Autophagy is a highly regulated membrane remodeling process that allows the lysosome-mediated degradation of cytoplasmic entities by sequestrating them in double-membrane autophagosomes. Autophagy is hence highly intertwined with the endocytic trafficking pathway, sharing similar molecular machinery. Atg14L, also known as Beclin 1-associated autophagy-related key regulator (Barkor), directly interacts with Beclin 1 through its coiled-coil domain and enhances phosphatidylinositol 3-phosphate kinase class III (PI3KC3) activity to induce autophagosome membrane nucleation, highlighting its essential role in the early stage of mammalian autophagy. Here, we report a novel function of Atg14L in the endocytic trafficking pathway wherein Atg14L binds to and colocalizes with the fusogenic SNARE effector protein Snapin to facilitate endosome maturation. Atg14L specifically binds to Snapin and this interaction effectively facilitates endosomal maturation without affecting autophagic cargo degradation. Consequently, atg14l knockdown significantly delayed the late stage of endocytic trafficking, as evidenced by the retarded kinetics of internalized surface receptor degradation. This phenotype was effectively complemented by wild-type Atg14L or Beclin 1-binding mutant, but not by its Snapin-binding mutant. Taken together, our study demonstrates that Atg14L functions as a multivalent trafficking effector that regulates endosome maturation as well as autophagosome formation, reflecting the complexity of the crosstalk between autophagic and endocytic vesicle trafficking in higher eukaryotes.
Macroautophagy (hereafter called autophagy) is a membrane trafficking process dedicated to the maintenance of homeostasis by allowing the self-cannibalization of intracellular constituents through the lysosomal degradation pathway (Klionsky, 2005). Unique de novo double-membrane structures termed isolation membranes (IMs) enwrap cytoplasmic contents to generate autophagosomes, which subsequently undergo stepwise maturation into hybrid-like organelles with degradative capabilities termed autolysosomes. Specifically, autophagosomes utilize the microtubule network to migrate and sequentially fuse with components of the endocytic pathway, such as early and late endosomes and multi-vesicular bodies (MVBs), producing amphisomes. Autophagosomes and/or amphisomes eventually fuse with lysosomes to generate autolysosomes, in which the sequestered cargo is degraded into building blocks for macromolecules and recycled to the cytoplasm (Chen and Klionsky, 2011; Simonsen and Tooze, 2009).
Despite profound differences in structure, autophagosomes topologically and mechanistically resemble endosomes during maturation as they share a number of vesicle-trafficking components (Orsi et al., 2010). For instance, both autophagic and endocytic trafficking processes require the class C vacuolar protein sorting (Vps-C)/homotypic fusion and protein sorting (HOPS) complex, a central regulator governing multiple trafficking events at the vacuolar/lysosomal compartments (Nickerson et al., 2009). In yeast, the Vps-C/HOPS complex functions as a tethering apparatus by promoting nucleotide exchange of Ypt7/Rab7 and drives membrane fusion by catalyzing the assembly of soluble N-ethylmaleimide sensitive factor (NSF) attachment protein receptor (trans-SNARE) complexes (Sato et al., 2000; Stroupe et al., 2006; Wurmser et al., 2000). Our previous studies have shown that UV radiation resistance-associated gene (UVRAG), initially characterized as a component of the class III posphatidylinositol-3-kinase (PI3KC3) complex during autophagosome formation, interacts with the Vps-C/HOPS complex and promotes autophagosome/endosome maturation, thus serving as a convergence point of the autophagic and endocytic pathways (Liang et al., 2006; Liang et al., 2008).
Consisting of three core components – hVps34, hVps15 and Beclin 1 – the activity and specificity of the PI3KC3 complex is tightly modulated by numerous regulators during autophagosome biogenesis and endocytic trafficking (Lindmo and Stenmark, 2006). It has been increasingly recognized that, as in yeast, the mammalian PI3KC3 complex containing UVRAG predominantly functions in autophagosome/endosome maturation, whereas the yeast Atg14-like (Atg14L)-harboring complex induces autophagosome membrane nucleation. Atg14L, also known as Beclin 1-associated autophagy-related key regulator (Barkor), was identified to directly interact with Beclin 1 through its coiled-coil domain and to be excluded from the PI3KC3 complex carrying UVRAG (Itakura et al., 2008; Matsunaga et al., 2009; Sun et al., 2008; Zhong et al., 2009). The N-terminus of Atg14L contains the conserved cysteine repeats necessary for endoplasmic reticulum (ER) localization and thereby recruits the PI3KC3 complex to the initiating IMs originated from the ER, a potential origin of autophagosomal membranes (Matsunaga et al., 2010). Moreover, the C-terminal amphipathic alpha helix in the Barkor/Atg14L autophagosome targeting sequence (BATS) domain was suggested to be essential for autophagosome targeting by stabilizing the highly curved early autophagic membrane in a phosphatidylinositol 3-phosphate (PI3P)-dependent manner (Fan et al., 2011; Lu et al., 2011). These support the notion that Atg14L is crucial for the early phase of autophagosome formation.
