The RBG-1–RBG-2 complex modulates autophagy activity by regulating lysosomal biogenesis and function in C. elegans

Multiple aspects of intracellular vesicle trafficking such as vesicle budding, transport, tethering and fusion are controlled by RAB small GTPases (Stenmark, 2009; Wandinger-Ness and Zerial, 2014). RAB proteins cycle between GTP-bound active and GDP-bound inactive forms, accompanied by membrane association and dissociation (Stenmark, 2009; Mizuno-Yamasaki et al., 2012; Blümer et al., 2013; Wandinger-Ness and Zerial, 2014). Guanine nucleotide exchange factors (GEFs) mediate the exchange of GDP for GTP, which in turn recruits downstream effectors (Stenmark, 2009; Mizuno-Yamasaki et al., 2012; Blümer et al., 2013; Wandinger-Ness and Zerial, 2014), whereas GTPase-activating proteins (GAPs) stimulate GTP hydrolysis to terminate the RAB effect (Stenmark, 2009; Mizuno-Yamasaki et al., 2012; Blümer et al., 2013; Wandinger-Ness and Zerial, 2014). GDP-dissociation inhibitors (GDIs) facilitate the extraction of membrane-associated GDP-bound RAB proteins into the cytosol. Targeting of the cytosolic RAB–GDI complex onto cognate membranes requires membrane-bound GDI displacement factors (GDFs) and also GEFs (Stenmark, 2009; Mizuno-Yamasaki et al., 2012; Blümer et al., 2013; Wandinger-Ness and Zerial, 2014). The cycling of RAB proteins between their GTP- and GDP-bound forms is critical for membrane dynamics. The concerted actions of GEF and GAP proteins ensure tight spatiotemporal control of the activity of RAB proteins in membrane trafficking (Stenmark, 2009; Mizuno-Yamasaki et al., 2012). RAB7 is located on late endosomes and lysosomes and participates in diverse dynamic membrane-based processes through the recruitment of different effector proteins. These processes including maturation of early endosomes to late endosomes, membrane fusion of late endosomes (homotypic fusion and also heterotypic fusion of late endosomes with lysosomes), lysosomal biogenesis, and positioning of late endosomes and lysosomes (Bucci et al., 2000; Hyttinen et al., 2013; Guerra and Bucci, 2016; Langemeyer et al., 2018). Little is known about the factors that regulate the dynamics of membrane-associated RAB7 to control lysosomal biogenesis.

Macroautophagy (hereafter referred to as autophagy) involves the formation of a double-membrane autophagosome, and its delivery to lysosomes (metazoans) or the vacuole (yeast) for degradation of sequestrated materials (Nakatogawa et al., 2009; Feng et al., 2014). In multicellular organisms, nascent autophagosomes fuse with endolysosomal vesicles to form amphisomes, a process known as autophagosome maturation, before eventually forming degradative autolysosomes (Zhao and Zhang, 2019). Upon degradation of sequestrated materials, lysosomes regenerate from autolysosomes to maintain lysosome homeostasis (Yu et al., 2010). Autophagosome maturation requires the concerted actions of RAB small GTPases, tethering factors and the SNARE complex (Zhao and Zhang, 2019). Fusion of autophagosomes with late endosomes and lysosomes is driven by the STX17–SNAP29–VAMP8 (VAMP7 in fly and worm) and YKT6–SNAP29–STX7 SNARE complexes (Itakura et al., 2012; Jiang et al., 2014; McEwan et al., 2015; Bas et al., 2018; Gao et al., 2018b; Matsui et al., 2018). The dynamics of GTP- and GDP-bound RAB7 is essential for formation of functional autolysosomes (Zhao and Zhang, 2019). Activated RAB7 recruits tethering factors such as PLEKHM1 and the HOPS complex to late endosomes and lysosomes to promote the stability and assembly of the trans-SNARE complex, mediating autophagosome–lysosome fusion (Jiang et al., 2014; McEwan et al., 2015). RAB7 is also targeted to autophagosomes by its GEF, the MON1–CCZ1 complex, which is recruited to autophagosomes via direct interaction with ATG8 to promote fusion of autophagosomes with lysosomes and vacuoles (Hegedűs et al., 2016; Vaites et al., 2017; Gao et al., 2018a). Inactivation of RAB7 by TBC-domain-containing RAB GAP protein Armus (TBC1D2), which is targeted to autophagosomes via interaction with the LC3 (MAP1LC3) protein family, also facilitates autophagosome maturation (Carroll et al., 2013).

