The identification and analysis of phosphorylation sites on the Atg1 protein kinase

Abstract

Autophagy is a conserved, degradative process that has been implicated in a number of human diseases and is a potential target for therapeutic intervention. It is therefore important that we develop a thorough understanding of the mechanisms regulating this trafficking pathway. The Atg1 protein kinase is a key element of this control as a number of signaling pathways target this enzyme and its associated protein partners. These studies have established that Atg1 activities are controlled, at least in part, by protein phosphorylation. To further this understanding, we used a combined mass spectrometry and molecular biology approach to identify and characterize additional sites of phosphorylation in the Saccharomyces cerevisiae Atg1. Fifteen candidate sites of phosphorylation were identified, including nine that had not been noted previously. Interestingly, our data suggest that the phosphorylation at one of these sites, Ser-34, is inhibitory for both Atg1 kinase activity and autophagy. This site is located within a glycine-rich loop that is highly conserved in protein kinases. Phosphorylation at this position in several cyclin-dependent kinases has also been shown to result in diminished enzymatic activity. In addition, these studies identified Ser-390 as the site of autophosphorylation responsible for the anomalous migration exhibited by Atg1 on SDS-polyacrylamide gels. Finally, a mutational analysis suggested that a number of the sites identified here are important for full autophagy activity in vivo. In all, these studies identified a number of potential sites of regulation within Atg1 and will serve as a framework for future work with this enzyme.

Introduction

The term autophagy refers to a collection of membrane trafficking pathways responsible for the turnover of cytoplasmic constituents within eukaryotic cells.^1–3 The cargo for these pathways includes bulk protein, invading bacterial pathogens, abnormal protein aggregates and damaged and superfluous organelles.^4–9 These materials are taken up in either a specific or nonspecific manner and are targeted ultimately to the lysosome for degradation.¹⁰ This transport is important for normal cellular homeostasis and recent studies have implicated these degradative processes in a number of human diseases.^11–14 These conditions include specific cancers, Crohn disease and neurological disorders, like Huntington disease. In many of these disorders, the autophagy pathway is being examined as a means of therapeutic intervention.^15–18 It is therefore important that we develop a thorough understanding of the components involved in these pathways and the manner in which these activities are regulated.

A number of studies indicate that these autophagy processes are highly regulated and that the Atg1 protein kinase appears to be a key element in this control.^19,20 In S. cerevisiae, Atg1 and its associated proteins are targeted by at least three different signal transduction pathways important for coordinating cell growth with nutrient availability.^21,22 Two of these pathways involving TORC1 and the cAMP-dependent protein kinase (PKA) inhibit this degradative process.^23–25 The TORC1 complex contains the target of rapamycin (TOR) proteins, serine/threonine-specific protein kinases that regulate growth in all eukaryotes.^26,27 PKA has been shown to directly phosphorylate Atg1 in this budding yeast, and both TORC1 and PKA phosphorylate Atg13, a positive regulator of Atg1 kinase activity.^28–30 The third signaling pathway stimulates autophagy and involves the AMP-activated protein kinase (AMPK) homolog, Snf1.³¹ The PKA, TORC1 and AMPK kinases have also been implicated in the regulation of autophagy in other organisms, including mammals.^32–36

The addition of particular phosphate residues is often used to modulate the activity of a given protein. This is certainly the case for the phosphorylation events identified thus far in the S. cerevisiae Atg1. For example, an autophosphorylation within the Atg1 activation loop, a conserved element in the kinase domain, has been shown to be necessary for both Atg1 kinase activity and the induction of macroautophagy.³⁷ Macroautophagy is a nonspecific process that is perhaps the best understood of the autophagy pathways;^38,39 for the remainder of this report, we will refer to this transport process as simply autophagy. During this degradative process, a double membrane grows out from a nucleation site in the cytoplasm, known as the phagophore assembly site (PAS).^40,41 This phagophore membrane encapsulates nearby material and ultimately packages it into a transport intermediate, called the autophagosome.^4,42 The phosphorylation of Atg1 by PKA disrupts the Atg1 association with the PAS, and thereby inhibits this autophagy process.²⁸ Finally, studies with other eukaryotes suggest that Atg1 might also be a substrate of the TORC1 signaling complex.^43–46 In all, this work indicates that the phosphorylation of Atg1 is important for the proper control of autophagy.⁴⁷ To better understand the extent of this regulation, we used a combination of mass spectrometry (MS) and molecular genetic methods to identify and characterize additional sites of phosphorylation on the S. cerevisiae Atg1. The positions of these modifications and their potential roles in the regulation of autophagy are discussed herein.

Results and Discussion

MS identification of potential phosphorylation sites on Atg1.

