Knock-down of odr-3 and ife-2 additively extends lifespan and healthspan in C. elegans

RNA interference of ife-2 but not cku-70 increases lifespan of the long-lived odr-3 mutants

To find new potential genetic interactions that could extend lifespan, we assessed the effect of a simultaneous depletion of ODR-3, IFE-2 and CKU-70. For this, we used the odr-3(n1605) putative null allele [37] and we knocked-down ife-2 and cku-70 by RNAi. The odr-3(n1605) animals exhibited at 20° C an increased mean (26.2%) and maximum (13.8%) lifespan compared with WT control animals. A significant mean lifespan extension was previously reported for odr-3(n1605) at 25° C, however the increase was very modest at 20° C [39]. Silencing of ife-2 by RNAi showed an 18.0% and 20.7% extension for the mean and maximum lifespan of WT, respectively (Figure 1A and Supplementary Table 1), which are in agreement with previously reported data for both ife-2(ok306) mutants and ife-2 downregulated animals [21, 42]. In our experiments, the RNAi knock-down of cku-70 in the WT worms produced only a marginal 4.4% increase for mean lifespan and was even slightly detrimental to maximum lifespan reducing it by 6.9% (Figure 1B and Supplementary Table 1), which is in agreement with previously reported data [46].

The odr-3(n1605); ife-2(RNAi) mutants exhibited more than 11% and 18% increase in mean lifespan compared with the odr-3(n1605) and ife-2(RNAi) single gene interventions, respectively (Table 1). Similarly, the maximum lifespan was also increased by more than 18% and 11% (Supplementary Table 1). Overall, compared with the WT controls, the combined odr-3 and ife-2 interventions extended mean and maximum lifespan by 40.3% and 34.5%, respectively (Figure 1A and Table 1 and Supplementary Table 1). This effect demonstrates an almost additive impact on mean lifespan, i.e. 40.3% increase compared to 44.2%, the sum of the two individual effects (Table 1). Similarly, the lifespan extension for odr-3; ife-2; EV (46.1%), i.e. worms exposed to a 1:1 mixture of HT115 (Empty Vector - EV) bacteria and of ife-2 RNAi clone, was greater than the sum of individual effects (26.2% + 11.2%), supporting the existence of additive/synergistic mechanisms (Table 1).

Next, we assessed the effect of cku-70 silencing in both odr-3(n1605) mutant animals and ife-2 knock-down animals. We observed that cku-70 knock-down dramatically decreased the extension of mean lifespan conferred by the odr-3(n1605) mutation, from 26.2% to only 13.6% increase comparative with WT (Figure 1B and Supplementary Table 1).

The simultaneous knock-down of ife-2; cku-70 by RNAi was performed by co-feeding worms with a mixture of the two RNAi bacterial clones. As such, for an appropriate comparison, the survival curves of double knock-down worms (which are presumably exposed to about half dsRNA for each gene) have been compared with those of single knock-down worms exposed to the same concentration of dsRNA for each of the corresponding genes (concentrations obtained by co-feeding the worms with the target RNAi clone and the control RNAi(EV) in a 1:1 ratio). In general, we obtained very small differences between the lifespan of worms fed only with the RNAi clone and worms fed with the mixture of RNAi clone / RNAi(EV) (Supplementary Figure 1A–1C), with the only notable difference being for odr-3; cku-70 for which the mix (and hence lower concentration of cku-70 RNAi bacteria) did not show a pronounced lifespan reduction (Supplementary Figure 1D).

In our assays, the lifespan of the double knock-down worms ife-2; cku-70 was 15% longer than that of WT animals (p = 3.0E-16), with a small increase compared to each of the single knock-down worms (3.5% and 8.22% longer lived than the ife-2; EV and the cku-70; EV animals, respectively) (Figure 1C and Supplementary Table 1). It should however be noted that these changes are very modest and that a slightly larger lifespan increase was obtained when worms were exposed to ife-2 RNAi alone, without EV mixing (18% compared to WT). Moreover, cku-70 knock-down had a negative effect on the lifespan of odr-3(n1605) mutants treated with ife-2 RNAi, decreasing the mean lifespan extension from 46.1% to 35.4% (Figure 1D and Supplementary Table 1).