The metazoan-specific autophagy protein EPG5 functions as a tethering factor to confer the fusion specificity of autophagosomes with late endosomes and lysosomes (Tian et al., 2010; Wang et al., 2016). EPG5 is a RAB7 effector and localizes on late endosomes and lysosomes. It directly interacts with autophagosomal-localized LC3 and facilitates assembly of the STX17–SNA29–VAMP8 SNARE complex (Wang et al., 2016). Loss of EPG5 function causes non-specific fusion of autophagosomes with other endocytic vesicles such as recycling endosomes, resulting in the formation of non-degradative autolysosomes (Wang et al., 2016). In addition to autophagy, EPG5 is essential for endocytic recycling and degradation (Zhao et al., 2013a). EPG5 knockdown slows endocytic degradation and delays endocytic recycling (Zhao et al., 2013a). Recessive mutations in human EPG5 are causatively linked to the multisystem disorder Vici syndrome, key features of which include agenesis of the corpus callosum, myopathy and combined immunodeficiency (Dionisi et al., 1998; Cullup et al., 2013). Epg5 knockout (KO) mice display some phenotypic similarities with individuals with Vici syndrome, including corpus callosum changes and myopathy (Zhao et al., 2013a,b). Epg5 KO mice, however, exhibit elevated basal lung inflammation (Lu et al., 2016). Accumulation of non-degradative autolysosomes has been observed in Vici syndrome patient tissues, in epg-5-depleted C. elegans and in Epg5-deficient mice (Cullup et al., 2013; Zhao et al., 2013a; Wang et al., 2016), but the mechanism underlying this accumulation remains unknown.

Here, we demonstrate that loss of function of rbg-1 and rbg-2, encoding RBG-1 and RBG-2 (RAB3GAP1 and RAB3GAP2 in mammals), respectively, ameliorates the autophagy defect in C. elegans epg-5 mutants. RAB3GAP1 and RAB3GAP2 form a complex that acts as a GAP for RAB3 and a GEF for RAB18 (Fukui et al., 1997; Nagano et al., 1998; Gerondopoulos et al., 2014). Loss of rbg-1 activity facilitates lysosomal biogenesis and function in a manner independent of its catalytic activity on RAB-3 and RAB-18. The reduced dynamics of membrane-associated RAB-7 and accumulation of non-degradative autolysosomes in epg-5 mutants are suppressed by simultaneous inactivation of rbg-1. Expression of the GDP-bound form of RAB-7 has the same effect as loss of function of rbg-1 activity on lysosomal biogenesis and epg-5 suppression. Our results indicate that RBG-1 regulates RAB-7 cycling to modulate lysosomal biogenesis and function.

To investigate the mechanism underlying the accumulation of non-degradative autolysosomes in epg-5 mutants, we performed genetic screens to identify mutations that suppress the accumulation of aggregates of the C. elegans p62 (also known as SQSTM1) homolog SQST-1 in epg-5(tm3425) null mutants. The bpIs267 line, in which SQST-1::GFP is specifically expressed in the hypodermis, was used to facilitate the screen. In wild-type animals, SQST-1::GFP is weakly expressed and diffusely localized in the cytoplasm (Fig. 1A,B), while in epg-5 mutants a large number of SQST-1 aggregates accumulate throughout development (Fig. 1C). From EMS screens, a mutation, bp1150, was isolated that dramatically reduced the number of SQST-1::GFP aggregates in epg-5 mutants (Fig. 1D,E). Accumulation of SQST-1::GFP aggregates was also suppressed by bp1150 in epg-5 mutants with transgenic expression of SQST-1 in other tissues, including the whole body using the endogenous sqst-1 promoter (Psqst-1::SQST-1::GFP), the intestine (Pges-1::SQST-1::GFP), muscle (Phlh-1::SQST-1::GFP) and neurons (Punc-119::SQST-1::GFP) (Fig. S1A–Y). Compared to epg-5 single mutants, accumulation of endogenous autophagy substrates, including SQST-1 and components of PGL granules, was also dramatically reduced in epg-5; bp1150 mutants (Fig. 1F–J; Fig. S1Z–I1). SQST-1::GFP aggregates in epg-5; bp1150 mutants were restored by simultaneous loss of autophagy activity, such as that caused by loss of lgg-1, the C. elegans homolog of Atg8 (GABARAP in vertebrates) (Fig. 1K–O). This suggests that the bp1150 mutation promotes autophagic degradation of SQST-1 in an EPG-5-independent manner during development.

A large number of non-degradative autophagic structures labeled by LGG-1 accumulate in epg-5 mutants (Tian et al., 2010). Accumulation of LGG-1 puncta in epg-5 mutant embryos and also in the hypodermis and intestine of epg-5 mutant larvae was reduced in animals also harboring the bp1150 mutation (Fig. 1P–T; Fig. S1J1–S1). Taken together, these results indicate that bp1150 suppresses the autophagy defect in epg-5 mutants.

We cloned the bp1150 mutation by transformation rescue experiments. A transgene expressing rbg-1 restored the accumulation of SQST-1::GFP aggregates in epg-5; bp1150 hypodermal cells (Fig. 2A–D). The bp1150 mutation was determined to contain a tryptophan-to-stop codon mutation at residue 718 in RBG-1 (Fig. S2A), and the mutation is hereafter termed rbg-1(bp1150). rbg-1(RNAi) also reduced SQST-1::GFP aggregates in epg-5 mutants (Fig. S2B), indicating that the suppression effect was caused by loss of rbg-1 activity. RBG-1 forms a complex with RBG-2. rbg-2(RNAi) suppressed the accumulation of SQST-1::GFP aggregates in epg-5 mutants (Fig. 2E–H; Fig. S2C,D), suggesting that RBG-1 functions with RBG-2 to regulate the degradation of autophagy substrates in epg-5 mutants.

