Suppressed translation and ULK1 degradation as potential mechanisms of autophagy limitation under prolonged starvation

ABSTRACT

Macroautophagy/autophagy is a well-organized process of intracellular degradation, which is rapidly activated under starvation conditions. Recent data demonstrate a transcriptional upregulation of several autophagy genes as a mechanism that controls autophagy in response to starvation. Here we report that despite the significant upregulation of mRNA of the essential autophagy initiation gene ULK1, its protein level is rapidly reduced under starvation. Although both autophagic and proteasomal systems contribute to the degradation of ULK1, under prolonged nitrogen deprivation, its level was still reduced in ATG7 knockout cells, and only initially stabilized in cells treated with the lysosomal or proteasomal inhibitors. We demonstrate that under starvation, protein translation is rapidly diminished and, similar to treatments with the proteosynthesis inhibitors cycloheximide or anisomycin, is associated with a significant reduction of ULK1. Furthermore, it was found that inhibition of the mitochondrial respiratory complexes or the mitochondrial ATP synthase function that could also take place in the absence of substrates, promote upregulation of ULK1 mRNA and protein expression in an AMPK-dependent manner in U1810 lung cancer cells growing in complete culture medium. These inhibitors could also drastically increase the ULK1 protein in U1810 cells with knockout of ATG13, where the ULK1 expression is significantly diminished. However, such upregulation of ULK1 protein is negligible under starvation conditions, further signifying the contribution of translation and suggesting that transcriptional upregulation of ULK1 protein will be diminished under such conditions. Thus, we propose a model where inhibition of protein translation, together with the degradation systems, limit autophagy during starvation.

Introduction

Autophagy is an intracellular degradation process, which is rapidly activated in response to various stress stimuli including nutrient deprivation. Recent reports have shown the regulation of autophagy under conditions of nutrient deprivation by several nuclear receptors and transcription factors, which control different steps of the autophagy process including autophagosome formation, autophagosome-lysosome fusion and degradation of autophagic substrates. A number of ATG genes that are transcriptionally upregulated and facilitate autophagy in response to starvation have been identified.^1-4 However, potential mechanisms and players that limit autophagy amplification under starvation conditions remain largely unknown.

The Atg1/ULK complex functions as the most upstream autophagy component in nutrient sensing and as a scaffold for the initiation of the phagophore assembly site (PAS) as well as the recruitment of downstream ATG proteins.^5,6 In mammals, there are several components of the ULK complex of which the serine/threonine ULK1 (unc-51 like autophagy activating kinase 1), RB1CC1/FIP200 (RB inducible coiled-coil 1), ATG13 and ATG101 have been identified. Under nutrient-rich conditions the MTOR complex phosphorylates the ULK1 and ATG13 proteins, leading to disruption of the ULK complex and inhibition of autophagy initiation.^7-9 Structural analysis revealed dephosphorylation of specific serine residues in the Atg13 protein that may enhance its interaction with Atg1 and Atg17 in yeasts, whereas in mammals ULK1 constitutively forms a complex with ATG13 irrespective of nutrient conditions.¹⁰ A key energy sensor, AMP-activated protein kinase (AMPK), directly regulates autophagic activity through phosphorylation of ULK1.^7,11 In addition to protein modification, several other mechanisms that regulate ULK1 expression and autophagy activity under starvation conditions have been proposed. Posttranslational K63 modification has been reported to modulate the rate of ULK1 turnover.¹² In addition, silencing of ATG13 has been suggested to play an important role in the stabilization of ULK1.¹³ Some other proteins, including HSP90 and CDC37 and the recently identified chaperone-like protein C1QBP/p32, also regulate ULK1 stability and autophagy.^14,15 It has also been reported that binding of Atg8 to Atg1/ULK1 in an Atg8-intearcting motif/LC3-interacting region delivers Atg1/ULK1 within an autophagosome for lysosomal degradation in yeast and mammals,¹⁶ suggesting a negative feedback loop in autophagy regulation.

Although many proteins that regulate the initiation of the autophagy pathway under conditions of nutrient deprivation have been identified, it is still not fully understood how, under the conditions of active autophagy, the autophagic amplification signal can be controlled in the long term. Here we report that under conditions of amino acid and serum deprivation the ULK1 protein is rapidly depleted. Such downregulation of ULK1 during starvation is associated with a rapid inhibition of general protein translation and involves multiple degradation pathways. Our study suggests that in addition to proteasomal and lysosomal degradation, another proteolytic system is involved in ULK1 elimination during starvation. Furthermore, we demonstrate that although ULK1 can be transcriptionally upregulated in response to inhibition of mitochondrial respiration (mitochondria activities are diminished under starvation conditions due to the lack of substrates) or the mitochondrial ATP synthase in U1810 cells, such upregulation of ULK1 is not observed under conditions of starvation or treatment with cycloheximide, further indicating the importance of new protein synthesis. Thus, we propose a model showing that in addition to already established modes of autophagy regulation during starvation, inhibition of protein translation together with degradation systems negatively regulate autophagy activity under these conditions, and we suggest ULK1 as one of the ATG proteins that is rapidly reduced and limits autophagy in response to starvation.

