PROTECT YOUR DNA WITH QUANTUM TECHNOLOGY
Orgo-Life the new way to the future Advertising by Adpathway-
Loading metrics
Open Access
Peer-reviewed
Research Article
- Lei Yuan,
- Peiyao Li,
- Aiying Ma,
- Chao Li,
- Fuqiao Liu,
- Yunmin Xie,
- Qian Zheng,
- Hui Wang,
- Hui Xiao
x
- Published: November 19, 2026
- https://doi.org/10.1371/journal.pbio.3004123
This is an uncorrected proof.
Abstract
Apoptotic cell clearance represents the final stage of apoptosis and involves phagosome formation, maturation, and digestion. Here, we found that SUMO modifications participate in apoptotic cell clearance by influencing phagosomal degradation activity in C. elegans. The SUMO E3 ligase, GEI-17, SUMOylates GDP dissociation inhibitor (GDI-1) at K270 to trigger phagosomal degradation. Upon SUMOylation, GDI-1 facilitates release of GDP-bound RAB-1, which is subsequently converted to the GTP-bound form in a manner dependent on GDI displacement factor, PRAF-3. RAB-1 conversion enables binding and translocation of RAB-7 from the endoplasmic reticulum to Golgi to promote phagosome maturation and degradation activity. In the absence of GDI-1 SUMOylation, GDI-1 binds to GDP-RAB-1, sequestering it in the cytoplasm, resulting in impaired phagosomal degradation. Furthermore, we found SUMOylation of GDI1 at a conserved site plays a crucial role in efferocytosis regulation in mammals. This study defines a previously unrecognized mechanism by which SUMOylation drives apoptotic cell clearance.
Citation: Yuan L, Li P, Ma A, Li C, Liu F, Xie Y, et al. (2026) GEI-17-mediated SUMOylation of GDP dissociation inhibitor GDI-1 regulates apoptotic cell clearance. PLoS Biol 24(11): e3004123. https://doi.org/10.1371/journal.pbio.3004123
Academic Editor: Nektarios Tavernarakis, Foundation for Research and Technology-Hellas, GREECE
Received: May 28, 2025; Accepted: November 10, 2026; Published: November 19, 2026
Copyright: © 2026 Yuan et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: All relevant data are within the paper and its Supporting information files and “The potential SUMOylated proteins identified by LC-MS/MS in C. elegans” is available from the ProteomeXchange Datasets (accession number: PXD084683). “The peptides identified for GDI-1 by LC-MS/MS in C. elegans” is available from the ProteomeXchange Datasets (accession number: PXD084683).
Funding: This work was supported by the National Natural Science Foundation of China (31871387 to H. X., 32370799 to H. X.), National Natural Science Foundation of China Youth Program (32300622 to L. Y.), Sanqin Bochuang Talent Support Program of Shaanxi Province (2024SQBC004 to L. Y.), Natural Science Foundation of Shaanxi Province (2024JC-YBMS-638 to H. W.), and the program of Innovative Research Team for the Central Universities (GK202302003 to H. X.). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Abbreviations: AD, Alzheimer's disease; AO, acridine orange; BMDM, bone marrow-derived macrophages; CBB, Coomassie Brilliant Blue; Chl, chloroquine; dsRNA, double-stranded RNA; ER, endoplasmic reticulum; ERES, endoplasmic reticulum exit site; FOV, field of view; GDFs, GDI displacement factors; GDI1, GDP dissociation inhibitor 1; GO, Gene Ontology; hphs, hour post-heat shock; IP, immunoprecipitated; IPTG, isopropyl β-d-thiogalactopyranoside; lf, loss-of-function; LG, linkage group; NGM, Nematode Growth Medium; PBS, phosphate-buffered saline; pre-GIC, pre-Golgi intermediate; RNAi, RNA interference; SIMs, SUMO-interacting motifs; ss, secreted soluble; UV, ultraviolet; Y2H, Yeast two-hybrid assays; Y3H, Yeast 3-Hybrid.
Introduction
Apoptosis plays an important role in the development of multicellular organisms and maintenance of homeostasis in tissue [1–3]. As the final step in apoptosis, cell corpse clearance is a highly dynamic process involving mutual recognition between apoptotic cells and phagocytic receptors, cytoskeletal rearrangement in phagocytes, phagosome formation in the phagocytes extending pseudopods to engulf apoptotic cells, and ultimately phagosomal degradation within phagocytes [4–6]. Efferocytosis is used to describe the phagocytosis of dying cells across the animal kingdom and from professional and nonprofessional phagocytes. Defects in apoptotic cell clearance have been implicated in the development of various severe chronic inflammatory or catastrophic autoimmune disorders (e.g., systemic lupus erythematosus) as well as neurodegenerative diseases (e.g., Alzheimer’s disease) [7,8].
During development of C. elegans, 131 somatic cells and ~50% of germ cells undergo apoptosis. During development, neighboring cells engulf apoptotic cells that are not specified and arise from different tissues at the moment of engulfment. In germline apoptosis, it is exclusively the sheath cells engulfing the apoptotic germ cells [9,10]. Apoptotic cell clearance is evolutionarily highly conserved, and apoptotic cells are degraded by phagocytosis by two parallel and partially redundant pathways. One is composed of CED-1/MEGF10, CED-6/GULP, and CED-7/ABC transporter membrane proteins [11–13], and the other is composed of CED-2/CrkII, CED-5/Dock180, CED-10/Rac, and CED-12/Elmo proteins [14,15]. Under the synergistic action of these two signaling pathways, the pseudopods of phagocytosis are extended to engulf apoptotic cells to form phagosomes [16,17]. Subsequently, phagosomes mature in the presence of a series of Rab GTP hydrolases (RAB-5, RAB-2, RAB-14, and RAB-7), and the mature phagosomes fuse with lysosomes to form phagolysosomes. Finally, apoptotic cellular components within the phagolysosome are degraded by lysosomal hydrolases [18–20].
Ubiquitin and SUMOylation are very important post-translational modifications of proteins, are implicated in various biological processes, including DNA damage repair, cell proliferation and apoptosis, etc [21–24]. Our previous research has demonstrated that in Drosophila melanogaster, the E3 ubiquitin ligase Pallbearer is involved in phagocytosis by ubiquitinating RpS6, resulting in the activation of the RAC2 GTPase, which subsequently leads to the upregulation of F-actin remodeling [25]. Our recent studies have revealed that the E3 ubiquitin ligase TRIM-21 and the E2 conjugating enzyme UBC-21 function synergistically to catalyze K48-mediated polyubiquitination modification of the phagocytosis receptor CED-1 in C. elegans, targeting proteasomal degradation and promoting apoptotic cell clearance [26]. These studies demonstrated that ubiquitination modifications are very important for apoptotic cell clearance. Whereas ubiquitylation is primarily associated with protein turnover, SUMOylation primarily facilitates or inhibits protein interactions and alters protein conformation or localization, especially with proteins that contain SUMO-interacting motifs (SIMs), thereby enabling transient binding to target proteins [27,28]. SUMOylation is a reversible process that is conserved in C. elegans. Analogous to ubiquitination, SUMOylation occurs through the action of an E1-activating enzyme (the Sae1/Sae2 heterodimer in humans and AOS-1/UBA-2 in worms), an E2-conjugating enzyme (Ubc9 in humans and UBC-9 in worms), and SUMO-specific E3 ligases (PIAS and MMS21 in humans and GEI-17 and MMS-21 in worms) [29,30]. However, some SUMOylation reactions do not involve E3 ligases [31]. The deconjugation of SUMO from its targets is regulated by SUMO-specific isopeptidases (SENP1, 2, 3, 5, 6, and 7 in humans and ULP-1, ULP-2, ULP-4, and ULP-5 in worms) [32,33]. Mammals contain three different SUMO proteins, whereas C. elegans expresses only a single SUMO orthologue, SMO-1, which renders the worm an expedient model organism to dissect this post-translational protein modification system. SUMOylation has been predominantly investigated in the development of C. elegans, and has been implicated in embryonic, vulval, and muscle development, as well as in DNA damage response [29,34,35]. Disturbances in SUMOylation are closely related to the pathogenesis of a variety of human diseases such as neurodegenerative diseases, diabetes mellitus, innate immune diseases, and tumors. Compared to the study of ubiquitination modification in apoptotic cell clearance, the study of SUMO modification mostly focuses on the process of apoptosis. However, it remains unclear whether SUMO modifications are involved in the clearance of apoptotic cells.
To determine whether and how SUMO modifications contribute to apoptotic cell clearance, we induced SUMO pathway knockdown in C. elegans and found that cell corpse degradation, lysosomal recruitment of LAAT-1, NUC-1, and CPL-1, and phagosomal acidification were all impaired. Proteomic analysis revealed that GDP dissociation inhibitor 1 (GDI-1) was the substrate for SUMO modification required for apoptotic cell clearance, specifically at residue K270. In addition, through numerous genetic and chemical manipulation approaches, we found that GDI-1 SUMOylation by the E3 ligase, GEI-17, modulated cell corpse removal via regulation of RAB-1 activity through a mechanism dependent on competing for binding with the GDI displacement factor, PRAF-3, in C. elegans. Specifically, SUMOylated GDI-1 promotes RAB-1 binding and transport of RAB-7 from the endoplasmic reticulum (ER) to Golgi apparatus to positively regulate apoptotic cell clearance. In addition, we found that the human homolog of GEI-17, PIAS1, functioned similarly in regulating phagocytosis in mammalian cells, and that it could reciprocally complement GEI-17 activity towards C. elegans proteins in vitro. Moreover, we found that SUMOylation of GDI1 also occurs in mammals, where it contributes to the regulation of efferocytosis. These results suggest that this SUMOylation-dependent apoptotic cell clearance pathway may be conserved between worms and humans. Our study thus defines a previously undocumented regulatory mechanism for cell corpse clearance via regulation of GDI-1 SUMOylation, suggesting possible therapeutic targets for further exploration to alleviate efferocytosis-associated diseases.
Results
The SUMO system participates in cell corpse clearance in C. elegans
To determine whether and how SUMOylation may participate in regulating apoptosis in C. elegans, we induced silencing of smo-1 by RNA interference (RNAi). Examination of cell corpse numbers showed that worms with smo-1 knockdown had higher numbers of cell corpses both in the germline, in an age-dependent manner, and in embryos at six developmental stages from comma to 4F compared to that in the nontargeted RNAi controls (Fig 1a – 1c). As ced-3 and ced-4 are key factors in apoptotic cell death in C. elegans, loss-of-function (lf) mutations in ced-3 or ced-4, caused an almost complete loss of apoptotic cell death [36]. Subsequent examination of cell corpses showed that apoptotic cell death was almost completely lost in the absence of either gene. Furthermore, ced-3(n717) or ced-4(n1126) mutants treated with smo-1 RNAi also had significantly fewer apoptotic germ cells (Fig 1d), suggesting that the germ cell corpses we observed in smo-1 knockdown animals resulted from physiological cell death. As the C. elegans SUMO modification system also includes the E1 activator, AOS-1/UBA-2, the E2 conjugating enzyme, UBC-9, and the E3 ligase homologs, GEI-17/MMS-21, we also induced knockdown of individual SUMOylation pathway components, and found that the number of apoptotic germline cells significantly increased compared to control (Fig 1e).
Fig 1. The SUMO system regulates apoptotic cell clearance in C. elegans.
(a) DIC (differential interference contrast) images of N2 treated with the control or smo-1 RNAi germline. Arrows indicate button-like cell corpses. The boxed region is magnified (2.5×) in the inset. Bars, 5 µm. (b and c) Quantification (mean ± SEM) of button-like cell corpses at different adult stages (h post L4) (b) and different embryonic stages (c) in N2/control and smo-1 RNAi worms. Fifteen adult worms and embryos were scored at each stage for each strain. (d, e and g, k) The development of adult or embryo stages corpses was quantified (mean ± SEM) in the indicated strains. Fifteen adult worms or embryos were scored at each stage for each strain. (f) Schematic illustration of the smo-1 (ok359), ubc-9 (tm2610), mms-21 (xwh80), and gei-17 (xwh81) homozygous mutations generated by CRISPR-Cas9 editing of the endogenous mms-21 and gei-17 loci. Amino acids near the mutated sites are indicated. (l) Time-lapse recording of cell corpses in N2/control and smo-1 RNAi germlines. Arrows indicate cell corpses. Bars, 5 µm. (m) Four-dimensional microscopy analysis of the persistence of 30 germ cell corpses duration were performed in N2/control and smo-1 RNAi. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, ***P < 0.001, NS, no significance. All error bars indicate means and SEM. The data underlying this Figure can be found in S1 Data.
To further investigate how the SUMO system affects programmed cell death in C. elegans, we analyzed smo-1(ok359) and ubc-9(tm2610) deletion mutants, which represent strong loss-of-function mutations (Fig 1f). In addition, we used CRISPR-Cas9 to generate homozygous mutant alleles of the SUMO E3 ligases, gei-17(xwh81) and mms-21(xwh80), carrying nonsynonymous point mutations leading to premature N-terminal stop codons (Fig 1f). Successful insertion of the stop codon was confirmed by sequencing the gei-17 and mms-21 mutants. In addition, performed RNAi of gei-17 for the gei-17 mutant and mms-21 for the mms-21 mutant, it was found that there was no increase in the number of apoptotic cells, suggesting that the gei-17 and mms-21 mutants were completely loss-of-function mutants (S1a, S1b Fig). The smo-1(ok359) mutation resulted in an embryonic apoptosis phenotype that precluded examination of cell corpses due to complete early embryonic lethality. However, we observed significantly increased numbers of cell corpses at multiple embryonic stages and in ubc-9(tm2610), gei-17(xwh81), and mms-21(xwh80) adult germline cells relative to that in WT controls (Fig 1g, 1h). Notably, gei-17(xwh81) mutants had higher embryonic cell corpse numbers than mms-21(xwh80) mutants, whereas more germline cell corpses were detected in the mms-21(xwh80) mutant than gei-17(xwh81). Although unable to generate homozygous gei-17(xwh81); mms-21(xwh80) double mutants, we obtained gei-17(xwh81); mms-21(xwh80) double heterozygotes, potentially due to redundant functions of these genes required for viability and normal development in C. elegans. Alternatively, RNAi knockdown of gei-17 resulted in significantly enhanced numbers of germ cell corpses in the mms-21 mutant, suggesting that the two E3 ligases might belong to independent pathways (Fig 1i). Taken together, these results demonstrated that SUMO modifications contributed to regulating apoptosis in C. elegans.
To investigate whether the increase in apoptotic cells was due to activation of apoptosis in a higher proportion of cells or defects in apoptotic cell clearance, we treated xwhIs49(Pced-1rde-1; rde-1) transgenic worms (with phagocyte-specific RNAi suppression) and xwhIs82(Pegl-1rde-1; rde-1) transgenic worms (with apoptotic cell-specific RNAi suppression) with smo-1 RNAi, ubc-9 RNAi, gei-17 RNAi, and mms-21 RNAi. We found that cell corpse counts significantly increased in Pced-1rde-1; rde-1, but not Pegl-1rde-1; rde-1 transgenic worms compared to their respective RNAi vector controls (Fig 1j, 1k), suggesting that silencing smo-1, ubc-9, gei-17, and mms-21 affected cell corpse removal rather than induction of apoptosis. To further verify this result, we obtained two transgenic worms: DCL569 (with germline-specific RNAi) [37] and Plim-7rde-1; rde-1(ne219) (with germline sheath cell-specific RNAi) [12]. Through knockdown of smo-1, ubc-9, gei-17, and mms-21 in DCL569 and Plim-7rde-1; rde-1(ne219) transgenic worms, we found that knockdown of smo-1, ubc-9, gei-17, and mms-21 in transgenic worms DCL569 did not increase the number of apoptotic cells in the germline compared to controls (S1c Fig), whereas Plim-7rde-1; rde-1(ne219) exhibited a phenotype of increased apoptotic cell numbers (S1d Fig). The results collectively indicate that the absence of the key component of SUMO was not responsible for the apoptosis of germ cells but rather for the removal of apoptotic cells from the germline.
We next conducted four-dimensional microscopy to measure the duration of embryonic or germ cell corpse persistence. In wild-type animals, the germ cell corpses seldom persisted for >60 min. By contrast, smo-1 knockdown led to significantly longer persistence in gonads for the large majority of germ cell corpses (>100 min) (Fig 1l, 1m). In addition, germ cell corpses persisted significantly longer in gei-17(xwh81) and mms-21(xwh80) mutants than in WT worms (S1e Fig). Additionally, cell corpse removal required significantly more time in gei-17(xwh81) and mms-21(xwh80) embryos compared to wild-type embryos (S1f, S1h Fig). However, the number of cell deaths in gei-17(xwh81) and mms-21(xwh80) embryos was indistinguishable from that in the wild type (S1g, S1i Fig). These findings indicated that SUMO modifications participated in the clearance of apoptotic cells rather than induction of apoptosis.
As phagocytosis of apoptotic cells is controlled by two partially redundant signaling pathways, CED-1/CED-6/CED-7/NRF-5/TTR-52/DYN-1 and CED-2/CED-5/CED-10/CED-12/PSR-1, in C. elegans, we next examined whether SUMO modifications functioned conjointly with engulfment genes to regulate cell corpse clearance. To this end, we treated the ced-1(e1735) and ced-6(n1813) engulfment mutants with smo-1 RNAi and found that smo-1 silencing led to significantly enhanced defects in germ cell corpse engulfment at different adult stages in both mutants (S1j Fig). Intriguingly, we found that smo-1 RNAi treatment also resulted in greater impairment of germ cell corpse engulfment in ced-5(n1812) and ced-12(n3261) mutants, the alternate regulatory pathway for cell corpse engulfment (S1j Fig). Additionally, smo-1 RNAi treatment also led to increased germ cell corpse abundance in the ced-1(e1735); ced-2(n1994) double mutant (S1k Fig). These findings suggested that SUMO modifications functioned independently of either pathway for cell corpse removal.
