New insights regarding SNARE proteins in autophagosome-lysosome fusion

ABSTRACT

Macroautophagy/autophagy refers to the engulfment of cellular contents selected for lysosomal degradation. The final step in autophagy is the fusion of autophagosome with the lysosome, which is mediated by SNARE proteins. Of the SNAREs, autophagosome-localized Q-SNAREs, such as STX17 and SNAP29, and lysosome-localized R-SNAREs, such as VAMP8 or VAMP7, have been reported to be involved. Recent studies also reveal participation of the R-SNARE, YKT6, in autophagosome-lysosome fusion. These SNAREs, with the help of other regulatory factors, act coordinately to spatiotemporally control the fusion process. Besides regulating autophagosome-lysosome fusion, some SNAREs, such as STX17, also function in other autophagic processes, including autophagosome formation and mitophagy. A better understanding of the functions of SNAREs will shed light on the molecular mechanisms of autophagosome-lysosome fusion as well as on the mechanisms by which autophagy is globally regulated.

Abbreviations: ATG: autophagy related; DNM1L: dynamin 1 like; ER: endoplasmic reticulum; GABARAP: GABA type A receptor-associated protein; GABARAPL1: GABA type A receptor associated protein like 1; IRGM: immunity related GTPase M; LAMP2: lysosomal associated membrane protein 2; MAP1LC3B/LC3: microtubule associated protein 1 light chain 3 beta; MTOR: mechanistic target of rapamycin kinase; PIK3R4: phosphoinositide-3-kinase regulatory subunit 4; PLEKHM1: pleckstrin homology and RUN domain containing M1; PRKN: PRKN RBR E3 ubiquitin protein ligase; RAB2A: RAB2A, member RAS oncogene family; RAB33B: RAB33B, member RAS oncogene family; RAB7A: RAB7A, member RAS oncogene family; RB1CC1: RB1 inducible coiled-coil 1; RTN3: reticulon 3; RUBCNL: rubicon like autophagy enhancer; SNARE: soluble N-ethylmaleimide-sensitive factor attachment protein receptor; SNAP29: synaptosomal associated protein 29; STX17: syntaxin 17; ULK1: unc-51 like autophagy activating kinase 1; VAMP7: vesicle associated membrane protein 7; VAMP8: vesicle associated membrane protein 8; YKT6: YKT6 v-SNARE homolog

KEYWORDS:

• Autophagosome-lysosome fusion
• autophagy
• mitophagy
• SNARE
• STX17

Introduction

Macroautophagy/autophagy is a conserved catabolic process in which selected cellular components are delivered to and degraded within lysosomes to maintain cellular homeostasis [1–3]. Upon starvation, the autophagic negative regulator, MTOR (mechanistic target of rapamycin), is inactivated and dissociates from dephosphorylated ULK1, which phosphorylates and forms a complex with ATG13 and RB1CC1/FIP200 (RB1 inducible coiled-coil 1) to initiate nucleation of the phagophore [1–3]. Another complex comprised of ATG14 (autophagy related 14), BECN1/Beclin 1 (an ortholog of yeast Vps30/Atg6), PIK3R4/Vps15, and PIK3C3 (phosphatidylinositol 3-kinase catalytic subunit type 3) serves to phosphorylate phosphatidylinositol to phosphatidylinositol-3-phosphate (PtdIns3P) to expand the phagophore during the nucleation phase. Elongation of the phagophore during the expansion phase is facilitated by two ubiquitin-like conjugation systems, the ATG12 conjugation system and the Atg8-family protein conjugation system. Finally, the completely closed autophagosome fuses with a lysosome, which is mediated by SNARE proteins, to form the autolysosome [1–3].

In this review, we will primarily discuss the roles of SNARE proteins in the autophagosome-lysosome fusion process and provide perspectives for future studies.

