GOP-1 promotes apoptotic cell degradation by activating the small GTPase Rab2 in C. elegans

Phagocytic removal of apoptotic cells is important for tissue remodeling, suppression of inflammation, and regulation of immune responses (Savill and Fadok, 2000; Savill et al., 2002). In Caenorhabditis elegans hermaphrodites, 131 somatic cells and ∼50% of germ cells die through apoptosis, and the resulting cell corpses are phagocytosed and cleared by neighboring cells in the soma or by gonadal sheath cells that encase the germ line. During this process, apoptotic cells expose the phosphatidylserine “eat me” signal on the surface and are recognized and engulfed by phagocytes through evolutionarily conserved pathways, leading to cytoskeleton reorganization and formation of membrane-bound vesicles, namely phagosomes (Pinto and Hengartner, 2012; Wang and Yang, 2016). Maturation of cell corpse–enclosing phagosomes, which in many ways parallels endosome progression and maturation of phagosomes containing foreign bodies, involves sequential interactions with early endosomes, late endosomes, and lysosomes to yield phagolysosomes, where apoptotic cells are degraded (Flannagan et al., 2012; Wang and Yang, 2016).

As the key regulators of membrane trafficking, Rab GTPases act at multiple steps to mediate various membrane-remodeling events, leading to maturation of phagosomes and formation of phagolysosomes capable of digesting phagosomal contents (Flannagan et al., 2012; Gutierrez, 2013). In worms, four Rabs (RAB-5, UNC-108/Rab2, RAB-14, and RAB-7) function in a stepwise manner to promote phagosome maturation and cell corpse degradation. RAB-5 transiently associates with early phagosomes to promote phosphatidylinositol 3-phosphate (PtdIns3P) generation, probably by activating the phosphoinositide-3 kinase VPS-34, whereas RAB-7 is recruited later to mediate phagolysosome formation, probably through HOPS complex components (Kinchen et al., 2008; Yu et al., 2008; Xiao et al., 2009). Progression of phagosome maturation requires transition from RAB-5–positive early phagosomes to RAB-7–positive late phagosomes. The GTPase-activating protein TBC-2 inactivates RAB-5 to release it from phagosomal membranes, thereby promoting progression of phagosome maturation through the RAB-5–positive stage (Li et al., 2009). In addition, SAND-1/Monl acts with CCZ-1/Ccz1 to regulate RAB-5–to–RAB-7 transition, and thus promotes progression from the RAB-5–positive to the RAB-7–positive stage (Kinchen and Ravichandran, 2010). It is unclear whether TBC-2– and SAND-1/CCZ-1–dependent mechanisms coordinate and whether additional mechanisms are involved in this process.

UNC-108/Rab2 and RAB-14/Rab14 act in parallel to promote cell corpse degradation through phagosome maturation, but the exact steps at which they function remains unclear (Lu et al., 2008; Mangahas et al., 2008; Guo et al., 2010). UNC-108/Rab2 transiently associates with apoptotic cell–containing phagosomes, which requires RAB-5 function, suggesting that it acts downstream of RAB-5 activation (Guo et al., 2010). As well as removing apoptotic cells, UNC-108/Rab2 regulates endosome-to-lysosome maturation and maturation of dense core vesicles (DCVs; Chun et al., 2008; Lu et al., 2008; Edwards et al., 2009; Sumakovic et al., 2009). How UNC-108/Rab2 is recruited to and activated on the target membrane in phagosome, endosome, and DCV maturation processes remains unaddressed.

As the molecular switches for a variety of membrane trafficking events, Rab GTPases oscillate between GDP-bound inactive and GTP-bound active forms under the control of multiple regulatory proteins. Prenylated GDP-bound Rabs in the cytosol or on membranes are bound to GDP dissociation inhibitor (GDI), which delivers Rabs to and retrieves them from the target membrane (Seabra and Wasmeier, 2004). The membrane targeting and subsequent activation of Rab proteins require dissociation of Rabs from the GDI complex, followed by exchange of GDP for GTP catalyzed by guanine nucleotide exchange factor (GEF; Pfeffer and Aivazian, 2004; Seabra and Wasmeier, 2004; Barr, 2013). GTP-bound active Rabs interact with effector proteins to achieve downstream functions and are subsequently inactivated by GTPase activating protein (GAP), which promotes GTP hydrolysis and therefore cycles Rabs to the GDP-bound inactive state (Barr and Lambright, 2010). GDI extracts GDP-bound Rabs from the target membrane to stabilize them in the cytosol or return Rabs to the original membrane for further rounds of membrane insertion and Rab activation. GDP-bound prenylated Rabs associate tightly with GDI, and disruption of certain Rab–GDI complexes may be facilitated by GDI displacement factor (GDF; Pfeffer and Aivazian, 2004). Yip3/PRA1 has GDF activity toward endosomal Rabs, and loss of its function affects membrane association of Rab9 (Dirac-Svejstrup et al., 1997; Sivars et al., 2003). On the other hand, SidM/DrrA, a Legionella type IV effector, regulates membrane cycling of Rab1 by performing both GDI displacement and nucleotide exchange functions, indicating that GDF and GEF activity can be promoted by a single protein (Ingmundson et al., 2007; Machner and Isberg, 2007; Schoebel et al., 2009; Suh et al., 2010; Zhu et al., 2010). Whether eukaryotic proteins can catalyze coupled GDI displacement and nucleotide exchange like SidM/DrrA remains to be determined.

Here we identified GOP-1, a conserved protein homologous to Drosophila melanogaster Ema and human CLEC16A, as the activator of UNC-108/Rab2 to promote apoptotic cell clearance and maturation of endosomes and DCVs. Loss of GOP-1 abrogates recruitment of UNC-108/Rab2 to phagosomes and endosomes, causing defects in maturation of phagosomes, endosomes, and DCVs. GOP-1 interacts specifically with GDP-bound and nucleotide-free UNC-108/Rab2, disrupts GDI-UNC-108 complexes, and promotes activation and membrane association of UNC-108/Rab2 in vitro. Expression of human CLEC16A substitutes for GOP-1 function in all three processes, indicating conserved roles of this protein family in Rab2 activation.

UNC-108/Rab2 and RAB-14 function in parallel to promote cell corpse degradation (Guo et al., 2010). We performed a genome-wide RNAi screen to search for genes whose loss of function enhances the persistent cell corpse phenotype of rab-14(lf) mutants and thus may act in the same pathway with UNC-108/Rab2 to remove apoptotic cells. This screen identified gop-1, which encodes a conserved protein homologous to human CLEC16A (C-type lectin domain family 16, member A) and Drosophila Ema (endosomal maturation defective; Fig. S1 A). All three proteins contain a conserved uncharacterized domain (FPL domain) at the N terminus, followed by a predicted short transmembrane span (Fig. S1 A). CLEC16A has a putative sugar-binding C-type lectin domain that is not conserved in GOP-1 or Ema (Fig. S1 A; Kim et al., 2010).