The RAB3GAP1-RAB3GAP2 complex acts as a GAP for RAB3 and a GEF for RAB18 (Fukui et al., 1997; Nagano et al., 1998; Gerondopoulos et al., 2014). Transgenes expressing RBG-1(R667A), which impairs the RAB3 GAP activity, or RBG-1(T16P) and RBG-1(E22V), which disrupt the RAB18 GEF activity, retained rescuing activity (Fig. 2I–L). Loss of function of rab-3 and rab-18 failed to restore the accumulation of SQST-1::GFP aggregates in epg-5; rbg-1 mutants (Fig. S2E), and did not reduce the number of SQST-1::GFP aggregates in epg-5 mutants (Fig. S2F). These results indicate that loss of rbg-1 activity suppresses the autophagy defect in epg-5 mutants in a RAB-3- and RAB-18-independent manner.

We next examined whether loss of function of rbg-1 suppresses the autophagy defect in other autophagy mutants with phenotypes that can be suppressed by enhanced autophagy activity, such as through mTOR inactivation. Loss of function of epg-7, which encodes the scaffold protein facilitating degradation of SQST-1, caused accumulation of SQST-1::GFP aggregates (Lin et al., 2013). rpl-43(bp399) mutants, which display impaired function of the ribosomal protein RPL-43, accumulate SQST-1::GFP aggregates in the intestine (Guo et al., 2014); however, SQST-1::GFP aggregates in both epg-7 and rpl-43 mutant larvae are suppressed by elevated autophagy activity (Lin et al., 2013; Guo et al., 2014). The rbg-1(bp1150) mutation did not reduce the number of SQST-1::GFP aggregates in epg-7 or rpl-43 mutants (Fig. S3A,B). Simultaneous depletion of rbg-1 did not suppress, but instead slightly increased, the SQST-1::GFP aggregates in bec-1(bp613) mutants, which exhibit partially impaired function of the C. elegans homolog of BECN1. (Fig. 2M–P). This is consistent with a previous study showing that loss of rbg-1 activity causes a mild autophagy defect (Spang et al., 2014).

We also determined whether loss of rbg-1 activity ameliorates the accumulation of SQST-1 aggregates caused by impaired lysosomal function. The number of SQST-1::GFP aggregates in mutants of cpl-1, encoding cathepsin L, was reduced in larvae also harboring the rbg-1(bp1150) mutation (Fig. 2Q–T). rbg-1(RNAi) failed to suppress the autophagy defect in mutants of scav-3 (Fig. S3C), which encodes a lysosomal membrane protein involved in maintaining lysosome integrity (Li et al., 2016). Thus, loss of rbg-1 activity appears to suppress the autophagy defect resulting from impaired lysosomal degradation.

The specific alleviation of the autophagy defect in epg-5 and cpl-1 mutants by loss of rbg-1 prompted us to examine lysosome biogenesis in rbg-1 mutants. mCherry-tagged lysosomal-localized DNase II NUC-1 and the late endosome marker GFP::RAB-7 label spherical structures in the wild-type hypodermis (Fig. 3A; Fig. S3D). In rbg-1 animals, NUC-1::mCherry and GFP::RAB-7 labeled spherical and tubular structures that were smaller and more abundant than those in wild-type animals (Fig. 3B,M,N; Fig. S3E). Ultrastructural analysis by transmission electron microscopy (TEM) also revealed that lysosomes in the hypodermis in rbg-1 mutants were smaller than those in wild-type animals (Fig. 3E,F).

We performed time-lapse analysis of the delivery of the sfGFP::mCherry tandem-tagged NUC-1 to monitor the transport of lysosomal enzymes. The NUC-1::sfGFP::mCherry reporter is driven by the heat shock promoter. The animals carrying this reporter are heat-shocked at 33°C for 30 min and then shifted to 20°C. The newly synthesized NUC-1 produces both red and green fluorescent signals. The GFP signal is quenched inside acidic lysosomes, and thus red-only signals represent acidified lysosomal structures. In wild-type animals, NUC-1::sfGFP::mCherry was detected 2 h after heat shock (Fig. 4A). The number of yellow puncta gradually decreased after 12 h and only red puncta persisted after 24 h (Fig. 4B–E,E1). In rbg-1 mutants, the number of yellow puncta was similar to that in wild-type animals at 2 h after heat shock (Fig. 4F,E1), but the number of red-only puncta was significantly higher at 8 and 12 h after heat shock (Fig. 4G–J,E1). These results indicate that loss of rbg-1 activity accelerates the delivery of NUC-1::sfGFP::mCherry and/or acidification of lysosomes. We also examined lysosomal acidification using the NUC-1::pHTomato reporter, which is driven by the heat shock promoter. pHTomato is highly sensitive to pH and the fluorescence intensity is weaker in lower pH compartments (Li and Tsien, 2012). At 24 h after heat shock, the fluorescence signal of lysosomal-localized NUC-1::pHTomato was weaker in rbg-1 mutants than in wild-type animals (Fig. 3I,J,O), indicating that lysosomal acidification is enhanced by loss of rbg-1 activity.

