Uth1p is involved in the autophagic degradation of mitochondria.

The absence of the outer mitochondrial membrane protein Uth1p was found to induce resistance to rapamycin treatment and starvation, two conditions that induce the autophagic process. Biochemical studies showed the onset of a fully active autophagic activity both in wild-type and Deltauth1 strains. On the other hand, the disorganization of the mitochondrial network induced by rapamycin treatment or 15 h of nitrogen starvation was followed in cells expressing mitochondria-targeted green fluorescent protein; a rapid colocalization of green fluorescent protein fluorescence with vacuole-selective FM4-64 labeling was observed in the wild-type but not in the Deltauth1 strain. Degradation of mitochondrial proteins, followed by Western blot analysis, did not occur in mutant strains carrying null mutations of the vacuolar protease Pep4p, the autophagy-specific protein Atg5p, and Uth1p. These data show that, although the autophagic machinery was fully functional in the absence of Uth1p, this protein is involved in the autophagic degradation of mitochondria.

The major cellular degradation pathways involved in protein and organelle turnover are autophagy and proteasome-mediated proteolysis. These processes are important for maintaining a controlled balance between anabolism and catabolism in order to have normal cell growth and development. These degradation pathways permit the cell to eliminate unwanted or unnecessary organelles and recycle the components for reuse. In eukaryotic cells, the lysosomes or the vacuole are major degradative organelles that contain a range of hydrolases able of degrading all the cellular constituents. During the last decade, autophagy has emerged as a crucial membrane trafficking process that transports bulk cytoplasm and sometimes entire organelles to the lysosome/vacuole for recycling in response to nutrient starvation or under specific physiological conditions (see Refs. 1 and 2 for reviews). Autophagy has two major forms, microautophagy and macroautophagy. Microautophagy operates by protruding or invaginating a portion of the vacuolar membrane to engulf cytosol or organelles. Only limited knowledge is available about microautophagy, which has been best characterized for the degradation of peroxisomes (3,4) and selective portions of the nucleus (5). Macroautophagy, of which the molecular aspects are better characterized, involves the non-selective sequestration of large portions of the cytoplasm into double membrane structures termed autophagosomes and their delivery to the vacuole for degradation.
Although the normal function of autophagy is thought to provide amino acids to starved cells, a fair amount of evidence suggests that it could also be required for the elimination of selected organelles, namely mitochondria, under peculiar conditions. In mammalian cells, autophagic removal of mitochondria has been shown to be triggered following a process of induction/blockade of apoptosis (6,7). Also, coregulation of apoptosis and autophagy by the Akt/protein kinase B-signaling pathway has been evidenced (8,9). Finally, molecular data suggest that autophagy-specific proteins might also be involved in the regulation of apoptosis (10). Because mitochondrial events are now regarded as a crucial step in apoptosis, this idea indicates that damaged mitochondria (following apoptotic induction) could control their own degradation by an autophagic process (11). This possibility is also supported by the observation that the overexpression of anti-apoptotic proteins, such as Bcl-2, is able to block autophagy (12) and that down-regulation of Bcl-2 induces autophagy (13). Also, the opening of the mitochondrial permeability transition pore, often considered to be a crucial event regulating the initial steps of apoptosis, induces the colocalization of mitochondrial and lysosomal markers, which is typical of an autophagic process in mammalian cells (14).
In yeast, a study reported the presence of mitochondria in autophagosomes following autophagy induction by nitrogen starvation (15). Under natural conditions, yeast often faces environmental changes requiring a modulation of its mitochondrial content (reviewed in Refs. 16, 17, and 18). For example, the shift from a respiratory carbon source to a fermentative carbon source or, more drastically, the shift from aerobic to anaerobic conditions is accompanied by a decrease of both the amount and the enzymatic equipment of mitochondria. Clearly, the inhibition of mitochondrial biogenesis is likely to play an important role in these changes (19,20) but might not be enough to explain the rapid disappearance of cohorts of mitochondria. One study, based on a mutant inactivated in the mitochondrial protease Yme1p, suggests that alterations of mitochondrial biogenesis are able to trigger mitochondrial autophagy (21,22). But hypotheses about the role of mitochondria in their own degradation remain highly speculative, because a molecular support for these hypotheses is still lacking (2).