We examined the appearance of cell corpses, which are identified by their raised button-like morphology and surface-exposed phosphatidylserine, in wild-type and gop-1 mutants (Fig. 1 A; Wang et al., 2007; Cheng et al., 2017). We found that three deletion mutants of gop-1 (tm5384, tm5694, and tm5654) and a nonsense mutant (yq79) that changes Arg 49 to a premature stop codon all contained significantly more cell corpses than wild type at various embryonic stages and in adult germline (Fig. 1, B and C; and Fig. S1, B–D). Because all gop-1 alleles showed similar cell corpse phenotypes, we performed all subsequent experiments in tm5384, a deletion mutation of gop-1 that causes an early stop codon after Ile 134 (Fig. S1 B). We found that cell deaths occurring 200–380 min after the first embryonic cleavage were similar in wild type and gop-1(tm5384), indicating that gop-1 mutants do not have excessive cell deaths (Fig. 1 D). In contrast, cell corpses persisted significantly longer in gop-1 mutants than in wild type, suggesting that cell corpse removal is defective in gop-1 mutants (Fig. 1, E and F). Expression of a GFP fusion of GOP-1 controlled by the gop-1 promoter or the ced-1 promoter, which drives gene expression in engulfing cells but not dying cells, fully rescued the persistent cell corpse phenotype in gop-1(tm5384) mutants, suggesting that GOP-1 acts in engulfing cells to promote cell corpse removal (Fig. S2, A and B). Consistent with this, GFP expression driven by the gop-1 promoter (P_gop-1GFP) was observed from embryonic stages throughout larval and adult stages in several engulfing cell types such as pharyngeal muscle cells, intestine cells, and gonadal sheath cells (Fig. S3, A–D′ and G–H′). GFP expression was also seen in coelomocytes and dorsal and ventral nerve cord, consistent with the function of GOP-1 in endosome and DCV maturation (see below; Fig. S3, E–F′, I, and I′). Expression of human CLEC16A driven by the ced-1 promoter efficiently rescued the cell corpse phenotype of gop-1(tm5384), indicating that CLEC16A can substitute for C. elegans GOP-1 in removing apoptotic cells (Fig. S2, A and B).

unc-108;rab-14 double mutants contain significantly higher numbers of cell corpses than either of the single mutants alone (Fig. 1, G and H; Guo et al., 2010), indicating that these genes function in parallel to promote apoptotic cell clearance. We found that gop-1(tm5384) contained similar numbers of embryonic and germ cell corpses as unc-108(n3263), rab-14(tm2095), or unc-108;gop-1 worms, whereas gop-1;rab-14 double mutants accumulated significantly more cell corpses than either of the single mutants alone or the unc-108;gop-1 double mutants, a phenotype very similar to unc-108;rab-14 (Fig. 1, G and H). These data suggest that GOP-1 functions in the same pathway with UNC-108 to remove cell corpses.

Cell corpses were similarly surrounded by the phagocytic receptor CED-1 or a plasma membrane reporter Myri-GFP in both wild type and gop-1(tm5384), suggesting that recognition and initiation of engulfment are not affected (Fig. 2, A and B). However, more gop-1(tm5384) phagosomes were labeled by the early phagosome markers RAB-5 and 2xFYVE, the PtdIns3P biosensor, whereas association of the late phagosome marker RAB-7 and the lysosomal reporter LAAT-1 was greatly reduced, suggesting impaired phagosome maturation (Fig. 2, A, B, and P). Similar results were also observed in unc-108(lf) and unc-108;gop-1 worms (Fig. 2, A, B, and P). The lumen of phagosomes enclosing apoptotic cells is gradually acidified during the maturation process, which requires parallel functions of UNC-108 and RAB-14 (Li et al., 2009; Guo et al., 2010). We found that significantly fewer germ cell corpses were labeled by Lysosensor green in gop-1(tm5384) worms, as in unc-108(n3263) or rab-14(tm2095) mutants, suggesting that gop-1 is required for phagosomal acidification (Fig. 2, C–E′ and I). Loss of gop-1 greatly enhanced the acidification defect in rab-14(lf) but not unc-108(lf) mutants (Fig. 2, F–I). The lysosomal membrane protein LAAT-1 was recruited to a majority of wild-type phagosomes, but the proportion of LAAT-1–positive phagosomes was significantly reduced in gop-1(tm5384), unc-108(n3263), or rab-14(tm2095) single mutants or unc-108;gop-1 double mutants (Fig. 2, J–M′ and P). In gop-1;rab-14 and unc-108;rab-14 double mutants, recruitment of LAAT-1::GFP to phagosomes was almost completely blocked (Fig. 2, N–P). These data indicate that GOP-1 functions in the same pathway with UNC-108 and in parallel to RAB-14 to promote phagosomal acidification and phagolysosome formation.

To further examine the phagosome maturation process in gop-1(lf) and unc-108(lf), we performed time-lapse analyses to follow the dynamics of RAB-5 and PtdIns3P on phagosomes. In wild type, RAB-5 and 2xFYVE appeared on phagosomes transiently, with a mean duration of 4.7 and 6.8 min, respectively (Videos 1 and 3 and Fig. 3, A–D). In gop-1(tm5384), however, RAB-5 and 2xFYVE persisted significantly longer on phagosomes, with a mean duration of 15.2 and 65 min (Videos 2 and 4 and Fig. 3, A–D). The phagosomal duration of RAB-5 and 2xFYVE was also prolonged in unc-108(n3263) mutants (Fig. 3, A–D). These data suggest that loss of GOP-1 and UNC-108 may affect progression of phagosome maturation through the RAB-5–positive stage. To examine this, we followed the dynamics of RAB-7 and LAAT-1, which are recruited to phagosomes at late maturation steps. In wild type, RAB-7 associated with phagosomes for a mean of 51 min, and RAB-7–positive phagosomes shrank gradually in size and eventually disappeared, indicating completion of cell corpse degradation and recycling of phagosomal membranes (Fig. 4 A). In unc-108(n3263) and gop-1(tm5384) worms, however, most phagosomes were transiently associated with RAB-7, with a mean duration of 18.1 and 21.3 min (Fig. 4 A). Moreover, unc-108 and gop-1 phagosomes failed to shrink, consistent with defects in phagosome maturation and cell corpse degradation (Fig. 4 A). MTM-1 is a plasma membrane–localizing PtdIns3P phosphatase that associates with forming phagosomes and is released after they seal (Cheng et al., 2015). In wild type, LAAT-1 appeared on the phagosomal surface 5.4 min after phagosome sealing, as indicated by MTM-1 release, and the LAAT-1–positive phagosome gradually shrank (Fig. 4 B). In unc-108(n3263) and gop-1(tm5384) mutants, MTM-1 release was unaffected, but LAAT-1 recruitment was severely impaired (Fig. 4, C–F). LAAT-1 was absent from 53% and 41% of phagosomes in unc-108(n3263) and gop-1(tm5384), respectively (Fig. 4, C and E), whereas 47% and 59% of unc-108(n3263) and gop-1(tm5384) phagosomes failed to recruit LAAT-1 until 26 and 23.9 min after MTM-1 release, which was significantly later than wild type (5.4 min; Fig. 4, B, D, and F). Together, these data suggest that loss of GOP-1 and UNC-108 affects progression of phagosome maturation through the RAB-5–positive stage, causing impaired RAB-7 association and phagolysosome formation.