SUMOylation promotes apoptotic cell clearance through phagosome maturation
Since phagosome assembly and subsequent clearance of apoptotic cells involves several steps in C. elegans, we investigated the possible role of SUMOylation in phagosomal recruitment of factors required for phagosome maturation. Fluorescence microscopy with a CED-1::GFP phagocytic receptor fusion reporter in germline cells of smo-1 knockdown worms indicated that cell corpses were surrounded by CED-1::GFP, similar to vector controls (Fig 2a, 2g), which suggested that recognition and initiation of engulfment were unaffected by smo-1 suppression. To determine whether SUMOylation affects phagosomal recruitment, we examined GFP/mCHERRY-labeled markers for different stages of phagosome maturation, including GFP::RAB-5, a small GTPase specific to early endosomes and phagosomes; GFP::RAB-7, a small GTPase required for late endosome–phagosome fusion [38]; LAAT-1::GFP, a lysosomal membrane protein [39]; NUC-1::mCHERRY, a lysosomal DNase [40]; and CPL-1::mChOint, a lysosomal cathepsin protease [41]. We found that the percentage of phagosomes surrounded by GFP::RAB-5 and GFP::RAB-7 in germ cells did not significantly differ between smo-1 RNAi-treated worms and vector controls (Fig 2b, 2c, 2g), whereas significantly fewer phagosomes were surrounded with LAAT-1::GFP under smo-1 knockdown (Fig 2d, 2g). Moreover, NUC-1::mCHERRY and CPL-1::mChOint signals were also significantly decreased on phagosomes in the germ line of smo-1 knockdown animals compared to that in control animals (Fig 2e, 2f, 2g). To further determine whether smo-1 RNAi affects phagosome maturation, we subjected dual-reporter transgenic worms incorporating engulfment and phagosomal markers to control and smo-1 RNAi. These strains included smIs34(Pced-1ced-1::gfp); qxIs68(Pced-1mcherry::rab-7), smIs34(Pced-1ced-1::gfp); qxIs352(Pced-1laat-1::mcherry), smIs34(Pced-1ced-1::gfp); qxIs257(Pced-1nuc-1::mcherry), and smIs34(Pced-1ced-1::gfp); yqEx620(Pced-1cpl-1::mChOint). By quantifying the proportion of CED-1::GFP-labeled phagosomes that also exhibited the corresponding mCHERRY fusion marker (mCHERRY::RAB-7, LAAT-1::mCHERRY, NUC-1::mCHERRY, or CPL-1::mChOint), we found that compared with the control, smo-1 RNAi did not significantly alter the mCHERRY::RAB-7/CED-1::GFP ratio but significantly decreased the LAAT-1::mCHERRY/CED-1::GFP, NUC-1::mCHERRY/CED-1::GFP, and CPL-1::mChOint/CED-1::GFP ratios (S2a, S2b Fig). These results are consistent with the single phagosomal marker assay results and further demonstrate that smo-1 RNAi affects the recruitment of LAAT-1, NUC-1, and CPL-1 to phagosomes. To ascertain which phases of apoptotic cell clearance were influenced by other components of the SUMO system, RNAi was performed against key SUMO components, including E1 (aos-1, uba-2), E2 (ubc-9), and E3 (mms-21, gei-17), in GFP/mCHERRY-labeled markers transgenic worms for different stages of phagosome maturation. Statistical analysis revealed that RNAi of key SUMO components did not affect phagosomes labeled with CED-1::GFP, GFP::RAB-5, and GFP::RAB-7 (S2c Fig). However, RNAi of aos-1, ubc-9, and gei-17 resulted in impaired phagosome recruitment of LAAT-1::GFP, NUC-1::mCHERRY, and CPL-1::mChOint (S2c Fig). RNAi of uba-2 had a significant impact on the phagosome recruitment of LAAT-1::GFP and NUC-1::mCHERRY (S2c Fig), while mms-21 exhibited a distinct effect on the phagosome recruitment of NUC-1::mCHERRY (S2c Fig). These data suggested that SUMO modifications regulate recruitment to phagosome in the later stages of maturation. We found that RNAi knockdown of gei-17 resulted in a significantly enhanced number of germ cell corpses in the mms-21 mutant (Fig 1i). As RNAi knockdown of gei-17 and RNAi of mms-21 influence different phases of apoptotic cell clearance, this further suggests that the two E3 ligases might belong to independent pathways.
Fig 2. SUMOylation promotes apoptotic cell clearance through phagosome maturation.
(a–f) Representative images of cell corpse labeling by CED-1::GFP (a), GFP::RAB-5 (b), GFP::RAB-7 (c), LAAT-1::GFP (d), NUC-1::mCHERRY (E), and CPL-1::mChOint (f). Cell corpses are indicated by arrows. Bars, 5 μm. (g) Quantification (mean ± SEM) of the labeling of cell corpses by different phagosomal markers shown in a–f. ≥ 100 cell corpses from at least 15 worms were analyzed for each marker; and data were derived from three replicates. Comparisons were performed between the N2/control and smo-1 RNAi worms for each marker. (h) AO staining of germ cell corpses from the indicated animals. Arrows indicate cell corpses in the N2/control and smo-1 RNAi worms. Bars, 5 μm. (i) Quantification (mean ± SEM) of AO-positive germ cell corpses in N2/control and smo-1 RNAi worms. ≥ 100 cell corpses from at least 15 worms were analyzed. The two-tailed unpaired t test was used for statistical comparison of two groups of samples. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. **P < 0.01, ***P < 0.001, NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data.
To test whether the persistence of cell corpses in smo-1 RNAi-treated worms was due to failed acidification of phagosomes containing cell corpses, a late stage of corpse clearance, acridine orange (AO) staining was used to identify compromised cell corpses. This analysis indicated that smo-1 (RNAi) worms had obviously decreased AO staining compared with control RNAi worms (Fig 2h, 2i), supporting the likelihood that impaired SUMO modification affected the acidification of phagosomes following cell corpse engulfment. Subsequently, we validated the AO results through experiments involving pHluorin (a pH-sensitive GFP) and pHrodo staining. Initially, we generated Plim-7::pHluorin transgenic worms via microinjection, with the lim-7 promoter facilitating gene-specific expression in germ sheath cells. Following RNAi treatment targeting smo-1 in Plim-7pHluorin transgenic worms, we observed a significantly increased proportion of phagosomes containing apoptotic cells exhibiting green fluorescence compared to controls (S2d Fig). This observation suggested that smo-1 RNAi treatment led to impaired phagosome acidification. Subsequently, we employed pHrodo Red AM dye to stain the gonads of worms, serving as an indicator of phagosome acidification in the presence of apoptotic cells. We dissected the gonads of three worm strains—N2, ced-1(e1735) (cell corpses were not phagocytosed in these mutants), and vps-18(tm1125) (cell corpses are not degraded properly in these mutants) [42]—and stained them with pHrodo. Following a 15-minute incubation at room temperature, fluorescence analysis revealed that phagosomes containing apoptotic cells in the ced-1(e1735) mutant were infrequently stained by pHrodo, whereas nearly all phagosomes containing apoptotic cells in the vps-18(tm1125) mutant were labeled by pHrodo (S2e Fig), indicating the efficacy of this method. By staining the gonads of worms subjected to smo-1 RNAi, we observed a significant reduction in pHrodo-labeled phagosomes containing apoptotic cells compared to the control group (S2e Fig). These findings further demonstrated that smo-1 deficiency resulted in defective phagosome acidification.
To further examine the impact of SUMOylation on post-engulfment phagosomal degradation, we next quantified endosomal/lysosomal degradation activity in smo-1 knockdown worms using arIs36(Pheat-shockssGFP) transgenic worms that expressed ssGFP under the control of a heat-shock-inducible promoter. Monitoring of ssGFP uptake and degradation by coelomocytes over 48 hours showed that ssGFP began to accumulate in the body cavity (pseudocoelom) of control RNAi worms within 1 hour post-heat shock (hphs) treatment, and was subsequently transported to coelomocytes (S3a Fig). The ssGFP signal then progressively decreased in the body cavity over 12 hphs, coinciding with increased GFP signal in coelomocytes. At 24 hphs, ssGFP was no longer detectable in the body cavity or coelomocytes (S3a Fig), indicating that ssGFP was taken in and degraded by coelomocytes. In smo-1 knockdown animals, ssGFP was transported to coelomocytes by 1 hphs, similar to that in controls, suggesting that SUMOylation does not affect ssGFP uptake into coelomocytes. However, ssGFP was still detectable in the body cavity as well as coelomocytes of smo-1 RNAi-treated animals at 48 hphs (S3a Fig), indicating that endosomal/lysosomal protein degradation was markedly impaired in coelomocytes of the smo-1 knockdown group. These results suggested that suppressing SUMOylation resulted in obviously attenuated endosomal/lysosomal degradation activity.
To directly observe the process of cell corpse degradation, we next induced smo-1 knockdown in transgenic ujIs113(Ppie-1H2B::mCherry; Pnhr-2HIS-24::mCherry) worms expressing germline-specific markers of chromatin H2B::mCHERRY and HIS-24::mCHERRY. We observed that chromatin was condensed in germ cell corpses and subsequently disappeared by around 30 min (31.00 ± 2.83 min, n = 5). On the contrary, we observed that chromatin was condensed in germ cell corpses in smo-1 RNAi worms, but then diffused throughout the corpse, rather than disappearing, in later stage phagosomes, with mCHERRY signal persisting for >70 min (80.00 ± 7.94 min, n = 5) (S3b, S3c Fig). These results suggested chromatin degradation was significantly delayed in the cell corpse-containing phagosomes of smo-1 knockdown worms. Collectively, these findings demonstrated that cell corpse accumulation associated with smo-1 knockdown was due to defects in cell corpse degradation in engulfing cells.
GDI-1 is the SUMOylation substrate potentially involved in apoptotic cell clearance
To identify the SUMOylation substrates potentially involved in apoptotic cell clearance in C. elegans, we generated an xwh83(ha::smo-1) worm strain by CRISPR-Cas9 gene editing that expressed smo-1 with an N-terminal HA tag as its sole copy of the SUMO gene (S4a Fig). We then immunoprecipitated (IP) SUMO-conjugated proteins with an anti-HA antibody and identified components of the isolated protein mixture by LC-MS/MS analysis (Fig 3a). Given the widespread expression of SUMO in C. elegans, and our use of a mixed worm population, we anticipated identifying targets across all stages of development and tissues. Based on the screening criterion FDR < 0.01, we identified proteins pulled down with ha::smo-1 and wild-type N2and subtracted the proteins pulled down with N2 from the total number of proteins pulled down with ha::smo-1. In total, this analysis uncovered 188 SUMO interactors (noncovalent) or SUMOylated proteins in C. elegans under normal developmental conditions (S1 Table).
Fig 3. GDI-1 is the SUMOylation substrates potentially involved in apoptotic cell clearance.
(a) HA IP was performed on ha::smo-1 worms, followed by the detection of proteins that interact with SMO-1 using LC-MS/MS. IB, immunoblot. (b) The potential SMO-1 interacting proteins were identified by protein mass. LFQ value for each hit is the sum. (c) A list of potential substrates for SUMOylation associated with lysosomes and vesicles, with data derived from the protein mass spectrometry results of SUMO-GFP. (d) Quantification (mean ± SEM) of germ cell corpses from the indicated strains. Cell corpses of each animal were scored in 15 animals at every time point (h post L4) as indicated. (e) Quantification (mean ± SEM) of cell corpses positive for phagosomal markers in indicated strains. A total of ≥100 cell corpses from at least 15 worms were analyzed for each marker; and data were derived from three replicates. (f) Quantification (mean ± SEM) of AO-positive germ cell corpses in the N2/control, smo-1, vha-9, vha-14, gdi-1, and arl-8 RNAi worms. (g) The interactions between GEI-17 and GDI-1 were examined by GST pull-down assays. GST and GST::GFP serve as negative controls. (h) Recombinant HA::GDI-1 is modified by hSUMO1 in vitro. Recombinant HA::GDI-1 was incubated for the indicated times (minutes) in modification reactions, including E1 (hSAE1 and hSAE2), E2 (hUBC9), hSUMO1, and ATP (lanes 5–8). HA IP was performed, followed by detection of SUMOylation with anti-HA antibodies. SUMOylation of GDI-1 was observed (lane 8 was compared with lanes 1–7). (i) Schematic illustration of the mutation and tag insertions generated by CRISPR-Cas9 editing of the endogenous gdi-1 loci. Amino acids near the insertion are indicated. (j) SUMOylation of GDI-1 was examined in N2, ha::smo-1, flag::gdi-1, and ha::smo-1; flag::gdi-1. FLAG IP was performed, followed by detection of SUMOylation with anti-HA antibodies. An isotype-matched rabbit IgG was used as the control. (k) GDI-1::GFP localization in larvae (top) and embryos (bottom) of xwhIs85 (Pgdi-1gdi-1::gfp) worms was imaged in the GFP channel using an Imager M2 microscope. Arrows indicate cell corpses in embryo images. The boxed region is magnified (2.0×) in the inset. Scale bars: 50 µm (top) and 2 µm (bottom). (l) Representative images (top) and quantification (bottom) of GDI-1::GFP localization on cell corpses in control and gei-17 RNAi germlines. Arrows indicate cell corpses. Bars, 5 µm. The two-tailed unpaired t test was used for statistical comparison of two groups of samples. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, ***P < 0.001. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images and the Coomassie Brilliant Blue-stained protein gel underlying this Figure can be found in S1 Raw Images.
We then examined Gene Ontology (GO) functional annotations of the SUMO-modified proteins (S4b Fig), and selected the RAB GDP dissociation inhibitor, gdi-1, due to its reported involvement in vesicular transport (Fig 3b), and based on our understanding that SUMO likely promotes apoptotic cell clearance through phagosome maturation in our above experiments. In addition to gdi-1, we examined six other candidates identified in similar proteomics studies in C. elegans and mammalian cells (Fig 3c and S1 Table), including LMP-1, a membrane protein found in cytoplasmic vesicles and lysosomal membranes; ARL-8, a small lysosomal GTPase that regulates axonal dendritic transport, phagosome-to-lysosome translocation, and synaptic vesicle paracrine transport; and five VHA family members (VHA-1, VHA-7, VHA-9, VHA-14, and VHA-19) that regulate lysosomal luminal acidification through proton-transporting ATPase activity [32,43]. Among these candidates, inactivation of vha-9, vha-14, arl-8, or gdi-1 by RNAi in WT worms produced an apoptotic phenotype similar to smo-1 knockdown (Fig 3d–3f).
To further elucidate the SUMO substrate involved in apoptotic cell clearance, we first examined VHA-9, VHA-14, ARL-8, and GDI-1 interactions with SMO-1 by Yeast two-hybrid assays (Y2H), and found that only ARL-8 and GDI-1 likely exhibited interactions with SMO-1 (S4c, S4d Fig). As SMO-1 can undergo direct covalent or noncovalent binding with substrates, or may interact with substrates indirectly through covalent or noncovalent binding with an intermediary interaction partner, we generated SMO-1 truncation variants with C-terminal deletions that either exposed the conserved bis-glycolide motif (GG’: active form) or deleted the bis-glycine motif (ΔGG: inactive form) [44]. Subsequent Y2H assays showed that WT and SMO-1 GG’, but not the SMO-1 ΔGG variant, were capable of interacting with ARL-8 or GDI-1 (S4d Fig). These results suggested that SMO-1 exhibited covalent, rather than noncovalent, direct or indirect interactions to modify ARL-8 or GDI-1. Co-IP and pull-down assays including two SUMO E3 ligases and ARL-8 or GDI-1 suggested that GDI-1 could exclusively interact with GEI-17, but not MMS-21, whereas ARL-8 did not interact with either E3 ligase (Figs 3g and S4e–S4g). Although the relatively high molecular weight of GST::GEI-17 (~120 kDa) inevitably resulted in some degradation products and its yield was lower than that of the control GST::GFP protein, which represents a certain limitation, GST::GEI-17 still efficiently pulled down MYC::GDI-1, whereas the control GFP did not. These results suggested that GDI-1 might serve as a substrate of GEI-17 [which has been shown to interact with SMO-1]. We therefore focused on GEI-17 function in apoptotic cell clearance.
To identify the regions of GDI-1 required for interaction with GEI-17, we generated a series of GDI-1 truncation constructs based on secondary structures identified in its amino acid sequence. Co-IP and GST pull-down assays showed that only the region between residues 1−150 of GDI-1 was required for binding with GEI-17 (S5a–S5i Fig). As SUMO-1/2/3 serve as precursor molecules in mammals, while SAE1/2 function as E1 activating enzymes, and UBC9 acts as the E2 binding enzyme, we next tested whether GDI-1 was a SUMO substrate through in vitro SUMOylation reactions using the human homologs of these proteins expressed and purified in E. coli. We found that recombinant hUBC9, hSAE1/hSAE2, and hSUMO1 could together SUMOylate GDI-1 (Figs 3h and S4h). Notably, the SUMOylation of GDI-1 was completely abolished by excluding either hSAE1/hSAE2 or hUBC9 (Fig 3h), indicating that GDI-1 was a bona fide SUMOylation target in vitro.
To determine whether GDI-1 was SUMOylated in C. elegans in vivo, we inserted a flag tag into the endogenous gdi-1 loci using CRISPR-Cas9, which resulted in the xwh84(flag::gdi-1) C. elegans strain (Fig 3i). We then crossed the xwh83(ha::smo-1) and xwh84(flag::gdi-1) strains to obtain progeny carrying both tagged proteins and performed anti-FLAG IP to test whether GDI-1 was directly modified with SMO-1. Western blot detection of HA in IP lysates of the flag::gdi-1; ha::smo-1 recombinant progeny revealed a high-molecular mass species (Fig 3j) that likely represented SUMOylated GDI-1 in these worms.