Tandem MS/MS spectrometry was used to identify potential sites of phosphorylation in Atg1. To facilitate a more comprehensive mapping of these sites, phosphopeptides were enriched from the in-gel digested Atg1 tryptic peptide pool by the use of a TiO₂ microcolumn. Both the original nonenriched sample and the TiO₂-bound fraction containing a majority of the phosphopeptides were subjected to several repeated runs of nanoLC-MS/MS analysis under data-dependent acquisition mode and the resulting MS/MS datasets were individually searched against the NCBInr database using the Mascot search engine. Combining the search results from all analyses provided a protein sequence coverage of 67% in total, with no close second ranking protein contaminant identified. The search results were therefore indicative of a relatively high purity of the Atg1 gel band such that all high and low scoring Atg1 tryptic phosphopeptides identified could represent positive hits in the first instance. The corresponding MS/MS spectra for each of the tentatively assigned phosphorylation sites were then manually verified and annotated for the sequence informative b and y fragment ions (see Fig. S1). In general, phosphorylated site assignment was considered reliable if the b and y ions flanking the implicated site could be detected, with mass shifts corresponding to a phosphorylated Ser or Thr, or one having eliminated its phosphate (-98 u; see Materials and Methods). By such criteria, all of the assigned sites could be manually verified except the two sites at Ser-551 and Ser-552, carried on the mono-phosphorylated peptide 536–575, which could not be unambiguously resolved. Notably, the mono-phosphorylated peptide 675–688 was identified thrice, corresponding to the same peptide eluting at slightly different retention times and being phosphorylated at three different positions, Ser-677, Ser-680 and Ser-683. Taken together, a total of 14 potential phosphorylation sites were identified, nine of which (Ser-34, Ser-304, Ser-533, Ser-551/552, Ser-621, Ser-680, Ser-769, Thr-129 and Thr-590) were not reported previously (Fig. 1A and B and Table 1).⁴⁸

Figure 1.

A phosphorylation site map for the S. cerevisiae Atg1. (A) The phosphorylation sites in Atg1 that were identified in this study. The 15 sites identified herein are indicated above the Atg1 schematic. All were identified by the MS/MS analysis except for S390 (indicated by a double asterisk). S515, indicated by a single asterisk, was identified previously as a site of PKA phosphorylation.²⁸ The two positions indicated below the Atg1 schematic were identified previously: S508 was also found to be a site of PKA phosphorylation and T226 is a conserved site of autophosphorylation within the Atg1 kinase domain.²⁸,³⁷ The conserved kinase and C-terminal domains of Atg1 are indicated with the light gray (KINASE) and dark gray (CTD) shading, respectively. (B) The list of tryptic phosphopeptides identified by the tandem MS/MS analysis. The data shown include the first and last positions of the peptide, the number of phosphates present (#P), the amino acid that was phosphorylated (Mod. AA), the Mascot score, the mass deviation in ppm (Mass dev) and the charge state of each peptide (m/z). The identified sites are grouped into two categories based on the relative confidence in their respective identifications; the lower scoring sites are listed in the bottom portion of the table. Phosphorylation sites assignments, including those for the three low scoring phosphopeptides (Thr-129, Ser-533 and Ser-680), were in general supported by the observed MS/MS ions, as manually interpreted and annotated in Figure S1.

Table 1.

A comparison of the phosphorylation sites identified in the S. cerevisiae Atg1

Present MS study	Albuquerque et al. study	Other work	Comments
S34
T129
		T226	Activation loop
S304
S365	S365
	S390	S390	Responsible for band shift
	S508	S508	PKA site
S515	S515	S515	PKA site
	S517
S533
S551/2
T590
S621			“LSTT” site
	S647
S677	S677		“LSAT” site
S680
S683	S683		“LSAT” site
S769
S783	S783

The first column lists the sites identified in the MS analysis performed here, the second column the sites identified in a previous MS study and the third column sites that were identified by other methods.²⁸,³⁷,⁴⁸ The sites underlined were identified in only one of the studies summarized here.

The phosphorylation sites identified here on the S. cerevisiae Atg1 were compared with those detected recently on the mouse Ulk1 protein.⁴⁹ Ulk1 is the mammalian protein kinase that exhibits the highest degree of sequence similarity to Atg1.⁵⁰ Although the phosphorylation sites are distributed throughout both proteins in a similar manner, we did not detect any obvious overlap between the sites thus far identified. However, many of the phosphorylation sites occur within the less conserved, serine-rich central region of Atg1. As a result, even if these sites were serving similar functional roles, it might be difficult to discern this from sequence alignments alone.

Ser-34 as a potential site of inhibitory control.

The MS/MS analysis identified Ser-34 within the Atg1 kinase domain as a site of phosphorylation (Fig. 1A and B). This position was of interest because it is located within a sequence element that is highly conserved in protein kinases. The residues within this motif form a glycine-rich loop that interacts with the ATP nucleotide and is responsible for positioning the γ-phosphate for catalysis (Fig. 2A).^51,52 Moreover, phosphorylation at the analogous position in several cyclin-dependent kinases has been shown to inhibit protein kinase activity.⁵³ Therefore, we tested whether the phosphorylation at Ser-34 in Atg1 might influence the autophagy process.

