As C. elegans has recently been introduced as a model for ether lipid deficiency, we sought to determine how the distribution of ether lipids in the nematode compares with humans. In other species, the majority of ether-linkages are found within phosphoethanolamine-containing lipids where the fatty acid bonded at the sn-1 position can contain an acyl moiety (phosphatidylethanolamine, PE), an O-alkyl ether bond (plasmanylethanolamines, O-PE), or an O-alkenyl ether bond (plasmenylethanolamines, P-PE) (Fig 1A). Both O-PE and P-PE populations have been identified in C. elegans; however, their abundance is not yet defined [4, 5]. Many aspects of lipid metabolism including the distribution of lipid classes can change depending on the developmental stage, the reproductive status and the age of the animal. Here, we established day 3 of adulthood as our primary collection stage as these animals have completed development and reproduction (see S1 Fig for brood analysis). Therefore, we exclusively examined the adult phospholipidome. Furthermore, we used an established temperature-sensitive sterile strain, fer-15;fem-1, to eliminate the presence of progeny that can contaminate the lipid profiles [5].

Because phosphocholine- and phosphoethanolamine-containing lipids are by far the most abundant phospholipids in the nematode as well as in humans, we considered these classes specifically and identified all the species within by their exact mass, retention time and predicted isotope distribution [5, 17]. In order to ensure equal recovery of lipids with phosphocholine- and phosphoethanolamine headgroups, a panel of external standards was implemented to correct for differences in ionization efficiency in the mass spectrometer (see Materials and Methods). In day 3 adults, phosphatidylcholines are the most abundant phospholipid class in the nematode making up more than up 82.4±1.4% of the major phospholipids; however, we did not detect any significant ether-linked lipids, namely plasmanylcholines or plasmenylcholines, within the phosphocholine-containing lipids (Fig 1B). These species were also not found in significant amounts by a distinct mass spectrometry analysis completed by Shi et al [4]. The absence of ether-linked PC represents a difference in the nematode phospholipidome as humans contain significant amounts of ether-linked PC, particularly in cardiac tissue [2]. Among the PE population, the typical di-acyl lipids represent the majority of the population; however, there are significant amounts of O-PE and P-PE, representing 22.7 ± 1.4% and 30.3 ± 0.9% respectively of the phosphoethanolamine-containing lipids (see S1 Dataset). In considering these major classes, the contribution of ether-linked lipids represents approximately 10% of the major phospholipids in the adult nematode, making these ether-linked species major contributors to the overall phospholipid landscape (Fig 1B).

We further defined the ether-linked population by considering the species of fatty acid present at each position of the glycerol backbone. In C. elegans, the fatty alcohol attached at the sn-1 position is a saturated C18 species in nearly all ether-linked molecules as we have previously reported [5] (Fig 1C, see S1 Dataset for complete phospholipid profile). In contrast, the sn-2 position can incorporate a number of different fatty acid species, with the majority of molecules containing a C18 fatty acid with one (37.1 ± 3.5% for O-PE and 50.8 ± 2.8% for P-PE) or two double bonds (41.9 ± 3.2% for O-PE and 20.1 ± 1.8% for P-PE) (Fig 1C). There are substantial amounts of C20 polyunsaturated chains in the ether lipid pools of adult nematodes, particularly in the P-PE lipids, accounting for 25.3 ± 2.3% of that population (Fig 1C). There are subtle differences in the distribution of lipids reported by Shi et al; however, both analyses show 18:0/18:1 and 18:0/18:2 as the primary fatty acid combinations in both O-PE and P-PE pools [4]. Furthermore, the numbers cannot be directly compared due to the harvest of different stage animals and the implementation of distinct methodologies. Our profiling corroborates the major features of the ether-linked lipids and further establishes the nematode as a model for the study of ether lipids [4].

