ATG-18 and EPG-6 are Both Required for Autophagy but Differentially Contribute to Lifespan Control in Caenorhabditis elegans

During macroautophagy, the human WIPI (WD-repeat protein interacting with phosphoinositides) proteins (WIPI1–4) function as phosphatidylinositol 3-phosphate effectors at the nascent autophagosome. Likewise, the two WIPI homologues in Caenorhabditis elegans, ATG-18 and EPG-6, play important roles in autophagy, whereby ATG-18 is considered to act upstream of EPG-6 at the onset of autophagy. Due to its essential role in autophagy, ATG-18 was found to be also essential for lifespan extension in Caenorhabditis elegans; however, this has not yet been addressed with regard to EPG-6. Here, we wished to address this point and generated mutant strains that expressed the autophagy marker GFP::LGG-1 (GFP-LC3 in mammals) and harbored functional deletions of either atg-18 (atg18(gk378)), epg-6 (epg-6(bp242)) or both (atg-18(gk378);epg-6(bp242)). Using quantitative fluorescence microscopy, Western blotting, and lifespan assessments, we provide evidence that in the absence of either ATG-18 or EPG-6 autophagy was impaired, and while atg-18 mutant animals showed a short-lived phenotype, lifespan was significantly increased in epg-6 mutant animals. We speculate that the long-lived phenotype of epg-6 mutant animals points towards an autophagy-independent function of EPG-6 in lifespan control that warrants further mechanistic investigations in future studies.


Bacteria Culture as Food Source
Bacterial cultures using E. coli OP50 (from CGC) were prepared overnight at 37 • C, 180 rpm in liquid LB medium (10 g/L tryptone, 5 g/L yeast extract, 5 g/L NaCl). Bacterial cultures were always prepared freshly from glycerol stocks and then seeded onto NGM plates.

Generation of Male Stocks by Heat Shock
Wild-type C. elegans male stocks were purchased from CGC (Caenorhabditis Genetics Center, University of Minnesota, Minnesota, MN, USA). For all other strains, males were generated by heat shock by incubating 20-30 hermaphrodites per 6-cm plate at 30 • C for 6 h, which were then transferred back to 20 • C, and incubated for 3 days. After 3 days, males were used for crossing with L4 hermaphrodites of the same strain.

Crossing Strains
To generate GFP::LGG-1 expressing strains carrying the desired mutations in atg-18 and/or epg-6, an adIs2122 extrachromosomal transgene expressing DA2123 strain was used for crossing (Table S1). The adIs2122 transgene encoded an N-terminal GFP::LGG-1 fusion protein. A 1.7 kb fragment of the lgg-1 promoter was cloned upstream of the GFP coding sequence, and a 2.2 kb fragment of the lgg-1 coding sequence was cloned downstream of the GFP sequence [11]. To remove background mutations, each strain was backcrossed with wild-type worms at least 3 times. For backcrossing, hermaphrodites carrying the allele of interest were crossed with wild-type (N2) males. Eight L4 hermaphrodites and at least 15-20 males were transferred to a small agar piece (about 1/3 of a 6-cm NGM agar plate) covered with E. coli OP50. The rest of the agar was removed using a sterile spatula before transferring the worms. The plate was incubated at 20 • C overnight, and the next morning, the hermaphrodites were transferred singly onto NGM plates covered with E. coli OP50. The plates were incubated at 20 • C, and after 3 days the progeny was inspected for males. As the presence of males represents successful mating, plates with the highest ratio of males were chosen for further analysis. From the F1 progeny, heterozygous L4 hermaphrodites were singled again, and incubated at 20 • C. After 2 days, the F1 hermaphrodites were lysed singly, and genotyped by single worm lysis (SWL)-PCR. From a plate, where the F1 hermaphrodite was heterozygous, F2 L4 hermaphrodites-from which 25% should be homozygous mutant-were transferred singly to fresh plates and incubated at 20 • C. After 2 days, the F2 hermaphrodites were genotyped by single worm PCR. The plates where the F2 hermaphrodite was homozygous mutant was kept, and F3 worms were genotyped with single-worm PCR again. When all F3 worms were homozygous mutant, the population was considered as backcrossed once. For the generation of double-mutant strains, crossing was done as described above, except that the hermaphrodites were not crossed with wild-type males, but with males carrying the allele of interest. polymerase (NEB, M0273) with standard Taq buffer (NEB, M0273L), 200 µM dNTP (NEB, N0446S), 0.5 µM of both primers (Sigma-Aldrich, Munich, Germany), and 5 µL DNA was used. For genotyping of epg-6(bp242) tetra-primer amplification-refractory mutation system PCR (tetra-primer ARMS-PCR) was applied [23]. For ARMS-PCR, the inner primers were used at 1 µM, the outer primers at 0.1 µM final concentration.

Brood-Size Assay
Stage L4 hermaphrodites were transferred singly to NGM plates seeded with E. coli OP50, and incubated at 20 • C. Hermaphrodites were transferred to a new plate once a day, until they laid eggs. Eggs were counted right after the hermaphrodite was transferred to a new plate, and L4/adult progeny was counted 2-3 days later. The brood size was determined as the number of progeny that reached L4/adult stage.

