The ATG8 protein family is composed of two subfamilies, the LC3 subfamily and the GABARAP subfamily, which are very close to each other in sequence; the ATG8 family is also referred to as the LC3/GABARAP family (Schaaf et al., 2016). C. elegans has two LC3/GABARAP family members, LGG-1 and LGG-2, which belong to the LC3 and GABARAP subfamilies, respectively (Figure 1A; Manil-Ségalen et al., 2014). Both LGG-1 and LGG-2 are attached to autophagosomes, except they each label autophagosomes of different maturity (Manil-Ségalen et al., 2014).

Figure 1 with 5 supplements see all

Download asset Open asset

Figure 1—source data 1

Relative mCherry::LGG-1 and mCherry::LGG-2 signal intesity over time in 1J.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig1-data1-v3.docx

Download elife-72466-fig1-data1-v3.docx

To determine whether autophagosomes interact with phagosomes that contain apoptotic cells in C. elegans embryos, we constructed the GFP-tagged LGG-1 and LGG-2 reporters that were expressed under the control of the ced-1 promoter (P_ced-1), a well-documented engulfing cell-specific promoter (Zhou et al., 2001; Lu et al., 2009). In embryos, we observed numerous GFP::LGG-1⁺ and GFP::LGG-2⁺ puncta (Figure 1D and G). Using our previously established time-lapse recording protocol (Lu et al., 2009), we observed the enrichment of GFP::LGG-1⁺ and GFP-LGG-2⁺ puncta to the surfaces of the phagosomes, including the phagosomes containing apoptotic cells C1, C2, and C3 (Figure 1B, D and G). C1, C2, and C3 are localized to the ventral surface of an embryo and undergo apoptosis at approximately 330 min post-1st cleavage (Lu et al., 2009). Apoptotic C1, C2, and C3 are each engulfed and degraded by a particular ventral hypodermal cell (Figure 1B). Inside the engulfing cell that expresses GFP::LGG-1 in the cytoplasm, the phagosomes containing C1, C2, and C3 (also referred to as phagosomes C1, C2, and C3 for convenience) appear like dark discs (Figure 1D and G). The GFP-labeled puncta were not observed inside the phagosomal lumen (Figure 1D and G, top panels).

C. elegans LGG-1 and LGG-2 were both reported to specifically attach to autophagosomes through their lipid tails (Manil-Ségalen et al., 2014; Wu et al., 2015). To verify that the observed GFP puncta are LGG-labeled lipid vesicles and not artifacts of protein aggregation, we tested two mutant constructs, GFP::LGG-1(G116A) and GFP::LGG-2(G130A), which bear mutations in the lipidation sites of the LGG proteins and are deficient for membrane targeting (Manil-Ségalen et al., 2014). We found that both GFP::LGG-1(G116A) and GFP::LGG-2(G130A) display a diffuse cytosolic localization pattern (Figure 1D, G, bottom panels), in stark contrast to the punctate pattern presented by GFP::LGG-1 and GFP::LGG-2. This result indicates that GFP::LGG-1 and::LGG-2 are membrane attached and thus label lipid vesicles.

The fluorophore within GFP is sensitive to acidic pH (pKa = 6.0) (Tsien, 1998), thus its signal diminishes when GFP is inside the acidic lumen of lysosomes and phagosomes. To further monitor the fate of the LGG⁺ puncta after they are recruited to the surfaces of phagosomes, we replaced GFP with mCherry (pKa <4.5) or mNeonGreen (mNG) (pKa = 5.1), both of which are more resistant than GFP to the acidic pH environment (Shaner et al., 2004; Shinoda et al., 2018), allowing fluorescence signal inside the phagosomal lumen to be detected. Co-expressed with the mCherry::LGG or mNG::LGG reporter are the CED-1::GFP or PH(hPLCγ)::mRFP reporters, which were used in the time-lapse experiments as markers for the extending pseudopods, allowing us to determine the moment when the pseudopods sealed up and a nascent phagosome was born (Figure 1E, F, H, I, top panels) (Zhou et al., 2001; Shen et al., 2013). In time-lapse image series of the clearance process of C1, C2, and C3, we found that like the GFP::LGG reporters, the mCherry::LGG and mNG::LGG reporters were enriched on the surfaces of phagosomes; moreover, unlike the GFP::LGG reporters, the mCherry and mNG-tagged reporters subsequently entered phagosomal lumen (Figure 1E, F, H, I). The fluorescence signal intensity increases over time with the continuous recruitment of the LGG⁺ puncta on phagosomal surfaces (Figure 1E, F, H, I, Figure 1—videos 1–3). We measured the intensity of the mCherry::LGG-1 and mCherry::LGG-2 signal in the center of the C3 phagosomes over time from the start of the phagosome (Materials and methods) (Figure 1C) and observed over 10-fold increases of the signal intensity within 60 min (Figure 1J). The appearance of the membrane-attached mCherry::LGG and mNG::LGG signal in the phagosomal lumen indicates that the LGG⁺ vesicles fuse with the phagosomal membrane (Figure 1K). They further suggest that these vesicles are composed of double membranes labeled with the LGG reporter molecules on both the outer and inner membranes (Figure 1K). If these LGG⁺ vesicles were single-membrane, no membrane-attached LGG reporter would end up inside the phagosomal lumen because, as a result of fusion, the reporter molecules will be retained on the phagosomal membrane (Figure 1K). Indeed, when CTNS-1, a lysosomal transmembrane protein, is tagged with mRFP, which is acid-resistant (pKa = 4.5 [Shaner et al., 2004]), on its C-terminus, the lysosome-phagosome fusion event resulted in the incorporation of the mRFP signal into the phagosomal membrane; as a result, the CTNS-1::mRFP signal is only observed on the phagosomal surface, not in the phagosomal lumen (Figure 1—figure supplement 1; Yu et al., 2008).

The observation that the LGG⁺ vesicles incorporated into phagosomes are likely double-membrane vesicles reminded us of canonical autophagosomes rather than LAP vesicles. In support of this model, previously, using electron microscopy and immune-gold staining, Manil-Ségalen et al have shown that in C. elegans embryos, LGG-1 or LGG-2-labeled vesicles are double-membrane autophagosomes (Manil-Ségalen et al., 2014). To further determine whether these vesicles represent canonical autophagosomes, we examined whether loss-of-function mutations of atg-7, atg-13, and epg-8, which are defective for the biogenesis of autophagosomes (Tian et al., 2010; Yang and Zhang, 2011), impaired the production of these vesicles. C. elegans atg-13 and epg-8 encode homologs of mammalian ATG13 and ATG14, respectively, which are essential for the biogenesis of autophagosomes but dispensable for that of LAP vesicles (Introduction) (Zhang and Baehrecke, 2015). atg-7 encodes a homolog of mammalian ATG7, a protein essential for conjugating a phospholipid tail onto the LC3 family proteins and thus for the biogenesis of both autophagosomes and LAP vesicles (Nakatogawa, 2020; Zhang and Baehrecke, 2015). We first scored whether the mCherry::LGG-1/–2 reporters were observed in the center of 15 phagosomes during the phagosome maturation process (Figures 1C and 2), which is indicative of the fusion of double-membrane mCherry::LGG⁺ vesicles to phagosomes. In wild-type embryos, the steady entry of mCherry into phagosomes over time (Figure 1E and H) results in the increase of the average mCherry::LGG-1 and -LGG-2 intensities to 9.2- and 8.4-fold of that at 0 min time point, respectively, at 60 min after the formation of a phagosome (Figure 2H and J). However, in atg-7(bp411) (Gomes et al., 2016) mutant embryos, hardly any mCherry signal was observed inside phagosomes (Figure 2A, D and G–J). At 60 min after the phagosome formation, the average mCherry signal intensities were merely 1.2 and 1.4-fold at 0 min time point (Figure 2H and J). These observations indicate a lack of LGG⁺ vesicles that fuse to phagosomes. In atg-13(bp414) and epg-8(bp251) mutant embryos, similar observations were made except that the defects were slightly weaker (Figure 2B, C, E, F and G–J).