In the present study, we used the autophagy-inducing drug rapamycin or physiological autophagy induced by nitrogen starvation to provide evidence that mitochondria are early targets of autophagic degradation in yeast, and we identified for the first time that a mitochondrial protein, Uth1p, is involved in this process.

MATERIALS AND METHODS
Yeast Strains, Plasmids, and Growth Conditions-W303-1B (MATa, ade2, his3, leu2, trp1, ura3, can1) was used as the wild-type strain. The ⌬uth1 strain was constructed from W303-1B as described previously (23). The ⌬atg5 strain was constructed by disruption of the ATG5 gene with the KanMX4 cassette in the W303-1B strain. ⌬pep4, derived from the wild-type strain BY4742 (⌴〈⌻a, his3, leu2, ura3, met15), was obtained from Euroscarf (Frankfurt, Germany). For the measurement of ALP 1 activity, the PHO8 locus was replaced with PHO8⌬60 in each strain by transformation with an HindIII fragment of the plasmid pTN9 (a gift from Y. Ohsumi, National Institute for Basic Biology, Okazaki, Japan) bearing PHO8⌬60 as described previously (24). For fluorescent microscopy experiments the cells were transformed with the plasmid pGAL-CLbGFP containing the presequence of mitochondrial citrate synthase fused to GFP under the control of a GAL1/10 promoter, as described (25), which was provided by J. P. di Rago (Institut de Biochimie et Génétique Cellulaires, Bordeaux, France).
Nitrogen starvation medium (SD-N) contained 0.175% yeast nitrogen base without amino acids and ammonium sulfate and 2% glucose or 2% lactate, respectively. To induce pGAL-CLbGFP expression, cells were grown overnight in an appropriate medium supplemented with 2% galactose. Rapamycin treatments were done with 0.2 g/ml rapamycin by using a stock solution of 100 g/ml rapamycin (Calbiochem) in 90% ethanol and 10% Triton X-100.
Plating Efficiency-The viability of cells was determined as plating efficiency. The cells were counted at indicates times and diluted, and aliquots with identical numbers of cells (5 ϫ 10 2 ) were plated on solid complete YPD (1% yeast extract, 2% peptone, and 2% dextrose) or YPL (1% yeast extract, 2% peptone, and 2% lactate) medium, respectively. The number of colonies was scored after 2-3 days of growth at 28°C. The viability of cells in each sample was expressed as the percentage of colonies produced by the plating of the same aliquots of culture at time 0.
Alkaline Phosphatase Assay-ALP activity assays using ␣-naphthyl phosphate (Sigma) as a substrate were performed on cells that had been treated with rapamycin (0.2 g/ml) for 2 h or nitrogenstarved for 15 h according to the method described (27). Fluorescence intensity was measured at 472 nm (excitation at 345 nm) in a Safas spectrofluorometer. Protein concentration was measured with the Lowry method (28). ALP activities were expressed as arbitrary fluorescence units/min/mg protein.

RESULTS
The Product of the UTH1 Gene Is Involved in the Response to Rapamycin-It has been shown previously that inactivation of the UTH1 gene induces resistance to rapamycin as well as to the heterologous expression of the human pro-apoptotic protein Bax, suggesting that Uth1p might be at the crossroads of different signals converging on mitochondria and leading to death (26).
Physiological autophagy induced by nitrogen starvation does not lead to cell death. On the contrary, strains impaired for autophagy, such as the ⌬atg5 mutant, do not survive nitrogen starvation (30). But overactivation of autophagy by rapamycin treatment induces growth arrest. As shown in Fig.  1, A and B, the growth of the ⌬uth1 mutant exhibits rapamycin resistance when cells were spotted on a solid medium.
FIG. 1. ⌬uth1 strain is resistant to rapamycin. A and B, drop test. Cells were dropped in batches of 1000, 500, 100, and 50 cells on YPD (A) or YPL (B) medium with or without rapamycin (10 ng ml Ϫ1 ). C and D, survival curves. Cells were grown aerobically in a YNB medium supplemented with 2% glucose (C) or 2% lactate (D). Rapamycin (0.2 g ml Ϫ1 ) was added and, at indicated times, aliquots of 500 cells were spread on YPD plates. The number of colonies was counted after a 3-day incubation at 28°C. Each curve represents the average of 10 independent experiments; S.D. Ͻ 1%. f, wild-type; Ⅺ, ⌬uth1; ‚, ⌬atg5.
