The LGG-1 protein is highly conserved from residue 1 to residue 116, sharing 92% and 74% similarity with the human GABARAP and the yeast Atg8, respectively (Manil-Ségalen et al., 2014). However, the GEVEKKE C-terminus of LGG-1 is unusual by its length and the presence of a non-conserved glycine residue in position 117 (Figure 1B, Figure 1—figure supplement 1). As consistent with other Caenorhabditis species as well as several nematodes and arthropods, the presence of a C-terminal di-glycine is reminiscent of other ubiquitin-like proteins such as SUMO and Nedd8 (Cappadocia and Lima, 2018; Jentsch and Pyrowolakis, 2000). These specificities raise the possibility that the C-terminus could confer particular functions to the precursor and the cleaved form.

To analyze the functions of LGG-1 P and LGG-1 I, a CRISPR-Cas9 approach was used to substitute the conserved glycine 116 by an alanine, and to generate three specific lgg-1 mutants with various C-terminus (Figure 1B). In theory, both lgg-1(G116A) and lgg-1(G116AG117A) mutants were expected to accumulate a P form due to the blockage of its cleavage by ATG-4 (Wu et al., 2012). Alternatively, the lgg-1(G116AG117*) mutant should produce a form I. Five supplementary lgg-1 frameshift mutants were isolated during the CRISPR experiments, resulting in deletion/insertion at the C-terminus (Figure 1B and Figure 1—figure supplement 1). Among them, lgg-1(ΔC100-123) and lgg-1(ΔC112-123) have been used in the present study. The allele lgg-1(tm3489), which deletes 48% of the open reading frame, was used as a negative control (Manil-Ségalen et al., 2014) as it is considered as a null mutant, and thereafter noted lgg-1(Δ).

To assess whether lgg-1(G116A), lgg-1(G116AG117A) and lgg-1(G116AG117*) alleles code for a precursor and form I, respectively, we performed a western blot analysis with two different LGG-1 antibodies (Al Rawi et al., 2011; Springhorn and Hoppe, 2019; Figure 1C). In basal conditions the wild-type LGG-1 was mainly present as form I (13.9 kDa) with a low amount of the faster migrating form II and no detectable precursor (14,8 kDa)(Figure 1C), while no band was observed in the allele lgg-1(tm3489) confirming that it is a bona fide null mutant. While the lgg-1(G116AG117*) mutant presented a major form I, the lgg-1(G116A) mutant accumulated both the expected precursor form and form I. This indicated that the cleavage of the LGG-1(G116A) precursor was still present although less efficient. In both mutants, an unexpected minor form was observed migrating differently from the lipidated form II, which was no longer detected. The lgg-1(G116AG117A) mutant accumulated the precursor form (96% of the protein) indicating that the cleavage observed in the LGG-1(G116A) was dependent on the presence of a second glycine in position 117.

The respective protein substitutions were further confirmed by mass spectrometry analyses after affinity purification of LGG-1(G116A) and LGG-1(G116AG117*) (Figure 1—figure supplement 2). The identification of C-terminal peptides validated the expected precursor form in LGG-1(G116A) and its cleavage after A116, and confirmed A116 as the last residue in LGG-1(G116AG117*). These latter forms are called hereafter ‘cleaved form’ and ‘truncated form’, respectively.

To confirm western blot analyses, we next performed immunofluorescence in the embryo to analyze the subcellular localization of LGG-1 protein from the various alleles. At the one-cell-stage and around 100 cells-stage, two selective autophagy processes have been well characterized, removing paternal mitochondria and maternal aggregates, respectively (Al Rawi et al., 2011; Sato and Sato, 2011; Zhang et al., 2009). The punctate staining, observed in the wild-type animals (Figure 1D) with two independent anti-LGG-1 antibodies, was characteristic for the autophagosomes formed during each process, and was absent in the lgg-1(Δ) mutant (Figure 1E). The five mutants lgg-1(G116A), lgg-1(G116AG117*), lgg-1(G116AG117A), lgg-1(ΔC100-123), and lgg-1(ΔC112-123) presented no puncta but a diffuse cytosolic staining. (Figure 1F–K), indicating that neither the precursor nor the form I are able to conjugate to the autophagosome membrane. The increase of the diffuse signal in lgg-1(G116A) and lgg-1(G116AG117*) embryos (Figure 1L) suggests that the protein is less degraded in these mutants. Moreover, no LGG-1(G116A) puncta were observed after depleting the tethering factor EPG-5 compared to the strong accumulation of puncta in LGG-1(wt) (Figure 1—figure supplement 2; Tian et al., 2010).

