Strains

C. elegans strains were maintained and cultured under standard laboratory conditions at 20°C with the food source E. coli OP50 as previously described [28]. Specific alleles of ogt-1 and oga-1 used in this study have previously been characterized, including ogt-1(ok1474), ogt-1(ok430) [22], oga-1(tm3642) [29], and oga-1(ok1207) [23]. The alleles ogt-1(jah01), oga-1(av82), and oga-1(av83) were generated during the course of this study using CRISPR/Cas9, as described below. The strain daf-16(mu86) was a gift from Dr. Michael Krause. We are grateful to Dr. Mark Van Gilst and Dr. Todd Lamitina for kindly gifting us the nhr-49(nr2041) allele and the OG1139 ogt-1(dr84[ogt-1::GFP] dr89[K957M]) strain, respectively. The strains RB754 aak-2(ok524), RB1988 acs-2(ok2457), TJ1052 age-1(hx546), BX153 fat-7(wa36), RB1206 rsks-1(ok1255), VC199 sir-2.1(ok434), and QV225 skn-1(zj15) were obtained from the Caenorhabditis Genetics Center (CGC, University of Minnesota, Minneapolis, MN) strain bank, which is funded by the NIH Office of Research Infrastructure Programs (P40 OD010440). All of these strains were backcrossed to our wild-type N2 Bristol laboratory strain a minimum of four times prior to analysis to ensure homogenous genetic background.

Genotyping was conducted using nested PCR, as described by the CGC, or using sequencing (Eurofins Genomics, Louisville, KY) in the case of point mutants. Primers for ogt-1(ok1474), ogt-1(ok430), oga-1(tm3642), and oga-1(ok1207) are described elsewhere [23, 29] and are available upon request. Primers for aak-2(ok524), acs-2(ok2457), rsks-1(ok1255), and sir-2.1(ok434) were derived from the sequences available on wormbase.org and CGC. Primers for daf-16(mu86), nhr-49(nr2041), ogt-1(dr84[ogt-1::GFP] dr89[K957M]), and the point mutants age-1(hx546), fat-7(wa36), and skn-1(zj15) were generated using Primer3 (Rozen and Skaletsky 2000) and are available on request.

Double mutants were generated by crossing, with the presence of deletion alleles verified either using nested PCR or by sequencing, in the case of point mutants. Double mutants were backcrossed a minimum of four times to our wild-type N2 Bristol strain before use in analysis.

Generation of CRISPR/Cas9 whole gene knockouts of ogt-1 and oga-1: Ogt-1(jah01), oga-1(av82), and oga-1(av83)

The alleles ogt-1(jah01), oga-1(av82), and oga-1(av83) were generated during the course of this study using CRISPR/Cas9 genome editing. These alleles are full gene deletions of the ORF of ogt-1 and oga-1, respectively. We used the direct-delivery protocol developed by the Seydoux lab [30] and screened for edits using the co-conversion dpy-10 method [31]. The crRNAs and repair template sequences are listed below. The repair template contains the deletion and perturbs the PAM sites to prevent Cas9 from re-cutting. From 15 injected P0s, we identified 3 “jackpots” and picked a total of 40 rollers from them. Using PCR genotyping, we identified and isolated three heterozygous lines from each. Homozygous animals were identified in the F2 generation and subjected to sequencing of the relevant ogt-1 or oga-1 region. For ogt-1, two lines contained the expected full deletion, which we subsequently named ogt-1(jah01). With oga-1, two of the three lines contained the expected deletion, while the third contained an additional ~17bp deletion. The two expected oga-1 deletions were named oga-1(av82), which were the strains used in our analysis. We also kept the larger oga-1 deletion and named it oga-1(av83).

ogt-1 sequences:

Targeted gene crRNA1 (ogt-1 deletion N-terminal)

5’–AAUUUCAGAAUUAUAUCGAAGUUUUAGAGCUAUGCUGUUUUG– 3’