Figure 2 with 2 supplements see all

Download asset Open asset

Figure 2—source data 1

Relative mCherry::LGG-1 and mCherry::LGG-2 singnal intesity over time in Figure 2G and I.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig2-data1-v3.docx

Download elife-72466-fig2-data1-v3.docx

Figure 2—source data 2

Relative mCherry::LGG-1 and mCherry::LGG-2 signal intensity at 60min-post engulfment in Figure 2H and J.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig2-data2-v3.docx

Download elife-72466-fig2-data2-v3.docx

We next examined whether the mCherry::LGG-1/-2⁺ vesicles are produced in cells of the atg-7, atg-9, atg-13, and epg-8 mutant embryos. In wild-type embryos at mid-embryonic developmental stages (~330, ~ 350, ~ 400 min post-1st embryonic cell division), numerous mCherry⁺ puncta were observed (Figure 2—figure supplements 1A and 2A). In the atg-7 mutant embryos, rarely any such puncta existed (Figure 2—figure supplements 1B and 2B), consistent with a previous report (Tian et al., 2010). atg-9 encodes ATG-9, the only transmembrane protein in the core autophagy machinery (Tian et al., 2010; Zhang and Baehrecke, 2015). ATG9 plays an essential role in the expansion of phagophore and the biogenesis of autophagosomes (Nakatogawa, 2020). In atg-9(bp564) mutant embryos that expressed mNG::LGG-1 or::LGG-2, much fewer mNG⁺ puncta that might represent autophagosomes were observed, and the mNG reporters label large aggregates (Figure 3—figure supplement 1), consistent with a previous report (Lin et al., 2013). In atg-13 and epg-8 mutant embryos, the numbers of mCherry-LGG⁺ puncta were also significantly reduced (Figure 2—figure supplement 1C-D and Figure 2—figure supplement 2C-D), suggesting that the atg-13(bp414) and epn-8(bp251) mutations severely impaired the biogenesis of autophagosomes, a phenotype that is consistent with previous reports (Tian et al., 2010; Yang and Zhang, 2011). The drastic reduction in the number of LGG-1/–2-labeled puncta in atg-7, atg-9, atg-13, and epg-8 loss-of-function mutants strongly indicates that these puncta belong to canonical autophagosomes.

To further confirm that autophagosomes fuse to phagosomes, we examined the subcellular localization of a mCherry::ATG-9 reporter expressed in engulfing cells. Besides the LGG proteins, ATG-9 is also an established autophagosome marker as it is the only transmembrane protein in autophagosomes (Lu et al., 2011b). In embryonic hypodermal cells that co-expressed mCherry::ATG-9 and either mNG::LGG-1 or::LGG-2, puncta labeled with mCherry are recruited to the surfaces of phagosomes (Figure 3). Furthermore, the mCherry signal gradually accumulates in the phagosomal lumen like the LGG reporters (Figure 3). During the phagosome maturation process, the co-localization between mCherry::ATG-9 and each of the two LGG reporters on the puncta on phagosomal surfaces and inside phagosomal lumen was nearly perfect (Figure 3). Together, the above observations verified that the ATG-9 and LGG double-positive autophagosomes are recruited to phagosomal surfaces and subsequently fuse to phagosomes containing apoptotic cells.

Figure 3 with 1 supplement see all

Download asset Open asset

Besides the C1, C2, and C3 phagosomes, in embryos co-expressing mCherry::LGG-1 or –2 and CED-1::GFP reporters, we also observed the recruitment of the mCherry signal to other phagosomes and the subsequent entry of the mCherry signal to the phagosomal lumen (Figure 1—figure supplement 2A-E). Time-lapse image series (Figure 1—figure supplement 2A-E) demonstrate the recruitment and fusion of LGG⁺ puncta to apoptotic cells C4 and C5 in the tail (Shen et al., 2013). In mid-stage (1.5-fold) embryos, multiple phagosomes are observed to either have LGG⁺ puncta attached to the surface (Figure 1—figure supplement 2 F(b, d), yellow arrows) or LGG⁺ signal inside the lumen (Figure 1—figure supplement 2 F(b,d), white open arrows). None of these phagosomes are C1, C2, C3, C4, or C5, judging by their positions in the embryo. These observations indicate that the fusion between autophagosomes and phagosomes is a general phenomenon during embryogenesis.

In each C. elegans adult hermaphrodite gonad, germ cells that die of apoptosis undergo cellularization and are engulfed by neighboring gonadal sheath cells (Gumienny et al., 1999). We used transmission electron microscopy (TEM) (Materials and methods) to examine whether there were any double-membrane vesicles attaching to phagosomes containing germ cells. In wild-type adults, due to the swift engulfment and degradation activities, on average only 2–4 germ cell corpses are observed in each gonadal arm in an adult hermaphrodite 48 hr post mid-L4 larval stage (Yu et al., 2008), making it very difficult to find any phagosomes in the 50 nm thin sections in TEM. We thus chose to analyze phagosomal surfaces in the rab-7 mutants, in which many germ cell corpses persist in the gonad due to the blockage of phagosome degradation (Yu et al., 2008). In TEM images, germ cell corpses are identified by their higher density than live germ cells and complete cellularization (Yu et al., 2008). The rab-7(ok511) mutants are maternal-effect embryonic lethal. rab-7(ok511) adult hermaphrodites are viable and produce dead embryos (Yu et al., 2008). In rab-7(ok511) mutant adult gonads, many undegraded phagosomes containing germ cell corpses are observed inside the sheath cell in thin (50 nm) TEM sections (Figure 4B and C), as reported previously (Yu et al., 2008). Furthermore, in the rab-7 mutant gonad, mCherry::LGG-1 expressed in sheath cells (under P_ced-1) is observed to accumulate on the surfaces of phagosomes (Figure 4A), like in rab-7 mutant embryos (see Results section ‘The small GTPase RAB-7 and the HOPS complex are essential for the fusion between autophagosomes and phagosomes’ below). This result is consistent with the observation of LGG⁺ puncta on phagosomal surfaces in embryos. In rab-7 mutant gonad, we did identify multiple double-membrane vesicles in close contact with the phagosomal surfaces in TEM thin sections (Figure 4D–H). Besides being composed of double membranes, these vesicles also resemble autophagosomes identified in C. elegans cells (Kovacs et al., 2013; Zhang et al., 2015). These vesicles vary in diameter from 200 nm to 800 nm, consistent with that of autophagosomes reported in the literature (Kovacs et al., 2013; Manil-Ségalen et al., 2014; Zhang et al., 2015 ). The above observations support our conclusion that the LGG⁺ puncta that are recruited to phagosomal surfaces are indeed canonical autophagosomes. They further suggest that the autophagosome-phagosome interaction is general rather than a cell-specific phenomenon.

Figure 4

Download asset Open asset

To examine whether the incorporation of autophagosomes into phagosomes affects the clearance of the engulfed apoptotic cells, we first quantified whether, in mutants of genes essential for the biogenesis of autophagosomes, apoptotic cells were un-degraded and thus persisted in embryos. In addition to the atg-7, atg-9, atg-13, and epg-8 mutants characterized above, we also characterized loss-of-function mutants of lgg-1 and lgg-2, and of atg-3, whose gene product is essential for the conjugation of a phosphatidylethanolamine (PE) tail to the LC3 family proteins (Tian et al., 2010; Wu et al., 2016), of atg-2 and atg-18, whose gene products function together with ATG-9 in the expansion of phagophore (Lu et al., 2011b), and of unc-51, which encodes a C. elegans homolog of ULK1, an autophagic protein kinase (Lu et al., 2011b). In twofold stage wild-type embryos, which are ~460 min post the first embryonic cell division, an average of 11.3 cell corpses were scored (Figure 5A). The mutant twofold stage embryos examined bore 46.9–91.2% more cell corpses (Figure 5), indicating that the clearance of cell corpses is defective. Together, the findings reported in Figures 2—5 indicate that autophagosomes made a substantial contribution to the clearance of cell corpses.

Figure 5 with 1 supplement see all

Download asset Open asset

Figure 5—source data 1

Cell corpse count and statistical analysis of atg mutants at 2-fold stage.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig5-data1-v3.docx

Download elife-72466-fig5-data1-v3.docx

To determine whether the lack of autophagosomes impairs the engulfment or degradation of cell corpses, we monitored the formation and degradation of phagosomes containing apoptotic cells C1, C2, and C3 (Figure 1B) in wild-type, atg-7, and lgg mutant embryos in real-time using an established protocol (Materials and methods) (Lu et al., 2011a; Lu et al., 2009). The CED-1::GFP expressed in engulfing cells labels the extending pseudopods and enables us to monitor the process of phagosome formation, starting from the budding and ending at the sealing of the pseudopods (Zhou et al., 2001; Yu et al., 2006). 2xFYVE::mRFP (also expressed under P_ced-1), a reporter for phagosomal surface PtdIns(3)P, enables us to monitor the shrinking of a phagosome, an indication of phagosome degradation (Yu et al., 2008; Lu et al., 2012; Lu et al., 2011a). In atg-7, lgg-1, and lgg-2 mutant strains that co-expressed CED-1::GFP and 2xFYVE::mRFP, we found that engulfment was completed in 4–8 min, just like in wild-type embryos (Figure 6A–D and F), indicating that defects in autophagosomes biogenesis do not affect the engulfment of cell corpses. However, the lifespans of phagosomes (Materials and methods) were much longer in atg-7, lgg-1, and lgg-2 mutants than in wild-type embryos (Figure 6A–D and G). All wild-type phagosomes have a lifespan between 40–60 min (Figure 6G). In atg-7, lgg-1, and lgg-2 mutants, the lifespan varied in a much more extensive range. Remarkably, 43.8%, 82.4%, and 53.3% of phagosomes in atg-7, lgg-1, and lgg-2 mutant embryos, respectively, have lifespans longer than 60 min (Figure 6G). These observations indicate that autophagosomes made a specific contribution to the efficient degradation of phagosomal contents.