We performed cellular fractionation of membrane vesicles to test whether LGG-1(G116A) and LGG-1(G116AG117*) are associated with autophagosomes. Compared with ER resident SEL-1 or ER-associated CDC-48, the LGG-1(wt) protein was detected in both the cytosolic fraction and the membrane pellet and could only be extracted with high salt or the detergent Triton X-100. In contrast to LGG-1(wt), both LGG-1(G116A) and LGG-1(G116AG117*) were absent in the membrane pellet fraction (Figure 1M), suggesting defective lipidation of both LGG-1 mutant proteins. In an alternative approach, we observed the localization of overexpressed GFP::LGG-1 and GFP::LGG-1(G116A) (Manil-Ségalen et al., 2014) after induction of autophagic flux by acute heat stress (aHS) (Chen et al., 2021; Kumsta et al., 2017). After aHS, GFP::LGG-1 formed numerous puncta that further accumulated when autolysosome formation was impaired by depletion of RAB-7 or EPG-5 (Figure 1—figure supplement 3). In contrast, in GFP::LGG-1(G116A), puncta were not reduced under any condition. Electron microscopy and immunogold labeling confirmed that GFP::LGG-1 was frequently detected to autophagosome membranes (Manil-Ségalen et al., 2014), whereas GFP::LGG-1(G116A) was rarely detected in association with autophagosomes and in these rare cases was predominantly localized in the lumen (Figure 1—figure supplement 3). Taken together, these results suggest that the G116A mutation does not allow conjugation of LGG-1 to the autophagosome membrane despite its cleavage. LGG-1(G116AG117A) represents only a precursor form and LGG-1(G116AG117*) only a truncated form, whereas LGG-1(G116A) produces both a precursor and a cleaved form. This allele series provides an ideal situation to study the respective roles of the precursor and form I in absence of lipidated form II.

The developmental phenotypes of the mutants lgg-1(G116A), lgg-1(G116AG117A), and lgg-1(G116AG117*) were explored in embryo, larvae, and adults and compared with lgg-1(Δ) and wild-type animals (Figure 2). We confirmed that lgg-1(Δ) homozygous animals present a massive lethality during late embryogenesis or first larval stage (Figure 2B and H; Manil-Ségalen et al., 2014). However, few escapers, circa 8% of the progeny, were able to reach adulthood and reproduce, allowing to maintain a lgg-1(Δ) homozygous population.

Neither lgg-1(G116A) nor lgg-1(G116AG117*) homozygous animals presented any observable defect in development (Figure 2C, D and H) or adulthood and they reproduced at a similar rate compared to wild-type animals (Figure 2I). In contrast, lgg-1(G116AG117A) and the five independent mutants harboring various deletions and frameshifts of the C-terminus presented a very strong lethality with the characteristic embryonic phenotype of lgg-1(Δ) animals (Figure 2E–H). Among them lgg-1(Δ112–123) presented a premature stop codon at position 112 and two others a frameshift in position 114 leading to an extension of the C-terminus (Figure 2 and Figure 1—figure supplement 1).

These data indicate that the cleaved LGG-1(G116A) and the truncated LGG-1(G116AG117*) forms, but not the precursor, are sufficient to recapitulate the normal development and viability, independently of membrane conjugation. These data suggest that cleavage of the C-terminus is necessary for LGG-1 developmental functions.

