|Home | About | Journals | Submit | Contact Us | Français|
Autophagy is a highly conserved degradative process in eukaryotic cells. This process plays an integral role in cellular physiology, and the levels of autophagy must be precisely controlled to prevent cellular dysfunction. The rapamycin-sensitive Tor kinase complex 1 (TORC1) has a major role in regulating the induction of autophagy; however, the regulatory mechanisms are not fully understood. Here, we find that Tap42 and protein phosphatase type 2A (PP2A) are involved in the regulation of autophagy in yeast. Temperature-sensitive mutant alleles of TAP42 revealed that autophagy was induced without inactivation of TORC1. Absence of the Tap42-interacting protein Tip41 abolished autophagy induction in the tap42 mutants, whereas overexpression of Tip41 activated autophagy. Furthermore, inactivation of PP2A stimulated autophagy and overexpression of a catalytic subunit of PP2A blocked rapamycin-induced autophagy. Our data support a model in which autophagy is negatively regulated by the Tap42-PP2A pathway.
Macroautophagy (hereafter autophagy) is a major degradative process, evolutionarily conserved in eukaryotes.1 Under certain stress conditions such as depletion of nutrients and/or growth factors, cells are stimulated to sequester portions of the cytoplasm, including organelles, within double-membrane vesicles. The completed vesicles, called autophagosomes, fuse with the degradative lysosome/vacuole, allowing access to the inner membrane compartment. The autophagosome inner membrane is then lysed and the contents are degraded; the resulting macro-molecules are released back into the cytosol for subsequent reuse.2 Autophagy is involved in a variety of cellular physiological events such as development, proliferation, remodeling, aging and defense against pathogens.3 Although controversial, autophagy may also contribute to cell death.4,5 Importantly, this process must be tightly regulated because either excess or insufficient autophagy can be deleterious.
The target of rapamycin (Tor) signaling pathway has a major role in controlling induction of autophagy,6,7 although the mechanism of regulation is not fully understood. Tor is a protein kinase whose activity depends on nutrient availability. The Tor signaling pathway is evolutionarily conserved in eukaryotes.8 Tor forms two functionally distinct protein complexes, Tor complex 1 and 2 (TORC1 and TORC2). While TORC2 primarily regulates polarization of actin cytoskeleton and the spatial aspect of cell growth, TORC1 has a direct role in regulation of autophagy, especially in response to low nutrient and/or growth factor conditions, which, similar to rapamycin, inactivates TORC1 activity.9 TORC1 also mediates the regulation of several downstream processes that are modulated in response to stress conditions, such as ribosome biogenesis, protein synthesis and cell cycle control.8,10 A paradigm for autophagy regulation was established based on the mechanism through which Tor acts in controlling cell growth.11,12 According to this model, Tor regulates the function of Tap42, which in turn controls the activity of the protein phosphatase type 2A (PP2A)-related phosphatase Sit4.13 Inactivation of TORC1 stimulates Sit4 activity by altering the nature of its association with Tap42.9,14
Tip41 was identified as a binding partner of Tap42 in yeast and has been suggested to function as a negative regulator of Tap42;15 when TORC1 is inactivated by nutrient depletion or rapamycin treatment, Tip41 associates with Tap42 more efficiently, resulting in inactivation of Tap42. The mammalian ortholog of Tip41, however, does not interact with a mammalian Tap42 ortholog, α4, but interacts with PP2A to regulate its activity.16–18
Although phosphatases are involved in the regulation of TORC1-dependent cellular events, the relationship between autophagy and TORC1-regulated phosphatases has not been examined in detail. In this study, we found that inactivation of Tap42 stimulated autophagy in nutrient-rich conditions without inactivation of TORC1. This autophagy induction was suppressed by the concurrent deletion of Tip41. In agreement, overexpression of Tip41 also led to autophagy upregulation. We also demonstrated that the PP2A phosphatase Pph21/Pph22, but not Sit4, was involved in the regulation of autophagy. Inactivation of PP2A stimulated autophagy, whereas overexpression of a catalytic subunit of PP2A blocked autophagy induced by rapamycin. Our data revealed a previously unknown negative role of the Tap42-phosphatase pathway in the regulation of autophagy.
