Phagolysosomes break down the membrane of a non-apoptotic corpse independent of macroautophagy

Cell corpses must be cleared in an efficient manner to maintain tissue homeostasis and regulate immune responses. Ubiquitin-like Atg8/LC3 family proteins promote the degradation of membranes and internal cargo during both macroautophagy and corpse clearance, raising the question how macroautophagy contributes to corpse clearance. Studying the clearance of non-apoptotic dying polar bodies in Caenorhabditis elegans embryos, we show that the LC3 ortholog LGG-2 is enriched in the polar body phagolysosome independent of membrane association or autophagosome formation. We demonstrate that ATG-16.1 and ATG-16.2, which promote membrane association of lipidated Atg8/LC3 proteins, redundantly promote polar body membrane breakdown in phagolysosomes independent of their role in macroautophagy. We also show that the lipid scramblase ATG-9 is needed for autophagosome formation in early embryos but is dispensable for timely polar body membrane breakdown or protein cargo degradation. These findings demonstrate that macroautophagy is not required to promote polar body degradation, in contrast to recent findings with apoptotic corpse clearance in C. elegans embryos. Determining how membrane association of Atg8/LC3 promotes the breakdown of different types of cell corpses in distinct cell types or metabolic states is likely to give insights into the mechanisms of immunoregulation during normal development, physiology, and disease.


Introduction
Throughout an organism's life, cells die and must be degraded in a contained manner.
Phagocytosis is an essential process by which dying cells, cellular debris, and pathogens are cleared from the extracellular space and degraded inside phagolysosomes.Disrupting engulfment and intracellular degradation of dying cells can cause inflammation and lead to the development of autoimmune diseases (1).For the contents of dying cells to be degraded, their plasma membrane must be broken down inside the phagolysosome, however how this is regulated remains unclear.
Previous work has implicated lipidation of the ubiquitin-like Atg8/LC3 family proteins in promoting phagolysosomal degradation of cell corpse membranes and their cargo (2).Depleting ATG-7, part of the Atg8/LC3 lipid conjugation machinery in C. elegans, delayed breakdown of the membrane of a non-apoptotic cell corpse and delayed disappearance of its contents in phagolysosomes, suggesting that lipidation of Atg8/LC3 family proteins promotes corpse degradation.Atg8/LC3 family proteins are better known for their roles during macroautophagy, when intracellular cargos are enclosed in a double membrane for degradation (3).Lipidation of Atg8/LC3 family proteins promotes breakdown of the inner membrane of autophagosomes during macroautophagy in mammalian cells (4), suggesting that lipidated Atg8/LC3 proteins have a conserved role in promoting the breakdown of internal membranes inside phagolysosomes and autolysosomes.
Several studies have found that when autophagosome formation or maturation was disrupted in C. elegans embryos, degradation of apoptotic corpses was significantly delayed within phagolysosomes (5)(6)(7).Recently, it was shown that autophagosome fusion with phagosomes promotes apoptotic corpse clearance mid-embryogenesis (8).Autophagosome puncta of fluorescent Atg8/LC3 family proteins LGG-1 and LGG-2, as well as a fluorescent transmembrane autophagy protein ATG-9, were observed next to phagosomes containing apoptotic cargo.As the bright LGG-1, LGG-2, and ATG-9 puncta disappeared, the markers were observed inside phagolysosomes, consistent with autophagosome fusion with the phagolysosome to transfer the internal membrane of the autophagosome into the phagolysosome.In addition, these autophagosome reporters were dispersed in the phagolysosome lumen, indicating prior breakdown of the corpse plasma membrane within the phagolysosome (8).These results suggest that autophagosome fusion contributes to the degradation of apoptotic cargo in phagolysosomes.
LGG-1 and LGG-2 reporters were also observed inside a non-apoptotic phagolysosome after corpse membrane breakdown (2), raising the question whether autophagosomes play a role in the early embryonic clearance of the second polar body by undifferentiated blastomeres.Polar bodies are born during meiosis and show TUNEL labeling independent of apoptotic regulators (9), as well as a loss of membrane integrity commonly associated with necrosis (2).C. elegans polar bodies are well-suited to study the dynamics of phagocytosis and cargo resolution, given the stereotyped timing of internalization by large embryonic blastomeres and membrane breakdown inside the phagolysosome (Fig 1).However, polar body membrane breakdown was not delayed when the ATG14 ortholog EPG-8 was deleted (2).ATG14 family proteins are thought to act as part of the PI3Kinase complex during canonical autophagy only (10), suggesting that the localization of Atg8/LC3 family proteins inside polar body phagolysosomes was independent of autophagosomes and canonical autophagy.Given the conflicting results between the role of autophagosomes in apoptotic and non-apoptotic cell clearance during different stages of C. elegans embryogenesis, we revisited whether canonical autophagy plays a role in breakdown of the polar body membrane using proteins that act upstream in the macroautophagy pathway, specifically homologs of Atg9 or ATG16L1.The lipid scramblase Atg9 acts early in autophagosome biogenesis, during the expansion of phagophores to engulf their intracellular cargo (11).The lipid transfer protein Atg2 transfers lipids to one leaflet and Atg9 acts as a lipid channel to distribute phospholipids across the leaflets of the phagophore bilayer, enabling the growth of isolation membranes.
Apoptotic corpses accumulate in atg-9 mutants (7), and C. elegans ATG-9 was recently shown to localize inside phagolysosomes clearing apoptotic corpses after autophagosome fusion (8).However, it is unknown whether ATG-9 plays a role during the phagolysosomal breakdown of non-apoptotic polar bodies.
In this study, we confirm that ATG-9 and ATG-16.2 are required for autophagosome biogenesis in C. elegans early embryos.In the absence of autophagosomes, we find no delay in corpse membrane breakdown, no disruption in LGG-2 localization inside the polar body phagolysosome, and no delay in degradation of a corpse cargo protein.We also confirm the role of membrane association of Atg8/LC3 by demonstrating that ATG-16.1 and ATG-16.2redundantly promote polar body membrane breakdown.This suggests that autophagosomes are not required for phagolysosome maturation in early embryos, and the association of Atg8/LC3 with a non-autophagic membrane is important for the phagocytic clearance of a non-apoptotic corpse.