To next investigate whether and how SUMO modification of GDI-1 was involved in apoptotic cell clearance in C. elegans, we generated a transgenic xwhIs85(Pgdi-1gdi-1::gfp) strain expressing GDI-1::GFP driven by the native gdi-1 promoter. Interestingly, fluorescence microscopy showed that GDI-1::GFP was mainly present in pharyngeal and body wall muscle cells (Fig 3k). Since apoptotic cells are engulfed by neighboring cells in C. elegans, including hypodermal cells, intestinal cells, muscle cells, and gonadal sheath cells, together with our finding that GDI-1 was mainly localized in pharyngeal and body wall cells, we hypothesized that GDI-1 could play a role in phagocytosis of neighboring cell corpses. Furthermore, we discovered that in the embryos of the xwhIs85 transgenic worms, GDI-1::GFP was localized to phagosomes containing apoptotic cells (Fig 3k). To test this possibility, we constructed a transgenic xwhIs86(Pced-1gdi-1::gfp) strain expressing GDI-1::GFP driven by the native ced-1 promoter. Subsequent fluorescence microscopy analysis showed that germ cell corpses were surrounded by GDI-1::GFP in the xwhIs86 worms. Moreover, we observed that significantly lower levels of GDI-1::GFP labeled germline corpses were recruited to phagosomes in gei-17 RNAi-treated worms compared to that in vector control animals (Fig 3l). These data suggested that SUMO modification of GDI-1 by GEI-17 indeed participated in apoptotic cell clearance and likely played an important role in the cell corpse degradation process.
SUMOylation of GDI-1 at K270 is required for apoptotic cell degradation
In order to identify the specific lysine residue on GDI-1 targeted for SUMOylation by the E3 SUMO ligase, we crossed the xwh83(ha::smo-1) and xwhIs86(Pced-1gdi-1::gfp) strains to obtain progeny co-expressing both tagged proteins for anti-HA co-IP assays. Western blot detection of HA in IP lysates of the recombinant progeny revealed a smear of high-molecular mass protein species (Fig 4a). Subsequent LC-MS/MS of the purified lysates (S2 Table) identified K113, K175, and K270 as predicted SUMO-acceptor sites on GDI-1 through SUMOplot analysis of the MS protein profile (Fig 4b). Site-directed mutagenesis converting the candidate lysine sites to arginines resulted in single, double, and triple GDI-1 SUMOylation site mutants, including K113R, K113R/K175R, and K113R/K175R/K270R. Y2H analysis of potential interaction between these mutants and SMO-1 showed that only GDI-1(K270R) lost the ability to interact with SMO-1 (Fig 4c).
Fig 4. SUMOylation of GDI-1 at K270 is required for apoptotic cell degradation.
(a) HA IP was performed on ha::smo-1; Pced-1gdi-1::gfp worms, followed by detection of the peptide of GDI-1 that interacts with SMO-1 by LC-MS/MS. (b) Predicted SUMOylation sites of GDI-1 using the SUMOplot web service (https://www.abcepta.com/sumoplot) (top), schematic illustration of the SUMOylation sites of GDI-1 (bottom). (c) Yeast two-hybrid assay of the interactions between GDI-1 (WT, K113R, K113R/175R, and K113/175/270R) and SMO-1. (d) In vitro SUMOylation of the wild-type and 270 lysine-to-arginine (K270R) mutant forms of GDI-1. The sumoylated-GDI-1 band is indicated by an arrowhead. SUMOylation of GDI-1(K270R) was not observed (lane 3 was compared with lanes 1, 2, and 4). (e) Schematic illustration of the mutation and tag insertions generated by CRISPR-Cas9 editing of the endogenous gdi-1 loci. Amino acids near insertion or mutation sites are indicated. (f) SUMOylation of GDI-1 was examined in worms carrying ha::smo-1 and flag::gdi-1 (WT and K270R). Long and short designate long and short exposure times, respectively. An isotype-matched rabbit IgG was used as the control. (g) The different stages (h post L4) of germ cell corpses were quantified (mean ± SEM) in the indicated strains. Fifteen adult worms were scored at each stage for each strain. (h and i) Cell corpses positive for phagosome markers in control/gdi-1 RNAi (h) and N2, gdi-1(K270R) (i) were quantified. A total of ≥100 cell corpses from at least 15 worms were analyzed for each marker, and data were derived from three replicates. The data exhibited by the control group are consistent with the data presented in Fig 2g. (j–l) OP50, PA14, and SL1344 survival assays in N2, gei-17, and gdi-1(K270R) animals. Three biological replicates were used with 100 worms per replicate. The two-tailed unpaired t test was used for statistical comparison of two groups of samples. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. Lifespan assays were analyzed using Log-Rank method. *P < 0.05, **P < 0.01, ***P < 0.001, NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images underlying this Figure can be found in S1 Raw Images.
To verify K270 as the SUMO modification site on GDI-1, we performed in vitro SUMOylation reactions with recombinant proteins expressed and purified from E. coli. We found that SUMO modification was indeed abolished in the GDI-1(K270R) variant (Fig 4d), thus supporting K270 as the likely SUMOylation site. We then generated the xwh87(flag::gdi-1(K270R)) worm strain by CRISPR-Cas9 editing that expressed a flag-tagged gdi-1 that had a K◊R conversion at the SUMOylation site (Fig 4e). Western blot assays of IP lysates revealed that the high molecular weight species were absent in the xwh87 strain (Fig 4f), thus demonstrating K270 as the site required for GDI-1 SUMOylation.
To further explore the role of GDI-1 SUMOylation in apoptotic cell clearance, CRISPR-Cas9 was used to introduce the K270R mutation at the endogenous gdi-1 locus, resulting in the xwh89(gdi-1(K270R)) mutant strain, which exhibited significantly increased numbers of germline cell corpses compared to that in N2 worms (Fig 4g). Moreover, RNAi silencing of smo-1 in these gdi-1(K270R) mutants did not significantly affect germ cell corpse number (Fig 4g), supporting K270 as the GDI-1 SUMOylation site necessary to activate apoptotic cell clearance. Given our above results showing that SUMOylation positively regulates phagosomal recruitment of LAAT-1 and acidification of the phagosome (Fig 2g), we next examined SUMO modification of GDI-1 was required for recruitment of phagosomal maturation factors. Fluorescence microscopy revealed both gdi-1 RNAi knockdown worms and gdi-1(K270R) mutants expressing LAAT-1::GFP, NUC-1::mCHERRY, or CPL-1::mChoint had a significantly lower proportion of germline cell corpses with these markers around phagosomes compared to WT (Fig 4h, 4i), which was consistent with our above results in smo-1 knockdown worms. These data indicated that SUMOylation of GDI-1 residue K270 promotes apoptotic cell clearance by influencing phagosome maturation in vivo.
We then investigated whether endosomal/lysosomal degradation activity was affected in gei-17 mutants and gdi-1(K270R) mutants by introducing the arIs36(Pheat-shockssGFP) expression vector into these strains. Quantification of ssGFP uptake and degradation in coelomocytes over a 36 h time course showed that ssGFP began to accumulate in the body cavity (pseudocoelom) of arIs36 worms by 1 hphs, and was subsequently transported to coelomocytes (S6a Fig). The ssGFP signal progressively decreased in the body cavity by 12 hphs, conjointly with increased ssGFP GFP signal in coelomocytes, and was no longer detectable in the body cavity or coelomocytes by 24 hphs (S6a Fig), suggesting its complete degradation by coelomocytes. Disruption of SUMO modification did not obviously affect initial ssGFP uptake into coelomocytes, but ssGFP could still be detected in both the body cavity and coelomocytes of gei-17 mutants or gdi-1(K270R) mutants at 36 hphs (S6a Fig), suggesting that loss of GDI-1 SUMOylation led to impaired endosomal/lysosomal activity.
To directly observe the process of cell corpse degradation, we introduced the ujIs113(Ppie-1H2B::mCherry; Pnhr-2HIS-24::mCherry) vector construct into WT and gdi-1(K270R) worms. Detection of mCherry signal by fluorescence microscopy showed condensed chromatin in germ cell corpses within early stage phagosomes and disappeared within ~30 min in WT worms. By contrast, mCHERRY signal persisted for longer than 90 min in the ujIs113; gdi-1(K270R) strain (S6b, S6c Fig), suggesting that chromatin degradation in cell corpses was substantially delayed in the absence of GDI-1 SUMOylation. These data cumulatively demonstrated that cell corpse accumulation in gdi-1(K270R) animals was due to defects in lysosomal degradation of engulfed cell corpses.
Although lacking innate immunity, the C. elegans immune system to some extent resembles the human innate immune system and can respond to pathogen infection by upregulating antimicrobial peptide expression. As apoptotic cell clearance has been linked to innate immune function in C. elegans, we also examined whether GDI-1 SUMOylation contributed to regulating innate immunity by analyzing the life span of WT N2, gei-17 knockout, and gdi-1(K270R) worms fed on OP50, P. aeruginosa PA14, and S. typhimurium SL1344, respectively. Although mutants fed with OP50, the standard nematode diet, exhibited similar life span to that of N2, both gei-17 and gdi-1(K270R) strains became infected and showed significantly reduced lifespans upon feeding with the pathogenic PA14 and SL1344 bacterial strains (Figs 4j–4l and S6d–S6f), suggesting that GDI-1 SUMOylation at K270 was required for C. elegans resistance to bacterial pathogens.
GDI-1 regulates RAB-1 activity during cell corpse clearance
To better define possible connections between GDI-1 SUMOylation and its function in apoptotic cell clearance, we investigated whether SUMO modification altered the protein stability, protein–protein interactions, subcellular localization, or catalytic activity of GDI-1. Western blots showed that endogenous GDI-1 protein levels did not significantly differ between the N2, gei-17, and gdi-1(K270R) strains (Fig 5a), suggesting that SUMOylation does not affect GDI-1 stability. As GDI family proteins regulate GDP/GTP exchange of RAB proteins by stabilizing membrane-associated, prenylated RABs in the cytosol, or restoring them to the membrane for activation, we next sought to identify the specific RAB binding partner of GDI-1 involved in apoptotic cell clearance. To this end, we individually expressed 31 RAB proteins from C. elegans [45] in Y2H assays and found that GDI-1 could interact with RAB-1, RAB-28, and RAB-Y1 (Figs 5b and S7a–S7f).
Fig 5. GDI-1 regulates RAB-1 activity during cell corpse clearance.
(a) Endogenous GDI-1 protein levels were examined by immunoblot analysis in N2, gei-17, and gdi-1(K270R) worms. Graphs show quantification of the protein level of GDI-1 in the indicate strains. (b) The yeast two-hybrid assay of the interaction between GDI-1 and RAB-1/RAB-28/RAB-Y1. (c and h) The interactions between GDI-1 and RAB-1 (c) and GDI-1 or RAB-1 and PRAF-3 (h) were examined by GST pull-down assays. GST and GST::GFP serve as negative controls. (d) Germ cell corpses positive for phagosome markers in N2 treated with control or rab-1/rab-28/rab-y1 RNAi were quantified. A total of ≥100 cell corpses from at least 15 worms were analyzed for each marker; and data were derived from three replicates. The data exhibited by the control group is consistent with the data presented in Fig 2g. (e, f, and i) The different stages (h post L4) of germ cell corpses were quantified (mean ± SEM) in the indicated strains. Fifteen adult worms were scored at each stage for each strain. (g) Four-dimensional microscopy analysis of cell corpse duration was performed in N2/control and N2/praf-3 RNAi worms. The persistence of 30 cell corpses from the embryos was monitored. (j) The yeast three-hybrid assay of GDI-1 and PRAF-3 competitively combined with RAB-1. Constitutive Gal4-BD GDI-1 expression and inducible PRAF-3 expression elements were cloned into pBridge. PRAF-3 expression was induced when methionine was dropout. -TLH, medium lacking Trp, Leu, and His; -MTLH, medium lacking Met, Trp, Leu, and His. (k) GDI-1 and PRAF-3 competitively combined with RAB-1 was examined by GST pulldown assays. Different concentrations of PRAF-3 were added to the system in which GDI-1 interacted with RAB-1, and RAB-1 and PRAF-3 were detected under the pull-down of GDI-1 using IB. From left to right lanes, PRAF-3 was added in amounts of 0, 1, 2, 3, 4, and 5 µg. The two-tailed unpaired t test was used for statistical comparison of two groups of samples. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, ***P < 0.001, NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images and the Coomassie Brilliant Blue-stained protein gels underlying this Figure can be found in S1 Raw Images.
Subsequent co-IP assays in HEK 293T cells expressing each respective candidate revealed that RAB-1 and RAB-Y1 could interact with GDI-1 (S7g–S7i Fig). GST pull-down assays showed that only recombinant RAB-1 protein could be pulled down by GST-GDI-1, and not by GST alone (Figs 5c and S7j, S7k), indicating that RAB-1 directly interacted with GDI-1. In addition, fluorescence microscopy showed that germline cell corpses of rab-1 RNAi knockdown worms expressing LAAT-1::GFP, NUC-1::mCHERRY, or CPL-1::mChoint had significantly decreased signal from these markers around phagosomes (Fig 5d), similar to the results of gdi-1 knockdown (Fig 4h). In addition, knockdown of rab-1 in gei-17 mutants did not significantly increase the number of germ cell corpses compared with N2/rab-1 RNAi and gei-17 mutants (Fig 5e), suggesting that gei-17 and rab-1 functioned in the same genetic pathway. These results collectively suggested that GDI-1 SUMOylation could promote phagosome maturation in apoptotic cell clearance via interaction with RAB-1.
As membrane-bound GDI displacement factors (GDFs) mediate RAB-GDP dissociation from the RAB-GDP-GDI complex to catalyze Rab insertion into cognate membranes, we therefore tested whether PRAF-3, the only C. elegans ortholog of mammalian GDFs, also participated in apoptotic cell clearance. We found that RNAi suppression of praf-3 led to an age-dependent increase in cell corpse numbers in germline cells compared to that in vector controls (Fig 5f). To test whether praf-3 knockdown increased cell corpse number through excessive apoptosis or defective phagosomal activity, we quantified cell corpses under praf-3 knockdown in transgenic Pced-1rde-1; rde-1 worms (with RNAi suppression of phagosome formation) and Pegl-1rde-1; rde-1 worms (with RNAi suppression of apoptosis). We observed praf-3 knockdown led to increased cell corpse abundance in Pced-1rde-1; rde-1, but not Pegl-1rde-1; rde-1 worms (S7l, S7m Fig), suggesting that praf-3 functioned in cell corpse clearance rather than induction of apoptosis. In addition, four-dimensional microcopy analysis of the timespan for germ cell corpse persistence showed that germ cell corpse degradation required an average of 36.5 ± 2.8 min in WT animals, whereas cell corpse degradation required 112.8 ± 4.1 min in praf-3 knockdown gonads (Fig 5g).
Subsequent GST pull-downs showed that recombinant HA::PRAF-3 could be immunoprecipitated with GST-GDI-1, but not with GST alone (Fig 5h), indicating that PRAF-3 could directly interact with GDI-1 in vitro. Furthermore, GST pull-downs and co-IP assays with RAB-1 showed that HA::PRAF-3 could also be immunoprecipitated by GST::RAB-1 or MYC::RAB-1, respectively (Figs 5h and S5n). As PRAF-3, GDI-1, and RAB-1 appeared to directly interact with each other, we also examined whether GDI-1 competed with PRAF-3 for binding with RAB-1 using Yeast 3-Hybrid (Y3H) and GST pull-down assays. We found that PRAF-3 could disrupt GDI-1 interaction with RAB-1 (Fig 5j) in Y3H assays. Similarly, adding increasing concentrations of PRAF-3 to GDI-1 incubations with RAB-1 resulted in a dose-dependent decrease in RAB-1–GDI-1 interactions (Fig 5k). These data suggested that PRAF-3 could compete with GDI-1 to bind RAB-1 to modulate apoptotic cell clearance in C. elegans.
GDI-1 SUMOylation regulates RAB-1-mediated vesicular transport of RAB-7 from ER to Golgi
Given the above direct GDI-1-RAB-1 interactions, we next investigated the function of GDI-1 SUMOylation in regulating its interactions with RAB-1. First, WB analysis of RAB-1 protein contents in gdi-1(K270R) mutants showed that total RAB-1 levels did not significantly differ from that in N2 (Fig 6a,6b). However, isolation of the plasma membrane and cytoplasmic protein fractions in N2, gdi-1(K270R), and praf-3(xwh92) mutants followed by WB detection of endogenous RAB-1 revealed that RAB-1 levels were significantly increased in the cytoplasmic fraction and concomitantly decreased in the membrane fractions from gdi-1(K270R) and praf-3 mutants (Fig 6c–6e). These results suggested that SUMO modification of GDI-1 affected the intracellular localization of RAB-1. However, GST pull-down assays showed no significant difference in recombinant MYC::RAB-1 levels pulled down by GST-GDI-1 versus GST-GDI-1(K270R) (S8a, S8b Fig). As in vitro GST pull-downs might not accurately reflect the function of SUMOylation in regulating protein-protein interactions in vivo, we therefore conducted co-IP assays in gdi-1(K270R) and gei-17(xwh81) mutants with impaired GDI-1 SUMOylation. Subsequent WB detection of RAB-1 co-IP extracts of GD-1 showed higher levels of GDI-1-RAB-1 interaction in the gdi-1(K270R) and gei-17(xwh81) mutants compared to that in N2 (Fig 6f, 6g). Similar results were obtained with Flag-tagged GDI-1 in flag::gdi-1(K270R) and gei-17; flag-gdi-1 worm lysates (S8c, S8d Fig). Based on these results, we hypothesized that unSUMOylated GDI-1 may undergo stronger interactions with GDP-bound RAB-1 in vivo, maintaining RAB-1 in the inactive state.
