Protein N-terminal Acetylation by the NatA Complex Is Critical for Selective Mitochondrial Degradation*

Background: Mitophagy is a catabolic mechanism that degrades mitochondria selectively, and protein N-terminal acetylation is a major modification of eukaryotic proteins. Results: Deletion of protein N-terminal acetyltransferase A (NatA) leads to suppression of mitophagy in yeast. Conclusion: Protein N-terminal acetylation by NatA is crucial for mitophagy. Significance: This is the first report on the link between mitophagy and protein N-terminal acetylation. Mitophagy is an evolutionarily conserved autophagy pathway that selectively degrades mitochondria. Although it is well established that this degradation system contributes to mitochondrial quality and quantity control, mechanisms underlying mitophagy remain largely unknown. Here, we report that protein N-terminal acetyltransferase A (NatA), an enzymatic complex composed of the catalytic subunit Ard1 and the adaptor subunit Nat1, is crucial for mitophagy in yeast. NatA is associated with the ribosome via Nat1 and acetylates the second amino acid residues of nascent polypeptides. Mitophagy, but not bulk autophagy, is strongly suppressed in cells lacking Ard1, Nat1, or both proteins. In addition, loss of NatA enzymatic activity causes impairment of mitochondrial degradation, suggesting that protein N-terminal acetylation by NatA is important for mitophagy. Ard1 and Nat1 mutants exhibited defects in induction of Atg32, a protein essential for mitophagy, and formation of mitochondria-specific autophagosomes. Notably, overexpression of Atg32 partially recovered mitophagy in NatA-null cells, implying that this acetyltransferase participates in mitophagy at least in part via Atg32 induction. Together, our data implicate NatA-mediated protein modification as an early regulatory step crucial for efficient mitophagy.

Mitochondria are essential organelles that supply most of the energy for a cell. This organelle concomitantly generates reactive oxygen species during respiration. Accumulation of reactive oxygen species eventually causes mitochondrial dysfunction that negatively affects cellular integrity and is thought to induce diverse pathology (1). Additionally, cells need to adjust mitochondrial quantity to maintain a suitable balance between ATP production and consumption (2). Therefore, degradation of dysfunctional and excess mitochondria is critical for cell homeostasis. To solve this problem, cells utilize mitophagy, a catabolic system via autophagy that isolates cytosolic components with double-membrane vesicles called autophagosomes and carries them into digestive compartments such as lysosomes (vacuoles in yeast) for degradation and recycling (3,4). Mitophagy is a selective pathway that specifically eliminates mitochondria. This process is conserved from yeast to humans and is relevant to cellular physiology (5,6).
In the budding yeast Saccharomyces cerevisiae, the mitochondria-anchored receptor Atg32 is induced in response to oxidative stress and is localized on the surface of mitochondria (7,8). Loss of Atg32 disrupts mitophagy, ultimately leading to mitochondrial genome instability (9). Atg32 interacts with Atg8, a ubiquitin-like protein conjugated to the phospholipid phosphatidylethanolamine and required for autophagosome formation, and with Atg11, a selective autophagy-specific scaffold protein required for assembly of core Atg proteins (7,8). Moreover, phosphorylation of Atg32 stabilizes the Atg32-Atg11 interaction (10 -12). Although the ternary complex is important for formation of mitochondria-specific autophagosomes (mitophagosomes), how Atg32 induction and mitophagosome formation are regulated remains poorly understood.
We show here that formation of mitophagosome is blocked by loss of N-terminal acetyltransferase A (NatA) catalyzing transfer of an acetyl group from acetyl-coenzyme A (acetyl-CoA) to the ␣-amino group of the N-terminal amino acid residue (13,14). In addition, Atg32 induction is partially suppressed in cells lacking NatA. Our findings suggest that NatA regulates selective mitochondrial degradation at least in part via mitophagosome formation and Atg32 expression.