Worm maintenance
Caenorhabditis elegans strains (Table S1) were maintained at room temperature (22-24˚C) on nematode growth media (NGM) plates seeded with OP50 bacteria, according to standard procedures (16).C. elegans grown for semi-quantitative PCR were maintained on peptonerich plates seeded with NA22 bacteria at 18.5°C.Genotyping DNA from lysed worms was amplified using OneTaq polymerase (New England BioLabs) and the primers listed in Table S2 that were designed based on sequences in WormBase (17).
MboI (New England BioLabs) was used to detect the point mutation in atg-9(bp564) and HinfI (New England BioLabs) was used to detect the point mutations in atg-

Microscopy
Gravid hermaphrodites were dissected in egg salts (94 mM NaCl, 32 mM KCl, 2.8 mM MgCl 2 , 2.8 mM CaCl 2 , 2 mM HEPES, pH 7.5) to isolate embryos and mounted on a 4% agarose pad for live imaging.Z-stacks were collected using a Zeiss Axio Observer 7 inverted microscope with Plan-APO 40X 1.4 NA oil objective with Excelitas Technologies X-Cite 120LED Boost illumination, and Hamamatsu ORCA-Fusion sCMOS camera controlled by 3i SlideBook6 software.Time-lapse imaging was acquired sequentially for mCherry and GFP every minute.mCh::LGG-2 autophagosome count Bright mCherry puncta in the cytosol of 8-to 15-cell embryos were counted in SlideBook (3i) or using the cell counter function in FIJI (20).Larger and dimmer LGG-2-positive puncta flanking dividing nuclei appeared to be centrosomes and were not included in the autophagosome count.Large clusters of puncta were challenging to distinguish and omitted from autophagosome counts in 1-cell embryos in Fig 7C.

Qualitative colocalization analysis
Images of live 8-to 15-cell embryos were analyzed in SlideBook (3i) for whether mCh::LGG-2 levels appeared higher at the second polar body than in the neighboring cytoplasm.Polar bodies and membrane breakdown were identified using GFP::H2B.

Fluorescence intensity measurements
The mean intensity of a circle of 1.59 µm diameter (10 pixel) for polar bodies and 3.98 µm (25 pixel) for AB or P1 blastomere nuclei was measured in the mCh::LGG-2 channel on the polar body or nuclei and three neighboring regions of the cytoplasm using FIJI.The results in Fig 5G , 6C-D, and 7F are reported after subtraction of the average cytoplasm mean from the polar body or blastomere nuclei mean.

Polar body internalization
Internalization was determined using the GFP::PH membrane marker as the timepoint in which the second polar body was fully engulfed by a blastomere, as previously described (21).Cell stages were identified as the beginning of each furrow ingression.

Polar body membrane breakdown
Polar body membrane breakdown was scored indirectly using mCherry histone markers, based on established protocol (21).Membrane breakdown was the first timepoint in which the histones dispersed to fill the entire polar body phagosome.