The extended longevity of odr-3; ife-2 double intervention might be independent of DAF-16

The FOXO family transcription factor DAF-16 is a transducer of many pro-longevity signaling pathways [47], thus it was natural to inquire to what extent the longevity of odr-3; ife-2 double inactivated animals require DAF-16. To answer this, we used the null daf-16(mu86) allele [32, 48] that affects coding of all DAF-16 isoforms, to generate daf-16(mu86); odr-3(n1605) double mutants, and carried out RNAi silencing assays for ife-2 and cku-70 in this strain.

We observed that the lifespan extension induced by the odr-3(n1605) mutation was not only suppressed by the daf-16(mu86) mutation, but also that the lifespan of the daf-16(mu86); odr-3(n1605) double mutants were even shorter than the lifespan of the daf-16 mutants alone (Figure 1E; mean and maximum lifespan decreased by 14.6% and 9.1% compared to daf-16; mean and maximum lifespan decreased by 26.3% and 16.7% compared with WT). A slight decrease of daf-16(mu86) lifespan induced by the odr-3(n1605) mutation was also previously reported [39]. Silencing of ife-2 on the other hand, extended the lifespan of daf-16 mutants (Figure 1E; mean lifespan 9.8% greater), although the mean and maximum lifespan were not completely reverted to the lifespan values of the WT (mean and maximum lifespan 5.3% and 4.2% lower than WT, respectively). The lifespan of the triple daf-16; odr-3; ife-2 inactivated animals did not significantly differ from daf-16 single mutants. This could be partly explained by the fact that DAF-16 is one of the main transducers of signaling pathways modulated by ODR-3 and IFE-2 activity. To clarify this aspect we examined the nuclear translocation of DAF-16::GFP in odr-3(n1605), ife-2(RNAi) and odr-3(n1605); ife-2(RNAi) animals, respectively (Supplementary Figure 2). Whereas odr-3(n1605) animals showed weak DAF-16::GFP nuclear accumulation in posterior intestinal cells, suggesting that ODR-3 could affect longevity partially through DAF-16 pathway, we did not observe consistent nuclear accumulation of DAF-16::GFP in ife-2(RNAi) and odr-3(n1605); ife-2(RNAi) animals. Therefore, ODR-3 and IFE-2 could affect lifespan by mediating parallel signaling pathways, which in the daf-16 background have antagonistic effects - odr-3 further decreasing lifespan and ife-2 partially increasing the daf-16(mu86) lifespan.

In contrast to a previous study that reported a slight decrease of daf-16(m26) lifespan by cku-70 knockdown at 25° C [46], in our experiments the lifespan of daf-16; cku-70 at 20° C was similar to that of daf-16 single mutants, and cku-70 knock-down did not significantly influence the lifespan of daf-16; odr-3, nor of daf-16; ife-2 mutants (Figure 1F, 1G). The quadruple daf-16; odr-3; ife-2; cku-70 mutants exhibited a lifespan similar to that of daf-16 single mutants (Figure 1H).

Overall, our results show that simultaneous inactivation of odr-3 and ife-2 produce an additive lifespan effect, while the additional cku-70 knock-down does not extend lifespan further (Figure 2). Since the lifespan effects observed are different in the daf-16 background (Figure 2), it is possible that the mechanisms through which odr-3; ife-2 animals achieve lifespan extension overlap with the pleiotropic mechanisms determined by daf-16 (Figure 2).

The ife-2(ok306) mutation also extends the lifespan of odr-3(n1605) mutant animals

One question is whether the ife-2(ok306) deletion mutation will produce similar effects as the ife-2(RNAi) in the odr-3(n1605) background. To answer this, we also generated the double odr-3(n1605); ife-2(ok306) and triple daf-16(mu86); odr-3(n1605); ife-2(ok306) mutants and conducted further lifespan assays. The mutant worms were cultured at the same temperature (20° C) as in the RNAi experiments and were fed OP50 bacteria. We observed similar trends (Figure 3A), i.e. a 13.6% and 35.7% increase in mean and maximum lifespan for the odr-3; ife-2 double mutant, compared to wild type, and an additive effect of the single mutations (the 13.6% increase in mean lifespan was comparable with the sum of the individual genetic effects of odr-3 and ife-2 mutations: 7.5% and 4.5%). It was however noticeable that in this experiment the impact of the double mutation on lifespan was smaller than in the case of ife-2(RNAi), which used an HT115 diet.