Experimental Procedures
Yeast Strains and Growth Conditions-Yeast strains and plasmids used in this study are described in Tables 1 and 2, respectively. Standard genetic and molecular biology methods were used for S. cerevisiae and Escherichia coli strains. Yeast cells were incubated in YPD medium (1% yeast extract, 2% peptone, and 2% dextrose), synthetic medium (0.17% yeast nitrogen base without amino acids and ammonium sulfate, 0.5% ammonium sulfate) with 0.5% casamino acids containing 2% dextrose (SDCA), or 0.1% dextrose plus 3% glycerol (SDGlyCA), supplemented with necessary amino acids. For mitophagy and pexophagy under respiratory conditions, cells grown to mid-log phase in SDCA were transferred to SDGlyCA and incubated at 30°C. For autophagy and mitophagy under starvation conditions, cells grown to mid-log phase in YPD were transferred to nitrogen-free medium (SD-N; 0.17% yeast nitrogen base without amino acids and ammonium sulfate, 2% dextrose) and incubated at 30°C. For starvation-induced pexophagy, cells grown overnight in YPD were transferred to oleate medium (0.17% yeast nitrogen base without amino acids ammonium sulfate, 0.5% ammonium sulfate, 1% casamino acids, 0.12% oleate (v/v), 0.2% Tween 40R (v/v), 0.1% yeast TABLE 1 Yeast strains used in this study extract) at a 1:10 dilution. Pexophagy was induced by transferring cells from oleate medium to SD-N.
Immunoblotting-Samples corresponding to 0.1 OD 600 units of cells were separated by SDS-PAGE followed by western blotting and immunodecoration. After treatment with enhanced chemiluminescence reagents, proteins were detected using a luminescent image analyzer (LAS-4000 mini; GE Healthcare). Quantification of the signals was performed using ImageQuant TL (GE Healthcare).
Quantitative RT-PCR-RNA was isolated by Master Pure Yeast RNA purification kit (Epicenter) from 2 OD 600 units of cells, according to the manufacturer's protocol. Equal amounts of total RNA were mixed with RNA-direct SYBR Green Real Time PCR Master Mix (TOYOBO) and the following primers: ATG32 forward 5Ј-TGTCACTGCAGCATACGAACAC and reverse 5Ј-CTGCTCAGTTGAAGAAGGAGATG; ACT1 forward 5Ј-TATCGTCGGTAGACCAAGACAC and reverse 5Ј-TCGTCCCAGTTGGTGACAATAC. The Applied Biosystems Step One Plus (Applied Biosystems) was used for quantitative RT-PCR analysis. For each gene, triplicate samples were calculated to average value and then normalized by averaged actin. The value in wild-type cells at the 0-h time point was set to 1. These numbers represent relative expression of ATG32 mRNA.

Mitophagy Is Markedly Suppressed in Cells Lacking NatA-
To elucidate the role of NatA in mitophagy, we observed transport of mitochondria to the vacuole, a lytic compartment in yeast. Mitochondria and vacuoles were visualized using a mitochondrial matrix-targeted GFP (mito-GFP) and mCherry fused at the C terminus of Vph1, a membrane-integrated subunit of the vacuolar ATPase (Vph1-mCherry), respectively, in cells lacking Prb1, a vacuole-localized serine protease. When cells are grown in nonfermentable glycerol medium (Gly), mitochondria become active in respiration and are eventually degraded by mitophagy (7). In vacuolar protease-deficient cells, mitochondria transported into the vacuole can be accumulated and observed as degradation intermediates. Cells grown in Gly for 48 h contained mitochondria that were overlapped with the vacuole in a manner dependent on Atg32 (Fig. 1A). We hardly detected mito-GFP signals colocalized with Vph1-mCherry patterns in cells lacking NatA subunits, the catalytic component Ard1 and ribosomal adaptor Nat1 (13,14), under the same conditions (Fig. 1A). These observations suggest that NatA is important for mitophagy.