Targeted gene crRNA2 (ogt-1 deletion C-terminal)

5’–GCUUGUGAAUAGAUUUUCGAGUUUUAGAGCUAUGCUGUUUUG– 3’

Repair template:

5’–Ccaatagttcaatttccaaatttcagaattatatcacacggcttgtgaatagattttcgaaggattttta– 3’

oga-1 sequences:

crRNA1 (oga-1 deletion N-terminal):

5’–gttcacagtcaatacttaat– 3’

crRNA2 (oga-1 deletion C-terminal:

5’–caaaatcgttgaatttgact– 3’

Repair template:

5’–ttttattaagagtgaaagttgatgacaacatttcctgattcaaattcaacgattttgttccaaccgggatat– 3’

Adult reproductive diapause

Adult reproductive diapause experiments were performed as described previously [13], with minor adjustments. The overall experimental approach for initiation of ARD is illustrated in S1 Fig. Briefly, nematodes were cultured at 20°C on nematode growth medium (NGM) with E. coli OP50 as the food source at a density of 3,000 per 60 mm tissue culture dish. Worms were synchronized using hypochlorite bleaching of gravid worms with a dilute alkaline hypochlorite solution, as previously described [32]. Eggs harvested from bleaching of gravid worms were then left with gentle rocking overnight at room temperature in M9 solution. Hatched L1s were then plated on NGM plus OP50 plates to complete synchronization. This synchronization process was repeated three times.

Following the third synchronization, 10,000 L1 worms (plate counting visually by sectors) were grown on a single 100 mm NGM + OP50 plate and visually inspected at intervals to ensure that over 90% of worms were in mid-L4 (approximately 48 hr, depending on growth of strain). Mid-L4 worms were then washed off the plate with M9 and collected in a 15 ml conical tube. The tubes were spun down at 20°C for 2 min at 1500 rpm. This washing was repeated 5 times with 5 ml M9 to ensure all bacteria was removed. Prior to the fifth and final spin, worms were incubated in 5 ml M9 with gentle rocking at room temperature for 30 min to remove bacteria from gut. Following the 30 min incubation, all media was aspirated and the nematodes were suspended in 1 ml of M9, counted, and plated on high agarose tissue culture plates without bacteria/food source (ARD plates). ARD plates were prepared and poured as follows: 1 L: 3 g NaCl and 15 g agarose autoclaved in 1 L water. After autoclave sterilization, filter sterilized solutions of the following were added: 1 ml of 1M CaCl₂, 1 ml of 1M MgSO₄, 25 ml potassium phosphate buffer (pH 6.0), and 1 ml cholesterol (8 mg/ml) [13]. Using this solution, 60 mm plates were filled to two-thirds full in the hood to ensure sterility. After plates were cooled and equalized to 20°C worms were added at a density of 3,000 worms per 60 mm plate. As a control, 3,000 of the synchronized and washed mid-L4s were plated separately from the worms that were put on the ARD plates and were instead plated on NGM plates with OP50 to allow analysis of lifespan and brood size in parallel to starved worms used to assess the ARD phenotype.

Worms were visually inspected every 24 hours (up to 72 hours, depending on strain) after plating and characterized as either in ARD, in L4, or bagged. At day 5, day 10, and day 30 after mid-L4 worms were plated on ARD plates, worms were subjected to Oil Red O (ORO) staining, carminic acid staining, lifespan analysis, and brood size counts as described below. Each experiment was repeated a minimum of three times from start to finish.

Lifespan, germline, and brood analysis

At least 10 worms from each ARD plate were collected individually at day 5, day 10, and day 30 post-plating on ARD plates. The worms were visually inspected to ensure ARD status and placed singly onto individual NGM plus OP50 plates and incubated at 20°C to monitor for restoration of the germline, analysis of overall number of progeny following ARD (brood size), and post-ARD lifespan. Restoration of the germline was followed visually at day 5, day 10, and day 30 using standard microscopy techniques on a Zeiss LSM 700 confocal microscope (Carl Zeiss Microscopy, LLC, Thornwood, NY) with a Plan-Apochromat 20x/0.8 objective lens. Brood size was monitored by carefully transferring the individually recovered worms every 24 hr to fresh NGM plus OP50 plates until the end of their reproductive phase, marked by the absence of eggs. Progeny of the post-ARD worms were counted after each transfer and added to total brood size reported for each animal. For all experiments, any ARD worm that failed to thrive in the first 48 hr or disappeared was censored.