Figure 6

Download asset Open asset

Figure 6—source data 1

Time of engulfment of C3 phagosomes.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig6-data1-v3.docx

Download elife-72466-fig6-data1-v3.docx

Figure 6—source data 2

Time of degradation of C3 cell corpses.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig6-data2-v3.docx

Download elife-72466-fig6-data2-v3.docx

LGG-1 and LGG-2 have distinct structural features and were observed to attach to different sub-populations of autophagosomes, which represent autophagosomes at different maturation stages (Manil-Ségalen et al., 2014; Wu et al., 2015). The sub-populations of autophagosomes labeled by LGG-1 or LGG-2 are incorporated into phagosomes (Figure 1). lgg-1 and lgg-2 single mutants are both inefficient in clearing cell corpses (Figure 5B). Furthermore, in lgg-1; lgg-2 double mutant embryos at the twofold stage, the Ced phenotype is further enhanced significantly (Figure 5B) -- the number of cell corpses is 22.0% and 27.1% larger than in the lgg-1 and lgg-2 single mutants, respectively. This enhanced phenotype indicates an additive effect and suggests that the autophagosomes labeled with LGG-1 or LGG-2 play parallel and partially redundant roles in promoting phagosome degradation. In embryos co-expressing GFP::LGG-1 and mCherry::LGG-2, we observed that both in the cytoplasm of the engulfing cells and on the surfaces of phagosomes, puncta that were either labeled with GFP or mCherry alone or with both GFP and mCherry (Figure 7A and B). In hypodermal cells, the average distribution of LGG-1⁺ LGG-2^-, LGG-1^- LGG-2⁺, and LGG-1⁺ LGG-2⁺ puncta is 35.8%, 40.7%, and 23.5%, respectively (Figure 7C). These observations indicate that in addition to the LGG-1⁺-only and LGG-2⁺-only sub-populations, a third, LGG-1⁺ LGG-2⁺ double-positive sub-population of autophagosomes exists. This sub-population likely corresponds to an intermediate stage in the maturation path of autophagosomes.

Figure 7

Download asset Open asset

Figure 7—source data 1

Percentage distribution of single and double-labeled puncta in the LGG-labeled population in the engulfing cells for C1, C2, and C3.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig7-data1-v3.docx

Download elife-72466-fig7-data1-v3.docx

To determine whether LGG-1 and LGG-2 act in engulfing cells to facilitate phagosome degradation, we examined whether the specific expression of each gene in engulfing cells, under the control of P_ced-1, would rescue the Ced phenotype of the corresponding mutant embryos. We tested the lgg cDNAs that are tagged with either gfp or mCherry for the rescuing activity by counting the number of cell corpses in 1.5-fold stage transgenic embryos. In the lgg-1 and lgg-2 null mutants, both the gfp- and mCherry-tagged corresponding lgg cDNA efficiently rescued the Ced phenotype (Figure 5—figure supplement 1). The gfp::lgg-1 and mCherry::lgg-1 transgenes lowered the number of cell corpses from on-average 152% of wild-type level observed in the lgg-1(tm3489) mutants to 118% and 123% of the wild-type level, respectively. Similarly, the gfp::lgg-2 and mCherry::lgg-2 transgenes lowered the number of cell corpses from on-average 143% of wild-type level observed in the lgg-2(tm5755) mutants to 109% and 112% of the wild-type level, respectively (Figure 5—figure supplement 1). These results indicate that lgg-1 and lgg-2 primarily act in engulfing cells to facilitate the clearance of apoptotic cells.

The small GTPase Rab7 is well known to specifically label late endosomes and lysosomes (Stenmark, 2009). In addition, Rab7 proteins in mammalian and C. elegans are also recruited from the cytoplasm to the phagosomal membrane shortly after the formation of a phagosome and mediate the fusion between the maturing phagosome and lysosomes and do that through their effector, the HOPs complex (Levin et al., 2016; Lu and Zhou, 2012). Furthermore, in yeast, Drosophila, and mammalian cells, Rab7 is directly recruited to the surfaces of autophagosomes and plays an important role in the fusion between autophagosomes and lysosomes (Szatmári and Sass, 2014; Gao et al., 2018; Hegedus et al., 2016; Vaites et al., 2018). C. elegans RAB-7 also plays an essential role in the fusion between autophagosomes and lysosomes (Manil-Ségalen et al., 2014). To examine whether RAB-7 is localized to LGG-1⁺ and LGG-2⁺ autophagosomes that fuse to C. elegans phagosomes, we generated two transgenic C. elegans strains that co-expressed the mCherry::LGG-1/ GFP::RAB-7 or mCherry::LGG-2/ GFP::RAB-7 pairs of reporters (Materials and methods). We observed that GFP::RAB-7 was localized to some but not all of the LGG-1⁺ or LGG-2⁺ puncta (Figure 8B–D). The LGG⁺ RAB-7⁺ double-positive autophagosomes were observed both freely distributed in the cytoplasm of ventral hypodermal cells (Figure 8B (a-c) and C) and on phagosomal surfaces (Figure 8B (d-f), D). In addition, Figure 8B (d-f) depicts that GFP::RAB-7 is both evenly distributed to the surface of a phagosome (d) as previously reported (Yu et al., 2008), and highly enriched on LGG-1⁺ autophagosomes that are recruited to the phagosomal surface (d-f, white arrows). Figure 8D shows a time-lapse series of a dynamic fusion event of an LGG-2⁺/RAB-7⁺ punctum (marked by the bottom arrow in the ‘0 min’ time point) to the phagosome: this punctum is first seen attached to the phagosome membrane and subsequently becoming part of the phagosomal surface at the ‘ + 6 min’ time point. Quantitative analysis of the green and red puncta distribution reveals that on average, 66.2% and 63.5% of LGG-1⁺ and LGG-2⁺ puncta observed in the cytoplasm of the engulfing cells are RAB-7⁺, respectively (Figure 8E). The LGG⁺ but RAB-7^- puncta (Figure 8B and C, white arrowheads) are likely immature autophagosomes not acquired RAB-7 yet. We also observed puncta that were LGG^- but RAB-7⁺ (Figure 8B and C, yellow arrows). These puncta represent intracellular organelles such as late endosomes or lysosomes. Different from the observation made in the cytoplasm, 100% of the autophagosomes that were observed on phagosomal surfaces were RAB-7⁺ (Figure 8B, D and F). This observation suggests that RAB-7 might play an essential role in the interaction between autophagosomes and phagosomes.

Figure 8 with 2 supplements see all

Download asset Open asset

Figure 8—source data 1

Percentage of distribution of LGG-1+ or LGG-2+ puncta that are also RAB-7+.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig8-data1-v3.docx

Download elife-72466-fig8-data1-v3.docx

Figure 8—source data 2

The time it takes for LGG-1 or LGG-2 autophagosomes fuse to with phagosomes.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig8-data2-v3.docx

Download elife-72466-fig8-data2-v3.docx

We examined whether the recruitment and fusion of mCherry::LGG-1⁺- or LGG-2⁺- autophagosomes to the C1, C2, and C3 phagosomes were normal in rab-7(ok511) null mutant embryos (Yu et al., 2008). The recruitment event can be evaluated by the level of enrichment of mCherry puncta on the surfaces of phagosomes, whereas the subsequent fusion event can be measured by the level of accumulation of the mCherry signal inside the phagosomal lumen. In rab-7(ok511)(m^-z^-) mutant embryos produced by rab-7(ok511)(m⁺z^-) mothers, robust enrichment of both the mCherry::LGG-1⁺ and::LGG-2⁺ puncta to phagosomal surfaces are prominent (Figure 8G–H), indicating that the recruitment of autophagosomes was normal. However, no mCherry signal was observed entering the phagosomal lumen (Figure 8G–H, Figure 8—videos 1; 2). We quantified the time between the moments a phagosome was just born and that when an obvious mCherry signal was observed inside the phagosomal lumen. In the wild-type embryos, this time is <30 min in 100% of the samples analyzed, and the median value is 14 and 18 min for LGG-1⁺ and LGG-2⁺ autophagosomes, respectively (Figure 1(E and H) and 8(I-J)). In stark contrast, in rab-7 mutants, the mCherry signal was not detected in the lumen inside any of the phagosomes for LGG-1 or LGG-2 reporters over a time span of 0–60 min after phagosome formation (Figure 8I–J). For over half of the samples, the observation period was extended beyond 90 min after phagosome formation, and still, no mCherry signal was observed in the phagosomal lumen within this period. These results strongly indicate that the function of RAB-7 is essential for autophagosome-phagosome fusion but not required for the recruitment of autophagosomes to phagosomal surfaces.

The HOPS complex acts as an effector for Rab7 (Balderhaar and Ungermann, 2013). In C. elegans, a null mutation in vps-18, which encodes a subunit of the HOPS complex, impairs phagosome maturation (Xiao et al., 2009). The lack of autophagosome-phagosome fusion observed in rab-7 mutants led us to subsequently examine the vps-18 null mutant embryos. We found that, like in rab-7 mutants, the recruitment of autophagosomes to phagosomes appeared normal in vps-18 mutants (Figure 8G–H). Also, like in rab-7 mutants, in vps-18 mutants, the fusion of LGG-1⁺ and LGG-2⁺ autophagosomes to phagosomes was severely defective (Figure 8G–J). However, whereas in rab-7 mutants, the accumulation of the mCherry signal in the phagosomal lumen was blocked completely in all samples, in some vps-18 mutant embryos, the entry of mCherry signal still occurred, albeit severely delayed (Figure 8I–J). In contrast, in other samples, the entry was blocked (Figure 8I–J), indicating a fusion defect that is less severe than that caused by the rab-7 null mutation and suggesting the existence of the residual HOPS function in vps-18 mutants. Together, our observations indicate that RAB-7 and the HOPS complex play a critical and specific role in driving the fusion between autophagosomes and phagosomes.