Egg Isolation Using Hypochlorite Treatment
Gravid adult worms were grown in high density and used for isolation of eggs as follows. Animals were washed off NGM plates with H 2 O, centrifuged (2 min, 2000 rpm, RT), supplemented with bleach solution (hypochlorite treatment), and shook until the worms lysed. Subsequently, tubes with lysed worms were filled with M9 buffer (3 g/L KH 2 PO 4 , 6 g/L Na 2 HPO 4 , 5 g/L NaCl, 1mM MgSO 4 ) and centrifuged (2000 rpm, 2 min). Finally, eggs were washed three times with M9 buffer.

L1 Survival Assay
Embryos were prepared using hypochlorite treatment (2-4% NaOCl, 1.5 M NaOH) of gravid adults, and collected in M9 buffer (3 g/L KH 2 PO 4 , 6 g/L Na 2 HPO 4 , 5 g/L NaCl, 1 mM MgSO 4 ) (1000 embryos/mL). The liquid culture was incubated without food at 20 • C, shaking (180 rpm). The first week every day, after that, 3 times a week 50 µL aliquots were placed on NGM plates seeded with E. coli OP50. The number of L1 larvae was counted right away. After 2-4 days at 20 • C, the numbers of larvae that could develop further and were able to pass the L1 larval stage were counted. The percentage of worms that survived starvation and could recover in the presence of food was calculated.

Lifespan Assay
Lifespan assays were performed at 20 • C. Synchronous animal populations were generated by hypochlorite treatment (2-4% NaOCl, 1.5 M NaOH) of gravid adults. Worms were allowed to grow on normal NGM plates seeded with E. coli OP50 until the L4 larval stage. Larval stage L4 animals were transferred to NGM plates supplemented with 5 mg/L FUDR (5-Fluoro-2 -deoxyuridin, Sigma-Aldrich, F0503). Twenty animals were transferred to one plate, and 3-5 plates were counted for each strain in one set. Animals were counted every 2-3 days and scored as dead when they stopped responding to gentle prodding with a platinum wire. Dead animals were removed from the plates. Animals dried out on the walls were censored. When lifespan assays were conducted in the absence of FUDR, animals were counted and transferred to fresh NGM plates every day. Statistical analyses of the lifespan assays were done with the "Online Application for the Survival Analysis of Lifespan Assays Performed in Aging" software using basic survival analysis [24].

Lifespan Assay with RNAi Treatment
For RNAi treatments E. coli HT115 expressing an empty pL4440 plasmid or pL4440 containing the EPG-6 targeting RNAi construct were seeded on LB agar plates containing 12.5 µg/mL tetracycline (Sigma-Aldrich, 87128) and 50 µg/mL ampicillin (Applichem, Darmstadt, Germany; No. A0839) and incubated overnight at 37 • C. For liquid culture, the bacteria were grown in LB medium supplemented with 50 µg/mL ampicillin for 10-16 h at 250 rpm and 37 • C. For the lifespan assays, the liquid culture was seeded onto NGM plates supplemented with 1 mM IPTG (Isopropyl β-D-1-thiogalactopyranoside, Sigma-Aldrich, I6758), 25 µg/mL carbenicillin (Applichem, Darmstadt, Germany; No. A1491), and 5 mg/L FUDR. The plates were incubated at room temperature overnight to allow induction of the RNAi expression by IPTG. Animals were synchronized as described above (see Section 2.9) before transfer to the plates covered with bacteria expressing EPG-6 targeting or control RNAi constructs. The animals were incubated at 20 • C throughout the experiment and transferred to fresh RNAi plates every 3-5 days. Viability was scored every second to third day, as described for the standard lifespan assay (see Section 2.9). The EPG-6 targeting construct was purchased from Source Bioscience (clone ID: DFCI11006A10) as part of the RNAi library of Marc Vidal's lab (Dana-Farber Cancer Institute, Boston, MA, USA) [25]. The empty pL4440 vector expressing the HT115 strain was a gift from Della David (German Center for Neurodegenerative Diseases, Tübingen, Germany).

Light Microscopy and Length Determination of C. elegans Strains
Animals were cultured on NGM plates supplemented with 5 mg/L FUDR (5-Fluoro-2 -deoxyuridin, Sigma-Aldrich, F0503) and imaged at 14 days of adulthood. Images were taken using a Nikon AZ100 microscope at 1.2× magnification. The length of the animals was determined by approximation of the nematode from the tip of the mouth to the end of the tail using the "polyline" function of NIS elements v4.

GFP::LGG-1 Puncta Assessments
Eggs were isolated from adult animals using hypochlorite treatment (2-4% NaOCl, 1.5 M NaOH) of gravid adults. For the microscopy of embryos, eggs were incubated in M9 buffer (3 g/L KH 2 PO 4 , 6 g/L Na 2 HPO 4 , 5 g/L NaCl, 1 mM MgSO 4 ) at 20 • C, and prepared for microscopy right after egg isolation (0 h), or 4 h later (4 h). For the microscopy of L1, L2, L3, L4, and adult worms, eggs were seeded on standard NGM plates with E. coli OP50, and incubated for 6, 28, 52, and 76 h at 20 • C. For mounting on microscopy slides an agar pad was used, and worms were anaesthetized using 0.1 M NaN 3 . Fluorescent (GFP) and DIC images were taken using an Axiovert 200M (Carl Zeiss Microscopy GmbH, Munich, Germany) and an EC Plan-Neofluar 100×/1.30 Oil M27 objective (embryos) or a Plan-Apochromat 63×/1.40 Oil Ph 3 objective (larvae, adult). Images were quantified using the stress granule counter plugin of ImageJ. To quantify puncta formation in embryos, we divided the embryos into three groups based on their developmental stage: early stage embryos (until the beam stage), and late stage embryos (after 1.5-fold until 3-fold). In L1-L4 larvae and adults, puncta formation was measured in the head and in the seam cells. In the case of the seam cells inclusion, count refers to inclusion count/cell. In the case of the head, the whole head was analyzed until the posterior bulb of the pharynx.