Fig 6. GDI-1 SUMOylation regulates RAB-1-mediated vesicular transport of RAB-7 from ER to Golgi.
(a) Endogenous RAB-1 protein levels were examined by immunoblotting in N2 and gdi-1(K270R) worms. (b) Graphs showing the quantification of the protein level of GDI-1 in N2 and gdi-1(K270R) worms. (c) Endogenous RAB-1 protein levels were examined by immunoblot analysis in N2, gdi-1(K270R), and praf-3 plasmalemma or cytoplasm. SCA-1 is a plasmalemma marker and TBB-1 is a cytoplasmic marker. (d and e) The graphs show quantification of the protein level of RAB-1 in the plasmalemma (d) and cytoplasm (e) using ImageJ software. (f) GDI-1 IP was performed, followed by detection of the interaction intensity between GDI-1 and RAB-1 in N2, gei-17, and gdi-1(K270R) worms with an anti-RAB-1 antibody. (g) The graph shows the protein level of RAB-1/GDI-1. The ratio of RAB-1 to GDI-1 was determined and normalized to 1-fold in N2. (h) RAB-1 IP was performed, followed by detection of the interaction between RAB-1 and RAB-7 in N2, qxIs68 (Pced-1mcherry::rab-7) worms with anti-mCHERRY antibody. An isotype-matched mouse IgG was used as the control. (i) Colocalization of GFP::RAB-7 and mCHERRY::RAB-1 in N2 embryos. Arrows indicate cell corpses. The boxed regions are magnified (2×) in the inset. Bars, 2 µm. (j) Quantification graph of GFP::RAB-7 and mCHERRY::RAB-1 colocalization on cell corpses in control and gei-17 RNAi embryos. Percentage refers to the ratio of mCHERRY::RAB-1 to GFP::RAB-7. (k and m) Endogenous RAB-7 protein levels were examined by immunoblotting in N2/gdi-1(K270R)/praf-3 (k) and control/rab-1 RNAi (m) worms. (l and n) Graphs show the quantification of the protein level of RAB-7 in N2/gdi-1(K270R)/praf-3 (l) and control/rab-1 RNAi (n) worms. (o) RAB-1 IP was performed, followed by the detection of the interaction intensity between RAB-1 and RAB-7 in qxIs66 (Pced-1gfp::rab-7) and gdi-1(K270R); qxIs66 (Pced-1gfp::rab-7) worms with an anti-GFP antibody. An isotype-matched mouse IgG was used as the control. (p) The graph shows protein levels of GFP::RAB-7/RAB-1. The ratio of GFP::RAB-7 to RAB-1 was determined and normalized 1-fold in qxIs66. (q) Endogenous RAB-7 protein levels were examined by immunoblot analysis in N2, gdi-1(K270R), and praf-3 in endoplasmic reticulum (ER) or Golgi apparatus. CNX-1 is an ER marker, SYX-6 is a Golgi marker, TBB-1 is a cytoskeleton marker, and Histon H3 is a nucleus marker. (r) The graph shows quantification of the protein level of RAB-7 in the ER and Golgi apparatus using ImageJ software. The unpaired t test was performed in this figure. *P < 0.05, **P < 0.01, ***P < 0.001. NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images underlying this Figure can be found in S1 Raw Images.
As RAB1 can reportedly interact with RAB7 in Drosophila [46], we also tested whether RAB-1 could interact with RAB-7 in C. elegans. We found that GST-RAB-7, but not GST, could pull down recombinant MYC::RAB-1 in vitro, indicating that RAB-7 could directly interact with RAB-1 (S8e Fig). In addition, co-IP assays in the transgenic qxIs68(Pced-1mcherry::rab-7) strain, expressing an mCHERRY::RAB-7 fusion reporter driven by the ced-1 promoter, showed that RAB-1 could be co-precipitated with RAB-7 from crude extracts using mCHERRY antibodies (Fig 6h). Fluorescence microscopy further revealed that mCHERRY::RAB-1 co-localized with a GFP::RAB-7 reporter on the surface of cell corpses in C. elegans (Fig 6i), and this co-localization increased in worms with gei-17 knockdown (Fig 6j). Furthermore, co-IP assays in HEK 293T cells also indicated that RAB-1 could interact with RAB-7 (S8f Fig). These results demonstrated that RAB-7 could indeed interact with RAB-1 in C. elegans.
To further investigate the mechanism through which SUMO modification of GDI-1 might regulate RAB-1 interaction with RAB-7, we performed GST pull-down assays and found that HA::RAB-7 protein could not be pulled down by GST-GDI-1 (S8g Fig), indicating that RAB-7 could not directly interact with GDI-1. Interestingly, endogenous RAB-7 levels were significantly increased in gdi-1(K270R), praf-3(xwh92), and rab-1 knockdown worms compared to that in controls (Fig 6k–6n). Moreover, exogenous GFP::RAB-7 expression was also significantly increased in both gdi-1(K270R); Pced-1gfp::rab-7 and rab-1 RNAi-treated Pced-1gfp::rab-7 worms (S8h–S8k Fig). WB analysis of RAB-7 contents immunoprecipitated by anti-RAB-1 antibody indicated that RAB-1 interaction with RAB-7 was reduced in qxIs66; gdi-1 (K270R) (Fig 6o, 6p), which might be due to stronger GDI-1 interaction with RAB-1 in the cytoplasm, where it lacks SUMO modification, consequently blocking RAB-1 binding with RAB-7. Fluorescence microscopy showed no significant difference in the percentage of mCHERRY::RAB-1 signal localized to the surface of cell corpses between Pced-1mcherry::rab-1; gdi-1 (K270R) and Pced-1mcherry::rab-1 worms (S8l Fig), whereas the percentage of mCHERRY::RAB-1 on the cell corpse surface was significantly lower in rab-7 RNAi-treated Pced-1mcherry::rab-1 animals compared with Pced-1mcherry::rab-1 controls (S8m Fig). These results suggested that RAB-7 may function downstream of RAB-1 activation in cell corpse clearance.
As RAB-7 is known to mediate lysosome recruitment and subsequent fusion to phagosomes in phagolysosome formation, we therefore investigated whether RAB-7 was degraded via the endo-lysosomal pathway. To this end, we blocked lysosomal acidification required for its enzymatic function by pretreating Pced-1gfp:: rab-7 worms with chloroquine (Chl). Western blots indicated that endogenous RAB-7 contents significantly increased compared to vehicle controls (S8n, S8o Fig). As the ubiquitin-proteasome pathway is the other major protein degradation pathway in eukaryotic cells, we also chemically inhibited this pathway with MG-132 and observed a similar increase in RAB-7 accumulation (S8n, S8o Fig). These results suggested that RAB-7 levels could be simultaneously modulated through both lysosome-mediated and ubiquitin-proteasome-mediated degradation. We then pretreated gdi-1(K270R); Pced-1gfp::rab-7 worms with Chl and MG-132, individually, and found that endogenous RAB-7 accumulated to high levels in the MG-132, but not in the Chl group (S8p, S8q Fig), indicating that the mutations in the SUMOylation sites of GDI-1 prevent RAB-7 degradation via the lysosomal pathway, forcing its degradation exclusively through the proteasomal pathway.
As mammalian Rab1A and Rab1B are predominantly found on the ER membrane and Golgi apparatus and function redundantly in vesicular transport from ER to Golgi, we next examined whether RAB-1 regulates vesicular transport of RAB-7 between the ER and successive Golgi compartments by isolating the ER and Golgi protein fractions from N2, gdi-1(K270R), and praf-3(xwh92) mutants. WB analysis showed that endogenous RAB-7 levels significantly increased in ER with a concomitant decrease in Golgi of gdi-1(K270R) and praf-3 mutants (Fig 6q, 6r), suggesting that SUMO modification of GDI-1 affects RAB-7 localization. Furthermore, we found that the same results were obtained when we performed gdi-1 RNAi in wild-type N2: RAB-7 protein levels increased in the ER and decreased in the Golgi (S9a, S9b Fig). However, when we performed gdi-1 RNAi in the gdi-1(K270R) background, no significant changes in RAB-7 protein levels were observed in either the ER or Golgi (S9c, S9d Fig). These results indicate that GDI-1 SUMOylation was a critical step in mediating RAB-7 transport. RAB-7 transport from the ER to the Golgi apparatus was inhibited in gdi-1(K270R) mutant worms. However, RNAi of smo-1 and gdi-1 did not affect phagosomes labeled with GFP::RAB-7 (Figs 2g and 4h). Thus, we propose that the level of RAB-7 on phagosomal surfaces was decreased in gdi-1(K270R) mutant worms. To test this hypothesis, we overexpressed Pced-1sfGFP1-10::rab-7 and Pced-1vps-41::sfGFP11 in wild-type nematodes and gdi-1(K270R) mutants. VPS-41 is a subunit of the well-known membrane tethering factor HOPS complex and functions as an effector protein for RAB-7 to promote phagosome-lysosome membrane fusion [47]. Our fluorescence intensity measurements of phagosomes containing apoptotic cells in gdi-1(K270R) mutants revealed significantly reduced fluorescence compared to that in wild-type (S9e, S9f Fig), indicating that the level of RAB-7 on phagosomal surfaces was decreased in gdi-1(K270R) mutants. However, we cannot completely exclude the possibility that phagosome recruitment of RAB-7 was unaffected. Additionally, we overexpressed recombinant mCHERRY::RAB7A in HEK293T cells with concurrent rab1a knockdown by shRNA. Upon extracting ER and Golgi, WB detection of mCHERRY indicated that RAB7A protein levels significantly increased in ER and again showed a concomitant decrease in Golgi under rab1a knockdown in HEK293T cells (S9g–S9k Fig). These cumulative results demonstrated that RAB1 mediates the vesicular transport of RAB-7 from ER to Golgi to promote apoptotic cell clearance.
SUMOylation of GDI1 modulates efferocytosis in mammals
Given the above evidence that GDI-1 SUMOylation by GEI-17 is required for its regulation of RAB-1 in C. elegans, we next investigated whether these regulatory interactions also occur in mammals. Co-IP assays in HEK293T cells co-expressing an HA-tagged human homolog of GDI-1, HA::hGDI1, a MYC-tagged human homolog of GEI-17, MYC::PIAS1, and recombinant human MYC::RAB1A, indicated that HA::hGDI-1 could interact with both recombinant MYC::PIAS1 and MYC::RAB1A (Fig 7a, 7b). Subsequent GST pull-down assays showed that the recombinant MYC::PIAS1 or MYC::RAB1A proteins could be pulled down by GST-hGDI1, but not by GST alone (S10a, S10b Fig), indicating that hGDI1 could directly interact with both PIAS1 and RAB1A in vitro. Interestingly, co-IP assays in HEK293T cells showed that hGDI1 could interact with C. elegans proteins, GEI-17 and RAB-1, while GDI-1 could interact with their human homologs, PIAS1 and RAB1A (S10c–S10f Fig), suggesting that these interactions were likely conserved between worms and humans.
Fig 7. SUMOylation of GDI1 modulates degradation in mammals.
(a andb) The interactions between hGDI1 and PIAS1 (a) and hGDI1 and RAB1A (b) were examined by CO-IP in HEK293T cells. (c) SUMOylation of the wild-type and 269 lysine-to-arginine (K270R) mutant forms of hGDI1. The SUMOylated-hGDI1 band is indicated by an arrowhead. (d) Phagocytosis rates were quantified in control, gdi1, and pias1 shRNA at 3 h. (e) Time course of apoptotic cell clearance in RAW264.7 phagocytes. Degradation rates were quantified in control, gdi1, and pias1 shRNA cells at 3, 6, and 9 h. (f) Representative images macrophages expressing the indicated vectors were incubated with dual-color apoptotic CharON Jurkat reporters (green: apoptosis-induced, acid-sensitive; red: acid-resistant). (g) Western blot analysis was performed to examine the expression of HA::GDI1 in the four experimental groups shown in Fig f. (h) The graph shows quantification of the degradation rate in Fig f. Data are representative of at least three independent experiments. For each experiment, five random imaging fields were analyzed. The unpaired t test was performed in this figure. ***P < 0.001. NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images underlying this Figure can be found in S1 Raw Images.
The above results indicate that the interactions of GDI-1 with GEI-17 and RAB-1 were evolutionarily conserved. Subsequently, we investigated whether mammalian GDI1 was modified by SUMOylation and whether the SUMOylation site of GDI-1 in C. elegans was conserved. Sequence comparison revealed that the amino acid sequence of GDI-1 is conserved in C. elegans, mice, and humans, where the lysine at position 270 of GDI-1 corresponds to the amino acid at position 269 of murine and human proteins (S10g Fig). Moreover, FLAG IP assays in HEK293T cells expressing HA::hSUMO1/2/3 and FLAG::hGDI1(WT and K269R) showed that hGDI1 was SUMOylated, whereas hGDI1(K269R) was not (Fig 7c). These findings suggest that hGDI1 was SUMOylated via lysine at position 269.
To determine whether SUMO modification plays a role in the degradation of apoptotic cells in mammals, we first tested the efficiency of shRNA (pGreen-puro) knockdown of pias1 and gdi1 in RAW264.7 macrophage cells by qPCR (S10h, S10i Fig).
We then introduced a stable Jurkat T cell line that expressed caspase-activated GFP and pH-insensitive mApple (CharON) [48] to trace apoptosis and efferocytosis, as the acid-resistant red fluorescence reliably labels all apoptotic cells, whereas the acid-sensitive green fluorescence is quenched upon lysosomal degradation, allowing a clear distinction between newly internalized (red⁺/green⁺) and fully degraded (red⁺/green⁻) apoptotic cells. After incubating UV-induced apoptotic Jurkat cells expressing the genetically encoded fluorescent reporter CharON for 2 h, we observed that green fluorescence was significantly reduced in the pias1 and gdi1 knockdown groups compared to the control group (Figs 7d,7e and S11a–S11c), indicating pias1 and gdi1 deficiencies impair apoptotic cell degradation. Degradation experiments with the CharON system demonstrated that wild-type GDI1 fully restored the defective degradation of apoptotic cells caused by gdi1-deficient cells, whereas the K269R mutant failed to restore this function (Fig 7f–7h).
Previous research has demonstrated that the release of metabolites resulting from the degradation of apoptotic cells modulates metabolic signaling in macrophages [49,50], thereby facilitating the resolution of inflammation and promoting continuous efferocytosis. To assess whether SUMO modification affected phagocytic efficiency, we incubated pias1 and gdi1 knockdown macrophages with Cell Tracker-labeled apoptotic cells for 6 hours and quantified the ratio of cell tracker-positive cells to total GFP-positive cells. This analysis revealed that efferocytosis was significantly reduced in RAW264.7, J774A.1, and BMDM cells with pias1 and gdi1 knockdown compared to that in controls (S12a–S12h Fig). Flow cytometry further illustrated this significant decrease in efferocytosis efficiency under pias1 and gdi1 knockdown (S13a–S13c and S14a Figs). To assess whether the SUMOylation of GDI1 affected phagocytic efficiency, we then overexpressed GDI1(WT and K269R) in the gdi1 knockdown background. Flow cytometric analysis revealed that while wild-type GDI1 successfully restored efferocytosis in gdi1-deficient cells, the K269R mutant failed to do so (S14b Fig). Collectively, these results demonstrate that GDI1 SUMOylation plays a crucial role in the regulation of efferocytosis in mammals.
Discussion
Ubiquitin and the ubiquitin-related modifier SUMO function in various aspects of the cellular stress response, such as targeting misfolded proteins for degradation, regulating transcription, and facilitating recovery from stress [22,51]. Previous studies have demonstrated that ubiquitination modifications are crucial for apoptotic cell clearance [26,52,53]. However, whether SUMO modifications are involved in apoptotic cell clearance remains unclear, although it has been established that SUMO modifications are implicated in apoptotic cell death in mammals [21]. Here, we found that the SUMO precursor molecule SMO-1, the E1 activator AOS-1/UBA-2, the E2 conjugating enzyme UBC-9, and the E3 ligase homologs GEI-17/MMS-21 are involved in the clearance of physiological apoptotic cells rather than in excessive apoptosis by using C. elegans, the suitable model organism to dissect SUMO modification. SUMOylation has been predominantly investigated in the development as well as in DNA damage response in C. elegans [34]. To the best of our knowledge, we provide the first evidence that SUMO modification regulates apoptotic cell clearance in C. elegans. We found that the two E3 ligases, mms-21 and gei-17, are most likely not in the same genetic pathway for apoptotic cell clearance. We focus here on the function of GEI-17 in apoptotic cell clearance. A further study with more focus on the function of mms-21 in apoptotic cell clearance is therefore suggested.
Previous proteomic analyses of SUMOylated worm proteins have relied on smo-1 transgenes that encode HIS-tagged SUMO fused to the FLAG epitope or to GFP [32,43]. Kim H and colleagues inserted HIS10 sequences directly after the initiator methionine codon of smo-1 without any other sequences, which is fully functional and can be utilized in conjunction with mutagenesis studies to identify specific SUMO acceptor lysines in substrate proteins [35,54]. Employing this affinity-based approach and mass spectrometry, they identified hundreds of C. elegans proteins that were likely to be direct SUMO conjugates. In the present study, we identified 188 potentially SUMO interactors (noncovalent) or SUMO-modified proteins in C. elegans under normal development conditions (S2 Table). This finding represents a significant expansion of the known SUMO regulatory network in this model organism. Furthermore, it was confirmed that lysine 270 of GDI-1 is an essential site for SUMO modification without affecting its protein levels, and that mutation at this site impairs degradation in apoptotic cells (Fig 8).
