TORC1 specifically inhibits microautophagy through ESCRT-0
- 702 Downloads
Nutrient starvation induces the degradation of specific plasma membrane proteins through the multivesicular body (MVB) sorting pathway and of vacuolar membrane proteins through microautophagy. Both of these processes require the gateway protein Vps27, which recognizes ubiquitinated cargo proteins at phosphatidylinositol 3-phosphate-rich membranes as part of a heterodimeric complex coined endosomal sorting complex required for transport 0. The target of rapamycin complex 1 (TORC1), a nutrient-activated central regulator of cell growth, directly phosphorylates Vps27 to antagonize its function in microautophagy, but whether this also serves to restrain MVB sorting at endosomes is still an open question. Here, we show that TORC1 inhibits both the MVB pathway-driven turnover of the plasma membrane-resident high-affinity methionine permease Mup1 and the inositol transporter Itr1 and the microautophagy-dependent degradation of the vacuolar membrane-associated v-ATPase subunit Vph1. Using a Vps277D variant that mimics the TORC1-phosphorylated state of Vps27, we further show that cargo sorting of Vph1 at the vacuolar membrane, but not of Mup1 and Itr1 at endosomes, is sensitive to the TORC1-controlled modifications of Vps27. Thus, TORC1 specifically modulates microautophagy through phosphorylation of Vps27, but controls MVB sorting through alternative mechanisms.
KeywordsTarget of rapamycin complex 1 (TORC1) Endosomal sorting complex required for transport (ESCRT) Multivesicular body Microautophagy Vps27
Living cells control growth in a homeostatic manner by tightly coupling protein synthesis and catabolism to the availability of nutrients. Upon nutrient starvation, cytosolic proteins are degraded through macroautophagy, while specific plasma membrane- and lysosomal membrane-resident proteins are catabolized through the multivesicular body (MVB) sorting pathway and microautophagy, respectively (Müller et al. 2015; Oku et al. 2017; Zhu et al. 2017). Macroautophagy requires a defined set of autophagy-related (Atg) proteins (Mizushima et al. 2011; Wen and Klionsky 2016) and is mechanistically different from MVB sorting, which relies on the endosomal sorting complex required for transport (ESCRT) machinery (Henne et al. 2011). Interestingly, recent studies have demonstrated that ESCRT also drives microautophagy at lysosomal (i.e., yeast vacuolar) membranes, although ESCRT-independent subtypes of microautophagy may also exist (Oku and Sakai 2018). In both processes, MVB sorting at endosomal membranes and ESCRT-driven microautophagy at vacuolar membranes, Vps27 forms a heterodimer with Hse1 (named ESCRT-0) that initiates the sorting of membrane-resident and associated proteins following their posttranslational ubiquitination. Accordingly, cargo recognition by ESCRT-0 is followed by invagination of the cargo-containing membrane and formation of intraluminal vesicles that are directly degraded within the vacuole (during microautophagy) or following the fusion of endosome-derived MVBs with the vacuole (during the ultimate step of the MVB pathway). How nutrients control these processes has largely remained elusive.