TBC-2 is the GAP of RAB-5 that inactivates RAB-5 to promote phagosome maturation. Loss of tbc-2 causes prolonged phagosomal accumulation of RAB-5 and PtdIns3P and impaired RAB-7 recruitment, phenotypes closely resembling those in gop-1(lf) and unc-108(lf) mutants (Li et al., 2009; Fig. 3, A–D). We found that loss of tbc-2 significantly enhanced the cell corpse phenotype in gop-1(lf) and unc-108(lf) embryos and further extended the duration of RAB-5 and 2xFYVE on unc-108(n3263) phagosomes (Fig. 3, A–D; and Table S1). Loss of tbc-2 in germ line, however, did not obviously affect clearance of germ cell corpses or alter the cell corpse numbers in gop-1 and unc-108 mutants, suggesting that TBC-2 plays a major role in removing embryonic but not germ cell corpses (Table S2). SAND-1/Mon1 regulates the switch from RAB-5–positive to RAB-7–positive phagosomes in the C. elegans germ line (Kinchen and Ravichandran, 2010). Loss of sand-1 significantly increased germ cell corpse numbers in gop-1(lf) and unc-108(lf) worms (Table S2). These data suggest that GOP-1 and UNC-108 function in parallel to TBC-2 and SAND-1 to promote phagosome progression through the RAB-5–positive stage.

GFP::GOP-1 displayed both cytoplasmic and vesicular localization patterns, with the majority of GOP-1–positive puncta containing UNC-108, suggesting colocalization of the two proteins (Fig. 5, A–A″). The GOP-1– and UNC-108–positive vesicles either overlapped with or were near puncta labeled by FAPP1-PH, the trans-Golgi marker, suggesting that they localize to the trans-Golgi or vesicles closely associated with it (Fig. 5, B–C″). GOP-1 was absent from lysosomes (NUC-1), whereas a few GOP-1–positive vesicles colocalized with or were near RAB-5–positive early endosomes and RAB-7–positive late endosomes (Fig. S3, J–L″).

We observed GFP::GOP-1 rings surrounding apoptotic cells, suggesting that GOP-1 associates with phagosomes (Fig. 5, D, D′, G, and G′). rab-5 RNAi but not rab-7 RNAi or loss of unc-108 disrupted phagosomal association of GOP-1, indicating that RAB-5 function but not RAB-7 or UNC-108 is required for recruiting GOP-1 to phagosomes (Fig. 5, E–K). Loss of tbc-2, which causes persistent activation of RAB-5 and sustained PtdIns3P accumulation on phagosomes, led to an increase in GOP-1–positive phagosomes (Fig. 5, F, F′, J, and K).

We next investigated whether GOP-1 is responsible for recruiting UNC-108 to phagosomes. UNC-108 was recruited to phagosomes shortly after MTM-1 release and stayed on them for ∼6.2 min in wild-type embryos (Video 5 and Fig. 6, A and F). This indicates that UNC-108 is recruited to and transiently associates with fully sealed phagosomes, consistent with its role in phagosome maturation. In gop-1(tm5384) embryos, however, MTM-1 dynamics were unaffected, but UNC-108 was not recruited to phagosomes (Video 6 and Fig. 6, B and F). Similarly, UNC-108 was enriched transiently on phagosomes enclosing germ-cell corpses in wild-type sheath cells but was absent from phagosomes in gop-1(tm5384) worms (Fig. 6, C–E and G). These data indicate that GOP-1 is required to recruit UNC-108 to phagosomes.

In addition to removing apoptotic cells, UNC-108/Rab2 plays important roles in endosome and DCV maturation. Loss of UNC-108 affects endocytic trafficking, causing accumulation of aberrant endosome/lysosome hybrids (Chun et al., 2008; Lu et al., 2008). We found that gop-1(tm5384) worms contained enlarged vacuoles in coelomocytes, a phenotype that was efficiently rescued by expression of GOP-1::GFP (Fig. 7, A and B; and Fig. S2, C–E and H). Similar to unc-108(lf), the large vacuoles in gop-1(tm5384) coelomocytes were labeled by both RME-8::GFP, the endosome marker, and LMP-1::GFP, a lysosomal membrane protein, suggesting that they are aberrant endosome/lysosome hybrids (Fig. 7, A–C′, E–G′, I, and J). Identical phenotypes were observed in unc-108;gop-1 double mutants (Fig. 7, D–D′, H–H′, I, and J). Texas red–conjugated BSA (TR-BSA), which was taken up by wild-type coelomocytes and transported to lysosomes within 30 min of injection into the body cavity, was trapped in the large vacuoles in gop-1(tm5384) coelomocytes, a phenotype resembling that of unc-108 mutants (Fig. S4, A and B; Lu et al., 2008). These data suggest that GOP-1 promotes endosome maturation like UNC-108.

UNC-108 regulates DCV maturation by preventing specific cargo from entering the late endosome system and therefore retaining them in the mature DCV (Edwards et al., 2009; Sumakovic et al., 2009). We found that loss of gop-1 led to greatly reduced NLP-21::VENUS, a soluble DCV cargo, and IDA-1::GFP, a DCV-specific membrane protein, in neuron axons and the appearance of big NLP-21::VENUS vesicles in neuron cell bodies, phenotypes that were indistinguishable from those of unc-108(lf) single mutants or unc-108;gop-1 double mutants (Fig. 7, K–T; and Fig. S4, C–G). In addition, loss of gop-1 caused increased GLR-1::GFP in the ventral nerve cord like unc-108(lf) and unc-108;gop-1, suggesting that GOP-1 also plays a role in trafficking of the glutamate receptor GLR-1 like UNC-108 (Fig. S4, H–L; Chun et al., 2008). Consistent with defects in DCV maturation and GLR-1 trafficking, gop-1(tm5384) mutants displayed an uncoordinated locomotion phenotype similar to unc-108(3263) single mutants and unc-108;gop-1 double mutants (Fig. S2 I). These data indicate that GOP-1 acts in the same pathway with UNC-108 to promote endosome and DCV maturation. Expression of human CLEC16A efficiently rescued the endosome maturation and locomotion defects of gop-1(tm5384), indicating that CLEC16A can replace GOP-1 in these processes (Fig. S2, F, H, and I).

As on phagosomes, UNC-108 appeared transiently on endosomes for ∼9.3 min in wild-type coelomocytes (Video 7 and Fig. 8 A). UNC-108 release coincided with or immediately followed the appearance of the lysosomal marker NUC-1, consistent with the role of UNC-108 in endosome maturation (Fig. 8 A). In gop-1(tm5384) mutants, however, endosomal recruitment of UNC-108 was abrogated (Video 8 and Fig. 8 B). In contrast, GOP-1::GFP associated with endosomes in both wild-type and unc-108(lf) coelomocytes (Fig. 8, C–E). These data indicate that GOP-1 is required to recruit UNC-108 to endosomes, whereas UNC-108 is dispensable for endosomal association of GOP-1.