To evaluate the defect in phagosome degradation caused by the rab-7(ok511) null mutation, which abolishes the fusion of both lysosomes and autophagosomes to phagosomes, we measured the phagosome lifespan in rab-7(ok511)(m^-z^-) embryos using the phagosome lifespan analysis used in Figure 6A–D. Compared to that observed in the atg-7, lgg-1, and lgg-2 mutant embryos, the phagosome lifespan is significantly longer in rab-7 mutants. In rab-7 mutants, 100% of phagosomes last longer than 60 min, among which 71.4% last longer than 80 min and 57% last longer than 90 min (Figure 6E and G). The difference in severity in the phagosome degradation defects displayed by rab-7 and autophagy mutants suggest that autophagosomes and lysosomes both contribute to phagosome degradation, and that autophagosomes provide a phagosome degradation activity in an additive and possibly independent manner to lysosomes.

The signaling pathway led by the phagocytic receptor CED-1 plays essential roles in initiating the maturation of phagosomes containing apoptotic cells (Yu et al., 2008). The CED-1 pathway is known to drive the incorporation of early endosomes and lysosomes to phagosomes (Yu et al., 2008; Yu et al., 2006). Here, we further examined whether the loss-of-function mutations in members of the CED-1 pathway affect the incorporation of autophagosomes to phagosomes. CED-1 promotes the recognition, engulfment, and degradation of apoptotic cells (Yu et al., 2008). In ced-1(e1735) null mutant embryos, although the recognition and engulfment are delayed or blocked due to the loss of CED-1’s engulfment activity, the majority of the cell corpses are eventually engulfed inside phagosomes (Yu et al., 2008). This is why in ced-1 mutants, there are phagosomes that contain C1, C2, or C3 available for analysis of phagosome degradation. The partially penetrant recognition and engulfment defects observed in the ced-1 mutants are due to the compensation of the activities by the ced-5/–10/–12 and the rab-35 pathways (Reddien and Horvitz, 2004; Yu et al., 2006; Haley et al., 2018). In ced-1(e1735) mutants expressing the mNG::LGG-1 or –2 reporters, we observed severe defects in the incorporation of autophagosomes into phagosomes. First of all, only a very dim mNG signal was observed inside the phagosomal lumen 50 min post phagosome formation, a time point well past the observed initiation time for autophagosomes/phagosome fusion in the wild-type condition (Figure 9A, C, F and G, Figure 9—video 1). Whereas in wild-type embryos, the median relative LGG-1 and –2 signal intensities are 6.3 and 5.3 at 50 min-post phagosome formation, respectively, in ced-1 mutant embryos, the median values are merely 2.1 and 1.6, respectively (Figure 9K–L). Secondly, unlike in rab-7 mutants, where autophagosomes were observed accumulating on phagosomal surfaces (Figure 8G–J), in ced-1 mutants, very few LGG-1- or LGG-2-labeled puncta were observed on phagosomal surfaces (Figure 9A and C).

Figure 9 with 2 supplements see all

Download asset Open asset

Figure 9—source data 1

Singal intesity over time of mNG::LGG-1 and mNG::LGG-1 in Figure 9F, G.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig9-data1-v3.docx

Download elife-72466-fig9-data1-v3.docx

Figure 9—source data 2

Recruitment of mNG::LGG-1 and mNG::LGG-2 to the surface of phagosomes in Figure 9I, J.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig9-data2-v3.docx

Download elife-72466-fig9-data2-v3.docx

Figure 9—source data 3

Relative mNG::LGG-1 and mNG::LGG-2 signal intensity at 50min-post engulfment.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig9-data3-v3.docx

Download elife-72466-fig9-data3-v3.docx

Further quantitative measurement of the samples presented in (Figure 9A and C) and 14 additional samples for each of the reporters confirmed that the mNG signal was not enriched on the surfaces of phagosomes, in contrast to wild-type embryos (Figure 9H–L). These results indicate a severe defect in the recruitment of autophagosomes to phagosomes. Unlike in the atg mutants that we have examined (Figure 2—figure supplements 1–2), in ced-1 mutant embryos, normal numbers of LGG-1⁺ or LGG-2⁺ puncta were observed (Figure 9—figure supplement 1), indicating that the biogenesis of autophagosomes is normal. Thus the recruitment defect observed in ced-1 mutants is not a consequence of the lack of autophagosomes; rather, it is likely a result of a defect in signaling between phagosomes and autophagosomes.

We further examined whether CED-6 and DYN-1, two other members of the CED-1 pathway, were also needed for the incorporation of autophagosomes into phagosomes. In the ced-6(n2095) and dyn-1(n4039) loss-of-function mutant embryos, the median relative LGG-1 and LGG-2 signal intensities in the center of phagosomes are much lower than that in wild-type samples, respectively, at 50 min-post phagosome formation mutants Figure 9B,D,F,G, K–L and Figure 10A, D E,F,I,J, although the defects are not as severe as in ced-1 mutants Figure 9A, C, F, G, K and L. Further observation discovered that the LGG-1 or LGG-2-labeled puncta were rarely observed on the surfaces of phagosomes in these mutants Figure 9B,D1,J and Figure 10A, F, demonstrating severe defects in the recruitment of autophagosomes to phagosomal surfaces.

Figure 10

Download asset Open asset

Figure 10—source data 1

Relative signal intesity over time mCherry::LGG-1 and mCherry::LGG-2 in Figure 10 D and I.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig10-data1-v3.docx

Download elife-72466-fig10-data1-v3.docx

Figure 10—source data 2

Relative mCherry::LGG-1 and mCherry::LGG-2 signal intensity 50min-post engulfment in Figure 10 E and F.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig10-data2-v3.docx

Download elife-72466-fig10-data2-v3.docx

In ced-1, ced-6, and dyn-1 mutants, the fusion between autophagosomes and phagosomes might also be defective. However, the severe recruitment defects resulted in the lack of LGG-labeled puncta on phagosomal surfaces, making it difficult to evaluate whether there are additional fusion defects and how severe the fusion defects are.

We also examined whether the pathway composed of the small GTPase CED-10 and its bipartite Guanine Nucleotide Exchange Factor (GEF) CED-5 and CED-12, which acts parallel to the CED-1 pathway in the engulfment of cell corpses, plays any role in promoting the incorporation of autophagosomes into phagosomes. In the ced-5(n1812) null mutant and ced-10(n1993) loss-of-function mutant embryos, due to the presence of the parallel engulfment pathways (the CED-1 and RAB-35 pathways), some C1, C2, or C3 phagosomes eventually form, although they usually suffer a severe delay, allowing the analysis of phagosome degradation (Yu et al., 2008; Yu et al., 2006; Haley et al., 2018). In these two mutants, the accumulation of the LGG-1 and LGG-2 signals on the surfaces of the phagosomes and the subsequent accumulation of signals inside phagosomal lumen were normal both in the time course and in the levels of signal enrichment (Figure 10B, C–E, G and H–J), indicating that both the recruitment and fusion of autophagosomes to phagosomes are normal. We thus conclude that, unlike the CED-1/–6 /DYN-1 pathway, the CED-5/–10 pathway is not involved in regulating the incorporation of autophagosomes to phagosomes (Figure 10K).

The incorporation of lysosomes into phagosomes is an essential force that drives the degradation of phagosomal contents (Levin et al., 2016). As a portion of the autophagosomes would fuse with lysosomes and become autolysosomes that retain lysosomal features (Morishita and Mizushima, 2019), the autolysosomes might contribute to phagosome degradation by depositing lysosomal luminal proteins to the phagosomal lumen. To test whether this is the case, we examined whether mutations that specifically impair autophagosome biogenesis would reduce the amount or speed of the incorporation of organelles with lysosomal features into phagosomes. C. elegans NUC-1 is an endonuclease belonging to the DNase II family and resides in the lysosomal lumen (Wu et al., 2000; Guo et al., 2010). Using a NUC-1::mCherry reporter expressed in engulfing cells as a lysosomal luminal marker, we quantified the level and rate of lysosome-phagosome fusion over time. In wild-type embryos, in phagosomes containing cell corpses C1, C2, and C3, we observed first the attachment of NUC-1::mCherry puncta on the surfaces of phagosomes and subsequently the accumulation of the mCherry signal inside the phagosomal lumen (Figure 11A). This dynamic process represents the recruitment and the subsequent fusion of lysosomal particles to phagosomes, which result in the delivery of NUC-1::mCherry into the phagosomal lumen. We next monitored NUC-1::mCherry inside the phagosomal lumen in lgg-1(tm3489), lgg-2(tm5755), and atg-7(bp411) embryos, which are defective in the biogenesis of autophagosomes. In these mutants, numerous mCherry::NUC-1 puncta are enriched on the surfaces of phagosomes C1, C2, and C3 and are subsequently fused to phagosomes (Figure 11B–D). As a result, the accumulation of the mCherry signal is observed in the phagosomal lumen in all samples like in wild-type embryos (Figure 11B–E). Quantitative analysis of 15 phagosomes for each genotype found that 60 min after phagosome formation, the average folds of increase of the luminal mCherry signal over the ‘0 min’ time point in lgg-1, lgg-2, and atg-7 mutants are not significantly different from that observed from wild-type samples (Figure 11F). In addition, the median value of the first time point when NUC-1::mCherry signal was detected inside the phagosomal lumen was not significantly different in all four genotypes (Figure 11G). These data strongly suggest that the incorporation of lysosomes into phagosomes in lgg-1, lgg-2, and atg-7 mutants is as efficient as in wild-type embryos, at least within the detection range of our assay. Therefore, defects in the biogenesis of autophagosomes, which potentially would indirectly result in the lack of autolysosomes, do not appear to significantly affect the incorporation of lysosomes into phagosomes.