We next quantified mitochondrial degradation using mitodihydrofolate reductase (DHFR) 3 -mCherry, a reporter located  prb1⌬, nat1⌬ prb1⌬, and ard1⌬ prb1⌬ cells grown to mid-log phase in dextrose medium (0 h) were shifted to respiration medium (Gly) for 48 h and observed using fluorescence microscopy. Cells lacking Prb1 are vacuolar protease-deficient and accumulate degradation intermediates within the vacuole. Scale bar, 5 m. DIC, differential interference contrast. B, mitochondria-targeted DHFR-mCherry-expressing (mito-DHFR-mCherry, depicted by arrow) wild-type, nat1⌬, ard1⌬, nat1⌬ ard1⌬, and atg32⌬ cells were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting. Generation of free mCherry (depicted by arrowhead) indicates transport of the marker to the vacuole. Pgk1 was monitored as a loading control. C, free mCherry in cells under respiratory conditions for 48 h in B was quantified in three experiments. Data represent the averages of all experiments, with bars indicating standard deviations. D, mitochondria-targeted DHFR-mCherry-expressing (mito-DHFR-mCherry, depicted by arrow) wild-type, nat3⌬, mak3⌬, nat4⌬, and nat5⌬ cells were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting as in B. E, mito-DHFR-mCherry-expressing wild-type, nat1⌬, ard1⌬, and atg32⌬ cells grown to mid-log phase in nutrient-rich medium were shifted to nitrogen starvation (ϪN), collected at the indicated time points, and subjected to western blotting as in B. F, free mCherry in cells under starvation conditions for 24 h in E was quantified in three experiments. Data represent the averages of all experiments, with bars indicating standard deviations. G, mito-DHFR-mCherry-expressing wild-type, nat1⌬, ard1⌬, and atg7⌬ cells were transformed with a plasmid encoding GFP-Atg8, grown to mid-log phase in rich dextrose medium, incubated for the indicated time points in starvation medium (ϪN), and subjected to western blotting. Generation of free GFP indicates progression of bulk autophagy. Atg7 is an E1 enzyme essential for autophagy. H, mito-DHFR-mCherry-expressing wild-type, nat1⌬, ard1⌬, and atg19⌬ cells were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting. Precursor form (p) of Ape1 is transported from the cytosol to the vacuole via the Cvt pathway, a selective autophagy-related process, and cleaved to be a mature form (m). Atg19 is an adaptor protein required for the Cvt pathway (35). I, Pot1-mCherry-expressing (a peroxisome marker) wild-type, nat1⌬, ard1⌬, and atg36⌬ cells were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting. Generation of free mCherry indicated transport of peroxisomes to the vacuole. J, free mCherry in cells under respiratory conditions for 48 h in I was quantified in three experiments. Data represent the averages of all experiments, with bars indicating standard deviations. K, Pot1-mCherry (a peroxisome marker) expressing wild-type, nat1⌬, ard1⌬, and atg36⌬ cells grown in oleic acid medium were shifted to nitrogen starvation (ϪN), collected at the indicated time points, and subjected to western blotting. Generation of free mCherry indicated transport of peroxisomes to the vacuole. Atg36 acts as an adaptor essential for pexophagy (16). L, free mCherry in cells under nitrogen starvation for 6 h in K was quantified in three experiments. Data represent the averages of all experiments, with bars indicating standard deviations.