Autophagosome-lysosome fusion

The fusion of autophagosomes and lysosomes is the final step of autophagy and requires the coordination of SNAREs, small GTPases, tethering factors, and other proteins [4]. In yeast, dissociation of Atg proteins from the autophagosomes, which is sensitive to the level of PtdIns3P, is a prerequisite for the initiation of autophagosome-lysosome fusion [5]. Depletion of Ymr1 (yeast myotubularin related), the sole yeast ortholog of the mammalian MTM/MTMR (myotubularin/myotubularin related) family of PtdIns3P phosphatases, causes accumulation of autophagosomes as a result of the Atg proteins’ dissociation defects [6]. However, a different role is determined in mammalian cells, as MTMR14 (myotubularin related protein 14) negatively regulates autophagosome formation [7]. Formation of PtdIns4P on the autophagosomes [8] and the proper ratio and amount of PtdIns(3,5)P₂ and PtdIns3P on the lysosomes [9] are reported to be required for autophagosome-lysosome fusion. The exact role of lipid composition in autophagy regulation merits further exploration.

Normally, interaction between the scattered autophagosomes in the cytoplasm and the perinuclear-enriched lysosomes depends on their bidirectional movement on microtubules, especially dynein-mediated autophagosome perinuclear movement [10]. Loss of dynein results in increased levels of LC3-II, which is associated with impaired autophagosome-lysosome fusion [11]. Depletion of KIF5B, the heavy chain of kinesin-1, triggers peripheral aggregation of lysosomes and perinuclear accumulation of autophagosomes, indicating an important role for kinesin-mediated transport in the correct positioning of these organelles [12]. Consistently, disrupting anterograde transport of lysosomes promotes their interactions with autophagosomes, and forced dispersion of lysosomes to the periphery blocks fusion [13]. Actin filaments have also been shown to be involved in regulation of autophagosome-lysosome fusion [14].

A major event during autophagosome-lysosome fusion is the SNARE complex-mediated fusion process. Autophagosome-localized STX17, SNAP29, and lysosome-localized VAMP8 [15] or VAMP7 [16] are believed to be the key factors involved in this process. A recent report shows that YKT6 also regulates autophagosome-lysosome fusion with the help of SNAP29, independent of the STX17 pathway [17]. However, the Drosophila Ykt6 homolog forms a temporary complex with Syx17/STX17 until the intervention of Vamp7, which has a higher binding capacity for Syx17 than Ykt6 [18]. The differences between these studies’ results may be due to differences in the model organisms used. Nevertheless, the significance of SNARE proteins’ involvement in autophagosome-lysosome fusion is well established.

To promote a systematic fusion process, other regulators, such as small GTPases and tethering factors, are also involved. RAB GTPases, such as RAB7A [19], RAB33B [20], and RAB2A [21], have all been reported to regulate autophagosome-lysosome fusion. PLEKHM1 (pleckstrin homology and RUN domain containing M1), a RAB7A effector, is able to recruit a tethering factor, the homotypic fusion and protein sorting (HOPS) complex, to lysosome membranes to control the fusion step [22]. The HOPS complex is reported to interact with STX17 and regulate assembly of the SNARE complex [23,24]. Another RAB7A effector, EPG5 (ectopic P-granules autophagy protein 5 homolog), binds to LC3 on autophagosomes and stabilizes the STX17-SNAP29-VAMP7/8 complex, facilitating autophagosome-lysosome fusion [25]. In addition, ATG14, which is well-known for its role in autophagosome initiation, also directly binds to the STX17-SNAP29 binary t-SNARE complex to promote its fusion with lysosomal VAMP8 [26]. Considering the complicated and delicate regulation of the SNARE-mediated fusion process, more regulators are likely involved and remain to be identified.

Assembly of SNARE complexes during autophagosome-lysosome fusion

In SNARE-mediated membrane fusion, four types of SNARE motifs, Q_a, Q_b, Q_c and R, provided by each of the SNAREs form a parallel four alpha-helix bundle [27]. In the case of autophagosome-lysosome fusion, the Q-SNAREs, STX17 and SNAP29, preform the Q_abc bundle on the autophagosome followed by its forming a complex with a lysosome-localized R-SNARE, such as VAMP8 [15]. The crystal structure shows that the Q196 of STX17, Q84 and Q230 of SNAP29, and R37 of VAMP8 constitute the ionic layer of the fusion-competent SNARE complex [26]. A recent study shows that SNARE priming mediated by NAPA/αSNAP, the adaptor for NSF (N-ethylmaleimide sensitive factor, vesicle fusing ATPase), also plays important roles in producing autophagosomes capable of fusing with the lysosomes, without affecting autophagosome formation or autophagosome localization of STX17 [28].

