Tissue-specific autophagy responses to aging and stress in C. elegans

A new assay for measuring autophagy in C. elegans

We developed an immunoblot-based assay of autophagy for C. elegans. Similar assays of autophagy are commonly used in yeast, plants, and mammals [20, 33–36]. In nematodes, autophagy measurement generally favors counting GFP::LGG-1 punctae or other fluorescent assays, with few instances of monitoring substrate cleavage [37, 38]. As further discussed below, we sought to overcome the substantial limitations of fluorescent punctum-based autophagy measurements in C. elegans by developing a biochemical assay of autophagy. Our goals were to facilitate tissue-specific quantifications of autophagy, develop an assay that would also be usable in aged as well as young worms and provide a validated, two-part tag that could be used to measure autophagic degradation of either bulk cytoplasm or specific organelles.

We tagged C. elegans LGG-1 with two fluorescent proteins connected by a flexible, protease-sensitive linker (dFP: dual Fluorescent Protein tag; Figure 1A). LGG-1 is incorporated into the growing autophagosome and the final step of autophagy, fusion of the lysosome and autophagosome, exposes LGG-1 to the lysosomal proteases. These proteases cleave the linker between the fluorescent tags, releasing monomeric FP (mFP) that is tightly folded and resistant to proteases (Figure 1B). An immunoblot probed with anti-FP antibody distinguishes between the full-length dFP::LGG-1 and the released mFP by size. Increased flux through the autophagic pathway results in an increase in the ratio of mFP to intact dFP::LGG-1. Therefore, this assay allows us to measure the flux through the autophagic pathway instead of the formation of autophagosomes that are marked with GFP::LGG-1. Importantly, this is also internally controlled, relying on the ratio of cleaved to uncleaved protein; this allows it to be used in comparisons between conditions with varying levels of protein expression. To verify that the dFP tag releases free mFP as a consequence of conditions known to increase autophagy we first validated the dFP tag in Saccharomyces cerevisiae cells using the LGG-1 homolog Atg8. When yeast expressing dFP-Atg8 were subjected to nitrogen starvation there was a time-dependent increase in mFP, showing that conditions that increased TOR-regulated autophagy also increased mFP. (Figure 1C). A quantification of the dFP and mFP bands, as well as the ratio between them, is shown for this and all other figures in Supplementary Table 1.

Figure 1. Monomeric FP released from dFP::LGG-1 correlates with autophagic flux. (A) Yeast and nematode constructs were made by fusing a double fluorescent protein module, dFP, to ATG8 (yeast) or LGG-1 (nematode). The dFP contained two fluorescent proteins (labeled FP) separated by a protease-sensitive flexible linker. (B) The dFP Atg8/LGG-1 constructs are recruited to the growing autophagosome. Upon fusion of the autophagosome with the lysosome, the dFP is cleaved by lysosomal proteases, releasing soluble FP monomers. These monomers are fairly stable, but will eventually be further degraded by proteaolysis in the lysosome. (C) S. cerevisiae expressing dFP::ATG8 showed a time-dependent increase in mFP abundance with 2 or 4 hours of nitrogen starvation. (D) The number of dFP punctae in seam cells of L3 nematodes expressing dFP::LGG-1 from extrachromosomal arrays was reduced when the nematodes were treated with bec-1 RNAi to impair autophagy, as has been previously observed for GFP::LGG-1 [14]. Error bars represent standard error, asterisk for p<0.05. (E) C. elegans with extrachromosomal arrays expressing either eft-3p::dFP::LGG-1 or lgg-1p::GFP::LGG-1 show mFP release. Full-length dFP::LGG-1 is present at 73kDa, GFP-LGG-1 at 40kDa, and the monomers of GFP or mFP are present at 25kDa. The accumulation of mFP was inhibited by treatment with bec-1 RNAi (lanes 2 and 4). The accumulation of mFP in animals grown on control RNAi food is shown in lanes 1 and 3.