Results

ULK1 protein levels are rapidly reduced during cell starvation

Deprivation of growth factors and amino acids is a commonly used condition to activate autophagy in different cellular systems.¹⁷ We aimed to identify the mechanisms allowing autophagy to be controlled in the long term under starvation (nitrogen-deprivation) conditions. To induce autophagy, several human cell lines, including U1810 and A549 lung cancer cells, 143B osteosarcoma cells and HEK293 kidney cells, were starved for different amounts of time in HBSS medium and the expression of several components of the initiator autophagy complex were analyzed (Fig. 1A). During starvation, the level of one of the autophagy initiation proteins, ULK1, was rapidly diminished (Fig. 1A and B). Such downregulation of ULK1 was already observed within 2 h of the initiation of starvation, while the other ULK complex components, including RB1CC1, ATG101 and ATG13, were not drastically changed at this time point (Fig. 1A). Because, based on some previous reports, several ATG proteins are transcriptionally upregulated under starvation conditions,^1-4 we checked the expression of ULK1 in cells under nitrogen-deprivation conditions. In contrast to the reduction of protein level, the level of ULK1 mRNA was significantly increased, suggesting that some other mechanism of the regulation of ULK1 protein expression during starvation is downstream of transcription (Fig. 1C). In all studied cell lines prolonged nitrogen deprivation (up to 16 h) induced an approximately 2–7% increase in cell death (compared to cells grown in complete medium), signifying the importance of understanding the role of autophagy regulation under such prolonged cellular stress conditions (Fig. 1D, Fig. S1).

Figure 1.

ULK1 is rapidly reduced under starvation conditions. (A,B) Different cell lines were starved for the indicated time points in HBSS medium and ULK1 complex components were detected in cell lysates by immunoblotting. (C) Expression of ULK1 mRNA in U1810 and A549 cells starved (8 h) in HBSS medium. (D) Detection of cell death in U1810, A549, HEK293 and 143B cell lines starved (16 h) in HBSS medium. Cells were stained with ANXA5 (AV) and propidium iodide (PI) and cell death was detected by FACS analysis as described in Materials and methods. The cells that are negative for ANXA5 and PtdIns are shown as live cells. Statistical significance: *p < 0.05. a.u., arbitrary units; Ctr, control.

Based on several previous reports,^14-16 the ULK1 protein could be degraded either as a part of the autophagosomal membrane or in proteasomes. Therefore, we tested the effect of lysosomal and proteasomal degradation inhibitors bafilomycin A₁ and MG132 on stabilization of the ULK1 protein. Cell pretreatment (30 min) with either of these inhibitors had an effect on stabilization of ULK1 expression within the first hours of starvation (Fig. 2A). For better understanding the contribution of macroautophagy to ULK1 degradation in the long term, the U1810 cell line with a knockout of ATG7 was established using the CRISPR system (Fig. 2B). Autophagy in these cells was efficiently blocked, as could be seen by the lack of MAP1LC3 lipidation and inhibition of its degradation during starvation. Irrespective of the status of ATG7, ULK1 protein levels were significantly reduced under conditions of prolonged nitrogen deprivation (Fig. 2B, Fig. S2A). Furthermore, we established an additional cell line with a knockout of ATG13, another ULK complex component (Fig. 2C, D). Similar to a previous report¹³ it was found that the ULK1 protein is significantly diminished in ATG13 knockout cells at the post-transcriptional level (Fig. 2D, Fig. S2B), and the ULK1 expression was further reduced in ATG13 knockouts in response to nitrogen deprivation (8 h, HBSS). Additionally, using the A549 cell line we showed that inhibition of lysosomal degradation with bafilomycin A₁ efficiently blocked autophagic flux and stabilized MAP1LC3 and SQSTM1 proteins for up to 24 h of cell starvation in HBSS medium. However, the level of ULK1 was robustly reduced under prolonged starvation, indicating the involvement of an additional mechanism(s)/degradation pathway(s) that regulate its expression and could have a significant impact on the process of autophagy (Fig. 2E).

Figure 2.

Degradation of ULK1 during starvation. (A) Cells were pretreated (30 min) with inhibitors of lysosomal (bafilomycin A₁[Baf A], 250 nM) or proteasomal (MG132 [MG], 5 μM) degradation followed by 4 h starvation in HBSS medium. Immunoblotting was performed to detect ULK1 expression. (B, C, D) Reduction of ULK1 protein expression in cells with knockout of ATG7 (B) or ATG13 (C, D) under starvation conditions as detected by immunoblotting. U1810 lung cancer cells were starved or not as indicated in the figure. The numbers under the blots correspond to densitometric analysis of ULK1 protein normalized to GAPDH. (E) A549 cells were starved (24 h) in HBSS medium with or without bafilomycin A₁, and autophagy-related proteins were detected by immunoblotting. The level of MAP1LC3 and SQSTM1 shows the efficiency of blockage of the autophagic flux. (F) Wild type, ATG7 and ATG13 knockout U1810 cells were starved in HBSS medium (8 h) with or without inhibitors of either proteasomal (MG132, 5 μM) or lysosomal (bafilomycin A₁, 250 nM) degradation. Immunoblotting was performed to detect ULK1 expression. Antibodies to SQSTM1 and ubiquitin were used to prove the efficiency of the inhibition of autophagic flux in ATG7 and ATG13 knockout cells, and proteasomal degradation, respectively. (G) Cleavage of ULK1 under starvation conditions. A549 or U1810 cells were starved (8 h) and ULK1 cleavage was detected by immunoblotting.

Furthermore, the effect of inhibition of proteasomal degradation on ULK1 stability was studied using the proteasomal inhibitor MG132 (Fig. 2F). The efficiency of inhibition of proteasomal degradation was confirmed by detecting an accumulation of ubiquitinated proteins, and lysosomal degradation by checking degradation of SQSTM1/p62. Notably, the level of ULK1 protein expression was not upregulated in HBSS medium by inhibiting proteasomal degradation in ATG7 or ATG13 knockout cells, further indicating the action of some additional mechanism(s) of regulation of ULK1 protein expression under starvation conditions. Under starvation the reduction of ULK1 was even more pronounced in both knockout cells, suggesting that some other degradation mechanism(s) could compensate for autophagic degradation when autophagy is suppressed over a prolonged period of time. Furthermore, a cleaved fragment of ULK1 was detected during nitrogen deprivation, suggesting the involvement of some protease in this degradation process (Fig. 2G). Because the level of ULK1 mRNA was upregulated in response to starvation but the protein level was not increased by inhibiting either lysosomal or proteasomal degradation pathways, it was further suggested that in addition to protein degradation some other factor such as protein translation is contributing to its expression under starvation conditions.