Since FUdR could cause an artefactual effect on the longevity of some mutants [49–51], we also conducted longevity experiments in the absence of FUdR. Similar to the previous results, we also observed an increased average lifespan for all strains even in the absence of FUdR (Figure 3B, odr-3:+8.2%, ife-2:+8.6% and odr-3;ife-2:+14.3%).

Next, we investigated the effect of the odr-3(n1605) and ife-2(ok306) mutations in the daf-16(mu86) background. Similar to the RNAi experiments, the ok306 extended the lifespan of daf-16 single mutants (Figure 3C, 6.25% increase), while the double daf-16(mu86); odr-3(n1605) and triple daf-16(mu86); odr-3(n1605); ife-2(ok306) mutants showed very similar lifespans as that of the daf-16 (Figure 3C).

The odr-3; ife-2 impaired animals display increased motility and pharyngeal pumping

Our finding that RNAi impairment of ife-2 in odr-3(n1605) animals increased lifespan prompted us to investigate the effect of their inactivation on healthspan. To assess healthspan, we focused on the evaluation of pharyngeal pumping and body movement. In C. elegans, these two physiological processes decrease with ageing, correlate between themselves and with other age-related declining properties, and ultimately can predict lifespan and healthy life [52].

In our experiments, up to late adulthood, individual or joint interventions in odr-3 and ife-2 did not produce obvious pathological changes in the phenotype, indicating that the effect on locomotion was caused by physiological age-related changes, rather than a specific disease. As such, the observed motility status, carried out along the lifespan assay, can be viewed as a measure of the healthspan of the population. To quantify this, we classified individuals in three motion stages based on their ability to move. Stage A (healthy, fully mobile worms) included animals in a physiological state that could move without any impediment, stage B (impaired worms) included animals with diminished locomotion, whereas stage C (frail worms) included animals found in a frailty state. Animals were scored daily and associated with one of the motion stages. The distribution of stages for each strain is presented graphically in Figure 4A–4D (for motility data of odr-3(n1605); ife-2(ok306) mutants, see Supplementary Figure 3A–3H; for the daf-16(mu86); odr-3(n1605); ife-2(ok306) mutants see Supplementary Figure 3I–3L).

Considering the average healthspan of worms in stage A, we observed a similar trend to lifespan. Simultaneous inactivation of odr-3 and ife-2 produced a synergistic healthspan effect, while the additional cku-70 knock-down does not extend healthspan further (Supplementary Figure 4).

Using the Kaplan-Meier method to estimate the fraction of mobile worms at each observation point, plotted against time, the average number of days the worms spent in each state was computed (Figure 4E and Supplementary Table 2). This allowed us to model the transitions from the healthy to impaired or frail (see Materials and methods), thus, determining if a change in the locomotion status was induced by the genetic interventions and whether in addition to lifespan, the ratio between healthspan and lifespan was also changed.

The WT animals spent on average 17.3 days in the mobile stage, which represent 84% of their mean lifetime, 1.5 days (7.3%) in an impaired stage and 1.8 days (8.7%) in the frailty stage. The odr-3 mutants remained mobile longer than WT animals (on average 20.8 days), however relative to their mean lifespan, they were fully mobile only for 80.0% of their lifetime, thus exhibiting a proportional (or greater) lifespan fraction in which they were impaired or frail (2.6 days, 10.0%, for both). ife-2 RNAi treated animals remained mobile on average more days (21.0 days, 86.4%) and exhibited a similar number of impaired (1.6 days, 6.6%) and frail days (1.7 days, 7.0%) (Supplementary Table 2).

For the odr-3; ife-2 animals, the longest-lived strain in our study, a corresponding increase of days with full (25.0 days, 86.5%) and impaired motility (2.2 days, 7.6%) was observed, while the number of days of frailty remained similar as for WT animals (1.7 days, 5.9%).

Next, we compared each fraction of being fully mobile with the corresponding fraction in the WT animals to find if the genetic interventions indeed conferred significant benefits for the quality of life, i.e. increasing the fraction of time spent in a mobile state and decreasing the impaired and frailty fractions. By doing this, we found that although the odr-3 mutants had extended longevity, this was not associated with a motility-based health benefit since the lifespan fraction in which worms lived as fully mobile actually decreased slightly by 4.8% (Supplementary Table 2), whereas the fraction of time spent in both the impaired and frailty stages increased by almost 40%. By contrast, ife-2 RNAi treated animals stayed mobile for a similar lifespan fraction as WT animals (3% longer), but they spent much less time as impaired or frail (these fractions were 9.6% and 19.5%, respectively, smaller than those for WT). Compared to WT control animals, silencing of ife-2 in odr-3 worms did not affect the mean lifespan fraction spent in the mobile and impaired stages, but decreased the lifespan fraction for the frailty period by more than 30%.