Originally identified as a neuronal SNARE protein, Snapin was suggested to modulate priming of synaptic vesicles for their synchronized fusion during neurotransmission (Pan et al., 2009). Further studies have extended the fusogenic function of Snapin in a wide array of membrane fusion events in non-neuronal cells such as endosome-mediated cytokinesis, phagosome maturation, and insulin exocytosis (Ilardi et al., 1999; Buxton et al., 2003; Gromley et al., 2005; Song et al., 2011; Tiwari et al., 2009). Snapin executes the fusogenic function in part by associating with proteins in vesicle SNARE (v-SNARE) complexes on donor membranes and drives the fusion with trans SNARE (t-SNARE) complexes on target membranes such as lysosomal membranes (Buxton et al., 2003; Lu et al., 2009; Pan et al., 2009). Moreover, Snapin has been recently implicated to coordinate the retrograde transport of late endosomes and lysosomal maturation by interacting with the dynein motor complex, thereby contributing to efficient autophagic-lysosomal function (Cai et al., 2010). These suggest that Snapin is a candidate molecular target for autophagy and lysosome regulation.
As part of our endeavor to delineate the biological functions of mammalian Atg14L, we screened novel interacting proteins of Atg14L using a yeast two-hybrid system. Our results showed that Atg14L directly binds to Snapin and that this interaction accelerates endosomal maturation without affecting autophagic cargo degradation. Our findings reveal that Atg14L, previously considered to be solely a Beclin 1-binding autophagy protein, plays a novel role in the late stage of endocytic trafficking in conjunction with Snapin.
Atg14L binds to and colocalizes with Snapin
To further explore the biological roles of Atg14L during autophagy and/or other cellular processes, we performed a yeast two-hybrid screen using full-length Atg14L as bait. Along with Beclin 1, which was previously shown to directly bind to Atg14L (Itakura et al., 2008; Matsunaga et al., 2009; Sun et al., 2008; Zhong et al., 2009), Snapin was identified as a putative interaction partner (supplementary material Fig. S1A). Co-immunoprecipitation verified the efficient and specific binding between endogenous Atg14L and Snapin as well as between epitope-tagged Atg14L and Snapin (Fig. 1A; supplementary material Fig. S1B). In search of an Atg14L mutant that specifically loses Snapin-binding ability while preserving Beclin 1-interacting activity, extensive binding mapping studies were conducted. Firstly, a series of Atg14L truncation mutants fused with mammalian glutathione S-transferase (GST) was constructed to define the region responsible for Snapin interaction (Fig. 1B). GST pull-down showed that Atg14L possesses three independent sites sufficient for Snapin binding: residues 71–106 (A), 270–320 (B) and 419–492 (C) (Fig. 1C). Accordingly, these three regions were individually deleted or in combination to identify the specific Snapin-binding site in the context of full-length Atg14L (Fig. 1D). Region 419–492 was modified to Δ419–480 since the last 10 amino acids of the Atg14L BATS domain are required for autophagosome targeting (Fan et al., 2011). Co-immunoprecipitation showed that any Atg14L mutant devoid of residues 419–480 lost Snapin-interacting activity (Fig. 1E). We further confirmed that the Atg14L Δ419–480 mutant, also denoted as ΔC or mSnapin, failed to interact with Snapin, while it still efficiently bound to the core components of class III PI3K complex, Beclin 1 and hVps34 (supplementary material Fig. S1C–F). Consistent with previous studies (Matsunaga et al., 2009; Sun et al., 2008), the Atg14L Δ71–180 (mBeclin 1) mutant carrying a deletion of its coiled-coil domain evidently lost Beclin-1-interacting activity, while retaining Snapin-binding ability (supplementary material Fig. S1C–F). Furthermore, confocal analysis revealed the substantial colocalizations of mCherry-fused Snapin with EGFP-fused Atg14L wild type (WT) [Pearson's correlation coefficient (PCC) = 0.778 Manders overlap coefficient (MOC) = 0.802] and EFGP–Atg14L (419–492) truncation mutant (PCC = 0.892, MOC = 0.899), while EFGP–Atg14L mSnapin mutant exhibited a minimal level of colocalization (PCC = 0.602, MOC = 0.645; Fig. 1F). Hence, these mutational analyses identified the Atg14L mutants defective in Snapin- or Beclin-1-binding ability, mSnapin and mBeclin 1, respectively, which were selected for the following functional studies.