Fig 8. Schematic summary of SUMOylation of GDI-1 promotes apoptotic cell clearance.
During engulfment, GDI-1-K270 is SUMOylated, SUMOylated GDI-1 promotes the release of RAB-1-GDP, and PRAF-3 converts RAB-1-GDP to RAB-1-GTP. The active form of RAB-1 aids in the vesicular transport of RAB-7 from the endoplasmic reticulum to the Golgi, and ultimately RAB-7 localizes to the phagosome, promoting phagosome maturation and recruitment of lysosomes, which in turn promotes phagosomal degradation. In the absence of SUMOylation, GDI-1 is unable to SUMOylate, leaving RAB-1 in its GDP form, and RAB-7 cannot be transported from the endoplasmic reticulum (ER) to the Golgi, excessively accumulating in the ER, and consequently suppressing phagosomal degradation.
GDI binds prenylated GDP-conjugated Rabs in the cytosol or membrane, mediating the transport of GDP-Rabs or their recovery from the target membrane [55–57]. The Rab family of proteins is a crucial regulator of membrane trafficking and plays a significant role in numerous cellular processes, including phagocytosis. Research has demonstrated that RAB-5 regulates the fusion of early endosomes, RAB-2 and RAB-14 regulate the maturation of phagosomes in parallel, and RAB-7 mediates the maturation of early endosomes to late endosomes and the recruitment and fusion of lysosomes during the degradation of apoptotic cells in C. elegans [58,59]. Our investigation revealed that GDI-1 interacts with RAB-1 and that knockdown of RAB-1 results in impaired phagosome recruitment to the lysosome, suggesting that RAB-1 and SUMO modification function in the same genetic pathway for apoptotic cell clearance. PRAF-3 (GDI displacement factor GDF in C. elegans) interacts with both GDI-1 and RAB-1, and GDI-1 binds RAB-1 competitively with PRAF-3. In the present study, we identified a novel mechanism that mutations in PRAF-3 affect the clearance of apoptotic cells. Notably, we demonstrated that PRAF-3 and SUMO modification of GDI-1 function in the same genetic pathway for the clearance of apoptotic cells in C. elegans (Fig 8).
RAB1 is located at the endoplasmic reticulum exit site (ERES) and pre-Golgi intermediate (pre-GIC), and mediates ER-Golgi transport [59]. Here, we found that SUMOylation of GDI-1 affects the localization of RAB-1, increasing its cytoplasmic localization and decreasing its membrane localization, and that impaired SUMOylation of GDI-1 leads to GDI-1 binding to more RAB-1. SUMOylation of GDI-1 appears to aid in the release of the GDP form of RAB-1, facilitating its membrane localization in its active form. Drosophila protein interaction mapping has revealed a possible interaction between RAB1 and RAB7 [46]. Previous research has demonstrated that SUMOylated Rab7 is recruited to bacterial phagosomes via Sulf, that overexpression of Rab7 promotes intracellular growth of strain Lp02rpsLWT in BMDMs, and that Legionella pneumophilus utilizes a lysosomal network for phagosome biogenesis in BMDMs [60]. Another study revealed that blockade of RAB7 SUMOylation by S. Typhimurium ensures the availability of long-lived but functionally compromised RAB7, which is beneficial to the pathogen. These studies demonstrated that SUMO modification of RAB7 exhibits distinct functions for different pathogens [61]. However, RAB-7 was not identified as a SUMOylation substrate in our and other similar proteomics studies in C. elegans, which may be attributed to species specificity. An alternative explanation could be that phagosomes containing bacteria differ from those containing apoptotic cells, and consequently, RAB7 functions differently in these two processes. The different mechanisms of SUMOylation of RAB7 under phagocytosed bacteria and apoptotic cells can be explored subsequently. To our knowledge, we discovered a novel mechanism that, in C. elegans, RAB-1 interacts directly with RAB-7, and both RAB-1 knockdown and impaired SUMOylation of GDI-1 led to an increase in the protein content of RAB-7. Aberrant SUMO modification of GDI-1 reduced the interaction between RAB-1 and RAB-7. In addition, when the SUMOylation site of GDI-1 was mutated and PRAF-3 was deleted, the localization of RAB-7 in the ER increased, whereas localization in the Golgi apparatus decreased. Although defects in SUMOylation did not affect phagosomes labeled with GFP::RAB-7, we found reduced binding of RAB-7 to VPS-41 in gdi-1(K270R) mutants, suggesting that the level of RAB-7 on phagosomal surfaces decreased. However, we cannot entirely rule out the possibility that the recruitment of RAB-7 to phagosomes remained unaffected. Normally, RAB-7 can be degraded by lysosomal and proteasomal pathways; however, RAB-7 is degraded only via the proteasomal pathway when the SUMOylation site of GDI-1 is mutated. Based on our findings, we propose that during apoptosis in C. elegans, GDI-1 is modified by SUMOylation, which promotes the release of the GDP form of RAB-1. Subsequently, in the presence of PRAF-3, RAB-1-GDP is converted to the GTP form, which then associates with RAB-7, thereby mediating vesicular transport of RAB-7 from the ER to the Golgi apparatus to promote the clearance of apoptotic cells (Fig 8). Our findings demonstrate that SUMO modification of GDI-1 plays a role in apoptotic clearance by regulating RAB-7 through interactions with RAB-1. This study represents the first identification of a function for SUMO modification of GDI-1 in an invertebrate, suggesting the possibility of similar regulatory mechanisms in other organisms.
In mammals, GDI1 is expressed predominantly in neural and sensory tissues, participates in vesicular transport, and may serve as a crucial regulator of several signaling pathways that govern synaptic plasticity, learning acquisition, and memory formation [62]. Furthermore, a 192P mutation in GDI1 results in X-linked nonsyndromic mental retardation in humans [63]. GDI1 expression is elevated in the cerebrospinal fluid of patients with the neurodegenerative disorder Alzheimer's disease (AD), and inhibition of GDI1 expression attenuates amyloid (Aβ)-induced neurotoxicity in AD [64,65]. We observed that hGDI1 interacted with PIAS1 and GEI-17, and GDI-1 interacted with PIAS1 and GEI-17. Knockdown of pias1 in macrophages resulted in a phagocytosis-deficient phenotype, consistent with that of C. elegans. The human genome contains more than 60 Rabs. Aberrant Rab function leads to abnormal distribution of subcellular structures, resulting in neurological disorders, inflammatory diseases, and tumorigenesis. Notably, hGDI1 can interact with RAB1A and RAB-1, and there is also an interactive relationship between GDI-1 and RAB1A and RAB-1. Additionally, the ER and Golgi localization of RAB7A in mammalian cells is regulated by RAB1A. Rab1 in mammals has significant functions in signaling, cell migration, and cell surface receptor presentation, and its dysregulated expression is closely associated with human diseases such as cancer, cardiomyopathy, and Parkinson's disease [66,67]. Given that hGDI1 interacts with RAB1A in mammals and that RAB7A is also regulated by RAB1A, we determined through sequence comparison analysis that the lysine at position 270 of GDI-1 is relatively conserved in mammalian hGDI1, and the lysine at position 269 of hGDI1 mediated its SUMOylation. The interactions and related functions of these genes are relatively conserved in mammals; therefore, elucidating the regulatory mechanism of SUMO modification of GDI-1 will have significant implications for understanding its role in the pathogenesis of neurodegenerative diseases such as AD, in addition to mechanistically revealing its role in apoptotic cell clearance.
In conclusion, we have provided evidence for the regulatory role of GDI-1 SUMOylation in apoptotic cell clearance in C. elegans and revealed a previously uncharacterized function of SUMO modification, which expands the known roles of this significant post-translational protein modification.
Methods
Ethics statement
All animal care and experiments were performed according to the “Guide for the Care and Use of Laboratory Animals,” which was approved by the Academic Committee of Shaanxi Normal University (Approval No. 202412048), and all manipulations were conducted in accordance with established guidelines.
C. elegans strains and genetics
The wild-type strain of C. elegans used in this study, as well as the strain used for genetics, was obtained from the Bristol strain N2, which was cultured on plates of Nematode Growth Medium (NGM) containing OP50 at 20 °C. Genetic manipulations were based on Brenner's research paper “The genetics of Caenorhabditis elegans” published in the journal Genetics in 1974 [68]. Nematode strains registered in our laboratory at wormbase were designated SNU (strain) and xwh (allele). The mutant alleles used in this study were as follows: linkage group (LG) Ⅰ: ced-1(e1735), ced-12(n3261), smo-1(ok359), and gei-17(xwh81); LG Ⅱ: mms-21(xwh12); LG Ⅲ: ced-4(n1162), and ced-6(n1813); LG Ⅳ: ced-2(n1994), ced-3(n717), ced-5(n1812), gdi-1(K270R) (xwh89), and praf-3(xwh92). The extrachromosomally inherited worm strains obtained by microinjection were subjected to UV irradiation to obtain stable genetically inherited nematode strains. All strains used in this study are listed in the S4 Table.
Cell lines
Human embryonic kidney 293T (HEK293T) and Mouse L929 cells were obtained from FuHeng Cell Center (Shanghai). Human Jurkat (Clone E6-1), Mouse J774A.1, and RAW264.7 cells were obtained from YiZeFeng Biotechnology (Shanghai) Co., LTD. All cell lines were cultured in DMEM (supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin liquid (100×)), except Jurkat cells, which were cultured in RPMI1640 medium (supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin liquid (100×)). The cells were then incubated at 37 °C with 5% CO2. All cell lines used in this study are listed in the S4 Table.
Mouse
Animal protocols were approved by the Academic Committee of Shaanxi Normal University. All mice were cared for according to the Shaanxi Normal University guidelines for the care and use of laboratory animals, and were in good general health based on appearance and activity. The strains of mice used in the experiments were Balb/c and C57BL/6J. The experimental mouse strains were kept in the Animal Experiment Center of Shaanxi Normal University according to standard methods.
RNAi experiments
RNA interference (RNAi) refers to the introduction of double-stranded RNA (dsRNA) into hermaphroditic C. elegans, which leads to sequence-specific and rapid degradation of endogenous mRNAs in order to knock down gene expression. The expression of target genes can be effectively blocked by RNAi and gene silencing can be transmitted between tissues throughout the adult and its progeny. We used a food-fed RNAi approach [13], which was performed by feeding E. coli HT115 with the corresponding RNAi vector for worm target genes. For genetic analysis, 10–20 adult worms were placed on NGM plates containing 1 mM isopropyl β-d-thiogalactopyranoside (IPTG) and 100 µg/mL ampicillin containing RNAi bacteria, and the apoptotic phenotypes of their progeny were counted. For biochemical experiments, 50–100 worms at different periods were placed on NGM plates containing RNAi bacteria, and the worms were collected with M9 buffer after 48 h. Proteins were extracted for subsequent experiments.
Quantification of embryonic and germline cell corpses
A control group was required for each RNAi treatment to confirm the achievement of RNAi effect and the confidence of the data. Three biological replications were performed for each RNAi treatment. The number of "button-like" shaped apoptotic cells in the head region of living embryos was counted using Nomarski DIC optics at different developmental stages (comma, 1.5 Fold, 2 Fold, 2.5 Fold, 3 Fold, and 4 Fold), with 15 embryos per nematode strain counted at each developmental stage. To determine the number of apoptotic germ cells, apoptotic cells were counted in the meiotic region of one gonadal arm of at least 15 nematodes at 12, 24, 36, 48, and 60 h after the L4 stage. The average number of embryonic and germline cell corpses was compared with those of other transgenic worms using an unpaired t test.
CRISPR/Cas9-mediated genome editing
The Cas9 editing involved in this project is a more mature method studied by previous researchers, in which the worm genome is edited to obtain specific allele knockout mutants, in situ insertions, or fixed-point mutations in C. elegans [69]. Briefly, target gene sgRNAs were analyzed at http://crispr.mit.edu. 1–3 sgRNAs were selected to ensure the knockdown or knock-in efficiency. The sgRNA sequences were introduced by PCR into the Cas9-expressing vector pDD162-peft-3-cas9-sgRNA empty. dpy-10 sgRNA plasmid 20 ng/μl, target sgRNA plasmid 20 ng/μl, target ssODN 2 μM, dpy- 10 ssODN 2 μM co-injected into 50 worms. The progeny were then identified by single-worm PCR to identify which were successfully edited. The target sgRNA and the repairing oligo sequences used in this study are listed in the S3 Table.
Time-lapse imaging and microscopy
The method was referred to the previous article with modifications [70]. Worms were placed in M9 buffer containing 2 mM levamisole, covered with a coverslip, and sealed with beeswax to keep them in liquid. Thirty germ cell corpses were tracked and photographed in each experiment. Germ cell corpses were tracked using a ZEISS Imager M2 orthogonal fluorescence microscope and recordings were taken every 5–10 min. During the experiment, worms were constantly checked for survival to ensure data reliability. To record the duration of embryonic cell corpses, early embryos (2-cell stage) were placed in egg salt buffer (118 mM NaCl and 48 mM KCl) and mounted on slides with agar pads. Images in a 25-section Z-series (1 μm/section) were captured every 1 min for 400 min using a ZEISS Imager M2 orthogonal fluorescence microscope.
Worms tagged with fluorescence were imaged by DIC and fluorescence using an Axio Imager M2 microscope (ZEISS). Images were processed and viewed using ZEN 2 pro software (ZEISS). Immersol 518F oil (Zeiss) was used. All images were captured at 20 °C.
To monitor the uptake and degradation of cell corpse chromatins in phagosomes and secreted soluble (ss)GFP in coelomocytes, worms were observed using fluorescence microscopy at the indicated times. Secreted soluble (ss)GFP worms (arIs36) were heat-shocked for 1 h at 33 °C and continued to grow at 20 °C.
Quantification of phagosomal markers
Transgenic worms with fluorescent marker genes for each stage of phagosome (CED-1::GFP, GFP::RAB-5, GFP::RAB-7, mCHERRY::RAB-7, LAAT-1::GFP, LAAT-1::mCHERRY, NUC-1::mCHERRY, and CPL-1::mChOint) were utilized for the relevant indicated strains, and clinical slides were prepared by picking the treated worms on slides with agar. One hundred apoptotic cells from at least 15 worms were counted for each worm species (three biological replications were performed for each set of experiments), and the apoptotic cells were observed under the fluorescence of ZEISS Imager M2 microscope to determine whether the apoptotic cells were surrounded by fluorescent phagosome marker genes. Given that the apoptotic cells in the adult gonad of C. elegans measure approximately 5 μm in diameter, we performed a thorough quantification by counting all visible apoptotic cells throughout the entire gonad, ensuring that different focal planes were considered. The fluorescent-labeling ratio of phagosomal genes is the number of fluorescently labeled apoptotic cells divided by the total number of apoptotic cells.
To detect phagosomal markers around internalized apoptotic corpses, dual-reporter assays were performed as follows. Double-reporter transgenic worms were generated by crossing smIs34(Pced-1ced-1::gfp) with qxIs68(Pced-1mcherry::rab-7), qxIs352(Pced-1laat-1::mcherry), qxIs257(Pced-1nuc-1::mcherry), and yqEx620(Pced-1cpl-1::mChOint), respectively. Each strain was subjected to control RNAi and smo-1 RNAi. For each double-reporter strain, the proportion of CED-1::GFP-labeled germ cell corpses that were also labeled with the respective mCHERRY fusion marker (mCHERRY::RAB-7, LAAT-1::mCHERRY, NUC-1::mCHERRY, or CPL-1::mChOint) was quantified. At least 100 CED-1::GFP-labeled phagosomes were scored per experiment, and all experiments were repeated independently three times.
AO staining and pHrodo staining
To detect phagosomal acidification by AO, adult worms were incubated in M9 medium containing 0.1 mg/ml AO and a small amount of OP50 in the dark for 1 h at room temperature. After treatment, worms were transferred to NGM plates at room temperature in the dark for 2 h and mounted on 2% agar pads. Images were captured using DIC and fluorescence microscopy. At least 100 apoptotic cells were counted in each strain.
To detect phagosomal acidification by pHrodo Red AM dye (Thermo Fisher Scientific, Cat#P35372), The gonads of adult worms (36 h post-L4) were dissected, and the dissected worms were placed in pHrodo staining solution (1 ml M9 solution containing 10 μl pHrodo AM Ester dissolved in 100 μl PowerLoad) and incubated at room temperature for 15 min. After treatment, worms were mounted on 2% agar pads. Images were captured using DIC and fluorescence microscopy.
Immunoprecipitation and LC-MS/MS analysis
Referencing experiments on worm intracellular proteins [44,52], the collected unsynchronized wild-type and transgenic worms (N2, ha::smo-1, and ha::smo-1; Pced-1gdi-1::gfp) were resuspended with lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% sodium deoxycholate, 1% NP-40, 10% glycerol, 1× mixture of protease inhibitors [Complete, EDTA free], and 1 mM PMSF), homogenized completely with tissue grinders, and centrifuged at 20,000 × g at 4 °C for 30 min. 4 mg proteins were co-incubated with the Anti-HA Magnetic Beads (Thermo Scientific, Cat#88837) for 6 h at 4 °C under rotary mixing. The bound protein beads were rinsed four times with TBS wash buffer (50 mM Tris-HCl, pH 7.6, 300 mM NaCl). The beads were then resuspended in 50 µl 2 × SDS loading buffer and boiled for 5 min at 100 °C for western blotting.
After immunoprecipitation, the samples from three replicate experiments were mixed together and identified by mass spectrometry (HOOGEN Biological Co., Shanghai). Mass spectrometry was performed as previously described [26]. The mass spectrometry proteomics data have been deposited in the ProteomeXchange Consortium (https://proteomecentral.proteomexchange.org) via the iProX partner repository with the dataset identifier PXD084683.