Polar body clearance
Polar bodies were tracked in time-lapse series using the ZF1::mCh::H2B marker.Clearance was determined as the first timepoint in which the largest polar body fragment was indistinguishable from background noise.

Semi-quantitative RT-PCR analysis
Frozen worm pellets were vortexed with glass beads (Sigma-Aldrich) and Trizol (Zymo Research).RNA purification was performed using the Direct-zol RNA Miniprep Plus kit (Zymo Research) and reverse transcription was performed using RNA to cDNA EcoDry Premix (TaKaRa).100ng cDNA was amplified through a 30-cycle PCR protocol using OneTaq polymerase (New England BioLabs) and the primers indicated in Table S2.SYBR Safe fluorescence (Invitrogen) was photographed using an iPhone 13 (Apple).The green fluorescence intensity was measured for each band and on empty regions of the gel using

Results and Discussion
To test whether ATG-9 and the ATG-16 paralogs are required for autophagosome biogenesis in early C. elegans embryos, we crossed a reporter for the LC3 homolog mCh::LGG-2 into atg-9, atg-16.1, or atg-16.2mutants with premature stop codons (Fig S1) (22)(23)(24).In control embryos, we observed an average of 27±15 mCh::LGG-2 puncta between  (25,26).These results confirm that ATG-9 is required for autophagosome formation in early embryos and that ATG-16.2 is required for LGG-2 localization to autophagosomes.As the atg-16.2nonsense mutants had a premature stop codon in their WD40 domain (Fig S1 ), we had predicted that this mutation would only disrupt non-canonical autophagy, similar to findings in mice (13,14).However, we observed a loss of the autophagosomes that clear sperm components in the early embryo (28,29).As premature stop codons can lead to nonsense-mediated decay (30), we asked whether atg-16.2 mRNA levels were altered by the premature stop codon in atg- 16  To examine whether the rare remaining autophagosome puncta in atg-9 and atg-16.2mutants were capable of fusing with the polar body phagolysosome, we crossed a GFP::H2B reporter into the mutant strains so that we could observe the second polar body chromosomes after phagocytosis.In the subset of mutant embryos with one or more mCh::LGG-2 puncta, puncta were not observed in the same cell as the second polar body in 92% of atg-9 (n=41) or 88% of atg-16.2(n=40) mutant embryos, which would preclude an autophagosome from fusing with the phagolysosome.In combination with the embryos that lacked any mCh::LGG-2 puncta, autophagosomes were only present in the same cell as the polar body in <8% of atg-9 or atg-16.2(W253*)mutant embryos, allowing us to use atg-9 and atg-16.2mutants to test the role of autophagosomes in polar body clearance.
To test whether macroautophagy contributes to phagocytic clearance of the second polar body, we first tested whether engulfment was normal in the absence of autophagosomes.A previous study had shown that LGG-1 promoted exposure of an engulfment signal on apoptotic cells, namely phosphatidylserine externalization (31), which has also been observed on dying polar bodies (2).We crossed mutants lacking autophagosomes to mCh::H2B reporters to label the polar body chromosomes and GFP::PH reporters to label the plasma membrane of the engulfing cell.In time-lapse series from control embryos, polar bodies were internalized 5±2 minutes after the 4-cell stage (Fig 1A-B,   S3), consistent with previous reports (2).We did not observe a significant delay in polar body uptake in atg-9 mutants or in atg-16.2;atg-16.1 double mutants (Fig S3).These data show that phagocytic engulfment of polar body corpses occurs without autophagosomes.
We next tested whether autophagosomes contribute to the degradation of protein cargo within the polar body, as atg-7 depletion to reduce Atg8/LC3 lipidation delayed cargo clearance by over an hour (2).We used the disappearance of an mCherry-tagged histone reporter in the polar body as a readout for protein cargo degradation (2,21).In control embryos, histone reporters disappeared 86±19 minutes after internalization (Fig 3A-B and   D), similar to previous results (2).When macroautophagy was disrupted in atg-9 mutants, the disappearance timing was not significantly different from control embryos (93±12 min, p> 0.1, Fig 3C and D).These data suggest that polar body clearance occurs independent of macroautophagy, similar to previous results with epg-8 deletion mutants (2).We next asked what happens to the timing of polar body membrane breakdown in the absence of autophagosomes, as atg-7 depletion to disrupt Atg8/LC3 lipidation delayed membrane breakdown after internalization almost two-fold (2).We used dispersal of the condensed histones within the phagolysosome as a proxy for membrane breakdown (  Open circles denote the end of a time-lapse movie when histone dispersal did not occur.Data are presented as mean ± std dev.Mutants were compared to controls using a one-tailed t-test with a Bonferroni correction for 3 comparisons: **p<0.01,***p<0.001,n.s.= not significant p>0.2. As the atg-16.2(W253*)allele only partially depletes atg-16.2mRNA levels (Fig S2 ), we wanted to confirm that a null allele of atg-16.2 also had no effect on membrane breakdown.We used the atg-16.2(ok3224)mutant in which amino acids 125 to 299 were deleted (∆) and the reading frame is shifted to disrupt both the CCD domain required for canonical autophagy and the WD40 domain required for non-canonical autophagy (Fig S1) (15).Deleting atg-16.2did not significantly delay breakdown of the polar body membrane inside the phagolysosome (p>0.2), with histone dispersal occurring 13±7 minutes after internalization (Fig 4G and K).This further confirms our findings with the nonsense allele of ATG-16.2 that polar body membrane breakdown is not promoted by macroautophagy.
We then wanted to confirm that membrane association of lipidated Atg8/LC3 promotes corpse membrane breakdown by disrupting both ATG-16.1 and ATG-16.2.In atg-16.1 atg-16.2double nonsense mutants, we found a significant delay in polar body membrane breakdown, with histone dispersal taking 17±9 minutes after internalization (Fig 4F and J, p<0.004).This almost two-fold delay is similar to previous results with atg-7 knockdown (2).
In double mutants using the atg-16.2deletion allele, we also found a significant over twofold delay in membrane breakdown (Fig 4I and K, p<0.0005), with histone dispersal occurring 23±7 minutes after internalization.However, the balanced deletion strain rapidly went sterile and was lost, limiting our data collection.These findings confirm the role of Atg8/LC3 membrane association in promoting timely corpse membrane breakdown.
In contrast to the double mutants, we did not observe a significant delay in polar body membrane breakdown in atg-16.1 single mutants (p>0.5), with histone dispersal occurring