To assess the significance of the changes in motility status, the Kaplan-Meier curves modeling the transitions from mobile to impaired and frailed were used and the relevant comparisons are included in the Supplementary Figure 5C. Overall, our findings show that the ife-2 RNAi treatment and the double intervention odr-3(n1605); ife-2(RNAi) have a beneficial effect on motility-assessed healthspan, significantly increasing the period of full motility (Supplementary Figure 5C; p < 2.0E-16) and exhibiting a decrease in the decrepit period of life. Similar data for the odr-3(n1605); ife-2(ok306) double mutant can be seen in Supplementary Figure 5D, 5E and for the daf-16(mu86); odr-3(n1605); ife-2(ok306) triple mutant in Supplementary Figure 5F.

To further determine whether healthspan is affected, we analyzed the decline of pharyngeal pumping with age, a process controlled by the cardiac-like pharyngeal muscle. As seen in Figure 4F, the number of pharyngeal movements strongly decreases with age in all cohorts. The changes for both single mutants, odr-3(n1605) and ife-2(ok306), are similar to those in the WT (no statistically significant difference observed when comparing to WT, at each day; one-way ANOVA). The odr-3(n1605); ife-2(ok306) double mutant showed a slower decline in pharyngeal contractions and a higher rate of pumping compared to WT. The improved healthspan was observed starting from day 5, when the average number of contractions was 17.8% higher than for WT (p = 0.002), and slightly increased, at day 10 being 21.4% higher (p = 0.14; ns). At day 15, the odr-3;ife-2 animals still showed approximately 42 contractions per minute (compared to 15 contractions per minute in the WT), a 187% increase (p odr-3(n1605) and ife-2(ok306) mutations showed a synergistic effect on the pharyngeal pumping, the increase observed in the double mutant being higher than the sum of the individual effects (187% increase compared to 66% + 55% increase for single mutants), corresponding to a fully synergic effect [18].

The odr-3; ife-2 double mutant displays an increased resistance to stress

Resistance to stress declines with age after early adulthood [53], and the loss of protein homeostasis together with the failure to activate cellular stress responses are among the earliest aging marks [54]. Elevated temperatures perturb the protein homeostasis due to accumulation of defective proteins, whereas production of reactive oxygen species (ROS) and H₂O₂ increase global cellular damage. Both insults activate the cellular stress responses, aiming to improve cellular fitness and organismal recovery. To find if the double intervention in odr-3 and ife-2 perturbs the stress response mechanisms, we monitored survival upon exposure to oxidative stress and acute heat stress. We induced oxidative stress by treatment with either paraquat, which generates reactive oxygen species, or NaN₃, which generates H₂O₂. Both treatments dramatically reduce survival of WT animals by more than 50% (Figure 5A, 5B). In contrast, mutation in the odr-3 increases survival of the animals treated with both paraquat and NaN₃ compared with treated WT, although not statistically significant for NaN₃. ife-2 mutants exhibited an increased survival upon both treatments, as previously reported [42]. The odr-3; ife-2 double mutants were less sensitive to ROS and H₂O₂ toxicity than WT animals, but they did not show an additive effect when compared with single mutants (Figure 5A, 5B).

Exposure to 35° C for 4h decreases the survival rate of WT animals by more than 50% (Figure 5C). Whereas the heat shock slightly increased the survival of odr-3 mutants, compared with WT (not statistically significant), it did not also affect the survival of ife-2 mutants or odr-3; ife-2 double mutants (Figure 5C), overall indicating that impairment of odr-3 and/or ife-2 do not affect the heat stress response.

Some genetic or non-genetic interventions that extend lifespan also reduce fecundity, implying a trade-off between longevity and reproduction [55–57]. We did not observe such an effect for odr-3 and ife-2 single and double mutants (Figure 5D). Thus, the reproductive period and the age of the peak egg-laying rate of single and double mutants coincided with that of WT animals (Figure 5D). Moreover, the brood size of the mutants is similar to that of WT animals (Figure 5E). In conclusion, simultaneous depletion of odr-3 and ife-2 extends lifespan without affecting fecundity, promotes muscle activity and maintains activation of stress response mechanisms, consistent with increased health.