Figure 2.

Phosphorylation at Ser-34 within Atg1 is a potential site of inhibition for the autophagy process. (A) A sequence alignment of the glycine-rich loop (G-loop) present in Atg1 and the mammalian Cdc2 enzymes. The consensus sequence for this motif is shown at the top and the residues known to be phosphorylated are underlined. (B) Assessing autophagy with the Pho8Δ60 alkaline phosphatase-based assay. The relative levels of autophagy for an atg1Δ strain carrying single-copy plasmids expressing the indicated variants of Atg1 are shown. Autophagy was induced by treating log phase cultures with 200 ng/ml rapamycin for 4 h. The data shown are the average of at least three independent experiments and the error bars indicate one standard deviation. (C) Autophagy activity as measured with the Ape1 processing assay. Cells overexpressing Ape1 were treated with 200 ng/ml rapamycin for the time indicated to induce the autophagy process. The relative level of the mature, processed form of the Ape1 protein (mApe1) was an indicator of the autophagy activity present. The bottom part shows the relative levels of the different myc-tagged Atg1 proteins in the strains examined. The electrophoretic conditions used for this latter control would not have resolved the two forms of Atg1 that comprise the doublet shown elsewhere. (D) An analysis of Atg1 autophosphorylation activity in vivo. The relative level of the autophosphorylated form of Atg1 (Atg1-P) was assessed in atg1Δ cells expressing the indicated variants of Atg1. Western blots were performed with extracts prepared from either log phase (L) or rapamycin-treated (R) cells. (E) An examination of the effects of altering Ser-34 on Atg1 autophosphorylation in vitro. The indicated myc epitope-tagged Atg1 proteins were immunoprecipitated and then either mock treated (IP) or incubated with λ phosphatase (PPase). An aliquot of the phosphatase-treated protein was then subjected to an in vitro autophosphorylation reaction (IVKA) with 2.5 mM ATP, as described in the Materials and Methods. The resulting reaction products were separated by SDS-polyacrylamide gel electrophoresis and the relative level of the autophosphorylated form of Atg1, Atg1-P, was assessed by western blotting. The (A, D and E) designations indicate the Atg1^S34A, Atg1^S34D and Atg1^S34E proteins, respectively.

To examine this possibility, we used site-directed mutageneses to change this serine to a variety of other amino acids, including the potential phosphomimetic residues, glutamic and aspartic acid. These latter two acidic residues can functionally substitute for a phosphorylated serine in some instances.⁵⁴ The ability of these variants to function in the autophagy process was assessed with an assay that measures the transport to the vacuole of an altered Pho8 alkaline phosphatase, known as Pho8Δ60.^55,56 Pho8Δ60 lacks the sequence information needed for normal transit through the secretory pathway, and this protein can only be delivered to the vacuole, and activated, via the autophagy pathway. We found that Atg1 proteins with a glycine, alanine, tyrosine and to some degree a phenylalanine, replacing Ser-34 were functional for autophagy (Fig. 2B). A similar result was obtained with a second assay that uses the Ape1 aminopeptidase I as a reporter for autophagy activity. Ape1 is normally delivered to the vacuole via an autophagy-related pathway known as cytoplasm-to-vacuole targeting (Cvt).⁵⁷ The overexpression of Ape1 saturates the Cvt system and results in the accumulation of a precursor form of this protein in the cytoplasm (Fig. 2C).⁵⁸ Upon the induction of autophagy, this protein is delivered to the vacuole and proteolytically processed to its active form.^23,59 We found that the Atg1^S34A variant was also functional in this second assay (Fig. 2C). In contrast, the Atg1 variants containing the phosphomimetics at position 34, Atg1^S34D and Atg1^S34E, exhibited dramatically reduced levels of autophagy activity in both of these assays (Fig. 2B and C). These results were therefore consistent with the phosphorylation at Ser-34 in Atg1 being inhibitory for the autophagy process.