NatA Is an Important Element in Mitophagy
in the matrix of mitochondria (11). Upon mitophagy, this fusion protein is processed to generate free mCherry that is appreciably protease-resistant, thereby indicating degradation of mitochondria. We found that accumulation of free mCherry was greatly decreased in cells lacking Nat1, Ard1, or both ( Fig.   1, B and C). By contrast, cells lacking the catalytic subunits of other N-terminal acetyltransferases, NatB-E (13,14), did not exhibit defects in degradation of mitochondria (Fig. 1D). We have previously demonstrated that mitophagy is induced during nitrogen starvation (15). Starvation-induced mitochondrial degradation was also suppressed in cells lacking Nat1 or Ard1 (Fig. 1, E and F). These results indicate that NatA participates in the regulation of mitophagy. We further investigated whether NatA affects the other autophagy-related pathways. First, we monitored processing of GFP-Atg8, an indicator of autophagy flux, and confirmed that free GFP was detected in cells lacking Nat1 or Ard1 under nitrogen deprivation (Fig. 1G). Therefore, it is likely that NatA is dispensable for starvation-induced autophagy. Next, we examined the cytoplasm-to-vacuole targeting (Cvt) pathway, a selective autophagy process that mediates transport of vacuolar enzymes such as Ape1, an amino peptidase, from the cytosol to the vacuole. Mature Ape1 transported to the vacuole was detectable in Nat1 or Ard1-null cells under respiratory conditions, indicating that NatA is dispensable for the Cvt pathway (Fig. 1H). Finally, we evaluated selective peroxisome autophagy (pexophagy). Degradation of peroxisomes was monitored using Pot1-mCherry, a marker localized in the peroxisomal matrix, under respiratory conditions that are also able to induce pexophagy (16). We found that accumulation of free mCherry was strongly depressed in cells lacking Nat1 or Ard1 (Fig. 1, I and J). Starvation-induced pexophagy was also inhibited in these mutant cells (Fig. 1, K and L). Together, our results raise the possibility that NatA is a common factor critical for selective elimination of organelles.
Enzymatic Activity of NatA Is Crucial for Mitophagy-To ask whether protein N-terminal acetylation by NatA is important for selective mitochondrial degradation, we investigated a NatA variant lacking enzymatic activity. Previous research unravels the structure of NatA in the fission yeast Schizosaccharomyces pombe, demonstrating that glutamine substitution for Glu-24 of Ard1 leads to loss of enzymatic activities but does not affect substrate binding capacity (17). This critical glutamate is conserved from budding yeast (Glu-26) to higher eukaryotes ( Fig. 2A). We thus monitored mitophagy using mito-DHFR-mCherry in cells expressing hemagglutinin (HA)-tagged Ard1 E26Q . Under respiratory conditions, mitochondrial degradation was inhibited to the levels similar to cells lacking Ard1 (Fig. 2, B and C). HA tagging of Nat1 and Ard1 did not significantly affect mitophagy (Fig. 2D). In addition, expression of Nat1 was not severely impaired in cells expressing the Ard1 E26Q mutant (Fig. 2E). Another previous study has reported that human Ard1 R82A , a variant containing a mutation in the acetyl-CoA binding domain, decreases enzymatic activities (18). Because this crucial arginine is also conserved among eukaryotes (Fig. 2F), we introduced a plasmid encoding HA-tagged Ard1 R125A (corresponding to human Ard1 R82A ) into ard1⌬ cells, and we monitored processing of mito-DHFR-mCherry under respiratory conditions. Cells expressing Ard1 R125A -3HA displayed a reduced accumulation of free mCherry (Fig. 2, G and H). Taken together, we conclude that protein N-terminal acetylation by NatA is critical for mitophagy in yeast.
Ribosomal Association of Ard1 Is Important for Efficient Mitophagy-Although Ard1 stably associates with the ribosome via Nat1 in yeast (19,20), it has also been reported that monomeric Ard1 purified from fission yeast is still catalytically active with altered substrate specificity in vitro (17) and that free Ard1 post-translationally acetylates ␤and ␥-actin in humans (21). To ask whether monomeric Ard1 dissociated from the ribosome contributes to regulation of mitophagy in S. cerevisiae, we monitored mitochondrial degradation in cells overexpressing Ard1 because the Ard1 levels are drastically reduced in nat1⌬ cells (Fig. 3A). We found that overexpression of Ard1 did not lead to recovery of mitophagy in the absence of the ribosome adaptor (Fig. 3, C and D). Hence, it seems likely that NatA-mediated cotranslational acetylation is crucial for mitophagy.