Strains and culture conditions

The following strains used in this study were provided by the Caenorhabditis Genetic Center (CGC): C. elegans wild-type Bristol strain (N2), CX3222 odr-3(n1605)V, RB579 ife-2(ok306), CF1038 daf-16(mu86)I, OH16024 daf-16(ot971 [daf-16::GFP])I, E. coli OP50 and HT115(DE3) strains. The C. elegans strains were maintained at 20° C using standard methods [84].

Multiple mutants were obtained by standard genetic methods and the presence of mutations was tested either by screening for characteristic phenotypes or via PCR genotyping. The homozygous odr-3(n1605) allele was confirmed by negative chemotaxis tests to isoamyl alcohol. To confirm the presence of homozygous daf-16(mu86) allele, high density populations were allowed dauer formation and subsequently tested for resistance to SDS 1%. Presence of ife-2(ok306) deletion was confirmed by PCR genotyping.

ife-2 and cku-70 RNAi

For the RNAi-mediated gene knock-down by feeding method, a slightly modified protocol of the Ahringer technique was used [85]. Briefly, bacteria were grown overnight in LB medium supplemented with 50 μg/ml ampicillin and seeded onto NGM plates supplemented with 25 μg/ml carbenicillin and 1 mM IPTG. The plates were kept at room temperature for two days before use. Several L4 hermaphrodites picked from plates seeded with OP50 were placed onto RNAi plates, transferred the next day to other fresh RNAi plates, allowed to lay eggs for 24 h, then removed. The L4 hermaphrodites developed from eggs laid onto RNAi plates were used for longevity and healthspan assays. For double RNAi experiments, the plates were prepared in a similar way, with the exception that the plates were seeded with a 1:1 mixture of both RNAi bacterial clones. The ife-2 and cku-70 RNAi clones were obtained from the Ahringer RNAi library (Source BioScience, Nottingham, UK); both clones were validated by sequencing. The HT115 bacteria transformed with the L4440 empty vector, HT115 (EV), was used as control for RNAi experiments unless otherwise specified. When ife-2; cku-70 double RNAi was employed, the control worms were grown on plates seeded with a 1:1 density mixture of HT115 (EV) bacteria and ife-2 or cku-70 RNAi clone, respectively, to maintain the same concentration of each double strand RNA as in the strains subjected to double RNAi.

Lifespan assays

Since the age at which a treatment is started can significantly influence the outcome [86], all worm cohorts used in this work have been age-synchronized (L4 larvae stage). For all RNAi experiments, age-synchronized L4 larvae were manually transferred to RNAi agar plates containing 15 μM 5-fluorodeoxyuridine (FUdR). For lifespan assays of mutant animals, age-synchronized L4 larvae were manually transferred to NGM plates, and seeded with OP50. In case of lifespan assays without FUdR, the worms were transferred to a new plate every day until they ceased laying eggs, then when needed. For mutant animals cultured with FUdR, a 15 μM FUdR concentration (same as in the RNAi experiments) was used. In all cases, worms were kept at 20° C and scored daily as dead or alive based on their response to a gentle touch with a wire. Worms that presented externalization of internal organs, died because of bagging, or crawled up the wall of the dish were censored. For RNAi experiments, the WT control and odr-3 animals were fed with HT115 (EV) bacteria. 85 worms were assayed per experiment.

While the lifespan assays were not conducted in a blinded manner, as suggested by Gruber et al., [87], the experiments were carried out by 3 operators, working with the data independently and results were evaluated for consistency. From the beginning of the study, all operators aimed to treat worm cohorts in an unbiased fashion and keep them in the same conditions.

Locomotion assay

Animals were scored for free movement and for a response to prodding with a wire, daily, during the lifespan assay until death. Worms were classified into three motion stages, based on ability to engage and coordinate the body wall muscle in a forward or backward movement according to a previously described method [88], with slight modifications. The three stages considered were: 1) state A, corresponding to a physiological, fully mobile state, which included animals that could move more than 0.5 cm (freely or upon prodding); 2) state B, representing impaired animals that responded to the prodding, but did not have enough strength to move more than 0.5 cm, and 3) state C, which encompassed animals in a frailty state that barely exhibited head or tail movements or twitches upon prodding.