The post-ARD (recovered) worms were then inspected daily after the reproductive period to determine lifespan. When worms no longer responded to gentle poking with a platinum pick, they were scored as dead. Lifespan counts for post-ARD animals includes the days in ARD, such that the lifespan post-re-feeding was added to the number of days the animals were starved. Lifespan was graphed, and statistics were performed as previously described using GraphPad and one-way ANOVA [29]. A minimum of 40 ARD animals from four independent experiments were used for analysis of each strain.

The controls also underwent the same analysis but without ever being placed on the ARD plates and, instead, were grown on standard NGM plus OP50 plates at 20°C.

Oil red O staining

The protocol for Oil Red O (ORO) staining was adapted from previously described methods [33] with minor adjustments. Prior to starting experiment, a stock solution of 0.5 g/100 ml ORO in isopropanol was allowed to equilibrate for at least two days at room temperature with gentle rocking and light protection. On the day of the experiment, a fresh dilution of ORO stock solution was diluted to 60% with water and rocked for at least 1 hr. Dilute ORO was then filter sterilized with a 0.22 μm filter.

Nematodes were washed off NGM plates with OP50 (control) or high agarose ARD plates with 1 ml of 1x PBS + 0.01% Triton X-100 (PBST). Worms were washed 3 times with 1 ml of PBST and allowed to settle by gravity after each washing. Cuticles were permeabilized by resuspending worms in 120 μl of PBST and 120 μl of 2x MRWB buffer (160 mM KCl, 40 mM NaCl, 14 mM Na₂EGTA, 1 mM spermidine-HCl, 0.4 mM spermine, 30 mM Na-PIPES pH 7.4, and 0.2% β-mercaptoethanol) with 2% paraformaldehyde. Worms were gently rocked for 1 hr at room temperature. Animals were allowed to settle by gravity, buffer was aspirated, and animals were washed with 1x PBST twice. Worms were then resuspended in 60% isopropanol and incubated 15 min with gentle rocking at room temperature to dehydrate. Worms were allowed to settle by gravity and remaining isopropanol was removed. 1 ml of 60% ORO was added and the worms were incubated overnight with gentle rocking. After the overnight incubation, ORO was removed and worms were resuspended in 100 μl of 1x PBST. 1 μl 1x DAPI was then added to the 20 μl PBST, as an internal control for permeabilization, and flicked to mix. Worms were then mounted on slides on an agarose pad and visualized with Zeiss LSM 700 confocal microscope using either a Plan-Apochromat 10x/0.45 or Plan-Apochromat 20x/0.8 objective lens, as described below.

Carminic acid staining

Carmine dyes such as carminic acid are used to complex with glycogen and short glucose polymers to produce a fluorescent signal responsive to the levels of glucose polymer. Thus, it is a useful tool to analyze glycogen storage [22]. Carminic acid was introduced to worms through feeding as previously described [22]. Briefly, OP50 was inoculated into LB broth with 1 mg/ml carminic acid and grown overnight at room temperature in a light protected vial. OP50 with carminic acid was then used to seed standard NGM plates. Animals were plated on carminic acid containing plates following initial mid-L4 harvest (controls) or post-ARD at day 5, day 10, and day 30 and were visualized with Zeiss LSM 700 confocal microscope using a Plan-Apochromat 20x/0.8 objective lens, as detailed below.

