, and adult) were subsequently quantified. The experiment was repeated in at least three independent biological replicates.

Pharyngeal pumping rates were measured in 1-day-old adult animals grown on E. coli BW25113 or the FAT-7-suppressing diets. The number of terminal bulb contractions was counted over a 30-s interval for each worm. For each condition, ten worms were assayed, and the experiment was performed in three independent biological replicates.

ROS levels were quantified using 2',7'-dichlorofluorescein diacetate (DCFHDA, Sigma-Aldrich #35845). A 50 mM DCFHDA stock was prepared in DMSO and stored at −20 °C. Before each experiment, a 50 µM DCFHDA working solution was freshly prepared in M9 buffer. Synchronized L1 larvae of N2 worms were obtained as described above and grown on the E. coli BW25113 control and mutant diets until the day-1 adult stage at 20 °C. Subsequently, 15–20 worms were transferred to 150 µL M9 buffer, followed by the addition of 150 µL of the DCFHDA working solution, resulting in a final DCFHDA concentration of 25 µM. Samples were incubated in the dark at room temperature for 5 h with gentle shaking. Next, the worms were pelleted, the supernatant was removed, and the worms were washed twice with PBS containing 0.01% Triton X-100. The prepared samples were then subjected to fluorescence imaging. The 2',7'-dichlorofluorescein (DCF) fluorescence was visualized using a GFP filter on a fluorescence microscope. At least five worms per condition were imaged, and three independent biological replicates were performed.

Fluorescence imaging was carried out as described previously (Gokul and Singh, 2022; Ravi et al, 2023). Briefly, animals were anesthetized using M9 buffer containing 50 mM sodium azide and placed on 2% agarose pads. The animals were then visualized using either a Nikon SMZ-1000 or SMZ18 fluorescence stereomicroscope. Quantification of fluorescence intensity was done using ImageJ software.

Synchronized L1 larvae were obtained from wild-type animals as described above and grown on the E. coli BW25113 control diet and four mutant diets, including ΔtktA, ΔpdeI, ΔyciA, and ΔallD, until the day 1 adult stage. Total RNA was extracted from three biological replicates using the RNeasy Plus Universal Kit (Qiagen, the Netherlands). Library preparation and sequencing were performed at Unipath Specialty Laboratory Ltd., India. cDNA libraries were sequenced on the NovaSeq 6000 platform using 150-bp paired-end reads.

RNA sequencing data were processed and analyzed using the Galaxy web platform (https://usegalaxy.org/↗), as described previously (Ghosh and Singh, 2024; Rao et al, 2025). Paired-end reads were first trimmed using the Trimmomatic tool and aligned to the C. elegans genome (WS220) with the STAR aligner. Gene expression levels were quantified using htseq-count, and differential expression analysis was performed using DESeq2. Genes with at least a twofold change and P < 0.01 were considered differentially expressed. Gene ontology enrichment analysis was conducted using the DAVID Bioinformatics Database. Venn diagrams were generated using the tools InteractiVenn (Heberle et al, 2015) and BioVenn (Hulsen et al, 2008). The enrichment factor and P values for overlap were generated using the hypergeometric P value calculator (https://systems.crump.ucla.edu/hypergeometric/↗). For the calculation, the total number of genes was set to 20,000.

Statistical analyses were performed with Prism 8 (GraphPad). All error bars represent the mean ± standard deviation (SD). An unpaired, two-tailed, two-sample t test was used when applicable, with statistical significance set at P < 0.05. In the figures, statistical significance is indicated by asterisks: *P < 0.05, **P < 0.01, and ***P < 0.001, relative to the relevant controls. Survival fractions were calculated using the Kaplan–Meier method, and statistical significance between survival curves was determined using the log-rank test. For oleic acid supplementation experiments across different mutant diets, a multivariable Cox regression analysis was performed with genotype and oleic acid supplementation as covariates, and corresponding hazard ratios and P values were calculated. All experiments were performed in triplicate unless indicated otherwise.