Previous data showed that inactivation of Tap42 does not block autophagy induction by rapamycin,19 suggesting that Tap42 does not positively regulate autophagy. To better understand whether Tap42 plays a role as a negative regulator of autophagy downstream of TORC1, we monitored the processing of the GFP-tagged autophagy-related (Atg) protein Atg8 in yeast cells expressing a temperature-sensitive allele of TAP42, tap42-109. Upon induction of autophagy, Atg8 is transported into the vacuole as a membrane component of the inner vesicle of the autophagosome, followed by degradation within the vacuole lumen. In contrast to the Atg8 protein, the GFP moiety is relatively resistant to proteolysis. Accordingly, the autophagic process can be monitored based on accumulation of free GFP processed from GFP-Atg8.20 Thus, this assay can reflect the autophagy activity quantitatively, in contrast to the previously reported method, i.e., the observation of autophagic bodies by light microscopy.19 At permissive temperature (23°C), the level of free GFP was barely detectable in the absence of rapamycin in cells expressing either TAP42 or the temperature-sensitive allele tap42-109 (Fig. 1A).14 On the other hand, treatment with rapamycin resulted in the obvious accumulation of GFP, indicating induction of autophagy in response to inactivation of TORC1. When cells were cultured at nonpermissive temperature (37°C), free GFP was generated in cells expressing TAP42 only in the presence of rapamycin. In contrast, a significant amount of free GFP was detected with cells expressing tap42-109 whether or not rapamycin was present. These results suggested that inactivation of Tap42 promoted processing of GFP-Atg8, and hence caused the induction of autophagy.
This finding concerning the role of Tap42 was in contrast to the previously reported data.19 Accordingly, we wanted to determine whether the autophagy induction that we observed was due to the normal pathway or represented some type of bypass mechanism. To verify that the observed induction depended on proteins that are known to regulate autophagy, we deleted the ATG1, ATG13 or ATG17 gene in cells expressing tap42-109 and monitored the processing of GFP-Atg8 (Fig. 1B). Deletion of ATG1 or ATG13 completely blocks autophagy, whereas atg17Δ mutants display greatly reduced levels of autophagy.21 At permissive temperature, deletion of either the ATG1 or ATG13 gene blocked GFP-Atg8 processing induced by rapamycin treatment, while a low level of free GFP was detected in the atg17Δ tap42-109 strain in the presence of rapamycin, in agreement with the partial defect in autophagy resulting from this mutation. Similarly, when Tap42 was inactivated at nonpermissive temperature, no free GFP was detected without rapamycin in the atg1Δ or atg13Δ background. Deletion of Atg17 also significantly reduced the processing of GFP-Atg8 caused by Tap42 inactivation. These results indicate that the processing of GFP-Atg8 induced by inactivation of Tap42 occurred in an autophagy-dependent manner.
To confirm the role of Tap42 in autophagy inhibition, we examined autophagy using another assay, the formation of phosphatidylethanolamine (PE)-conjugated Atg8 (Atg8–PE). Upon autophagy induction, in addition to an elevation of its expression, Atg8 is conjugated to PE and associated with autophagosome membranes, resulting in an increase of Atg8–PE.22 In both TAP42 and tap42-109 cells treated with rapamycin, Atg8 and Atg8–PE levels increased at 23°C or 37°C (Fig. 1C). When shifted to 37°C in the absence of rapamycin, Atg8 and Atg8–PE were significantly enhanced in tap42-109 cells compared to wild-type TAP42 cells. This result suggested that Tap42 inactivation induced autophagosome formation.
