ERK7 is a negative regulator of protein secretion in response to amino-acid starvation by modulating Sec16 membrane association

Abstract

RNAi screening for kinases regulating the functional organization of the early secretory pathway in Drosophila S2 cells has identified the atypical Mitotic-Associated Protein Kinase (MAPK) Extracellularly regulated kinase 7 (ERK7) as a new modulator. We found that ERK7 negatively regulates secretion in response to serum and amino-acid starvation, in both Drosophila and human cells. Under these conditions, ERK7 turnover through the proteasome is inhibited, and the resulting higher levels of this kinase lead to a modification in a site within the C-terminus of Sec16, a key ER exit site component. This post-translational modification elicits the cytoplasmic dispersion of Sec16 and the consequent disassembly of the ER exit sites, which in turn results in protein secretion inhibition. We found that ER exit site disassembly upon starvation is TOR complex 1 (TORC1) independent, showing that under nutrient stress conditions, cell growth is not only inhibited at the transcriptional and translational levels, but also independently at the level of secretion by inhibiting the membrane flow through the early secretory pathway. These results reveal the existence of new signalling circuits participating in the complex regulation of cell growth.

Figure 1

Overview of the microscopy-based RNAi kinase screen. The primary screen was performed in 384-well plates (in duplicate) using Fringe-GFP S2 cells and dsRNAs transcribed from the HFA library targeting 254 genes (245 kinases). The cells were immunolabelled with anti-GFP and anti α-tubulin antibodies as well as Hoechst and were viewed by widefield microscopy. Forty-three candidates (scored in the two plates or only in one) were identified, whereas 112 depletions did not lead to a Golgi phenotype. Fifty depletions led to an unclear phenotype because of a phenotypic discrepancy between the two plates examined (a Golgi phenotype was observed ‘in one plate only’). The phenotype of 49 depletions was ‘not determined’ because the data were not recorded properly (out of focus or lack of cells). The validation screen was performed using different dsRNAs transcribed from a second generation RNAi library (HD2) to target 30 out of 43 candidates. It was performed in Fringe-GFP S2 cells seeded in 24-well plates that were immunolabelled with Sec16/PDI/Dapi and viewed by confocal microscopy. In all, 26 out of 30 candidates were validated. The depletion phenotypes of 11 candidates were characterized (using a third set of independent dsRNAs) and 8 were cloned, localized and overexpressed leading to the identification of ERK7.

Figure 2

Examples of different phenotypic groups from the confirmation/validation screen. (A) Visualization of tER-Golgi units (Sec16 and Fringe-GFP, respectively) upon different RNAi depletions by confocal microscopy. Typical pattern of tER-Golgi units in mock-treated cells (−dsRNA). The very strong (+ds cdc2), strong (+ds sticky; +ds CG10177) and moderate (+ ds CG32703) MG phenotype (more and smaller Golgi spots) are presented as well as the LS phenotype (less spots, +ds wallenda/wnd). The pictures represent 2D projections of confocal sections. Scale bar: 5 μm. (B) The number of S2 cells after a 5-day incubation with the indicated dsRNAs expressed as percentage relative to the number of mock-treated cells. Red and blue columns indicate genes whose depletion led to a significant decrease or increase in cell proliferation, respectively. Error bars represent s.d. from at least three independent experiments. Conditions with P<0.01, 0.01<P<0.05 and 0.05<P<0.10 are indicated with triple, double and single asterisks, respectively. (C, D) Cell-cycle distribution of live S2 cells after 5 days incubation with the indicated dsRNAs determined by staining their DNA content. The population of G1 (M1), S/G2/M (M2) or sub-G1 (M3) cells in each condition was quantified by FACS analysis. Percentage of gated cells in S/G2/M phase (4N) (normalized to the respective value of mock-treated cells, which was considered as 100%) of one representative experiment (C). Red and blue columns indicate genes whose depletion leads to a significant decrease or increase in the percentage of cells in S/G2/M phase, respectively. cdc25 and myb depletions (n=3) lead to an average of 152.61%±7.85 (P-value of 0.010) and 144.62%±12.41 (P-value of 0.036), respectively. For CG10738 (#1 and #2) depletion (n=3), the average is 71.60%±7.12 with a P-value of 0.014. Representative examples are shown in (D). Note the increase in G1 population upon depletion of CG10738 kinase. (E) Efficiency of anterograde transport of Delta S2 cells incubated for 5 days with the indicated dsRNAs, followed by 1-h induction of Delta with CuSO₄ and 75 min chase to allow its transport to plasma membrane. Fixed cells were labelled for Delta and dGMAP (cis-Golgi marker). Scale bars: 5 μm.

Figure 3

ERK7 overexpression induces Sec16 dispersion and disassembly of tER sites. (A) IF localization of ERK7-V5 (green) and effect of its overexpression on Sec16 (red). Note that ERK7 is largely cytosolic but that its expression leads to Sec16 dispersion. (B) Overexpression of kinase-dead ERK7-V5^K54R and ERK7-V5^TDY>ADF (green) does not lead to Sec16 dispersion (red). 2D projections of confocal sections are presented in (A) and the first panel in (B). (C) Quantification of Sec16 dispersion upon expression of WT and K54R ERK7-V5. (D, E) Localization of Sec23 by Immunoelectron microscopy (IEM) in untransfected cells (D) and ERK7-V5 expressing cells (E). Note that the tER-Golgi units in (D) (between brackets) are largely absent in (E) (arrowheads), Sec23 (red circles) is dispersed and largely absent from the remnants of tER-Golgi units, and ERK7 is sometimes present in small aggregates (arrows). Scale bars: 5 μm (A, B); 200 nm (D, E).