The target of rapamycin complex 1 (TORC1) is a central integrator of nutritional information that regulates various anabolic and catabolic processes to ensure cellular homeostasis in eukaryotic cells (Albert and Hall 2015; Eltschinger and Loewith 2016; Saxton and Sabatini 2017). In yeast, TORC1 inhibits macroautophagy via direct phosphorylation of Atg13, which reduces its affinity for and promotes dissociation from Atg1, a protein kinase that centrally induces macroautophagy when in complex with Atg13 (Kamada et al. 2010). TORC1 also controls the MVB pathway-driven degradation of plasma membrane proteins (Dauner et al. 2017; Dobzinski et al. 2015; Jones et al. 2012; MacGurn et al. 2011), and conceivably also the microautophagic turnover of vacuolar membrane proteins (Oku et al. 2017; Rahman et al. 2018; Zhu et al. 2017), although the underlying mechanisms have remained incompletely understood. In this context, we recently identified two spatially and functionally distinct pools of TORC1 that reside on the limiting membranes of vacuoles and some endosomes that we coined signaling endosomes (Hatakeyama and De Virgilio 2019; Hatakeyama et al. 2019). Interestingly, vacuolar TORC1 plays a prominent role in regulating protein synthesis via its proximal effector Sch9 (Ejzykowicz et al. 2017; Hatakeyama et al. 2019; Jin et al. 2014; Urban et al. 2007). Endosomal TORC1, in contrast, phosphorylates Atg13 to antagonize initiation of macroautophagy, which occurs at perivacuolar foci called preautophagosomal structures that are in close proximity to endosomal membranes (Hatakeyama et al. 2019; Suzuki et al. 2013). Endosomal TORC1 is further specifically commissioned to phosphorylate the bulk of Vps27 (at 7 Ser/Thr residues) and thereby antagonize the ESCRT-driven microautophagic degradation of the vacuolar membrane protein phosphatase Pho8. Whether the TORC1-mediated phosphorylation of Vps27 also affects cargo sorting through the MVB pathway, however, is currently not known.
Here, we show that TORC1 inhibits sorting and degradation of both specific plasma membrane proteins (via the MVB pathway) and of the vacuolar membrane protein Vph1 (via microautophagy). Using a Vps277D variant that mimics the TORC1-triggered phosphorylated state of Vps27, we further show that cargo sorting during microautophagy, but not MVB sorting, is sensitive to the TORC1-controlled modifications of Vps27. Thus, TORC1 appears to specifically tune microautophagy, but not MVB sorting, through phosphorylation of Vps27.
Materials and methods
Yeast strains, plasmids, and growth conditions
Saccharomyces cerevisiae strains used in this study, all derived from BY4741/2 (Brachmann et al. 1998), are RKH119 (MATa, vps27∆::KanMX, his3∆1, leu2∆0, ura3∆0), RKH495 ([RKH119] VPH1-EGFP::HIS3), RKH463 ([RKH119] MUP1-EGFP::HIS3), RKH483 ([RKH119] ITR1-GFP::HIS3 met15∆0), and RKH360 ([RKH119] URA3::PPHO5-GFP-CPS1 met15∆0). The plasmids used in this study are: pRS415 (CEN, ARS, LEU2) and pRS416 (CEN, ARS, URA3) (Sikorski and Hieter 1989); p1379 (CEN, ARS, URA3, MET15), p3505 (CEN, ARS, LEU2, VPS27), and p3550 (CEN, ARS, LEU2, VPS27S155D/S157D/T159D/S274D/S277D/S279D/S280D) (Hatakeyama et al. 2019); and p1770 (CEN, ARS, HIS3, MET15), p3628 (CEN, ARS, URA3, VPS27) and p3629 (CEN, ARS, URA3, VPS27S155D/S157D/T159D/S274D/S277D/S279D/S280D) (this study). Yeast strains were made prototrophic by complementing auxotrophic markers with empty vector plasmids and grown to exponential growth phase in synthetic dextrose medium (0.17% yeast nitrogen base, 0.5% ammonium sulfate, and 2% glucose) at 30 °C. In Fig. 2a, b, leucine (14.8 mg l−1 final concentration) was supplemented to the medium to support the expression of Mup1-GFP.
Images of live fluorescent cells were captured with an inverted spinning disk confocal microscope (VisiScope CSU-W1, Puchheim, Germany) that is equipped with a scientific grade 4.2 sCMOS camera and a 100 × 1.3 NA oil immersion Nikon CFI series objective (Egg, Switzerland), and processed using ImageJ software. For each strain and condition, at least 60 cells were recorded and the pictures show representative cells in each case.