Microscopy & image analysis

Microscope images were acquired on a Zeiss LSM 700 confocal microscope using either Plan-Apochromat 10x/0.45 or a Plan-Apochromat 20x/0.8 objective lens. In all cases: photomultiplier tube (PMT) offset was adjusted so that zero photons corresponded to a pixel value of zero; and PMT gain was adjusted to fully utilize the available 8-bit dynamic range while avoiding saturating the images. The same settings were used across all images being compared.

When acquiring the (non-fluorescence) ORO images, sequential scans were performed with the 488nm, 555nm, and 639nm lasers while using the transmitted light detector to produce an RGB image like what one would acquire using a color camera under widefield illumination from white light. Differential Interference Contrast (DIC) optics were also in place while imaging to provide contrast sufficient to distinguish sample anatomy at the same time as recording ORO. Either the 10x/0.45 or the 20x/.8 objective lens was used for ORO imaging—there was no difference in quantitation because the numerical aperture of the condenser was limited to <0.45, and kept constant across experiments.

When acquiring fluorescence images of carminic acid, a 555nm laser was used for excitation. The emission was passed through a laser blocking filter and a wide open pinhole to produce a non-confocal image. The emission was further filtered through a SP640 (640 nm short pass) emission filter, to produce a widefield fluorescence image appropriate for quantitating total signal from the sample volume. Simultaneously with the fluorescence channel, a DIC image was also acquired. Only images acquired with the 20x/0.8 lens were used for quantitation.

For quantitative image analysis, Fiji software [34] was used, employing in-house macros customized to streamline the processing for the user. For both ORO and carminic acid, regions of interest (ROIs) were manually traced very roughly around the edges of each worm, further refined by automatic segmentation, then the average intensity per pixel was calculated. This mean pixel value was divided by the diameter of the worm, measured manually by the user, to provide a relative value for quantitating mean carminic acid fluorescence per unit volume. For the ORO (RGB) images, a more in-depth analysis was performed, converting the intensities to relative absorbance (pseudo-fluorescence) at each pixel location, then combining with automated physical measurements of the worm from the ROI, to calculate an average ORO density per unit volume. Quantitative comparisons could then be made (S1 Methods). For presentation, all images had their intensities re-scaled to maximize the dynamic range and visual appeal of the images. Images of ORO and fluorescence that were being compared were rescaled by the same amount in those channels.

Diverse nutrient sensors are required for ARD entry

In the initial characterization of ARD, only one allele impacting ARD entry was identified:the nhr-49(nr2041) mutant allele of NHR-49, a protein involved in fatty acid metabolism [13]. Subsequent studies have identified additional pathways which regulate ARD in a variety of ways, and one has called into question the impact of nhr-49 on ARD entry [35, 36]. We sought to identify additional regulators of ARD dynamics, focusing on the three distinct phases of ARD: initiation (proportion of animals entering ARD vs bagging or L4 fate), maintenance (germline shrinkage and retained embryos), and recovery (post-ARD lifespan, germline regrowth, and brood size) [13] (Fig 1A). Using a candidate gene approach, we focused on key nutrient-responsive pathways. The pathways profiled included AMP/energy signaling (AAK-2/AMPK), insulin-like signaling (AGE-1/PI(3)K, DAF-16/FoxO), TOR (RSKS-1/S6K), stress signaling (SKN-1/Nrf), carbohydrate metabolism/hexosamine signaling (OGT-1/OGT and OGA-1/OGA), and sirtuins (SIR-2.1/SIRT1). The alleles analyzed were chosen based on their viability under the experimental conditions and that they have been previously characterized for their roles in nutrient signaling and longevity.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. ARD stages and entry defects in nutrient sensing mutants.