Pharyngeal pumping assay

For the pharyngeal pumping assay, separate worm cohorts were cultured, in three independent experiments, each with 60 animals grown on FUdR-supplemented plates seeded with OP50 bacteria. Pumping was monitored and recorded at 1, 5, 10 and 15 days post-L4 moult by filming 13-15 randomly selected worms at each time point. Time lapse recordings were obtained on the Zeiss SteREO Discovery. V20 stereomicroscope (Carl Zeiss AG, Jena, Germany) using the AxioVs40 V4.8.2.0 software (Carl Zeiss AG), using the 1X Plan Apo S objective, at a magnification of 150X; the time lapse captures were converted into.mp4 files using an in-house developed script. Pharyngeal contractions were then accurately counted during a 30-seconds interval. For each strain the average number of pharyngeal movements per minute, standard deviation and standard error of the mean were computed.

Stress assays and fecundity

Tolerance to heat and oxidative stress was tested for late L4 animals, since responses to both stresses become repressed early in adulthood (starting as early as 4 hours post L4) [89], suggesting that collapse of cellular stress response could represent an early molecular event in the aging process.

Heat stress assay

To obtain synchronized populations, five hermaphrodites were let to lay eggs for about two hours on three replicate plates. The larvae were reared at 20° C up to late L4 stage, shifted to 35° C for four hours, then returned to 20° C. The percentage of alive animals was scored 48 h later. More than 420 animals were tested for each strain in three individual experiments.

Oxidative stress assay

Tolerance to oxidative stress was tested by exposure to paraquat and NaN₃. Both assays were essentially performed as previously described [42]. Briefly, synchronized L4 larvae were transferred to NGM plates containing 2 mM paraquat and survival was scored on day 5 of exposure. For NaN₃ tolerance, synchronized L4 larvae were collected, washed with M9, incubated for one hour with 0.5 M freshly made NaN₃ in M9, washed with M9 and placed on a NGM plate to recover. Survival rates were determined after 24 hours.

Brood size

To analyze the total brood size, L4 worms were placed on NGM plates seeded with OP50 and transferred every 12h to a new plate until they ceased laying eggs. The number of eggs laid by each worm was counted after removal of the parent.

DAF-16::GFP nuclear translocation

To verify DAF-16 activation we used the CRISPR allele of daf-16 tagged at the C-terminus with GFP [90]. Since DAF-16 translocation from cytoplasm to nucleus is induced by heat stress and common drugs used to anesthetize the worms [91], the young adult worms were fixed in 4% paraformaldehyde. The worms were grown at 20o C on RNAi plates, at L4 stage 15 μM FUdR was added and next day, the young adults were fixed and immediately visualized. Images were acquired with a Zeiss LSM 710 laser scanning confocal microscope using 10x objective, Argon 488 laser, and identical acquisition setting.

Statistical analysis

Comparisons between the lifespan values of different strains were carried out by analyzing Kaplan-Meier survival curves. For the statistical analysis and graphical representation of the curves, the R package “survival” was used (https://cran.r-project.org/package=survival).

For the statistical analysis of locomotion, animals were scored by motion stage and plotted as motility curves, similar to the lifespan curves, to evaluate the decline rate of motility in each strain. While in the survival analysis an event represents the death of one worm, in the locomotion analysis an event was defined as the transition between motility states, modeling the population dynamics from increased to low motility (Supplementary Figure 4). Since three motility categories exist, two different motility analyses were conducted, corresponding to two types of events: i) transitions from the mobile state A to either the impaired state B or to the frail state C, and ii) transitions from either state A or B to state C. In the manuscript, only the results from the motility analysis of transitions from A vs. cumulated B and C is included, however the two analyses produced very similar results (data not shown).

For all survival and locomotion analyses, the statistical significance was tested using the log-rank test (Mantel-Cox). Comparisons were performed against WT, WT(EV) or daf-16, as appropriately. If not otherwise specified, p<0.0001 was considered significant. The p-values were corrected for multiple testing using the Benjamini-Hochberg method, at alpha=0.05.

One-way ANOVA was used for the analysis of pharyngeal pumping, heat, paraquat and sodium azide stress resistance assays. All mutants were compared to wild type and statistical significance was assessed using Dunnett’s test.