Reduction of ULK1 protein during starvation is associated with inhibition of protein translation

Using A549 and U1810 lung adenocarcinoma cell lines, the protein translation was measured in cells starved in HBSS medium and compared with the effect of the translation inhibitors anisomycin and cycloheximide by detecting the incorporation of puromycin. Nitrogen deprivation had a rapid and pronounced effect on the inhibition of proteosynthesis, which was also associated with the reduced phosphorylation of RPS6KB/p70S6K, a substrate of MTOR (Fig. 3A). In contrast, the MTOR activity was increased in cells treated with cycloheximide or anisomycin, suggesting that unused accumulated amino acids activated the MTOR complex. To study whether the ULK1 protein was the one for which the levels were specifically reduced under starvation conditions, the expression of several other ATG proteins was analyzed in immunoblotting (Fig. 3B). Although some of the ATG proteins were reduced during starvation, potentially via a lysosomal degradation pathway, ULK1 showed a significant reduction in all conditions of blocked protein translation. Moreover, inhibition of protein translation with the MTOR inhibitor Torin1 had a similar effect on ULK1 protein expression (Fig. 3C). Furthermore, cell starvation in HBSS medium did not have a substantial additive effect on the reduction of ULK1 protein expression in combination with cycloheximide, further suggesting the involvement of similar mechanisms for the regulation of ULK1 expression under both conditions (Fig. 3D).

Figure 3.

Inhibition of protein translation contributes to rapid ULK1 downregulation in lung cancer cells under starvation conditions in HBSS. (A) U1810 and A549 lung cancer cell lines were treated with the inhibitors of proteosynthesis, cycloheximide (CHX) or anisomycin (AM), or starved for 2–8 h in HBSS medium. Protein translation was assessed by measuring puromycin incorporation by immunoblotting. Ponceau S staining was used as a marker for equal loading. (B) Expression of ATG proteins in U1810 and A549 cells treated for 2–8 h with inhibitors of protein translation or starved in HBSS medium was detected by immunoblotting. (C) A549 cells were treated (8 or 24 h) with the MTOR inhibitor Torin1 (250 nM). ULK1 expression was detected by immunoblotting. (D) A549 cells were starved (4 h) in HBSS medium with or without cycloheximide (20 μg/ml). ULK1 expression was detected by immunobloting. (E, F) Protein translation and ULK1 protein expression is rapidly restored after replenishment of HBSS with a complete medium. U1810 cells were grown in complete medium, starved in HBSS medium (8 h) or starved in HBSS medium followed by 1 h of incubation in complete medium. Protein translation was assessed by measuring puromycin incorporation by immunoblotting. (G) Expression of ULK1 mRNA in U1810 cells starved in HBSS medium (8 h) or starved in HBSS medium followed by 1 h of incubation in complete medium. (H) U1810 cells were starved overnight and polysome profiling was performed according to the protocol described in the Materials and methods. (I) Distribution of ULK1 mRNA between 3 fractions obtained after centrifugation in a sucrose gradient. Expression of mRNA was measured by qPCR and RNA18S RNA was used for normalization of ULK1 mRNA expression in different fractions. Ctr, control; Starv, HBSS medium.

To further confirm the dependence of ULK1 protein expression on translation under starvation conditions, U1810 cells were starved for 8 h in HBSS medium with subsequent incubation for 1 h in complete medium (Fig. 3E and F). As can be seen, protein translation was blocked and then rapidly restored when replacing HBSS with a fresh complete medium. This restoration was associated with a dramatic increase in the level of the ULK1 protein, despite the reduction of the level of ULK1 mRNA after supplementation with growth factors and amino acids (Fig. 3F and G), further suggesting a significant contribution of protein translation to the regulation of the level of ULK1 in cells under starvation conditions. Furthermore, we performed polysome profiling to determine the distribution of ULK1 mRNA among different fractions (Fig 3H and I). For qPCR, mRNA was prepared from 3 different fractions. Thus, fraction 1 contained the ribosomal subunits and light polysomes, fraction 2 consisted primarily of light polysomes, and fraction 3 of heavy polysomes. Based on the obtained results, there was a significant change in the distribution of ULK1 mRNA among heavy and light fractions of polysomes, with a shift of ULK1 mRNA from fraction 3 (heavy polysomes) to fraction 2 (light polysomes) under starvation conditions in comparison to fed conditions (Fig 3I). Taken together, these data demonstrate that although upregulation of ATG genes under starvation conditions could lead to amplification of the autophagy process, inhibition of the downstream process of protein translation by starvation contributes to a limiting of autophagy. Thus, in addition to degradation systems, inhibition of general protein translation has a further significant impact on the reduction of ULK1 protein expression during starvation.