To further test whether the inactivation of Tap42 induces autophagy, we used one additional assay that does not directly monitor Atg8. Precursor aminopeptidase I (prApe1) is transported into the vacuole as a specific cargo by the cytoplasm to vacuole targeting pathway in nutrient-rich conditions, where it is turned into the mature form by cleavage of the propeptide.23 During starvation conditions, prApe1 is transported by autophagy, and it remains a selective cargo due to the use of a receptor. The precursor and mature (mApe1) forms can be resolved by SDS-PAGE and detected by immunoblotting. In order to measure autophagy-specific delivery of prApe1 to the vacuole, we used the vac8Δ mutant to eliminate the background of mApe1 resulting from the constant delivery of prApe1 to the vacuole during nutrient-rich conditions. Vac8 is required for transport of prApe1 to the vacuole during vegetative growth, but is not essential for starvation-induced bulk autophagy; the vac8 Δ mutant accumulates prApe1 in vegetative conditions, whereas it is rapidly processed into mApe1 upon induction of autophagy.24 At 23°C, TAP42 and tap42-109 cells harboring a deletion of VAC8 accumulated prApe1, but processed this protein in response to rapamycin treatment (Fig. 1D). After being shifted to 37°C, TAP42 cells deleted for VAC8 could mature prApe1 only with rapamycin treatment, whereas VAC8-deleted tap42-109 cells showed maturation of prApe1 even in the absence of rapamycin. Importantly, additional deletion of ATG1 blocked maturation of prApe1 in tap42-109 cells in any condition, indicating that maturation of prApe1 is due to an autophagic process in the vac8Δ background. This result further demonstrated that inactivation of Tap42 induced the autophagic delivery of cargos to the vacuole.
Thus, by three different assays we were able to show that the alteration of Tap42 activity resulted in autophagy induction. To further address the discrepancy between our present results and the previously published data (that utilized the tap42-11 mutant) we next examined whether this autophagy-related phenotype was specific to the tap42-109 allele. We found that two other temperature-sensitive alleles of TAP42, tap42-11 and tap42-106,13,14 also allowed delivery of prApe1 to the vacuole at the nonpermissive temperature in the absence of rapamycin in the vac8Δ background (Fig. 1D). Similarly, all three mutant alleles showed elevated GFP-Atg8 processing under these same conditions (Fig. 2A).
We extended our analysis by examining the ability to induce autophagy based on the quantitative Pho8Δ60 assay. Pho8 is an alkaline phosphatase present in the vacuole membrane, whereas the mutant Pho8Δ60 version is unable to enter the secretory pathway and remains in the cytosol; Pho8Δ60 can be delivered to the vacuole via autophagy, resulting in its proteolytic maturation, which can be measured enzymatically.25 In the strain background used in these studies, the shift to 37°C caused a reduction in Pho8Δ60-dependent alkaline phosphatase activity in the wild type (Fig. 2B), and a similar reduction was seen with the tap42-106 mutant. In all cases, however, the cells expressing mutant forms of TAP42 showed a higher level of Pho8Δ60-dependent alkaline phosphatase activity at nonpermissive temperature compared to the wild-type strain. Taken together, these results indicated that autophagy induction due to Tap42 inactivation is not specific to the tap42-109 mutant allele, and that Tap42 negatively regulates autophagy.
Atg1 kinase activity is upregulated upon binding to Atg13, which is thought to be important for the induction step of autophagosome biogenesis.19 Accordingly, we tested whether autophagy induced by Tap42 inactivation required Atg1 kinase activity. We monitored GFP-Atg8 processing and vacuolar accumulation of free GFP in atg1Δ cells expressing two kinase-defective mutants, Atg1K54A and Atg1K54A,M102A; the K54A mutation results in a partial block in autophagy, whereas the double mutant displays a more severe defect.19,26,27 At 23°C, unlike wild-type Atg1, both mutants blocked the processing of GFP-Atg8 in TAP42- or tap42-109-expressing cells treated with rapamycin (Fig. 3A). At 37°C, TAP42 and tap42-109 cells with the Atg1K54A mutant displayed reduced processing relative to the wild type in the presence of rapamycin, whereas the Atg1K54A,M102A mutant showed a complete block. Similarly, cells expressing wild-type Atg1 showed vacuolar accumulation of GFP-Atg8 as detected by fluorescence microscopy, whereas the fluorescent signal was greatly reduced or completely absent in tap42-109 cells expressing the Atg1 mutants in the absence of rapamycin at the nonpermissive temperature (Fig. 3B); the vacuolar signal in the presence of wild-type Atg1 was due to autophagy induction because it was below the level of detection at 23°C when only the Cvt pathway was operating. These results indicated that Atg1 kinase activity is required for autophagosome formation induced by inactivation of Tap42.