This resistance is more marked than that of the autophagydeficient ⌬atg5 mutant.
The effect of rapamycin was further assayed on the viability of cells grown on glucose (fermentation-dependent growth; Fig.  1C) or lactate-supplemented medium (respiration-dependent growth; Fig. 1D). Rapamycin induces a loss of viability of wildtype and ⌬atg5 strains, but the ⌬uth1 strain exhibited a strong resistance. It is noteworthy that this resistance was more marked on lactate-supplemented medium (Fig. 1D) where yeast cells develop large and fully active mitochondria, because this carbon source is exclusively oxidized by mitochondrial metabolic pathways (31). Whatever the carbon source, under nitrogen starvation conditions, a better survival rate was observed for the ⌬uth1 strain in comparison with the wild-type, the autophagy-deficient ⌬atg5 mutant, and the ⌬pep4 mutant, which is deficient in the major vacuolar hydrolase, proteinase A (Fig. 2).
Autophagic Machinery Is Not Affected in ⌬uth1 Cells-We were interested in examining the cellular autophagic activity in the ⌬uth1 mutant to know if the observed resistance to rapamycin was due to a general blockage of the autophagic process. Several biochemical methods for assaying autophagy have been established. One improved method (24) is to follow the vacuolar delivery of the Pho8⌬60p, a cytosolic variant of the Pho8p alkaline phosphatase in which the membrane-spanning region and the vacuolar targeting sequence have been deleted. Such an N-terminal truncated Pho8⌬60p variant cannot be delivered to the vacuolar compartment via the normal secretory pathway and is only imported to the vacuole upon the induction of autophagy, when cells are shifted to nitrogen starvation conditions or treated with rapamycin. The vacuolar delivery of randomly sequestrated Pho8⌬60p in autophagic vesicles results in cleavage of the C-terminal propeptide in a proteinase B-dependent manner and activation of its enzyme activity. In our study, ALP activity was assayed in wild-type cells and ⌬uth1 and ⌬atg5 mutant strains grown on glucose- (Fig. 3A) or lactate-supplemented (Fig. 3B) minimal medium after a 1-h rapamycin treatment or 15 h of nitrogen starvation, respectively. The level of autophagy was indicated by the difference in enzymatic activity found at time 0 (control) and under these two conditions.
At time 0 all tested strains exhibited approximately the same basal activity, which was probably due to the participation of Pho13p, another alkaline phosphatase that is constitutively active, in the reaction (32). In the ⌬atg5 strain, where the autophagic process was impaired and, consequently, Pho8⌬60p cannot be delivered to the vacuole, no increase of ALP after rapamycin treatment or under starvation conditions was ob- served. Under the same conditions, ALP activity was significantly increased both in the wild-type and the ⌬uth1 strains whatever the condition of induction. The highest autophagic activity was found in starved ⌬uth1 mutant cells, when ALP activity was increased five times.
Another method for measuring autophagic activity is to follow the processing of the vacuolar aminopeptidase I (API) in starved cells (33). API is synthesized on soluble ribosomes as a 61-kDa inactive precursor that is transported to the vacuole in nutrient-rich conditions by a constitutive non-classical vesicular transport mechanism, the cytoplasm-to-vacuole targeting pathway, and is submitted to proteolytic maturation leading to a 50-kDa form. Under starvation conditions or in presence of rapamycin, API is transported to the vacuole by macroautophagy. Total protein extracts were prepared from cultures of wild-type and ⌬uth1 strains grown on glucose-containing minimal medium and after 15 h of nitrogen starvation. An immunoblot with an antibody specific for API was done to monitor precursor API processing (Fig. 4). Under starvation conditions, only the mature form was visualized in wild-type and ⌬uth1 strains showing the activation of autophagy. Taken together, these two sets of experiments (ALP activity and API maturation) showed that the autophagy process was unaffected in the ⌬uth1 mutant.
The Product of UTH1 Is Required for Vacuolar Degradation of Mitochondria-Targeted GFP-According to the localization of Uth1p in the external mitochondrial membrane (34), the degradation of mitochondria was investigated further.