(A) ARD stages including initiation, remodeling/maintenance, and recovery. Initiation occurs when worms are removed from a food source during mid-L4. With wild type worms, approximately 1/3 enter ARD, 1/3 have hatched offspring inside (bagging), and 1/3 maintain an L4 state during the first 72 hours of starvation. Once ARD is initiated, worms enter the maintenance phase wherein the germline shrinks and the characteristic retained embryo(s) are observed. Recovery occurs after starved worms are re-fed. During this stage the germline expands and a normal lifespan commences (adapted from [13]). (B) The percentage of worms entering ARD as opposed to bagging or L4 fate were quantified for the respective mutants. For single mutants ARD initiation was significantly defective in the following strains, as compared to N2: nhr-49(nr2041), aak-2(ok524), oga-1(av82), ogt-1(jah01), ogt-1(ok430), ogt-1(ok1474), and sir-2.1(ok424). The double mutants nhr-49(nr2041);oga-1(av82), nhr-49(2041);ogt-1(ok1474) and ogt-1(jah01);oga-1(av82) also showed low ARD entry. A minimum of 300 worms from 3 different experiments were counted and statistics were performed in Graph Pad Prism using an Ordinary one-way ANOVA. P-value **** = <0.0001, ** = 0.001, NS = not significant. Box and wisker plots show median, quartiles, and range.

https://doi.org/10.1371/journal.pone.0274076.g001

ARD initiation was clearly discernible by 48 to 72 hours after mid-L4 worms were placed on ARD plates and could be distinguished from the bagging or L4 fates by the morphology of the worms, particularly the presence of one to two embryos retained in the uterus of adult worms [13] (S1 Fig). We found that age-1(hx546), daf-16(mu86), rsks-1(ok1255), and skn-1(zj15) did not significantly alter ARD initiation, suggesting that insulin-like signaling, the TOR pathway, and stress response are not major contributors to entry (Fig 1B), in agreement with another recent report [35]. In contrast, aak-2(ok524), ogt-1(ok430), ogt-1(ok1474), sir-2.1(ok434), and our positive control nhr-49(nr2041) all contribute to the initial sensing and response to starvation during ARD entry. These mutants displayed a significant reduction in the ability of nematodes to successfully initiate ARD (Fig 1B, S2 Fig).

We were intrigued to find that ogt-1 mutants exhibited severe ARD entry defects similar to nhr-49(nr2041) (Fig 1B, S2 Fig). We verified this finding by generating and analyzing targeted full gene deletions, ogt-1(jah01) and oga-1(av82). These new CRISPR/Cas9 alleles were confirmed to have similar phenotypes as the previously characterized ogt-1 and oga-1 alleles, including lifespan and brood size under normal laboratory conditions (S1 Table). As with the other two ogt-1 alleles, we observed a significant ARD entry defect with ogt-1(jah01) (Fig 1B). Because of this, we used the ogt-1 alleles interchangeably. Both the oga-1(ok1207) and newly-generated oga-1(av82) alleles show a decrease in mean ARD entry, but this effect was milder than for ogt-1 and was only significant with the CRISPR allele (Fig 1B, S2 Fig). For this reason, we continued with the full gene deletion oga-1(av82) for further analysis. The milder and allele-specific phenotypes of the oga-1 mutants suggested that removal of O-GlcNAc does not have as strong of an influence on ARD as addition of O-GlcNAc. Considering this and recent findings demonstrating non-catalytic roles of OGT-1 in C. elegans [37, 38], we tested ARD entry of worms with the catalytic-dead mutation ogt-1(dr89[K957M]). Interestingly, we found that this point mutant line entered ARD at similar rates to N2, demonstrating this starvation response phenotype does not require the ability of OGT-1 to transfer O-GlcNAc, suggesting it instead depends upon a non-catalytic role of the protein (Fig 1B).

Associated with ARD initiation, we noted significant differences between the mutant strains in the distribution of animals adopting either the L4 or bagging fate. Both L4 and bagging had a higher degree of variability than ARD entry, in general, suggesting that the ARD entry may be more tightly regulated than the other two fates (S2 Fig). Interestingly, some of the mutant alleles showing reduced ARD entry (nhr-49(nr2041), ogt-1(ok1474), and oga-1(ok1207)) did so while producing sizable numbers of animals adopting both the L4 fate and bagging. The other strong entry defective alleles, aak-2(ok524) and sir-2.1(ok434), produced predominantly L4 arrested larvae with few bagging or in ARD (S2 Fig), suggesting there may be alternative mechanisms at play in the regulation of entry between mutants and, therefore, distinct pathways.