Figure 11

Download asset Open asset

Figure 11—source data 1

NUC-1::mCherry signal intensity over time in Figure 11E.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig11-data1-v3.docx

Download elife-72466-fig11-data1-v3.docx

Figure 11—source data 2

Time of fusion of NUC-1::mCherry and relative signal intesity at 60 min-post in Figure 11 F and G.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig11-data2-v3.docx

Download elife-72466-fig11-data2-v3.docx

In the wild-type embryos that co-express the mNG::LGG-1 or –2 reporters with the NUC-1::mCherry reporter in engulfing cells, we observed that, in addition to the lysosomal particles that were mNG^- but mCherry⁺, there were two kinds of mNG⁺ puncta that were recruited to phagosomal surfaces and subsequently fused to phagosomes. Those puncta that are not labeled with mCherry represent autophagosomes that are not fused to lysosomes; on the other hand, the mNG⁺ mCherry⁺ puncta represent autolysosomes (Figure 12A–B). Quantification of these two kinds of puncta on the surfaces of the C1, C2, and C3 phagosomes reveals that 40.6% and 36.5% of LGG-1⁺ and –2⁺ puncta are autolysosomes, respectively (Figure 12C). Whereas the LGG⁺ NUC-1⁺ autolysosomes are capable of contributing both autophagosomal and lysosomal materials to phagosomes, the LGG⁺ NUC-1^- population of autophagosomes are likely to deliver certain autophagosome-specific material to phagosomes.

Figure 12

Download asset Open asset

Figure 12—source data 1

Percentage distribution percentage of autolysosomes (GFP+ mCherry+ puncta) among LGG-1+ or -2+ puncta on the surface of C1, C2, and C3 phagosomes.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig12-data1-v3.docx

Download elife-72466-fig12-data1-v3.docx

Acidification of the phagosomal lumen is a critical event for the degradation of phagosomal content. To examine whether defects in autophagosome biogenesis would affect the acidification of phagosomes, we developed an acidification reporter for the phagosomal lumen. This reporter, P_his-72 his-72::gfp::mCherry, expresses a HIS-72 (histone H3.3)::GFP::mCherry fusion protein in all cells (Ooi et al., 2006), including cells that undergo programmed cell death. We quantified the GFP/mCherry signal ratio in the phagosomal lumen over time and normalized it over the GFP/mCherry signal ratio at the ‘0 min’ time point when a nascent phagosome just formed. We refer to this normalized value as an ‘acidification index’. In a typical phagosome in a wild-type embryo, the acidification index reduces continuously from 1.0 (t = 0 min post engulfment) to 0.4 (t = 51 min) (Figure 13A and D). In addition, in 100% of the phagosomes (n = 6), the acidification index values at the 51 min time point are <0.6, among which in 35% of the samples, the index values are <0.4 (Figure 13E). On the contrary, in cup-5(n3264) mutant embryos, the acidification of the phagosomal lumen is severely defective (Figure 13B, D and E). The acidification index values of all the phagosomes (n = 13) at t = 51 min are >0.6, and in 53.8% of the phagosomes, the index values are >0.9, indicating a minimal reduction of the GFP signal. cup-5 encodes a lysosomal TRP channel homologous to human Mucolipin IV (Treusch et al., 2004; Campbell and Fares, 2010). CUP-5 and its mammalian homologs play conserved and essential roles in supporting lysosomal biogenesis and functions (Treusch et al., 2004; Campbell and Fares, 2010). The severe acidification defect observed here emphasizes the essential role of lysosomes for phagosome degradation. Compared to the cup-5 mutant embryos, in atg-7(bp411) null mutant embryos, we observed a moderate phagosome acidification defect (Figure 13C, D and E). The average acidification index values for wild-type, atg-7(bp411), and cup-5(n3264) mutants are 0.443, 0.679, and 0.912, respectively, at t = 51 min. In atg-7(bp411) mutants, the acidification index values in 69.2% of the phagosomes at t = 51 min are >0.6, yet the values of only 15.3% phagosomes are >0.9. This moderate defect is distinct from that observed from the cup-5 mutants.

Figure 13

Download asset Open asset

Figure 13—source data 1

The acidification index curves of phagosomes over time in Figure 13D.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig13-data1-v3.docx

Download elife-72466-fig13-data1-v3.docx

Figure 13—source data 2

Distribution of the acidification index measured at 51-min-post the formation of phagosomes in Figure 13E.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig13-data2-v3.docx

Download elife-72466-fig13-data2-v3.docx

The overall phagosome degradation efficiency in wild-type and autophagy mutant embryos has been measured by the reduction of phagosomal size over time (Figure 6). To further evaluate the hydrolytic activities inside the phagosomal lumen in the autophagy mutant embryos, we measured the rate of the degradation of chromatin DNA in engulfed apoptotic cells. We found that, unlike a previous report that concluded that the NUC-1 endonuclease acts in cells undergoing programmed cell death to generate chromatin DNA fragments (Wu et al., 2000), NUC-1 acts in engulfing cells, in the phagosomal lumen to degrade chromatin DNA of engulfed apoptotic cells (Pickett J., Auld N., Lucas, L., Pena-Ramos O., and Zhou, Z., unpublished results). During the above study, we established HIS-72::mCherry as a reporter for chromatin DNA. Histone b H3.3, together with other histones and the DNA wrapping around them, form the nucleosome, the fundamental subunit of the chromatin. In Z-section images, inside the phagosomal lumen, the HIS-72::mCherry signal appears as a condensed red disc, outlining the nucleus of the engulfed apoptotic cell. We observed that in the wild-type embryos, the degradation of the apoptotic cell-chromatin DNA inside the phagosomal lumen allowed the reduction of the size of the mCherry⁺ disc over time (Figure 14A and D). Within 60 min after the formation of a phagosome, the average diameter of the mCherry⁺ disc reduces to 49% of the 0 min value (Figure 14E). On the contrary, in the nuc-1 mutant embryos, the diameter of the mCherry⁺ disc remains the same over the entire period of time-lapse observation (> 80 min) (Pickett J., Auld N., Lucas, L., Pena-Ramos O., and Zhou, Z., unpublished observation), indicating the lack of degradation of chromatin DNA when NUC-1 activity is absent.

Figure 14

Download asset Open asset

Figure 14—source data 1

The ratio of the nuclear diameter curves of phagosomes over time in Figure 14D.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig14-data1-v3.docx

Download elife-72466-fig14-data1-v3.docx

Figure 14—source data 2

Distribution of the ratio of nuclear diameters measured at 60min-post phagosome formation in Figure 14E.

https://cdn.elifesciences.org/articles/72466/elife-72466-fig14-data2-v3.docx

Download elife-72466-fig14-data2-v3.docx

We observed that, as predicted for a mutant severely defective in lysosomal biogenesis and functions, in cup-5(n3264) mutant embryos, within 60 min post phagosome formation, the average diameter of the mCherry⁺ disc remains at 89% of the 0 min-value (Figure 14B and E). In comparison, in atg-7(bp411) mutant embryos, that value is reduced to 66% of the 0 min-value; in addition, the diameters of the mCherry⁺ discs in 25% of the phagosomes are >80% of the 0 min-value. These results indicate that in atg-7 mutants, there is a defect in the digestion of chromatin DNA, yet this defect is relatively moderate.