To address the functionality of LGG-1 precursor and form I, we analyzed autophagy-related processes that have been well characterized during C. elegans life cycle (Leboutet et al., 2020). Selective autophagy was studied in the early embryo, where a stereotyped mitophagy process occurs. The degradation of selective cargos was observed in live embryos using specific labeling of the paternal mitochondria (HSP-6::GFP and mitoTracker, Figure 3A–F and Figure 3—figure supplement 1). In lgg-1(Δ) animals, the cargos accumulated while they were degraded in the wild-type situation. In lgg-1(G116A), but neither in lgg-1(G116AG117*) nor in lgg-1(G116AG117A) mutants, paternal mitochondria were degraded, suggesting that the LGG-1(G116A) protein maintained autophagic activity.

Bulk autophagy was then studied by starvation of the first stage larvae (Figure 3G). While lgg-1(G116AG117*) and lgg-1(G116AG117A) mutants displayed a marked decrease of survival, lgg-1(G116A) mutants showed no difference compared the wild-type animals. Moreover, the longevity of adults, which depends on bulk autophagy, was similar for lgg-1(G116A) and wild-type animals (Figure 3H), but strongly reduced for lgg-1(G116AG117*) mutants.

The autophagic capacity of LGG-1(G116A) protein, but not LGG-1(G116AG117*) or LGG-1(G116AG117A), was further documented by the elimination of apoptotic corpses in the embryo (Figure 3I, and Figure 3—figure supplement 1; Jenzer et al., 2019).

Overall, these data demonstrate that, despite its defect to localize to autophagosomes, LGG-1(G116A) achieves both selective and bulk autophagy during physiological and stress conditions. This is the first in vivo evidence that the autophagy functions of LGG-1/GABARAP can be uncoupled from its membrane conjugation. The non-functionality of LGG-1(G116AG117A) suggests that the precursor form is not responsible of LGG-1(G116A) autophagy activity. Despite an identical protein sequence, the truncated LGG-1(G116AG117*) is not functional in autophagy, indicating that the cleavage of the C-terminus from the precursor is essential for the functionality of LGG-1(G116A). Moreover, the normal development of lgg-1(G116AG117*) animals demonstrates that the developmental functions of LGG-1 are independent of its autophagic functions. Interestingly, the expression in S. cerevisiae of LGG-1(wt) and LGG-1(G116A), but not LGG-1(G116AG117*), slightly improved the nitrogen starvation survival of atg8Δ mutant (Supplementary data and Figure 3—figure supplement 2), suggesting that the LGG-1(G116A) retains a partial autophagy functionality in the yeast.

Our previous study has shown a partial redundancy of LGG-1 and LGG-2 during starvation survival, and longevity (Alberti et al., 2010), which raises the possibility of functional compensation of lgg-1(G116A) by LGG-2. To test this possibility, we used the large deletion mutant lgg-2(tm5755), which is considered as a null (Manil-Ségalen et al., 2014), and constructed the double mutant strains lgg-1(G116A); lgg-2(tm5755) and lgg-1(G116AG117*); lgg-2(tm5755).

Similar to the single mutants lgg-1(G116A) and lgg-2(tm5755), the double mutant lgg-1(G116A); lgg-2(tm5755) animals were viable and presented no morphological defect (Figure 4A–F). These data indicate that the correct development of lgg-1(G116A) is not due to a compensative mechanism involving lgg-2.

Next, we analyzed the autophagy functions in lgg-1(G116A); lgg-2(tm5755) animals. If LGG-2 compensates for LGG-1(G116A) in autophagy, lgg-1(G116A); lgg-2(tm5755) animals should behave similarly to lgg-1(G116AG117*); lgg-2(tm5755) (of note lgg-1(Δ); lgg-2(tm5755) animals are not viable). The lgg-1(G116A); lgg-2(tm5755) animals presented a decrease for both survival to starvation and longevity compared to lgg-1(G116A) single mutant. However, they survived better than lgg-1(G116AG117*); lgg-2(tm5755) animals (Figure 4G and H). These results indicate that the functionality of LGG-1(G116A) in bulk autophagy partially relies on LGG-2. Selective autophagy during early embryogenesis was then quantitatively analyzed in the double mutant strains (Figure 4I–K). Surprisingly, paternal mitochondria were eliminated in the lgg-1(G116A); lgg-2(tm5755) animals indicating that LGG-1(G116A) was sufficient for the allophagy process. This suggests that paternal mitochondria could be degraded by autophagosomes devoid of both LGG-1 and LGG-2. However, a delay in the degradation was observed compared to lgg-1(G116A) animals suggesting that the autophagy flux is reduced. These results revealed a partial redundancy between LGG-1 and LGG-2 in autophagy, but demonstrated at the same time that LGG-1(G116A) fulfills developmental functions and maintains some autophagy activity independent of LGG-2.