Western blot analysis
HEK293T cells and worms were resuspended in lysis buffer containing a protease inhibitor cocktail and homogenized completely using a sonic crusher or tissue grinder. Homogenized samples were centrifuged at 20,000 × g for 15 min at 4 °C to remove debris and protein concentration was quantified by BCA method, then 30 µg samples were boiled for 5 min at 100 °C in 5× SDS loading buffer for further analyses. Proteins were loaded onto SDS-PAGE, transferred onto PVDF membranes, blocked with 5% skimmed milk, and incubated overnight with the primary antibodies (β-actin loaded control) at 4 °C. After washing with TBS buffer, probed with HRP-conjugated secondary antibodies, and developed with the Immobilon Western chemiluminescent HRP substrate (Millipore). Protein bands were visualized using a MiniChemi610 (SAGE creation). All antibodies used in this study are listed in the S4 Table.
Yeast two-hybrid analysis
Genes indicated for detecting protein interactions were cloned into the pGADT7 and pGBKT7 vectors. Y2HGold was cultured in liquid YPDA medium supplemented with 0.25 μg/ml kanamycin and 40% glucose, when the OD600 was shaken to 0.3–0.6, prepared for the Y2HGold chemically competent cell. 240 μL 50% PEG3350, 36 μL 1 M LiAC, 10 μL Salmon Testes DNA, and 0.7 μg pGADT7 and pGBKT7 plasmids were added sequentially into Y2HGold chemically competent cell. Transformed Y2HGold chemically competent cell were coated on SD-Leu-Trp medium for 3–5 days at 30 °C. Single colonies were transferred to SD-Leu-Trp liquid medium at 280 rpm overnight in a shaker at 30 °C, and the yeast solutions were dropped onto SD-Leu-Trp and SD-Leu-Trp-His plates to test for interactions.
Co-IP assay
Co-immunoprecipitation experiments were performed as previously described [26], with some modifications. To detect protein interactions, HEK 293T cells in a 10 cm plate were transfected with 8 µg of each plasmid and 48 µl 1 µg/µl PEI (Polysciences, 24765-1). 48 hours after transfection, the cells were lysed with cell lysis buffer (25 mM Tris-HCl, pH 7.6, 150 mM NaCl, 10 mM MgCl2, 1% NP-40, and EDTAfree protease inhibitors), sonicated for 15 s, and centrifuged at 20,000 × g at 4 °C for 20 min to remove debris. 2 mg total protein of was incubated with anti-HA (Thermo, 88837) or anti-C-MYC (Thermo, 88842) magnetic beads at 4 °C for 2–3 h. The precipitates were 1× SDS loading buffer with 5% 2-mercaptoethanol boiled for 5 min at 100 °C. The samples were analyzed by WB.
To detect the interactions of GDI-1 with RAB-1 and RAB-1 with RAB-7, unsynchronized worms raised on 10 NGM plates were harvested. The samples were resuspended in lysis buffer containing a protease inhibitor cocktail, homogenized completely with tissue grinders, and centrifuged at 20,000 × g at 4 °C for 30 min to remove debris for Co-IP. Supernatants were incubated with anti-FLAG M2 affinity gel (Sigma-Aldrich, A2220) or Protein A/G Mix Magnetic Beads (Millipore, LSKMAGAG) at 4 °C for 4−6 h. When used with Protein A/G Mix Magnetic Beads, lysates were first incubated with primary antibody (GDI-1 or RAB-1 polyclonal antibody) at 4 °C for 2−4 h. The precipitates were boiled in 1× SDS loading buffer with 5% 2-mercaptoethanol and analyzed using WB. Rabbit IgG (Rabbit IgG control Polyclonal antibody: proteinthch, 30000-0-AP) or mouse IgG (Mouse IgG: proteinthch, B900620) matched to the host species and isotype of the primary antibody were used as the controls.
Purification of recombinant proteins and GST pull down
Plasmids expressing each protein were generated using GST or HIS tags. The recombinant proteins GST, GST::ARL-8, GST::GDI-1, GST::GEI-17, GST::GDI-1(K269R), GST::RAB-1, GST::RAB-7, and GST::hGDI1 were expressed in E. coli BL21 strain induced by IPTG. GST-tagged proteins in BL21 were lysed in GST lysis buffer (25 mM Tris-HCl, pH 7.6, 150 mM NaCl, 0.5% NP-40) with 1 mM PMSF and 1 mg/ml lysozyme and sonicated until the lysates were clear. GST-tagged proteins were incubated with Glutathione Sepharose 4 B (GE Healthcare, 17-0756) beads for 2–3 h at 4 °C, washed three times with GST lysis buffer, and eluted with 20 mM or 30 mM glutathione reduced (AMERSCO, 0399). 6HIS::MYC::MMS-21::6HIS, 6HIS::MYC::ARL-8::6HIS, 6HIS::MYC::GDI-1::6HIS, 6HIS::MYC::GDI-1(1-150 aa)::6HIS, 6HIS::MYC::GDI-1(150-444 aa)::6HIS, 6HIS::MYC::RAB-1::6HIS, 6HIS::MYC:RRAB-28::6HIS, 6HIS::MYC::RABY1::6HIS, 6HIS::HA::RAB-7::6HIS, 6HIS::HA::PRAF-3::6HIS, 6HIS::MYC::PIAS1::6HIS, and 6HIS::MYC::RAB1A::6HIS were expressed in E. coli BL21 (DE3) strain induced by IPTG. HIS-tagged proteins were lysed using the same approach for the GST-tagged proteins. After sonication, the HIS-tagged proteins were incubated with Ni-NTA Superflow (QIAGEN) beads for 2–3 h at 4 °C and eluted with 10/30/100/150 mM imidazole (Merck) in GST lysis buffer. Purified proteins were stored in GST lysis buffer containing 20% glycerol at −80 °C.
GST pull-down experiments were performed as previously described [26] with some modifications. 2 µg purified GST-tagged proteins immobilized on Glutathione Sepharose 4 B beads were incubated with 0.5 µg HIS-tagged proteins in binding buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% NP-40) for 2 h at 4 °C. The resins were washed three times in washing buffer (25 mM Tris-HCl, pH 7.4, 300 mM NaCl, 0.1% NP-40), and the bound proteins were analyzed by WB. The image obtained by Coomassie Brilliant Blue (CBB) staining represents 2.5% of the mixed preparation of GST-tagged and His-tagged proteins without co-incubation, which was used as the input to indicate the loading amount of the sample.
SUMOylation of GDI-1 in vitro and in vivo
SUMOylation experiments in vitro were performed as previously described [31], with some modifications. To examine SUMOylation in vitro, 2 µg HA::GDI-1 or HA::GDI-1(K270R), 0.2 µg hSAE1/SAE2, 0.5 µg hUBC9, and 4 µg MYC::hSUMO1 were incubated in SUMOylation buffer (20 mM HEPES, pH 7.4, 100 mM NaCl, 5 mM MgCl2, 2 mM ATP) at 37 °C at the indicated time points to allow formation of soluble oligomers. The reactions with the control and HA::GDI-1(K270R) were incubated for 1 h at 37 °C.
To confirm the SUMOylation of GDI-1 in vivo, unsynchronized worms (ha::smo-1, flag::gdi-1, ha::smo-1; flag::gdi-1, flag::gdi-1(K270R), and ha::smo-1; flag::gdi-1(K270R)) were collected and lysed in lysis buffer (25 mM Tris-HCl, pH 7.4, 150 mM NaCl, 0.1% sodium deoxycholate, 1% NP-40, 10% glycerol) containing 1× mixture of protease inhibitors [Complete, EDTA free] and 20 mM N-Ethylmaleimide (Sigma–Aldrich, E3876). Protein (4 mg) was immobilized on an anti-FLAG M2 affinity gel at 4 °C for 4 h, washed three times with washing buffer (25 mM Tris-HCl, pH 7.6, 300 mM NaCl) and analyzed by WB.
Lifespan assay
Lifespan experiments were performed as previously described [71], with some modifications. Lifespan assays by P. aeruginosa PA14 infection were conducted at 25 °C, and S. typhimurium SL1344 infection was conducted at 20 °C. The L4 stage was designated as day 0, at least 100 worms were used for each experiment, and animals were transferred every 1–2 days to fresh plates to eliminate overcrowding by progeny until they laid no further eggs. Worms were marked as dead when they did not move, pump, or respond to the stimuli. Survival curves were plotted using GraphPad Prism 8.
Yeast three-hybrid assay
The method was referred to the previous article with modifications [44]. The yeast three-hybrid assay in this subject was used to detect GDI-1 and PRAF-3 competitive binding to RAB-1. gdi-1 alone, or together with praf-3, were cloned into pBRIDGE. rab-1 was cloned into pGADT7. The pGADT7 and pBRIDGE plasmids were transformed into the Y2HGold strain and cultured on SD-Leu-Trp double-dropout solid culture medium. On the SD-Leu-Trp-His dropout medium, GDI-1 interacted with RAB-1 on SD-Met-Leu-Trp-His dropout medium to induce the expression of PRAF-3 and test for competitive binding.
Extraction of membrane proteins
Membrane protein extraction was performed using a Membrane Protein Extraction Kit (Bestbio, BB-3103). Unsynchronized N2, gdi-1(K270R), and praf-3 worms were collected and lysed with 400 μL Reagent A from the kit, and incubated at 4 °C for 25 min with rotation. After centrifugation, the supernatant was incubated at 37 °C for 8 min and centrifuged again, after which the supernatants containing cytoplasmic proteins and the lower layers containing membrane proteins were collected. The membrane proteins were resuspended by adding Reagent B. An ice bath was performed for 2 min, followed by a water bath at 37 °C for 5 min. After centrifugation, 50 μL Reagent C was added to dissolve the lower layer of solution, which is the membrane proteins. Proteins were quantified using the bicinchoninic acid method and analyzed by WB.
Extraction of ER and Golgi proteins
Extraction of ER and Golgi proteins was performed using Endoplasmic Reticulum Protein Extraction Kit (Solarbio, EX1260) and Golgi Protein Extraction Kit (Solarbio, EX1240). Unsynchronized N2, gdi-1(K270R), and praf-3 worms or control/rab1a knockdown HEK 293T cells were washed with cold PBS. For the extraction of ER, samples dissolved in 400 μL of Reagent A were homogenized using a homogenizer or sonication. After centrifugation at 1,000 × g for 5 min at 4 °C, the supernatants were centrifuged again at 11, 000 × g for 10 min. Subsequently, centrifugation at 50,000 × g for 45 min was performed, and the precipitates were resuspended in Reagent B and centrifuged again at 50,000 × g for 45 min. Finally, the precipitates were resuspended in 50 μL of protein extract C, which is an ER protein.
For the extraction of Golgi, samples dissolved in 400 μL of Reagent A were homogenized using a homogenizer or sonication. The samples were centrifuged at 3,000 × g for 10 min and 10 μL of Reagent WT was added. The samples were centrifuged at 20,000 × g for 20 min, and 500 μL of Reagent B was added to the precipitates and mixed thoroughly. After centrifugation at 20,000 × g for 30 min, 50 μL of Reagent C was added to the precipitates, which were Golgi proteins. The extracted ER and Golgi proteins were quantified by BCA method and analyzed by WB.
Antibodies
GDI-1, RAB-1, SCA-1, TBB-1, RAB-7, CNX-1, and SYX-6 antibodies were generated in mice by injecting recombinant proteins. BALB/c mice at 8–12 weeks were intraperitoneally injected with 80 μg of purified protein in a 1:1 mixture with Freund’s Adjuvant, Complete (Sigma–Aldrich, F5881). After 2 weeks, mixtures of 80 μg of purified protein with Freund’s Adjuvant, Incomplete (Sigma–Aldrich, F5506) were intraperitoneally injected at 2-week intervals for a total of three injections. Mice were blooded from eye sockets and sera were collected for antibody detection.
Quantitative real-time PCR
Quantitative real-time PCR experiments were performed as previously described [26] with some modifications. RNA was extracted from the cells by chloroform extraction, followed by ethanol precipitation and DNase treatment. Complementary DNA was synthesized from 1 μg of RNA using HiScript III RT SuperMix for qPCR (+ gDNA wiper) (Vazyme, CAT# R323) according to the manufacturer’s instructions. Quantitative RT-PCR was performed using the ChamQ Universal SYBR qPCR Master Mix (Vazyme, CAT# Q711).
shRNA-mediated gene silencing and lentivirus packaging
The targeted shRNAs were cloned into pGREEN-puro plasmids. Scrambled RNA control and shRNAs were transfected into cells using electrotransfer or lentivirus. After 48 h, the cells were used for subsequent analyses.
For lentiviral packaging, 293 TN cells were co-transfected with the targeted shRNA pGREEN-puro plasmid and pSBH155/156 packaging plasmids using PEI. After 24 h, the media were changed to DMEM (supplemented with 10% FBS and 1% penicillin-streptomycin liquid (100×)), and the cell culture fluids were harvested at 48 h and 72 h after infection, respectively which were the lentivirus infection solutions.
Bone marrow-derived macrophages
Bone marrow cells were isolated from 8 weeks C57BL/6J mice. BMDM were differentiated and cultured for one week in DMEM supplemented with 10% fetal bovine serum, 1% penicillin, and 20% L929 cell supernatant. The cells were then incubated at 37 °C with 5% CO2.
Electrotransfer
Mouse RAW264.7 cells transfection were performed using Neon Transfection Kit (Invitrogen, MPK10096) by Invitrogen Neon Transfection System. RAW264.7 cells were rinsed with PBS and resuspended with R Buffer, 100 μL 4 × 106 RAW264.7 cells and 6 μg plasmids were transfected under electric shock conditions at 1,700 V, 25 ms, 1 pules. Transfected cells were cultured in DMEM containing 10% FBS, and after 48 h of incubation, subsequent experiments were performed.
Efferocytosis assay and flow cytometry
To induce apoptosis, CharON-expressing Jurkat cells were exposed to ultraviolet (UV) irradiation at a dose of 150 J/m2 and subsequently incubated for 3 h at 37 °C under 5% CO2 to allow apoptosis progression. The resulting apoptotic Jurkat cells were harvested, resuspended in sterile 1× PBS, and added to RAW264.7 macrophages previously transfected with target-specific shRNA at a ratio of 5:1 (apoptotic cells: macrophages).
For time-lapse confocal imaging, macrophages were co-cultured with UV-induced apoptotic CharON-expressing Jurkat cells carrying a fluorescent reporter gene. After 2 h of co-incubation, noninternalized apoptotic cells were gently removed by washing the macrophages twice with phosphate-buffered saline (PBS). For microscopy analysis, 1 mL of DMEM was added to each culture dish for observation. Efferocytosis was quantified as the percentage of macrophages containing internalized red fluorescent CharON Jurkat cells among the total number of macrophages per field of view (FOV). Degradation was assessed as the ratio of macrophages containing internalized CharON Jurkat cells exhibiting only red fluorescence (indicative of loss of the green signal due to lysosomal degradation) to those containing cells with both red and green fluorescence, expressed as a percentage per FOV. In our statistical analysis, we excluded instances where macrophages were either just beginning to interact with apoptotic cells or were in the process of extending pseudopodia for engulfment. These scenarios were excluded as they did not involve the complete internalization of apoptotic cells.
Jurkat cells were stained with 1 μM CellTracker Red CMTPX (Invitrogen, E34552) for 30 min, and then irradiated with 150 J UV, followed by incubation at 37 °C 5% CO2 for 3 h to generate apoptosis. The apoptotic Jurkat cells were resuspended in 1x PBS, and added to RAW264.7, J774A.1, or BMDM cells transfected with the pias1-shRNA at a ratio of 10:1, followed by incubation at 37 °C for 6 h. After incubation, the macrophages were rinsed twice with PBS to remove unbound apoptotic Jurkat cells. For microscopic imaging, 1 ml of DMEM was added to the cell culture dish. Apoptotic Jurkat cell uptake was quantified as the percentage of Apoptotic Jurkat cell-positive macrophages out of the total number of macrophages per field of view. For flow cytometric experiments, the macrophages were digested with trypsin, resuspended in 200 µL PBS, and analyzed on a BECKMAN COULTER CytoFLEX S flow cytometer. Data analysis was performed using CytExpert software.
Statistical analyses
Statistical analyses were performed using Microsoft Office Excel, and bar graphs or other statistical plots were produced using GraphPad Prism Software. Data were derived from at least three biological replicates. The details of statistical tests, sample size, and methodology of statistics were provided in individual Figure legends. The two-tailed unpaired t test was used for statistical comparison of two groups of samples. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. Lifespan assays were analyzed using Log-Rank method. *p < 0.05; **p < 0.01; ***p < 0.001.
Supporting information
S1 Fig. SUMO modifications functions independently of either pathway for cell corpse removal.
(a–d) The development of adult stages corpses was quantified (mean ± SEM) in the indicated strains. Fifteen adult worms were scored at each stage for each strain. (e) Four-dimensional microscopy analysis of germ cell corpse duration was performed in N2, mms-21 (xwh80), and gei-17 (xwh81). The persistence of 30 germ cell corpses was also monitored. (f and h) The number of embryo corpses was measured by four-dimensional microscopy in N2, gei-17 (xwh81), (f), and mms-21 (xwh80) (h). Embryos (2-cell stage) were isolated from adult worms and imaged every minute for 400 min. (g and i) Four-dimensional microscopy analysis of somatic cell corpse persistence in N2, gei-17 (xwh81), (g), and mms-21 (xwh80) (i). Thirty somatic cell corpses were analyzed for each genotype. (j and k) Quantification (mean ± SEM) of germ cell corpses from the indicated strains. Cell corpses of each animal were scored in 15 animals at every time point (h post L4) as indicated. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, ***P < 0.001. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data.
https://doi.org/10.1371/journal.pbio.3004123.s001
(TIF)
S2 Fig. SUMOylation affects phagosome acidification.
