The hookup model of the HOPS complex in autophagosome-lysosome fusion

ABSTRACT

Macroautophagy/autophagy is a highly conserved process that involves the degradation of proteins, damaged organelles, and other cytoplasmic macromolecules. Autophagosome-lysosome fusion is critical for successful substrate degradation and is mediated by SNARE proteins. The fusion process requires additional vesicle docking and tethering-regulating factors. Our recent work has uncovered a functional model of autophagosome-lysosome fusion. We demonstrated that the six-subunit homotypic fusion and vacuole protein sorting (HOPS) complex can be assembled by two subcomplexes, the VPS39-VPS11 subcomplex (HOPS-2) and the VPS41-VPS16-VPS18-VPS33A subcomplex (HOPS-4). VPS39 binds with RAB2 on the autophagosome and VPS41 binds with RAB39A on the lysosome, which then promotes membrane tethering and autophagic SNARE-mediated membrane fusion. Moreover, we have revealed that ALS- and FTD-related C9orf72 is a guanine exchange factor (GEF) for RAB39A. In this punctum, we discuss how the C9orf72-RAB39A-HOPS axis function regulates autophagosome-lysosome fusion.

Genetic manipulation research indicates that the HOPS complex is required for autophagosome-lysosome fusion in mammals, but the mechanisms by which HOPS docks on heterotypic vesicles and mediates tethering between autophagosome and lysosome membranes remained unclear. Additionally, the in vitro biochemical reconstitution of a functional autophagic membrane fusion system containing the HOPS complex has not yet been accomplished.

In yeast vacuole fusion, the HOPS complex interacts with Ypt7 and assembles SNAREs to tether vacuole membranes and mediate the fusion process. The machinery mediating mammalian autophagosome-lysosome membrane fusion is not established yet. First, we found lysosome-resident RAB39A is the most enriched RAB GTPase associated with the autophagic SNARE STX17, and directly interacts with HOPS [1]. Additionally, the overexpression of RAB39A enhances the interaction between STX17 and VPS33A, and the association of SNAREs, indicating that RAB39A has the capacity to promote the assembly of fusion machinery. RAB7, the mammalian homolog of Ypt7 may recruit HOPS onto lysosomes in autophagic fusion, but we found that overexpression of RAB39A displaces RAB7 from HOPS subunits and strengthens the binding between RAB2 and VPS39. These results imply that RAB39A, rather than RAB7, contributes to the formation of the RAB-HOPS-SNARE fusion machinery and enhances its stability.

Unlike the tightly associated yeast HOPS complex, the mammalian HOPS complex falls apart into two subcomplexes, HOPS-2 (VPS39-VPS11) and HOPS-4 (VPS41-VPS16-VPS18-VPS33a), which can be assembled in vitro. To biochemically reconstitute in vitro autophagic membrane fusion including SNAREs, HOPS, and RABs, we first attempted to reconstitute the HOPS complex with RAB2 and RAB7. However, this attempt was unsuccessful. But when we reconstituted RAB39A as a replacement for RAB7 on VAMP8-decorated proteoliposomes, with pre-incubation with the HOPS-4 subcomplex, and RAB2 on STX17-SNAP29-decorated proteoliposomes, with pre-incubation with the HOPS-2 subcomplex, the membrane fusion is significantly promoted. Moreover, cryo-EM images showed that HOPS cooperates with RAB39A and RAB2 leading to the clustering of liposomes. Through all these experiments, we successfully reconstituted a functional RAB39A-HOPS-RAB2 complex, and found that HOPS subcomplexes anchored on opposing membranes through RAB interactions can hook up with each other, and assemble into a six-subunit complex, which in turn tether proteoliposomes and facilitates membrane fusion. Therefore, we propose a “hook-up” model to describe this induced assembly of the HOPS complex for autophagic membrane tethering and fusion.