Previously, C. elegans autophagy genes have been implicated in facilitating the clearance of apoptotic cells. Autophagy occurring in apoptotic cells was proposed to facilitate the exposure of phosphatidylserine, which is the ‘eat me’ signal that attracts engulfing cells to the surfaces of apoptotic cells in mice embryonic bodies and C. elegans (Qu et al., 2007; Jenzer et al., 2019). Multiple autophagy genes were reported to participate in the clearance of apoptotic cells in C. elegans engulfing cells (Cheng et al., 2013; Huang et al., 2013; Li et al., 2012; Jenzer et al., 2019). They were reported to work together with Class II PtdIns3 kinase PIKI-1 to facilitate the production of PtdIns(3)P on phagosomal surfaces (Cheng et al., 2013), to promote the recruitment of RAB-5 and RAB-7 to the phagosomal surfaces (Li et al., 2012), and to promote phagosome degradation (Jenzer et al., 2019). All the above findings suggest that the autophagy genes examined might have separate functions in the degradation of phagosomal contents unrelated to their canonical roles in autophagosome biogenesis. Our work reported here, on the other hand, presents a different discovery regarding how the autophagy machinery regulates phagosome degradation.

We found that the canonical double-membrane autophagosomes actively participate in the degradation of apoptotic cells inside phagosomes. In the time-lapse recording series, we first observed that vesicles labeled with LGG-1/or –2 reporters are recruited to phagosomal surfaces; subsequently, the LGG-tagged reporters enter the phagosomal lumen. We propose that the entry of the LGG reporters is a result of the fusion of the outer membrane of a double-membrane vesicle with the membrane of a phagosome based on two lines of evidence. First, the fusion between these two membranes would allow the release of the inner vesicle into the phagosomal lumen. Moreover, the RAB-7 GTPase and its effector—the HOPS complex, are well-known membrane tethering factors that promote the fusion between two intracellular membranes (Stenmark, 2009; Balderhaar and Ungermann, 2013). Their essential roles in supporting the entry of the LGG reporters into the phagosomal lumen that we have discovered strongly indicate that this event is a result of the fusion between the outer membrane of autophagosomes and phagosomes. Formally, other possible explanations of the entry of the LGG reporters exist, such as that phagosomes engulf the LGG-labeled vesicles. However, currently, there are no reported findings that phagosomes would engulf intracellular vesicles.

Using the immuno-electron microscopy technique, Manil-Segalen et al previously found that the LGG-1 and –2 are specifically enriched on double-membrane autophagosomes in C. elegans embryos (Manil-Ségalen et al., 2014). Five critical sets of evidence we report here further indicate that the LGG-1- and –2-labeled vesicles that are incorporated into phagosomes are double-membrane autophagosomes, not the single-membrane LAP vesicles. First, only when the vesicles are composed of double membranes and when the LGG-1/ or –2 reporter molecules label both the inside and outside membranes, the entry of the reporters into phagosomal lumen is possible (Figure 1K). Conversely, if the LGG-1 or –2 labeled vesicles are of a single membrane, after vesicle-phagosome fusion, the reporter signal would remain on the phagosomal membrane, just like the membrane-bound lysosomal marker CTNS-1::mRFP (Figure 1—figure supplement 1). Second, these vesicles are also labeled with GFP-tagged ATG-9, an integral membrane protein inserted into the autophagosomal membranes. Third, we found that besides in embryos, in the gonad of adult hermaphrodites, many LGG-1-labeled vesicles are also attached to the surfaces of phagosomes that contain germ cell corpses in the rab-7 mutant background in which membrane fusion is presumably blocked. Also, in the rab-7 mutant adult hermaphrodites, thin-section transmission electron micrographs reveal multiple double-membrane vesicles that are in close contact with the phagosomal membranes. These vesicles resemble autophagosomes in morphology (Kovacs et al., 2013; Zhang et al., 2015). Fourth, the production of these vesicles relies on genes pivotal for autophagosome biogenesis, such as atg-3, atg-7, and atg-9. Last but not least, genes encoding the C. elegans homologs of mammalian ULK1, ATG13, and ATG14, which are dispensable for the generation of LAP vesicles in mammalian cells (Green et al., 2016), are required for the formation of the LGG-labeled vesicles discussed here and for the efficient clearance of apoptotic cells, further verifying that these vesicles are distinct from the single-membrane LAP vesicles. Together, these five sets of findings demonstrate that double-membrane autophagosomes, not single-membrane LAP vesicles, are being incorporated into phagosomes.

In C. elegans, LC3-dependent phagocytosis was reported to function in the clearance of the midbody, a structure that is essential for the completion of cytokinesis, and the second polar body generated during female meiosis (Fazeli et al., 2018; Fazeli et al., 2016). The clearance of midbodies and second polar body is considered independent of autophagosomes based on the observation that neither unc-51 nor epg-8 is required for the above processes (Fazeli et al., 2018; Fazeli et al., 2016). However, whether the LGG markers attached to phagosomes carrying midbodies or polar bodies represent double-membrane or single-membrane vesicles have not been investigated (Fazeli et al., 2018; Fazeli et al., 2016). Whether LAP vesicles exist in C. elegans cells still needs to be demonstrated. On the other hand, we found that in atg-13 mutant embryos, in which the biogenesis of the LAP vesicles is not supposed to be affected, the LGG-labeled puncta largely disappeared, indicating that these puncta are primarily autophagosomes instead of LAP vesicles (Figure 2—figure supplement 1 –2). This observation does not support the existence of a substantial population of the LAP vesicles in C. elegans embryos, at least not in the cells in which the LGG reporters are expressed.

The observation of the LGG-1 or –2 signals inside the phagosomal lumen is made using the mCherry reporter. The mCherry protein is resistant to acidic pH (Shaner et al., 2004). Phagosome luminal pH value reduces from 5.5 to 6.0 to 4.5–5.5 after incorporating lysosomes (Vieira et al., 2002), leading to the inactivation of the fluorophore in the commonly used GFP reporter (Tsien, 1998). In fact, in C. elegans engulfing cells, GFP-tagged LGG-1 or –2, unlike the mCherry-tagged reporters, are only observed on the surfaces but not in the lumen of phagosomes, presumably due to the acidic environment of the phagosomal lumen (Figure 1). It is thus conceivable that in other experimental systems, a mCherry-tagged LC3 marker might disclose a previously overlooked fusion between autophagosomes and phagosomes.

Based on the evolutionary conservation of almost all molecular mechanisms, in particular the strong conservation of the known mechanisms that control phagocytosis and autophagy in metazoans (Lu and Zhou, 2012; Zhang and Baehrecke, 2015), we propose that in mammalian cells, besides LAP vesicles, canonical autophagosomes might also contribute to the degradation of apoptotic cells in phagosomes. The demonstration of an autophagosome-phagosome fusion event in C. elegans, a well-established model organism for studying cell-death-related events, opens a path to investigate whether the event and mechanism of autophagosome-phagosome fusion are evolutionarily conserved. In addition, the autophagosome-phagosome fusion might contribute to the degradation of other kinds of phagosomal cargos besides apoptotic cells.

LGG-1 and –2 are close homologs that each labels autophagosomes. We have observed three separate classes of autophagosomes that are incorporated into phagosomes: LGG-1⁺ LGG-2^-, LGG-1^- LGG-2⁺, or LGG-1⁺ LGG-2⁺. These subpopulations might represent autophagosomes at different stages of maturation (Manil-Ségalen et al., 2014). Remarkably, we have found that the lgg-1; lgg-2 double mutants display a further enhanced Ced phenotype than that displayed by each of the lgg-1 or lgg-2 single mutants. Together, these results indicate that autophagosomes at different stages of maturity all contribute to the degradation of apoptotic cells. These results are also consistent with our finding that the autophagosome-phagosome fusion is primarily independent of lysosome-phagosome fusion (Figure 11) (Also see the next section).

We have found that the blockage of the biogenesis of autophagosomes or the recruitment or fusion of autophagosomes to phagosomes results in a significant delay in the degradation of phagosomal content and, consequently, the persistent appearance of engulfed cell corpses. Among other possible mechanisms, the incorporation of autophagosomes into phagosomes presumably delivers certain substances to the lumen and/or membrane of phagosomes, substances that are important for the degradation of phagosomal contents. Here, we discuss several possible candidates and mechanisms supporting the roles of autophagosomes in phagosome degradation.

We have observed that the LGG⁺ puncta inside engulfing cells are a collection of two subpopulations: autophagosomes that are already fused to lysosomes and become autolysosomes (LGG⁺ NUC-1⁺) and autophagosomes that are not fused to lysosomes (LGG⁺ NUC-1^-) (Figure 12). The autolysosome populations, which occupied 40.7% and 36.5% of the LGG-1⁺ and LGG-2⁺ puncta observed on phagosomal surfaces, respectively, might contribute both autosomal and lysosomal materials to the phagosomes. In atg-7 mutants, in which the biogenesis of autophagosomes is virtually blocked, the acidification of the phagosomal lumen and the degradation of the nuclear DNA of engulfed apoptotic cells are modestly defective (Figures 13 and 14). These defects might be partially attribute to the lack of autolysosomes that are incorporated into phagosomes. However, in cells that maintain a stable lysosome population, the lack of autophagosomes to fuse with lysosomes is not expected to influence the overall activities of lysosomes. In atg-7, lgg-1, and lgg-2 mutant embryos, we found that the lack of autophagosomes does not significantly affect the efficiency of lysosome-phagosome fusion (Figure 11). This result could be due to that in an engulfing cell, the population of autolysosomes, comparing to that of lysosomes, is a rather minor population. It supports the notion that the autophagy pathway primarily regulates activities that are additive to the lysosomal activities for promoting phagosome degradation, although the assay might not be sensitive enough to detect a modest decrease. Moreover, in rab-7(m^-z^-) mutant embryos in which the fusion of both the lysosomes and the LGG⁺ particles to phagosomes are blocked, the phagosome degradation defect is much more severe than in autophagosome biogenesis mutants. These results suggest that the collected population of LGG⁺ vesicles contribute to phagosome degradation in a manner in parallel to lysosomes. In addition to the autolysosomal subpopulation, the LGG-1⁺ NUC-1^- (59.3% of the LGG⁺ puncta) or LGG-2⁺ NUC-1^- (63.5% of the LGG2⁺ puncta) subpopulations that fuse to phagosomes are likely to contribute certain unique, lysosome-independent activity(ies) to the phagosomes.