Upregulation of ULK1 in response to inhibition of mitochondrial respiration is diminished in cells under starvation conditions

Previously, it has been established that ULK1 is required for mitophagy upon nutrient deprivation.¹¹ Due to the deficiency of substrates, starvation conditions are usually accompanied by reduced mitochondria activities that potentially could lead to upregulation of ULK1 mRNA. Indeed, using U1810 lung adenocarcinoma cells we found that the ULK1 level was significantly increased in response to inhibition of mitochondrial respiratory complexes. Thus, cell treatments with inhibitors of the mitochondrial complex I (rotenone), complex II (thenoyltrifluoroacetone/TTFA) or complex III (antimycin A or myxothiazol) were able to increase the level of the ULK1 protein. It is known that inhibition of mitochondrial respiration might cause energetic stress leading to the activation of AMPK, which could subsequently promote the activation of autophagy. Indeed, upregulation of ULK1 in cells was associated with an increase of AMPK phosphorylation (Fig. 4A). Similar to the nitrogen deprivation conditions, ULK1 mRNA was significantly upregulated in cells treated with antimycin A (Fig. 4B). Furthermore, it was found that inhibition of mitochondrial ATP synthase with oligomycin caused a robust upregulation of ULK1, and such effect was blocked by the AMPK inhibitor, compound C (Fig. 4C). Similar to antimycin A, oligomycin caused upregulation of ULK1 at the level of transcription, which was blocked by compound C (Fig. 4D). Furthermore, using the siRNA approach, a reduction of ULK1 in U1810 cells with silenced AMPK was found, suggesting a significant contribution of AMPK to the expression of ULK1 (Fig. 4E). The effect of antimycin A on ULK1 was less pronounced in cells with silenced AMPK, further signifying the importance of AMPK in ULK1 protein upregulation (Fig. 4F).

Figure 4.

Inhibition of mitochondrial respiratory chain complexes or mitochondrial ATP-synthase leads to transcriptional upregulation of ULK1. (A) U1810 lung cancer cells were treated (8 h) with the inhibitors of mitochondrial respiration antimycin A (Ant A; 5 μM), myxothiazol (Myx; 2 μM), TTFA (50 mM) or rotenone (Rot; 250 nM), and the levels of the ULK1 and AMPK proteins were detected by immunoblotting. (B) Upregulation of ULK1 mRNA level in cells treated (8 h) with antimycin A was detected by quantitative PCR. (C) Inhibition of mitochondrial ATP-synthase with oligomycin leads to upregulation of ULK1. U1810 cells were treated (24 h) with oligomycin (OM; 5 μg/ml), compound C (Comp C; 10 μM) or their combination. Expression of ULK1 and phosphorylation of AMPK were detected by immunoblotting. (D) Compound C blocks transcriptional upregulation of ULK1 in U1810 cells treated with oligomycin. U1810 cells were treated (8 h) with oligomycin (5 μg/ml), compound C (10 μM) or their combination, and the level of ULK1 mRNA was detected by immunoblotting. (E) Silencing of AMPK is associated with a reduction of ULK1 protein expression. U1810 cells were transfected with siRNA targeting AMPK and the level of protein expression was detected by immunoblotting. (F) AMPK is involved in upregulation of ULK1 in cells treated with antimycin A. AMPK was silenced using siRNA and cells were treated with antimycin A (8 h, 5 μM). The expression of proteins was detected by immunoblotting. Statistical significance: *p < 0.05. Ctr, control.

To check whether an increase of ULK1 was associated with mitophagy activation, U1810 cells were treated with activators of mitophagy; CCCP, antimycin A or oligomycin (Fig. 5A). Based on the colocalization profile, CCCP dramatically increased the level of mitophagy; however, in contrast to oligomycin and antimycin A, it did not have a significant effect on upregulation of ULK1 (Fig. 5B), suggesting that the effect of oligomycin and antimycin A on ULK1 upregulation was independent of mitophagy. To further prove that the effect of mitochondria inhibitors on ULK1 expression was not due to protein stabilization but rather upregulation of its mRNA, protein stability was checked in cells pretreated with antimycin A followed by treatment with cycloheximide (Fig. 5C, D). Inhibition of protein translation caused a dramatic decrease of ULK1 protein, suggesting that RNA transcription rather than protein stabilization was required for upregulation of ULK1 in cells with repressed mitochondrial respiration. Furthermore, similar to treatment with cycloheximide, under starvation conditions the level of ULK1 was rapidly reduced (Fig. 5E). We also attempted to restore ULK1 protein expression under starvation conditions using treatment with antimycin A. Interestingly, an increase of the ULK1 protein in response to antimycin A treatment was observed in both control and ATG13 knockout cells (Fig. 5F, Fig. S2C), the latter being cells where the ULK1 protein level was already significantly diminished (Fig. 2D). Although the level of ULK1 mRNA was upregulated, we did not observe an increase of the ULK1 protein in response to antimycin A under starvation conditions, further signifying the importance of protein synthesis. Taken together, inhibition of mitochondrial respiration (that also takes place under starvation conditions) contributes to the upregulation of ULK1 mRNA; however, it is not sufficient to upregulate ULK1 protein during starvation.

Figure 5.