SNAP29, which lacks any transmembrane domain or membrane targeting motifs, is recruited to autophagosomes by interacting with STX17 after STX17 is transferred to autophagosomes [15]. STX17 has two tandem transmembrane domains containing glycine zipper-liker motifs, both of which are required for the autophagosome localization of STX17 [15]. However, the two transmembrane domains have low hydrophobicity, leading to a relatively large amount of cytosol-localized STX17. Combined with the observation that STX17 only presents on completely closed autophagosomes, autophagosomal STX17 likely originates from the cytosolic pool, but the detailed mechanism of its translocation remains undescribed [15]. Autophagosomal recruitment of STX17 seems to occur immediately before or after autophagosome closure when the elliptic elongating phagophore becomes spherical in shape [29]. It is reported that the autophagosome targeting of STX17 is affected in the absence of the lysosomal protein, LAMP2 [30], as well as autophagosomal IRGM (immunity related GTPase M) with the help of LC3 [31]. However, the role of Atg8-family members in recruiting STX17 to the autophagosome is controversial, as STX17 still presents on autophagosomal structures in cells deficient in all eight members of the Atg8 protein family during mitophagy [32]. Consistently, STX17-positive autophagosome-like structures are still formed in the absence of ATG conjugation systems, but at a much lower rate [29]. One plausible explanation is that LC3 alone is not sufficient for STX17 autophagosomal transfer but assists the recruitment and occludes premature binding of other fusion machineries to the SNARE domain of STX17 [31].

SNARE bundle formation is also regulated by post-transcriptional modifications, which either alter ionic charges critical for protein interaction or result in steric hindrance within the SNARE complex [33,34]. For example, a recent study shows that K219 and K223, two sites in the SNARE domain of STX17 that face the outside of the alpha-helical bundle, are acetylated by CREBBP (CREB binding protein) under normal conditions and are deacetylated by HDAC2 (histone deacetylase 2) upon autophagy stimulation [35]. SNAP29 and VAMP8 display a binding preference for deacetylated STX17 over acetylated STX17, indicating that the capacity for recruiting SNAP29, and thus subsequent SNARE complex formation, is the mechanism underlying the requirement of STX17 deacetylation during autophagosome-lysosome fusion [35]. The O-GlcNAc-modification on SNAP29 also hinders the formation of the STX17-SNAP29-VAMP8 SNARE complex, possibly due to steric hindrance caused by this modification [36].

Key players of the SNARE systems in autophagosome-lysosome fusion

STX17

STX17 was the first Q_a-SNARE found to be involved in autophagosome-lysosome fusion [15]. The conformation of STX17 is unique, as it contains two transmembrane domains in its C terminus; these form a hairpin structure that results in both the N terminus and C terminus localizing to the cytosol (Figure 1). STX17 is broadly expressed in multiple tissues and is primarily considered to be an endoplasmic reticulum (ER)-localized SNARE [37]. The C terminal tail of STX17 is required for its ER localization, while the N terminal domain, especially the conserved tyrosine residue Y156, regulates its exit from the ER into the ER-Golgi intermediate compartment (ERGIC) [38]. The Y156 residue can be phosphorylated by ABL1/c-Abl and dephosphorylated by TC48, an isoform of PTPN2 (protein tyrosine phosphatase non-receptor type 2), to modulate the interaction of STX17 with COPI vesicles [39]. Thus, STX17 is thought to cycle between the ER and ERGIC, which maintains the architecture of ERGIC and the Golgi apparatus [38]. Overexpression of STX17 redistributes COPB1 (COPI coat complex subunit beta 1) and blocks anterograde transport of VSVG (vesicular stomatitis virus G-protein) in the ERGIC [38]. Nuclear localization of STX17 is reportedly observed in several cancer cell lines, suggesting the existence of non-SNARE functions of STX17 [40]. STX17 has also been reported to present on the mitochondria, mitochondria-associated membranes (MAM), and in the cytosol [15,41]. Therefore, STX17 exhibits a diversity of subcellular localizations, indicating the complexity of its functions and mechanisms of regulation (Figure 2).