What could be the unique activities or mechanisms provided by the autophagosomes? Through fusion to phagosomes and releasing the inner vesicles to the phagosomal lumen, autophagosomes might deliver certain protein or lipid molecules to facilitate the degradation of phagosomal cargos. For example, the outer membrane of the autophagosomes might provide important signaling molecules to the phagosomal membrane. One such candidate is PtdIns(3)P, a membrane signaling molecule essential for many membrane trafficking events. On the phagosomal surfaces, PtdIns(3)P recruits PtdIns(3)P-binding proteins, including the sorting nexins SNX-1, SNX-6, and LST-4/SNX-9 and the HOPs complex, which subsequently drive multiple membrane remodeling events that promote phagosome maturation (Lu et al., 2011a; Lu and Zhou, 2012). In C. elegans, PtdIns(3)P is presented on phagosomal membrane over time in a two-peak pattern (Lu et al., 2012). Autophagosomes are coated with PtdIns(3)P (Nakatogawa, 2020). Judging by the timing of autophagosome-phagosome fusion, autophagosomal membranes are likely to contribute PtdIns(3)P for the rise of the second peak of PtdIns(3)P.

Alternatively, the autophagosome-phagosome fusion incorporates membrane materials to the phagosomal membrane and thus should increase the size of phagosomal membrane. Increasing the overall amount of phagosomal membrane might aid the degradation of the apoptotic cell by facilitating the recruitment of lysosomes to the phagosomal surfaces. We previously reported that phagosomes extend transient lipid tubules to capture lysosomal particles in the cytoplasm and bring these particles back to the phagosomes for fusion (Yu et al., 2008; Lu et al., 2011a). The increased phagosomal membrane material might support the extension of these lipid tubules. A substantial body of future investigation is required to discover the mechanisms employed by autophagosomes that facilitate phagosome degradation.

Autophagosomes are incorporated into phagosomes in two sequential steps: (1) they are recruited to the surfaces of phagosomes, detected by the enrichment of punctated LGG-1 or –2 fluorescence reporters on phagosomal surfaces, and (2) they subsequently fuse to phagosomes, detected by the enrichment of the LGG-1 or –2 reporter signals inside the phagosomal lumen. RAB-7 and its effector, the HOPS complex, are known to act as tethering factors that facilitate the fusion of various intracellular organelles, including autophagosomes, to lysosomes (Manil-Ségalen et al., 2014; Spang, 2016). Closely related to this study, RAB-7 and VPS-18 are pivotal for the fusion between lysosomes and phagosomes and for the degradation of apoptotic cells inside phagosomes in C. elegans (Yu et al., 2008; Xiao et al., 2009; Kinchen et al., 2008). Here, we observed that 100% of the LGG⁺ puncta attached to phagosomal surfaces were also labeled with RAB-7. We further discovered that RAB-7 and VPS-18 played essential and specific roles in the fusion but not for the recruitment of autophagosomes to phagosomes. Our finding adds a new pair of organelles that depend on the RAB-7/HOPS complex for fusion to each other. As RAB-7 is enriched on the surfaces of both phagosomes and autophagosomes (Yu et al., 2008 and Figure 8A-D), and as the HOPS complexes in mammalian cells and Drosophila are known to interact with the SNARE complex, the membrane fusion machinery (Jiang et al., 2014; Takáts et al., 2014), we propose that the RAB-7/HOPS complex acts on the surfaces of phagosomes and autophagosomes to facilitate autophagosome-phagosome fusion via promoting the interaction between the SNARE complexes on phagosomes and autophagosomes (Figure 8K). Currently, the specific SNARE proteins that catalyze the autophagosome-phagosome fusion have not been identified.

The CED-1 signaling pathway, which initiates the maturation of phagosomes that bear apoptotic cells, is essential for the enrichment of GTP-bound RAB-7 to the surfaces of phagosomes (Yu et al., 2008). Here we have found that CED-1, CED-6, and DYN-1 drive the incorporation of autophagosomes to phagosomes, in addition to driving the incorporation of early endosomes and lysosomes previously discovered (Yu et al., 2008; Yu et al., 2006). In ced-1 mutants, for example, the recruitment of autophagosomes to the surfaces of phagosomes is almost completely blocked. Due to the severe recruitment defect, whether the ced-1 null mutation further impairs the fusion between autophagosomes and phagosomes cannot be readily visualized. However, since a ced-1 null mutation impairs the recruitment of RAB-7 to and the production of PtdIns(3)P on phagosomal surfaces (Yu et al., 2008; Lu et al., 2012), and since the recruitment of the HOPS complex to the surfaces of intracellular organelles requires both RAB7 and PtdIns(3)P (Balderhaar and Ungermann, 2013; Stroupe et al., 2006; Jeschke and Haas, 2018), we predict that the CED-1 pathway would also control the RAB-7/HOPS complex-mediated autophagosome-phagosome fusion (Figure 10). Identifying the CED-1 signaling pathway as the driving force for the incorporation of autophagosomes to phagosomes helps to reveal the molecular mechanisms behind the crosstalk between autophagy and phagocytosis.

C. elegans strains were grown at 20 °C as previously described (Wood, 1988) unless indicated otherwise. The N2 Bristol strain was used as the wild-type control strain. Mutations are described in Riddle et al., 1997 and by the Wormbase (http://www.wormbase.org) unless noted otherwise (Key Resources Table): LG1: ced-1(e1735), epg-8(bp251); LGII: lgg-1(tm3489), rab-7(ok511), vps-18(tm1126); LGIII: atg-13(bp414), ced-6(n2095), cup-5(n3264); LGIV: atg-3(bp412), atg-7(bp411), ced-5(n1812), ced-10(n1993), lgg-2(tm5755 and tm6474); LGV: atg-9(bp564), atg-18(gk378), unc-76(e911), unc-51(e369); LGX: atg-2(bp576), dyn-1(n4039). dyn-1(n4039) homozygous mutants, which are zygotic embryonic lethal, were maintained by an extrachromosomal array carrying a wild-type dyn-1 gene and a co-expressed P_egl-13gfp marker (Key Resources Table) (Yu et al., 2006). dyn-1(n4039) homozygous embryos losing the rescuing transgene were identified as the embryos not carrying P_egl-13gfp. The rab-7(ok511) and lgg-1(tm3489) homozygous strains are both maternal-effect embryonic lethal, and the mIn1 balancer maintained each allele with an integrated pharyngeal GFP marker (Key Resources Table) (Edgley et al., 2006). To obtain rab-7(ok511) m^-z^- homozygous embryos, GFP^- rab-7(ok511) homozygous hermaphrodites were isolated among the progeny of the strain VC308, and their progeny were collected as embryos. The same protocol was used to collect lgg-1(tm3489) m^-z^- embryos from the strain GK738 (Key Resources Table). Double mutants between lgg-1(tm3489)/mIn1 and lgg-2(tm6474) were generated by standard genetic crosses.

Extrachromosomal arrays were generated by the microinjection of plasmids with the co-injection marker p76-18B [punc-76(+)] into unc-76(e911) mutants (Bloom and Horvitz, 1997; Jin et al., 1999). Non-Unc mutants were identified as transgenic animals. Integrated transgenic arrays were generated by gamma irradiation (Jin et al., 1999). Integrated arrays generated in this study are as follows (Key Resources Table): LGI: enIs87[P_ced-1PH(hPLCγ)::mrfp and P_ced-1mNG::lgg-1]; LGII: enIs82[P_ced-1 ced-1::gfp and P_ced-1 mCherry::lgg-1]; LGV: enIs85[P_ced-1 PH(hPLCγ)::mrfp and P_ced-1 mNG::lgg-2]; LGX: enIs83[P_ced-1 ced-1::gfp and P_ced-1 mCherry::lgg-2].