Upregulation of ULK1 in U1810 lung cancer cells is blocked under starvation conditions. (A) U1810 cells were treated (24 h) with the mitochondria uncoupler CCCP (10 μM), antimycin A (Ant A; 5 μM) or oligomycin (OM; 5 μg/ml), and mitophagy was detected by confocal microscopy using staining of lysosomes with LysoTracker Red and mitochondria with MitoTracker Green. Colocalization profiles show a significant activation of mitophagy induced with CCCP but not antimycin A or oligomycin treatments. (B) Expression of ULK1 in U1810 cells treated (24 h) with oligomycin, CCCP or antimycin A was detected by immunoblotting. (C, D) Upregulation of de novo protein synthesis rather than stabilization of ULK1 takes place in cells treated with antimycin A. U1810 cells were untreated or treated for 24 h with antimycin A (5 μM) followed by treatment for the indicated time points with the proteosynthesis inhibitor cycloheximide (CHX; 20 μg/ml). ULK1 expression was detected by immunoblotting and the fold changes of protein expression were quantified by densitometry. (E) ULK1 protein levels are rapidly reduced in U1810 cells starved in HBSS medium. U1810 cells were treated with antimycin A (24 h, 5 μM) and then either starved in HBSS (4 h) or treated with cycloheximide (4 h, 20 μg/ml). ULK1 expression was detected by immunoblotting. (F) Upregulation of ULK1 took place in ATG13-knockout U1810 cells treated with antimycin A (24 h, 5 μM); however, there was no upregulation under cell starvation (24 h). AMPK phosphorylation and ULK1 protein expression was detected by immunoblotting. Statistical significance: *p < 0.05. a.u., arbitrary units.

ULK1 is required for basal autophagy and autophagic flux under starvation conditions

Previous studies revealed the importance of ULK1 for autophagy under short-term starvation.^18,19 Using siRNA-mediated knockdown of genes encoding ULK complex components including ULK1, ATG13 and RB1CC1 as well as siRNA-targeting the gene for the autophagy receptor protein SQSTM1, we confirmed that all 3 ULK complex components have a comparable effect on basal autophagic flux (Fig. 6A, Fig. S3A). To further confirm the importance of ULK1 for autophagy under starvation conditions, U1810 cells with either overexpressed or silenced ULK1 were starved for 4–8 h and the autophagic flux was assessed by measuring the ratio of MAP1LC3-II:MAP1LC3-I or degradation of MAP1LC3 and SQSTM1 via immunoblotting (Fig. 6B, C). A decrease of lipidation and degradation of MAP1LC3 was observed in cells with silenced ULK1, indicating the importance of the ULK1 protein for autophagic lipidation and degradation of MAP1LC3 during starvation. In addition, the degradation of SQSTM1 was also partially reduced. Staining with MAP1LC3 antibody revealed the reduced formation of autophagosomes in cells with silenced ULK1, signifying the importance of ULK1 for autophagy under starvation conditions in HBSS (Fig. S3B). Furthermore, the autophagic flux was measured in U1810 and A549 cell lines using an inhibitor of lysosomal degradation, bafilomycin A₁. Under prolonged starvation both SQSTM1 and MAP1LC3 proteins were significantly reduced, and much less accumulation of MAP1LC3-II and SQSTM1 under prolonged starvation further confirmed a decreased autophagic flux (Fig. 6D, E). Taken together, these data indicate the importance of ULK1 for basal and nitrogen deprivation-induced autophagy and demonstrate that the reduction of ULK1 expression is one of the factors that limit autophagy under conditions of prolonged starvation. A proposed general model for regulation of ULK1 expression under starvation conditions and its contribution to autophagy is presented in Fig. 7.

Figure 6.

ULK1 is required for basal and induced autophagic flux under starvation conditions. (A) Silencing of ULK complex components (ULK1, RB1CC1 and ATG13) inhibits the basal autophagic flux. To assess the level of autophagy, the expression of SQSTM1 was analyzed by immunoblotting. (B) Overexpression of HA-ULK1 increases the basal and induced autophagic flux. Autophagic flux was assessed by adding bafilomycin A₁ (Baf A; 50 nM, 2 h) before cells were collected for immunoblotting. (C) Silencing of ULK1 reduces the degradation of MAP1LC3 and SQSTM1 under starvation conditions. ULK1 was silenced using siRNA and the expression of SQSTM1 and MAP1LC3 were detected by immunoblotting. The numbers below the blots correspond to densitometric analysis of SQSTM1 and MAP1LC3-II proteins normalized to GAPDH. (D, E) U1810 and A549 lung cancer cell lines were starved for the indicated time points in HBSS medium. Autophagic flux was assessed by adding bafilomycin A₁ (50 nM, 2 h) before cells were collected for immunoblotting. ctr, control.

Figure 7.

A proposed model for autophagy limitation under the condition of nitrogen deprivation. ULK1 is activated under starvation, leading to initiation of autophagy signaling. Lack of amino acids significantly diminishes general protein translation, contributing to the reduction of protein expression. Several degradation pathways are involved in elimination of ULK1. Therefore, activated autophagy serves as a negative feedback loop in regulation of ULK1 protein expression. In addition, proteasomal and some yet-unidentified proteolytic system(s) are involved in ULK1 degradation during starvation. Thus, acting together, both protein translation and degradation systems contribute to the regulation of the level of ULK1 expression, suggesting a negative loop for the limitation of autophagy signaling under the condition of nitrogen deprivation.

Discussion

Here we suggest a new model for autophagy regulation during starvation, a condition that is widely used to activate autophagy in various in vitro and in vivo models. We identified the ULK1 protein as one of the autophagy components that is rapidly reduced in response to growth factor and amino acid deprivation. Although in some cellular systems ULK1 is degraded by autophagy, we demonstrated that in the long term, contribution of the autophagic/lysosomal degradation pathway is negligible. Moreover, we revealed that the reduction of the ULK1 protein under starvation conditions in HBSS also takes place in cells with inhibited proteasomal degradation, suggesting the involvement of some additional proteolytic pathway(s) in the degradation of the ULK1 protein under nitrogen-deprivation conditions. Furthermore, we proposed that inhibition of protein translation, in addition to these degradation pathways, contributes to the level of ULK1 expression under starvation conditions.