The presence of an alkyl or alkenyl bond in a phospholipid changes the properties of that molecule, specifically allowing tighter packing and reducing fluidity [2]. Isolation of loss-of-function mutations in ether lipid synthesis genes in C. elegans revealed an increase in the amount of saturated fatty acids in all lipid classes which were proposed to compensate for the increased fluidity predicted by ether lipid loss [4]. We sought to further examine the perturbations of the major phospholipids upon ether lipid deficiency using an RNAi model in order to examine the immediate effects of short-term ether lipid deficiency versus the long-term, multigenerational perturbation of the mutant animals. RNAi against fard-1, the proposed rate-limiting gene in ether lipid synthesis, was previously found to result in significant changes in ether lipid indicators including an accumulation of 18:0 fatty acid and a depletion of 18:0 dimethylacetal [4]. Here, we directly measured the ether lipids by HPLC-MS/MS and present a complete characterization of the membrane landscape when L1 stage animals were fed fard-1 RNAi until day 3 of adulthood (Fig 2A). In this developmental fard-1 RNAi model, we saw a dramatic depletion of ether-linked lipids with over 95% reduction in the abundance of both O-PE and P-PE classes (Fig 2C).

We also examined the defined lipid alterations, namely altered PE levels and increased saturated fatty acid content, in our developmental fard-1 model. Indeed, the fard-1 RNAi treatment drove a similar increase in the relative abundance of PE lipids (Fig 2C) similar to what has been observed in fibroblasts derived from RCDP patients and in fard-1 loss-of-function mutants [4, 18]. Additionally, we observe an accumulation of saturated C18 fatty acids in the total phospholipid compartment and in the neutral lipid population by gas chromatography-mass spectrometry (GC-MS) (see S2 Fig). The accumulation of saturated fatty acids in the neutral lipids is greater than in the phospholipid fraction indicating that the RNAi perturbation of ether lipid synthesis results in an impact on the overall fatty acid saturation within the nematode. The increase in saturated C18 may be a passive consequence of reduced ether lipid production as excess C18 chains would be available in the absence of ether lipid synthesis. Alternatively, it may represent an active remodeling of the nematode’s fatty acid profile to adapt to ether lipid absence, which would predict additional changes in the distribution of phospholipids.

To better understand this compensatory response, we examined the individual phospholipid species by direct sampling with HPLC-MS/MS. In doing so, we focused on the intact phospholipid molecules which allows for the determination of the overall amount of saturation within the lipid and not only the distribution of the chains throughout the class, which would be representative of data obtained from GC-MS. There are significant changes in 11 of the 21 major PE species (see S1 Dataset). To understand the impact on the saturation of the membrane, we binned the phospholipids by the number of double bonds contained within both of their fatty acid tails and found a major shift in the degree of saturation within the PE population (Fig 2D). Specifically, in L4440 control RNAi-treated animals, the largest proportion of PE lipids contain a total of 4 to 5 double bonds; however, in fard-1-treated nematodes, PE species containing 2 to 3 double bonds were the largest contributors to the pool (Fig 2D). Species with more than 4 double bonds are all significantly decreased, suggesting a remodeling of the population towards a less fluid state (Fig 2D). Interestingly, the alterations in PE parameters cannot be accounted for solely by the increase in saturated C18 as PE molecules such as 38:6 and 38:7 are significantly altered in the RNAi-treated animals and do not include C18:0 fatty acid tails (see S1 Dataset). The remodeling of the PE population, therefore, is driven by increased saturated fatty acid as well as by additional changes in the distribution of phospholipid species. Ultimately, these altered parameters may stabilize the membrane upon ether lipid loss.

In contrast, although there are minor changes in individual PC species, the overall characteristics of the PC population are relatively unchanged with fard-1 RNAi treatment (Fig 2D and S1 Dataset). This lack of PC remodeling is significant as PCs make up a far greater proportion of the membrane landscape (>80% PC versus <20% PE; Fig 1B). Taken together, the remodeling of the PE pool may ultimately reduce the fluidity of the membrane; however, it also must be considered that the PE remodeling may also contribute to local membrane structures and/or signaling pathways. In summary, HPLC-MS/MS demonstrates that although the amount of saturated fatty acid is increased in multiple lipid classes, there is more of an impact on the level of saturation in the PE class than in the PC populations, indicating that PE pools are more affected by ether lipid deficiency.

