ULK1 phosphorylates Sec23A and mediates autophagy-induced inhibition of ER-to-Golgi traffic

Abstract

Background: Autophagy is an inducible autodigestive process that allows cells to recycle proteins and other materials for survival during stress and nutrient deprived conditions. The kinase ULK1 is required to activate this process. ULK1 phosphorylates a number of target proteins and regulates many cellular processes including the early secretory pathway. Recently, ULK1 has been demonstrated to phosphorylate Sec16 and affects the transport of serotonin transporter at the ER exit sites (ERES), but whether ULK1 may affect the transport of other cargo proteins and general secretion has not been fully addressed.

Results: In this study, we identified Sec23A, a component of the COPII vesicle coat, as a target of ULK1 phosphorylation. Elevated autophagy, induced by amino acid starvation, rapamycin, or overexpression of ULK1 caused aggregation of the ERES, a region of the ER dedicated for the budding of COPII vesicles. Transport of cargo proteins was also inhibited under these conditions and was retained at the ERES. ULK1 phosphorylation of Sec23A reduced the interaction between Sec23A and Sec31A. We identified serine 207, serine 312 and threonine 405 on Sec23A as ULK1 phosphorylation sites. Among these residues, serine 207, when changed to phospho-deficient and phospho-mimicking mutants, most faithfully recapitulated the above-mentioned effects of ULK1 phospho-regulation.

Conclusion: These findings identify Sec23A as a new target of ULK1 and uncover a mechanism of coordinating intracellular protein transport and autophagy.

Keywords: Autophagy; COPII; ER exit sites; Sec23; ULK1.

Fig. 1

ULK1 interacts with Sec23A and affects ERES morphology. a Sec23A but not TRAPPC12 was pulled down by ULK1 in co-immunoprecipitation experiment. Myc-ULK1, FLAG-FIP200 were co-transfected with the indicated GFP fusion DNA constructs. Myc-ULK1 was immunoprecipitated with anti-c-Myc antibody and the presence of GFP-Sec23A or GFP-TRAPPC12 was determined. b ULK1 affects the morphology of ERES. Myc-ULK1 wild type or kinase dead (KD) mutant were co-transfected with GFP-Sec23A into Hela cells. ULK1 transfected cells were determined by immunostaining of TRITC conjugated anti-Myc antibody. ERES was detected with Sec31A staining. Asterisks in the Merge pictures indicate cells transfected with ULK1

Fig. 2

Autophagy changes the morphology of ERES. a Hela cells were cultured in growth medium or amino acid starved in EBSS to activate autophagy for indicated time. Cells were stained with ERES by anti-Sec31A antibody. Fluorescence intensity profiles of the ERES of the cells marked with asterisks were quantified. A line was drawn from a cell chosen in the image to generate fluorescence pixel intensity. b ERES in HEK293 (top) and COS cells (bottom) in complete growth medium or in EBSS (AA starvation). c Hela cells incubated with complete medium or EBSS for amino acid starvation for 0.5 or 2 h were stained with ERES (Sec31A, green in merge) and autophagosomes (LC3B, red in merge). d HEK293 cells were transfected with GFP-Sec23A (green). Transfected cells were cultured in growth medium or EBSS for 2 h before fixation and staining with Sec31A (red)

Fig. 3

Autophagy and ULK1 caused retention of transport marker VSVG in ERESs. a Hela cells were transfected with GFP-VSVG-tsO45 and the cells were kept at 39.5C overnight. Then the cells were cultured in (a) normal growth medium or (b) induced with autophagy by amino acid starvation for 2 h before the incubation temperature was shifted to 32 °C for the indicated time to chase the VSVG out of the ER. c Quantification of the ERES-associated VSVG signals. The fractions of VSVG signal in ERES at different time points after 32 °C chase were measured by Manders Coefficients. Number of cells ≥10; Error Bars = S.D. d Hela cells were transfected with GFP-VSV-G tsO45 and Myc-ULK1 or (e) its kinase dead mutant at 1:4 ratio and the cells were kept at 39.5 °C. After being chased at 32 °C for different times as indicated, the cells were fixed and stained with Sec31A for ERES. f Quantification of the amount of ERES-retained VSVG signal in ULK1 or kinase dead mutant. The fractions of VSVG signal in ERES at different time points after 32 °C chase were measured by Manders Coefficients. Number of cells ≥ 8; Error Bars = S.D., p < 0.05

Fig. 4

ULK1 phosphorylates Sec23 and changes its binding to other COPII components. a Myc-ULK1 WT was co-transfected with indicated COPII components (GFP-tag) pairwise in HEK293T cell. ULK1 was immunoprecipitated by anti-Myc antibody and the presence of co-precipitated COPII components were detected by blotting with anti-GFP antibody. b Myc-ULK1 and Myc-Sec23A were separately overexpressed and purified from HEK293T cells by immune-isolation using anti-Myc antibody. Some of ULK1 transfected cells were put under amino acid starvation to further activate the ULK1 kinase activity. Beads loaded with ULK1 and Sec23A protein were mixed to allow ULK1 to phosphorylate Sec23A in vitro. Phosphorylated Sec23a was detected by Mouse anti-serine/threonine antibody. Total Sec23a was detected by Rabbit anti-Sec23a antibody. c-e HEK293T cells were transfected with the indicated plasmids and treated by EBSS or complete medium prior to harvest. AA star. Means amino acid starvation. After harvest, Sec23A was immunoprecipitated and detected by using monoclonal anti-GFP antibody. Sec31A (c), Sar1b (d) and Sec24D (e) were detected by mouse anti-Myc antibody

Fig. 5

Identification of ULK1 phosphorylation sites on Sec23A. a Candidate ULK1 phosphorylation sites on Sec23A. b Various phospho-deficient mutants were tested for ULK1 phosphorylation. This experiment was similar to that in Fig. 3b with ULK1 activated, but in addition to wild type Sec23A, four alanine mutants of Sec23A were also used. c Phospho-deficient and phospho-mimicking mutants of Sec23A were analyzed for their morphology. Hela cells were transfected with wild type or indicated mutants of Myc-Sec23A and the Myc-Sec23A mutants were stained with anti-c-Myc antibody.Amino acid starvation was induced to cell for 2 h before fixation and staining (lower rows). Cell in the non-starved condition was cultured in complete growth medium before fixation (upper rows)

Fig. 6

The interactions between the ULK1 phosphorylation mutants of Sec23A and endogenous Sec31A. a S207 and T405 mutants of Sec23A were tested for their interaction with Sec31A in normal growth condition. The indicated Myc-His-Sec23A mutants were transfected into cells and tested for their ability to bring down endogenous Sec31A in co-IP experiment. b Wildtype Sec23A, S207A and T405A mutants were tested for the interaction with Sec31A in growth medium or in amino acid starved medium EBSS. c Combinations of S207 and T405 double mutants were tested for the interaction with Sec31A. Overexpression of the indicated protein was carried in HEK293T cells, followed by immunoprecipitation of Myc-His-Sec23A by anti-Myc antibody. Co-precipitated endogenous Sec31A was detected by anti-Sec31A antibody

Fig. 7

Expression of the Sec23A mutant compromised the integrity of ERGIC. Various indicated Sec23A mutants in DsRed vector were transfected to HeLa cells and the status of ERGIC was investigated by staining with marker ERGIC-53. DsRed and ERGIC-53 signals are pseudo-colored in red and green, respectively, in the merge panel (bottom)