Figure 1. Schematic illustration of autophagosome-lysosome fusion-related SNARE proteins. The schematic drawing of the protein structures for STX17, SNAP29, VAMP7, VAMP8, and YKT6. TM, transmembrane domain; NPF, the Asp-Pro-Phe (NPF) motif; CCAIM, the C terminal prenylation consensus site; N, the N terminus

STX17 is thought to be transferred to autophagosomes upon nutrient starvation, and a large number of autophagosomes accumulate upon STX17 depletion [15]. Mechanistically, STX17 and the Q_bc-SNARE SNAP29 interact on the autophagosomes and mediate autophagosome-lysosome fusion by further coupling with lysosome-localized VAMP8 [15] (Figure 2). Essential for autophagosome localization, the two transmembrane domains of STX17 are also required for autophagic regulatory function, as quadruple mutations in G244, G248, G264 and G268, which disrupt the organization of the transmembrane domains, fail to rescue defects in LC3-II accumulation caused by STX17 knockdown [15].

Figure 2. The diverse functions of STX17 in autophagy-related processes. Under normal conditions, STX17 recruits DNM1L at the mitochondria-ER-microtubule contact site to drive mitochondrial fission [49]. Mitochondria-localized STX17 also controls mitochondria quality by MDV [51], by PRKN-dependent mitophagy through regulation of the PGAM5-FUNDC1 interaction [52], or by PRKN-independent mitophagy in the absence of FIS1 [53]. Upon starvation, dephosphorylated MAP1B-LC1 releases STX17 from the microtubule [55]. STX17 binds to ATG14 at MAM and initiates formation of the phagophore [41]. Golgi apparatus-localized STX17, which is phosphorylated by TBK1, transfers to the phagophore assembly site to promote assembly of the RB1CC1-ATG13-ULK1 complex [47]. On the completely closed autophagosome, STX17 is recruited from the cytosol by LC3 and IRGM [31], followed by SNARE-mediated fusion with the help of various regulators, such as the HOPS complex (including VPS33A) [23,24], ATG14 [26], DIPK2A [67] and ULK1 [48]. Upon ULK1 phosphorylation by PRKCA, ULK1 is degraded [48]

During STX17-mediated autophagosome-lysosome fusion, the fusion process is precisely regulated. The HOPS complex has been reported to interact with STX17 and translocate to STX17-positive autophagosomes [23,24]. The recruitment of HOPS to autophagosomes is precisely regulated by RUBCNL/PACER (rubicon like autophagy enhancer), whose autophagosome targeting is also regulated by STX17 [42]. Knockdown of the HOPS components, VPS33A, VPS16, or VPS39, blocks autophagic flux and leads to autophagosome accumulation [24]. The N terminal domain of STX17 is required for its interaction with VPS33A [43]. Overexpression of an STX17 mutant lacking the N terminal domain causes a dominant-negative effect in HeLa cells, with accumulation of a large number of undigested autophagosomes occurring as a result of a defect in HOPS binding [43]. HOPS may act as a scaffold for organizing the fusion machinery [24]. PLEKHM1 interacts with HOPS through the RUN domain and is colocalized at vesicle contact sites with RAB7A and HOPS [22]. PLEKHM1 is able to contact the autophagosomes as it binds to autophagosome-localized LC3 through the LC3-interacting region (LIR). Similar to PLEKHM1, BIRC6/BRUCE, a lysosome-localized inhibitor of apoptosis protein (IAP), is recently shown to promote autophagosome-lysosome fusion by bridging STX17 and LAMP2-labeled lysosomes [44]. Unlike PLEKHM1, which nonspecifically interacts with all six Atg8-family members, BIRC6 interacts with autophagosomes through a binding preference for GABARAP and GABARAPL1. Notably, VPS33A belongs to the Sec1/Munc18 (SM) protein family, the well-known SNARE assembly regulators, and thus may promote the fusion process by stabilizing the SNARE bundle [45,46].