Interestingly, this detailed analysis also revealed a slight delay in the elimination of paternal mitochondria in lgg-1(G116A) animals compared to wild-type (Figure 4K). Although the cleaved LGG-1 is sufficient for autophagy, this observation suggests that loss of membrane targeting could affect the dynamics of autophagy flux.

The autophagic flux and the dynamics of autophagosome formation were compared between lgg-1(G116A), lgg-1(G116AG117*) and lgg-1(Δ) animals. We first focused on the early embryo where the autophagy process is stereotyped and the nature of the cargos and the timing of degradation are well characterized. Moreover, the autophagosomes sequestering the paternal mitochondria were clustered and positive for LGG-2 (Figure 5A; Manil-Ségalen et al., 2014). In lgg-1(Δ) mutant, LGG-2 autophagosomes were not detected as a cluster but were spread out in the whole embryo as single puncta that persisted after the 15 cells stage (Figure 5B and E). This indicated that individual LGG-2 structures could be formed in absence of LGG-1, but were not correctly localized and not degraded properly, presumably because of the role of LGG-1 in cargo recognition (Sato et al., 2018) and of its latter involvement in the maturation of autophagosomes, respectively. The pattern of LGG-2 was somehow different in lgg-1(G116A) and lgg-1(G116AG117*) mutants, forming sparse structures of heterogeneous size, which persisted longer (Figure 5C–F). These data suggested that the cleaved and the truncated LGG-1 could both promote the recruitment of LGG-2 to autophagic structures, but display an altered autophagic flux. The analysis of the colocalization between paternal mitochondria and LGG-2 did not reveal an increase in lgg-1(G116A) or lgg-1(G116AG117*) mutants (Figure 5G–J and Figure 5—figure supplement 1). These data suggested that the elimination of paternal mitochondria in lgg-1(G116A) animals was not due to the enhanced recruitment of LGG-2. A western blot analysis of worm lysates indicated that there was no increase of LGG-2 expression in lgg-1(Δ), lgg-1(G116A), and lgg-1(G116AG117*) mutants (Figure 5K).

The autophagic structures in lgg-1(G116A) and lgg-1(G116AG117*) embryos were further characterized by electron microscopy and compared with wild-type and lgg-1(Δ) mutant embryos (Figure 6). In wild-type animals, autophagosomes containing cytoplasmic materials (referred as type 1) and the characteristic paternal mitochondria (Zhou et al., 2016) were observed in early embryos (Figure 6A). At that stage, rare autophagosomes containing partially degraded material were present (referred as type 2). As expected, almost no autophagosome was observed in lgg-1(Δ) embryos and paternal mitochondria were non-sequestered (Figure 6B). In lgg-1(G116A) embryos, the numbers of type 1 and type 2 autophagosomal structures increased. The autophagosomes appeared to be closed and contained various cellular materials and membrane compartments (Figure 6C–E). This confirmed that LGG-1(G116A) was sufficient to form functional autophagosomes but with delayed degradation. On the other hand, lgg-1(G116AG117*) embryos presented non-sequestered paternal mitochondria (Figure 6F) and multi-lamellar structures containing cytoplasm but no membrane organelles (type 3 Figure 6G and K). The analysis of the double mutant strains lgg-1(G116A); lgg-2(tm5755) and lgg-1(G116AG117*); lgg-2(tm5755) revealed the presence of types 1 and 2 autophagosomes, but less frequent than in the single lgg-1 mutants (Figure 6H–K). This data confirmed that LGG-1(G116A) alone was able to initiate the formation of autophagosomes but less efficiently in absence of LGG-2. Type 3 structures were only observed in lgg-1(G116AG117*) and lgg-1(G116AG117*); lgg-2(tm5755), suggesting a neomorphic function of the truncated LGG-1(G116AG117*) protein that induced a non-functional compartment.