Depletion of Tip41, an interaction partner of Tap42 that negatively regulates Tap42 upon binding, suppresses the growth defect seen in tap42-11 cells.15 To provide further confirmation for the role of Tap42 in regulating autophagy, and to investigate the potential role of Tip41 in Tap42 regulation, we decided to determine whether Tip41 affects autophagy induced by Tap42 inactivation. We first determined that TIP41 deletion could rescue the growth of cells with the tap41-109 allele (Fig. 4A). As expected, the tap42-109 cells with or without Tip41 were essentially indistinguishable from wild-type TAP42 cells at the permissive temperature of 24°C. At a semipermissive temperature of 30°C or nonpermissive temperature of 37°C, tap42-109 cells displayed a substantial block in growth. The deletion of TIP41 effectively rescued the growth of tap42-109 cells to a degree comparable with TAP42 cells. These results suggested that TIP41 deletion could suppress the growth defect in tap42-109 cells, consistent with the previously published result with tap42-11 cells.15
Next, we examined whether depletion of TIP41 could suppress induction of autophagy in tap42-109 cells, using the GFP-Atg8 processing assay (Fig. 4B). When grown at 37°C, unlike the tap42-109 TIP41 strain, TIP41-deleted tap42-109 cells did not show significant processing of GFP-Atg8 when no rapamycin was applied. The deletion of ATG1 again served as a negative control. We also observed that in the absence of rapamycin, the accumulation of Atg8–PE due to Tap42 inactivation was blocked when TIP41 was deleted (Fig. 4C). These data suggested that Tip41 was required for autophagy induced by inactivation of Tap42.
The above results revealed a potential role for Tip41 as a positive regulator of autophagy. To test our hypothesis, we over-expressed Tip41 under the control of the GAL1 promoter, which is induced by galactose. As shown previously,15 in the presence of galactose, cell growth was impaired due to Tip41 overexpression and hence Tap42 inactivation (Fig. 5A and B). We observed a low yet significant level of GFP-Atg8 processing in the cells overexpressing Tip41 even in the absence of rapamycin (Fig. 5C). This processing of GFP-Atg8 was not due to galactose per se, because in the presence of galactose, cells with the vector control did not show any processing of GFP-Atg8 unless the cells were treated with rapamycin. We note that induction of autophagy was relatively minor with Tip41 overexpression, and was detected only starting with the 6 h time point, even though a substantial level of Tip41 was detected by 4 h of growth in galactose. We interpret this to indicate that Tip41 by itself is only able to partially inhibit Tap42, compared to the temperature sensitive tap42 mutants. Thus, Tip41 serves to amplify the signal when Tor is inactivated, but it does not normally act to shut off Tap42 function when Tor is active. We subsequently confirmed that GFP-Atg8 was processed through autophagy following Tip41 overexpression because no processing was observed in autophagy-defective atg1Δ cells even if Tip41 was overexpressed (Fig. 5D).