We also measured the kinase activity associated with these Atg1 variants with an in vivo assay that takes advantage of previous observations made with this enzyme. In particular, Atg1 has been found to migrate as a doublet on SDS-polyacrylamide gels and previous work indicates that the presence of the upper or slower-migrating, band is due to an autophosphorylation reaction.^29,37 The relative level of this upper band can therefore be used as a measure of Atg1 kinase activity in vivo.²⁹ As found previously, the relative amount of the wild-type Atg1 in the slower-migrating form increased dramatically upon rapamycin treatment and a similar result was observed here with the Atg1^S34A variant (Fig. 2D). In contrast, cells containing the Atg1^S34D and Atg1^S34E proteins lacked this upper band and resembled mutants known to be defective for kinase activity (Fig. 2D and E).^29,37 We have also shown that this slower-migrating form of Atg1 can be generated in vitro and that this conversion is dependent upon a functional Atg1 kinase domain.^28,29,37 Here, we tested whether these S34 alterations would have any effect upon this in vitro activity. The Atg1 variants were precipitated, treated with phosphatase and then incubated with unlabeled ATP. We found that the slower-migrating form of Atg1 was regenerated in vitro with both the wild-type Atg1 and the Atg1^S34A variant (Fig. 2E). However, this autophosphorylation reaction did not occur with either the Atg1^S34D or Atg1^S34E variant (Fig. 2E). These latter variants were localized normally to the PAS and interacted with Atg13 in a co-immunoprecipitation assay (Fig. S2A and B). In all, these results suggested that the phosphorylation at Ser-34 specifically inhibited the kinase activity associated with Atg1. Future work will be directed at identifying the protein kinase responsible for this phosphorylation.

Analysis of phosphorylation site mutants.

The MS analysis here identified 12 additional sites of phosphorylation that have not yet been analyzed. A number of these sites and additional positions of phosphorylation were also identified in a recent MS/MS analysis of Atg1.⁶⁰ Here, we tested the consequences of replacing each of these 12 positions with the nonphosphorylatable residue, alanine, for both the induction of autophagy and Atg1 kinase activity. We also examined several variants with multiple alterations that typically involved residues that were nearby in the linear sequence. However, one particular variant had alterations at three sites, S621A/S677A/S683A, that shared a common sequence around the identified site of phosphorylation: LS₆₂₁TT, LS₆₇₇AT and LS₆₈₃AT. Using the Pho8Δ60-based assay for autophagy, we found that four of the variants exhibited 40% less activity than the wild type and an additional six mutants displayed 25% to 30% less activity (Fig. 3A). The four variants with the strongest defects were the single mutants S552A and S677A, the double mutant, S551A/S552A and the triple mutant, S621A/S677A/S683A. We also assessed Atg1 kinase activity with both in vivo and in vitro assays. We found that all of the altered proteins exhibited autophosphorylation activity in vivo based on the presence of the slower-migrating form of Atg1 in rapamycin-treated cells (see representative images in Fig. 3B). All of the variants tested also exhibited kinase activity in an in vitro autophosphorylation assay (Fig. 3C). The most significant defects were observed with the Atg1^S621A variant that had only 50% the activity of the wild-type and the Atg1^S680A variant that reproducibly exhibited more activity than the wild-type protein (Fig. 3C). This latter result was interesting in light of the autophagy data above where the presence of the S680A alteration led to increased activity in a protein that already possessed the two changes, S677A and S683A (Fig. 3A).

Figure 3.

An analysis of the autophagy and kinase activities associated with alterations of the Atg1 phosphorylation sites identified in this study. (A) Autophagy activity as measured with the Pho8Δ60 alkaline phosphatase-based assay. The relative levels of autophagy present in an atg1Δ strain expressing the indicated Atg1 protein variants are shown. Autophagy was induced by treating log phase cultures with 200 ng/ml rapamycin for 4 h. The values are shown as a percentage of the activity observed with the wild-type Atg1. The data shown are the average of at least three independent experiments and the error bars indicate one standard deviation. (B) An assessment of Atg1 autophosphorylation levels in vivo. The relative levels of the autophosphorylated form (Atg1-P) of each of the indicated Atg1 variants was assessed by western blotting. The extracts were prepared from either log phase (L) or rapamycin-treated (R) cells. (C) Atg1 autophosphorylation in vitro. The indicated Atg1 proteins were immunoprecipitated from rapamycin-treated cells and subjected to an in vitro autophosphorylation assay with [γ-³²P] ATP, as described in the Materials and Methods. The reaction products were separated on an SDS-polyacrylamide gel and the relative amount of radioactivity incorporated into Atg1 was assessed with a phosphorimager. The data presented represent the average of at least two independent experiments where the phosphorylation levels varied by less than 15% between the different experimental replicates.

In summary, about half of the single alterations reduced autophagy by 25% to 40% and therefore these sites may be important for the normal control of this degradative response in vivo. In addition, there might be some level of cooperation amongst these phosphorylation sites that we have not yet uncovered. The appropriate combination of alterations, perhaps including sites that were not identified here, could result in a stronger defect in the autophagy process. Finally, it is important to point out that we have only looked at macroautophagy here and that some of these alterations might have a greater effect upon other types of autophagy processes.

Autophosphorylation at Ser-390 is responsible for the anomalous migration of Atg1 on SDS-polyacrylamide gels.