In epg-5 mutants, in addition to small spherical structures, NUC-1::mCherry labeled abnormally enlarged structures (Fig. 3C). GFP::RAB-7 also labeled enlarged structures, and there were fewer GFP-positive tubular structures (Fig. S3F). In epg-5; rbg-1 double mutants, the number of abnormally enlarged GFP::RAB-7-labeled or NUC-1::mCherry-labeled structures was reduced, accompanied by formation of more small spherical and tubular structures, resembling the phenotype in rbg-1 single mutants (Fig. 3D,M,N; Fig. S3G). EM analysis also showed that epg-5 mutants contained enlarged lysosomal structures and small lysosomes with abnormal appearance in the hypodermis (Fig. 3G; Fig. S3M,N). Both of these phenotypes were suppressed by simultaneous depletion of rbg-1 (Fig. 3H). We further determined whether loss of rbg-1 activity restored the lysosomal function in epg-5 mutants. Compared to wild-type animals, epg-5 mutants exhibited a defect in NUC-1::sfGFP::mCherry delivery and/or acidification, and many yellow puncta persisted at 24 h after heat shock (Fig. 4K–O,E1). In epg-5; rbg-1 double mutants, the delivery of NUC-1::sfGFP::mCherry was similar to rbg-1 mutants and the GFP signal disappeared at 24 h after heat shock (Fig. 4P–T,E1). Compared to wild-type animals, the fluorescence signal of NUC-1::pHTomato was stronger in epg-5 mutants (Fig. 3K,O), indicating that lysosomal acidification is impaired in epg-5 mutants. The enhanced fluorescence intensity of NUC-1::pHTomato in epg-5 mutants was suppressed by loss of function of rbg-1 (Fig. 3L,O). Taken together, these results provide evidence that loss of rbg-1 activity suppresses the defect in lysosomal acidification in epg-5 mutants.

In epg-5 mutants, non-degradative autolysosomes are labeled by markers for a variety of endosomal vesicles (Wang et al., 2016). The recycling endosomal marker GFP::RAB-10 labeled a few punctate structures in the hypodermis of wild-type and rbg-1 animals (Fig. S3H,I). GFP::RAB-10 labeled enlarged puncta in epg-5 mutants, and this was suppressed in epg-5; rbg-1 double mutants (Fig. S3J–L). In wild-type and rbg-1 animals, BFP::LGG-1 formed a few small puncta that were separable from GFP::RAB-10-labeled structures (Fig. 5A,B,G). In epg-5 mutants, BFP::LGG-1 labeled big puncta that largely co-localized with the enlarged NUC-1::mCherry- and GFP::RAB-10-labeled structures (Fig. 5C,G) (Wang et al., 2016). The number of enlarged LGG-1-labeled structures was dramatically reduced and their co-localization with GFP::RAB-10-labeled punctate structures also decreased in epg-5; rbg-1 double mutants (Fig. 5D,G). These results indicated that formation of abnormally enlarged vesicles with mixed identities in epg-5 mutants is suppressed by loss of rbg-1 activity.

To determine the relative levels of GDP- and GTP-bound RAB-7, we examined the levels of RAB-7 in worm extracts that are pulled down by RILP, a RAB7 effector that specifically binds to GTP-bound RAB7 (Frasa et al., 2010). RAB-7 is conserved from C. elegans to mammals. GST–RILP interacted with RAB-7(Q68L), which mimics the active GTP-bound RAB-7, but not with RAB-7(T23N), which mimics the GDP-bound inhibitory form (Fig. S3O,P). The GST pulldown results showed that levels of GTP-bound RAB-7 were lower in rbg-1 mutants than in wild-type worms (Fig. 6A,B), suggesting a concomitant increase in levels of the GDP-bound form of RAB-7 in rbg-1 mutants. Levels of GTP-bound RAB-7 in epg-5; rbg-1 mutants were similar to rbg-1 single mutants (Fig. 6A,B). The purified RBG-1–RBG-2 complex did not show evidence of RAB-7 GEF activity (Fig. S3Q). Consistent with this, no interaction was detected between GST–RAB-7(T23N) and TRX–RBG-1 (Fig. S3R). These results suggest that RBG-1 does not act as a GEF to regulate RAB-7 activity.

In wild-type worms, the ring-like structures of GFP::RAB-7 enclosed the punctate structures formed by NUC-1, indicating that GFP::RAB-7 is membrane-associated (Fig. S3S). We performed a fluorescence recovery after photobleaching (FRAP) assay to examine the mobility of membrane-associated GFP-RAB-7. In wild-type and rbg-1 mutant hypodermis, the GFP–RAB-7 fluorescence signal recovered within 5 s after bleaching (Fig. 6C,D). However, the GFP–RAB-7 signal barely recovered in epg-5 mutants (Fig. 6C,D). Simultaneous loss of function of rbg-1 rescued the recovery of GFP–RAB-7 fluorescence in epg-5 mutants after bleaching (Fig. 6C,D). These results indicate that loss of rbg-1 activity promotes the mobility of RAB-7 on late endosomes and lysosomes in epg-5 mutants.