Next, we determined whether GOP-1 physically interacts with UNC-108. GOP-1 interacted with UNC-108 but not RAB-5, RAB-7, or RAB-14 in yeast two-hybrid assays, suggesting that GOP-1 is a specific binding partner of UNC-108 (Figs. 9 A and S5 A). Using various mutant forms of UNC-108 that fix the Rab in different nucleotide-binding states, we found that GOP-1 interacted with GDP-bound UNC-108(S20N) and the nucleotide-free mutant of UNC-108(N119I), but not wild-type or GTP-bound UNC-108(Q65L) (Fig. 9 A). Consistent with this, significantly more HIS-tagged GOP-1 was pulled down by GST-UNC-108(S20N) and GST-UNC-108(N119I) than wild-type UNC-108 or UNC-108(Q65L) (Fig. 9 B). These data indicate that GOP-1 interacts specifically with the inactive GDP-bound and nucleotide-free UNC-108, consistent with its role in membrane recruitment and activation of UNC-108. All GOP-1–positive phagosomes contained UNC-108, and GFP::GOP-1 coprecipitated with Cherry::UNC-108 in worms expressing both proteins, indicating that GOP-1 associates with UNC-108 in vivo (Fig. 9, C–D′′′).

The activation of Rab proteins requires dissociation of Rab from the GDI–Rab complex, followed by GEF-catalyzed nucleotide exchange. We purified GDI-1–UNC-108 complexes from yeast cells in which Rabs can be prenylated (Fig. S5 B). Wild-type UNC-108 but not UNC-108 (ΔCC), which lacks the two prenylatable C-terminal cysteines, associated with GDI-1, suggesting that UNC-108 in complex with GDI-1 is prenylated (Figs. S2 J and S5 C; Musha et al., 1992; Farnsworth et al., 1994). We incubated the GDI-1–UNC-108 complex with GOP-1 immobilized on GST beads and found that UNC-108 bound strongly to GST-GOP-1 but not GST-only beads, whereas GDI-1 remained in the supernatant (Fig. 9 E). When purified UNC-108–GDI-1 was immobilized on magnetic beads and incubated with soluble GOP-1, GOP-1 was coprecipitated with the UNC-108–coated magnetic beads, whereas the association of GDI-1 with the beads was reduced (Fig. S5 D). These data suggest that GOP-1 disrupts the Rab–GDI complex, causing dissociation of UNC-108 from GDI-1. Next, to examine whether GOP-1 promotes membrane recruitment of UNC-108, we performed a flotation assay using PC liposomes containing PtdIns3P to mimic PtdIns3P-positive phagosomes (Fig. 10 A). Incubation of GOP-1 with UNC-108–GDI-1 led to association of UNC-108 but not GDI-1 with PC + PtdIns3P liposomes, and this was enhanced by addition of GTPγS, which prevents GTP hydrolysis (Fig. 10 A). Recombinant GOP-1 or UNC-108 by itself, however, did not associate with liposomes (Fig. S5, E and F). We found that GOP-1 strongly facilitated loading of GTPγS onto UNC-108 when incubated with UNC-108–GDI-1 in the presence of ³⁵S-labeled GTPγS (Fig. 10 B). Collectively, these data suggest that GOP-1 promotes activation and membrane association of UNC-108. To further examine this, we mutated lysine 120, which resides in the highly conserved guanine base-binding motif, in UNC-108 (Fig. S2 J; Pereira-Leal and Seabra, 2000). The analogous mutation in Ypt7, which weakens the affinity for the nucleotide and thus facilitates nucleotide exchange, bypasses the necessity for its GEF (Kucharczyk et al., 2001; Cabrera and Ungermann, 2013). We found that K120E, but not wild-type UNC-108, associated with phagosomes and endosomes in the absence of GOP-1 (Fig. 10, C–G). Moreover, expression of UNC-108(K120E) rescued the germ-cell corpse phenotype and partially reversed the endosome maturation defect in gop-1(tm5384) mutants, consistent with GOP-1 acting as an activator of UNC-108 (Fig. S2, A, G, and H).

We identified gop-1 as a novel regulator of apoptotic cell clearance that promotes phagosome maturation by activating UNC-108/Rab2. GOP-1 transiently associates with apoptotic cell–containing phagosomes, a process that requires RAB-5 function and may be facilitated by membrane PtdIns3P. Because GOP-1 did not interact with RAB-5 or associate with PtdIn3P (Fig. S5, A and E), its recruitment to phagosomes may be mediated by effectors of RAB-5 or PtdIns3P. UNC-108/Rab2 is recruited to apoptotic cell–containing phagosomes by GOP-1, which promotes UNC-108 activation and thus stabilizes the Rab on the phagosomal surface. Our data suggest that GOP-1–UNC-108 promotes phagosome progression from the RAB-5–positive stage to the RAB-7–positive stage, a process that also involves TBC-2–mediated RAB-5 inactivation and SAND-1–mediated RAB-7 recruitment (Li et al., 2009; Kinchen and Ravichandran, 2010). We found that loss of tbc-2 and sand-1 significantly enhanced the persistent cell corpse and phagosome maturation phenotypes in gop-1(lf) and unc-108(lf) mutants, suggesting that parallel mechanisms are involved. Consistent with this, GOP-1 and UNC-108 do not associate with TBC-2 or affect its recruitment to phagosomes (unpublished data). Thus, multiple mechanisms are engaged to control phagosome maturation through the RAB-5–positive stage. Future work should identify UNC-108/Rab2 effectors and reveal the UNC-108/Rab2-dependent mechanism in this process.

UNC-108 regulates multiple processes including apoptotic cell degradation and endosome and DCV maturation, whereas mammalian Rab2 controls protein secretion and recycling through the Golgi apparatus. How Rab2 is targeted to and activated on the specific membrane compartment in these processes remains unknown. Here we present five lines of evidence that GOP-1 is an upstream activator of UNC-108/Rab2. First, unc-108 and gop-1 mutants exhibit identical phenotypes in apoptotic cell removal and maturation of endosomes and DCVs; these defects are unaltered in unc-108;gop-1 double mutants, indicating that GOP-1 acts together with UNC-108 in all three processes. Second, loss of GOP-1 disrupts phagosomal and endosomal association of UNC-108, whereas UNC-108 is dispensable for recruiting GOP-1 to membranes. Third, GOP-1 binds to the inactive GDP-bound and nucleotide-free UNC-108, but not the active GTP-bound UNC-108. Fourth, GOP-1 disrupts UNC-108–GDI-1 complexes and promotes GTP loading onto and membrane association of UNC-108. Fifth, mutation of lysine 120 of UNC-108, which resides in the conserved guanine base–contacting motif and may facilitate nucleotide exchange, bypasses the necessity for GOP-1 in cell corpse removal and endosome maturation. Thus, GOP-1 promotes UNC-108 activation in multiple processes.