Altogether, these data indicate that the cleaved, but not the truncated, LGG-1 form I is able to form functional autophagosomes with a delayed degradation.

To better understand the function of LGG-1 form I during autophagy flux, we next analyzed a developmental aggrephagy process (Figure 7). The Zhang lab has demonstrated that aggregate-prone proteins are degraded through autophagy in C. elegans embryo through liquid-liquid phase separation promoted by the receptor SEPA-1 and regulated by the scaffolding protein EPG-2 (Lu et al., 2011; Tian et al., 2010; Wu et al., 2015; Zhang et al., 2018; Zhang et al., 2009). Initiation and elongation of autophagosomes were analyzed by quantifying the colocalization between ATG-18/WIPI2 and LGG-2 (Figure 7A–E) during autophagosome formation. ATG-18, the worm homolog of the omegasome marker WIPI2 (Polson et al., 2010), acts at an early step of biogenesis (Lu et al., 2011). Puncta labelled with ATG-18 only, both ATG-18 and LGG-2, or LGG-2 only were considered as omegasomes, phagophores, and autophagosomes, respectively. In lgg-1(RNAi) animals the number of omegasomes increased while the proportion of phagophore decreased compared to the wild-type embryos (Figure 7A, B and E). This indicates that the initiation of autophagy was triggered in absence of LGG-1, but the biogenesis of autophagosome was defective. lgg-1(G116A) animals showed no difference with the wild-type (Figure 7C and E) supporting that both initiation and phagophore extension are normal with the cleaved LGG-1. Similar to the lgg-1(RNAi), lgg-1(G116AG117*) animals were defective in the phagophore extension (Figure 7D and E). These data confirmed that the cleaved LGG-1(G116A), but not the truncated LGG-1(G116AG117*), is functional for the early step of autophagosome biogenesis. Moreover, RNAi depletion demonstrated that the function of LGG-1(G116A) in aggrephagy pathway was dependent on UNC-51/Ulk1 and the scaffolding protein EPG-2 (Supplementary data and Figure 7—figure supplement 1).

Quantification of SEPA-1::GFP in late embryo showed that lgg-1(G116A) mutant was able to perform aggrephagy but not lgg-1(G116AG117*) or lgg-1(G116AG117A) mutants (Figure 7F–K). However, the elimination was decreased compared to wild-type confirming that LGG-1(G116A) was less efficient for selective cargo degradation.

Finally, the interactions between cargoes and autophagosomes were studied in lgg-1(G116A) mutant and wild-type embryos by analyzing the colocalization between SEPA-1 and LGG-1 or LGG-2. In wild-type embryos, immunofluorescence analyses showed the presence of LGG-1 and LGG-2 autophagosomes in contact with SEPA-1 aggregates (Figure 7L and N). In lgg-1(G116A) embryos, LGG-2 positive autophagosomes were observed but no LGG-1 dots, in line with the absence of lipidation (Figure 7M and O). A part of LGG-2 puncta was present close to SEPA-1 aggregates, however, they were less numerous and the overlap between LGG-2 and SEPA-1 signals was weaker (Figure 7P). These data suggested that LGG-1(G116A) was able to maintain the function of LGG-1 for initiation and extension of autophagosomes but was partially deficient for cargo sequestering.

Altogether, the analyses of LGG-1(G116A) indicate that many of the functions of LGG-1 in autophagy can be achieved by the cleaved, non-lipidated form I. However, the lipidation of LGG-1 appears to be important for the coordination between cargo recognition and autophagosome biogenesis and for the correct degradation of the autophagosome.