Mitophagosome Formation Is Impaired in Cells Lacking NatA-In which stage of mitophagy does NatA participate? It is noteworthy that the Nat1 and Ard1 protein levels declined during respiratory growth and became exceedingly low at the 48-h time point (Fig. 3, A and B). We thus speculated that NatA affects the early phase of mitophagy, in particular mitophagosome formation through which mitochondria are enclosed by the isolation membrane. To test this possibility, we used cells lacking Ypt7, a Rab family GTPase essential for homotypic vacuole fusion. Although autophagosomes are formed normally in this mutant, they do not fuse to the vacuole and accumulate in the cytosol (22). When ypt7⌬ cells grown in fermentable medium (Dex) are shifted to respiratory medium (Gly), their mitochondrial shape changes from tubules to fragments (Fig.  4A). Conversely, upon a shift from respiratory to fermentable  Atg32 induction is partially suppressed in nat1⌬ and ard1⌬ cells. A, Atg32-3HA-expressing wild-type, nat1⌬, and ard1⌬ cells were grown in glycerol medium (Gly), collected at the indicated time points, and subjected to western blotting. B, peaks of Atg32-3HA expression in A were quantified in three experiments (wild type, 18 h; nat1⌬, 24 h; ard1⌬, 24 h). Data represent the averages of all experiments, with bars indicating standard deviations. *, p Ͻ 0.01. C, mito-DHFR-mCherry-expressing wild-type, nat1⌬, and ard1⌬ cells were grown for the indicated time points in glycerol medium (Gly). ATG32 mRNA expression was analyzed by real time PCR and normalized to ACT1 mRNA expression. Data represent the averages of all experiments, with bars indicating standard deviations. D, Atg32-3HA-and mito-DHFR-mCherry-expressing wild-type, nat1⌬, and ard1⌬ cells transformed with empty vectors or two plasmids encoding Atg32-3HA were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting. Generation of free mCherry (depicted by arrowhead) indicates transport of mitochondria to the vacuole. E and F, free mCherry in nat1⌬ and ard1⌬ cells under respiratory conditions for 48 h in D was quantified in three experiments. Data represent the averages of all experiments, with bars indicating standard deviations. G, mito-DHFR-mCherry-expressing atg32⌬, nat1⌬ atg32⌬, and ard1⌬atg32⌬ cells were transformed with a plasmid encoding GFP-Atg32 with the constitutive TEF2 promoter, grown to log phase in glucose medium, and investigated using fluorescence microscopy. Scale bar, 5 m. H, pep4⌬ prb1⌬ and pep4⌬ prb1⌬ ard1⌬ cells expressing chromosome-or plasmid-encoded versions of Atg32 or Atg32-HA were grown in glycerol medium and subjected to coimmunoprecipitation using anti-HA antibody-conjugated agarose. Eluted immunoprecipitates (IP) and detergentsolubilized mitochondria-enriched fractions (input) were analyzed by western blotting. I, mito-DHFR-mCherry-expressing wild-type and atg32⌬ cells transformed with an empty vector or a plasmid encoding Atg32-3HA or Atg32 V2P (V2P)-3HA were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting. Generation of free mCherry (depicted by arrowhead) indicates transport of mitochondria to the vacuole.
conditions, mitochondrial tubular networks are reformed by their fusion, but mitophagosomes remain isolated from intact mitochondria. Accordingly, we can detect mitophagosomes as dot-like structures (7). We found mitochondrial dots in cells grown under fermentable conditions for 3 h after respiratory growth (Fig. 4B). By contrast, cells containing mitochondrial dots were few in the absence of Nat1 or Ard1 (Fig. 4C). Hence, these observations suggest that NatA plays a key role in the early step of mitophagosome formation.
Atg32 Expression Levels Are Low in nat1⌬ and ard1⌬ Cells-Based on the previous findings that Atg32 acts as a mitophagy receptor to generate mitophagosomes (7,8), we investigated whether NatA affects Atg32 functions. At first, to examine Atg32 expression in NatA-null cells under respiratory conditions, HA-tagged Atg32 was expressed from the endogenous locus and monitored by western blotting. In light of previous reports that the Atg32 levels are transiently elevated in respiring cells at mid-log phase (7), and that growth of cells lacking NatA is slow in nonfermentable medium (13), we prepared samples from nat1⌬ and ard1⌬ cells at later time points than wild-type cells and compared their highest amounts of Atg32. Cells lacking Nat1 or Ard1 exhibited a partial reduction in the Atg32 protein levels (Fig. 5, A and B). Additionally, transcriptional levels of ATG32 mRNA in these cells were quantified by real time PCR. We found that ATG32 mRNA expressions in nat1⌬ and ard1⌬ cells were less than half of those in wild-type cells during respiratory growth (Fig. 5C), raising the possibility that NatA may regulate transcription of ATG32.