The presence of the slower-migrating form of Atg1 in SDS-polyacrylamide gels requires a functional Atg1 kinase domain and is lost upon phosphatase treatment.^29,37 These and other results suggested that this anomalous migration was due to a specific autophosphorylation event. However, the phosphorylation site responsible for this shift has not yet been identified. All of the alanine variants analyzed above exhibited the characteristic doublet for Atg1 when analyzed by gel electrophoresis. Therefore, we used a combined deletion and mutagenesis approach here to identify this site of autophosphorylation. Since the kinase domain is located at the N-terminus of Atg1, we examined the autophosphorylation of Atg1 fragments with increasingly larger deletions from the C-terminus and found that the N-terminal 420 (Atg1_1-420) or 450 (Atg1_1-450) residues were sufficient for autophosphorylation in vitro (Fig. 4A). The Atg1 kinase domain is known to contain at least one site of autophosphorylation at Thr-226 within the activation loop.³⁷ To test whether the observed incorporation of radioactivity into the Atg1_1-420 fragment was due solely to this site, we took advantage of previous observations indicating that an Atg1 variant with a phosphomimetic at position 226, such as Atg1^T226E, was able to undergo an autophosphorylation reaction in vitro.³⁷ Therefore, we asked whether the Atg1_1-420 fragment would still incorporate radioactivity when Thr-226 was replaced with a glutamic acid. Interestingly, we found that the T226E version of this fragment was functional in this in vitro assay indicating that there was a second site of autophosphorylation within the N-terminal 420 amino acids of Atg1 (Fig. 4B).

Figure 4.

Autophosphorylation at Ser-390 is responsible for the slower-migrating form of Atg1 observed on SDS-polyacrylamide gels. (A) An in vitro autophosphorylation assay with N-terminal fragments of Atg1. The indicated Atg1 deletion constructs were immunoprecipitated from yeast cell extracts, incubated with [γ-³²P] ATP and then separated on SDS-polyacrylamide gels. The relative amount of radioactivity incorporated into each Atg1 fragment was assessed by autoradiography. The lower part shows a western blot depicting the relative amount of each construct present in the kinase reactions. The positions of the relevant protein bands are indicated with an asterisk. (B) The identification of a second site of autophosphorylation within the N-terminal 420 amino acids of Atg1. Atg1_1-420 fragments with either a threonine (Wt) or glutamic acid (T226E) at position 226 were precipitated from cell extracts and then subjected to an in vitro autophosphorylation assay with [γ-³²P] ATP. The relative amount of radioactivity incorporated into each fragment was determined by autoradiography. The lower part shows a western blot depicting the relative amount of each construct present in the kinase reactions. (C) Alterations of Ser-390 influence Atg1 mobility in an SDS-polyacrylamide gel. Western blots were performed with extracts prepared from either log phase (L) or rapamycin-treated (R) cells expressing the indicated variants of Atg1. D211A, a kinase-defective Atg1. (D) The gel mobility of the Atg1^S390A and Atg1^S390D variants were not affected by either a phosphatase treatment or autophosphorylation reaction. The indicated Atg1 proteins were precipitated and then either mock treated (IP) or incubated with λ phosphatase (PPase). An aliquot of the phosphatase-treated protein was then subjected to an in vitro autophosphorylation reaction (IVKA) with 2.5 mM ATP, as described in the Materials and Methods. The resulting reaction products were separated by SDS-polyacrylamide gel electrophoresis and the relative mobility on SDS-polyacrylamide gels was assessed by western blotting. The (A and D) designations indicate the Atg1^S390A and Atg1^S390D proteins, respectively.

To map this site, we set out to systematically replace the serine and threonine residues present within this fragment with an alanine. This procedure was initiated from position 420 and moved progressively towards the N-terminus of the protein. We found that the replacement of Ser-390 in the full-length protein, but none of the other residues tested, resulted in the loss of the slower migrating form of Atg1 (Fig. 4C and D). In addition, replacing this residue with the potential phosphomimetics, aspartic or glutamic acid, resulted in an Atg1 protein that migrated only as the upper band on SDS-polyacrylamide gels (Fig. 4C and D). These data therefore suggested that Ser-390 might be the site of phosphorylation responsible for the band shift observed with active Atg1. However, the absence of the slower-migrating form of the Atg1^S390A variant could also have been due to a loss of kinase activity. To test this possibility, we carried out in vitro autophosphorylation assays with the Atg1^S390A and Atg1^S390D variants. These variants exhibited similar levels of kinase activity that was somewhat less than that seen with the wild-type protein (Fig. 5A). This decrease was not unexpected as these variants would be predicted to have one less site of autophosphorylation. As a second measure of kinase activity, we asked whether these S390 variants were capable of autophosphorylation at Thr-226. For this analysis, we used an antibody that specifically recognizes the phosphorylated form of this position and found that each of these variants was indeed phosphorylated at Thr-226 (Fig. 5B and Fig. S3).³⁷ In all, these data suggested that the phosphorylation at Ser-390 is both necessary and sufficient for the presence of the slower-migrating form of Atg1 detected on SDS-polyacrylamide gels. Although our MS analysis here did not identify Ser-390 as a site of phosphorylation, this modification was detected in a previous large-scale analysis of the yeast phosphoproteome (Table 1).⁴⁸

Figure 5.