Nematode strains were grown on nematode growth media [for 500 ml H2O: 1.5 g NaCl (Sigma-Aldrich, 60142), 1.5 g bactopeptone (Becton-Dickinson, 211677), 0.5 ml cholesterol (Sigma-Aldrich, C8667) 5 mg/ml, 10 g bacto agar (Becton-Dickinson, 214010) supplemented with 500 µl CaCl2 (Sigma-Aldrich, C3306) 1 M, 500 µl MgSO4 (Sigma-Aldrich, M5921) 1 M, 10 ml KH2PO4 (Sigma-Aldrich, P5655) 1 M, 1650 µl K2HPO4 (Sigma-Aldrich, 60356) 1 M] and fed with Escherichia coli strain OP50.

A CRISPR-Cas9 approach optimized for C. elegans was used, based on a dpy-10 co-CRISPR protocol (Paix et al., 2015). All reagents are in 5 mM Tris-HCl pH 7.5. Crispr.mit.edu and CRISPOR (http://crispor.org) web tools were used to choose a Cas9 cleavage site (NGG) close to the edit site, the best sequence of the crRNAs (50 to 75% of GC content), and for off-target prediction. 1 µL of CrRNA(s) (8 µg/µL or 0.6 nmole/µL) and repair template(s) (1 µg/µL) designed for lgg-1 and dpy-10 genes were mixed with 4.1 µL of S. pyrogenes Cas9 (20 pmole/µL) and 5 µL of universal tracrRNA (4 µg/µL 4 µg/µL or 0.17 nmol/µL molarity) in 0.75 µL Hepes (200 mM) 0.5 µL KCl (1 M) and water up to 20 µL. The mix was heated for 10 min at 37 °C and injected in the gonad of young adult hermaphrodites. Progenies of injected animals were cloned and genotyped by PCR. Mutants were outcrossed three times and lgg-1 gene was sequenced to check for the specific mutations.

For starvation experiments, adult hermaphrodites were bleached to obtain synchronized L1 larvae. L1 were incubated in 0.5 mL sterilized M9 at 20 °C on spinning wheel. At each time point, an aliquot from each sample tube was placed on a plate seeded with E. coli OP50. The number of worms surviving to adulthood was counted 2 or 4 days after. Life span was performed on more than 100 animals for each genotype with independent duplicates and analyzes using Kaplan-Meier method and Log-Rank (Mantel-Cox) test.

RNAi by feeding was performed as described (Kamath et al., 2003). Fourth‐larval stage (L4) animals or embryos were raised onto 1 mM isopropyl‐D‐β‐thiogalactopyranoside (IPTG)‐containing nematode growth media (NGM) plates seeded with bacteria (E. coli HT115[DE3]) carrying the empty vector L4440 (pPD129.36) as a control or the bacterial clones from the J. Ahringer library, Open Biosystem.