Next we sought to validate dFP::LGG-1 in nematodes. The most widely-used method for monitoring autophagy in C. elegans relies on counting GFP::LGG-1 punctae [8, 16, 24]. We validated our dFP::LGG-1 reporter by comparing it to this generally accepted system. Autophagy requires bec-1, the nematode orthologue of beclin/Vps30 [14, 20, 24]. We found that the number of fluorescent dFP::LGG-1 punctae in the seam cells of third-stage (L3) larvae decreased when bec-1 was depleted by RNAi (Figure 1D), just as has been shown previously using the GFP::LGG-1 reporter [14, 24]. This result indicates that the dFP tag is effectively incorporated into growing autophagosomes.

We next wanted to validate that mFP release from the dFP reporter was also dependent upon autophagy. To do this we depleted bec-1 by RNAi in animals ubiquitously expressing dFP::LGG-1. We found that bec-1 RNAi reduced the amount of mFP released at steady-state from dFP::LGG-1 and GFP::LGG-1, just as we had expected (Figure 1E). We also observed a band consistent with cleavage of GFP from the GFP::LGG-1 fusion protein that was similarly sensitive to bec-1(RNAi). Importantly, bec-1 RNAi did not significantly affect the amount of full-length protein and only decreased the amount of cleaved (mFP or GFP) protein. Together, these results from punctae and immunoblot assay validate our dFP autophagy reporter.

Autophagy varies substantially between tissues

To measure autophagy in different tissues, we constructed strains with dFP::LGG-1 expressed under the control of tissue-specific promoters. These drove expression specifically in the intestines, hypodermis, muscles, pharynx, or neurons (Figure 2A). Tissue-specific lines were created using both high-copy extrachromosomal arrays and single-copy insertion into the genome. We assayed mixed-age populations of each strain by immunoblot and found that the intestines had the highest relative abundance of the mFP cleavage product. This was the case whether worms carried a single copy integrant (Figure 2B) or an extra-chromosomal array (Figure 2C, uncropped immunoblot for this and all other blots shown in Supplementary Figure 1) of the dFP::LGG-1 transgene. Importantly, we could not detect any experimental differences between the behavior of dFP::LGG-1 at low and high levels of expression in this assay, and quantifying the ratio of mFP to dFP in each tissue yielded relatively similar values independent of the amount of overexpression (Supplementary Table 1). Subsequent experiments employed animals expressing dFP reporters from transgene arrays.

Figure 2. Tissue-specific expression of dFP::LGG-1 reveals variance in basal autophagy. (A) C. elegans strains were created with dFP::LGG-1 expressed under control of a promoter that is ubiquitously active (eft-3p), or under the control of promoters driving expression specifically in the intestine (vha-6p), hypodermis (dpy-7p), muscles (myo-3p), pharynx (myo-2p), or neurons (rab-3p). An image of each strain shows fluorescence consistent with the expected tissue-specific expression. Animals were imaged at approximately late L3 larval stage; scale bar is 100mm. Visualizing the fluorescent protein using an anti-FP immunoblot revealed expression in mixed-age populations of strains with the reporter present in either (B) single-copy chromosomal insertion or (C) tandem, multi-copy extrachromosomal array. In both expression conditions mFP is most abundant in the intestine with some mFP present in the hypodermis. A full blot is shown in (B) to provide context, while subsequent images focus on the upper and lower bands of interest; uncropped blots are shown in Supplementary Figure 1. (D) dFP::LGG-2 expressed from the same ubiquitous, intestinal, or neuronal promoter showed lysosomal processing primarily in the intestines.

LGG-1 is one of two C. elegans orthologues of mammalian LC3. Given the possibility of isoform-specific function [39–41], we asked whether the other LC3 orthologue LGG-2 might control different aspects of autophagic flux. Both proteins are broadly expressed, with largely overlapping expression patterns. There is slightly more LGG-2 present in neurons, raising the possibility of tissue-specific preferences of isoform incorporation into autophagosomes [42]. We therefore created a dFP::LGG-2 construct to measure autophagy that involves this protein. We found that dFP::LGG-2 had a tissue-specific signal of autophagic flux indistinguishable from that of dFP::LGG-1. Specifically, expression of full-length dFP::LGG-2 was detectable in all tissues, but dFP::LGG-2 released mFP primarily in the intestines and at very low levels in the neurons (Figure 2D). Thus, it appears that tissue-specific differences in mFP release reflect cell type-specific differences in the overall rate of autophagic flux, rather than reflecting functional variations between the two LGG proteins in C. elegans.