We further investigated whether expression of the GDP-bound mutant RAB-7(T23N) suppresses the autophagy and lysosome defect in epg-5 mutants. A low concentration of constructs expressing RAB-7(T23N) or RAB-7(Q68L) was injected and stable transgenic lines were obtained for analysis. No evident accumulation of SQST-1::GFP aggregates and GFP::LGG-1 puncta was detected in animals expressing RAB-7(T23N) or RAB-7(Q68L) (Fig. S4A–C). Expression of RAB-7(T23N), but not RAB-7(Q68L), suppressed the accumulation of SQST-1::GFP and GFP::LGG-1 puncta in epg-5 mutants (Fig. 6E–N; Fig. S4D). The increased number of SQST-1::GFP aggregates in cpl-1 mutants was also suppressed by RAB-7(T23N), but not by RAB-7(Q68L) (Fig. 6O–S). Expression of RAB-7(T23N) and RAB-7(Q68L) failed to suppress the autophagy defect in scav-3 mutants (Fig. S4E).

We next examined whether expression of RAB-7(T23N) affects lysosomal biogenesis and function in epg-5 mutants. NUC-1::mCherry labeled more small spherical structures in animals expressing RAB-7(T23N) (Fig. 6U; Fig. S4G). Expression of RAB-7(Q68L) had no evident effect on the pattern of NUC-1::mCherry-labeled structures (Fig. S4H). The delivery rate of NUC-1::sfGFP::mCherry was faster in animals expressing RAB-7(T23N), and resembled rbg-1 mutants (Fig. 4U–Y,E1). LGG-1 formed a few puncta that were separate from GFP::RAB-10-labeled structures in animals expressing RAB-7(T23N) (Fig. 5E,G). Expression of RAB-7(T23N), but not RAB-7(Q68L), suppressed the accumulation of abnormal NUC-1::mCherry-labeled structures in epg-5 mutants (Fig. 6T–X). The defective delivery rate of NUC-1::sfGFP::mCherry and the co-localization of BFP::LGG-1 with NUC-1::mCherry and GFP::RAB-10 in epg-5 mutants were also suppressed by RAB-7(T23N) expression (Fig. 4Z–E1; Fig. 5F,G). These results indicated that expression of RAB-7(T23N), similar to loss of function of rbg-1, suppresses impaired autolysosome degradation in epg-5 animals.

Different from RAB-7(T23N) expression, partial reduction of rab-7 activity using RNAi enhanced the accumulation of SQST-1::GFP aggregates in epg-5 single mutants and epg-5; rbg-1 double mutants (Fig. S4J,K). rab-7(RNAi) animals contained numerous small NUC-1::mCherry-labeled structures, suggesting a defect in lysosomal formation (Fig. S4I). These results indicate that suppression of the defect in epg-5 mutants by expression of RAB-7(T23N) does not occur through reduction of RAB-7 activity.

Here, we have uncovered a novel role of the RBG-1–RBG-2 complex in modulating lysosomal biogenesis and function. Compared to wild-type animals, rbg-1 mutants contain more spherical and tubular late endosomes and lysosome structures. Loss of rbg-1 activity accelerates lysosomal maturation and degradation, as shown by the enhanced delivery of the lysosomal enzyme NUC-1, attenuation of the impaired degradation in cpl-1 mutants, and suppression of the autophagy defect and accumulation of enlarged non-degradative autolysosomes in epg-5 mutants. The RAB3GAP1-RAB3GAP2 complex has been shown to act as a GAP for RAB3 to regulate neurotransmitter release and synaptic plasticity (Sakane et al., 2006; Müller et al., 2011). The RAB3GAP1-RAB3GAP2 complex also localizes to subdomains of the endoplasmic reticulum (ER) and functions as a GEF for RAB18 (Gerondopoulos et al., 2014). This complex is also localized on lipid droplets (LDs) and is required for targeting and activation of RAB18 on LDs for LD growth and maturation (Xu et al., 2018). The RAB3GAP1-RAB3GAP2 complex has been shown to modulate autophagosomal biogenesis and its loss of function enhances protein aggregation in muscle cells in a manner dependent on the GAP activity but not on RAB3 itself (Spang et al., 2014). The Drosophila ortholog of RBG-1 (CG31935) is also required for autophagy in muscle (Zirin et al., 2015). Consistent with this, we found that loss of rbg-1 activity slightly increases accumulation of SQST-1 aggregates in bec-1 hypomorphic mutants. The suppression effect of RBG-1 on the autophagy defect in epg-5 mutants is independent of its GAP and GEF activity. Thus, the RBG-1–RBG-2 complex has dual functions in autophagy.

RAB7 regulates multiple processes involving late endosomes and also plays a key role in lysosomal biogenesis (Bucci et al., 2000; Wang et al., 2011; Hyttinen et al., 2013; Guerra and Bucci, 2016; Wong et al., 2018). Biogenesis and maintenance of the lysosomal compartment requires the concerted actions of lysosome fission, and also of heterotypic fusion of late endocytic vesicles with pre-existing lysosomes. RAB7 controls homotypic fusion of late endosomes and their heterotypic fusion with lysosomes. The proper function of RAB7 requires a precise balance of the activated and inactivated forms (Frasa et al., 2010; Carroll et al., 2013). Overexpression of the GTP-bound form of RAB7 increases the fusion and thus the size of lysosomes, and also causes their perinuclear localization, while expression of the inhibitory mutant form of RAB7 results in dispersal of lysosomes (Bucci et al., 2000). Elevated RAB7-GTP levels in Vps34-deficient cells, which results from failed recruitment of the PI(3)P-binding RAB7 GAP Armus (also known as TBC1D2) to late endosomes, causes the formation of enlarged late endosomes and impaired lysosomal maturation (Jaber et al., 2016).