A construction expressing GFP fused downstream of the targeting sequence of mitochondrial citrate synthase (mtGFP) (25) was introduced in wild-type, ⌬uth1, and ⌬atg5 strains in order to follow the behavior of mitochondria in situ. The autofluorescent probe FM4-64, which specifically labels the vacuolar membrane in yeast cells (29), was used simultaneously. Under control conditions, wild-type cells grown on a lactate and galactose-supplemented medium exhibit a well differentiated mitochondrial network, and FM4-64 labeling shows the presence of one or two large vacuoles (Fig. 5). Rapamycin treatment of wild-type cells induced the disappearance of the mitochondrial network and the appearance of patches predominantly at the periphery of the cells. Most cells lost the classical vacuolar membrane FM4-64 staining, and only few cells showed faintly stained vacuoles. The proportion of cells in which FM4-64 fluorescence was lost from vacuolar membranes to give rise to diffused cytoplasmic staining corresponded with the proportion of dead cells (85%; see Fig. 1D). The remaining 15% of the cells, which were still viable after a 2-h rapamycin treatment (Fig. 1D), exhibited intensely punctuated FM4-64 staining, with some of them colocalizing with mtGFP as shown in Fig. 5. The same perturbation of vacuolar membrane staining after rapamycin treatment was observed for glucose-grown cells treated with rapamycin (data not shown).
The expression of mtGFP in ⌬uth1 and ⌬atg5 mutants also revealed a normal mitochondrial network (Fig. 5). Opposite to the case with wild-type, the addition of rapamycin to both mutants did not induce any rapid disorganization of the mitochondrial network that remained intact after a 2-h treatment (Fig. 5), and the network was marginally affected after a 6-h treatment (not shown). No colocalization of mtGFP and FM4-64 was ever observed in these strains. Interestingly, FM4-64 staining of vacuolar membranes showed the presence of a large number of small fragmented vacuoles in ⌬uth1 as described in some mutants defective in vacuolar biogenesis and autophagy, such as in the ⌬atg5 strain (Fig. 5) (35).
To verify that rapamycin treatment did not induce any side effects by affecting mtGFP synthesis or import, cells were incubated in the presence of cycloheximide (an inhibitor of protein synthesis) or carbonyl cyanide m-chlorophenylhydrazone (CCCP; an uncoupler dissipating transmembrane electrical potential). For the time of the experiment, the mitochondrial network remained unaffected by these treatments in both the wild-type (Fig. 6) and the ⌬uth1 strains (not shown).
Under conditions of physiologically induced autophagy by FIG. 4. The absence of Uth1p does not affect starvation-induced processing of precursor API (prAPI). Wild-type and ⌬uth1 cells were grown in a YNB medium containing 2% glucose to an A 600 of 1 and then transferred to the nitrogen starvation medium (SD-N) identically as described in Fig. 2. Protein extracts were prepared from these cultures after 15 h of starvation, and API processing was assessed by immunoblot using a polyclonal antiserum specific for API. Processing at starvation point (ϪN) was compared with that observed in the non-starved control culture (SD). mAPI, mature API.
FIG. 5. The product of the UTH1 gene is required for rapamycin-induced degradation of mtGFP. Wild-type, ⌬uth1, and ⌬atg5 strains were grown in a YNB medium supplemented with 2% lactate. mtGFP expression (green) was induced by adding 2% galactose. After 12 h, FM4-64 (at a final concentration of 40 M; red) was added for 4 h. Then rapamycin (0.2 g/ml) was added or not added, and cells were examined under a fluorescence microscope after 2 h of treatment as described under "Materials and Methods." The arrow in the overlay picture points to mtGFP/FM4-64 co-staining. VIS corresponds to differential interference contrast (DIC). nitrogen starvation, the appearance of patches of mtGFP at the periphery of the cells and the concomitant disappearance of the mitochondrial network were also observed in wild-type cells (Fig. 7). Moreover, co-staining with FM4-64 revealed the presence of some portion of mtGFP in the vacuoles (Fig. 7). Both phenomena were completely absent in control cells growing in rich medium (Fig. 5). In accordance with the deficiency of autophagy, no signs of mitochondria degradation were ever observed in ⌬atg5 mutant cells under the same conditions (Fig.  7). In the culture of a ⌬uth1 strain starved for 15 h, the following two distinct behaviors were observed. (i) About half of the cells exhibited normal FM4-64 staining and an intact mitochondrial network like that of the ⌬atg5 mutant. (ii) The other half exhibited some hallmarks of a weak autophagy process, with patches of mtGFP at the periphery of the cells and vacuolar delivery of mtGFP but also the visible presence of a residual mitochondrial network (Fig. 7).