Statistical Analysis
Statistical analyses were carried out using SPSS Statistics Version 25.0 (IBM, Ehningen, Germany). The data distribution was tested using the Shapiro-Wilk test. If the data were normally distributed, group differences were analyzed using one-sided ANOVA and subsequent Tukey's post-hoc test, as indicated in the figure legend. When data were not normally distributed, the groups were analyzed using a Kruskal-Wallis test with sequential Dunn-Bonferroni to test for significances between the strains. Statistical analysis for Western blot results when only two strains were compared was carried out using a two-tailed Student's t-test.
We observed that although the number of eggs laid by wild-type and mutant hermaphrodites differed significantly only in the atg-18(gk378);epg-6(bp242) double mutant ( Figure 1B), and most of the atg-18(gk378), epg-6(bp242), and atg-18(gk378);epg-6(bp242) eggs could hatch, the brood size (progeny that reached L4 larval or adult stage) of atg-18(gk378) and atg-18(gk378);epg-6(bp242) mutants significantly decreased ( Figure 1C). In line with previous reports [10,11], our results show that loss of ATG-18 causes lethality at an early developmental stage, which is even higher in the absence of both ATG-18 and EPG-6, suggesting that ATG-18 and EPG-6 are essential survival factors under fed conditions. Of note, we did not observe a difference in brood size when comparing wild-type and epg-6(bp242) animals ( Figure 1C). , atg-18(gk378), epg-6(bp242), and atg-18(gk378);epg-6(bp242) strains were taken using a 1.2× objective and representative images are shown, n ≥ 9. (B) Single hermaphrodites were transferred to nematode growth medium (NGM) plates and the total number of eggs during the reproductive period per hermaphrodite was counted (n > 4). (C) The number of eggs hatching and reaching the larval stage 4 (L4)/adult stage was counted, and the percentages, based on the total egg number, was calculated. Mean values (+STDEV) of the number of eggs (B) and the brood size (C) are shown, and significant differences compared to wild-type indicated ** p < 0.01, *** p < 0.001, one-way ANOVA.
In L2/L3 larvae, GFP:: LGG-1 puncta formation was very similar to L1 larvae: enlarged autophagic structures accumulated in the head, whereas a reduced number of autophagic structures with aberrant size was observed in the seam cells of atg-18(gk378);adIS2122 and epg-6(bp242);adIS2122 L2/L3 larvae ( Figure 4C-E). Accumulation of enlarged puncta in the head could also be observed in atg-18(gk378);adIS2122 and epg-6(bp242);adIS2122 L4 larvae ( Figure 5C-E). Although in the seam cells (A) Light microscopy images of Caenorhabditis elegans wild-type (N2), atg-18(gk378), epg-6(bp242), and atg-18(gk378);epg-6(bp242) strains were taken using a 1.2× objective and representative images are shown, n ≥ 9. (B) Single hermaphrodites were transferred to nematode growth medium (NGM) plates and the total number of eggs during the reproductive period per hermaphrodite was counted (n > 4). (C) The number of eggs hatching and reaching the larval stage 4 (L4)/adult stage was counted, and the percentages, based on the total egg number, was calculated. Mean values (+STDEV) of the number of eggs (B) and the brood size (C) are shown, and significant differences compared to wild-type indicated ** p < 0.01, *** p < 0.001, one-way ANOVA.
In summary, aberrantly enlarged GFP:: LGG-1 decorated autophagic structures were observed during larval development (L1-4) in the head of atg-18 and epg-6 mutants (Figures 3-5), suggesting a block in the autophagic flux in the absence of ATG-18 or EPG-6. Of note, in seam cells, GFP::LGG-1 puncta were almost non-visible in nutrient-rich conditions; however, they become visible, e.g., in ATG mutant strains due to a block in autophagy [21,28], or during challenging conditions, such as starvation, due to the induction of autophagy [21]. In our case, and according to previous findings [10,11], we suggest that the decrease in GFP::LGG-1 puncta numbers in seam cells (Figures 3-5) likewise suggest that autophagy is impaired in the atg-18 and epg-6 mutant strains. In summary, aberrantly enlarged GFP:: LGG-1 decorated autophagic structures were observed during larval development (L1-4) in the head of atg-18 and epg-6 mutants (Figures 3-5), suggesting a block in the autophagic flux in the absence of ATG-18 or EPG-6. Of note, in seam cells, GFP:: LGG-1 puncta were almost non-visible in nutrient-rich conditions; however, they become visible, e.g., in ATG mutant strains due to a block in autophagy [21,28], or during challenging conditions, such as starvation, due to the induction of autophagy [21]. In our case, and according to previous findings [10,11], we suggest that the decrease in GFP:: LGG-1 puncta numbers in seam cells (Figures 3-5) likewise suggest that autophagy is impaired in the atg-18 and epg-6 mutant strains.  