Video S1. LGG-2 accumulates in mitotic nuclei as the nuclear envelope breaks down.
A 2-cell C. elegans embryo developing to the 4-cell stage shows the timing of LC3 reporter mCh::LGG-2 accumulation inside the nucleus, as well as to centrosomes and spindle microtubules.Mitotic stage is visible in the merged image with histone GFP::H2B reporter.A z-series was recorded every 45 seconds, six 1.5 µm z-step images were max projected, and the projections are displayed at 5 fps using Imaris.
To determine whether the LGG-2 observed within polar body phagolysosomes was from the dying or engulfing cell, we created a degron-tagged LGG-2 reporter as a selective labeling approach.Degron tagging allows for the specific labeling of proteins protected by membranes as cytosolic degron-tagged proteins will be targeted for proteasomal degradation (18).As LGG-2 is found on both the cytosolic face of the outer autophagosome membrane and the luminal face of the inner membrane of autophagosomes, we expected a loss of LGG-2 from only the cytosolic face of autophagosomes that have sealed prior to the onset of degradation.To initiate degradation in embryonic blastomeres prior to polar body uptake at the 4-cell stage, we tagged the mCh::LGG-2 reporter with the C-terminal phosphodegrons (CTPD) from OMA-1, which are phosphorylated during the first cell division in C. elegans embryos.Phosphorylation leads to the degradation of CTPD-tagged proteins during the 2and 3-cell stages (18,33), but CTPD-mediated degradation does not occur inside polar bodies (18).LGG-2 in embryonic blastomeres.These results reveal that the degron partially reduced the labeling of autophagosomes and degraded cytosolic pools of LGG-2 in the engulfing cells.