Snapin enhances autophagosome maturation
We firstly examined whether Snapin plays an important role in the autophagy pathway. Upon autophagy induction, soluble microtubule-associated protein light chain 3 (LC3-I) is converted to a lipidated form (LC3-II) that preferentially associates with the growing autophagosome membrane in puncta formations (Kabeya et al., 2000). Thus, we monitored the effect of Snapin expression on autophagosome biogenesis with fluorescent microscopy after transiently expressing green fluorescent protein (GFP)-tagged LC3 (GFP–LC3) in snapin+/+ and snapin−/− mouse embryonic fibroblasts (MEFs). Compared with WT MEFs, snapin-deficient MEFs displayed significant increases in both the percentage of GFP–LC3-puncta-positive cells and the number of GFP–LC3 puncta per cell with or without rapamycin treatment (Fig. 2A). Furthermore, LC3-II and an autophagic substrate p62/SQSTM protein levels were substantially increased in snapin-deficient MEFs compared with WT MEFs, whereas reintroduction of snapin transgene into snapin-deficient MEFs reduced LC3-II and p62/SQSTM to levels comparable to those in WT MEFs (Fig. 2B). These results collectively indicate that autophagosomes are substantially accumulated upon the ablation of Snapin expression. Because autophagosomes can be accumulated due to either upregulation of autophagosome formation or blockade of autophagosome maturation, we sought to distinguish between the two possibilities by treating the cells with vacuolar H+ ATPase inhibitor bafilomycin A1 that abrogates autophagosome maturation (Fig. 2C). Bafilomycin A1 treatment increased LC3-II and p62/SQSTM protein levels in snapin+/+ MEFs, while snapin−/− MEFs exhibited no detectable change (Fig. 2C), suggesting that the aberrant accumulation of autophagosomes after Snapin deletion results from the impairment of autophagosome maturation rather than hyper-induction of autophagosome formation. Collectively, these results indicate that Snapin is pivotal for autophagy, especially during autophagosome maturation.
Snapin promotes endocytic trafficking
Since Snapin appears to serve as an important modulator of the late endocytic fusion machinery besides its established role in regulating synaptic vesicle fusion, we sought to explore the role of Snapin in endocytic trafficking. We firstly examined the degradation kinetics of epidermal growth factor receptor (EGFR) as the receptor-ligand complex is internalized from the cell surface into endosomes, which ultimately fuse with lysosomes for degradation. The results showed that EGFR degradation was severely delayed in snapin-depleted MEFs compared with WT MEFs (Fig. 3A). Subsequently, HeLa cells transfected with an empty vector or Snapin were exposed to Alexa-Fluor-488-conjugated EGF to visualize the effect of Snapin expression on the internalization, endocytic transport, and lysosomal degradation of EGFR associated with the exogenous fluorescent ligand (Fig. 3B). At 15 minutes post-stimulation, EGF was detected in small puncta structures in vector-expressing cells (PCC = 0.460, MOC = 0.388), whereas in Snapin-expressing cells it was present in cytoplasmic aggregates that were considerably colocalized with lysosomal-associated membrane protein 1 (LAMP1)-positive compartments (PCC = 0.496, MOC = 0.521). This difference became more pronounced at 30 minutes post-incubation (vector control, PCC = 0.605, MOC = 0.632; Snapin overexpression, PCC = 0.694, MOC = 0.644). Yet, there were no noticeable differences in the amounts of internalized EGF between snapin+/+ and snapin−/− MEFs at 15 minutes post-stimulation (supplementary material Fig. S2A,B). To further confirm the role of Snapin in endolysosomal trafficking, we then assessed the lysosomal proteolysis kinetics of a self-quenched red BODIPY dye-conjugated bovine albumin serum (DQ-Red BSA) after small hairpin RNA (shRNAmir)-mediated gene silencing of snapin in HeLa cells (Fig. 3C; supplementary material Fig. S2C). Consistently, DQ-Red BSA degradation rate was attenuated in HeLa cells expressing Snapin-specific shRNAmir compared with those carrying scrambled shRNAmir, indicating that ablation of the Snapin expression compromises endolysosomal trafficking. Altogether, these findings suggest that Snapin facilitates the trafficking of endocytic cargo to late endosomes and lysosomes for subsequent decomposition.
Atg14L–Snapin interaction is not required for autophagosome maturation
Given that Snapin positively regulates autophagosome maturation (Fig. 2), we next investigated whether the Atg14L–Snapin interaction functions in the late stage of autophagy. In an attempt to specifically monitor the role of Atg14L–Snapin binding during autophagosome maturation independent of the Atg14L–Beclin 1 interaction, we complemented atg14l knockout mouse embryonic stem cells (ESCs) with Atg14L Snapin- or Beclin 1-binding-deficient mutant (mSnapin or mBeclin 1, respectively; supplementary material Fig. S1) and monitored the steady-state levels of LC3-II and p62/SQSTM in atg14l−/− ESCs stably expressing vector, Atg14L WT, mSnapin, or mBeclin 1. Compared with atg14l+/+ ESCs, LC3-II was reduced and p62/SQSTM was markedly accumulated in atg14−/− ESCs in the absence and presence of rapamycin; this phenotype was readily rescued by introducing the atg14l WT transgene into atg14l-deficient ESCs (Fig. 4A). Interestingly, ectopic expression of the Atg14L mSnapin mutant effectively rescued autophagy defects, whereas the mBeclin 1 mutant expression failed to do so. These suggest that the Atg14L–Snapin interaction has little or no effect on autophagic cargo degradation. This phenotype was further confirmed by transiently introducing monomeric red fluorescent protein (mRFP)–GFP tandem fluorescently tagged LC3 (tfLC3) into atg14l ESCs (Fig. 4B). tfLC3 can be used as a probe to distinguish autophagosome formation from maturation due to the distinct pKa values of the two fluorescent proteins. In short, tfLC3 is stable in newly formed autophagosomes, giving rise to red and green fluorescence, whereas green fluorescence is lost in acidic autolysosomal compartments due to the high pKa value of GFP, thus showing only red fluorescence (Kimura et al., 2007). No detectable difference was observed between atg14l−/−-WT ESCs and atg14l−/−-mSnapin ESCs; the proportion of vesicles that solely exhibited red fluorescence was similar over the time course of rapamycin treatment. These results collectively indicate that the Atg14L–Snapin interaction is not required for autophagosome maturation.