To ask whether these reductions of Atg32 expression levels impact mitochondrial degradation, we introduced two lowcopy plasmids encoding Atg32-3HA into NatA-null cells and monitored processing of mito-DHFR-mCherry. When grown in nonfermentable medium, Nat1 or Ard1-null cells overexpressing Atg32 displayed partial recovery of mitochondrial degradation (Fig. 5, D-F). These data are consistent with the idea that partial impairment of Atg32 induction may be one of the factors for severe mitophagy defects in cells lacking NatA.
We also confirmed Atg32 localization to mitochondria and interaction with Atg8 and Atg11 in the early stage of mitophagy (7,11). Fluorescence live cell imaging and coimmunoprecipitation assays were performed using GFP-and HA-tagged Atg32, respectively. GFP-Atg32 was normally localized to mitochondria in cells lacking Nat1 or Ard1 (Fig. 3G), and Atg32-3HA coprecipitated with Atg11 and Atg8 in the absence of Ard1 (Fig.  3H). It is likely that NatA does not regulate organelle targeting and protein-protein interactions of Atg32.
In yeast, NatA acetylates the second amino acids, Ser, Ala, Thr, Val, Gly, and Cys, of nascent polypeptides whose N-terminal methionine residues are removed by methionine peptidase (23)(24)(25). Atg32 contains a valine at the second amino acid position, raising the possibility that Atg32 is acetylated by NatA. We thus substituted proline for Val-2 of Atg32, introduced a plasmid encoding Atg32 V2P -3HA into Atg32-null cells, and monitored processing of mito-DHFR-mCherry under respiratory conditions. A protein containing proline at the second amino acid position is not acetylated by all N-terminal acetyltransferases in accordance with the XPX rule (26). We found that mitophagy is not altered in cells expressing Atg32 V2P -3HA (Fig.  3I). Thus, NatA is unlikely to directly control Atg32.
Atg8 and Atg11 Are Expressed Normally in the Absence of NatA-Finally, we examined whether loss of NatA affects Atg8 and Atg11, two Atg32-interacting proteins essential for mitophagy, under respiratory conditions. The protein levels of Atg8 and Atg11 were not decreased in cells lacking Nat1 or Ard1 (Fig. 6, A and B). Moreover, the phospholipid conjugation of Atg8 occurred properly in NatA-null cells (Fig. 6A). These data are in agreement with the results that NatA is dispensable for the Cvt pathway under respiratory conditions (Fig. 1H). Because Atg11 has alanine at the second amino acid position that can be acetylated by NatA, we introduced a plasmid encoding Atg11 A2P into cells lacking Atg11 and monitored mitophagy under respiratory conditions. Cells expressing Atg11 A2P did not exhibit significant changes in mitochondrial degradation (Fig. 6C). Therefore, it seems likely that NatA acts in mitophagy irrespectively of Atg8 and Atg11.

Discussion
In this study, we demonstrated that NatA plays a critical role in selective degradation of mitochondria in yeast. This catabolic process requires NatA catalytic activity and association with the ribosome, indicating that nascent polypeptide N-terminal acetylation of NatA target(s) is a key step for activation of mitophagy. Notably, mitophagosome formation is compromised in nat1⌬ and ard1⌬ cells, suggesting that NatA partici- anti-Atg11 C FIGURE 6. Atg8 and Atg11 are not significantly affected in cells lacking NatA. A, wild-type, nat1⌬ ard1⌬, and atg7⌬ cells were grown in glycerol medium (Gly), collected at the indicated time points, and subjected to western blotting. B, wild-type, nat1⌬ ard1⌬, and atg11⌬ cells were grown in glycerol medium (Gly), collected at the indicated time points, and subjected to western blotting. C, mito-DHFR-mCherry (depicted by arrow)-expressing wild-type and atg11⌬ cells transformed with an empty vector or a plasmid encoding Atg11 or Atg11 A2P (A2P) were grown for the indicated time points in glycerol medium (Gly) and subjected to western blotting. Generation of free mCherry (depicted by arrowhead) indicates transport of mitochondria to the vacuole.