Rab activation requires its disassociation from the Rab–GDI complex and subsequent nucleotide exchange catalyzed by GEF. We found that GOP-1 disrupts UNC-108–GDI-1 complexes and promotes loading of GTP onto and membrane association of UNC-108 by using Rab–GDI as the substrate. GOP-1 binds strongly to the nucleotide-free form of UNC-108, a hallmark of GEF; loss of GOP-1 completely disrupts phagosomal and endosomal targeting of UNC-108, whereas UNC-108(K120E), a mutation that may facilitate nucleotide exchange of Rab, reverses the membrane targeting defect of UNC-108 in gop-1 mutants. Together, these data suggest that GOP-1 may act as a GEF of UNC-108 that serves both GDI displacement and nucleotide exchange functions. However, by in vitro GEF assays, we failed to detect GEF activity of GOP-1 toward UNC-108 (unprenylated or prenylated) in the presence or absence of liposomes. Notably, many GEFs, including those of Ras, Ran, Rho, Arf, and Rab, are regulated through autoinhibitory mechanisms (Cherfils and Zeghouf, 2013). For example, RABEX-5, the RAB5 GEF, forms a complex with RABAPTIN-5, an effector of RAB5 that suppresses the autoinhibition of RABEX-5 (Horiuchi et al., 1997; Delprato and Lambright, 2007). We suspect that the nucleotide exchange activity of GOP-1 may be regulated in vivo by additional modifications, cofactors, or autoinhibitory elements and is thus not revealed in vitro by recombinant GOP-1. Further work, especially structural studies of GOP-1, should clarify this issue.

GOP-1 shares sequence homology with Drosophila Ema and human CLEC16A, both of which associate with endolysosomal compartments and/or the Golgi complex, like GOP-1, and are implicated in endosomal maturation and autophagosomal trafficking (Kim et al., 2010, 2012; Soleimanpour et al., 2014; van Luijn et al., 2015). These findings indicate conserved functions of this gene family in membrane trafficking, a process extensively regulated by Rab GTPases. Our data indicate that GOP-1 promotes activation of Rab GTPase 2 in multiple processes. Moreover, expression of CLEC16A in worms efficiently rescues the persistent cell corpse, endosomal maturation, and locomotion phenotypes of gop-1(lf) mutants, indicating that CLEC16A can substitute for GOP-1 in all these processes. Thus, GOP-1 family proteins may serve as a novel family of Rab activators to regulate multiple membrane trafficking events. In genome-wide association studies, CLEC16A is associated with several autoimmune diseases such as type 1 diabetes mellitus, multiple sclerosis, and primary adrenal insufficiency, but how its deficiency contributes to disease pathogenesis remains largely unknown (Hakonarson et al., 2007; Skinningsrud et al., 2008; Hoppenbrouwers et al., 2009; Márquez et al., 2009). Our findings reveal the essential role of GOP-1 in the activation of Rab GTPase 2, which may contribute to the pathogenesis of diseases associated with CLEC16A deficiency.

Strains of C. elegans were cultured and maintained using standard protocols. The N2 Bristol strain was used as the wild-type strain. The following strains were used in this work: linkage group I (LG I), unc-108(n3263); LG II, tbc-2(qx20); LG III, gop-1(tm5384, tm5654, tm5694, yq79); LG IV, sand-1(ok1963); LG V, unc-76(e911); and LG X, rab-14(tm2095).

Standard microinjection methods were used to generate transgenic animals carrying extrachromosomal arrays (qxEx), and genome-integrated arrays (qxIs) were acquired by γ-ray irradiation to achieve stable expression from arrays with low copy numbers. The reporter strains used in this study are as follows: smIs76 (P_hspAnnexin-V::GFP), qxIs477 (P_ced-1mCherry::MTM-1(C378S)), qxIs408 (P_ced-1GFP::RAB-5), qxIs456 (P_ced-1mCherry::RAB-5), qxIs66 (P_ced-1GFP::RAB-7), qxIs68 (P_ced-1mCherry::RAB-7), qxIs234 (P_ced-1GFP::UNC-108), qxIs466 (P_ced-1mCherry::UNC-108), qxIs257 (P_ced-1NUC-1::nCherry), qxIs390 (P_ced-1myriGFP), qxIs458 (P_ced-1GFP::GOP-1), qxEx4387 (P_gop-1GOP-1::GFP), qxIs354 (P_ced-1LAAT-1::GFP), smIs34 (P_ced-1CED-1::GFP), qxEx4320 (P_gop-1GFP), qxEx6226 (P_ced-1CLEC16A-FLAG), qxEx7115 (P_ced-1GFP::UNC-108(K120E)), qxEx7118 (P_unc-122GFP::UNC-108(K120E)), qxEx7264 (P_gop-1CLEC16A-FLAG), qxEx5088 (P_ced-1GFP::GOP-1+P_ced-1mCherry::FAPP1-PH), qxEx5638 (P_ced-1GFP::UNC-108+P_ced-1mCherry::FAPP1-PH), qxEx5199 (P_ced-1GFP::GOP-1+P_ced-1mCherry::TRAM-1), and qxEx5636 (P_unc-122GFP::UNC-108).

We obtained opIs334 (P_ced-1YFP::2xFYVE) from K.S. Ravichandran (University of Virginia, Charlottesville, VA) and M.O. Hengartner (University of Zurich, Zurich, Switzerland); ceIs72 (P_unc-129IDA-1::GFP) from K.G. Miller (Oklahoma Medical Research Foundation, Oklahoma City, OK); nuIs183 (P_unc-129NLP-21::Venus) and nuIs24 (P_glr-1GLR-1::GFP) from J.M. Kaplan (Massachusetts General Hospital, Boston, MA); bIs34 (P_cc1RME-8::GFP) from H. Fares (University of Arizona, Tucson, AZ); pwIs50 (LMP-1::GFP) from B. Grant (Rutgers University, New Brunswick, NJ); and itIs37 (P_pie-1mCherry::H2B) from Z. Bao (Sloan Kettering Institute, New York, NY).