Plasmid construction lgg-1 and lgg-2 cDNAs were PCR amplified from a mixed-stage C. elegans cDNA library (Haley et al., 2018). To generate the P_ced-1 gfp::lgg-1 and P_ced-1 gfp::lgg-2 plasmids, lgg-1 and lgg-2 cDNAs were cloned into the XmaI and KpnI sites of plasmid pZZ956 (P_ced-1 5’gfp) (Haley et al., 2018). P_ced-1 mCherry::lgg-1 or –2 were constructed by replacing the cDNA of gfp with that of mCherry (Shaner et al., 2004). P_ced-1 mNG (mNeonGreen)::lgg-1 or –2 were constructed by replacing the cDNA of gfp with that of mNeonGreen (Shaner et al., 2013). To generate P_ced-1gfp::lgg-1(G116A) and P_ced-1gfp::lgg-2(G130A), the QuickChange Site-directed Mutagenesis Kit (Stratagene, La Jolla, CA) was used to introduce the above mutations into the constructs. To construct P_ced-1 atg-9::mCherry, the atg-9a open reading frame was PCR-amplified from a C. elegans mixed-stage cDNA library and inserted between P_ced-1 and gfp in pZZ829 (P_ced-1 3’gfp) (Haley et al., 2018). The gfp cDNA was then replaced by the mCherry cDNA (Shaner et al., 2004). All plasmids contain an unc-54 3’ UTR.

In a mixed population of the strain VC308, isolate mid-L4 stage hermaphrodite larvae that were rab-7(ok511) and place them in a new plate. Forty-eight hours later, these worms, which are adults, were processed for transmission electron microscopic studies as previously described (Yu et al., 2008; Yu et al., 2006). Cross-sections of the gonadal region were generated. In 50 nm thin sections, germ cell corpses and the neighboring gonadal sheath cells were identified as previously described. A phagosome containing a germ-cell corpses was identified in serial 50-nm-thin sections that cover the entire length of each germ-cell corpse were analyzed by determining whether a germ-cell corpse is entirely inside the vacuole inside the gonadal sheath cells. Phagosomal surfaces were carefully monitored in a thin section electron micrograph for the presence of double-membrane vesicles.

Cell corpses display a highly refractive button-like morphology under Differential Interference Contrast (DIC) microscopy. An Axionplan two compound microscope (Carl Zeiss, Thornwood, NY) equipped with Nomarski DIC optics, an AxioCam digital camera, and AxioVision imaging software was used for DIC microscopy. Using a previously established protocol (Lu et al., 2009), we quantified the number of cell corpses in the head region at the 1.5-fold and 2-fold stage embryos, which are ~420 and ~ 460 min post-first cleavage.

A DeltaVision Elite Deconvolution Imaging System (GE Healthcare, Inc) equipped with a DIC imaging apparatus and a Photometrics Coolsnap two digital camera was used to capture fluorescence, and DIC images Applied Precision SoftWoRx 5.5 software was utilized for deconvolving and analyzing the images (Lu et al., 2009). To observe the amount of autophagosomes in ventral hypodermal cells that express P_ced-1, fourteen serial Z-sections in 0.5 μm interval between adjacent optical sections, starting at the ventral surface of embryos at mid-embryonic stages, were collected. The 2D projection image of each Z-stack was generated and compared among different genetic backgrounds. To track fluorescence markers on pseudopods, on the surfaces of phagosomes, or inside phagosomal lumen during the clearance process of cell corpses C1, C2, and C3, embryos were monitored on their ventral surface starting at ~310 min post-first cleavage using an established time-lapse imaging protocol (Lu et al., 2009). Twelve to 16 serial Z-sections (at 0.5 μm intervals) were captured every 2 min, with recordings typically lasting between 60 and 180 min. Embryos that exhibited normal elongation and movement were considered developing properly. The moment engulfment starts is defined as when the extension of pseudopods around C1, C2, or C3 is first observed. The moment a nascent phagosome form is defined as when the pseudopods around a cell corpse join and make a full closure. The life span of a phagosome is defined as the time interval between the moments when the nascent phagosome forms and when the phagosome shrinks to one-half of its original diameter.

The time spans of the engulfment and degradation processes of cell corpses C1, C2, or C3 were measured as previously established (Haley et al., 2018). Briefly, a pseudopod marker, either CED-1::GFP or PH(PLCγ)::GFP, was monitored over time. The moment engulfment starts is defined as when the budding pseudopods around C1, C2, or C3 is first observed. The moment a nascent phagosome form is defined as the moment when the pseudopods around a cell corpse join and make a full closure. The period between the budding and the sealing of the pseudopods is the time span of engulfment. To measure phagosome duration, a co-expressed phagosomal surface marker mCherry::2xFYVE was used to track the diameter of the phagosome over time. The life span of a phagosome is defined as the time interval between the time points when a nascent phagosome was born and when the phagosome shrank to one-half of its original diameter.

To monitor the degradation process of the chromatin DNA of apoptotic cells inside phagosomes, we measured the size of the nucleus of engulfed C1, C2, or C3 over time by following the ubiquitously expressed P_his-72 his-72::mCherry reporter. Briefly, the ‘0 min’ timepoint of phagosome formation was determined by the appearance of the ‘bottom-like’ structure under DIC optics. The nuclear diameter of each phagosome was measured as the diameter of the mCherry⁺ disc inside the phagosome. The measurement is performed in regular time intervals until 60 min after the phagosome formation.

To measure the efficiency of recruitment of autophagosomes to phagosomes, we quantified the intensity of mCherry- or mNG-labeled LGG-1 or LGG-2 on the surfaces of phagosomes. First, we identified the boundary of a phagosome and the ‘0 min’ time point when a nascent phagosome just formed by observing co-expressed CED-1::GFP or PH(PLCγ)::GFP. At a particular time point, Tn or T0, the surface of a phagosome was outlined by two closed polygons (Figure 9H). The total signal intensities, as well as the areas of the polygons, were recorded. The unit signal intensity of the ‘donut-shape’ area between the two polygons was calculated as follows:

Unit Intensity (UI_phagosome) = (Intensity_{external polygon} – Intensity_{internal polygon})/(Area_{external polygon} – Area_{internal polygon}). The Unit Background Intensity (UI_background) was measured from a polygon outside the embryo was calculated as follow: UI_background = Intensity_background/Area_background.

At the time point Tn, the relative signal intensity (RInt_Tn) = (UI_phagosome – UI_background)_Tn / (UI_phagosome – UI_background)_T0. The RInt_Tn value 1.0 indicates no enrichment of LGG-1- or LGG-2-labeled autophagosomes on phagosomal surfaces comparing to the ‘0 min’ time point.

In embryos expressing P_his-72 his-72::gfp::mCherry, we quantified the fluorescence intensity of the mCherry and GFP signals at the center of phagosomes in regular time intervals. The ‘0 min’ timepoint of phagosome formation was determined by the appearance of the ‘bottom-like’ structure under DIC optics. At each time point, the total mCherry and GFP signal intensity of a fixed area (3 × 3 pixels) at the center of a phagosome (Int_mCherry, Int_GFP) was recorded. The acidification index (A_idx) at a particular time point (Tn) is defined as: A_idx = (Int_GFP/Int_mCherry)_Tn / (Int_GFP/Int_mCherry)_T0. The A_idx value 1.0 indicates no phagosome acidification compared to the ‘0 min’ value.

In wild-type embryos co-expressing either mNG::LGG-1 or mNG:: LGG-2 together with NUC-1::mCherry (mCh), we counted the numbers of mNG⁺ mCherry^- puncta and the mNG⁺ mCherry⁺ puncta observed on the surfaces of phagosomes. For each set of reporters, C1, C2, and C3 phagosomes were quantified in each embryo, and the numbers of puncta were added. Seven embryos were scored (n = 7), each representing one sample. The percentage of autolysosomes (%_AL) for each embryo was calculated as follow: %_AL = 100*((# of mNG⁺ mCh⁺ puncta)/ [(# of mNG⁺ mCh⁺) + (# of mNG⁺ mCh^-)]). The average value and standard deviation were calculated and reported.

In wild-type embryos co-expressing P_ced-1 mCherry-lgg-1 or -lgg-2 together with P_ced-1 gfp::rab-7, we counted the numbers of mCh⁺ GFP^- and mCh⁺ GFP⁺ puncta in the engulfing cells for C1, C2, and C3 in two different locations: (1) on the surfaces of the phagosomes, and (2) in the cytoplasm. For each set of co-expressed reporters, 9 engulfing cells were scored. N_mCh+ (the total number of mCh⁺ puncta) = # of mCh⁺ GFP^- puncta + # of mCherry⁺ GFP ⁺ puncta. % mCh⁺ GFP⁺ = 100 * (# mCh+ GFP + puncta)/ N_mCh+. The average values and standard deviation values were calculated and reported.

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

We thank Ying Wang and Lidan Gao for technical support and Ryan Haley for helpful comments. We thank H Robert Horvitz for advice and the use of the electron microscope. We thank Erika Hartwieg for the technical support for the TEM study. We thank the C. elegans Genetics Center (CGC), funded by the NIH Office of Research Infrastructure Programs (P40 OD010440), for providing some strains. We also thank the National BioResource Project of Japan and Dr. Shohei Mitani for providing some strains. NIH R01GM067848 supports this work.

1. Preprint posted: July 16, 2021 (view preprint)
2. Received: July 24, 2021
3. Accepted: January 3, 2022
4. Accepted Manuscript published: January 4, 2022 (version 1)
5. Version of Record published: January 18, 2022 (version 2)
6. Version of Record updated: January 21, 2022 (version 3)

© 2022, Peña-Ramos et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.