The autophagic flux can be assessed by measuring lipidation of GFP:: LGG-1 by Western blotting [21,27] where non-lipidated, cytosolic GFP-LGG-1 is distinguished from GFP-LGG-1 conjugated to phosphatidylethanolamine (GFP-LGG-1-PE) at autophagosomal membranes based on different migration patterns ( Figure 6A). Using GFP::LGG-1 also offers the possibility to compare the levels of cleaved GFP. GFP is cleaved from LGG-1 after the autophagosome fuses with the lysosome, and proteins enclosed in the autolysosome are degraded. In the presence of ammonium chloride (NH 4 Cl), which increases the lysosomal pH and thereby slows down autophagic degradation, cleaved GFP accumulates ( Figure 6A) [21,27]. Hence, relative abundances of cleaved GFP provide a measure for functional autophagy flux assessments. To see whether the autophagic flux was impaired in atg-18 and epg-6 mutant strains, we starved atg-18(gk378);adIS2122 and epg-6(bp242);adIS2122 L1 larvae overnight in the presence or absence of NH 4 Cl, and extracted proteins for Western blotting ( Figure 6B). Both atg-18(gk378);adIS2122 and epg-6(bp242);adIS2122 mutant strains showed a strong accumulation of GFP::LGG-1 and GFP::LGG-1-PE when compared to wild-type ( Figure 6B), indicating either a block in autophagy or strong upregulation of GFP::LGG-1 after loss of ATG-18 and EPG-6. Addition of NH 4 Cl caused an accumulation of cleaved GFP in the wild-type due to a block in starvation-induced autophagy as expected, but this was not found in either the atg-18(gk378);adIS2122 or the epg-6(bp242);adIS2122 mutant sample ( Figure 6B,C; Data S1). The lack of cleaved GFP accumulation following NH 4 Cl treatment in the context of accumulated GFP::LGG-1 and GFP-LGG-1-PE further supports the idea that loss of either ATG-18 or EPG-6 leads to a block in autophagy. ). Both atg-18(gk378);adIS2122 and epg-6(bp242);adIS2122 mutant strains showed a strong accumulation of GFP:: LGG-1 and GFP:: LGG-1-PE when compared to wild-type ( Figure 6B), indicating either a block in autophagy or strong upregulation of GFP:: LGG-1 after loss of ATG-18 and EPG-6. Addition of NH4Cl caused an accumulation of cleaved GFP in the wild-type due to a block in starvation-induced autophagy as expected, but this was not found in either the atg-18(gk378);adIS2122 or the epg-6(bp242);adIS2122 mutant sample ( Figure 6B,C; Data S1). The lack of cleaved GFP accumulation following NH4Cl treatment in the context of accumulated GFP:: LGG-1 and GFP-LGG-1-PE further supports the idea that loss of either ATG-18 or EPG-6 leads to a block in autophagy.  LGG-1 puncta accumulated in L4 larvae in the head but decreased in numbers in seam cells. GFP::LGG-1 puncta formation was analysed in the head and seam cells of wild-type, atg-18(gk378), and epg-6(bp242) L4 larvae expressing the adIS2122 transgene. The scheme in (A) indicates the developmental larval stage (L4) where GFP::LGG-1 puncta assessments were conducted, and the images in (B) show head and mid-body regions, the latter with seam cells, as examples. Fluorescence microscopy images were quantified using the stress granule counter plugin of ImageJ. The average number (C) and size (D) of GFP::LGG-1 puncta were calculated. Representative images are shown (E). Scale bar: 20 µm. Significances calculated using the Kruskal-Wallis test, ns: not significant, * p < 0.05, ** p < 0.01, *** p < 0.001. For more information, see Data S1.
When C. elegans eggs hatch in the absence of food, L1 larvae enter an arrested stage, which wildtype larvae survive for up to three weeks and upon re-feeding they recover and develop further [21]. As proper autophagic activity is required for the survival of this starvation period [11,21,29], we aimed to further investigate the physiological roles of ATG-18 and EPG-6 in C. elegans by measuring the survival and recovery of L1 larvae upon starvation. (Figure 7). Following synchronization, L1 larvae (atg-18(gk378), epg-6(bp242), and atg-18(gk378);epg-6(bp242)) were continuously starved over a period between 1 to 29 days and their survival was assessed by counting larvae that recovered from starvation and progressed through larval stage L2 in the presence of food ( Figure 7A,B). LGG-1-PE and cleaved GFP is detected by Western blotting using GFP::LGG-1 expressing strains along with wild-type N2 control (A, right). Cleaved GFP, which can also be visualized by Western blotting, accumulates upon lysosomal inhibition due to a decelerated degradation using NH4Cl (A, right). L1 animals of wildtype (Control), atg-18(gk378), and epg-6(bp242) expressing the adIS2122 transgene were starved overnight in the presence (+) or absence (-) of NH4Cl, and proteins were extracted for anti-GFP Western blotting (B). A representative short exposure is shown and GFP::LGG-1 and GFP::LGG-1-PE migrations are indicated (B, upper panel). Likewise, a longer exposure is shown and the migration of cleaved GFP indicated (B, lower panel). Protein extracts from equal numbers of L1 larvae were used for anti-GFP western blotting and the appearance of cleaved GFP quantified, n = 3 (C). Significances calculated using two-tailed Student's t-test, * p < 0.05.