Yeast strain BCY123 (MATa, Can1, ade2, trp1, Ura3-52, his3, leu2-3, 112, pep4::his⁺, prb1::leu2⁺, bar1::HisG⁺, and lys2::pGAL1/10-GAL4⁺) was provided by J. Zhou (Institute of Biochemistry and Cell Biology, Shanghai Institutes for Biological Sciences, Shanghai, China).

gop-1 deletion alleles were isolated from a trimethylpsoralen and ultraviolet-mutagenized library, as described previously (Gengyo-Ando and Mitani, 2000). gop-1 mutant alleles were identified by nested PCR with the following primers: 5′-TGGGTCACTATGGAAGCCGA-3′ and 5′-GCTACGCATGTGTCCCTTTA-3′ as first-round primers and 5′-AAGCATTACGAGCTATCGCA-3′ and 5′-CTTACTGTCTGAAGAGGAGT-3′ as second-round primers for tm5384 and tm5694, or 5′-GCGTAGCTTGTTGTGTGGGT-3′ and 5′-CGTGCTGTTTGGCGACCCTA-3′ as first-round primers and 5′-CTCGCCGCGCTTTGTTTGAT-3′ and 5′-CGTTACCTTGGTGAGCCGTT-3′ as second-round primers for tm5654. The yq79 mutation was isolated from a forward genetic screen for additional regulators of cell corpse clearance. yq79 was mapped to the left arm of LG III between genetic map positions −3.79 (Snp-WBVar00058951) and −1.57 (Snp-WBVar00060777) by single nucleotide polymorphism mapping. The complementation test indicated that yq79 and tm5384 affected the same gene. The sequence of the gop-1 gene was determined in all gop-1 alleles. The tm5384 deletion allele contains a 511-bp deletion that causes a frame shift after Tyr 111, resulting in an early stop codon after Ile 134. The tm5654 allele of gop-1 contains a 1,185-bp deletion that removes most of exon 8, all of exons 9–13, and part of exon 14, resulting in a frame shift after Val 698. The tm5694 allele of gop-1 contains a 739-bp deletion that removes part of exon 3, all of exons 4 and 5, and part of exon 6, resulting in a truncated protein lacking aa 140–340. yq79 contains a C-to-T mutation that causes a premature stop codon after Arg 49. All mutants were backcrossed with N2 animals at least four times before further analyses.

The number of somatic cell corpses in the head region of living embryos and the number of germ cell corpses in one gonad arm at various adult ages were scored as described (Gumienny et al., 1999; Wang et al., 2002). The cell corpses were identified by their raised button–like morphology using Nomarski optics. At least 15 animals were scored at each stage in each strain. To examine embryonic cell deaths and cell corpse duration, embryos at the two-cell stage (cell death occurrence) or precomma stage (embryonic cell copse duration) or adult animals (24 h after L4/adult molt, germ cell corpse duration) were mounted on agar pads. Images in 40 z-sections (1.0 µm/section, cell death occurrence) or 30 z-series (1.0 µm/section, cell corpse duration) were captured every 2 min for 8 h (cell death occurrence) or 2–4 h (cell corpse duration) using an Axioimager M1 microscope (ZEISS) equipped with an AxioCam monochrome digital camera (ZEISS). Images were processed and viewed using Axiovision Rel 4.7 software (ZEISS).

To examine phagosome maturation, differential interference contrast (DIC) and fluorescence images were captured using an Axioimager A1 microscope (ZEISS) as described before (Lu et al., 2008). The percentage of somatic and germ cell corpses labeled by various phagosomal markers was quantified in embryos at the 1.5-fold stage or adult animals (24 or 48 h after L4/adult molt) by dividing the total number of cell corpses by the number of labeled ones. Corpses were examined at multiple focal planes. At least 15 animals were scored in each strain.

DIC and fluorescent images were captured with an Axioimager A1 equipped with epifluorescence (Filter Set 13 for GFP [excitation, BP 470/20; beam splitter, FT 495; emission, BP 503–530] and Filter Set 20 for Cherry [excitation, BP 546/12; beam splitter, FT 560; emission, BP 575–640]) and an AxioCam monochrome digital camera. Images were processed and viewed using Axiovision Rel. 4.7 software. A 100× Plan-Neofluar objective (NA 1.30) was used with Immersol 518F oil (ZEISS). For confocal images, a ZEISS LSM 5 Pascal inverted confocal microscope with 488 (emission filter, BP 503–530) and 543 (emission filter, BP 560–615) lasers was used, and images were processed and viewed using LSM Image Browser software (ZEISS). All images were taken at 20°C.

The dynamic association of MTM-1::Cherry(C378S), GFP::UNC-108, GFP::RAB-5, YFP::2xFYVE, GFP::RAB-7, LAAT-1::GFP, and NUC-1::Cherry with phagosomal or endosomal membranes was recorded by spinning-disk time-lapse recording as described previously (Cheng et al., 2015). In brief, adult animals (24–36 h after L4/adult molt) or embryos at the precomma stage were mounted on agar pads in M9 buffer in the presence (adult worms) or absence (embryos) of 5 mM levamisole to prevent animals from moving without affecting the germline or the gonad. Fluorescent images were captured using a 100× objective (NA 1.45) with immersion oil (Type NF) on an inverted fluorescence microscope (Nikon Eclipse Ti-E) with an UltraView spinning-disc confocal scanner unit (PerkinElmer) with 488 (emission filter 525 [W50]) and 561 (dual-band emission filter 445 [W60] and 615 [W70]) lasers. Images in 15–20 z-sections (1.0 µm/section) were captured every 1 min for 2 h (embryos), 2–5 h (adult animals), or 2–3 h (coelomocytes in adult animals) at 20°C. Volocity software (PerkinElmer) was used to view and analyze images. The numbers of phagosomes or endosomes that were followed and quantified are indicated in the figures and figure legends.

In vivo pulse-chase experiments were performed as described (Zhang et al., 2001). In brief, TR-BSA was injected into the body cavity in the pharyngeal region at 1 mg/ml. Injected worms were transferred to seeded NGM plates at RT, and the uptake by coelomocytes was recorded at the indicated time points (5, 15, 30, and 60 min). At each time point, the injected worms were transferred to an ice-cold NGM plate to stop the intracellular trafficking of endocytosed molecules before examination by epifluorescence microscopy.

Live animals (12 h after L4/adult molt) expressing NLP-21::VENUS, IDA-1::GFP, or GLR-1::GFP were mounted on agar pads in M9 in the presence of 5 mM levamisole. All images were taken with equal exposure time with an LSM 5 Pascal inverted confocal microscope (ZEISS) with the 488 laser (emission filter, BP 503–530). Images were processed and viewed using LSM Image Browser software. To examine NLP-21::VENUS, IDA-1::GFP, and GLR-1::GFP patterns, the posterior region of the dorsal nerve cord (NLP-21, IDA-1) or ventral nerve cord (GLR-1) was oriented toward the objective as described (Sumakovic et al., 2009). Images in five z-series (1.0 µm/section) were captured. Maximum stack projections of images were obtained using LSM Image Browser, and the fluorescence intensity was measured by ImageJ and normalized to wild type. To examine NLP-21::VENUS in neuronal cell bodies in ventral nerve cord, images of entire cell bodies were captured in eight z-series (1.0 µm/section). The diameter of each NLP-21::VENUS–positive vesicle was quantified using ImageJ.

Lysosensor green staining was performed as described before (Guo et al., 2010). In brief, adult worms (48 h after L4/adult molt) were dissected in gonad dissection buffer (60 mM NaCl, 32 mM KCl, 3 mM Na₂HPO₄, 2 mM MgCl₂, 20 mM Hepes, 50 µg/ml penicillin, 50 µg/ml streptomycin, 100 µg/ml neomycin, 10 mM glucose, 33% FCS, and 2 mM CaCl₂) with 1 µM Lysosensor green DND-189 (Invitrogen) and examined by fluorescence microscopy.