Next, we wondered whether the function of RAB39A in autophagy is related to its GTPase function. By expressing different mutants of RAB39A, we found its membrane distribution depends on activation of GTP loading and C-terminal prenylation. Besides, the interaction between the inactive mutant of RAB39A with HOPS subunits is largely decreased compared with wild-type RAB39A, and the autophagic vesicle-localized HOPS declines remarkedly when inactive RAB39A is overexpressed. These results suggested that HOPS recruitment onto autophagic vesicles may be dependent on RAB39A activation. Additionally, we found that the GDP-loaded RAB39A is not able to rescue the autophagy flux defect in RAB39A knockout cells, suggesting the autophagy-promoting function of RAB39A is dependent on its GTP loading. The in vitro lipid mixing assay results further support this hypothesis. Moreover, we investigated the regulatory mechanism of RAB39A GTP loading. According to the literature, the DENN-domain enriched protein C9orf72 can function as a GEF toward RAB39B with SMCR8-WDR41, and RAB39B shares a high degree of amino acid sequence identity with RAB39A. Thus, we speculated that C9orf72 may be a GEF of RAB39A as well. We first confirmed that C9orf72 can bind RAB39A both in vivo and in vitro, and the GDP-bound form of RAB39A enhances this interaction. Using an in vitro Mant-GDP releasing assay, we found that C9orf72 strongly facilitates the release of Mant-GDP from RAB39A in a concentration-dependent manner alone without other catalytic subunits. To further dissect the role of C9orf72 in autophagy, we generated a C9orf72 knockout cell line, which exhibits inhibited autophagy flux and autophagosome-lysosome fusion defects, similar to RAB39A knockout cells. More importantly, the autophagy defects in C9orf72 knockout cells can be rescued by the RAB39A GTP-loaded form, but not the wild-type form, in line with our hypothesis that C9orf72 regulates autophagosome-lysosome fusion through activating RAB39A as a GEF.

Overall, our work has revealed that through proper pairing with RAB39A and RAB2 on membranes, HOPS-4 and HOSP-2 subcomplexes hook-up to form a functional HOPS-6 complex for membrane tethering and fusion between autophagosomes and lysosomes. The activation of RAB39A by C9orf72 is essential for the anchoring of the HOPS complex (Figure 1). It remains unclear whether there are other regulatory mechanisms for the induced assembly of the HOPS complex. More in vivo and in vitro analysis should be conducted, including the investigation of the cellular distribution of HOPS subcomplexes with or without autophagy stimulation, and analyzing the complex structure of the functional reconstituted HOPS-RABs complex. Additionally, the structure analysis of the HOPS-RABs protein complex would aid in explaining the mechanism of how the HOPS complex binds to distinct RAB GTPases in autophagic vesicle fusion. Given the data in several studies, the endogenous C9orf72 protein level is impaired in a portion of C9orf72-mutated ALS-FTD patients, it is reasonable to hypothesize that the autophagosome-lysosome fusion impairment could play an important role in onset or progression of ALS-FTD diseases, and elevation of autophagosome-lysosome fusion may help the clearance of toxic protein aggregates. Therefore, the C9orf72-RAB39A-HOPS axis could have therapeutic value in neurodegenerative disease.

Figure 1.

A schematic model for HOPS-mediated autophagosome-lysosome fusion. Lysosome-resident RAB39A is activated by C9orf72, and subsequently recruits the HOPS-4 subcomplex, whereas autophagosome-localized GTP-RAB2 recruits the HOPS-2 subcomplex. HOPS-2 and HOPS-4 “hook-up” to form HOPS-6 to tether autophagosomes and lysosomes, which leads to promote autophagosome-lysosome fusion driven by the STX17-SNAP29-VAMP8 SNARE complex.

Funding Statement

This work was supported by grants 2019YFA0508602 (MOST (Ministry of Science and Technology of the People’s Republic of China)), M-0140 (NSFCSino-German Mobility Programme), and 92254307, 91754205 (NSFC (National Natural Science Foundation of China)) to Q.Z.

Disclosure statement

No potential conflict of interest was reported by the author(s).