The amount of mFP present at steady-state is a function of both autophagosome maturation (which would release mFP) and lysosomal degradation (which would remove mFP). We used chloroquine treatment to address the possibility that the accumulation or absence of mFP reflects the efficiency of lysosomal proteolysis rather than autophagosome maturation [21]. Chloroquine accumulates in the lysosome and prevents acidification, impairing protease function. We reasoned that if mFP accumulated in animals treated with chloroquine, this would confirm that mFP was released from dFP in the mature autophagosome as well as indicate that an absence of mFP at steady-state was due to efficient degradation of mFP by lysosomal proteases. We exposed synchronized populations of L4 nematodes to chloroquine and measured the release of mFP in different tissues. Chloroquine treatment caused an increase in mFP in all tissues. This increase is visible in the gel (Figure 3A) and quantification of the increase is also provided (Supplemental Table 1). Chloroquine treatment increased the ratio of mFP:dFP 1.5-fold or more above the control baseline in every tissue assayed. There was variation in the magnitude of the effect observed using this assay, but this effect was robust and readily replicated across tissues (Figure 3B). These data demonstrate that autophagy occurs in all tissues, but that mFP may be cleared from the lysosome too rapidly to be detected at steady-state in some tissues, such as neurons. We conclude that, as expected, the steady-state mFP levels we observe are influenced both by forward movement through the autophagic pathway and the activity of hydrolytic enzymes in the lysosome.

Figure 3. All tissues have active autophagic flux. (A) Synchronized populations of L4 animals exposed to chloroquine for 18 hours showed accumulation of mFP in all tissues. (B) Average fold change over three experiments showing that all tissues increase mFP with chloroquine treatment, as shown by a fold change of greater than one, as marked by the horizontal line. Error bars represent standard error, tissue-specific chloroquine responses are not significantly different from each other (p>0.2).

Tissue-specific upregulation of autophagy in response to stress

Autophagy is a key physiologic response to starvation, and the number of GFP::LGG-1 punctae increases in nematodes that receive no or very little food [16, 24]. Microscopic assays focus on only a readily-observed subset of cells within a starved organism, however, and studies rarely attempt to compare the effects of autophagy across multiple tissues in response to nutrient limitation. To address whether tissues differ in their upregulation of autophagy following starvation, we analyzed autophagy in animals that had been removed from food for five days. We observed at least a two-fold up-regulation in the relative amount of mFP, as compared to full-length dFP::LGG-1, in every tissue after starvation (Figure 4A). The change in flux was most dramatic in the intestines, but all tissues responded to starvation with an increase in autophagic flux. These results show for the first time that the regulation of autophagy during starvation is similar across all tissues, and is not specific to the minor subset of cells and tissues used for visualization of GFP::LGG-1 punctae.

Figure 4. Starvation and anoxia induce different patterns of autophagy upregulation. (A) Animals lysed at L1, before starvation, showed low levels of basal autophagy (left panel, and band intensities quantified in Supplementary Table 1). After 5 days of starvation there was an increase in mFP in all tissues (right panel). Fold increase indicates change in normalized mFP after starvation. Representative gel is shown but experiments consistently showed the largest increases in intestine, muscles and neurons, with a positive but smaller fold increase in other tissues. (B) Exposing nematodes to 18 hours of anoxia at mid-larval stage (L2/L3) causes an increase in autophagic flux in the intestine but does not affect mFP levels in other tissues. All lanes are from the same experiment run in the same SDS-PAGE gel, with delineations indicating areas of the image that were separately adjusted for brightness and contrast for the sake of clarity (uncropped immunoblots shown in Supplementary Figure 1). (C) A graph of the average fold change of normalized mFP in each tissue after anoxia. N=3-7, asterisk indicates p<0.05 for the mFP ratio of the ubiquitous reporter compared to the mFP ratio in the hypodermis, pharnynx or muscle, and p<0.06 for a comparison of the mFP ratios in ubiquitous and neuronal expression. The ubiquitous and intestinal reporters were statistically similar (p=0.8).