Atg14L is required for endosome maturation
Provided that the autophagic and endocytic pathways converge at the endosome prior to lysosome-mediated degradation, we next asked whether Atg14L was functionally linked to endolysosomal transport through its interaction with Snapin. Firstly, we observed that Atg14L and Snapin were colocalized with the late endosomal proteins LAMP1 or CD63, but not with the early endosomal proteins EEA1 or Rab5 (supplementary material Fig. S3). We then tested the role of Atg14L expression in endocytic vesicle trafficking by silencing atg14l expression and examining the EGF-stimulation-mediated endocytic transport of EGFR to lysosomes. This showed that silencing atg14l expression in HeLa cells resulted in marked retardation of EGFR degradation, while it induced no gross alterations of organelle structures and distributions (Fig. 5A; supplementary material Fig. S4A). By striking contrast, silencing atg5 expression, another autophagy essential gene, had little or no effect on EGFR degradation kinetics (supplementary material Fig. S4B). Using Alexa-Flour-488-conjugated EGF, we also ascertained whether silencing atg14l expression affected the internalization, endocytic trafficking, and lysosome-mediated degradation of EGFR. Small interfering RNA (siRNA)-mediated knockdown of atg14l expression in HeLa cells substantially reduced the colocalization of Alexa-Fluor-488-conjugated EGF with the LAMP1-positive lysosomal compartments compared with scrambled siRNA treatment (Fig. 5B). However, there was no detectable difference in the amounts of internalized EGF between the scrambled siRNA-treated cells and the Atg14L-specific siRNA-treated cells at 15 minutes post-incubation (supplementary material Fig. S4C,D), indicating that the levels of EGF endocytosis are not affected by atg14l gene knockdown. However, the lysosomal proteolysis kinetics of DQ-Red BSA was considerably lower in cells treated with atg14l-specific siRNA compared with scrambled siRNA (Fig. 5C; supplementary material Fig. S4E). Altogether, these results suggest that Atg14L is a potential positive regulator of endocytic trafficking.
Atg14L facilitates endolysosomal trafficking through its interaction with Snapin
To determine whether the Atg14L–Snapin interaction functions in the endolysosomal pathway, we continued testing the effects of Snapin expression on EGF-stimulated endosomal transport under atg14l-silenced conditions. Ectopic expression of Snapin profoundly facilitated EGFR degradation in cells treated with scrambled shRNAmir, whereas no obvious effect was observed in cells treated with atg14l-specific shRNAmir (Fig. 6A), indicating that Snapin-mediated endocytic trafficking depends on Atg14L expression. To further support that Atg14L–Snapin interaction is required for endosomal transport, HeLa cells stably expressing the atg14l-specific shRNAmir were reconstituted with GST-fused Atg14L WT, mSnapin or mBeclin 1 transgene, each carrying silent mutations at the target sequence to confer resistance to the atg14l-specific shRNAmir. Introduction of atg14l-specific shRNAmir-resistant GST–Atg14L WT or GST–Atg14L mBeclin 1 mutant, but not the GST–Atg14 mSnapin mutant, restored efficient EGFR degradation kinetics (Fig. 6B), highlighting the importance of Atg14L–Snapin binding in the endocytic trafficking of EGFR. Given that three regions of Atg14L (aa71–106, aa270–320 and aa419–492) were individually sufficient for Snapin binding, we next examined whether the expression of any of these regions disrupted the Atg14L interaction with Snapin and thus exerted a dominant-negative effect on EGFR endocytic trafficking. Indeed, expression of the GST-fused Atg14L (419–492) truncation mutant suppressed the Atg14L–Snapin interaction, thereby causing a significant delay in EGFR degradation without affecting autophagic flux (supplementary material Fig. S5A,B; Fig. 6C). These results provide compelling evidence that Atg14L participates in endolysosomal trafficking through its interaction with Snapin.