Tap42 regulates two types of phosphatase complexes, PP2A and the PP2A-related phosphatase Sit4. The latter is suggested to positively regulate autophagy through dephosphorylation of Ure2 and Gln3.28 PP2A catalytic subunits consist of the partially redundant proteins Pph21 and Pph22, and have not previously been implicated in autophagy regulation. Therefore, we decided to examine whether these phosphatases are involved in regulation of autophagy using two temperature-sensitive mutants, pph22-172 and sit4-102.14 At 23°C, rapamycin treatment effectively induced variable levels of GFP-Atg8 processing in all the strains expressing wild-type Pph22, Sit4, or the two mutant alleles (Fig. 6A). This result suggests that inactivation of Tor was able to induce autophagy even when Sit4 was inactive, or alternatively that the sit4-102 mutant retains partial activity. When shifted to 37°C, pph22-172 cells, but not sit4-102 cells, showed a significant level of free GFP even without rapamycin, similar to the result seen with tap42-109 cells. This result suggested that PP2A, but not Sit4, is involved in the inhibition of autophagy.
Next, we examined whether overexpression of phosphatases repressed autophagy. We used the GAL1 promoter to drive the expression of the SIT4, PPH21 and PPH22 genes. The effect on autophagy was measured by the GFP-Atg8 processing assay following rapamycin treatment (Fig. 6B). When overexpressed, either Pph21 or Pph22 significantly delayed GFP-Atg8 processing compared to the vector control. On the other hand, overexpression of Sit4 had little effect on the processing of GFP-Atg8. These results together suggested that the PP2A phosphatase negatively regulates autophagy.
To examine whether such regulation of autophagy requires phosphatase activity, we examined GFP-Atg8 processing upon overexpression of the Pph22H186N mutant. This mutant has a substitution of asparagine for histidine at the phosphatase active site, which is highly conserved in PP2A catalytic subunits among species, resulting in loss of phosphatase activity.29 Unlike the wild-type Pph22, overexpression of Pph22H186N showed a similar level of GFP-Atg8 processing in response to rapamycin as the vector control (Fig. 6C). Thus, the phosphatase activity of PP2A is required for its inhibition of autophagy.
Finally, we extended our analysis by further examining the role of Sit4 effectors in autophagy induction. As indicated above, Sit4 acts in part by activating the Gln3 transcription factor. We examined the role of Gln3 or the related GATA-type transcription factor Gat1 by monitoring GFP-Atg8 processing. Deletion of both GLN3 and GAT1 had no effect on autophagy induction (Fig. 6D), in agreement with our finding that Sit4 is not needed for autophagy.
Autophagy is a degradative process that is inducible primarily under nutrient starvation conditions in yeast. TORC1 is a major nutrient sensor that controls a range of downstream reactions through the regulation of phosphatase complexes in response to nutrient conditions. Although it is known that TORC1 is a prominent negative regulator of autophagy, the mechanism through which it controls autophagy is poorly understood. A previous report suggested that TORC1 acts through Tap42, and that Tap42-dependent regulation of Sit4 controls the induction of autophagy.30 However, another study by Kamada et al.19 showed that the temperature-sensitive tap42-11 mutant does not block autophagy induction in response to rapamycin, and therefore noted that Tap42 is not involved in autophagy regulation by TORC1 inactivation.
To clarify the role of Tap42 in the regulation of autophagy, we investigated the function of Tap42 using a series of temperature-sensitive TAP42 alleles. Consistent with previous data, our results showed that the TAP42 mutants including tap42-11 displayed normal induction of autophagy at nonpermissive temperature when treated with rapamycin. Accordingly, we conclude that Tap42 does not act as a positive regulator of the autophagy that is induced upon TORC1 inactivation. Importantly, we found that inactivation of Tap42 stimulates a significant level of autophagy without inactivation of TORC1 by rapamycin, suggesting that Tap42 functions as a negative regulator for autophagy (Fig. 7). One explanation for the discrepancy in the present results and those published previously19 is that the tap42-11 allele has a less severe loss of function at nonpermissive temperature than the tap42-106 or tap42-109 alleles.14 Furthermore, the assay used previously to monitor autophagy, detection of autophagic bodies, is less quantitative than the methods employed in the present analysis.