Oxygen deprivation (anoxia; operationally defined as < 10 ppm O₂) is similar to starvation in that cellular energy generation is severely limited. Moreover, as has been demonstrated for starvation, autophagy is required for nematodes to survive an anoxic shock [8]. We used our dFP::LGG-1 reporter to reveal the effect of anoxia on autophagic flux in different tissues. After 18 hours of anoxia, we observed an increase in mFP release that was primarily due to changes in the intestinal mFP accumulation (Figure 4B). In contrast, the ratio of mFP to full-length dFP::LGG-1 remained constant in other tissues of animals exposed to anoxia (Figure 4C). We conclude that, in marked contrast to the more general response to starvation, anoxia induces tissue-selective upregulation of autophagy in the intestine.

Measuring autophagy of soluble cytoplasmic proteins

A potential confounding factor in experiments based on measuring GFP::LGG-1 is that overexpressing LC3 is reported to cause abnormal autophagosome formation and pathway dynamics [25]. We observed similar measurements of autophagic flux whether we expressed dFP::LGG-1 from multi-copy extrachromosomal arrays as when the reporter was expressed at a more natural level from single-copy integration into the genome (Figure 2), arguing that overexpression of dFP::LGG-1 was neither inhibitory nor stimulatory in our system. However, to circumvent this potentially confounding effect of LGG protein overexpression, we expressed dFP alone as a soluble cytoplasmic protein, without fusing it to any proteins directly involved in autophagy. We reasoned that mFP would be released from any of the cytoplasmic dFP taken into autophagosomes during bulk autophagic turnover, but that expressing mFP alone should not affect the formation of auto-phagosomes.

The dFP construct driven by an eft-3 promoter was expressed in all tissues, as expected (Figure 5A). We observed release of mFP in animals that expressed the soluble dFP reporter, consistent with our hypothesis that it would be taken up into autophagosomes non-specifically (Figure 5B). The release of mFP was enhanced by exposure to anoxia (Figure 5B), similar to our observation of increased mFP release from dFP::LGG-1 (Figure 4B). This result underscores that autophagic flux is induced in response to anoxia. The result further indicates that the dFP::LGG-1 fusion protein does not interfere with the assembly or maturation of autophagosomes.

Figure 5. Cytoplasmic flux through the autophagic pathway. (A) Ubiquitously-expressed soluble dFP was present in all tissues, as visualized with fluorescence. Scale bar 100mm. (B) This reporter showed a substantial increase in mFP release when the nematodes were exposed to anoxia. This demonstrates an increase of bulk autophagic flux without overexpression of any protein involved in autophagy.

dFP reporters to visualize mitophagy

Our observation that mFP is released from solubly-expressed dFP suggests that this tag could be used to measure autophagy of specific proteins or organelles, even if the tagged protein or signal sequence is very small. This is in contrast to monomeric GFP, which cannot be used alone or with very small proteins because the size of the protein is identical whether it is in the cytoplasm or the lysosomal environment. As a proof of concept, we took advantage of this feature of the dFP reporter to monitor autophagy of mitochondria (mitophagy).

We fused dFP to the outer mitochondrial protein TOMM-7 to measure turnover of mitochondria by mitophagy. TOMM-7 is the C. elegans orthologue of Tom7, a key member of the mitochondrial import protein complex [43]. The dFP::TOMM-7 reporter was expressed in all tissues under control of the eft-3 promoter. Visualization of the fluorescent protein in the body-wall muscle was consistent with its localization to mitochondria (Figure 6A). We confirmed this observation using subcellular fractionation. As expected, dFP::TOMM-7 was more efficiently retained in the mitochondrially-enriched membrane fraction than was the soluble dFP control (Figure 6B). Probing with an antibody against the mitochondrial voltage-dependent anion channel protein (VDAC) confirms that our fractionation effectively concentrated mitochondria. Although most of the soluble dFP was in the cytoplasmic fraction, some soluble dFP (and released mFP) were observed in the high-speed spin pellet. This likely reflects that fact that some lysosomes were included in pellet fraction after this high-speed spin. Consistent with this interpretation, we also observed some free mFP in the pellet from lysate of animals expressing dFP::TOMM-7, though less than is present from animals expressing soluble dFP. We conclude that dFP::TOMM-7 is correctly targeted to mitochondria.