A key advantage of an RNAi-based model for ether lipid depletion is that the knockdown can be initiated at specific times during the lifespan of the nematode (Fig 2A). We, thus, can consider for the first time the specific role of ether lipid synthesis in the adult after the formation of membranes during development is complete. To do this, nematodes were grown on control (L4440 empty vector) bacteria until day 1 of adulthood, and, at this point, the animals were moved to fard-1 RNAi (Fig 2A). The distribution of ether lipids from animals treated with adult-only fard-1 RNAi were measured with HPLC-MS/MS, revealing a shift in the distribution of ether-linkages (Fig 2E). Specifically, the adult-only fard-1-treated animals have a 39.3% reduction in O-PE and 17.3% reduction in P-PE. The reduction in P-PE abundance was not statistically significant; however, this may be due to the relatively short duration of RNAi treatment (Fig 2E, see full list in S2 Dataset). Longer adult-only RNAi feeding did result in further perturbation of the ether lipid population as indicated by increased C18:0 abundance; however, these animals are significantly older (day 8 of adulthood) making direct comparisons of lipid profiles problematic (see S3 Fig).

As with the developmental fard-1 RNAi, we examined the impact that adult-only fard-1 depletion had on the PE lipids. Adult-only fard-1 RNAi resulted in more limited alterations in the lipids with significant changes in only 4 PE species compared to altered abundance of 11 PE species after developmental RNAi (see S2 Dataset). Moreover, the changes in individual species had a minimal impact on the overall saturation found in the lipid classes, and, in fact, there were no statistically significant changes in the distribution of double bonds within either PC or PE pools (Fig 2F). Despite this, there are trends within each population that mirror those seen with the remodeling of PE after developmental RNAi treatment. For instance, there is an increase in phospholipids with 2 to 3 double bonds and a decrease in species with 4 to 5 bonds (Fig 2F). The lack of significant remodeling may indicate an inability for the appropriate changes to be orchestrated in the adult animal.

Because the O-PE and P-PE abundances were less impacted after adult only fard-1 RNAi, we confirmed the efficacy of our RNAi strategy by qRT-PCR. The expression of fard-1 is decreased by nearly 60% in the adult-only RNAi protocol which is nearly the same reduction as the 65% observed during the developmental RNAi treatment (Fig 2B). The difference in RNAi efficacy can be explained by less efficient RNAi during adulthood in C. elegans; however, another consideration is the length of the exposure to RNAi. The developmental fard-1 RNAi-treatment results in the nematodes being fed RNAi for 4 days with the adult-only treatment resulting in only 2 days of exposure (Fig 2A). This setup was established in order to compare animals of the same age, as many parameters of lipid metabolism vary depending on the age of the animal. The duration of RNAi exposure should be considered by comparing distinct lengths of RNAi treatment with residual expression and ether lipid abundance in future studies.

To ascertain how ether lipid synthesis genes are expressed throughout the wild-type animals lifecycle, we measured the expression of fard-1 along with acl-7 and ads-1, additional ether lipid synthesis genes, in L1 stage animals versus adults. The expression of all the ether lipid synthesis genes was greatest in L1 animals and decreased significantly in day 3 adults. In fact, fard-1 expression was at least 13-fold higher in the L1 animals (Fig 3). The lower expression in adults argues that the less dramatic phenotypes after adult-only RNAi are a result of a reduced capacity to synthesize ether lipids in the adult stage and not exclusively because of the decreased RNAi efficiency. The trajectory of ether lipid synthesis gene expression highlights the importance of ether lipid production during development and helps to explain the more extreme phenotypes observed in developmental RNAi-treated populations.