Figure 6

The disassembly of ER exit sites is rescued by loss of ERK7. (A) IF localization of Sec16 in mock-treated (−dsRNA) or ERK7-depleted S2 cells (+ds ERK7) in the absence of serum. (B) IF localization of Sec16 (green) in S2 cells overexpressing ERK7^K54R-V5 in the absence of serum. Transfected cells are marked by the V5 labelling (red) and the nuclei are stained with DAPI. (C) Quantitation of Sec16 dispersion upon serum starvation in non-transfected, WT ERK7-V5 and ERK7^K54R-V5 transfected cells and in cells depleted of ERK7, and upon amino-acid starvation in WT cells and cells expressing ERK7^K54R-V5. Note that removing ERK7 activity partially rescues the loss of ER exit sites upon starvation. (D) IF visualization of the ER exit site marker Sec31 upon serum and amino-acid starvation in mock- and MAKP15-depleted HeLa cells. (E) Quantitation of the number of ER exit sites in the conditions indicated in (D) using Sec31 and Sec16 as ER exit sites markers. Note that removing MAPK15 totally rescues the loss of ER exit sites upon starvation. (F) Western blot of lysates of S2 cells overexpressing ERK7-V5 in full medium for 4 h followed by a 2-h chase in the presence of cycloheximide in full medium, in full medium supplemented with the proteasome inhibitor MG132, and in Ringer/glucose (−AA −serum). Note that ERK7-V5 (upper band) is severely reduced in cells incubated in full medium when compared with starved cells. The band marked by an asterisk is recognized by the V5 antibody and corresponds to a ERK7-V5 truncation occurring normally in cells. Error bars in (C) and (E) represent SD from three and five independent experiments, respectively. Scale bars: 5 μm (A, B) and 25 μm (D).

Figure 4

Serum and amino-acid starvation induces tER site disassembly. (A) IF localization of Sec16 in WT S2 cells grown in full medium and in serum-free medium for 7 h. Arrows indicate cells in which Sec16 pattern is significantly affected. Effects are quantified in Figure 6C using similar criteria as in Figure 3C. (B) Western blot of homogenates of S2 cells cultured in full medium (+ serum) and in the absence of serum (−) using anti-Sec16 and anti α-tubulin (Tub) antibodies. Note that the level of Sec16 remains similar after starvation. (C) IF localization of Sec16 in WT S2 cells starved of amino acids and serum. Note the haze and the large aggregates that are found in almost all of the cells. Effects are quantified in Figure 6C. (D, D′) Time lapse of tER sites disassembly using GFP-Sec23 and Sec16-GFP transfected S2 cells incubated in the absence of amino acids at t=0. Projections of the cells were made every 9–12 min. Note that in cells where expression is low (two upper cells in D′), the GFP punctuates disappear in 30–40 min of starvation and are replaced by a haze. In cells expressing higher level (D, D′, asterisks), the tER sites also disappear but are replaced by aggregates, a situation that recapitulates what is observed with endogenous Sec16. (E) Localization of endogenous Sec16 and Sec23 by IEM in S2 cells incubated in full medium, in the absence of serum and in the absence of amino acids and serum. The arrowhead indicates the remnant of Golgi that can be found in few serum starved cells whereas most of them exhibit no identifiable tER-Golgi units. In amino-acid starved cells, the asterisks indicate the Sec16 and Sec23 membrane-free aggregates that often associated with a cloud of small electron luscent vesicles. Note that this is very different from the tER-Golgi unit morphology observed in S2 cells grown in full medium. Scale bars: 5 μm (A–C) and 200 nm (E).

Figure 5

Amino-acid starvation is mediated by a TORC1-independent pathway. (A, A′) IF visualization of Sec16 in cells incubated in the absence of serum followed by insulin or simultaneous serum withdrawal and insulin addition (A′). Note that Insulin neither rescues nor protects from serum starvation. (B) IF visualization of Sec16 (green) in S2 cells grown full medium supplemented by rapamycin (Rap) for 1.5 h. Note that Sec16 localization is unaffected by drug treatment. (C) Western blot of Phospho-S6K on lysates of S2 cells incubated in DMSO, rapamycin, insulin, insulin+rapamycin showing the effectiveness of the drugs (see also Kondylis et al, 2011). (D) IEM localization of Sec16 in S2 cells treated with rapamycin for 2 h. Note that the structure of the early secretory pathway is intact and Sec16 localizes as in non-treated cells (compared with Figure 4E). (E) IF visualization of Sec16 in mock or raptor depleted S2 cells for 5 days. Note that raptor depletion does not affect Sec16 localization. (F) Quantitation of the different treatments on S2 cells with respect to Sec16 phenotypes. Error bars represent s.d. from five independent experiments (+DMSO, +Rap, + Ins in full serum and in − serum) and three independent experiments for the rest of the conditions. Scale bars: 5 μm.

Figure 7

Response of Sec16 truncations to starvation. (A) Schematic representation of the V5-tagged Sec16 truncations (Sec16 (1–1945); N-ter; MiniSec16, minimum region of Sec16 necessary for its localization to tER sites (aa 690–1535) as described in Ivan et al, (2008); ΔNC1-Sec16; ΔNC1-ΔCter; and ΔNC1-Δ64) with regards to their ability to localize to tER sites and to disperse/aggregate upon serum and amino-acid starvation. The asterisks indicate that the Sec16 construct used in 75 amino acids shorter than endogenous Sec16 (2021). (B) IF visualization of V5-tagged Sec16 truncations (using an anti-V5 antibody) in transfected cells in full medium, serum-free medium (−serum) for 7 h, and serum and amino acid-free medium (−AA –Serum) for 4 h. (C) Quantitation of dispersion/aggregation of V5-tagged Sec16 truncations upon serum and AA starvation. Error bars represent SD from three independent experiments. Scale bars: 5 μm.