Several recent publications have reported the transcriptional regulation of autophagy under starvation conditions.^1-4 However, not all of these studies have been focused on the effects downstream of mRNA transcription. Potentially, upregulation of transcription factor activities contributes to an increase in the mRNA level of respective ATG genes during starvation; however, many other factors may control autophagy at the posttranscriptional level under different starvation conditions. Herein we demonstrate that in addition to protein modification and degradation, autophagy is attenuated under conditions of prolonged starvation by the inhibition of protein translation. In particular, we suggest that this could be due to the reduction of ULK1 protein expression, a protein that has a lower stability than several other studied ATG proteins under starvation conditions. However, inhibition of protein translation associated with prolonged starvation can potentially affect other autophagy-related proteins, which might also have an impact on the formation of autophagic vesicles. Because in many cellular systems, autophagy is induced by specific inhibition of the MTOR complex, which is required for cap-dependent protein translation, it can be proposed that such an effect may also make some contribution to the expression of ATG proteins and autophagy limitation during long-term star-vation.

It should be noted that although ULK1 is important for autophagy activity in different cells and under different conditions, there are some alternative ways to induce autophagic degradation independently from ULK1. For example, the MAP1LC3 protein is susceptible to lysosomal degradation under conditions of glucose starvation when the MTOR complex remains active, and in these conditions the autophagy pathway is independent of ULK1.^20,21 However, as has been previously reported, in contrast to glucose starvation, ULK1 is required for autophagy under the condition of amino acid deprivation. Such autophagy activation during amino acid deprivation is mediated by the interaction between the ULK1 and ATG5 complexes in a short RB1CC1/FIP200-binding domain of ATG16L1.²² The HBSS medium used in our experimental system for autophagy induction contained glucose but was lacking growth factors and amino acids. Similar to the previous reports, in this condition ULK1 was essential for autophagic degradation of MAP1LC3 and to some extent for degradation of SQSTM1 proteins. We suggest that the decrease of the level of the ULK1 protein is one of the factors reducing autophagic degradation under conditions of amino acid and growth factor deprivation. Thus, inhibition of protein synthesis as a response to starvation will lead to a rapid ULK1 reduction and limit the ULK1-dependent autophagic degradation.

In addition to autophagy, several other autophagy-independent functions of the ULK1 protein have been reported, thus opening further perspectives to investigate the connection of autophagy functions with the other cellular processes via regulation of ULK1 protein expression. Although AMPK and MTOR regulate autophagy by phosphorylating ULK1, the latter can, in turn, phosphorylate both kinases and therefore serves as a feedback loop in regulating their activities, suggesting some other important functions of the ULK1 protein in addition to autophagy regulation. Thus, phosphorylation of RPTOR by ULK1 negatively regulates MTOR complex 1²³ and has also been proposed as a mechanism connecting ULK1 activity and cell proliferation.²⁴ Similarly, ULK1/2 phosphorylate all 3 subunits of AMPK and negatively regulate its activity, thereby terminating signaling that initiates autophagy.²⁵ If so, based on our results, it could be suggested that downregulation of ULK1 under starvation conditions could lead to more efficient activation of AMPK.

In summary, here we report that amplification of ULK1-dependent autophagy is reduced under starvation conditions and we suggest a potential contribution of protein translation inhibition in addition to degradation systems. Further studies are required to understand whether or not such autophagy control during starvation may contribute to the inhibition of cell death via autophagy.

Materials and methods

Reagents and antibodies

Mitochondrial inhibitors: Antimycin A (Sigma-Aldrich, A8674), myxothiazol (Sigma-Aldrich, T5580), TTFA (Sigma-Aldrich, T27006), rotenone (Sigma-Aldrich, R8875). Proteasomal inhibitor: MG132 (Enzo Life Sciences, BML-PI102). Proteosynthesis inhibitors: Cycloheximide (Sigma-Aldrich, c6255), anisomycin (Sigma-Aldrich, A9789). Lysosomal inhibitor: Bafilo-mycin A₁ (LC laboratories, B-1080). AMPK inhibitor: Compound C (Sigma-Aldrich, P5499). Antibodies: Rabbit anti-ULK1 (Cell Signaling Technology, 4773), anti-ATG13 (Cell Signaling Technology, 6940), anti-RB1CC1/FIP200 (Sigma-Aldrich, SAB4-200135), anti-MAP1LC3 (MBL International, PM036), anti-phosphorylated (p)-ULK1 (S757; Cell Signaling Technology, 6888), anti-p-ULK1 (S555; Cell Signaling Technology, 5869), anti-ATG14 (Novus Biologicals, NBP1-88877), anti-ubiquitin (Santa Cruz Biotechnology, sc-8017), anti-ATG5 (Cell Signaling Technology, 2630), anti-ATG4B (MBL International, M134-3), anti-ATG3 (MBL International, M133-3), anti-BECN1/Beclin 1 (Cell Signaling Technology, 3738), anti-ATG7 (Cell Signaling Technology, 2631), anti-AMPK (Cell Signaling Technology, 5832), anti-p-PRKAA/AMPKα (Y172; Cell Signaling Technology, 2535), anti-GAPDH (Trevigen, 2275-PC-100), anti-ACTA1/actin (Sigma-Aldrich, A2066), anti-TUBA4A/tubulin (Sigma-Aldrich, T6074), anti-TP53/p53 (Santa Cruz Biotechnology, SC-126), anti-SQSTM1 (Santa Cruz Biotechnology, 28359), anti-puromycin (Merk Millipore, MABE343), anti-AKT (Cell Signaling Technology, 9272), anti-p-AKT (Cell Signaling Technology, 4056). Secondary antibodies: Anti-mouse IgG (Pierce, 31430), anti-rabbit IgG (Pierce, 31460).