(a) Representative images of phagosome labeling by CED-1::GFP and mCHERRY::RAB-7, CED-1::GFP and LAAT-1::mCHERRY, CED-1::GFP and NUC-1::mCHERRY, or CED-1::GFP and CPL-1::mChOint. Phagosomes are indicated by arrows. Bars, 5 μm. (b) Quantification (mean ± SEM) of the labeling of phagosome by mCHERRY::RAB-7/CED-1::GFP, LAAT-1::mCHERRY/CED-1::GFP, NUC-1::mCHERRY/CED-1::GFP, and CPL-1::mChOint/CED-1::GFP shown in (a). ≥ 100 CED-1::GFP labeled germ cell corpses were analyzed for phagosomal markers; and data were derived from three replicates. Comparisons were performed between the N2/control and smo-1 RNAi worms for each marker. (c) Cell corpses positive for phagosome markers in control, aos-1, uba-2, ubc-9, mms-21, and gei-17 RNAi were quantified. ≥100 cell corpses from at least 15 worms were analyzed for each marker; and data were derived from three replicates. (d) The pHluorin labeled germ cell corpses from the indicated animals. Arrows indicate cell corpses in the Plim-7pHluorin/control and smo-1 RNAi worms. Bars, 5 μm. The graphs show the percentage referred to the ratio of germ cell corpses labeled by pHluorin in Plim-7pHluorin/control and smo-1 RNAi worms. (e) Representative images of germ cell corpse staining by pHrodo in indicated strains. Quantification of corpse staining is shown on the right. The two-tailed unpaired t test was used for statistical comparison of two groups of samples. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data.
https://doi.org/10.1371/journal.pbio.3004123.s002
(TIF)
S3 Fig. SUMOylation regulates phagosome degradation.
(a) Animals expressing ssGFP controlled by a heat-shock promoter treated with control and smo-1 RNAi were heat shocked for 60 min at 33 °C, and the uptake and degradation of ssGFP in coelomocytes were monitored at the indicated time points. Left, accumulation of ssGFP in the body cavity (400×), and arrows indicate coelomocytes that are enlarged in the right pictures (1,000×). Bars, 10 μm. (b) Time-lapse chasing of HIS-24::mCherry positive phagolysosomes in N2/control and smo-1 RNAi worm germlines. The time point at which the HIS-24::mCherry mostly rings the cell corpse was set to 0 min. Arrows indicate the continuous presence of cell corpses (DIC, top row) and HIS-24::mCherry (fluorescence, bottom row). DIC and fluorescence images were obtained at certain time intervals. Bars, 5 µm. (c) The persistence of the HIS-24::mCherry fluorescence signal in five germ cell corpses in N2/control and smo-1 RNAi worms was monitored. An unpaired t test was performed in this figure. ***P < 0.001. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data.
https://doi.org/10.1371/journal.pbio.3004123.s003
(TIF)
S4 Fig. Identification of GDI-1 as a potential substrate of GEI-17 for apoptotic cell clearance.
(a) Schematic illustration of the HA tag insertion generated by CRISPR-Cas9 editing of endogenous smo-1 loci. Amino acids near the insertion site are indicated. (b) Grouping of identified SUMO targets into broad functional classes. (c and d) Yeast two-hybrid assay of the interactions between VHA-9 or VHA-14 and SMO-1 (c) GDI-1 or ARL-8 and SMO-1 WT/GG’/ΔGG (d). BD-53 + AD-T serves as the positive control, while BD-Lam + AD-T functions as the negative control. SMO-1 GG’, residues after C-terminal di-glycine were deleted. SMO-1 ΔGG, di-glycine residues in SMO-1 were deleted. -TL, medium lacking Trp and Leu; -TLH, medium lacking Trp, Leu, and His. (e and f) The interactions between ARL-8 or GDI-1 and MMS-21 (e) and GEI-17 and ARL-8 (f) were examined by GST pull-down assays. GST and GST::GFP serve as negative controls. (g) The interactions between ARL-8 or GDI-1 and GEI-17 or MMS-21 were examined by CO-IP in 293T cells. (h) Coomassie brilliant blue (CBB) stained gel of MYC::hSUMO1, hSAE1, hSAE2, hUBC9, and HA::GDI-1 following affinity purification using Ni-NTA agarose. The raw uncropped western blot images and the raw Coomassie Brilliant Blue-stained protein gels underlying this Figure can be found in S1 Raw Images.
https://doi.org/10.1371/journal.pbio.3004123.s004
(TIF)
S5 Fig. The 1–150aa of GDI-1 interacts with GEI-17.
(a–g) The interactions between various truncated forms of GDI-1 amino acids and GEI-17 were examined by CO-IP in 293T cells. (h and i) The interactions between GDI-1(1−150aa) and GEI-17 (h) and GDI-1(150−444aa) and GEI-17 (i) were examined by GST pull-down assays. GST and GST::GFP serve as negative controls. The raw uncropped western blot images and the Coomassie Brilliant Blue-stained protein gels underlying this Figure can be found in S1 Raw Images.
https://doi.org/10.1371/journal.pbio.3004123.s005
(TIF)
S6 Fig. GDI-1 SUMOylation at K270 affects Endosomal/Lysosomal degradation and phagosome degradation.
(a) N2, gei-17, and gdi-1(K270R) animals expressing ssGFP controlled by a heat-shock promoter were heat shocked for 60 min at 33 °C, and the uptake and degradation of ssGFP in coelomocytes were monitored at the indicated time points. Left, accumulation of ssGFP in the body cavity (400×), and arrows indicate coelomocytes that are enlarged in the right pictures (1,000×). Bars, 10 μm. (b) Time-lapse chasing of cell corpses in DIC and HIS-24::mCherry positive phagolysosomes in N2 and gdi-1(K270R) germlines. The time point at which the HIS-24::mCherry ring was first detected on the cell corpses was set as 0 min. Bars, 5 µm. (c) The persistence of the HIS-24::mCherry fluorescence signal in five germ cell corpses in N2 and gdi-1(K270R) worms was monitored. (d–f) Time to 50% mortality of C. elegans in OP50 (d), PA14 (e), and SL1344 (f). An unpaired t test was performed in this figure. *P < 0.05, **P < 0.01, ***P < 0.001. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data.
https://doi.org/10.1371/journal.pbio.3004123.s006
(TIF)
S7 Fig. Identification of RAB-1 as an interaction protein for GDI-1.
(a–f) Yeast two-hybrid assay for screening RABs that interact with GDI-1. (g–i and n) The interaction between GDI-1 and RAB-1 (g), GDI-1 and RAB-28 (h), GDI-1 and RAB-Y1 (i), and PRAF-3 and RAB-1 (n) were examined by CO-IP in 293T cells. (j and k) The interactions between GDI-1 and RAB-28 (j) and GDI-1 and RABY1 (k) were examined by GST pull-down assays. GST and GST::GFP serve as negative controls. (l and m) Different stages (h post L4) of germ cell corpses were quantified (mean ± SEM) in the indicated strains. Fifteen adult worms were scored at each stage for each strain. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, ***P < 0.001. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images and the Coomassie Brilliant Blue-stained protein gels underlying this Figure can be found in S1 Raw Images.
https://doi.org/10.1371/journal.pbio.3004123.s007
(TIF)
S8 Fig. RAB-1 interacts with RAB-7 and regulates RAB-7 protein levels in conjunction with SUMOylation of GDI-1.
(a and c) The interaction intensity between RAB-1 and GDI-1 (WT/K270R) was examined by the GST pulldown assay (a) and FLAG IP in flag::gdi-1, flag::gdi- 1(K270R), and gei-17; flag::gdi-1 worms (c). (b and d) The graphs show the protein levels of RAB-1 and GDI-1. (e and f) The interaction between RAB-7 and RAB-1 was examined by GST pulldown assay (e) and CO-IP in HEK293T cells (f). (g) The interaction between RAB-7 and GDI-1 was examined by the GST pulldown assay. GST and GST::GFP serve as negative controls. (h and j) Exogenous RAB-7 protein levels were examined by immunoblot analysis in Pced-1gfp::rab-7 and gdi-1(K270R); Pced-1gfp::rab-7 (h) and Pced-1gfp::rab-7 control/rab-1 RNAi (j) worms. (i and k) The graphs show quantification of the protein level of RAB-7 in Pced-1gfp::rab-7 and gdi-1(K270R); Pced-1gfp::rab-7 (i) and Pced-1gfp::rab-7 control/rab-1 RNAi (k) using ImageJ. (l and m) The graphs show the percentage referred to the ratio of germ cell corpses labeled by mCHERRY::RAB-1 in Pced-1mcherry::rab-1 and gdi-1(K270R); Pced-1mcherry::rab-1 (l) and Pced-1mcherry::rab-1 control/rab-7 RNAi (m) worms. (n and p) Exogenous RAB-7 protein levels were examined by immunoblot analysis in Pced-1gfp::rab-7 (n) and gdi-1(K270R); Pced-1gfp::rab-7 (p) treated with 25 µM Chloroquine (lysosome inhibitor) or 10µM MG-132 (proteasome inhibitor). (o and q) The graphs show quantification of the protein level of RAB-7 in Pced-1gfp::rab-7 (o) and gdi-1(K270R); Pced-1gfp::rab-7 (q) treated with 25µM Chloroquine or 10µM MG-132 using ImageJ software. The unpaired t test was performed in this figure. One-way ANOVA followed by Dunnett’s posttest was performed for statistical comparison of multiple groups of samples. *P < 0.05, **P < 0.01, NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images and the Coomassie Brilliant Blue-stained protein gels underlying this Figure can be found in S1 Raw Images.
https://doi.org/10.1371/journal.pbio.3004123.s008
(TIF)
S9 Fig. SUMOylation of GDI-1 at K270 regulates RAB-1-mediated vesicular transport of RAB-7 from ER to Golgi.
(a and c) Endogenous RAB-7 protein levels were examined by immunoblot analysis in N2/control or gdi-1 RNAi (a) and gdi-1(K270R)/control or gdi-1 RNAi (c) in endoplasmic reticulum (ER) or Golgi apparatus. CNX-1 is an ER marker, SYX-6 is a Golgi marker, TBB-1 is a cytoskeleton marker, and Histon H3 is a nucleus marker. (b and d) Graphs show quantification of the protein level of RAB-7 in the ER and Golgi apparatus in N2/control or gdi-1 RNAi (b) and gdi-1(K270R)/control or gdi-1 RNAi (d) using ImageJ software. (e) Representative images of germ cell corpse labeled by sfGFP(1–10)::RAB-7 + VPS-41::sfGFP(11) in indicated strains. Arrows indicate cell corpses. Bars, 5 µm. (f) The graphs show the fluorescence intensity value on the phagosome using ImageJ software. (g) rab1a mRNA in 293T cells treated with control or rab1a shRNA was determined by qRT-PCR. (h and j) Exogenous RAB7A protein levels were examined by immunoblot analysis in 293T cells treated with control or rab1a shRNA in the ER (h) or Golgi (j). Calnexin is the ER marker; Syntaxin6 is the Golgi marker. (i and k) The graphs show quantification of the protein level of RAB7A in ER (i) and Golgi (k) using ImageJ. The unpaired t test was performed in this figure. *P < 0.05, **P < 0.01, NS, no significance. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images underlying this Figure can be found in S1 Raw Images.
https://doi.org/10.1371/journal.pbio.3004123.s009
(TIF)
S10 Fig. The interaction between GDI-1 and GEI-17 and RAB-1 is conserved.
(a–d) The interaction between hGDI1 and GEI-17 (a), hGDI1 and RAB-1 (b), GDI-1 and PIAS1 (c), and GDI-1 and RAB1A (d) were examined by CO-IP in HEK293T cells. (e and f) The interaction between hGDI1 and PIAS1 (e) and hGDI1 and RAB1A (f) was examined by the GST pulldown assay. GST and GST::GFP serve as negative controls. (g) Sequence alignment of C. elegans (C.e), mice (Mus.), and human (Homo.) GDI-1. Identical residues are shaded in black, and similar ones in gray. The asterisk indicates 270 lysine of C.e GDI-1. (h and i) pias1 (h) and gdi1 (i) mRNA in RAW264.7 cells treated with control or pias1 shRNA (h) and control or gdi1 shRNA (i) were determined by qRT-PCR. The unpaired t test was performed in this figure. **P < 0.01. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data. The raw uncropped western blot images and the Coomassie Brilliant Blue-stained protein gels underlying this Figure can be found in S1 Raw Images.
https://doi.org/10.1371/journal.pbio.3004123.s010
(TIF)
S11 Fig. Representative images of degradation experiments in gdi1 and pias1 knockdown.
(a–c) Role of gdi1 and pias1 in macrophage degradation of apoptotic cells. Macrophages expressing control, gdi1, and pias1-targeting shRNA (blue) were incubated with dual-color apoptotic CharON Jurkat reporters (green: apoptosis-induced and acid-sensitive; red: acid-resistant) at 3, 6, and 9 h.
https://doi.org/10.1371/journal.pbio.3004123.s011
(TIF)
S12 Fig. SUMOylation and GDI1 regulates phagocytosis in mammals.
(a–c) Representative images of control and pias1 shRNA-treated RAW264.7 (a), J774A.1 (b), and BMDM (c) phagocytes (green) engulfing apoptotic Jurkat cells (red). Bars, 100 µm. Arrows indicate the phagocytosis of apoptotic cells. The boxed region is magnified (2×) in the insets (bars, 25 µm). (d–f) Quantification of the percentage of control or pias1 shRNA treated RAW264.7 (d), J774A.1 (e), and BMDM (f) phagocytes performing apoptotic Jurkat cells engulfment from three independent assays. (g) Representative images of control and gdi1 shRNA-treated RAW264.7 phagocytes (green) engulfing apoptotic Jurkat cells (red). Bars, 100 µm. Arrows indicate the phagocytosis of apoptotic cells. The boxed region is magnified (2×) in the insets (bars, 25 µm). (h) Phagocytosis rates of control and gdi1 shRNA-treated RAW264.7 phagocytes performing apoptotic Jurkat cells engulfment from three independent assays. Quantification of the percentage of control or gdi1 shRNA treated RAW264.7 phagocytes performing apoptotic Jurkat cells engulfment from three independent assays. The unpaired t test was performed in this figure. *P < 0.05, **P < 0.01, ***P < 0.001. All error bars indicate the mean ± SEM. The data underlying this Figure can be found in S1 Data.
https://doi.org/10.1371/journal.pbio.3004123.s012
(TIF)
S13 Fig. SUMOylation regulates phagocytosis in mammals.
(a–c) Flow cytometry plots evaluating the phagocytosis of CellTracker Red-labeled apoptotic cells by RAW264.7 cells (a), J774A.1 cells (b), and bone marrow-derived macrophages (BMDM) (c) transfected with control shRNA or pias1 shRNA. For each panel, the first column displays the FSC-A versus SSC-A gating strategy used to select the target macrophage population (P1 gate), excluding debris. The second column shows FITC-A (GFP, representing shRNA-transduced macrophages) versus PE-A (CellTracker Red, representing apoptotic cells) quadrant scatter plots. In these scatter plots, the quadrants represent: Q1-UL (GFP−/PE+), nontransduced macrophages or free apoptotic bodies; Q1-UR (GFP+/PE+), successfully transduced macrophages that have engulfed apoptotic cells (the phagocytic population); Q1-LL (GFP−/PE−), nontransduced macrophages without phagocytosis; and Q1-LR (GFP+/PE−), successfully transduced macrophages that did not undergo phagocytosis. The third and fourth columns show representative histograms of FITC-A and PE-A fluorescence intensity, respectively, with the P2 gate indicating GFP-positive (transduced) cells and the P3 gate indicating PE-positive cells. The raw flow cytometry data files (.fcs) underlying this Figure can be found in S2 Data.
https://doi.org/10.1371/journal.pbio.3004123.s013
(TIF)
S14 Fig. SUMOylation of GDI1 regulates phagocytosis in mammals.
(a) Flow cytometry plots comparing the phagocytic capacity of RAW264.7 cells transfected with control shRNA or gdi1 shRNA. In this panel, the FITC-A signal represents the GFP fluorescence expressed by the shRNA lentiviral vector, indicating successfully transduced macrophages. (b) Flow cytometry analysis of RAW264.7 cells undergoing rescue experiments. Cells were co-transfected with gdi1 shRNA (which lacks fluorescent tags) and one of the following plasmids: sfGFP-N1 (empty vector control), sfGFP-N1-gdi1 (wild-type rescue), or sfGFP-N1-gdi1(K209R) (mutant rescue). Here, the FITC-A signal represents the GFP fluorescence from the sfGFP-N1 plasmid, indicating successful transfection of the rescue constructs. For both panels (a) and (b), the first column displays the FSC-A versus SSC-A gating strategy for target macrophages (P1). The second column shows FITC-A (GFP) versus PE-A (CellTracker Red-labeled apoptotic cells) quadrant scatter plots. The quadrants are defined as follows: Q1-UR (GFP+/PE+) represents successfully transduced/transfected macrophages that have engulfed apoptotic cells (the phagocytic population); Q1-LR (GFP+/PE−) represents GFP-positive macrophages without phagocytosis; Q1-UL (GFP−/PE+) represents nontransduced cells or free apoptotic bodies; and Q1-LL (GFP−/PE−) represents double-negative cells. The third and fourth columns display representative histograms of FITC-A and PE-A fluorescence, respectively, where the P2 gate indicates GFP-positive cells, and the P3 gate indicates PE-positive cells. The raw flow cytometry data files (.fcs) underlying this Figure can be found in S3 Data.
https://doi.org/10.1371/journal.pbio.3004123.s014
(TIF)
S1 Raw Images. The uncropped western blot images and Coomassie Brilliant Blue-stained gel images for all main and supporting figures.