The worms were collected after centrifugation in M9 and then mixed with the lysis buffer described previously (Springhorn and Hoppe, 2019) (25 mM tris-HCl, pH7.6; 150 mM NaCl; 1 mM ethylenediaminetetraacetic acid (EDTA) 1% Triton X-100; 1% sodium deoxycholate (w/v); 1% SDS (w/v)) containing glass beads (Sigma-Aldrich 425–600 µm G8772100G). They were then denatured using Precellys 24 machine at 6000 rpm with incubation for about 5 min twice to cool down the sample. The protein extracts are then centrifuged at 15,000 rpm and separated on a NuPAGE 4%‐12% Bis‐Tris gel (Life Technologies, NP0321BOX). The non‐specific sites are then blocked after the incubation for one hour with PBS Tween 0.1% (Tris Base NaCl, Tween20) BSA 2%. Blots were probed with anti‐LGG‐1 (1:3000 rabbit Ab#3 Springhorn and Hoppe, 2019 or 1:200 Ab#1 Al Rawi et al., 2011), anti‐LGG‐2 (1:200 rabbit), anti‐Tubulin (1:1000 mouse; Sigma, 078K4763), anti-SEL-1 (1:8000, rabbit), anti-CDC-48.1 (1:5000, rabbit) and revealed using HRP‐conjugated antibodies (1: 5000 promega W401B and 1:10,000 promega W4021) and the Super Signal Pico Chemiluminescent Substrate (Thermo Fisher Scientific, 34579). Signals were revealed on a Las3000 photoimager (Fuji) and quantified with Image Lab software. For cellular fractionation, 4000 age-synchronized worms (day 1 of adulthood) were collected from NGM/OP-50 plates, washed three times with M9 buffer and transferred to NGM plates without OP-50 to induce starvation. Worms were starved at 20 °C for 7 hr, and then transferred to fractionation buffer (50 mM Tris-HCl pH 7.4, 150 mM NaCl, 1 mM DTT, 1 mM PMSF, and protease inhibitor cocktail). For cell lysis, worms were homogenized 50 times using a Dounce homogenizer and sonicated for 20 s at 60% amplitude. Cell lysates were centrifuged at 500 RCF and 4 °C for 5 min to remove cell debris and the nuclear fraction. The supernatant was centrifuged again at 20,000 RCF and 4 °C for 90 min to separate soluble (cytosolic) and insoluble (membrane) fractions. Supernatant and pellet were separated and the pellet was resuspended in 150 μL of fractionation buffer. From this step, input samples were prepared for Western blot analysis. Subsequently, 30 μL of the pellet sample was mixed with 3 μL each of fractionation buffer, 3 μL 5 M NaCl, and 3 μL Triton X-100. Treated pellet samples were incubated on ice for 1 hour and then centrifuged at 20,000 RCF and 4 °C for 60 min. The resulting supernatants and pellets were again separated and analyzed by Western blotting.

Fifty adult hermaphrodites were cut to release the early embryos on a previously poly-L-lysinated slide (0.1%). Late embryos were deposited using a flattened platinum wire and bacteria as glue. Embryos were prepared for immunofluorescence staining by freeze-fracture and methanol fixation 30 min at –20 °C, incubated 40 min in 0.5% Tween, 3% BSA, PBS solution, and washed twice 30 min in 0.5% Tween PBS solution. Incubation overnight at 4 °C overnight with the primary antibodies anti-LGG-1(rabbit 1:100) anti-GABARAP (rabbit 1:200) (1: 200), anti-LGG-2 (rabbit 1:100) was followed by two washes, 2 hr incubation at room temperature with the secondary antibodies, Alexa488 and Alexa 568 (1: 1000), and two washes. Embryos were mounted in DABCO and imaged on an AxioImagerM2 microscope (Zeiss) equipped with Nomarski optics, coupled to a camera (AxioCam506mono) or a confocal Leica TCS SP8 microscope with serial z sections of 0.5–1 µm. Images were analyzed, quantified and processed using ImageJ or Fiji softwares.

For live imaging samples were mounted on a 2% agarose pad and larvae were immobilized by 40 mM sodium azide. For MitoTracker staining, adult worms were transferred to NGM agar plates containing 3.7 µM of Red CMXRos (Molecular Probes, Invitrogen) and incubated for overnight in the dark.

One-day adults were transferred to M9 20% BSA (Sigma‐Aldrich, A7030) on 1% phosphatidylcholine (Sigma-Aldrich) pre-coated 200 µm deep flat carriers (Leica Biosystems), followed by cryo‐immobilization in the EMPACT‐2 HPF apparatus (Leica Microsystems; Vienna Austria) as described (Jenzer et al., 2019). Cryo‐substitution was performed using an Automated Freeze‐substitution System (AFS2) with integrated binocular lens, and incubating chamber (Leica Microsystems; Wetzlar, Germany) with acetone. Blocks were infiltrated with 100% EPON, and embedded in fresh EPON (Agar Scientific, R1165). Ultrathin sections of 80 nm were cut on an ultramicrotome (Leica Microsystems, EM UC7) and collected on a formvar and carbon‐coated copper slot grid (LFG, FCF‐2010‐CU‐50). Sections were contrasted with 0,05% Oolong tea extract (OTE) for 30 min and 0.08 M lead citrate (Sigma‐Aldrich, 15326) for 8 min. Sections were observed with a Jeol 1400 TEM at 120 kV and images acquired with a Gatan 11 Mpixels SC1000 Orius CCD camera.