Conclusions
Together, our data suggest that phagocytic degradation of the second polar body requires the association of lipidated LC3 with membranes that are not autophagosomes.This finding is consistent with normal residual body clearance during spermatogenesis in autophagy mutant worms (6), which suggests that autophagosomes do not contribute to the phagocytic clearance of large cell fragments by gonadal sheath cells.However, our findings in early embryonic blastomeres contrast with the role of autophagosome fusion in the degradation of apoptotic corpses mid-embryogenesis (8).One possibility is that apoptotic and non-apoptotic corpses induce different clearance mechanisms.However, polar bodies and residual bodies expose phosphatidylserine and depend on the same signaling pathways as apoptotic cells for their uptake (2,6).In addition, polar body and residual body phagosomes mature using similar Rab GTPase pathways as apoptotic phagosomes, leaving it unclear whether different phagocytic cargos have different needs for autophagosomes.
Alternatively, the difference in the role of autophagy between polar body and apoptotic corpse clearance may lie in the engulfing cells.Early embryonic blastomeres and mid-embryonic differentiated cells may have different metabolic states or types of autophagosomes and autolysosomes.The early embryonic cells that engulf the polar body have been isolated in an eggshell for <2 hours, while the apoptotic corpse phagosomes fused with autophagosomes after >5 hours of isolation inside the eggshell (8).The additional time without external nutrients could lead to starvation-induced autophagy, consistent with the abundance of autophagosomes and autolysosomes at mid-embryonic stages.Indeed, ~40% of autophagosomes fusing with apoptotic phagolysosomes contained a lysosomal nuclease, indicating that autolysosomes also fuse with phagolysosomes (8).In contrast, early embryonic cells contain several cargos that are specifically targeted for autophagic degradation, including sperm components degraded by allophagy and RNA granules degraded by aggrephagy (28,29,34).Therefore, what distinguishes the autophagosome dependence of phagocytic clearance may be the metabolic state of the cell or the prevalence of lysosomes containing autophagic cargo.
Further research is needed to determine how lipidated Atg8/LC3 family proteins promote polar body clearance, including which membrane ATG-16 proteins localize lipidated Atg8/LC3 to for corpse membrane breakdown.Although Atg8/LC3 lipidation factors promote breakdown of the polar body membrane, we rarely observed LGG-1 or LGG-2 localization to the phagosome prior to membrane breakdown with fluorescent protein-tagged reporters (2).Furthermore, all LGG-1 and LGG-2 reporters examined showed colocalization of Atg8/LC3 proteins with histone markers after corpse membrane breakdown rather than the surface labeling of phagosomes predicted for LC3-associated phagocytosis (This study and (2).We also observed dynamics for mCh::LGG-2 independent of membrane association by ATG-16 proteins (Fig 5 -6), but ATG-16 proteins promote timely corpse membrane breakdown (Fig 4).One possibility is that the N-terminal fluorescent protein tags on the Atg8/LC3 reporters disrupt Atg8/LC3 association to phagosome membranes but not autophagosome membranes, consistent with ATG16L1 interacting with different proteins while localizing Atg8/LC3 proteins during autophagy or LAP/CASM (13,14).Alternatively, as Atg8/LC3 reporters increase their fluorescence in response to nuclear membrane breakdown during mitosis as well as corpse membrane breakdown inside phagolysosomes, the observed changes in fluorescence of mCh::LGG reporters could be due to interactions of soluble Atg8/LC3 proteins between mixing organelle compartments, obscuring the membrane-associated Atg8/LC3 proteins that function in corpse clearance.Thus, it appears that a low level of membrane-associated Atg8/LC3 is sufficient to promote corpse membrane breakdown inside phagolysosomes.

Fig 1 .
Fig 1. Engulfment of the second polar body by embryonic blastomeres and corpse

FIJI.
The background fluorescence was subtracted from the band fluorescence and then normalized to the average intensity of the wild-type N2 bands.Image processing Single z planes are shown, except for Fig 2A-E, where 26 z planes at 0.77 µm steps were projected (maximum intensity), Fig 3A, where 8 z planes at 1.5 µm steps were projected (maximum intensity) and Fig 6A-B, where 3 z planes at 1.5 µm steps were projected for mCh::LGG-2 (maximum intensity) using Slidebook.Images were cropped, rotated, and their brightness was adjusted in Adobe Photoshop.Statistics Statistical significance of polar body membrane breakdown timing was determined by performing a Student's one-tailed t-test with unequal variance.All significant p-values were adjusted using Bonferroni corrections for multiple comparisons.Data is represented as mean ± standard deviation.Statistical significance of percent LGG-2 colocalization with the second polar body was determined by performing a one-tailed Fisher's exact test.Data Exclusions Images in which the polar body was out of view or time lapse series where embryonic development was arrested or significantly delayed were excluded.LGG-2 images that included a polar body with condensed chromosomes (before corpse membrane breakdown) were excluded from the colocalization analysis in Fig 5G.Polar bodies internalized before the 3-cell stage were excluded from the internalization analysis in Fig S3.

Fig
Fig S1.ATG-9 and ATG-16 alleles used in this study.

Fig
Fig S2. atg-16.2 mRNA levels are reduced by a premature stop codon in the WD40

Fig 3 .
Fig 3. Macroautophagy does not promote degradation of polar body protein cargo.

Fig 6 .
Fig 6.LGG-2 accumulates in mitotic nuclei as the nuclear envelope breaks down.