Uth1p Is Required for the Early Vacuolar Degradation of Mitochondrial Proteins-The time course of autophagy induction by rapamycin was also followed as the increase in the content of Atg8p, which is essential for the formation of autophagosomes (36,37). In accordance with this observation, we observed an increase of the intracellular content of Atg8p after the addition of rapamycin in all tested strains growing under strict respiratory conditions (Fig. 8).
The amounts of several mitochondrial proteins localized in the outer (the mitochondrial porin Por1p) and inner (the cytochrome c oxidase subunit 2, Cox2p) mitochondrial membranes were followed by Western blot analyses in comparison with proteins localized in cytosol phosphoglycerate kinase (Pgk1p) and actin (Act1p) (Fig. 8). In the wild-type strain, the content of both mitochondrial proteins was strongly depressed 2 h after the beginning of rapamycin treatment and further decreased over a 6-h treatment. Under the same conditions, the amounts of cytosolic proteins were only marginally affected. When the wild-type cells were grown on glucose, this more rapid decrease of mitochondrial proteins was also observed (not shown).
To verify that the rapamycin-induced decrease of the two mitochondrial proteins was actually caused by vacuolar proteolysis and not by a blockade of their synthesis, the experiment was reproduced in a ⌬pep4 strain (Fig. 8). The amounts of Por1p and Cox2p did not decrease after a 6-h rapamycin treatment, whereas in fact a slight increase was actually observed.
This lack of degradation in ⌬pep4 cells confirms that the protein degradation observed in wild-type cells is executed in the vacuoles.
Rapamycin treatment of the autophagy-deficient ⌬atg5 strain grown on lactate induced a higher increase of Atg8p in wild-type (Fig. 8); this is in accordance with the fact that the process of autophagosome formation and the subsequent vacuolar degradation of Atg8p is blocked in this mutant. Concerning mitochondrial proteins, no decrease in the amounts of Por1p and Cox2p was observed, but, as in the ⌬pep4 mutant, a slight increase was recorded instead.
From this set of experiments it was seen that rapamycin treatment of cells grown under strict respiratory conditions induces the early degradation of mitochondrial proteins by a process that requires the following: (i) the formation of autophagosomes; and (ii) vacuolar proteolysis. During the short course of the experiment cytosolic proteins remain poorly affected, suggesting that fully differentiated organelles like mitochondria are early targets of the rapamycin-induced process. Also, Por1p and Cox2p, which are localized on different membranes, are degraded together, suggesting that whole mitochondria are targeted by the process.
In accordance with the resistance of the ⌬uth1 strain to rapamycin (Fig. 1), degradation of the two mitochondrial proteins was largely impaired in this strain (Fig. 8). On other hand, the product of ATG8 increased up to a level similar to that of the wild-type strain (Fig. 8), confirming the conclusion drawn from the measurement of ALP activity (Fig. 3) and API processing (Fig. 4) that the autophagic process is still active. This supports the view that the resistance of the ⌬uth1 strain is not caused by a general defect in the autophagic machinery but that the mitochondrial degradation targeted by this machinery is delayed. DISCUSSION Data reported in this paper allowed us to identify conditions where a rapid degradation of mitochondria by rapamycin treatment or nitrogen starvation growth could be visualized. Within 2 h of rapamycin treatment, mitochondrial proteins are de- graded in a way that depends on the following: (i) autophagosomes formation, because it is prevented by atg5-inactivation; and (ii) vacuolar degradation, because it is prevented by pep4 inactivation. This degradation of mitochondrial proteins correlates with the disappearance of the mitochondrial network and the appearance of a mitochondrial/vacuole co-staining. This mitochondrial degradation does not occur in strains impaired for autophagy such as the ⌬atg5 mutant. Opposite to the case with the ⌬atg5 mutant, the autophagic machinery remains fully active in ⌬uth1 mutant, but mitochondria degradation by this process is delayed. These results showed that the mitochondrial protein Uth1p is involved in autophagic mitochondrial degradation.