Our understanding of STX17 in autophagy regulation was significantly expanded by the identification of MAM-localized STX17 [41], pointing to a role for STX17 in autophagosome formation. MAM is a newly identified hot spot for origination of autophagosomes. STX17 binds to ATG14 and recruits it to the MAM after starvation. In STX17 knockdown cells, recruitment of ATG14 to MAM is hindered, and ATG5-positive dots are accumulated [41]. Interestingly, ATG14 is also involved in autophagosome-lysosome fusion through binding to the SNARE domain of STX17, stabilizing the STX17-SNAP29 binary t-SNARE complex on autophagosomes [26]. The autophagosome initiation function of STX17 is further supported by the recent observation that STX17 is phosphorylated by TBK1 at S202 [47]. Upon starvation, phosphorylated STX17 (S202) translocates from the Golgi apparatus to the phagophore assembly site to orchestrate the assembly of the ATG13-RB1CC1-ULK1 complex [47]. In addition, the ULK1-STX17 interaction is also necessary for the later step of autophagosome-lysosome fusion by enhancing STX17-SNAP29 coupling [48]. PRKCA/PKCα phosphorylates ULK1 at S423, which decreases ULK1-STX17 interactions. Phosphorylated ULK1 has an attenuated ability to regulate autophagosome-lysosome fusion and is degraded via chaperone-mediated autophagy.

STX17 is able to present on the mitochondria; this presentation is dependent on both its transmembrane domains and C terminal domain [49]. Replacement of the K254 residue of STX17 is sufficient to disrupt its mitochondrial localization [49]. Consistently, the transmembrane domains and the C terminal domain are also responsible for STX17’s regulation of mitochondrial dynamics, as a mutant with their depletion fails to rescue the mitochondria elongation caused by STX17 knockdown, which also disrupts DNM1L/DRP1 (dynamin 1 like) puncta formation [49]. However, there is disagreement regarding whether STX17 is involved in mitophagy regulation. In HeLa cells, STX17 knockdown blocks hypoxia- or deferiprone (DFP)-induced mitophagy as a result of autophagosome-lysosome fusion defects [50]; however, STX17 depletion does not affect the levels of mitochondrial proteins in oligomycin/antimycin A (OA)-induced mitophagy under mCherry-PRKN (parkin RBR E3 ubiquitin protein ligase)-expressing conditions [32]. Another study that uses GFP-PRKN-expressing U2OS cells induced by carbonyl cyanide m-chlorophenylhydrazone (CCCP) shows that STX17 is dispensable for depolarization-induced mitochondrial degradation as well as PRKN translocation [51]. Only recently is the role of STX17 in PRKN-mediated mitophagy verified; the detachment of PGAM5 from STX17 facilitates the interaction between PGAM5 and FUNDC1 [52]. In this case, mitophagy is inhibited upon depletion of STX17 or PGAM5. A recent study further shows that STX17 is also involved in PRKN-independent mitophagy. STX17 interacts with the TPR2 domain of FIS1 (fission, mitochondrial 1) through its N terminal domain and transmembrane domain, and STX17 overexpression induces mitophagy upon FIS1 depletion in a PINK1/PRKN-independent manner [53]. STX17 recruits ATG14 and, subsequently, the core autophagy proteins, to the mitochondria and MAM to form mitophagosomes, which finally fuse with lysosomes through RAB7A and TBC1D15 (TBC1 domain family member 15). STX17 is found to present on mitochondrial-derived vesicles (MDVs), which are shed from the mitochondria to protect cells against oxidative stress [51,54]. STX17, together with SNAP29 and VAMP7, mediate fusion between MDVs and lysosomes. In summary, STX17 clearly plays important roles in mitochondrial quality control (Figure 2).

Considering the diverse distribution of STX17 in the cell, questions remain regarding how the protein senses upstream signals and then transfers to autophagosomes. Autophagosome-localized STX17 primed for autophagosome-lysosome fusion is believed to come from the cytosolic pool following autophagosome closure [15]. Surprisingly, STX17 is able to bind to Atg8-family members, including LC3B and GABARAP, through two LIR motifs (LIR1_172–175 and LIR2_189–192), mutations to which reduce the autophagosomal localization of STX17 [31]. Another LC3-interacting protein, IRGM, binds to the transmembrane domains of STX17, the regions required for its insertion into autophagosome membranes, to facilitate the recruitment of STX17 to autophagosomes by forming a STX17-LC3B-IRGM holocomplex [31].