To understand the role of ether lipid synthesis in adults, we examined the production of new ether lipids with a ¹⁵N-labeling strategy. The ¹⁵N-label is incorporated in PC and PE headgroups which both contain a single nitrogen atom, and the accumulation of ¹⁵N can thus be used to measure newly formed lipids [5]. Indeed, during adulthood, ¹⁵N accumulates at distinct rates within all of the measured phosphoethanolamine-containing lipids. Ether lipids with a combination of 18:0 and 18:1 fatty acid tails were used in this comparison as they are highly abundant in both alkyl and alkenyl forms, and these species met the standards for analysis with ¹⁵N-incorporation as previously discussed [5]. We found higher replacement of O-PE (18:0/18:1) than of P-PE (18:0/18:1) in adults (32.5±3.2% versus 16.2±4.4%) (Fig 4A). The faster replenishment of the O-PE compared to P-PE population is somewhat surprising as there are no defined roles for O-PE other than as a precursor for P-PE synthesis. Therefore, the conversion of O-PE to P-PE may represent a regulated and rate-limiting step of ether lipid production, and thus, the precursor alkyl population may accumulate in significant quantities to rejuvenate P-PE pools when needed. This model is supported by the specific decrease in O-PE pools that is observed with adult-only fard-1 RNAi treatment (Fig 2E). The production of both O-PE and P-PE species occurs at a lower rate compared to PE (36:2) as well as other PE populations previously measured [5]. A direct extension of this analysis would be to measure de novo production of ether-linked lipids in larval stages of the nematode; however, the rapid growth of the larval animals makes it difficult to distinguish new membrane formation from maintenance. The expression of ether-linked lipid synthesis genes in the early animal would suggest highly abundant production during development (Fig 3).

Adult-only fard-1 RNAi treatment resulted in a significant reduction in the amount of ¹⁵N found in the both O-PE and P-PE classes, indicating that the synthesis of new ether lipid species was significantly compromised by this adult-only RNAi strategy as predicted (Fig 4A). The altered production was specific to ether-linked lipids, as the synthesis of representative species from PC and PE were not compromised (Fig 4A). In fact, the production of PE (38:5) was significantly higher in the fard-1 RNAi-treated populations consistent with remodeling of the PE populations upon ether lipid synthesis deficiency (Fig 4A). The abundance of PE (38:5) is elevated in adult-only fard-1-treated animals; however, the increase is not statistically significant, and the increased ¹⁵N-incorporation may reflect faster turnover of this population (S2 Dataset). In summary, the adult-only RNAi results in dramatic reductions in O-PE synthesis (73% decreased) and P-PE synthesis (80% decreased). Interestingly, the overall abundance of P-PE remains largely intact despite the reduced capacity to make these lipids. It is possible that the O-PE population is converted to P-PE as a mechanism to preserve P-PE pools, and longer ether lipid deficiency is needed to observe a more dramatic reduction in P-PE. Regardless, adult-only RNAi of fard-1 is a viable model to probe the function of new ether lipid synthesis in adult membranes.

The RNAi treatments described here have distinct impacts on ether lipid metabolism that will be informative for future studies interrogating the function of ether-linked lipids in disease models. We summarize the biochemical perturbations observed with each RNAi treatment including the abundance of ether-linked lipids, their production and the impact of ether lipid deficiency on other lipid pools (Fig 4B). This comprehensive analysis allowed for the identification of key trends from a complex dataset and found, for example, that the PE remodeling was much more extensive after developmental RNAi treatment. Along with exploring the impact of ether lipid synthesis at different life stages, these models will also be useful in distinguishing the consequences of reduced ether lipid abundance in comparison to compromised synthesis. For instance, ether lipid deficiency results in sensitivity to oxidative stress; however, it is unclear whether ether lipid abundance or synthesis is the primary driver of this impact. The adult-only model can be useful in dissecting these parameters.