Cell lysate preparation and immunoblot analyses
Cells were treated with 6.7% w/v trichloroacetic acid (final concentration) for at least 10 min on ice, pelleted, washed with ice-cold 70% ethanol, dried, dissolved in urea buffer [50 mM Tris–HCl (pH 7.5), 5 mM EDTA, 6 M urea, 1% SDS, 0.4 mM Pefabloc (Sigma-Aldrich), and PhosSTOP (Roche, one tablet per 10 ml)], and disrupted with glass beads using a Precellys homogenizer. After being heated at 65 °C for 10 min twice, with the second heating performed in Laemmli SDS sample buffer, samples were subjected to SDS-PAGE and transferred to nitrocellulose membranes. Subsequent to blocking with blocking buffer (1% bovine serum albumin and 5% milk powder in phosphate-buffered saline), membranes were immunoblotted with the following primary antibodies diluted in the blocking buffer: mouse anti-GFP (Roche, product number 118144600001; 1:1000 dilution), rabbit anti-Vps27 (lab stock; 1:1000 dilution) or rabbit anti-Adh1 (Calbiochem, product number 126745; 1:200,000 dilution). After washing with Tris-buffered saline containing 0.3% Tween 20, membranes were incubated with mouse (BIO-RAD, product number 170-6516; 1:3000 dilution) or rabbit (BIO-RAD, product number; 170-6515; 1:3000 dilution) secondary antibodies conjugated with horseradish peroxidase, washed again and developed with ECL (GE Healthcare). All immunoblot analyses were carried out in triplicates (biological replicates) of which one representative blot is shown in each case.
TORC1 controls microautophagic degradation of Vph1 via ESCRT-0
TORC1 does not modulate MVB sorting of Mup1 and Itr1 through Vps27 phosphorylation
TORC1-mediated phosphorylation of Vps27 does not affect steady state MVB sorting
Our analyses of Vph1-GFP indicated that Vps277D functioned normally in MVB sorting in exponentially growing cells (Fig. 1a). To further substantiate this notion, we also examined the localization of the vacuolar peptidase Cps1, which is continuously delivered to the vacuolar lumen through ESCRT-driven MVB sorting in exponentially growing cells (Brune et al. 2019; Odorizzi et al. 1998). GFP-Cps1 accumulated in the class E compartment and to some extent on the vacuolar surface in exponentially growing vps27∆ cells as previously described (Mageswaran et al. 2015; Odorizzi et al. 1998), but normally reached the vacuolar lumen in both Vps27- and Vps277D-expressing cells (Fig. 2e). Thus, the steady state flux through the MVB pathway appears not to be restrained by TORC1-mediated phosphorylation of Vps27.
The data presented here not only corroborate our recent discovery according to which TORC1 controls microautophagy through phosphorylation of Vps27, but also indicate that the respective regulatory mechanism does not modulate the flux through the MVB pathway. The underlying mechanistic details of this differential importance of Vps27 phosphorylation for microautophagy and MVB sorting is currently elusive, but may be related to the observation that Vps27 is mainly targeted by TORC1 on endosomal membranes, which are very rich in phosphatidylinositol 3-phosphate (PI3P) when compared to vacuolar membranes (Marat and Haucke 2016). Accordingly, we speculate that TORC1-mediated phosphorylation of Vps27 may reduce its capacity to act as a coincidence detection module that binds ubiquitinated cargo on membranes through both its ubiquitin-interacting and PI3P-binding domains (Henne et al. 2011). In such a scenario, Vps27 may be able to dock onto ubiquitinated cargo on PI3P-rich endosomes, but may be compromised to do so on vacuolar membranes as long as it is kept in its phosphorylated state by endosomal TORC1. The binding and dissociation constants between Vps27 and its ligands combined would then define the fraction of Vps27 that is released from endosomes to probe ubiquitinated cargoes on the vacuolar membrane. This model could be addressed in the future by studying the respective biochemical parameters in vitro (using purified variants of Vps27 and differently decorated liposomes) and microscopic imaging of Vps27 mobility in vivo [using for instance tandem fluorescent protein timers as described (Khmelinskii et al. 2012)].