Figure 6. Mitochondrial and cytoplasmic flux through the autophagic pathway. (A) The mitochondrial outer membrane protein TOMM-7 tagged with dFP was incorporated into mitochondrial patterns in the muscles. Scale bar 50mm. (B) Cells from animals expressing either soluble dFP or dFP::TOMM-7 were subfractionated into cytoplasmic and mitochondria-enriched fractions. The presence of VDAC in the pellet confirms mitochondrial enrichment (lower panel). Full-length dFP::TOMM-7 is enriched in the mitochondrial fraction as compared to soluble dFP. (C) Animals expressing dFP::TOMM-7 exposed to anoxia showed an increase in mFP release, indicating that autophagy brings mitochondria to the lysosome in response to anoxic shock.

We measured mitophagy in animals exposed to anoxia using the dFP::TOMM-7 reporter. The amount of mFP released was increased in animals exposed to anoxia, similar to what we measured using the dFP::LGG-1 reporter (Figure 6C). This result suggests that mitochondria are degraded by autophagy in anoxia. These studies show the potential power of using the dFP tag to assay the flux of specific proteins and organelles through the autophagic pathway in different tissues.

Autophagy and lysosomal function in ageing

Genetic data have consistently shown that autophagy is required for increased lifespan in a variety of genetic models [44]. However, it is technically challenging to quantify the effect of normal aging on autophagy in C. elegans. Autofluorescence increases with age, particularly in the intestine, making it difficult or impossible to count GFP::LGG-1 punctae. However, the level of autophagic activity changes as animals age, and autophagy genes are required for increased lifespan in many genetic long-lived models [44, 45].

We measured the release of mFP from dFP::LGG-1 expressed in different tissues in aging C. elegans to test the hypothesis that autophagic flux changes with age in a tissue-specific pattern. We observed an increase in soluble mFP as animals aged, though the effect in pharynx and neurons was not as dramatic as in other tissues (Figure 7A). The overall expression decreased from some promoters in older animals, such as in the hypodermis, but the ratiometric nature of our assay still allowed a measurement of autophagic flux. This result demonstrated an age-dependent increase in autophagic flux that varied between tissues. To determine more precisely when during aging the increase in autophagic flux occurred, we measured steady-state levels of mFP in animals for 12 days after L1 using both ubiquitously-expressed and intestinally-expressed dFP::LGG-1. The ratio of mFP to full-length dFP::LGG-1 increased dramatically in early adulthood, at three to six days after L1 (Figure 7B). Importantly, chloroquine treatment increased the amount of mFP in all tissues (Figure 7C), demonstrating that lysosomes of older animals remain acidified and proteolytically proficient. Moreover, the increase in autophagy measured here is not influenced by the overexpression of LGG-1, as we observed an age-related increase in mFP release from the soluble dFP bulk autophagy reporter (Figure 7D). Thus, our results strongly support the hypothesis that autophagy increases over the course of normal aging.

Figure 7. Autophagic flux increases with age. (A) L1 animals have low levels of mFP in all tissues (left panel), while animals aged to five days after L1 show an increase in mFP in most tissues (right panel, quantifications shown in Supplementary Table 1). Representative blots (of n=4 experiments) are shown. (B) Animals expressing either ubiquitous or intestinal dFP::LGG-1 show a shift in autophagy that occurs in early adulthood, with a dramatic increase in the amount of mFP accumulation between three and six days of life. (C) Animals aged five days and treated with chloroquine for 18 hours accumulate more mFP in all tissues, showing that mFP accumulation is not due to a dramatic change in lysosome function. All lanes are from the same experiment run in the same SDS-PAGE gel, with delineations indicating areas of the image that were separately adjusted for brightness and contrast for the sake of clarity (uncropped immunoblots are shown in Supplementary Figure 1). (D) Release of mFP from soluble dFP reporter also shows a shift in mFP accumulation with age, with increasing amounts of mFP at five days of age.