Ether lipids have been implicated in the response to oxidative stress in both mammals and nematodes; therefore, we considered the impact of stressed conditions on the ether lipids using HPLC-MS/MS analysis [2, 4]. To do so, we examined whether the ether lipid populations are directly impacted by exposure to 100mM paraquat (PQ) in wild-type animals. PQ produces elevated levels of oxidative stress, and the sacrificial antioxidant model of ether lipid function predicts that increased oxidative species would consume oxidant-labile molecules and specifically ether lipids. However, HPLC-MS/MS analysis demonstrated that there are no changes in the overall abundance of O-PE or P-PE populations after stress exposure (Fig 5A). We hypothesized that the relatively short exposure of PQ might not result in a detectable depletion of ether lipids; therefore, we also measured the phospholipids in wild-type animals exposed to a longer period of stress. Because PQ is ultimately toxic to the nematode, the 4-day PQ exposure was completed with a reduced concentration of PQ (25mM), and again there were no significant changes in O-PE and P-PE pools (see S4 Fig).

We hypothesized that new ether lipid synthesis could compensate for the consumption of ether lipids by reactive oxygen species (ROS) through the upregulation of de novo ether lipid production. As the adult-only fard-1 RNAi treatment dramatically compromised ether lipid synthesis, this treatment would severely limit the ability to make new ether lipids that could replace those consumed by ROS. In these adult-only fard-1 treated animals, there was reduced ether lipid abundance as described (Fig 2C); however, there was not any additional reduction due to the PQ exposure (Fig 5A). Therefore, it is clear that increased de novo synthesis of ether lipid pools is not responsible for the maintenance of ether lipid levels in wild-type animals exposed to stress. The stable abundance of ether lipids, even in the absence of new production, suggests that their primary role may not necessarily be as global antioxidants in the nematode.

Not only did we quantify shifts in the relative abundance in phospholipid classes with PQ treatment, but we also monitored for changes in the major phospholipid species in wild-type adults. There are no significant changes for any individual O-PE or P-PE species after PQ exposure (Fig 5B, see S3 Dataset for complete list). However, there are a number of significant changes in individual PC and PE species following oxidative stress treatment (Fig 5C). Most notably, we see a significant reduction of PC molecules that contain more than 7 double bonds, specifically PC 40:8, PC 40:9 and PC 40:10 (Fig 5C). This group is comprised of lipids that contain two C20 PUFAs, which are particularly sensitive to oxidative stress, and their depletion supports the efficiency of the PQ treatment implemented here. In the major PE lipids, there are significant increases in PE 38:5 and PE 38:6 populations after treatment (Fig 5C). Overall, the changes in the phospholipid landscape suggest that the primary effect of PQ exposure is a depletion of PC species with multiple PUFA tails and not the oxidant-mediated reduction of ether lipid pools.

We further probed whether compromised ether lipid synthesis impacted the properties of the other major phospholipids and their response to PQ exposure. This analysis was done in adult-only fard-1 RNAi-treated animals where the starting phospholipid composition is much more similar to wild-type adults. In these fard-1 animals, there were significant differences in the major PC and PE species; however, the altered species are the same as observed with control RNAi suggesting that the changes are independent of the ether lipid synthesis pathway (Fig 5C). Any subtle changes in the O-PE or P-PE populations would be magnified in fard-1 treated animals; however, the ether lipids were not significantly different than the controls after PQ treatment (Fig 5B). Our direct analysis of the biochemical changes following PQ treatment suggests that the primary effects of PQ stress are a decrease of PC species with C20 chains and a general increase in PE species, and both of these effects appear to be independent of ether lipid synthesis (Fig 5D).