Cell culture

Human lung carcinoma A549 (ATCC, CCL­185), HEK293 embryonic kidney cells (ATCC, CRL-1573), 143B bone osteosarcoma cell line (ATCC, CRL­8303) and U1810 lung carcinoma (from the collection at Uppsala University, Sweden) were cultured in RPMI 1640 medium (Gibco, Thermo Fisher Scientific, 52400-041) supplemented with 10% (v/v) fetal bovine serum (Gibco, Thermo Fisher Scientific, 10270), 100 U/ml penicillin and 100 mg/ml streptomycin (Gibco, Thermo Fisher Scientific, P0781) and maintained at 37°C with 5% CO₂ in an air atmosphere. Starvation conditions were achieved by culturing cells in HBSS medium (Gibco, Thermo Fisher Scientific, 24020-133) for the indicated hours.

Construction of lentiCRISPR_ATG7 and lentiCRISPR_ATG13 knockout plasmids

ATG7 and ATG13 knockout cell lines were generated by adapting a CRISPR-Cas9 system using lentiCRISPR v1 vector (Addgene, 49535; deposited by Dr. Feng Zhang) according to the protocol of Dr. Feng Zhang.²⁶ ATG7 and ATG13 gRNA oligos of 20 base pair nucleotides were designed by excluding potential off-target effects using the following databases: http://crispr.mit.edu/ developed by the Zhang lab in 2013, and http://blast.ncbi.nlm.nih.gov/Blast.cgi. The following sequences of primers were used: ATG7_KO1 FW: CACCGGAAGCTGAACGAGTATCGGC; RW: AAAC-GCCGATACTCGTTCAGCTTCC; ATG7_KO2 FW: CACC-GGCTGCCAGCTCGCTTAACAT; RW: AAACATGTTAAG-CGAGCTGGCAGCC; ATG13_KO1 FW: CACCGGTGATTG-TCCAGGCTCGGCT; RW: AAACAGCCGAGCCTGGACAA-TCACC; ATG13_KO2 FW CACCGGATTTCACTTAAGAC-TTCTG; RW: AAACCAGAAGTCTTAAGTGAAATCC. The sequences of purified plasmids, along with a self-ligated empty vector as a negative control, were validated using LKO.1 5′ primer in the KIGene sequencing service at Karolinska Institutet. HEK293 cells were used to produce viruses. Corresponding cell lines were infected with viruses and knockout cells were selected using puromycin treatment.

Overexpression of ULK1

Cells were transfected with the HA-hULK1 plasmid encoding human ULK1 by using Lipofectamine LTX with plus reagent (Invitrogen, 15338-100) according to the manufacturer's instructions. Twenty-four h after transfection, medium was exchanged and the cells were starved in HBSS medium (8 h) with or without bafilomycin A₁. The HA-hULK1 plasmid was obtained from Addgene (31963; deposited by Dr. Do-Hyung Kim).²⁷

Small interfering RNA (siRNA) transfection

Cells were seeded in 6-well plates and in 24 h transfected with INTERFERin siRNA Transfection Reagent (Polyplus-transfection, 409). For each transfection, siRNAs (si_ULK1 s15965, si_ATG13 s18881, si_FIP200 s18995, si_SQSTM1 s16961, Silencer Select, Life Technologies) were mixed with 3 μl of INTERFERin in OPTI-MEM medium (Gibco, Thermo Fisher Scientific, 51985-026). After a 10-min incubation at room temperature (RT) the complexes were added to the cells. The final concentration of siRNA in the medium was 20 nM. Medium was replaced and treatments were administered 24 h after transfection.

Immunoblotting

Cells were collected using trypsin (Sigma-Aldrich, 52400-041), washed with phosphate-buffered saline (PBS; Sigma-Aldrich, P3813) and lysed using a complete Lysis-M buffer (Roche, 04719956001), supplemented with protease inhibitor (Roche, 05892970001) and phosphatase inhibitor (Roche, 049068-37001) cocktails. Protein concentrations were determined using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific, 23228) according to the manufacturer's instructions. Then, samples were mixed with Laemmli buffer and boiled for 5 min at 98°C. Equal amounts of proteins were separated by SDS-PAGE and blotted onto nitrocellulose membranes (Bio-Rad, 1620112). Membranes, blocked for 1 h with 5% skimmed milk in PBS, were incubated overnight with primary antibodies diluted in PBS (Sigma-Aldrich, D8537) containing 2% BSA (Sigma-Aldrich, A4503) and 0.05% Tween-20 (Sigma-Aldrich, P1379). The recognized proteins were detected using horseradish peroxidase-labeled secondary antibodies anti-mouse IgG and anti-rabbit IgG, and the enhanced chemiluminescence Clarity Western ECL Substrate (Bio-Rad, 1705061).

Measurement of proteosynthesis

A nonradioactive method based on the incorporation of puromycin was used to monitor protein synthesis.²⁸ Briefly, puromycin (1 μg/ml) was added for 10 min to the medium of cultured cells. Then, cells were washed twice with PBS and collected with trypsin. The proteosynthesis level was assessed by immunoblot using anti-puromycin antibody.