Protein kinase and autophagy activities associated with alterations of Ser-390 in Atg1. (A) Atg1 autophosphorylation in vitro. The indicated Atg1 proteins were immunoprecipitated from rapamycin-treated cells and subjected to an in vitro autophosphorylation assay with [γ-³²P] ATP, as described in the Materials and Methods. The reaction products were separated on an SDS-polyacrylamide gel and the relative amount of radioactivity incorporated into Atg1 was assessed with a phosphorimager. The data presented represent the average of at least two independent experiments where the phosphorylation levels varied by less than 10% between the different experimental replicates. (B) Assessing the levels of Thr-226 phosphorylation in vivo. The indicated Atg1 proteins were immunoprecipitated from rapamycin-treated cells, separated on SDS-polyacrylamide gels, and then analyzed by western blotting with the indicated antibodies. The protein signals were detected with a LI-COR Odyssey infrared detection system. The relative signal detected with the anti-pT226 phosphospecific antibody (α-pT226) is a measure of the level of autophosphorylation at Thr-226. The (A, D and E) designations indicate the Atg1^S390A, Atg1^S390D and Atg1^S390E proteins, respectively. (C–E) Autophagy activity associated with atg1Δ cells expressing the indicated Atg1 proteins was assessed with the Pho8Δ60 processing (in C), Ape1 maturation (in D) and GFP-Atg8 processing (in E) assays, as described in the Materials and Methods. Autophagy was induced by treating log phase cells with 200 ng/ml rapamycin for 4 h (in C) or the times indicated (in D and E).

Finally, we tested whether these alterations at Ser-390 affected the autophagy process. For these experiments, we used the Pho8Δ60 and Ape1 assays described above as well as a third assay that assesses the processing of an Atg8-GFP fusion protein. This fusion is delivered to the vacuole via the autophagy pathway and the Atg8 moiety is then degraded by the hydrolases present within this compartment.⁶¹ GFP is more resistant to this degradation and the relative level of the free GFP that accumulates has been used as a measure of autophagy activity.⁶¹ We found that cells containing either the Atg1^S390A or Atg1^S390D variant exhibited autophagy activity in each of these three assays (Fig. 5C–E). These data therefore suggested that the phosphorylation at Ser-390 does not significantly influence the macroautophagy process induced by rapamycin treatment. However, as with the other sites identified here, it will be important to assess whether this modification has a role in other autophagy-related processes.

Conclusions

In summary, this study identified 15 sites of phosphorylation on the S. cerevisiae Atg1 protein, nine of which had not been identified previously. A functional analysis suggested that the phosphorylation at one of these sites, Ser-34 in the conserved Gly-rich loop of the kinase domain, is inhibitory for Atg1 kinase activity and the autophagy process. Future work will be directed at identifying the protein kinase responsible for this modification and the physiological significance of this inhibition. Alterations at many of the other sites identified here resulted in diminished autophagy activity in vivo suggesting that these phosphorylation events could also be important for the proper control of the macroautophagy process. In all, the phosphorylation sites identified represent potential avenues of regulation that will need to be explored in future work.

Materials and Methods

Strains and growth media.

Standard E. coli growth conditions and media were used in this study. The yeast rich growth medium, YPAD, consists of 2% glucose, 1% yeast extract, 2% Bacto-peptone and 500 mg/L adenine-HCl. The SC-glucose minimum growth medium has been described in reference ⁶² and ⁶³. The yeast strains used were TN125 (MATa ade2 his3 leu2 lys2 trp1 ura3 pho8::pho8Δ60), YYK126 (TN125 atg1Δ::LEU2) and BY4741 (MATa his3Δ1 leu2Δ0 ura3Δ0 met15Δ0).^24,56,64

Plasmid construction.

Site-directed mutagenesis was performed with a gap-repair strategy.⁶⁵ The ATG1-containing plasmids, pPHY115 and pPHY2376, were digested with the restriction enzyme PshAI, and then co-transformed into yeast cells with PCR products containing the desired mutations and flanking sequence complimentary to the gapped plasmid. The resulting plasmids were extracted and screened for the desired mutations. Plasmid pPHY1115 was originally named pRS316-3xmycATG1 and was generously provided by Dr. Yoshinori Ohsumi. Plasmid pPHY2376 is a pRS423-based construct where Atg1 has three copies of the HA epitope at its N-terminus; this plasmid was originally provided by Dr. Daniel Klionsky. The HA-tagged N-terminal Atg1 fragments were generated by a PCR reaction using pPHY2376 plasmid DNA as template and were cloned into the pRS423 vector, containing the ATG1 promoter. The GST-Atg1 fusion protein was overexpressed from the GAL1/10 promoter in a plasmid that was generously provided by Dr. Michael Snyder.