In this report, we demonstrate that Beclin 1-binding autophagy protein Atg14L binds to SNARE-associated protein Snapin to promote endosomal maturation in mammalian cells. This finding is surprising given that Atg14(L) in both yeast and mammals has been primarily regarded as a specificity factor that enables the PI3KC3 complex to function in autophagy induction. There are indeed two distinct subsets of PI3KC3 complexes in yeast: the Vps38-containing PI3KC3 complex acts in vacuolar protein sorting, a phenomenon equivalent to late endosomal fusion with lysosomes in mammals, whereas the Atg14-containing PI3KC3 complex targets the complex to the pre-autophagosomal structure (Klionsky et al., 2008; Kihara et al., 2001). Likewise, recent studies have established mammalian Atg14L as a predominantly positive regulator of the PI3KC3 complex via its coiled-coil domain-mediated interaction with Beclin 1, highlighting the importance of Atg14L in the early stage of mammalian autophagy (Itakura et al., 2008; Matsunaga et al., 2009; Sun et al., 2008; Zhong et al., 2009). Further studies have elaborated the underlying mechanisms of Atg14L-dependent autophagosome formation: the N-terminal cysteine repeats of Atg14L direct the PI3KC3 complex to produce PI3P at local ER membranes and presumably cradle the precursor membranes of autophagosomes (Matsunaga et al., 2010). Its C-terminal BATS domain then binds to PI3P at the sprouting membranes to sense and/or stabilize the membrane curvature (Fan et al., 2011; Lu et al., 2011). These studies comprehensively demonstrate the vital roles of Atg14L in the early stage of autophagosome biogenesis. Our results have revealed an additional function of Atg14L in endosomal maturation in concert with Snapin in mammalian cells. We found that although the Atg14L–Snapin interaction is not required for the internalization of endocytic cargos, it promotes their endomembrane trafficking and lysosome-dependent degradation. Abrogation of Atg14L–Snapin binding led to significant retardation of stimulation-induced EGFR degradation, underscoring the importance of this interaction in the late stage of the endolysosomal pathway. In contrast, EGFR degradation kinetics were not altered when silencing other autophagy essential gene atg5, indicating that endolysosomal defects observed in atg14l knockdown cells is a specific phenotype of Atg14L rather than a general effect arising from global autophagic failure. These results indicate that Atg14L is a multivalent trafficking effector that regulates not only autophagosomal formation, but also late endosomal maturation, which may concurrently facilitate the transport of autophagic and endocytic cargos to digestive compartments.
Our findings provide compelling evidence that the function of Atg14L is not exclusively confined to the regulation of autophagosome formation but extends to the modulation of dynamic membrane trafficking at the late stage of the endolysosomal pathway. Indeed, the multivalent activity of mammalian Vps34 regulators in membrane trafficking has been well exemplified by UVRAG, a putative mammalian homolog of yeast Vps38 (Itakura et al., 2008). Our previous studies have demonstrated that UVRAG not only interacts with Beclin 1 to promote PI3KC3 lipid kinase activity for autophagosome formation, but also targets the HOPS complex and activates Rab7 GTPase activity to trigger autophagosomal or endosomal fusion with lysosomes (Liang et al., 2006; Liang et al., 2008). These findings collectively imply that, unlike their yeast counterparts, mammalian Vps34 regulators, Atg14L and UVRAG, are bestowed with pleiotropic functions owing to the complexity of endolysosomal systems in higher eukaryotes. Exploring specific spatiotemporal cues for Atg14L- versus UVRAG-mediated membrane trafficking in the endocytic and autophagic pathways would provide a better understanding of their seemingly parallel, yet specific actions.