The adult-only fard-1 RNAi treatment is a useful tool to examine compromised ether lipid populations as developmental fard-1 RNAi or loss-of-function mutation results in a near complete depletion of the ether lipid pools which is not reflective of many disease states. There is also significant remodeling of the PE population in developmental fard-1 RNAi animals that alters the baseline lipid distribution and may influence how these membranes would respond to insults, when compared to the wild-type membrane. The adult-only fard-1 RNAi treatment also affords an opportunity to examine the phenotypic consequences of perturbation of ether lipid synthesis depletion in a relatively uncompromised membrane. Unlike the developmental fard-1 RNAi (Fig 6B) and the fard-1 loss-of-function mutants [4], the adult-only fard-1 treated animals are only slightly and not significantly sensitive to oxidative stress (Fig 6A). Additionally, adult-only RNAi did not impact longevity (Fig 6C). Taken together, survival in stress and control conditions is dependent on larval expression of fard-1. We hypothesize that its primary role may be to generate an initial membrane landscape that enables juvenile animals to respond to complex sets of environmental stressors, which is supported by the higher expression of ether lipid synthesis genes in larval animals.

There are other stress conditions that may require the synthesis of ether lipids in adult animals, and we examined the role of fard-1 in heat stress and osmotic stress (Fig 6D and 6E). Although fard-1 is not required for survival in elevated salt conditions, it is clear that adult-only fard-1 plays a significant role in surviving elevated temperatures to the same extent as that was observed with developmental RNAi treatment (Fig 6E). Thus, the adult-only model will be of great utility in exploring how the phospholipid population and, in particular, the ether lipids respond to alternative insults.

The temperature-sensitive sterile strain, CF512, fer-15(b26);fem-1(hc17), was obtained from the Caenorhabditis Genetics Center (CGC, Minneapolis, MN) and maintained on E. coli (OP50) unless otherwise noted. CF512 animals were used for all experiments to allow for isolation of adults without larval contamination as previously described [5].

HT115 bacteria, transformed with empty vector control RNAi (L4440) or fard-1 RNAi from the Ahringer library, were grown in approximately 100mL LB (50μg/mL carbenicillin; 15μg/mL tetracycline) overnight. RNAi plates were prepared as previously described [5]. In short, bacteria were plated at a density of 0.15g per 10cm NGM-CI plate. L1 animals or young adult worms (44 hours from L1) were plated at a density of 2,500 worms per plate and incubated at 25°C for 48 hours.

For fatty acid analysis, total lipids were extracted and major lipid groups (phospholipid and neutral lipid) were purified by solid phase exchange chromatography. The resulting lipids were resuspended in sulfuric methanol to generate fatty acid methyl esters as described [25]. FAMEs were analyzed by gas chromatography/mass spectrometry (GC/MS) (Agilent 5975GC, 6920MS).

Total lipid fractions were isolated from frozen worm populations (5,000–10,000 animals) via chloroform:methanol (2:1) extraction [5]. Total lipid extracts were dried under nitrogen stream and then dissolved in 200–300μL of acetonitrile/2-propanol/water (65:30:5 v/v/v) dilution buffer [26]. As previously described, 10μL of each sample was injected onto a onto an HPLC system (Accela 600) equipped with a C18 Hypersil Gold 2.1 x 50mm, 1.9μm column (25002–052130; Thermo Scientific) and connected to an LTQ-Orbitrap mass spectrometer (Thermo Scientific) [5]. Analysis of the HPLC-MS/MS data was conducted using the software Lipid Data Analyzer (LDA) Version 1.6.2.5 [17]. The program utilizes exact mass, retention time, and predicted isotope distribution from full-profile, negative-ion-mode MS1 scans to identify lipids. The exact mass lists for PC, PE, O-PE and P-PE were provided to the program using the same molecular ranges previously described [5]. A 0.1% cutoff was applied in order to focus on the major phospholipid species in control animals, but, although it is only 0.08% abundant in control animals, we included the species P-18:0/20:2 in the final P-PE mass list, in order to compare the alkenyl form of the ether-linked lipid with its alkyl O-18:0/20:2 counterpart.