Isolation of the GST-tagged Atg1 protein for the MS analysis.

Yeast cells were grown to mid-log phase in an SC medium containing 2% raffinose and expression of the GST-Atg1 fusion protein was induced by transferring the cells to a medium containing 5% galactose for 4 h.⁶⁶ The expression of this fusion protein was under the control of the GAL1 promoter. The GST-tagged Atg1 was isolated as described with the following modifications.^28,66 Briefly, the cells were collected by centrifugation and lysed by agitation with glass beads in Lysis buffer (50 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EGTA, 0.1% TritonX-100, 0.1% β-mercaptoethanol, 0.5 mM PMSF and protease inhibitors). The GST-Atg1 protein was then incubated with 100 µg of anti-GST antibody (Cell Signaling, 2622) at 4°C overnight and the immune complexes were collected on Protein A-Sepharose beads (GE Healthcare, 17-0780-01) for 1 h. These beads were washed five times each with Wash buffer (50 mM Tris-HCl, pH 7.5, 500 mM NaCl, 1 mM EGTA, 0.1% TritonX-100, 0.1% β-mercaptoethanol, 0.5 mM PMSF and protease inhibitors) and TBS. The bound proteins were then eluted, separated on an SDS-polyacrylamide gel and visualized by staining with Coomassie G-250 (Bio-Rad, 161-0786). The corresponding GST-Atg1 band was excised from this gel and used for the mass spectrometry analysis.

In-gel digestion and TiO₂ enrichment of phosphopeptides for MS analysis.

The GST-Atg1 band was excised, reduced, alkylated with iodoacetamide, in-gel digested with sequencing grade trypsin (Promega, V5111), and then extracted from the gel, as described in reference ⁶⁷. To enrich for phosphopeptides, the tryptic peptides were re-dissolved in 20 µl DHB solution (200 mg 2,5-dihydroxybenzoic acid (Sigma, G-5254) in 1 ml of 80% acetonitrile (Merck 100003), 5% trifluoroacetic acid (Riedel-de Haën, 61030)) and loaded onto a TiO₂ microcolumn (GL Sciences Inc., 5020-08520-5u-TiO₂), as described in reference ⁶⁸. The columns were washed three times with 20 µl DHB solution and five times with 20 µl of a solution of 5% trifluoroacetic acid and 80% acetonitrile to remove non-specific binding and DHB. The bound peptides were then eluted with 3 × 20 µl of 0.3 N NH₄OH (pH > 10.5) and evaporated to dryness in a SpeedVac for further MS analysis.

MS and MS/MS analyses.

NanoLC-MS/MS analyses were performed on either an API-Q-TOF Ultima™ mass spectrometer (Waters/Micromass) coupled with a CapLC system (Waters/Micromass) or a QSTAR XL (Applied Biosystems/MDS Sciex) mass spectrometer fitted with an Agilent 1100 LC system (Agilent Technologies). The TiO₂ enriched peptides were dissolved in 5 µl of 0.1% formic acid and subsequently loaded onto an in-house packed C18 trap column (1.5 cm, 360 µm outer diameter, 150 µm inner diameter, Nucleosil 120-5 C18) at a flow rate 10 µl/min. The retained peptides were then eluted and separated by an analytical C18 capillary column (25 cm × 75 µm id) packed with the same material at a flow rate 300 nL/min. The nanoLC gradient was first held at 100% solvent A for 4 min, then raised with linear gradients from 0–10% solvent B in 0.5 min, from 10–40% solvent B in 25.5 min, from 40–80% solvent B in 0.5 min, and then held at 80% B for 9.5 min. Solvent A and B were 0.1% formic acid and 95% ACN in 0.1% formic acid, respectively.

Data-dependent acquisition on API-Q-TOF was performed under the software control of MassLynx™ 4.0. One sec MS survey scans were followed by MS/MS acquisitions on the three most abundant ions detected above the intensity threshold set at 15 counts/sec. Charge-state recognition was used to exclude singly charged precursor ions, and the selected ions were dynamically excluded for 120 sec. Each MS/MS acquisition was terminated after 4 sec or when the intensity of the precursor fell below 3 counts/sec. The resulting MS and MS/MS spectra were processed by ProteinLynx™ Global Server (PGS) 2.0 (Micromass/Water) software to generate a single Mascot-searchable peak list (.pkl) file for protein database search. Information-dependent acquisition (IDA) analyses on the QSTAR mass spectrometer and subsequent data processing were performed using the Analyst QS 1.1 software. One sec MS survey scans were followed by 2 sec MS/MS scans on the three most abundant multiply charged ions, with dynamic exclusion set at 120 s.