Intriguingly, unlike the UVRAG–HOPS interaction, the Atg14L–Snapin binding had no pronounced impact on autophagy (Fig. 4). This appears quiet counterintuitive, considering the primary role of Atg14L per se in autophagy as well as the observed effect of Snapin on autophagosome maturation (Fig. 2). It is possible that Snapin is involved in autophagosome formation as a SNARE component based on the recent findings delineating the role of SNAREs in autophagosome biogenesis (Moreau et al., 2011; Nair et al., 2011). However, we found that disrupting Atg14L–Snapin interaction by the Atg14L (419–492) truncation mutant led to a dominant-negative effect on EGFR degradation without affecting autophagy (Fig. 6C; supplementary material Fig. S5B). Moreover, a recent study (Cai et al., 2010) has shown the roles of Snapin in the regulation of late endosomal trafficking and subsequent lysosome maturation via interaction with Dynein intermediate chain (DIC). Specifically, Snapin contributes to lysosome maturation by recruiting lysosomal hydrolases-containing late endosomes to the dynein motor complex for their trafficking along microtubules, thereby bringing late endosomes and immature lysosomes into sufficient proximity to allow highly efficient fusion events. Interestingly, this study also shows that autolysosome formation is intact under Snapin-depleted conditions while cargo clearance is severely diminished, indicating that Snapin is involved in autophagy pathway by ensuring lysosome maturation for the final degradation of autophagic cargo. Importantly, the study shows that Snapin interaction with the DIC affects the ‘migration’ of late endosomes along microtubules rather than the ‘fusion’ of late endosomes or autophagosomes with lysosomes. However, since atg14l expression showed no effect on the Snapin–DIC interaction (supplementary material Fig. S6A). On the other hand, the role of Snapin in SNARE-mediated membrane fusion suggests that the Atg14L–Snapin interaction may regulate endosomal fusions by targeting the SNARE machinery. Indeed, Snapin has been shown to associate with the LAMP1-positive late endosomal compartments and interact with the late endosomal SNARE proteins Syntaxin 8, Vti1b and VAMP8 (Lu et al., 2009). Likewise, we found that Atg14L was colocalized with Syntaxin 8 and Vti1b (supplementary material Fig. S6B). Hence, it is tempting to speculate that Atg14L regulates endolysosomal trafficking, such as homotypic fusion between late endosomes or heterotypic fusion with lysosomes, by targeting Snapin together with late endosomal SNARE complex. Finally, given the roles of PI3P in several endosomal functions (Lindmo and Stenmark, 2006), Atg14L–Snapin binding may rely on coincident recognition of PI3P. Our mapping study showed that Snapin interaction requires the C-terminal residues 419–480 of Atg14L (Fig. 1), largely overlapping with the BATS domain (residues 413–492) that preferentially binds to PI3P on highly curved membranes (Fan et al., 2011). Moreover, Snapin has also been reported to directly interact with PI3P (Tiwari et al., 2009). Therefore, it is possible that Atg14L and Snapin bind to PI3P and utilize these lipid moieties to trigger endosomal maturation. Yet, further studies are necessary to elaborate the molecular mechanisms of Atg14L-mediated endolysosomal trafficking.
In summary, our data indicate that Atg14L targets two independent molecular machineries, the PI3KC3–Beclin 1 complex and the SNARE-associated Snapin, to accomplish its dual roles in autophagosome formation and endosome maturation, respectively. Our study suggests that Atg14L functions as a multifaceted trafficking effector that orchestrates the autophagic and endocytic pathways depending on its binding partners, which allows proper coordination of these two distinct, yet interconnected trafficking processes.
Materials and Methods
HEK293T, HeLa and mouse embryonic fibroblast (MEF) cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum and 1% penicillin–streptomycin (Invitrogen). Snapin WT and KO MEFs were generously provided by Dr Zu-Hang Sheng (National Institutes of Health, MD) and immortalized with LXSN-E6/E7 retroviral vector containing human papilloma virus 16 E6 and E7 oncogenes using a standard protocol of selection using 200 µg/ml of neomycin (Sigma). Transient transfections were performed with Lipofectamine 2000™ (Invitrogen) or calcium phosphate (Clontech) following the manufacturer's instructions. HeLa and MEF stable cell lines were established by transfection followed by selection using 2 µg/ml of puromycin (Invitrogen). Mouse embryonic stem cells were graciously provided by Dr Tamotsu Yoshimori (Osaka University, Japan) and cultured in DMEM supplemented with 20% fetal bovine serum (Invitrogen), 1% penicillin–streptomycin (Invitrogen), 2 mM L-glutamine (Invitrogen), 0.1 mM non-essential amino acids (Invitrogen), 1 µM 2-mercapthoethanol (Invitrogen), and 1000 units/ml of Leukemia Inhibitory Factor (Sarkar et al., 2007) (Millipore) on plastic plates coated with 0.1% gelatin (Millipore). Mouse ES stable cell lines were established by electroporation followed by selection using 2 µg/ml of puromycin.
Full-length cDNA corresponding to the coding sequence of human atg14l gene was purchased from Open Biosystems and subcloned into the pGBKT7 vector between NdeI and SalI restriction sites to serve as bait in yeast two-hybrid screening. All plasmids for transient and stable expression in mammalian cells were derived from the pEF-IRES-puro vector encoding C-terminal FLAG or AU1 epitope tag or the pEBG GST fusion vector. Full-length and truncation mutant cDNA fragments corresponding to the coding sequence of atg14l gene were amplified from the template cDNA by polymerase chain reaction (PCR) and subcloned into the pEF-IRES-puro-FLAG vector between EcoRI and XbaI restriction sites or the pEBG vector between BamHI and NotI restriction sites. Because the atg14l gene contains a BamHI restriction site in its open reading frame, the PCR amplicon was digested with BglII that generates different but compatible overhangs with BamHI. Atg14L deletion mutants were generated by two-step PCR-directed mutagenesis. Full-length cDNA fragment corresponding to the coding sequence of the snapin gene was purchased from Open Biosystems and subcloned into the pEF-IRES-puro vector containing FLAG or AU1 epitope tag between AflII and XbaI restriction sites. pEGFP-LC3 plasmid was kindly provided by Dr Tamotsu Yoshimori (Osaka University, Japan), and ptfLC3 plasmid was purchased from Addgene. pEGFP-Atg14L plasmid was generously provided by Qing Zhong (University of California, Berkeley, CA), and Atg14 mutants (Δ419–480) and (419–492) were subcloned into the pEGFP-C2 vector between restriction sites BglII and EcoRI. Snapin was subcloned into pmCherry-C1 vector between restriction sites BglII and EcoRI.