The GEF for regulating RAB-7 on lysosomes has yet to be identified. The MON1–CCZ1 complex, which is a RAB5 effector, is involved in the recruitment and activation of RAB7 on late endosomes (Kinchen and Ravichandran, 2010; Nordmann et al., 2010; Poteryaev et al., 2010). The MON1–CCZ1 complex dissociates from late endosomes prior to their fusion with lysosomes, and thus is not involved in RAB7 activation on lysosomes (Poteryaev et al., 2010; Yasuda et al., 2016). The RBG-1–RBG-2 complex appears to have been newly revealed to act as a RAB-7 regulator to modulate lysosome biogenesis and function. Levels of GTP-bound RAB-7 are lower in rbg-1 mutants than wild-type animals. Expression of RAB-7(T23N) mimics loss of rbg-1 activity in facilitating lysosomal biogenesis. One possible function of wild-type RBG-1 could be to target the RAB-7 GEF to lysosomes, or to promote its activity on lysosomes.

In the autophagy pathway, autophagosomes and/or amphisomes fuse with lysosomes to form degradative autolysosomes. Upon degradation of the sequestrated materials, lysosomes are regenerated from autolysosomes (Yu et al., 2010). The activity of RAB7 is precisely controlled to ensure the progression of autophagic flux. Depletion of the RAB7 GAP Armus impairs autolysosome formation and acidification (Carroll et al., 2013). RAB7 knockout cells show accumulation of autolysosomes, indicating that RAB7 is essential for autolysosome maturation (Kuchitsu and Fukuda, 2018). In EPG5-depleted cells or epg-5 mutant animals, GFP::RAB7 forms abnormally enlarged punctate structures (Wang et al., 2016). The reduced mobility of GFP::RAB-7 on late endosomes and lysosomes in epg-5 mutants may recruit effectors such as HOPS to promote fusion. Accumulation of active GTP-bound RAB-7 on autolysosomes also depletes the availability of cycling and/or mobile RAB-7 and causes a defect in non-lysosomal functions of RAB-7, such as delivery of lysosomal enzymes. Loss of rbg-1 activity or expression of the GDP inhibitory form of RAB-7 promotes the dynamics of membrane-associated RAB-7 in epg-5 mutants. The release of RAB-7 from autolysosomes promotes the transport of lysosomal enzymes and/or lysosomal membrane proteins to increase the acidification of lysosomes. These effects restore the autolysosome function and ameliorate the autophagy defect in epg-5 mutants. Accumulation of non-degradative autolysosomes has been observed in Vici syndrome patient tissues (Cullup et al., 2013). Our study provides new insights into potential therapeutic treatments of Vici syndrome.

Strains of C. elegans were cultured and maintained on nematode growth medium (NGM) according to standard protocols (Brenner, 1974) at 20°C unless indicated otherwise. The Hawaiian strain CB4856 was used for polymorphism mapping.

The following strains were used in this work: N2 Bristol (wild-type), epg-5(tm3425), rbg-1(bp1150), bec-1(bp613), epg-7(tm2508), rpl-43(bp399), cpl-1(qx304), scav-3(qx193), bpIs267(P_hyp-7SQST-1::GFP, unc-76), bpIs262(P_ges-1SQST-1::GFP, unc-76), bpIs151(P_sqst-1SQST-1::GFP, unc-76), bpIs193(P_hlh-1SQST-1::GFP, unc-76), bpIs328(P_unc-119SQST-1::GFP, unc-76), adIs2122(P_lgg-1GFP::LGG-1, rol-6(su1006)), bpEx342(P_hyp-7BFP::LGG-1, P_myo-2::GFP), qxIs686(P_hyp-7GFP::RAB-10, unc-76), qxIs66(P_ced-1GFP::RAB-7, unc-76), qxIs257(P_ced-1NUC-1::mCherry, unc-76), qxIs750(P_hspNUC-1::pHTomato, P_ord-1::GFP), qxIs612(P_hspNUC-1::sfGFP::mCherry, unc-76) and bpIs395(P_nfya-1RAB-7(T23N), rol-6(su1006)). Further details are provided in Table S1.

bec-1(bp613) contains an aspartic acid to asparagine mutation at amino acid 347. The autophagy defect in bec-1(bp613) is much weaker than in bec-1(ok700) null mutants.

Mapping using genetic and SNP markers placed bp1150 on chromosome X, between 21.23 cM and 21.97 cM. A transgene containing fosmid WRM0637bE01 or the single gene rbg-1 restored the accumulation of SQST-1 aggregates in epg-5; bp1150 mutants.

Constructs were injected into worms at a concentration of ∼20 ng/μl to obtain stable transgenic lines. Transgenes analyzed in this study include: P_hyp-7::RBG-1, P_hyp-7::RBG-1(R667A), P_hyp-7::RBG-1(T16P), P_hyp-7::RBG-1(E22V), P_nfya-1::RAB-7(T23N) and P_nfya-1::RAB-7(Q68L). Further details are provided in Table S1.