To further examine the role of Tap42, we determined whether Tip41 has a role in regulating autophagy. Previous genetic analysis has shown that Tip41, which was originally identified as a Tap42-interacting protein, inhibits Tap42 function.15 Deletion of TIP41 did not affect autophagy induced by nutrient starvation or rapamycin treatment in wild-type cells and did not show induction of autophagy without any stimulation (our unpublished data). However, we found that deletion of TIP41 prevented autophagy induced by inactivation of Tap42, and overexpression of Tip41 induced autophagy in vegetatively growing conditions. These results suggest a role for Tip41 that antagonizes Tap42 function with regard to autophagy (Fig. 7). We propose that the Tap42-109 protein retains partial negative activity at nonpermissive temperature and that Tip41 is needed to alter the activity sufficiently to allow autophagy induction.
Previous findings suggest that Tap42 may remain associated with the PP2A and PP2A-related phosphatases upon Tor inactivation, and that it can be a positive regulator of its associated phosphatases in yeast.31,32 Our present results do not address the question of whether Tap42 remains associated with Pph21/Pph22, but they do suggest that Tap42 is needed for PP2A activity. PP2A has essential functions in determining cell integrity such as in cell wall synthesis and actin cytoskeleton organization, and is therefore important during cell growth; loss of PP2A activity causes cell lethality.33 Thus, in growing conditions, PP2A might have a relatively high level of phosphatase activity to maintain these cellular events. Previous data suggest a positive function for the Tap42-PP2A complex in anabolic activities during growth;12 reciprocally, we propose a model in which the Tap42-PP2A pathway plays a negative role in regulating the catabolic autophagy process (Fig. 7). Upon inactivation of TORC1, PP2A is inactivated, which stimulates autophagy. Recent data suggest that the mechanism controlling the activity of the Tap42-PP2A complex may involve multiple steps;31,34 one model is that Tap42-PP2A becomes activated after dissociation from TORC1. This complex soon becomes inactive, however, as the phosphatase dephosphorylates Tap42, a step that presumably correlates with the induction of autophagy. Future studies including the identification of the downstream effector(s) of the PP2A phosphatase will provide more information on the signaling mechanism controlling autophagy.
The S. cerevisiae strains used in this study are listed in Table 1. Yeast strains were grown or incubated in YPD, synthetic medium containing 2% glucose (SMD), 2% galactose and 0.5% raffinose (SMG), or starvation medium (SD-N) as described previously.21 Gene deletions were performed by a PCR-based procedure.21,35,36 If necessary, rapamycin (Fermentek, catalog code R) was used at a final concentration of 0.2 µg/ml.
Plasmids pRS405, pRS406, pATG1(415), pATG1K54A(415) and pATG1K54AM102A(415) have been described previously.19,27,37,38 GFP-tagged Atg8 was expressed under the control of the CUP1 promoter either from a plasmid, pCUGFP-AUT7(416),39 or following integration using pRS405-based pTY425, or pRS406-based pTY426. For construction of plasmid pTY006, triple tandem repeats of a hemagglutin (HA) epitope sequence was PCR-amplified from pFA6a-3xHA-kan 36 and ligated into the HindIII and BamHI sites of pYES2 (Invitrogen, V825-20). N-terminally 3xHA-tagged Tip41, Sit4, Pph21 and Pph22 were expressed under the control of the GAL1 promoter from the pTY006-based plasmids, pTY702, pTY703, pTY706 and pTY707, respectively. The plasmid pTY707-H186N encoding Pph22H186N was generated by PCR-based site-directed mutagenesis as described previously.27
Yeast cells were grown in SMD medium at 30°C to early log-phase, and rapamycin was added to each culture to induce autophagy. For starvation conditions, cells were shifted to SD-N medium. At the indicated times, cells were collected and proteins were precipitated by addition of TCA. Protein extracts were subjected to SDS-PAGE followed by immunoblotting with appropriate antiserum or antibodies.
The Pho8Δ60-dependent alkaline phosphatase assay was performed as described previously.25
We thank Dr. James R. Broach (Princeton University) for providing strains. This work was supported by National Institutes of Health Public Health Service grant GM53396 to D.J.K.