Cloning

The dual fluorescent protein (dFP) tag was constructed using tags that were variants of Cerulean and Venus fluorescent proteins (gifts of Erik Snapp, [78]) connected by a linker with GPG, Tev and Cathepsin-sensitive sequences (GGACCAGGCGAGA ATTTGTATTTTCAGGGTCGTCTTGTCAAGTTCCTT GTTGGAGGACCAGGC).

For yeast expression, 990bp of the S. cerevisiae ATG8 upstream sequence was cloned into pRS415 followed by the dFP tag cassette and then the ATG8 gene sequence followed by 965bp of downstream sequence. For expression in nematodes, dFP was fused to LGG-1 using plasmids constructed using PCR and Gateway cloning technology (Invitrogen). Plasmids with promoter sequences were kindly provided by Dr. Michael Ailion (University of Washington, Department of Biochemistry). The tissue-specific promoters were eft-3p (ubiquitous, also known as eef-1A.1), [71]vha-6p (intestine) [72], dpy-7p (hypodermis) [73], myo-2p (pharynx) [74], myo-3p (muscle) [75] and rab-3p (neurons) [76]. The dFP tag was cloned into the pDONR221 vector and the LGG-1 genomic sequence and 300bp of downstream sequence was cloned into pDONR-P2R-P3 vector. Gateway recombination yielded each promoter upstream of the dFP::LGG-1 coding sequence. Plasmid constructs were verified by sequencing.

For the cytoplasmic dFP expression, dFP was cloned behind the eft-3 promoter and followed by the let-858 3′UTR sequence. The C. elegans lgg-2 and tomm-7 genes were cloned into the pDONR-P2R-P3 vector with 572bp or 111 bp of 3′ UTR sequence, respectively. These vectors were recombined with the eft-3 promoter and dFP plasmids described above to generate each dFP fusion protein expression construct.

Nematode strains

Arrays were made by injecting Unc EG6699 (ttTi5605 II; unc-119(ed3) III; oxEx1578) worms with the expression constructs and either yeast genomic DNA or an empty yeast expression plasmid to total 100 ng/μL. Offspring with arrays showed normal movement and green fluorescence in the target tissues. Worms with a single, chromosomal insertion of dFP-tagged LGG-1 were made following an established Mos1 mediated Single Copy transgene Insertion (MosSCI) protocol [71] Clonal lines were established for all lines except the LGG-2 arrays, which were mixed populations from a single successful injection. Identical experimental results were obtained both with and without outcrossing for all lines. Worms expressing GFP::LGG-1 (adIs2122) were from the Caenorhabditis Genetics Center.

Imaging

Images were taken using an Olympus Fluoview FV1200 confocal microscope using a 40x objective and a pinhole setting of 115μm. Images are of a single image slice of worms at approximately larval stage 3 visualized with 515nm excitation and 527nm emission. To count dFP::LGG-1 punctae, L3 animals were placed in a Nikon Eclipse 90i microscope, and the punctae visible in hypodermal seam cells through a GFP filter were counted.

Immunoblots and antibodies

Animals were harvested by centrifugation and lysed in SDS sample buffer by repeated cycles of freezing and boiling, cleared by centrifugation at 8,000 × g for 5 minutes, and the supernatant was run on a 12% electrophoresis polyacrylamide gel containing sodium dodecyl sulfate (SDS-PAGE) and transferred to a nitrocellulose membrane (Bio-Rad Laboratories). The membrane was blocked for one hour in Odyssey blocking buffer (Li-Cor Biosciences). The membrane was incubated with anti-GFP monoclonal antibody (Roche Applied Science), for detecting either FP or GFP, polyclonal anti-actin (gift from William T. Wickner) or anti-VDAC (Pierce) overnight. After washing in PBST buffer (137 mM NaCl, 2.7 mM KCl, 20 mM dibasic sodium phosphate, 2mM monobasic potassium phosphate, 0.1% Tween-20), the membranes were incubated with the species-appropriate, Alexa-conjugated secondary antibody (IR Dye 800 goat anti-mouse from Rockland Immunochemicals or Alexa Fluor 680 goat anti-rabbit from Invitrogen), washed in PBST, and visualized using the Odyssey Imaging System (Li-Cor Biosciences). ImageJ software (National Institutes of Health; http://imagej.nih.gov/ij/index.html) was used for the densitometry calculations, and for statistical comparison we used a two-tailed t-test.