Based on the similarity of phenotypes for the single mutants of nhr-49 and ogt-1, in addition to their known overlapping roles in metabolism, we next chose to explore the genetic interaction between these mutants by analyzing the ARD initiation phenotypes in double mutants. Three double mutant lines were generated to test the interactions between nhr-49, ogt-1, and oga-1. nhr-49(nr2041);oga-1(av82), nhr-49(nr2041);ogt-1(ok1474), and ogt-1(jah01);oga-1(av82) each showed severe ARD entry defects similar to nhr-49(nr2041) single mutants (Fig 1B). Because the ARD entry rate is so close to zero for nhr-49 and ogt-1 mutants, it was difficult to discern between an additive effect of parallel pathways or the effect of the more severe allele in a single pathway. Therefore, our epistasis experiments were unable to determine whether nhr-49, oga-1, and ogt-1 function in the same pathway to regulate ARD entry or each have independent roles.

Distinct nutrient sensors are required for ARD recovery

Successful maintenance of ARD can be determined by examining germline shrinkage and embryo preservation that occurs after entry into ARD. Characteristics of the recovery stage of ARD is marked by dynamic germline regrowth, production of progeny (brood size), and post-ARD lifespan [13]. Thus, to further define which nutrient sensors regulate each of these stages of ARD, we next looked at the dynamics of the germline, brood size, and lifespan.

Despite significant changes in ARD entry, we found that all mutant strains exhibited robust germline shrinkage once ARD was established, analogous to wild-type worms and the retained embryos were observable throughout the starvation period (Fig 2, left panel]. Further, when worms were re-fed for two days following 30 days of starvation (post-ARD), a robust regrowth of the germline was observed in all strains (Fig 2, right panel). Thus, in this set of nutrient sensors, dynamic germline shrinkage and regrowth is a general feature of ARD once it is established, as is the maintenance of the retained embryos [13, 36]. These results suggest that the maintenance of ARD and germline shrinkage may be independent of nutrient sensing and support the hypothesis that the distinct phases of ARD may be genetically separable [13].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Germline shrinkage during ARD maintenance and regrowth during recovery.

Representative images (with magnified panels for detail] showing the germline (outlined in yellow] of worms starved (left panel] at day 30 of ARD, with uterine embryos noted with white arrows. The right panel shows worms that have been re-feed for 48 hours after 30 days of ARD, with white dashed lines to indicate the germline, where healthy embryos and/or oocytes can be seen. All strains showed robust shrinkage of the germline during starvation and regrowth after feeding. The 50 μm scale bar shown in the top row applies to all images in each column.

https://doi.org/10.1371/journal.pone.0274076.g002

As noted previously, post-ARD animals retain their ability to reproduce, albeit to a significantly diminished degree. However, this was found to be largely dependent on the amount of sperm retained in the hermaphrodites, with selfing producing significantly less offspring than mating post-ARD [13]. In our analysis of brood size following ARD, we found that there was a high level of variability in the number of progeny, even between replicates of the same strains within the same experiment (S1 Table). Surprisingly, reduced brood size did not correlate with reduced sperm counts (S3 Fig) nor with the germline dynamics we observed (Fig 2). This suggests that the regulation of ARD recovery as measured by brood size is complex and may not correlate to sensing and responding to nutrient changes but may involve other processes. We therefore chose to focus on post-ARD lifespan as a measure of ARD recovery.