The locomotion assay was performed as described before (Sumakovic et al., 2009). Adult animals were transferred to unseeded NGM plates to give an initial adjustment period of 60 min. Locomotion was then assayed by counting the number of body bends (corresponding to a whole 360° sine wave) in a period of 3 min.

The bacteria feeding protocol was used in RNAi experiments. In each experiment, L4 larvae (P0) were cultured on the RNAi plates. The F1 progeny at the L4 stage were transferred to fresh RNAi plates and aged for 48 h before examination of germ cell corpses. Embryonic cell corpses were scored in the F2 progeny. For rab-5 RNAi, 30 L2 or L3 larvae (P0) were cultured on the RNAi plates, and germ cell corpses were scored 48 h after L4/adult molt at the same generation because most F1 progeny die before reaching the adult stage as a result of inactivation of rab-5. To quantify somatic cell corpses in rab-5 RNAi worms, L4 larvae (P0) were cultured on the RNAi plates, and cell corpses were examined in the F1 embryos. The DNA sequences that are targeted by the dsRNAs in the RNAi experiments are as follows: rab-7 (W03C9, 32,688–34,320 nt), rab-5 (F26H9, 20,787–22,312 nt), unc-108 (F53F10, 9,645–10,724 nt), and gop-1 (C34E10, 27,648–28,758 nt).

Yeast two-hybrid analyses were performed using the Matchmaker yeast two-hybrid system according to the manufacturer’s instructions (Takara Bio, Inc.). Specific cDNAs to be tested were cloned into the pGADT7 and pGBKT7 vectors to produce Gal4 transcription activation domain (AD) and DNA binding domain (BD) fusion proteins. Specific pairs of constructs expressing AD and BD fusion proteins were transformed into the yeast strain AH109, and transformants were selected on synthetic complete (SC) medium lacking leucine and tryptophan (SC-Leu-Trp). Individual clones were streaked on SC-Ade-His–Leu-Trp plates to test for the activation of the reporter genes HIS3 and ADE2. Self-interaction of GOP-1 or UNC-108 was also tested.

Full-length cDNA of unc-108 (WT/S20N/Q65L/N119I) and gop-1 was cloned into the pET41b vector, and GST-fusion proteins were produced in E. coli BL21(DE3), then purified using glutathione-Sepharose beads (Sigma-Aldrich). Full-length cDNA of gop-1 was also cloned into the pET21b vector to produce HIS-tagged proteins in E. coli BL21(DE3), which were purified using Ni-NTA resin (QIAGEN). pCMV-GOP-1::FLAG was transfected into HEK293T cells to produce GOP-1-FLAG protein, which was purified using anti-FLAG M2 agarose (Sigma-Aldrich).

The wild-type and mutant forms of GST::UNC-108 were immobilized on glutathione-agarose beads, and 3 mM GDP and GTP were loaded onto GST-UNC-108(S20N)– and GST-UNC-108(Q65L)–containing beads, respectively, to lock UNC-108 in the GDP- and GTP-bound conformation. GOP-1-HIS was incubated with immobilized GST-UNC-108(WT), GST-UNC-108(S20N), GST-UNC-108(Q65L), or GST-UNC-108(N119I) for 1 h at 4°C in a buffer containing 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 1 mM DTT, and 5 mM MgCl₂. The resins were washed three times in wash buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM DTT, 5 mM MgCl₂, and 0.1% CHAPS), and the bound proteins were resolved with SDS-PAGE and detected by immunoblotting using anti-His antibody. The protein level of GST-UNC-108 was revealed by Coomassie blue staining.

Adult worms expressing both GFP::GOP-1 and mCherry::UNC-108 or mCherry::UNC-108 alone were lysed with FastPrep-24 (MP Biomedicals) in worm lysis buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM MgCl₂, and protease inhibitor cocktail (EDTA-free; Roche). The homogenized solutions were centrifuged at 1,500 rpm for 10 min at 4°C to remove debris. 500-µl precleared solutions were incubated with 10 µl GBP beads (GFP binding protein; ChromoTek) for 2 h at 4°C. After extensive washing with 50 mM Tris-HCl, pH 7.5, and 150 mM NaCl, the bound proteins were resolved with SDS-PAGE and revealed by Western blotting with antibodies against GFP (Roche) and mCherry (HX1810; Huaxingbio).

UNC-108–GDI-1 complex was purified from yeast cells as described before (Machner and Isberg, 2007). In brief, pYES3-FLAG-UNC-108 and pYES2-MYC-HIS-GDI-1 constructs were cotransformed into BCY123 (MATa, Can1, ade2, trp1, Ura3-52, his3, leu2-3, 112, pep4::his⁺, prb1::leu2⁺, bar1::HisG⁺, and lys2::pGAL1/10-GAL4⁺) and selected on SC selective medium lacking tryptophan and uracil (SC-Trp-Ura). A single colony of the transformed yeast was inoculated into SC selective medium containing 2% glucose and grown overnight at 30°C with shaking. Yeast cells were transferred into induction medium (SC selective medium containing 2% galactose and 1% raffinose) at a starting OD₆₀₀ of 1.0 to induce protein expression for 18 h. Cells were then harvested and lysed by FastPrep in a buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 2 mM MgCl₂. The cell lysate was centrifuged and first incubated with Ni-NTA resin for 2 h to separate the FLAG-UNC-108-MYC-HIS-GDI-1 complex from unbound yeast proteins. The bound proteins were eluted with 250 mM imidazole and further incubated with anti-FLAG M2 agarose (Sigma-Aldrich) overnight to separate the FLAG-UNC-108-MYC-HIS-GDI-1 complex from free MYC-HIS-GDI-1. The bound UNC-108–GDI-1 complex was eluted by addition of 3xFLAG peptides (Sigma-Aldrich).

The GTP loading assay was performed as described previously, with modifications (Zhu et al., 2010). 15 pmol UNC-108–GDI-1 complex was incubated with 2 pmol ³⁵S-labeled GTPγS in reaction buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 2 mM MgCl₂ at RT. 15 pmol GOP-1-FLAG was then added to initiate the nucleotide-exchange reaction. Samples were taken at the indicated time points, and the reaction was stopped by adding 500 µl ice-cold wash buffer (50 mM Tris-HCl, 50 mM NaCl, and 2 mM MgCl₂). [³⁵S]GTPγS bound to UNC-108 was analyzed by filter binding followed by scintillation counting.

Recruitment of UNC-108 onto liposomes was analyzed as previously described, with modifications (Wang et al., 2011). In brief, 10 pmol UNC-108–GDI-1 complex was incubated with 1 mM PC liposomes containing 8% PtdIns3P without or with GOP-1 (10 pmol) or GTPγS (100 pmol) as indicated for 2 h at 4°C, followed by a 30-min incubation at RT in a buffer containing 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 2 mM MgCl₂. The samples were analyzed by sucrose gradient ultracentrifugation. The top layer containing floating PC liposomes was resolved by SDS-PAGE and analyzed by Western blotting.