All MS/MS data generated were searched against the NCBInr database (20080327/6350093 entries) with the Mascot search engine (version 2.2.0.7, Matrix Science) limited to the following criteria: peptide mass tolerance, 50 ppm; MS/MS ion mass tolerance, 0.25 Da; allowed up to one missed cleavage; considered variable modifications of serine, threonine and tyrosine phosphorylation, methionine oxidation and cysteine carboxy-amidomethylation. All identified phosphopeptides and inferred phosphorylation sites were additionally verified and annotated manually (Fig. S1). Taken as an example, manual inspection of the MS/MS spectrum acquired on the triply charged precursor ion for peptide ²⁹EIG KGS FAT VYR⁴⁰ at m/z 469.9 (Fig. S2A) showed that the mono-phosphorylated peptide was indeed phosphorylated exclusively at Ser-34 and not Thr-37. The y₇ (pSFATVYR), y₈ (GpSFATVYR) and y₁₀ (GKGpSFATVYR) ions were detected exclusively as (y₇-98), (y₈-98) and (y₁₀-98) ions at m/z 825.38 (1+), 882.36 (2+) and 534.27 (2+), respectively. In contrast, the y₄ and y₅ ions harboring the Thr-37 site were detected without any phosphate or the loss of it at m/z 538.28 (1+) and 609.31 (1+), respectively.

Autophagy assays.

Autophagy activity was analyzed with three independent assays that have been described previously in reference ²⁸ and ³⁷. In general, autophagy was induced by treating mid-log phase cells with 200 ng/ml rapamycin for the indicated times. The alkaline phosphatase-based assay measures the delivery and subsequent activation in the vacuole of a modified form of the Pho8 alkaline phosphatase, known as Pho8Δ60.^55,56 The level of autophagy activity was determined by the difference in alkaline phosphatase activity detected between extracts prepared from log phase and rapamycin-treated cells. The data presented here were the average of at least three independent experiments. The Ape1 maturation assay assesses the processing of overexpressed aminopeptidase I or Ape1, by the autophagy pathway.²³ The elevated levels of Ape1 saturate the capacity of the Cvt system and result in the accumulation of the unprocessed precursor in the cytoplasm.⁵⁸ The relative levels of the two forms of Ape1 were assessed by western blotting with an anti-Ape1 antibody provided by Dr. Daniel Klionsky. In the GFP-Atg8 assay, the relative level of free GFP generated by proteolysis in the vacuole is used to indicate autophagy activity.^23,59 The GFP-Atg8 fusion protein and the free GFP were detected by western blotting with an anti-GFP antibody (Clontech 632381).

Western blotting and immunoprecipitation.

Protein samples were prepared as described in reference ²⁹. The proteins were separated on SDS-polyacrylamide gels, transferred to a nitrocellulose membrane and then probed with the appropriate primary and secondary antibodies (GE Healthcare, NA9310-1ML). The primary antibodies used were the anti-HA (Roche 11867423001), anti-myc (Cell Signaling 2276), anti-GFP (Clontech 632375) and anti-Ape1 antibodies (Santa Cruz Biotechnology, Inc., sc-26740). A chemiluminescent substrate (Thermal Fisher Scientific Inc., 34095) or LI-COR Biosciences Odyssey infrared imaging system was then used to detect the reactive bands or fluorescence intensities. Immunoprecipitations were performed as described in reference ²⁸ and ⁶⁹. HA-tagged proteins were precipitated on an anti-HA affinity matrix (Roche 11815016001) whereas myc-tagged proteins were immunoprecipitated with a monoclonal anti-myc antibody and subsequently collected on Protein A-Sepharose beads. The generation and characterization of the anti-pT226 antiserum (Lampire Biological Laboratories) has been described in reference ³⁷. Briefly, this antiserum was generated in rabbits immunized with a phosphorylated peptide, FLP NTS LAE (pT) LCG SPL Y, that corresponds to the sequence of the Atg1 activation loop.

In vitro kinase assays (IVKAs).

After immunoprecipitation, Atg1 proteins were bound to beads and incubated with 10 µCi [γ-³²P] ATP or 2.5 mM cold ATP for 30 min in kinase buffer (25 mM MOPS, 1 mM EGTA, 100 µM Na₃VO₄, 15 mM MgCl₂ and 15 mM ρ-nitrophenylphosphate).²⁴ Because these assays were performed with Atg1 immunoprecipitates, the final activities could have been influenced by other proteins present in the samples. The reaction products were subsequently separated by SDS-polyacrylamide gel electrophoresis and the relative level of radioactivity incorporated into the Atg1 proteins was assessed with a Typhoon Trio phosphorimager. The input level of each protein was assessed by western blotting with the appropriate antibody.

Abbreviations

AMPK

AMP-activated protein kinase

Cvt

cytoplasm-to-vacuole targeting

MS

mass spectrometry

PAS

phagophore assembly site

PKA

cAMP-dependent protein kinase

prApe1

precursor Ape1

TOR

target of rapamycin