Real-time quantitative PCR

RNA was extracted using an RNeasy mini kit (Qiagen, 74106). One µg of total RNA extract was reverse-transcribed using a Transcriptor First Strand cDNA Synthesis Kit (Roche, 04379012001) according to the manufacturer's instructions. Gene expression levels were assessed in a 7500 Real-Time PCR System (Applied Biosystems, Stockholm, Sweden) on 10 ng of cDNA mixed with FastStart Universal SYBR Green Master (Rox; Roche, 04913914001) and 100 nM of primers. The reaction mixtures were subjected to an initial denaturation step, at 95°C for 10 min, and 40 cycles of amplification, each cycle consisting of a denaturation step at 95°C for 15 s and an annealing/extension step at 60°C for 1 min. A melting curve analysis was used to confirm primer specificity and to ensure the absence of primer-dimer formation. The relative expression levels of each gene are presented as the fold increase relative to untreated cells after normalization against GAPDH, HPRT1 and ACTB. The following sets of primers from Life Technologies were used: ULK1 (FW: ACGACTTCCAGGAAATGGCTA, RW: GGAAGAGCCTGATGGTGTCCT), ACTB (FW: GGA-CTTCGAGCAAGAGATGG, RW: AGCACTGTGTTGGCGTACAG), MAP1LC3B (FW: GAACGATACAAGGGTGAGA AGC, RW: TACACCTCTGAGATTGGTGTGG), HPRT (FW: AATTATGGACAGGACTGAACGTCTTGCT, RW: TCCAGCAGGTCAGCAAAGAATTTATAGC); GAPDH (FW: GTCA-ACGGATTTGGTCGTATT, RW: AGTCTTCTGGGTGGCA-GTGAT) and 18S (FW: GTCGCTACTACCGATTGGATG, RW: CAAGTTCGACCGTCTTCTCAG)

Immunohistochemistry

Cells were seeded and grown on coverslips in 6-well plates. Twenty-four h after transfection with the desired siRNAs, cells were fixed in 4% paraformaldehyde (15 min, RT). Afterwards coverslips were handled as reported previously.²⁹ Briefly, cells were permeabilized with digitonin (100 µg/µl; Sigma-Aldrich, D141) and stained with anti-LC3 or anti-SQSTM1 antibodies diluted in PBS-2% BSA. The next day, the slides were incubated (1 h, RT) with Alexa Fluor 488-conjugated donkey-anti-mouse (Molecular Probes, A21206; 1:500) or donkey-anti-rabbit (Molecular Probes, A21203; 1:500) secondary antibodies. Nuclei were counterstained with Hoechst 33342 (10 μg/ml; Molecular Probes, H3570) and the slides were mounted in Vectashield mounting medium (Vector Laboratories, H-1000) and examined under a Zeiss LSM 510 META confocal laser scanner microscope (Carl Zeiss, Jena, Germany).

Detection of cell death using ANXA5/annexin V-propidium iodide (PI) staining

ANXA5-PtdIns double staining was carried out using an Annexin-V-FLUOS Staining Kit (Roche Applied Science, 11988549001) according to the manufacturer's protocol. The cells were analyzed by flow cytometry (BD Accuri, Becton Dickinson) and the data were evaluated using BD CSampler software.

Polysome profiling

Cells were seeded and grown in 15-cm dishes until 80–90% confluent and then treated with cycloheximide (100 µg/µl, 5 min), washed with PBS + cycloheximide, and collected. Cells were resuspended in 0.425 mL of hypotonic buffer (5 mM Tris-HCl, pH 7.5, 2.5  mM MgCl₂, 1.5 mM KCl, 100 μg ml⁻¹ cycloheximide, 2 mM DTT, 100 U ml⁻¹ RNAsin (Promega, N2511), 1x protease inhibitor cocktail EDTA-free [Roche, 11836-170001]). Subsequently, final concentrations of 0.5% Triton X-100 (Sigma-Aldrich, X100) and sodium deoxycholate (Sigma-Aldrich, D6750) were added, and the tubes were briefly vortexed then centrifuged for 5 min at 13,000 rpm (16,000 g). Samples were adjusted using hypotonic buffer buffer and equal OD₂₅₄ amounts from each sample were loaded onto 10–50% linear sucrose (Sigma-Aldrich, S7903) gradients and centrifuged at 171,000 g (50,000 rpm) using a Beckman L-60 ultracentrifuge (Optima, USA) with a type 70.1Ti rotor for 2 h at 4 °C. Gradients were fractionated from the top in 48 fractions, 200 μl each, and RNA was measured manually by NanoDrop 1000 spectrophotometer (Thermo Fisher Scientific). Next, several OD fractions were combined and RNA was extracted using TRIzol Reagent (Invitrogen, 15596-026) according to the manufacturer's protocol. Isolated RNA was resuspended in RNase-free water and 10 μg from each fraction reverse transcribed using SuperScript III Reverse Transcriptase (Invitrogen, 18080-044) according to the manufacturer's protocol. The cDNA was then used for real-time quantitative PCR as previously described. The relative expression of RNA18S/18S RNA was used for normalization and calculations of the distribution of ULK1 mRNA expression among the different fractions.

Statistical analysis

All data are presented as the means of triplicate assays ± s.d. One-tailed t tests were used to determine whether differences were statistically significant. Values of p < 0.05 were considered significant.

Abbreviations

AMPK

AMP-activated protein kinase

PBS

phosphate-buffered saline

RT

room temperature

TTFA

thenoyltrifluoroacetone

ULK1

unc-51 like autophagy activating kinase 1.

Disclosure of potential conflicts of interest

No potential conflicts of interest were disclosed.

Acknowledgments

The authors thank Vladimir Gogvadze for his helpful discussions, Alexander Sellgren and Eduarda Lopes for providing technical assistance, and Belen Espinosa for help in performing experiments with ULK1 overexpression, presented in Fig. 6B.

Funding

This project was supported by grants from the Swedish and Stockholm Cancer Societies, the Swedish Research Council, the Swedish Childhood Cancer Foundation, Karolinska Institutets Forskningsstiftelser and the Elsa Goljes Foundation. BZ was supported by the Russian Science Foundation (14-25-00056). GA was supported by the Erasmus traineeship program.