How STX17 serves to precisely coordinate a diversity of subcellular functions remains undefined. Likely, under non-starvation conditions, STX17 associates with the microtubules, the ER, and the mitochondria to regulate mitochondrial division by recruiting DNM1L [49]. MAP1B-LC1 (microtubule associated protein 1B-light chain 1), a proteolytic product of the polyprotein MAP1B, is able to link STX17 to microtubules by binding both [55]. Starvation causes the dephosphorylation of MAP1B-LC1_T217, which decreases its interaction with STX17. Subsequently, released STX17 further regulates autophagy initiation by binding to ATG14 [55]. Thus, STX17 performs different functions by binding to diverse partners (Figure 2).

SNAP29

SNAP29, the Q_ab-SNARE, belongs to the SNAP25-protein family, members of which possess two SNARE domains separated by a linker region (Figure 1) [37,56]. SNAP29 lacks a transmembrane domain and thus transiently associates with the membrane of organelles, including the Golgi apparatus, endosomes, and lysosomes, by interacting with other endocytic factors, such as syntaxins [37,56]. The wide distribution of SNAP29 allows it to function in a variety of membrane trafficking processes, from endolysosomal trafficking to secretion [57]. SNAP29 was also found to localize to the outer kinetochore and to function in organizing kinetochore formation [58].

In Drosophila fat bodies, Snap29 depletion results in complete inhibition of starvation-induced autolysosome formation as well as in accumulation of Atg8-positive autophagosomes, similar to Syx17 or Vamp7 depletion, indicating a role for Snap29 in autophagosome-lysosome fusion [16]. In starved mammalian cells, SNAP29 colocalizes with STX17-positive LC3 structures, and SNAP29 knockdown leads to LC3-II accumulation and inhibition of autophagic flux [15].

As discussed above, the autophagic activity of SNAP29 is regulated by O-GlcNAc modification [36]. OGT (O-linked N-acetylglucosamine (GlcNAc) transferase) mediates O-GlcNAcylation of SNAP29 at multiple sites (S2, S61, T130, and S153); these modifications are removed by OGA (O-GlcNAcase). Upon starvation, the levels of O-GlcNAcylation on SNAP29 are reduced, and unmodified SNAP29 forms a more stable complex with STX17 and VAMP8 to mediate autophagosome-lysosome fusion. Expression of the O-GlcNAc-defective SNAP29 mutant, SNAP29_{S2A/S61G/T130A/S153G}, enhances autophagic activity and facilitates autophagic degradation of protein aggregates even under nutrient-rich conditions [36]. Further work in type I diabetic rats confirms the role of O-GlcNAc modification of SNAP29 in vivo, highlighting its functions and molecular mechanisms in disease [59]. A recent report demonstrates that DAPK3 (death associated protein kinase 3), a member of the DAPK family of proteins involved in gestational diabetes mellitus (GDM), enhances the STX17-SNAP29 interaction by binding to SNAP29 [60]. Interestingly, silencing DAPK3 modulates autophagy only in high-glucose conditions [60]. It remains undescribed if there is a link between the O-GlcNAc modification of SNAP29 and DAPK3 function. Thus, fine tuning of SNAP29 modifications may represent a potential target for clinical application.

VAMP8 and VAMP7

Lysosome-localized R-SNAREs function as couplers for autophagosomal STX17-SNAP29; these include VAMP8 in mammals and Vamp7 in Drosophila [15,16]. The trafficking of VAMP8 (or VAMP7) to lysosomes is regulated by RAB21 and MTMR13, depletion of which reduces VAMP8 (or VAMP7) lysosomal distribution and attenuates autophagosome-lysosome fusion [61].

A recent study using high-resolution fluorescence lifetime imaging (FLIM)-FRET reveals that STX17 heterodimerizes with VAMP7 in situ, highlighting a functional role for VAMP7 in autophagosome clearance in mammalian cells [45]. VAMP7 appears to be more reliable than VAMP8 in coupling with the STX17-SNAP29 complex due to its greater proximity to STX17 [45]. If not the dominant R-SNARE in autophagosome-lysosome fusion, VAMP7 is at least an alternative SNARE in this process.