Although TORC1 does not apparently modulate MVB sorting through Vps27, it antagonizes the endocytosis and MVB pathway-driven degradation of certain plasma membrane permeases [e.g., Mup1 and Itr1 (studied here), or the high-affinity tryptophan permease Tat2 (Beck et al. 1999)] via alternative means. It may do so in part by inhibiting the effector protein kinase Npr1, which controls the ubiquitination state, and consequently the trafficking through the endocytic pathway, of various plasma membrane proteins via mechanisms that implicate arrestin-related ubiquitin ligase adaptors and their associated ubiquitin ligases (Beck et al. 1999; Jones et al. 2012; MacGurn et al. 2011; Merhi and André 2012). The respective mechanisms remain largely to be elucidated, but TORC1 can regulate the access of plasma membrane proteins to the endocytic pathway rather specifically and, depending on the protein, in opposite ways. In this context, it is worth noting that both Mup1 and Itr1 also undergo substrate-induced endocytosis (Lai et al. 1995; Lin et al. 2008; Menant et al. 2006), which, analogous to the situation for Gap1 (Ghaddar et al. 2014), is likely caused by substrate-induced conformational changes that promote the recognition of these permeases by arrestin-like adaptors of the Rsp5 ubiquitin ligase (Guiney et al. 2016; Hatakeyama et al. 2010; Nikko and Pelham 2009). In our experimental setting, cells were not exposed to methionine and only to rather low amounts of inositol (11 µM) in the medium, which is why both permeases were quite stably concentrated at the plasma membrane in exponentially growing cells. Conceivably, however, TORC1 inhibition may increase the cytoplasmic levels of these metabolites as a result of the inhibition of methionine- and inositol-consuming anabolic processes. If this were the case, it would be interesting to address the question whether high cytoplasmic levels of these or other metabolites can, similarly as proposed in the case of the transporter-related Ssy1 amino acid sensor (Wu et al. 2006), induce conformational changes in permeases, which then would trigger their endocytosis. Intriguingly, TORC1 downregulation may then, perhaps like treatment of cells with cycloheximide (Lin et al. 2008; Nikko and Pelham 2009), control permease sorting indirectly by causing specific metabolites to accumulate in the cytoplasm. Finally, TORC1 may also control MVB sorting by other means such as for instance by promoting the expression and/or stability of Ist1 that functions as an ESCRT inhibitor (Jones et al. 2012).
The v-ATPase acts upstream of TORC1 in flies, mammals, and yeast (Dechant et al. 2014; Eskes et al. 2018; Hatakeyama et al. 2019; Stauffer and Powers 2016; Zoncu et al. 2011). Our finding that TORC1 prevents microautophagic degradation of the v-ATPase subunit Vph1, therefore, indicates the existence of a positive feedback control mechanism that may contribute to the robustness of TORC1 signaling. The degradation of Vph1-GFP upon inhibition of TORC1, however, is rather slow and only affects a minor fraction of the total pool of Vph1-GFP over the period studied (which also explains why the 3-h rapamycin treatment does not visibly enhance the GFP signal within the vacuoles). This raises the question whether Vph1 is specifically targeted for degradation by a vacuolar membrane-resident ubiquitin ligase or whether, due to its abundance within the vacuolar membrane, follows the microautophagic degradation of other proteins. Thus, future progress in this area will not only require the analysis of the ubiquitinated residues within Vph1 (Swaney et al. 2013), but also the identification of additional bona fide microautophagic substrates. The phosphomimetic Vps277D allele that specifically compromises Vps27-dependent microautophagy will likely be useful in these efforts, which should ultimately clarify the importance of microautophagic processes for the survival of starved cells.
This work was supported by the Canton of Fribourg and the Swiss National Science Foundation (310030_166474/1; CDV). We thank Christian Ungermann for the GFP-Cps1 strain (Markgraf et al. 2009) and Marie-Pierre Péli-Gulli for critically reading the manuscript.
- Hatakeyama R, Péli-Gulli MP, Hu Z, Jaquenoud M, Garcia Osuna GM, Sardu A, Dengjel J, De Virgilio C (2019) Spatially distinct pools of TORC1 balance protein homeostasis. Mol Cell 73(325–338):e328Google Scholar
- Sikorski RS, Hieter P (1989) A system of shuttle vectors and yeast host strains designed for efficient manipulation of DNA in Saccharomyces cerevisiae. Genetics 122:19–27Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.