pates in the early phase of mitophagy. In addition, Atg32 induction is partially suppressed in the absence of Nat1 or Ard1. In particular, the transcriptional levels of ATG32 are significantly reduced. When we up-regulate Atg32 expression in these mutant cells, mitochondrial degradation is not fully recovered to a wild-type level. We thus speculate that NatA regulates mitophagy through Atg32 and other unknown factors (but not Atg8 and Atg11) (Fig. 7).
Recently, the Ume6-Sin3-Rpd3 complex, a transcriptional regulator, has been reported to suppress Atg32 expression (27). It remains possible that NatA might acetylate Sin3 and Rpd3 containing a serine and a valine at the second amino acid position, respectively. As N-terminally acetylated residues act as degradation signals (28), Sin3, Rpd3, or both could become stabilized, leading to reduction of Atg32 levels in NatA-null cells.
Bulk autophagy normally occurs in NatA-null cells, indicating that NatA does not act on the core autophagy machinery. Moreover, NatA is unlikely to serve as a common regulator of selective autophagy because it is dispensable for the Cvt pathway. Interestingly, pexophagy is suppressed in NatA-null cells, raising the possibility that NatA regulates selective organelle autophagy via a common mechanism. It has recently been reported that the MAPK signaling pathway, Wsc1-Pkc1-Bck1-Mkk1/2-Slt2, is important for mitophagy and pexophagy in yeast (29). However, these proteins acting in the MAPK pathway have so far not been identified as NatA targets (23). Alternatively, it remains possible that NatA affects distinct molecules specific for either mitophagy or pexophagy.
Does NatA-mediated protein modification affect mitochondrial shaping during mitophagy? Fragmentation is a crucial step for mitochondria to be efficiently surrounded by the isolation membrane in mammalian cells (30 -32). In yeast, it has been reported that Atg11 interacts with Dnm1, a dynamin-related GTPase required for mitochondrial fission, and that mitophagy is suppressed in dnm1⌬ cells (33). However, mitochondrial fragmentation is almost normal in NatA-null cells under respiratory conditions (Figs. 1A and 4B). It is therefore unlikely that altered mitochondrial dynamics causes impairment of mitophagy in the absence of NatA. Because more than half of all proteins are acetylated by N-terminal acetyltransferases in yeast, NatA acetylates a lot of mitochondrial proteins, including import receptor, ATP synthase subunits, and ribosomal proteins (23). We thus do not exclude the possibility that changes in multiple mitochondrial functions could lead to mitophagy defects in NatA-null cells. Moreover, a recent study has reported that ERMES, a protein complex tethering the endoplasmic reticulum to mitochondria, is crucial for mitophagosome formation (34). The endoplasmic reticulum-mitochondrial tethering does not seem to be completely disrupted in NatA-null cells, because mitochondrial morphology defects, typical phenotypes of the ERMES mutants, are not seen in cells lacking Ard1 or Nat1 (Fig. 4B).
To our knowledge, this is the first report that protein N-terminal acetylation is linked to mitophagy. To clarify NatA-mediated activation of mitochondrial degradation in more detail, further studies are needed to identify NatA substrate(s) involved in mitophagy, to understand Atg32 induction mechanisms under respiratory conditions, and to investigate whether mitochondrial functions alter in NatA-null cells. These future approaches will elucidate regulation of mitophagy via protein N-terminal acetylation.
Author Contributions-K. O. designed the experiments, conceived and coordinated the study, and wrote the paper. A. E. designed, performed, and analyzed the experiments and wrote the paper. All authors reviewed the results and approved the final version of the manuscript.