VAMP7 is a member of the longin family of SNARE proteins. Longin SNAREs, including VAMP7, YKT6, and SEC22B, all contain a coiled-coil longin domain in their N terminus (Figure 1); this domain is an important regulatory domain for trans-SNARE complex fusion and thus membrane fusion [62]. The longin domain of VAMP7 is able to interact with its own SNARE domain to autoinhibit the protein itself [63]. However, the longin domain also regulates the distribution and function of VAMP7 by binding to various partners [64–66]. Interestingly, an isoform of VAMP7, VAMP7B, lacks the longin domain of the predominant form of VAMP7 (VAMP7A), and is able to inhibit autophagosome-lysosome fusion by ameliorating the interaction of VAMP7A and STX17 [67]. VAMP7B binds to the SNARE domain of STX17 competitively with VAMP7A through its C terminal domain, while DIPK2A (divergent protein kinase domain 2A), a lysosome-localized protein, disrupts their interaction. Thus, DIPK2A promotes autophagosome-lysosome fusion by enhancing the STX17-VAMP7A interaction [67].

Apart from its role in mediating autophagosome-lysosome fusion, VAMP7 is also involved in early autophagic process regulation. Knockdown of VAMP7 results in decreased LC3-II levels both in the presence and absence of the autophagy inhibitor bafilomycin A₁ (Baf A1), indicating an inhibition of autophagosome formation [68]. Similarly, VAMP7 knockout attenuates autophagosome formation, leading to accumulation of dysfunctional mitochondria in pancreatic β-cells [69]. Moreover, VAMP7 colocalizes and interacts with ATG9A on recycling endosomes [70]. AGFG1/Hrb, which binds to the longin domain of VAMP7, recruits VAMP7 and ATG9A to recycling endosomes from the plasma membrane, followed by VAMP7-, STX16 (syntaxin 16)-, and SNAP47-mediated fusion of ATG9A-resident vesicles, which finally contributes to autophagosome formation [70]. VAMP7, ATG16L1, and ATG5 are colocalized on LC3-negative autophagic precursors as well as on LC3-positive phagophores [68]. VAMP7 promotes the homotypic fusion of ATG16L1 precursors and regulates the size of ATG16L1 vesicles, which are important for autophagic membrane expansion. The ATG16L1-positive pool of VAMP7 also originates from the plasma membrane in a AGFG1-dependent manner [68], perhaps partially from the ATG9A pathway as well [70]. VAMP7 has also been reported to recruit LC3 to Yersinia pseudotuberculosis-containing vacuoles (YCVs) in infected macrophages, paving the road for this pathogen during its autophagic journey [71]. All of these studies expand our understanding of VAMP7 in autophagy regulation.

YKT6

YKT6, an R-SNARE protein, has recently received increased attention in the field of autophagosome-lysosome fusion [17,18]. YKT6 contains a longin domain in its N terminus, similar to VAMP7, but does not contain a transmembrane domain (Figure 1) [72]. Thus, membrane targeting of YKT6 depends on its C terminal prenylation consensus site (CCAIM), which is reversibly modulated by palmitoylation [72]. YKT6 exhibits variable functions, and yeast YKT6 is reported to be required for autophagosome formation [73].

In Drosophila, large amounts of autophagosomes accumulate in Ykt6-depleted cells as a result of failed autophagosome-lysosome fusion [18]. Lysosome-localized Ykt6 is able to form a SNARE complex with Syx17-Snap29, until participation of Vamp7, which outcompetes Ykt6 from the temporary SNARE complex [18]. Mutation in the SNARE domain does not affect the autophagic ability of Ykt6, indicating that Ykt6 acts as a regulator rather than as the fusion machinery in this process [18]. The role of YKT6 in mammalian cells differs from that in Drosophila, where it mediates the autophagosome-lysosome fusion process by coupling with SNAP29 and STX7 in a STX17-independent manner [17]. In this case, YKT6 is recruited to the autophagosomes, but not to LC3-negative lysosomes, through its longin domain. Thus, YKT6-SNAP29-STX7 likely functions as a second functional SNARE complex in addition to the STX17-SNAP29-VAMP8 complex in mammals [17]. Studies in yeast further demonstrate the functional conservation of YKT6, which is localized on the autophagosomes, in autophagosome-vacuole fusion using in vitro autophagosome-vacuole fusion assays [74,75]. Though differences remain between species, the involvement of YKT6 in regulating the autophagosome-lysosome fusion process is unequivocal.