For RNAi injection experiments, single-stranded RNA was transcribed from T7 and SP6 promoter-flanked PCR templates. ssRNAs were annealed and injected into wild-type animals or animals carrying the indicated transgene. The F1 progeny at the indicated developmental stages were examined for the phenotype.

The primers used for RNAi synthesis were: lgg-1 5′-CACTAGTAATACGACTCACTATAGGGATGAAGTGGGCTTACAAGG-3′, 5′-CACTAGATTTAGGTGACACTATAGAACCTTCTTTTCGACCTCTCCT-3′; rbg-1 5′-CACTAGTAATACGACTCACTATAGGGCTTGAACCGAACAACGAAGG-3′, 5′-CACTAGATTTAGGTGACACTATAGAAGCATCAAGTGGACCTTCATG-3′; rbg-2 5′-CACTAGTAATACGACTCACTATAGGGGGAAATGTTCGGGTATGTGC-3′, 5′-CACTAGATTTAGGTGACACTATAGAACATCAATATCCTCTTGAGAC-3′; rab-18 5′-CACTAGTAATACGACTCACTATAGGGGCGTGGATTTCCGCGTAACT-3′, 5′-CACTAGATTTAGGTGACACTATAGAACTAGCATCCACACATTCC-3′; rab-7 5′-CACTAGTAATACGACTCACTATAGGGATGTCGGGAACCAGAAAGA-3′, 5′-CACTAGATTTAGGTGACACTATAGAACTGAGATTCTCTAGCCAACG-3′.

Mutations in rab-3 were generated by CRISPR/Cas9. sgRNA was injected into epg-5; rbg-1 animals with a co-injection marker and F1 progeny were examined. The sgRNA sequence was 5′-CCAATCTTCGAAGGGTTC-3′. In rab-3(bp1558) mutants, five nucleotides (TCGTC) were deleted, which introduced a premature stop codon at amino acid 101 (RAB-3 contains 220 amino acid residues).

All genes were PCR-amplified and cloned into the pET.32M.3C vector to produce TRX–His6-tag-fused recombinant proteins, or the pGEX.6p-1 vector to produce GST-tag-fused recombinant proteins. Point mutations were introduced by the site-directed mutagenesis approach (Landt et al., 1990).

Soluble recombinant proteins were expressed in Escherichia coli BL21-CodenPlus (DE3) and purified on glutathione Sepharose 4B beads (for GST-tagged proteins, GE Healthcare) and Ni-NTA agarose beads (for His6-tagged proteins, Qiagen). After extensive washing, the proteins were eluted with a buffer containing 1×PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄) and 10 mM L-glutathione (Sigma-Adrich, 70-18-8) (for GST-tagged proteins), and 50 mM HEPES (pH 7.9), 500 mM imidazole and 500 mM NaCl (for His6-tagged proteins). The eluted proteins were then loaded onto a desalting column (GE Healthcare, 17-0851-01) and finally eluted with 1×PBS buffer.

We failed to purify E. coli expressed RBG-2. Adult worms expressing FLAG–RBG-1 and FLAG–RBG-2 were collected and lysed in lysis buffer containing 25 mM HEPES pH 7.4, 150 mM NaCl, 0.5% NP40, 1 mM EDTA, 10% glycerol, 1 mM DTT and protease inhibitor cocktail (EDTA-free; Roche, B14003). The worm extracts were clarified by centrifugation at 18,000 g for 30 min at 4°C. The FLAG-tagged proteins were isolated from the clarified worm lysate using 100 μl anti-FLAG M2 affinity gel (Sigma-Aldrich, A2220) for 4 h at 4°C. The beads were washed five times with worm washing buffer (25 mM HEPES pH 7.4, 300 mM NaCl, 1 mM EDTA, 0.5% NP40, 10% glycerol), and ten times with GEF reaction buffer (20 mM HEPES pH 6.8, 1 mg/ml BSA, 150 mM NaCl, 1 mM MgCl₂). The proteins were separated on 10% SDS-PAGE gels for analysis by immunoblotting. The proteins bound to FLAG beads were snap-frozen in liquid nitrogen for storage at −80°C and prepared for the nucleotide exchange assay.

Glutathione Sepharose 4B beads (GST, GE Healthcare, 17-0756-05) were washed in pulldown buffer (140 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 1.8 mM KH₂PO₄ and 1% NP-40) three times. 20–30 μg GST and GST-tagged proteins were incubated with 20 μl GST beads in 500 μl pulldown buffer for 0.5 h at 4°C. The supernatants were then removed. The GST beads were washed three times with pulldown buffer. 50 μg Trx-tagged proteins were incubated with the GST-tagged protein-conjugated beads in 500 μl pulldown buffer for 1 h at 4°C. The beads were then washed three times with pulldown buffer and the bound proteins were boiled in 5×SDS sample loading buffer and visualized by means of immunoblotting. To determine the interaction of RILP or RBG-1 with constitutively active GTP-bound mutant RAB proteins in vitro, GTP and Mg²⁺ were added to the buffer. For the interaction with dominant-negative GDP-bound mutant RAB proteins, GDP was added to the buffer. To examine the level of GTP-bound RAB-7 in wild type and mutants, GST–RILP bound to GST beads was incubated with worm extracts. Antibody against RAB-7 (1:1000; Chen et al., 2010) and secondary antibody (peroxidase AffiniPure goat anti-rat IgG, Jackson ImmunoResearch, 112-035-003) were used to detect the level of endogenous GTP-bound RAB-7 in worm extracts.