First, we assessed movement abilities of all larvae right after starvation (scheme in Figure 7B, results in Figure 7C, left panel). We found that the majority of wild-type animals showed prominent movement after being starved for 1, 2 or 3 days ( Figure 7C, left panel, animals not moving in black) and the majority also recovered from these periods of starvation in the presence of food ( Figure 7C, right panel, animals not recovering in black). In contrast, less immediate movement upon 1 to 3 days LGG-1-PE and cleaved GFP is detected by Western blotting using GFP:: LGG-1 expressing strains along with wild-type N2 control (A, right). Cleaved GFP, which can also be visualized by Western blotting, accumulates upon lysosomal inhibition due to a decelerated degradation using NH 4 Cl (A, right). L1 animals of wild-type (Control), atg-18(gk378), and epg-6(bp242) expressing the adIS2122 transgene were starved overnight in the presence (+) or absence (-) of NH 4 Cl, and proteins were extracted for anti-GFP Western blotting (B). A representative short exposure is shown and GFP:: LGG-1 and GFP:: LGG-1-PE migrations are indicated (B, upper panel). Likewise, a longer exposure is shown and the migration of cleaved GFP indicated (B, lower panel). Protein extracts from equal numbers of L1 larvae were used for anti-GFP western blotting and the appearance of cleaved GFP quantified, n = 3 (C). Significances calculated using two-tailed Student's t-test, * p < 0.05.
When C. elegans eggs hatch in the absence of food, L1 larvae enter an arrested stage, which wild-type larvae survive for up to three weeks and upon re-feeding they recover and develop further [21]. As proper autophagic activity is required for the survival of this starvation period [11,21,29], we aimed to further investigate the physiological roles of ATG-18 and EPG-6 in C. elegans by measuring the survival and recovery of L1 larvae upon starvation. (Figure 7). Following synchronization, L1 larvae (atg-18(gk378), epg-6(bp242), and atg-18(gk378);epg-6(bp242)) were continuously starved over a period between 1 to 29 days and their survival was assessed by counting larvae that recovered from starvation and progressed through larval stage L2 in the presence of food ( Figure 7A,B).
First, we assessed movement abilities of all larvae right after starvation (scheme in Figure 7B, results in Figure 7C, left panel). We found that the majority of wild-type animals showed prominent movement after being starved for 1, 2 or 3 days ( Figure 7C, left panel, animals not moving in black) and the majority also recovered from these periods of starvation in the presence of food ( Figure 7C, right panel, animals not recovering in black). In contrast, less immediate movement upon 1 to 3 days of starvation was observed in the absence of EPG-6 ( Figure 7C, left panel, not moving animals are shown in dark red), as well as reduced survival in the presence of food ( Figure 7C, right panel, not recovering animals are shown in dark red). This result is in line with previous reports [10]. Less movement ( Figure 7C, left panel, animals not moving in light red) and poorer recovery in the presence of food after starvation periods between 1 to 3 days ( Figure 7C, right panel, animals not recovering in light red) was even more prominent in the absence of ATG-18, and severe in the absence of both EPG-6 and ATG-18 ( Figure 7C, left panel, not moving animals are shown in gold; right panel, not recovering animals are shown in gold). Moreover, when we assessed the survival of L1 larvae over a longer period (up to 29 days) of starvation ( Figure 7D; Data S1), differences in the capacities to recover became even more apparent when comparing wild-type animals with mutant strains deficient for EPG-6 and/or ATG-18 ( Figure 7D). of starvation was observed in the absence of EPG-6 ( Figure 7C, left panel, not moving animals are shown in dark red), as well as reduced survival in the presence of food ( Figure 7C, right panel, not recovering animals are shown in dark red). This result is in line with previous reports [10]. Less movement ( Figure 7C, left panel, animals not moving in light red) and poorer recovery in the presence of food after starvation periods between 1 to 3 days ( Figure 7C, right panel, animals not recovering in light red) was even more prominent in the absence of ATG-18, and severe in the absence of both EPG-6 and ATG-18 ( Figure 7C, left panel, not moving animals are shown in gold; right panel, not recovering animals are shown in gold). Moreover, when we assessed the survival of L1 larvae over a longer period (up to 29 days) of starvation ( Figure 7D; Data S1), differences in the capacities to recover became even more apparent when comparing wild-type animals with mutant strains deficient for EPG-6 and/or ATG-18 ( Figure 7D).

Figure 7.
Survival of L1 larvae upon starvation decreased in the absence of ATG-18 and EPG-6. The overall experimental set-up for the survival analysis upon starvation conditions (A). L1 larvae hatched in nutrient-free medium and were incubated in this medium for the indicated time. The larvae were then tested for recovery on bacteria-covered plates, while their movement was recorded. Subsequently, the percentage of animals that recovered from starvation in the presence of food was calculated (B). The percentages of animals that showed movement upon starvation of 1, 2 or 3 days (C, left panel) is shown, as well as the percentages of animals recovering and surviving from the indicated starvation periods in the presence of food: e.g., 1 (+2) stands for 1 day of starvation followed by 2 days of feeding (C, right panel). Likewise, over a period of 1-29 days, the percentages of surviving animals were recorded where atg-18 and/or epg-6 mutant L1 larvae showed a decreased tolerance to starvation (D). The overall experimental set-up for the survival analysis upon starvation conditions (A). L1 larvae hatched in nutrient-free medium and were incubated in this medium for the indicated time. The larvae were then tested for recovery on bacteria-covered plates, while their movement was recorded. Subsequently, the percentage of animals that recovered from starvation in the presence of food was calculated (B). The percentages of animals that showed movement upon starvation of 1, 2 or 3 days (C, left panel) is shown, as well as the percentages of animals recovering and surviving from the indicated starvation periods in the presence of food: e.g., 1 (+2) stands for 1 day of starvation followed by 2 days of feeding (C, right panel). Likewise, over a period of 1-29 days, the percentages of surviving animals were recorded where atg-18 and/or epg-6 mutant L1 larvae showed a decreased tolerance to starvation (D).