For siRNA-mediated gene silencing, HeLa cells were transfected with stealth RNAi™ siRNA duplex specific for human Atg14L (5′-CCACUGCAUACCCUCAGGAAUCUAA-3′) and non-specific scrambled siRNA (Invitrogen) using Lipofectamine RNAiMAX™ (Invitrogen) according to the protocol provided by the manufacturer. At 48–72 h post-transfection, gene silencing was confirmed by immunoblotting and the cells were analyzed for EGF endocytosis and DQ-Red BSA dequenching assays. For shRNAmir-mediated gene silencing, HeLa cells were transduced with pGIPZ lentiviral shRNAmir specific for human Atg14L (5′-CAGCATGTAAATTTAGATCAA-3′), Snapin (5′-CCACAGAACTGTGCCGCATAA-3′), Atg5 (5′-CTTGGAACATCACAGTACA-3′) or non-silencing shRNAmir control (Open Biosystems) for 72 h following the manufacturer's instruction. The stable transfectants were selected with 2 µg/ml of puromycin. Atg14L–GST full-length, Δ419–480 (mSnapin), and Δ71–480 (mBeclin 1) resistant to pGIPZ lentiviral shRNAmir-mediated gene silencing were generated by two-step PCR-directed mutagenesis using the following mutagenic primer (the shRNAmir targeting sequence is capitalized): 5′-gtttttctcagCACGTGAACCTGGACCAGttacaaccac-3′.
Yeast two-hybrid screen
The human leukocyte Matchmaker™ cDNA library (Clontech) was screened using full-length human Atg14L as bait. Yeast transformation was performed using Matchmaker GAL4 Two-Hybrid System 3 (Clontech) according to the protocol provided by the manufacturer.
Immunoprecipitation and in vivo GST pulldown
HEK293T cells were harvested and lysed in NP40 buffer [50 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM EDTA, 0.5% Nonidet P-40] supplemented with a complete protease inhibitor cocktail (Roche). The whole-cell extracts were pre-clearing with Sepharose beads for 2 h at 4°C. For immunoprecipitation, the whole-cell extracts were incubated with 1–2 µg of the indicated antibodies at 4°C for 8–12 h. After adding protein A/G agarose beads, incubation was continued for an additional 2 h. For in vivo GST pulldown, the whole-cell extracts were incubated with 50% slurry of glutathione-conjugated Sepharose beads (GE Healthcare) for 2 h at 4°C. The precipitates were then extensively washed with lysis buffer and eluted by boiling with Laemmli sample buffer (Sigma) for 5 min.
Polypeptides were resolved by SDS-PAGE (SDS-PAGE) and transferred onto a PVDF membrane (Bio-Rad). Immunodetection was achieved with anti-FLAG (Sigma), anti-GST (Santa Cruz Biotech), anti-AU1 (Covance), anti-V5 (Invitrogen), anti-β-Actin (Santa Cruz Biotech), anti-Atg14L (Sigma and homemade), anti-Snapin (Synaptic Systems), anti-Beclin 1 (BD Biosciences), anti-LC3 (Cosmo Bio), anti-p62 (MBL), anti-EGFR (Millipore), anti-Atg5 (MBL), or anti-DIC (Millipore) antibody for 8–12 h at 4°C. To produce Atg14L polyclonal antibody, the peptide (APGCGPRPLARDLVDSVDDAEGLYVAVERCP) was immunized to rabbits and the antiserum from rabbits was subjected to affinity purification. The membrane was washed three times with PBS-T and incubated with the appropriate secondary antibody for 1 hr at room temperature. The proteins were visualized by chemiluminescence reagent (Denville) and detected by Fuji Phosphor Imager.
Confocal immunofluorescence microscopy
HeLa cells were fixed with 4% (w/v) paraformaldehyde (PFA) in phosphate buffered saline (PBS) for 30 min, permeabilized with 0.2% (v/v) Triton X-100 in PBS for 10 min, and blocked with 10% goat serum (Invitrogen) in PBS for 1 h. Primary antibody staining was performed using anti-FLAG (Sigma), anti-GFP (Santa Cruz Biotech), anti-LAMP1 (Abcam), anti-CD63 (Santa Cruz Biotech), anti-EEA1 (BD Biosciences), anti-Rab5 (Santa Cruz Biotech), Syntaxin 8 (BD Biosciences) or Vti1b (BD Biosciences) antibody diluted in PBS containing 1% goat serum for 1–2 h at room temperature. The cells were washed three times with PBS and incubated with the appropriate secondary antibody diluted in PBS containing 1% goat serum for 1 h at room temperature. The cells were then washed three times with PBS and nucleus was stained using Hoechst 33342 (Invitrogen). After slide mounting using fluorescent mounting medium (Dako), the confocal images were acquired using a Nikon Eclipse C1 laser-scanning microscope (Nikon Instruments Inc.) fitted with a 100× Nikon objective and Nikon Element image software.