To examine GDI displacement activity, GST and GST-GOP-1 immobilized on glutathione-agarose beads were incubated with FLAG-UNC-108-MYC-HIS-GDI-1 complex in reaction buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 2 mM MgCl₂) for 10 min at 25°C followed by centrifugation. The supernatant was collected, and the beads were washed five times. The unbound proteins in the supernatant and proteins bound to beads were subjected to SDS-PAGE and detected by Western blotting. In another experiment, FLAG-UNC-108-MYC-HIS-GDI-1 complex immobilized on FLAG M2 magnetic beads (Sigma-Aldrich) was incubated with GOP-1-HIS for 60 min at 25°C in reaction buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, and 2 mM MgCl₂), followed by centrifugation. The supernatant was collected, and the beads were washed five times. The unbound proteins in the supernatant and proteins bound to beads were subjected to SDS-PAGE and detected by Western blotting.

The SD was used as y-axis error bars for bar charts plotted from the mean value of the data. Data derived from different genetic backgrounds were compared by Student’s unpaired t test, one-way analysis of variance (ANOVA) followed by Tukey’s posttest, or two-way ANOVA followed by Bonferroni posttest, as indicated in the figure legends. Data were considered statistically different at P < 0.05.

To generate P_ced-1GFP::GOP-1, gop-1 genomic DNA was amplified by primers PYWW96/PYWW97 and cloned into pPD49.26-P_ced-1GFP1 through the KpnI site. To generate P_gop-1GFP, the promoter region of gop-1 (5,020 bp) was amplified using primers PJHY24/PJHY25 and cloned into pPD49.26-GFP through the SphI and BamHI sites. P_gop-1GOP-1::GFP was constructed by cloning the gop-1 genomic sequence into pPD49.26-P_gop-1GFP at the XmaI and KpnI sites. To generate P_ced-1hCLEC16A-FLAG and P_gop-1hCLEC16A-FLAG, human CLEC16a was amplified from a human cDNA library using primers PPFG656/PPFG631 and cloned into pPD49.26-P_ced-1 and pPD49.26-P_gop-1, respectively, through the KpnI site. To generate pGADT7-GOP-1, full-length gop-1 cDNA was amplified from a C. elegans cDNA library using primers PPFG581/PJHY75 and cloned into pGADT7 at the NdeI and BamHI sites. pET21b-GOP-1 was constructed by ligating gop-1 cDNA amplified using primers PPFG581 and PJHY118 into the pET21b vector at the NdeI and BamHI sites. To generate pET41b-GOP-1, full-length gop-1 cDNA was amplified using primers PPFG626/PJHY75 and cloned into pET41b at the SpeI and BamHI sites. pCMV-GOP-1-FLAG was constructed by inserting gop-1-flag amplified from pGADT7-GOP-1 using primers PYWW96/PPFG648 into pCMY-MYC(-) through the KpnI site. pYES3-FLAG-UNC-108 was constructed by inserting FLAG-UNC-108 amplified by primers PPFG246/PJHY111 into the pYES3 vector at the KpnI and PmeI sites. To generate pET28a-GDI-1, full length cDNA of gdi-1 was amplified from a C. elegans cDNA library using primers pJHY94/PJHY95 and inserted into the pET28a vector at the NdeI and BamHI sites. pYES2-MYC-HIS-GDI-1 was generated by ligating MYC-HIS-GDI-1 amplified from pET28a-GDI-1 by primers PJHY135 and PPFG81 into the pYES2 vector at the KpnI and XhoI sites. To generate pGBKT7-UNC-108, unc-108 cDNA was amplified from a C. elegans cDNA library using primers PPFG152 and PXCW232 and inserted into pGBKT7 at the NcoI–BamHI sites. The S20N, Q65L, and N119I mutations were introduced into pGBKT7-UNC-108 by site-directed mutagenesis (QuikChange; Stratagene) using the primer pairs PPFG37/PPFG38 (S20N), PPFG39/PPFG40 (Q65L), and PPFG51/PPFG52 (N119I). pET41b-UNC-108 (WT/S20N/Q65L/N119I) was constructed by inserting UNC-108 (WT/S20N/Q65L/N119I) amplified from pGBKT7-UNC-108 (WT/S20N/Q65L/N119I) using primers PXCW259/PJHY130 into pET41b at the SpeI–BamHI sites. pGBKT7-RAB-5, pGBKT7-RAB-7, and pGBKT7-RAB-14 were generated by ligating rab-5, rab-7, and rab-14 cDNAs amplified from a C. elegans cDNA library using primers PWZ7/PWZ8 (RAB-5), PWZ13/PWZ14 (RAB-7), and PWZ23/PWZ24 (RAB-14) into pGBKT7 through the NdeI and BamHI sites (Table 1).

Fig. S1 shows that gop-1 encodes an evolutionarily conserved protein. Fig. S2 shows that expression of human CLEC16A rescues the phagosome, endosome, and locomotion defects of gop-1 mutants. Fig. S3 shows that gop-1 is widely expressed. Fig. S4 shows that gop-1 mutants display defects in endosome and dense core vesicle maturation. Fig. S5 shows that GOP-1 interacts with inactive UNC-108 and disrupts the UNC-108-GDI-1 complex. Video 1 shows that RAB-5 transiently associates with apoptotic cell–containing phagosomes. Video 2 shows that loss of gop-1 causes prolonged association of RAB-5 with phagosomes. Video 3 shows that PtdIns3P transiently accumulates on phagosomes. Video 4 shows that loss of gop-1 causes persistent accumulation of PtdIns3P on phagosomes. Video 5 shows that UNC-108 transiently associates with apoptotic cell–containing phagosomes. Video 6 shows that loss of gop-1 abrogates phagosomal recruitment of UNC-108. Video 7 shows that UNC-108 transiently associates with endosomes in coelomocytes. Video 8 shows that loss of gop-1 abolishes recruitment of UNC-108 to endosomes. Table S1 shows that loss of tbc-2 enhances the somatic cell corpse phenotype of gop-1 and unc-108 mutants. Table S2 shows that loss of sand-1 enhances the germ cell corpse phenotype of gop-1 and unc-108 mutants.

We thank Drs. K. Ravichandran, M.O. Hengartner, B. Grant, H. Fares, K.G. Miller, J.M. Kaplan, and Z. Bao for strains and Dr. Isabel Hanson for editing services.

Some strains were provided by the Caenorhabditis Genetics Center, which is funded by the National Institutes of Health Office of Research Infrastructure Programs (P40OD010440). This work was supported by the National Science Foundation of China (31325015), the National Basic Research Program of China (2013CB910100 and 2014CB849702), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB19000000), Chinese Ministry of Science and Technology (2016YFA0500203), and an International Early Career Scientist grant from the Howard Hughes Medical Institute to X. Wang.

The authors declare no competing financial interests.

Author Contributions: J. Yin and Y. Huang performed and designed experiments along with X. Wang and P. Guo, and S. Hu performed some of the genetic and cell biological experiments. S. Yoshina, N. Xuan, Q. Gan, S. Mitani, and C. Yang contributed to the generation of strains. X. Wang wrote the paper with input from J. Yin, C. Yang, and S. Mitani.