Post-ARD lifespan is a means to quantify the survivability of ARD, or the ability to recover and thrive following re-feeding after time spent in ARD [13]. Thus, we analyzed the lifespan of worms recovered after 30 days of ARD using standard lifespan assays. We found that aak-2(ok524), daf-16(mu86), rsks-1(ok1255), and skn-1(zj15) mutants had significantly reduced post-ARD lifespans (Fig 3). A concurrent study also found these genes reduced post-ARD lifespan, with the exception of skn-1 [35]. Based on tour analysis, we suggest that the TOR pathway most impacts recovery.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Lifespan of mutant strains in recovery after 30 days of ARD and summary model.

Lifespan of wild-type and mutant strains in control husbandry conditions (blue bars) compared to those recovered after 30 days of ARD (gray bars). Lifespan after recovery from ARD was significantly defective in aak-2(ok524), daf-16(mu86), and rsks-1(ok1255). The mutant skn-1(zj15) also had a significantly reduced lifespan post-ARD but not as strongly as the others. A minimum of 8 worms from independent experiments were analyzed for lifespan and statistics were performed in Graph Pad Prism using one-way ANOVA. P-value **** = <0.0001, * = 0.0448. Error bars represent standard deviation.Overall, we found that whereas nhr-49, oga-1, ogt-1, and sir-2.1 play a role in ARD initiation, daf-16, rsks-1, and skn-1 function primarily in recovery from ARD, with aak-2 playing a role in the regulation of both states. Conversely, none of the alleles that we explored significantly impacted maintenance of the ARD.

https://doi.org/10.1371/journal.pone.0274076.g003

Mitochondrial β-oxidation but not fatty acid desaturation is required for ARD entry

Due to the strong phenotypes we observed for ARD entry, the role for O-GlcNAc signaling, and the similarity in phenotypes for nhr-49 and ogt-1 mutants, we next explored this entry pathway in more detail. We chose to focus on ogt-1 and not oga-1 for the rest of the experiments due to the variability of oga-1 alleles and the weaker entry phenotype.

NHR-49 transcriptionally regulates both fatty acid desaturation via fat-7, and mitochondrial β-oxidation via acs-2 [39] (Fig 4A). To determine if these NHR-49-dependent pathways is relevant for ARD entry we analyzed acs-2(ok2457) (mitochondrial beta-oxidation) and fat-7(wa36) (fatty acid desaturation) mutants for ARD phenotypes. Whereas the acs-2 mutant and acs-2;ogt-1 double mutant had a significant defect in ARD entry, mutant fat-7 entry was comparable to wild-type. Similar to the nhr-49, oga-1, and ogt-1 single mutants, we did not observe a significant difference in lifespan of the fat-7 or acs-2 mutants post-ARD (Fig 4C). This data diverges slightly from a report which showed acs-2 and fat-7 mutants had reduced post-ARD lifespan [35]. These effects were relatively minor, and fat-7 was not tested alone but in combination with a fat-6 mutation [35], which may explain why our findings differ. Our findings further implicate lipid metabolism and suggest that acs-2 regulates ARD entry to a greater extent than recovery.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Fatty acid beta-oxidation, but not desaturation, plays a role in ARD entry.

(A) NHR-49 regulates both fatty acid beta-oxidation (ACS-2) and desaturation (FAT-7) via genetically separable pathways that are known to generate distinct phenotypic outputs. (B) ARD entry was significantly impaired with acs-2(ok2457) but not with the fat-7(wa86) mutant. The double mutant acs-1(ok2457); ogt-1(ok1474) was also significantly impaired for ARD entry. P-value **** = <0.0001 in comparison with N2 as determined using one-way ANOVA, NS = not significant. (C) Lifespan under standard conditions and with refeeding at day 30 of ARD was not significantly altered in these mutants as determined by one-way ANOVA. Error bars represent standard deviation.