Functional autophagy and lifespan control was closely correlated in C. elegans, and loss of ATG-18 has previously been shown to decrease lifespan [11,12]. However, the role of EPG-6 in adult lifespan regulation had not yet been addressed. Hence, we investigated whether epg-6 mutant animals would also display a short-lived phenotype in our approach. We performed lifespan assessments by counting living animals on NGM/E. coli OP50 plates in the presence of 5-fluorodeoxyuridine (FUDR), a compound that prevents egg hatching ( Figure 8A-C). Our results confirmed that atg-18 mutants have a significantly decreased lifespan (light-red label) compared to the wild-type (black label), (Figure 8B,C). Unexpectedly, epg-6 mutant worms (dark-red label) lived significantly longer than wild-type worms ( Figure 8B,C). Loss of EPG-6 in atg-18 mutant background (atg-18(gk378);epg-6(bp242), gold label) suppressed the short-lived phenotype of atg-18 mutants and shifted the lifespan curve to wild-type levels ( Figure 8B,C). Likewise, downregulation of EPG-6 by RNAi in atg-18 mutant animals prolonged their lifespan ( Figure S2). These results suggest opposing roles for ATG-18 and EPG-6 in lifespan control. Supporting our data, decreased GFP:: LGG-1 puncta formation was observed in the seam cells upon atg-18 RNAi treatment [16] and in atg-18 mutants [18]. Enlarged GFP::LGG-1 punctate   To assess if the compound FUDR may influence the lifespan, we performed lifespan measurements also in the absence of FUDR by transferring animals every day onto fresh plates in order to separate them from their progeny ( Figure 8D-F). Under these conditions, loss of EPG-6 function did not extend lifespan when compared to wild-type animals, but also did not show a short-lived phenotype as again seen for atg-18 mutant animals ( Figure 8E,F). These results suggest that the anticipated autophagy-independent role of EPG-6 in lifespan control may be connected to germline function, which is counteracted by the presence of FUDR.

Discussion
In line with previous observations [10], we provide evidence that in the absence of PROPPIN [4,30] members ATG-18 and EPG-6, both of which are considered to function as PI3P effectors at the nascent autophagosome in C. elegans [31,32], autophagosome formation and the autophagic flux was negatively influenced from early embryo throughout all larval stages (L1-4) (Figures 2-6). We demonstrated that epg-6 loss-of-function mutants were less fitted to survive L1 starvation periods [10] and ATG-18 deficient strains had almost completely lost their capability to recover from L1 starvation ( Figure 7). Moreover, we found that epg-6 mutants not only survived starvation better than atg-18 mutants, but while atg-18 loss-of-function mutants displayed a short-lived phenotype [11,12], loss of EPG-6 function promoted longevity in C. elegans (Figure 8).
By characterizing the atg-18, epg-6 and atg-18;epg-6 mutant strains we observed high lethality at early larval development (L1) in the absence of ATG-18 (Figure 1), suggesting that ATG-18 is essential even under favourable conditions. To assess the abundance of formed autophagic structures, we measured GFP::LGG-1 puncta formation throughout development of C. elegans. In early embryos GFP::LGG-1 puncta count and size decreased in atg-18 and epg-6 mutants compared to wild-type (Figure 2), indicating an inhibition of autophagosome formation. It was reported that in atg-18 mutant embryos small, in epg-6 mutant embryos enlarged LGG-1 puncta accumulate [10,14,33]; however, we observed enlarged GFP::LGG-1 puncta accumulation of both atg-18 and epg-6 mutants in late-stage embryos. The difference to reported studies may be due to the different embryonic stages investigated. As GFP::LGG-1 puncta abundance changes dynamically during embryonic development (see Figure 2, early, late embryos of N2), the developmental stage presents a critical parameter when assessing LGG-1 puncta formation by fluorescence microscopy.
During postembryonic development we observed enlarged GFP:: LGG-1 structures both in the head and seam cells of mutant atg-18 and epg-6 L1 through L4 larvae; however, in the head, puncta accumulated, while in the seam cells, puncta count decreased (Figures 3-5).
Supporting our data, decreased GFP:: LGG-1 puncta formation was observed in the seam cells upon atg-18 RNAi treatment [16] and in atg-18 mutants [18]. Enlarged GFP::LGG-1 punctate structures in the head and seam cells during larval development suggests that autophagic flux was impaired. In addition, in the seam cells, autophagy initiation might also be inhibited leading to decreased puncta number. In summary, our quantitative assessment of GFP::LGG-1 puncta by fluorescence microscopy showed that aberrant autophagosomal membranes accumulated in the absence of ATG-18 and EPG-6, and in the seam cells, the initiation of autophagosome formation, a feature attributed to a function upstream of LGG-1 (or LC3 in mammals) [10], was also inhibited.