Animal culture and treatments. Worms were grown at 20° on NGM plates with E. coli OP50, or on enriched plates seeded with E. coli NA22 cells, as indicated. To developmentally synchronize populations, embryos were isolated using a standard bleaching protocol [77]. To synchronize aging populations, isolated embryos were grown to fourth-stage larvae (L4) and then moved to plates containing 50μM 5-Fluoro-2′-deoxyuridine (FuDR) and grown for the indicated number of days.

In starvation experiments, isolated embryos were hatched in M9 without food overnight. The arrested first-stage larvae (L1) were incubated in 3mL of M9, rotating at 20° for a total of 5 days, with one change of M9. A control sample of L1 animals was harvested, lysed, and frozen at the beginning of the starvation period.

For exposure to anoxia (operationally defined as < 10 ppm O2), isolated embryos were placed directly on OP50 bacteria and grown in room air at 20° for approximately 30 hours to stage L3. Animals were then collected in M9 media, washed twice, and placed on unseeded plates at 20° in room air or in chambers continuously flooded with humidified 100% nitrogen (Airgas, Seattle, WA) for 18 hours [55]. Compressed gas tanks were certified standard to contain < 10 ppm O₂.

For chloroquine treatment, isolated embryos were grown on OP50 bacteria for approximately 55 hours at 20° and then washed and placed into M9 containing 4x concentrated OP50 and 20mM chloroquine diphosphate salt (Sigma-Aldrich) and incubated, rotating, for 18 hours. For chloroquine treating older worms, animals were grown on OP50 from embryo to L4 when they were moved to NA22 plates containing FuDR. Five days after initial isolation the animals were washed off the plates and incubated for 18 hours in M9 medium containing 4x concentrated OP50 bacteria and 50μM FuDR, with or without chloroquine for 18 hours.

Yeast growth and starvation

BY4742 yeast were transformed with dFP::ATG8 using a lithium acetate transformation protocol [79] Yeast were cultured overnight in synthetic medium lacking leucine to ensure plasmid retention and upon reaching an optical density (OD600nm wavelength) of 0.8 were resuspended in synthetic medium without leucine, or a nitrogen starvation medium (0.175 wt/vol nitrogen base without ammonium sulfate or amino acids and 2% glucose) for the indicated time. Yeast cells were then harvested by pelleting, resuspending in lysis buffer (50mM Hepes pH 7.4, 200mM NaCl, 10% glycerol, 0.5% Triton-X) with SDS PAGE loading buffer (final concentrations of 62mM Tris pH 6.7, 69 mM sodium dodecyl sulfate, 10% glycerol with bromophenol blue) in the presence of protease inhibitors, heated to 100° for 10 minutes, vortexed with glass beads to break apart the cell wall, cleared from the beads, and centrifuged at 16,000 × g for two minutes. The supernatant lysate was separated by SDS-PAGE and analyzed by immunoblot.

RNAi

E. coli expressing bec-1 RNAi or the empty control plasmid L4440 were struck onto plates containing 50μg/ml carbenicillin and 15μg/ml tetracycline. A single colony from that plate was cultured overnight in ampicillin-containing medium and plated onto NGM-lite plates that contained 3 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and 50μg/ml carbenicilin. These RNAi plates were allowed to dry overnight and stored at 4°. Animals were placed onto RNAi bacteria as embryos and grown for approximately 48 hours at 20° before harvesting for analysis by immunoblot. For hypodermal punctae counts, animals were placed on RNAi at L4, allowed to lay eggs, and the progeny were grown to L3 before imaging.