Animals expressing GFP–RAB-7 were cultured at 20°C and young adult animals (24 h after L4 stage) were used for analysis. The FRAP assay was performed using a confocal microscope (LSM 880 Meta plus Zeiss Axiovert zoom, Zeiss) using a 63× oil immersion objective lens (Plan-Apochromatlan, Zeiss) at room temperature. The selected regions of interest were bleached at 488 nm and the fluorescence intensities in these regions were collected every 2.5 s. The pre-bleach intensity was used to normalize the fluorescence intensity. Image intensity was measured by mean regions of interest (ROI), calculated using ZEN software and analyzed by Prism (GraphPad).

The samples were subjected to SDS-PAGE and signals were detected with diluted polyclonal rat anti-RAB-7 (Chen et al., 2010) and monoclonal mouse anti-Trx-tag (Genscript, A00180-40) primary antibodies (diluted 1:1000, unless specifically indicated) and peroxidase AffiniPure goat anti-rat IgG (Jackson ImmunoResearch, 112-035-003), HRP-goat anti-mouse (Jackson ImmunoResearch, 115-035-003) secondary antibodies (diluted 1:200).

Embryos were collected by cutting 50–100 gravid animals. The embryos were permeabilized using the freeze-cracking method, and fixed with methanol and acetone (−20°C), then blocked with PBS (140 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, 2 mM KH₂PO₄, 1% BSA) for 1 h and incubated with polyclonal rat anti-PGL-3 (1:1000; Zhang et al., 2009), polyclonal rabbit anti-SEPA-1 (1:10,000; Zhang et al., 2009), polyclonal rat anti-SQST-1 (also known as T12G3.1; 1:1000; Tian et al., 2010) and polyclonal rat anti-LGG-1 (1:1000; Tian et al., 2010) primary antibodies overnight at 4°C. The embryos were then washed three times with PBST (PBS+0.2% tween 20) and incubated for 1 h with FITC-conjugated AffiniPure goat anti-rat IgG (Jackson ImmunoResearch, 112-095-003) or rhodamine-conjugated goat anti-rabbit IgG (Jackson ImmunoResearch, 111-025-003) and Alexa Fluor 594 AffiniPure goat anti-rat IgG (Jackson ImmunoResearch, 112-585-003) secondary antibodies (diluted 1:200) at 20°C. The embryos were then washed three times with PBST and stained with DAPI before examination.

2′-(3′)-bis-O-(N-methylanthraniloyl)-GDP (Mant-GDP) (Jena Bioscience, Germany) was loaded with 10 nmol of hexahistidine-GST-RAB in a buffer (20 mM HEPES pH 6.8, 1 mg/ml BSA, 20 mM EDTA pH 8.0, 40 mM Mant-GDP) at 30°C for 30 min. Next, 25 mmol MgCl₂ was added and Zeba spin columns (Fisher Scientific) were used to exchange the sample into reaction buffer (20 mM HEPES pH 6.8, 1 mg/ml BSA, 150 mM NaCl, 1 mM MgCl₂). This step removes the free Mant-GDP. Nucleotide exchange was measured using 1 nmol of the loaded RAB and the amount of GEF specified in the figure legends in a final volume of 100 μl reaction buffer by monitoring the quenching of fluorescence after release of Mant-GDP using a Tristar LB 941 plate reader (Berthold Technologies, UK) under control of MikroWin software (Qiagen, 9234701). Samples were excited at 350 nm and emission was monitored at 440 nm. GTP was added to a final concentration of 0.1 mM to start the exchange reaction at 30°C. Data were collected using a Safire multimode microplate spectrophotometer (Tecan) and plotted and analyzed using Microsoft Excel (Delprato et al., 2004).

Young adult animals (24 h post-L4 larval stage) expressing NUC-1::mCherry were mounted on agar pads in buffer with 5 mM levamisole. Fluorescent images in 15–20 z-series (0.5 μm/section) were taken with a spinning-disk confocal scanner unit (UltraView, PerkinElmer) using a 100× oil immersion objective (CFI Plan Apochromat Lambda, NA 1.45, Nikon) at room temperature. Serial optical sections were analyzed. The volume of NUC-1::mCherry-positive lysosomes was analyzed with Velocity software (PerkinElmer). At least 10 animals for each genotype were quantified and the data was analyzed using Prism (GraphPad).

Images were taken using Zeiss Axio Imager M2 and LSM880 microscopes. Imaging was carried out on animals at the same developmental stage, and on a similar body region of each animal. The same exposure time and magnification were used for comparison. Immunoblots were representative of three independent experiments. Images were analyzed using Image J. Statistical analysis was carried out using GraphPad Prism5 software and significant differences between datasets were determined by performing Student's t-test (two-tailed). The results are shown as mean±s.e.m. *P<0.05, **P<0.01, ***P<0.001.

We are grateful to Dr Isabel Hanson for editing work and members of Hong Zhang's lab for helpful discussion.