To further investigate whether autophagic flux was impaired, lipidation of GFP:: LGG-1 was analyzed by Western blotting and also upon lysosomal inhibition ( Figure 6A) [17,21]. Increased LGG-1 and LGG-1-PE levels have been reported for both atg-18 and epg-6 mutant strains [10,14,32], which we also observed ( Figure 6B). We also showed that cleaved GFP decreased in atg-18 and epg-6 mutants, and while cleaved GFP was stabilized by inhibiting lysosomal degradation in wild-type larvae, this was not the case in epg-6 and atg-18 mutants ( Figure 6B). These results further support the idea that autophagic flux was impaired in the absence of both ATG-18 and EPG-6 in C. elegans ( Figure 6). Consistent with the observation by Lu and colleagues [10], we demonstrated that epg-6 mutant animals survive starvation in L1 larval stage poorly. Furthermore, we demonstrate that atg-18 mutant animals tolerate starvation even worse, and in the absence of both ATG-18 and EPG-6 worms cannot survive starvation. Hence, both ATG-18 and EPG-6 are necessary for L1 larvae survival under starvation conditions and for the recovery from L1 starvation periods in the presence of food ( Figure 7). As autophagic activity increases in response to starvation in order to degrade polymers and release monomers for recycling purposes, loss of essential autophagy factors, like ATG-18 and EPG-6, likely leads to decreased survival of starvation. Our results support the general concept that autophagy functions as an evolutionarily conserved pathway engaged to secure survival upon starvation, as early demonstrated in autophagy-deficient mice models [1,34].
Autophagy has an essential role in lifespan regulation as impairment of autophagy is connected to decreased lifespan, hence, autophagy is considered an anti-ageing and pro-survival mechanism [35]. Our results provide evidence that ATG-18 and EPG-6 have opposing roles in lifespan control (Figure 9). analyzed by Western blotting and also upon lysosomal inhibition ( Figure 6A) [17,21]. Increased LGG-1 and LGG-1-PE levels have been reported for both atg-18 and epg-6 mutant strains [10,14,32], which we also observed ( Figure 6B). We also showed that cleaved GFP decreased in atg-18 and epg-6 mutants, and while cleaved GFP was stabilized by inhibiting lysosomal degradation in wild-type larvae, this was not the case in epg-6 and atg-18 mutants ( Figure 6B). These results further support the idea that autophagic flux was impaired in the absence of both ATG-18 and EPG-6 in C. elegans ( Figure  6). Consistent with the observation by Lu and colleagues [10], we demonstrated that epg-6 mutant animals survive starvation in L1 larval stage poorly. Furthermore, we demonstrate that atg-18 mutant animals tolerate starvation even worse, and in the absence of both ATG-18 and EPG-6 worms cannot survive starvation. Hence, both ATG-18 and EPG-6 are necessary for L1 larvae survival under starvation conditions and for the recovery from L1 starvation periods in the presence of food ( Figure  7). As autophagic activity increases in response to starvation in order to degrade polymers and release monomers for recycling purposes, loss of essential autophagy factors, like ATG-18 and EPG-6, likely leads to decreased survival of starvation. Our results support the general concept that autophagy functions as an evolutionarily conserved pathway engaged to secure survival upon starvation, as early demonstrated in autophagy-deficient mice models [1,34].
Autophagy has an essential role in lifespan regulation as impairment of autophagy is connected to decreased lifespan, hence, autophagy is considered an anti-ageing and pro-survival mechanism [35]. Our results provide evidence that ATG-18 and EPG-6 have opposing roles in lifespan control ( Figure 9). Figure 9. Model for the role of ATG-18 and EPG-6 in autophagy and longevity in C. elegans. Both ATG-18 and EPG-6 are considered to function in autophagy as PI3P effectors on the nascent autophagosome, equivalent to the role of their homologues in mammals [7]. In this context, EPG-6 is considered to function downstream of ATG-18 during autophagosome formation [7,10]. A novel, autophagy-independent role of EPG-6 is postulated in lifespan control, which is likely connected to germline signaling.
According to previous reports, we observed decreased lifespan in the absence of ATG-18 [3], while in the absence of EPG-6 we found that lifespan was significantly increased and atg-18;epg-6 double-mutants lived also longer than atg-18 worms ( Figure 8). As autophagy is considered to counteract ageing, and autophagic degradation is impaired in the absence of EPG-6, we suggest that the lifespan promoting effect of EPG-6 is autophagy-independent. As ATG-18 was recently reported EPG-6 ATG-18 Autophagy Longevity Figure 9. Model for the role of ATG-18 and EPG-6 in autophagy and longevity in C. elegans. Both ATG-18 and EPG-6 are considered to function in autophagy as PI3P effectors on the nascent autophagosome, equivalent to the role of their homologues in mammals [7]. In this context, EPG-6 is considered to function downstream of ATG-18 during autophagosome formation [7,10]. A novel, autophagy-independent role of EPG-6 is postulated in lifespan control, which is likely connected to germline signaling.