For transmission electron microscopy (TEM) microscopy, the standard procedure was applied. Briefly, the cells were fixed by ½K and osmium tetroxide solutions, stained uranium salt, dehydrated, and infiltrated with Epon plastic (TED PELLA, Redding, CA). The samples were then embedded in Epon blocks and polymerized overnight. After ultra thin sections were prepared (MT6000, Servall Instrument), samples were picked up on grids, followed by image analysis using JEOL JEM 2100 electron microscope.
For autophagy induction, cells were washed two times with PBS and incubated with rapamycin (Sigma) at a final concentration of 2 µM in serum-free DMEM (and 1000 units/ml of LIF in the case of mouse ES cells) for 2–8 h (3 h in general). For inhibition of autophagosome maturation, cells were incubated with bafilomycin A1 (Sigma) at a final concentration of 0.1 µM in complete medium (containing 1000 units/ml of LIF in the case of mouse ES cells) for 1 h. Autophagy was assessed by monitoring redistribution of GFP–LC3 or tfLC3, LC3 conversion, and p62 degradation. For GFP–LC3 redistribution, MEFs were transfected with GFP–LC3 plasmid. At 16 h post–transfection, GFP–LC3 was detected under normal conditions and in the presence of rapamycin using an inverted fluorescence microscope. The percentage of GFP–LC3-positive cells with puncta staining was determined in three independent experiments. To quantify GFP–LC3-positive autophagosomes, six random fields representing 200 cells were counted. For tfLC3 redistribution, mouse ES cells were electroporated with ptfLC3 plasmid. After treatment with rapamycin for the indicated time points at 24 h post-electroporation, the cells were fixed with 4% PFA and processed for confocal microscopy analysis. To monitor LC3 conversion or p62 degradation, cells were treated on ice for 30 min, lysed with 1% Triton X-100 buffer (50 mM Tris-HCl [pH 8.0], 150 mM NaCl, 1 mM EDTA, 1% Triton X-100) supplemented with a complete protease inhibitor cocktail (Roche), and subjected to immunoblot analysis with anti-LC or anti-p62 antibody.
EGF endocytosis and EGFR degradation assays
For EGF uptake, cells were washed twice with PBS and starved in serum-free DMEM for 2 h at 37°C. The cells were then washed with ice-cold PBS and incubated on ice in uptake medium (20 mM HEPES and 2% BSA in DMEM) containing 5 µg/ml of EGF complexed to Alexa-Fluor-488–streptavidin (Invitrogen). After incubation on ice for 1 h, the cells were washed three times with ice-cold PBS to remove unbound ligand. The cells from one well were fixed with 4% PFA to give the total amount of bound ligand and the remaining wells were transferred to a 37°C incubator. At each indicated time point, the cells were fixed with 4% PFA and processed for confocal microscopy analysis.
For EGFR degradation assay, cells were cultured to ∼80% confluency, washed twice with PBS and starved in serum-free DMEM for 2 h at 37°C. EGFR endocytosis was stimulated by adding 200 ng/ml of EGF (Sigma) in DMEM containing 20 mM HEPES and 0.2% BSA. At each time point after EGF stimulation, the cells were harvested and the amount of protein in each sample was measured using the BCA protein assay kit (Pierce) following the manufacturer's protocol. Equal amount of the extracts was subjected to immunoblot analysis to quantify the total EGFR protein level.
BSA dequenching assay
HeLa cells were incubated with 20 µg/ml of DQ™-Red BSA (Invitrogen) in complete DMEM at 37°C for the indicated time periods (Vázquez and Colombo, 2009). After washing twice with ice-cold PBS to remove excess probe, the cells were harvested and resuspended in 4% PFA. Red-fluorescent DQ-Red BSA was then analyzed by flow cytometry using FACSCanto II (BD Biosciences) and FlowJo software (Treestar, Inc.).
Statistical analyses were performed using Student's paired t-test with a two-tailed distribution. Values are expressed as mean ± s.d. of at least three independent experiments, unless otherwise noted. A P-value of ≤0.05 was considered statistically significant.
We thank Stacy Lee for the manuscript preparation and all of the members of the J.U.J. laboratory for their support and discussion.
This work was partly supported by the Hastings Foundation (grant number 22-2107-1000 to J.U.J.), the Fletcher Jones Foundation [grant number 22-2101-1003 to J.U.J.], the Global Reaserch Laboratory Program from National Research Foundation of Korea (grant number K20815000001 to JUJ), and National Institute Health [grant numbers CA82057, CA31363, DE019085, AI073099, and AI083025 to J.U.J; CA140964 and AI083841 to C.L.], and American Cancer Society (grant number RGS-11-121-01-CCG to C.L.). Deposited in PMC for release after 12 month.