https://doi.org/10.1371/journal.pone.0274076.g004

Neutral lipid changes associated with ARD entry

Changes in lipid metabolism play a prominent role in multiple pro-longevity pathways [40]. NHR-49 and the O-GlcNAc cycling enzymes have well-established roles in the regulation of fatty acid metabolism [39, 41, 42]. To determine the effects of ARD-regulating pathways on lipid storage,we used the lysochrome dye Oil Red O (ORO), which stains triglycerides (TAGs) [33]. We examined the various strains for ORO staining in control (fed conditions) and after 30 days in ARD (prolonged exposure). Fed ogt-1(ok1474) worms showed significantly lower ORO staining compared to N2s (Fig 5A and 5B), as has been reported previously with ogt-1 mutants [22]. ORO staining is present throughout the body of worms in fed conditions, but after 30 days in ARD, staining is greatly diminished throughout the body with the exception of the retained embryo (Fig 5A and 5B). Between the fed and ARD conditions, ogt-1(ok1474) animals did not show a significant decrease in quantified ORO staining, likely due to the already-low lipid stores in this strain. Additionally, the quantification was normalized by volume (see S1 Methods), which may underestimate the effect of ARD on total triglycerides per animal, as worms in ARD are significantly smaller than fed worms.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Oil Red O staining reveals redistribution of lipid stores in ARD.

(A) Representative images showing the distribution of Oil Red O in control (top) and starved animals after day 30 of ARD (bottom) in the ogt-1-dependent pathway. Retention of TAGs appeared to be strongest in the retained embryos in all mutant strains (white arrows). Scale bar indicates 50 μm and is consistent across each row. (B) Fiji Image analysis-based quantification of ORO staining (see S1 Methods) of wildtype and the mutant strains of the ogt-1-dependent pathway in normal husbandry (control) and in day 30 of ARD. Using two-way ANOVA analysis, both variables (genotype and feeding condition) were found to have significant effects, and these variables interact significantly. For each strain, mean ORO density decreased after 30 days of ARD: most dramatically for nhr-49(nr2041) and not significantly for ogt-1(ok1474). In normal husbandry conditions, ogt-1(1474) showed significantly less ORO signal than did N2 worms. P-value **** = <0.0001, *** = <0.001, ** = <0.01. Error bars represent standard deviation.

https://doi.org/10.1371/journal.pone.0274076.g005

Notably, we observed strong depletion of lipid stores as early as 24 hours of starvation (S4A Fig), suggesting it is an early response. This is supported by another study which also noted decreased ORO signal in ARD worms at 48 hours of starvation [35]. As a control, we also monitored cuticle penetrance of a small molecule dye using DAPI which revealed comparable penetrance between strains (S4B Fig). We further observed that in other strains analyzed, those without major effects on ARD entry (age-1(hx546), fat-7(wa36), and skn-1(zj15)) showed higher levels of ORO staining throughout the animal in ARD than wild-type worms in ARD (Fig 5, S5 Fig). Strains with significant ARD entry defects (acs-2(ok2457), sir-2.1(ok424), and oga-1(tm3642)) showed reduced ORO staining when in ARD, with characteristic retention of lipids in the one or two uterine embryos (S5 Fig). This correlation suggests lipid stores play an important role in the initiation of this form of diapause.

Based on the interplay between carbohydrate and fatty acid metabolism, as well as the role of O-GlcNAc in carbohydrate metabolism, we next looked at glycogen storage using carminic acid staining. Under control husbandry conditions, our analysis suggested a change in the pattern of glycogen storage in nhr-49(nr2041) and ogt-1(ok1474) worms, despite fluorescent signal having no significant difference between strains (S6 Fig). After 30 days of ARD, these stains all showed higher carminic acid signal normalized by volume. This may be partially attributable to the reduction in size associated with ARD, and suggests glycogen stores are relatively unaffected or increased by ARD. Interestingly, nhr-49(nr2041) worms in ARD showed dramaticly higher carminic acid staining (a 3-fold increase compared to N2) than did non-ARD worms (S6 Fig). With the findings of lipid reduction in ARD, this suggests glycogen stores may be maintained or increased at the expense of lipid stores while in ARD. These results suggest ARD has effects on glycogen storage, and glycogen storage does not correlate with the ARD entry defects we see in nhr-49 and ogt-1 mutant strains.