For phospholipid quantification, we employed a correction factor in order to correct for varying ionization efficiencies for the different lipid classes in our analysis. Because all species are not available, our correction factor was determined by the following representative species from Avanti Polar Lipids: 1,2-dilauroyl-sn-glycero-3-phosphoethanolamine (PE 12:0/12:0), 1,2-dilauroyl-sn-glycero-3-phosphocholine (PC 12:0/12:0), 1-(1Z-octadecenyl)-2-oleoyl-sn-glycero-3-phosphoethanolamine (P-PE 18:0/18:1), 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine (PE 18:1/18:1), 1,2-dioleoyl-sn-glycero-3-phosphocholine (PC 18:1/18:1), and 1,2-dilignoceroyl-sn-glycero-3-phosphocholine (PC 20:4/20:4). When possible we included phospholipids with various fatty acids to account for differences in chain length and shape; however, we did not see any correlation with short versus long chains or different degrees of unsaturation. In summary, for each HPLC analysis, we used a correction factor 0.67580026 to account for the higher ionization efficiency of the PC’s over the PE population.

Nematodes (~50μL pellet) were harvested after 0 hours (L1 stage), 44 hours (day 1 adult) and 92 hours (day 3 adults) on OP50. RNA was isolated using TRIzol (Invitrogen) and purified by RNeasy Mini Kit (Qiagen) and DNase I solution (Qiagen), and cDNA was made by ProtoScript First Strand cDNA Synthesis Kit (New England Biolabs). A 10μL reaction consisted of 5ng of cDNA, 1X Power SYBR Green PCR Master Mix (Life Technologies), and 5pmol each of the forward and reverse primers. The qRT-PCR reactions were executed on ABI 7900HT Fast Realtime PCR platforms and the cycle threshold (CT) values were exported from the Sequence Detection System 2.3 software. qRT-PCR efficiency values were calculated from mean CT values and were then used to calculate the relative expression of each gene as described [27], normalized to the expression of the housekeeping gene, cdc-42.

To quantify the ability of the animals to survive in high oxidative stress conditions, 100–200 day 3 adult worms were exposed to 100mM paraquat (PQ) in M9 buffer. Samples were placed on horizontal shaker at room temperature, and survival percentage was determined in a blinded manner. Residual bacteria may have been present after transfer; however, bacteria was not added to the PQ treatment for either fard-1 or L4440 RNAi conditions. Animal death was determined when worms did not respond after 2 gentle prods with a worm pick. HPLC-MS/MS analysis requires a minimum of 5,000 adult animals; therefore, larger-scale PQ treatment was performed at the same final concentration. From the HPLC-MS/MS studies, an aliquot was counted to obtain corresponding stress sensitivity numbers.

Lifespan assays were conducted as follows: L1s were plated onto OP50 and allowed to grow for 44 hours. The day 1 adults were then transferred to NGM-CI plates seeded with fard-1 or L4440 RNAi bacteria. The worms were then counted each day and transferred to fresh plates every 2 to 3 days. To test for sensitivity to osmotic stress, day 3 animals (~50) were transferred to high salt plates (500mM NaCl) and scored every 24 hours for viability as described above. The NaCl plates were seeded with fard-1 or L4440 control bacteria. Thermotolerance was determined by moving day 3 adults on seeded plates to a 35°C bath for 24 hours. Animals were allowed to recover for 24 hours at 25°C before assayed for viability.

¹⁵N feeding plates were prepared as previously described [5]. Briefly, LB (¹⁴N) and ¹⁵N–Isogro (Sigma) cultures were inoculated with OP50 and grown overnight. Once harvested, the 0.15g of bacteria were plated on individual agarose plates. Approximately 5,000 synchronized worms were harvested after 18 hours on isotope feeding plates, and lipids were extracted for analysis on HPLC-MS/MS. ¹⁵N-isotope incorporation rates were determined as previously described using the Thermo Qual Browser Software in Xcalibur version 2.2 (Thermo Scientific). New phospholipid synthesis was calculated by the following formula: [(Σm+1,…,m+4)/(Σm+0,…,m+4)]*100 as reported in [5].