One mg of total proteins from C. elegans lysate were incubated on ice 10 min in 800 µL of TUBE lysis buffer [50 mM sodium fluoride, 5 mM tetra-sodium pyrophosphate, 10 mM β-glyceropyrophosphate, 1% Igepal CA-630, 2 mM EDTA, 20 mM Na₂HPO₄, 20 mM NaH₂PO₄, and 1.2 mg/ml complete protease inhibitor cocktail (Roche, Basel, Switzerland)] supplemented with 200 µg of purified LC3 traps or GST control (Quinet et al., 2022). After cold centrifugation at 16,200 g for 30 min, supernatant was harvested and added to 400 µl of prewashed glutathione-agarose beads (Sigma), and incubated for 6 hr rotating at 4 °C. Beads were centrifugated at 1000 g for 5 min at 4 °C (Beckman Coulter Microfuge 22 R, Fullerton, CA, USA), washed five times using 10 column volumes of PBS-tween 0.05%. Elution was done in 100 µL of (Tris pH7.5, 150 mM NaCl, 1% Triton, 1% SDS) at 95 °C during 10 min, and supernatant was harvested.

Protein samples affinity purification were prepared using the single-pot, solid-phase-enhanced sample-preparation (SP3) approach as described (Hughes et al., 2019). Samples were mixed with 10 µl of 10 µg/µl solution of Sera-Mag SpeedBeadsTM hydrophilic and hydrophobic magnetics beads (GE healthcare, ref 45152105050250 and 65152105050250) with a bead to sample ratio of 10:1. After a binding step in 50% ethanol in water, and three successive washes with 80% ethanol in water, sample were digested with 100 µl of a 5 ng/µl sequencing grade modified trypsin solution (PROMEGA). Fifty µl of Trypsin-generated peptides were vacuum dried, resuspended in 10 µl of loading buffer (2% acetonitrile and 0.05% Trifluoroacetic acid in water) and analyzed by nanoLC-MSMS using a nanoElute liquid chromatography system (Bruker) coupled to a timsTOF Pro mass spectrometer (Bruker). Briefly, peptides were loaded on an Aurora analytical column (ION OPTIK, 25cm x75µm, C18, 1.6 µm) and eluted with a gradient of 0–35% of solvent B for 100 min. Solvent A was 0.1% formic acid and 2% acetonitrile in water, and solvent B was 99.9% acetonitrile with 0.1% formic acid. MS and MS/MS spectra were recorded and converted into mgf files. Proteins identification were performed with Mascot search engine (Matrix science, London, UK) against a database composed of all LGG-1 sequences including the wild-type and mutant sequences. Database searches were performed using semi-trypsin cleavage specificity with five possible miscleavages. Methionine oxidation was set as variable modification. Peptide and fragment tolerances were set at 15 ppm and 0.05 Da, respectively. A peptide mascot score threshold of 13 was set for peptide identification. C-terminal peptides were further validated manually.

All experiments were done at least three times. All data summarization and statistical analyses were performed by using either the GraphPad-Prism or the R software (https://www.r-project.org/). The Shapiro-Wilk’s test was used to evaluate the normal distribution of the values and the Hartley Fmax test for similar variance analysis. Data derived from different genetic backgrounds were compared by Student t test, Anova, Kruskal-Wallis or Wilcoxon-Mann-Whitney tests. The Fisher’s exact test was used for nominal variables. Longevity was assessed using Log-Rank (Mantel-Cox) test. Error bars are standard deviations and boxplot representations indicate the minimum and maximum, the first (Q1/25th percentile), median (Q2/50th percentile) and the third (Q3/75th percentile) quartiles. NS (Not Significant) p>0.05; * 0.05>p > 0.01, **0.01>p > 0.001, *** 0.001>p > 0.0001 and **** p<0.0001. Exact values of n and statistical tests used can be found in the figure legends.