According to previous reports, we observed decreased lifespan in the absence of ATG-18 [3], while in the absence of EPG-6 we found that lifespan was significantly increased and atg-18;epg-6 double-mutants lived also longer than atg-18 worms ( Figure 8). As autophagy is considered to counteract ageing, and autophagic degradation is impaired in the absence of EPG-6, we suggest that the lifespan promoting effect of EPG-6 is autophagy-independent. As ATG-18 was recently reported to regulate lifespan cell non-autonomously [17,18], it is of interest to assess in future studies whether EPG-6 may control lifespan in C. elegans independent of its role in autophagy, also in a cell non-autonomous manner. As increased lifespan upon knockdown of some autophagy genes was reported before [13], it is of intense interest to assess autophagy-dependent and autophagy-independent functions of ATGs in lifespan control, and also, whether autophagy may promote ageing under some conditions. Interestingly, our lifespan analyses revealed an FUDR-dependent lifespan regulating mechanism in the absence of EPG-6 ( Figure 8). Fluorodeoxyuridine (FUDR) is a thymidylate synthase inhibitor, which inhibits DNA synthesis in developing embryos. In early publications, it was reported that FUDR allows synchronization without interfering with the lifespan of C. elegans [36]; however, recent publications suggest that FUDR can prolong the lifespan of both wild-type and certain mutant animals [37][38][39]. Our experiments support this observation, as FUDR-treated wild-type animals indeed showed extended lifespan compared to untreated animals ( Figure 8) and we also observed lifespan extension of epg-6 mutants upon FUDR treatment. In this context, it is of great interest that FUDR was demonstrated to activate stress response pathways [40]. With this notion taken into consideration, it may be that the long-lived phenotype observed in the absence of EPG-6 function is provoked by FUDR conferring stress conditions, and that EPG-6's loss-of-function under non-stress conditions would not influence lifespan, which in fact was found when we assessed the lifespan of epg-6 mutant animals in the absence of FUDR. In fact, several ATG mutant strains have been identified where the expected short-lived phenotype was not present under non-stress conditions [16,41,42].
It also bears mentioning that germline ablation increases the lifespan of C. elegans in a DAF-2/DAF-16 dependent manner [43], and increased lifespan upon FUDR treatment was reported for FEM-3, a protein involved in oogenesis [38]. As the lifespan extension of the epg-6 mutant animals was observed under FUDR treatment conditions, it is tempting to speculate that the autophagy-independent function of EPG-6 could likely be found in the germline, as this is a prominent pathway affected by FUDR treatment. Here, EPG-6 may have a lifespan-repressing function as part of the overall ageing program.
Autophagy-independent functions of ATG proteins, as here suggested for EPG-6 in lifespan control, have been reported for several ATG proteins [44][45][46][47]. In fact, the serine-/threonine-specific protein kinase ULK1, UNC-51 in C. elegans, activated upon energy or nutrient deprivation in order to induce autophagy via signaling through the PI3KC3 [1,2,45], was originally identified as being required for extending axons along the dorsoventral axis in C. elegans [48]. The UNC-51 protein interacts with UNC-14 [49,50] and regulates the intracellular localization of the netrin receptor UNC-5 in C. elegans [51], thereby controlling axon guidance [52] independent of UNC-51's role in autophagy [46,53,54]. Moreover, it has been reported that constitutive autophagy in the central nervous system in mice is executed in the absence of Ulk1 (and Ulk2) [46], a process referred to as non-canonical autophagy, whereby autophagy proceeds in the absence of canonical ATG proteins [44]. In this context, it was demonstrated that Ulk1/UNC-51 regulates ER-to-Golgi trafficking [46]. Further autophagy-independent functions of ATG proteins have been described. By way of example, ATG7, a critical member of two ubiquitin-like conjugation systems in autophagy and required for LC3 lipidation [1,45], was shown to positively regulate the expression of cell cycle arrest genes in complex with the tumor suppressor p53 and in response to metabolic stress in mice [55]. Further, it was also found that ATG16L1, functioning in association with ATG5-ATG12 in promoting the lipidation of LC3 at autophagosomal membranes [1,45] can also promote LC3 lipidation at endolysosomal membranes during non-canonical autophagy [56].
The examples above refer to autophagy-independent functions of ATG proteins, where it has been shown that homologous proteins in different species carry out equivalent biological functions. With regard to the role of ATG-18 and EPG-6, this assignment still awaits clarification. Whereas C. elegans expresses the two PROPPIN members ATG-18 and EPG-6, three PROPPIN members, Atg18, Atg21 and HSV2, have been identified in yeast, and the four WIPI proteins (WIPI1 through WIPI4) present the PROPPIN members in humans [4]. Caenorhabditis elegans ATG-18 belongs to one paralogous group of the PROPPIN family together with yeast Atg18 and Atg21, and human WIPI1 and WIPI2. Caenorhabditis elegans EPG-6 belongs to the other PROPPIN group together with yeast HSV2 and human WIPI3 and WIPI4 [8]. Based on this phylogenetic relationship it can be anticipated that PROPPIN members belonging to the same paralogous group share equivalent biological functions. This, however, seems different with regard to the PROPPIN family as follows. The ancestor of the PROPPIN family, Atg18 in Saccharomyces cerevisiae [57], is essential for autophagy and interacts with Atg2 [57][58][59] for tethering the forming autophagosome to the endoplasmic reticulum in yeast [60]. Although in C. elegans, ATG-18 also fulfills an essential role in autophagy, its paralogue EPG-6 was shown to interact with ATG-2 in order to control autophagosome expansion downstream of ATG-18 [10]. This has also been shown for human WIPI4 that specifically binds to ATG2 and controls the expansion of the nascent autophagosome, downstream of WIPI1 and WIPI2 [7]. Hence, the ancestral yeast Atg18 function is replicated in EPG-6 in C. elegans and in WIPI4 in humans. Our results here point out that ATG-18 and EPG-6 should also fulfill non-redundant roles in lifespan regulation, as in C. elegans, both are required for autophagy which is considered to counteract ageing. In future studies it will be of interest to decipher the differential contribution of PROPPIN members in lifespan regulation, not only in C. elegans, but also in other model organisms, including mice.