This file contains the uncropped western blot images for Figs 3a, 3g, 3h, 3j, 4a, 4d, 4f, 5a, 5c, 5h, 5k, 6a, 6c, 6f, 6h, 6k, 6m, 6o, 6q, 7a–7c, 7g, S4e–S4g, S5a–S5i, S7g–S7k, S7n, S8a, S8c, S8e–S8h, S8j, S8n, S8p, S9a, S9c, S9h,S9j, and S10a–S10f, as well as the Coomassie Brilliant Blue-stained gel images for Figs 3g, 5c, 5h, S4e, S4f, S4h, S5h, S5i, S7j, S7k, S8e, S8g, S10a, and S10b.
https://doi.org/10.1371/journal.pbio.3004123.s019
(PDF)
Acknowledgments
We thank Dr. Chonglin Yang, Dr. Di Chen, and the C. elegans Genetic Center for C. elegans strains, and Dr. Cheng-Gang Zou for providing P. aeruginosa PA14 and S. typhimurium SL1344 strains.
References
- 1. Doran AC, Yurdagul A Jr, Tabas I. Efferocytosis in health and disease. Nat Rev Immunol. 2020;20(4):254–67. pmid:31822793
- 2. Boada-Romero E, Martinez J, Heckmann BL, Green DR. The clearance of dead cells by efferocytosis. Nat Rev Mol Cell Biol. 2020;21(7):398–414. pmid:32251387
- 3. Schilperoort M, Ngai D, Sukka SR, Avrampou K, Shi H, Tabas I. The role of efferocytosis-fueled macrophage metabolism in the resolution of inflammation. Immunol Rev. 2023;319(1):65–80. pmid:37158427
- 4. Yang C, Wang X. Lysosome biogenesis: regulation and functions. J Cell Biol. 2021;220(6):e202102001. pmid:33950241
- 5. Nagata S. Apoptosis and clearance of apoptotic cells. Annu Rev Immunol. 2018;36:489–517. pmid:29400998
- 6. Kawano M, Nagata S. Efferocytosis and autoimmune disease. Int Immunol. 2018;30(12):551–8. pmid:30165442
- 7. Romero-Molina C, Garretti F, Andrews SJ, Marcora E, Goate AM. Microglial efferocytosis: diving into the Alzheimer’s disease gene pool. Neuron. 2022;110(21):3513–33. pmid:36327897
- 8. Poon IKH, Ravichandran KS. Targeting efferocytosis in inflammaging. Annu Rev Pharmacol Toxicol. 2024;64:339–57. pmid:37585658
- 9. Wang X, Yang C. Programmed cell death and clearance of cell corpses in Caenorhabditis elegans. Cell Mol Life Sci. 2016;73(11–12):2221–36. pmid:27048817
- 10. Xie C, Chen G, Li M, Huang P, Chen Z, Lei K, et al. Neurons dispose of hyperactive kinesin into glial cells for clearance. EMBO J. 2024;43(13):2606–35. pmid:38806659
- 11. Mapes J, Chen Y-Z, Kim A, Mitani S, Kang B-H, Xue D. CED-1, CED-7, and TTR-52 regulate surface phosphatidylserine expression on apoptotic and phagocytic cells. Curr Biol. 2012;22(14):1267–75. pmid:22727702
- 12. Zhou Z, Hartwieg E, Horvitz HR. CED-1 is a transmembrane receptor that mediates cell corpse engulfment in C. elegans. Cell. 2001;104(1):43–56. pmid:11163239
- 13. Chen D, Xiao H, Zhang K, Wang B, Gao Z, Jian Y, et al. Retromer is required for apoptotic cell clearance by phagocytic receptor recycling. Science. 2010;327(5970):1261–4. pmid:20133524
- 14. Wang X, Wu Y-C, Fadok VA, Lee M-C, Gengyo-Ando K, Cheng L-C, et al. Cell corpse engulfment mediated by C. elegans phosphatidylserine receptor through CED-5 and CED-12. Science. 2003;302(5650):1563–6. pmid:14645848
- 15. Reddien PW, Horvitz HR. CED-2/CrkII and CED-10/Rac control phagocytosis and cell migration in Caenorhabditis elegans. Nat Cell Biol. 2000;2(3):131–6. pmid:10707082
- 16. Fazeli G, Levin-Konigsberg R, Bassik MC, Stigloher C, Wehman AM. A BORC-dependent molecular pathway for vesiculation of cell corpse phagolysosomes. Curr Biol. 2023;33(4):607-621.e7. pmid:36652947
- 17. Lee Y, Hamann JC, Pellegrino M, Durgan J, Domart MC, Collinson LM, et al. Entosis controls a developmental cell clearance in C. elegans. Cell Rep. 2019;26(12):3212–20. pmid:30893595
- 18. Guo P, Hu T, Zhang J, Jiang S, Wang X. Sequential action of Caenorhabditis elegans Rab GTPases regulates phagolysosome formation during apoptotic cell degradation. Proc Natl Acad Sci U S A. 2010;107(42):18016–21. pmid:20921409
- 19. Li W, Zou W, Zhao D, Yan J, Zhu Z, Lu J, et al. C. elegans Rab GTPase activating protein TBC-2 promotes cell corpse degradation by regulating the small GTPase RAB-5. Development. 2009;136(14):2445–55. pmid:19542357
- 20. Poteryaev D, Fares H, Bowerman B, Spang A. Caenorhabditis elegans SAND-1 is essential for RAB-7 function in endosomal traffic. EMBO J. 2007;26(2):301–12. pmid:17203072
- 21. Li P, Jing H, Wang Y, Yuan L, Xiao H, Zheng Q. SUMO modification in apoptosis. J Mol Histol. 2021;52(1):1–10. pmid:33225418
- 22. Zheng N, Shabek N. Ubiquitin ligases: structure, function, and regulation. Annu Rev Biochem. 2017;86:129–57. pmid:28375744
- 23. Qin Y, Li Q, Liang W, Yan R, Tong L, Jia M, et al. TRIM28 SUMOylates and stabilizes NLRP3 to facilitate inflammasome activation. Nat Commun. 2021;12(1):4794. pmid:34373456
- 24. Carvalho CA, Abu-Shach UB, Raju A, Vershinin Z, Levy D, Boxem M, et al. SUMO-mediated regulation of H3K4me3 reader SET-26 controls germline development in C. elegans. PLoS Biol. 2025;23(1):e3002980. pmid:39761316
- 25. Xiao H, Wang H, Silva EA, Thompson J, Guillou A, Yates JR Jr, et al. The Pallbearer E3 ligase promotes actin remodeling via RAC in efferocytosis by degrading the ribosomal protein S6. Dev Cell. 2015;32(1):19–30. pmid:25533207
- 26. Yuan L, Li P, Jing H, Zheng Q, Xiao H. trim-21 promotes proteasomal degradation of CED-1 for apoptotic cell clearance in C. elegans. Elife. 2022;11:e76436. pmid:35929733
- 27. Zhao X. SUMO-mediated regulation of nuclear functions and signaling processes. Mol Cell. 2018;71(3):409–18. pmid:30075142
- 28. Vertegaal ACO. Signalling mechanisms and cellular functions of SUMO. Nat Rev Mol Cell Biol. 2022;23(11):715–31. pmid:35750927
- 29. Princz A, Pelisch F, Tavernarakis N. SUMO promotes longevity and maintains mitochondrial homeostasis during ageing in Caenorhabditis elegans. Sci Rep. 2020;10(1):15513. pmid:32968203
- 30. Pelisch F, Tammsalu T, Wang B, Jaffray EG, Gartner A, Hay RT. A SUMO-dependent protein network regulates chromosome congression during Oocyte Meiosis. Mol Cell. 2017;65(1):66–77. pmid:27939944
- 31. Kaminsky R, Denison C, Bening-Abu-Shach U, Chisholm AD, Gygi SP, Broday L. SUMO regulates the assembly and function of a cytoplasmic intermediate filament protein in C. elegans. Dev Cell. 2009;17(5):724–35. pmid:19922876
- 32. Drabikowski K, Ferralli J, Kistowski M, Oledzki J, Dadlez M, Chiquet-Ehrismann R. Comprehensive list of SUMO targets in Caenorhabditis elegans and its implication for evolutionary conservation of SUMO signaling. Sci Rep. 2018;8(1):1139. pmid:29348603
- 33. Tsur A, Bening Abu-Shach U, Broday L. ULP-2 SUMO protease regulates e-cadherin recruitment to adherens junctions. Developmental Cell. 2015;35(1):63–77. pmid:26412237
- 34. Fergin A, Boesch G, Greter NR, Berger S, Hajnal A. Tissue-specific inhibition of protein sumoylation uncovers diverse SUMO functions during C. elegans vulval development. PLoS Genet. 2022;18(6):e1009978. pmid:35666766
- 35. Kim H, Ding Y-H, Zhang G, Yan Y-H, Conte D Jr, Dong M-Q, et al. HDAC1 SUMOylation promotes Argonaute-directed transcriptional silencing in C. elegans. Elife. 2021;10:e63299. pmid:34003109
- 36. Yuan JY, Horvitz HR. The Caenorhabditis elegans genes ced-3 and ced-4 act cell autonomously to cause programmed cell death. Dev Biol. 1990;138(1):33–41. pmid:2307287
- 37. Zou L, Wu D, Zang X, Wang Z, Wu Z, Chen D. Construction of a germline-specific RNAi tool in C. elegans. Sci Rep. 2019;9(1):2354. pmid:30787374
- 38. Kinchen JM, Doukoumetzidis K, Almendinger J, Stergiou L, Tosello-Trampont A, Sifri CD, et al. A pathway for phagosome maturation during engulfment of apoptotic cells. Nat Cell Biol. 2008;10(5):556–66. pmid:18425118
- 39. Liu B, Du H, Rutkowski R, Gartner A, Wang X. LAAT-1 is the lysosomal lysine/arginine transporter that maintains amino acid homeostasis. Science. 2012;337(6092):351–4. pmid:22822152
- 40. Lyon CJ, Evans CJ, Bill BR, Otsuka AJ, Aguilera RJ. The C. elegans apoptotic nuclease NUC-1 is related in sequence and activity to mammalian DNase II. Gene. 2000;252(1–2):147–54. pmid:10903446
- 41. Xu M, Liu Y, Zhao L, Gan Q, Wang X, Yang C. The lysosomal cathepsin protease CPL-1 plays a leading role in phagosomal degradation of apoptotic cells in Caenorhabditis elegans. Mol Biol Cell. 2014;25(13):2071–83. pmid:24829385
- 42. Xiao H, Chen D, Fang Z, Xu J, Sun X, Song S, et al. Lysosome biogenesis mediated by vps-18 affects apoptotic cell degradation in Caenorhabditis elegans. Mol Biol Cell. 2009;20(1):21–32. pmid:18923146
- 43. Becker J, Barysch SV, Karaca S, Dittner C, Hsiao H-H, Berriel Diaz M, et al. Detecting endogenous SUMO targets in mammalian cells and tissues. Nat Struct Mol Biol. 2013;20(4):525–31. pmid:23503365
- 44. Gao K, Li Y, Hu S, Liu Y. SUMO peptidase ULP-4 regulates mitochondrial UPR-mediated innate immunity and lifespan extension. Elife. 2019;8:e41792. pmid:30642431
- 45. Gallegos ME, Balakrishnan S, Chandramouli P, Arora S, Azameera A, Babushekar A. The C. elegans Rab family: identification, classification and toolkit construction. PLoS One. 2012;7(11). pmid:WOS:000311821000045
- 46. Guruharsha KG, Rual J-F, Zhai B, Mintseris J, Vaidya P, Vaidya N, et al. A protein complex network of Drosophila melanogaster. Cell. 2011;147(3):690–703. pmid:22036573
- 47. Sasaki A, Nakae I, Nagasawa M, Hashimoto K, Abe F, Saito K, et al. Arl8/ARL-8 functions in apoptotic cell removal by mediating phagolysosome formation in Caenorhabditis elegans. Mol Biol Cell. 2013;24(10):1584–92. pmid:23485564
- 48. Raymond MH, Davidson AJ, Shen Y, Tudor DR, Lucas CD, Morioka S, et al. Live cell tracking of macrophage efferocytosis during Drosophila embryo development in vivo. Science. 2022;375(6585):1182–7. pmid:35271315
- 49. Schilperoort M, Ngai D, Katerelos M, Power DA, Tabas I. PFKFB2-mediated glycolysis promotes lactate-driven continual efferocytosis by macrophages. Nat Metab. 2023;5(3):431–44. pmid:36797420
- 50. Gerlach BD, Ampomah PB, Yurdagul A Jr, Liu C, Lauring MC, Wang X, et al. Efferocytosis induces macrophage proliferation to help resolve tissue injury. Cell Metab. 2021;33(12):2445-2463.e8. pmid:34784501
- 51. Dikic I, Schulman BA. An expanded lexicon for the ubiquitin code. Nat Rev Mol Cell Biol. 2023;24(4):273–87. pmid:36284179
- 52. Liu J, Li M, Li L, Chen S, Wang X. Ubiquitination of the PI3-kinase VPS-34 promotes VPS-34 stability and phagosome maturation. J Cell Biol. 2018;217(1):347–60. pmid:29092895
- 53. Yuan L, Li P, Zheng Q, Wang H, Xiao H. The ubiquitin-proteasome system in apoptosis and apoptotic cell clearance. Front Cell Dev Biol. 2022;10:914288. pmid:35874820
- 54. Kim H, Ding Y-H, Lu S, Zuo M-Q, Tan W, Conte D Jr, et al. PIE-1 SUMOylation promotes germline fates and piRNA-dependent silencing in C. elegans. Elife. 2021;10:e63300. pmid:34003111
- 55. Bar DZ, Charar C, Gruenbaum Y. Small GTPases in C. elegans metabolism. Small GTPases. 2018;9(5):415–9. pmid:27854154
- 56. Gilbert PM, Burd CG. GDP dissociation inhibitor domain II required for Rab GTPase recycling. J Biol Chem. 2001;276(11):8014–20. pmid:11116150
- 57. Müller MP, Goody RS. Molecular control of Rab activity by GEFs, GAPs and GDI. Small GTPases. 2018;9(1–2):5–21. pmid:28055292
- 58. Mangahas PM, Yu X, Miller KG, Zhou Z. The small GTPase Rab2 functions in the removal of apoptotic cells in Caenorhabditis elegans. J Cell Biol. 2008;180(2):357–73. pmid:18227280
- 59. Zhang J, Jiang Z, Shi A. Rab GTPases: the principal players in crafting the regulatory landscape of endosomal trafficking. Comput Struct Biotechnol J. 2022;20:4464–72. pmid:36051867
- 60. Li C, Fu J, Shao S, Luo Z-Q. Legionella pneumophila exploits the endo-lysosomal network for phagosome biogenesis by co-opting SUMOylated Rab7. PLoS Pathog. 2024;20(5):e1011783. pmid:38739652
- 61. Mohapatra G, Gaur P, Mujagond P, Singh M, Rana S, Pratap S, et al. A SUMOylation-dependent switch of RAB7 governs intracellular life and pathogenesis of Salmonella Typhimurium. J Cell Sci. 2019;132(1):jcs222612. pmid:30510112
- 62. Lohmer LL, Clay MR, Naegeli KM, Chi Q, Ziel JW, Hagedorn EJ, et al. A sensitized screen for genes promoting invadopodia function in vivo: CDC-42 and Rab GDI-1 direct distinct aspects of invadopodia formation. PLoS Genet. 2016;12(1):e1005786. pmid:26765257
- 63. D’Adamo P, Menegon A, Lo Nigro C, Grasso M, Gulisano M, Tamanini F, et al. Mutations in GDI1 are responsible for X-linked non-specific mental retardation. Nat Genet. 1998;19(2):134–9. pmid:9620768
- 64. Morè L, Künnecke B, Yekhlef L, Bruns A, Marte A, Fedele E, et al. Altered fronto-striatal functions in the Gdi1-null mouse model of X-linked intellectual disability. Neuroscience. 2017;344:346–59. pmid:28057534
- 65. Liu L, Liu L, Lu Y, Zhang T, Zhao W. GDP dissociation inhibitor 1 (GDI1) attenuates β-amyloid-induced neurotoxicity in Alzheimer’s diseases. Neurosci Lett. 2024;818:137564. pmid:38013121
- 66. Kiral FR, Kohrs FE, Jin EJ, Hiesinger PR. Rab GTPases and membrane trafficking in neurodegeneration. Curr Biol. 2018;28(8):R471–R86. pmid:29689231
- 67. Cooper AA, Gitler AD, Cashikar A, Haynes CM, Hill KJ, Bhullar B, et al. Alpha-synuclein blocks ER-Golgi traffic and Rab1 rescues neuron loss in Parkinson’s models. Science. 2006;313(5785):324–8. pmid:16794039
- 68. Brenner S. The genetics of Caenorhabditis elegans. Genetics. 1974;77(1):71–94. pmid:4366476
- 69. Paix A, Wang Y, Smith HE, Lee C-YS, Calidas D, Lu T, et al. Scalable and versatile genome editing using linear DNAs with microhomology to Cas9 Sites in Caenorhabditis elegans. Genetics. 2014;198(4):1347–56. pmid:25249454
- 70. Gan Q, Wang X, Zhang Q, Yin Q, Jian Y, Liu Y. The amino acid transporter SLC-36.1 cooperates with PtdIns3P 5-kinase to control phagocytic lysosome reformation. J Cell Biol. 2019;218(8):2619–37. pmid:31235480
- 71. Wan J, Yuan L, Jing H, Zheng Q, Xiao H. Defective apoptotic cell clearance activates innate immune response to protect Caenorhabditis elegans against pathogenic bacteria. Virulence. 2021;12(1):75–83. pmid:33372828






















English (US) ·
French (CA) ·