This protein, Uth1p, has been initially found in a genetic screen aiming to identify proteins involved in the regulation of the yeast life span (38). Furthermore, it was found to be a member of the so-called "SUN" family, which includes four proteins having a very high degree of identity (23,39). The product of UTH1 exhibits some remarkable properties. It is mainly localized in the mitochondrial outer membrane (34) but was also found in the cell wall (34,40). This alternative cell wall localization is likely to have a physiological meaning, because it can be selectively suppressed by point mutations in the protein. 2 The absence of Uth1p does not have dramatic consequences under normal growth conditions; it induced a 15-25% decrease of mitochondrial cytochromes as well as other mitochondrial enzymes and an increase of growth yield (23). This higher metabolic efficiency, linked to a general but slight slowing down of mitochondrial functions, is likely to explain the extended life span as well as the modified response to oxidative stress exhibited by this mutant (41).
More recently, Uth1p was found to be involved in yeast cell death induced by the heterologous expression of the human proapoptotic protein Bax (26). As in mammalian cells, Bax targets the mitochondrial outer membrane, where it induces the formation of a large channel likely involved in the mitochondria-to-cytosol relocalization of cytochrome c (42)(43)(44). Bax-induced characteristics in yeast resemble apoptotic like hallmarks induced by H 2 O 2 treatment (45) in which a caspase-like activity might be involved (46). However, cytochrome c release is not mandatory to Bax-induced cell death in yeast (47). In addition, because Uth1p is not involved in the mitochondrial targeting of Bax or in cytochrome c release activity (26), we proposed that Bax was able to activate an alternative deathinducing pathway involving Uth1p. As a matter of fact, Uth1p overexpression was found to induce cell death, and uth1 inactivation induced a resistance of growth to rapamycin (26). These observations suggested that Uth1p might be involved in a death pathway related to autophagy.
The observation reported in the present paper that Uth1p is involved in the autophagic degradation of mitochondria induced by rapamycin or nitrogen starvation supports this hypothesis. Clearly, Uth1p does not participate in the autophagic machinery, because the biochemical hallmarks related to this process (ALP activation and API processing) are as active as those in wild-type, in contrast to the behavior of the autophagydeficient strain ⌬atg5. More likely, Uth1p might be involved in the recognition of mitochondria by the autophagic machinery. However, it should be noted that, both from viability and fluorescence microscopy experiments, about half of ⌬uth1 cells are still able to undergo normal mitochondria autophagy, suggesting that Uth1p may not be the only protein involved in the process.
The degradation process termed autophagy covers distinct, although related, phenomena. The best characterized, and also the less specific, is the macroautophagy of large parts of the cytoplasm including the organelles (reviewed in Ref. 2). However, more selective phenomena were also evidenced. As an example, microautophagy of the nucleus via the physical contact between nuclei and vacuoles was shown to involve selective protein-protein interactions such as Nvj1p on the nuclear envelope and Vac8p on the vacuolar membrane (5). This process does not involve the well established components of macroautophagy signaling but is still under the control of the TOR (target of rapamycin) kinase pathway. Another example of selective autophagy was observed in the methylotrophic yeast Pichia pastoris, where peroxisomes can be selectively degraded FIG. 8. Degradation of mitochondrial proteins is dependent on Uth1p. A, wild-type, ⌬uth1, ⌬atg5, and ⌬pep4 cells were grown in a YNB medium supplemented with 2% lactate. Rapamycin (0.2 g/ml) was added and, at indicated times, whole proteins were extracted from the same number of cells and analyzed by immunoblot as described under "Materials and Methods." B, protein quantification was done on non-saturated blots using a scanner and ImageJ software. Proteins were quantified at time 0 (control, non-treated culture) and after a 2-h rapamycin treatment. Data are the ratio between treatment over control and are average (Ϯ S.D.) from 3-5 independent experiments. by vacuoles in a way independent from macroautophagy (3). The degradation of other organelles has not been investigated to date, but tools such as selectively targeted GFP and specific antibodies are now available for such studies.
Concerning yeast mitochondria, since the first report that they could be engulfed by the macroautophagy machinery (15), one study has suggested that a selective mechanism of mitochondria autophagy also occurred in yeast, i.e. inactivation of the mitochondrial AAA-type protease Yme1p leads to a vacuoledependent mitochondria-to-nucleus transfer of genetic material (21,22). However, the molecular basis of this transfer has not been identified to date.
The finding that Uth1p is a component involved in the early selective degradation of mitochondria is now a basis for the identification of the other components involved in the autophagic degradation of mitochondria. This might be of general interest, not only in yeast but also, given the recent evidence of connections between apoptosis and autophagy, in mammalian cells (48).