Perspectives

Along with the discovery of an increasing number of regulators in autophagosome-lysosome fusion, the mechanism of the process is becoming clearer. However, questions still remain regarding how different SNARE couplers and regulatory factors cooperate to ensure the process proceeds efficiently and precisely and how the fusion activity of SNARE proteins is spatiotemporally regulated.

Owing to its diverse localization in the cell, STX17 is a “Jack of all trades” in numerous cellular functions. How STX17 balances its variable roles in different cellular pathways remains unclear. One possibility is that STX17 exerts these functions through different domains. For example, the roles of STX17 in autophagy and mitochondria fission regulation diverge from each other, as elongated mitochondria can be rescued by a STX17 mutant defective in autophagosome-lysosome fusion [76]. However, how these distinct functions are precisely regulated remains unclear.

It is well-known that organelle interactions play important roles in cellular functions [77]. The function of STX17 is likely regulated by signaling from different organelles. For example, SIGMAR1 (sigma non-opioid intracellular receptor 1), an ER-resident chaperone protein, interacts with STX17, along with ATG14 and VAMP8, and promotes autophagosome-lysosome fusion during mitophagy [78]. The STX17 and ATG14 interaction is disrupted by BCAP31 (B cell receptor associated protein 31), an ER membrane protein with an ER stress-modulating function [79]. Dysfunctional mitochondrial respiration can lead to transfer of STX17 from the ER to the mitochondria [53]. STX17 has also been reported to interact with ACSL3 (acyl-CoA synthetase long chain family member 3), a key enzyme in lipid droplet biogenesis, at the ER-mitochondria interface through its SNARE domain [80]. STX17 knockdown disrupts ACSL3 localization from the ER to the lipid droplet, attenuating lipid droplet maturation [80]. In summary, a better understanding of its functions will provide more insights into how STX17 is coordinated in various cellular processes.

A recent study reveals an ER-to-lysosome-associated degradation (ERLAD) pathway, in which some ER-localized proteasome-resistant cargos, such as the Z variant of SERPINA1/alpha1-antitrypsin Z/ATZ, are segregated into ER-derived single membrane vesicles with the help of a reticulophagy receptor, RETREG1/FAM134B, followed by direct fusing with lysosomes through STX17-VAMP8-mediated fusion [81]. ERLAD differs from reticulophagy in its lack of dependence on autophagosome biogenesis machinery. Additional reticulophagy receptors, such as RTN3 (reticulon 3) [82], ATL3 [83], and TEX264 [84], have been identified recently. It will be interesting to investigate whether STX17, as well as other SNARE proteins, coordinate with these factors and also regulate ER quality control.

Our current understanding of the mechanism of autophagosome-lysosome fusion largely benefits from traditional biochemistry, molecular, and cell biology techniques, such as analyzing autophagic protein turnover, localization, and pH alteration [85]. In vitro-purified proteoliposome fusion assay, which is based on Förster/fluorescence resonance energy transfer/FRET, also represents a reliable means for analyzing SNARE-mediated vesicle fusion during autophagosome-lysosme fusion [26,86]. However, the need to reveal the spatiotemporal dynamics and regulatory mechanisms of autophagosome-lysosome fusion necessitates the development of new tools to describe the fusion process more precisely and quantitively. With the improvements in the resolution of optical microscopes, biocellular processes, such as organelle interactions, can be captured and quantified accurately by super-resolution microscopy methods. For example, a recent study analyzes mitophagy events by quantifying mitochondria-lysosome contacts with structured illumination microscopy/SIM [87]. Combined with single synapse imaging, time-lapse imaging or fluorescence recovery after photobleaching/FRAP, the analyses of SNARE protein assembly dynamics during synaptic vesicle exocytosis have been well-established by stochastic optical reconstruction microscopy/STORM [88], stimulated emission depletion/STED fluorescence microscopy [89], and total internal reflection fluorescence/TIRF microscopy at single-molecule levels [90]. Application of these techniques to autophagosome-lysosome fusion may explain the mechanisms underlying the accurate regulation of SNARE assembly in this process.

Disclosure statement

The authors declare no competing interests.

Additional information

Funding

This work was supported by the National Natural Science Foundation of China (NSFC) [91954124 and 31871353], the National Key Research and Development Program of China, Stem Cell and Translational Research [2016YFA0100501], and the China Postdoctoral Science Foundation [2020M672116].