MAPK signaling to the early secretory pathway revealed by kinase/phosphatase functional screening

Abstract

To what extent the secretory pathway is regulated by cellular signaling is unknown. In this study, we used RNA interference to explore the function of human kinases and phosphatases in controlling the organization of and trafficking within the secretory pathway. We identified 122 kinases/phosphatases that affect endoplasmic reticulum (ER) export, ER exit sites (ERESs), and/or the Golgi apparatus. Numerous kinases/phosphatases regulate the number of ERESs and ER to Golgi protein trafficking. Among the pathways identified, the Raf-MEK (MAPK/ERK [extracellular signal-regulated kinase] kinase)-ERK cascade, including its regulatory proteins CNK1 (connector enhancer of the kinase suppressor of Ras-1) and neurofibromin, controls the number of ERESs via ERK2, which targets Sec16, a key regulator of ERESs and COPII (coat protein II) vesicle biogenesis. Our analysis reveals an unanticipated complexity of kinase/phosphatase-mediated regulation of the secretory pathway, uncovering a link between growth factor signaling and ER export.

Figure 1.

Classification of kinase/phosphatase hits. (A) Assignment to protein families. TKL, Tyr kinase–like; STE, related to sterile kinases; CK1, casein kinase 1; AGC, kinase family containing PKA, PKG, and PKC; CAM kinase, calmodulin kinases; CMGC, kinase family containing CDK, MAPK, GSK, and CKL. (B) Phenotypic classification of hits. (C–F) Immunofluorescence of phenotypes. HeLa cells transfected with control siRNA (C) or siRNA to MAPK4 (example of the ER class; D), SCYL1 (example for the mixed class; E), and MAPK14 (example for the Golgi class; F) are shown. Cells were fixed 72 h after transfection and double stained with antibodies against the indicated proteins. Bars, 10 µm.

Figure 2.

Network of kinase hits indicates a role for MAPK signaling in the regulation of ER export. (A) All ER-class hits were assembled into a network using the STRING database. (B) Assembly of all of the Golgi-class hits into a network using the STRING database. Edges in both networks indicate an interaction that is experimentally verified or an association into a signaling pathway. (C) ER-, mixed-, and Golgi-class hits specifically enriched in signaling pathways deduced from the DAVID database.

Figure 3.

BiNGO analysis of Golgi-class hits from four anchored networks. (A) Four networks were generated with the Golgi-class hits anchored to GOLGA2, TMED2, ARF1, and polarity regulators (Par3, Par6, Cdc42, Pasl1, PKC-ζ, and Scrib). Networks were merged using the Cytoscape software and analyzed using the BiNGO 2.3 plugin. GO term cellular processes were scored at a significance value of at least 0.001. The Benjamini-Hochberg false discovery rate correction was used. (B) The network of cellular processes displayed in a hierarchical layout. Large PDF versions of these networks and their underlying data (which can be opened and visualized in Cytoscape) can be found in the online supplemental material.

Figure 4.

Kinases and phosphatases of the Golgi class inhibit orientation of the Golgi toward a scratch wound. HeLa cells were transfected with siRNAs to the indicated genes. After 72 h, a scratch wound was introduced. 6 h later, cells were fixed and stained for giantin to label the Golgi. (A) Representative images. (B) Quantification of two experiments with at least 35 cells each displayed as the percentage of cells exhibiting Golgi orientation to and away from the wound. (C) 48 h after the indicated knockdown, a wound was introduced, and cells were fixed 24 after later. Cells were labeled with FITC-tagged Con A. Note that STYX and AKAP28 knockdown cells fail to close the wound, in contrast to control cells. (A and C) White lines indicate the position of the scratch. Bars: (A) 10 µm; (C) 200 µm.

Figure 5.

A role for MAPK signaling in the regulation of ERES number and ER to Golgi transport. (A) HeLa cells were transfected with the indicated siRNAs and fixed after 72 h. Anti-Sec31–labeled ERESs were quantified (means ± SD; n = 3). Asterisks indicate statistically significant differences (*, P < 0.05) from control (unpaired t test). (B) Cells were transfected with control or ERK2 siRNA. After 72 h, cells were lysed, and ERK1/2 was detected by immunoblotting. (C) HeLa cells were transfected with control or ERK2 siRNA (siERK2). After 48 h, the ERK2 siRNA–treated cells were left untreated (mock) or were transfected with Flag-tagged ERK2 cDNA or Flag-tagged ERK2-K52R cDNA. Cells were fixed 24 h later and stained for Flag and Sec31, and ERESs were counted. Asterisks indicate statistically significant differences (*, P < 0.05). (D) ER–Golgi transport of endogenous transferrin receptor was measured 72 h after transfection with the indicated siRNAs. Cells were subjected to pulse-chase experiments with [³⁵S]Met. Transferrin receptor was immunoprecipitated, treated with Endo H, separated by SDS-PAGE, and visualized by autoradiography. C0 and C30: control siRNA transfected and chased for 0 min and 30 min, respectively. ERK2: knockdown cells chased for 30 min. The bar diagram shows means ± SD (n = 3). The asterisk indicates statistically significant difference (*, P < 0.05) from control (unpaired t test). (E) ER–Golgi transport of endogenous α1-antitrypsin in HepG2 cells. 2 d after plating, cells were subjected to pulse-chase experiments with [³⁵S]Met. Cells were treated with DMSO (C) or the MEK inhibitor PD98059 (+PD) for 30 min before the pulse. Cells were pulsed for 10 min and chased for 15 min. C0 and C15: DMSO-treated cells chased for 0 min and 15 min, respectively. +PD: PD98059-treated cells chased for 15 min. The bar diagram is an evaluation of two independent experiments. (B, D, and E) The black lines on the blot indicate the position of the 50-kD (B and E) or 75-kD (D) molecular mass marker.

Figure 6.

ERK2 stimulates COPII vesicle budding. HeLa microsomes were incubated for 30 min with ATP and an ATP-regenerating system, GTP, and cytosol of ERK2-silenced HeLa cells in the presence or absence of purified ERK2. COPII vesicles (V) were collected by centrifugation and immunoblotted for anti–ERGIC-53 and anti–CLIMP-63 (ER marker). The asterisk represents a statistically significant difference (*, P < 0.05; paired t test). Error bars indicate ±SD. M, 33% microsome input.

Figure 7.

Changes in Sec16 dynamics upon inhibition or activation of the MAPK pathway. (A) Cells were transfected with control siRNA or siRNA to ERK2. After 48 h, cells were transfected with GFP-Sec16 cDNA, and 24 h later, FRAP analysis was performed. Of the two control siRNA–transfected cells, one dish was treated with solvent (control) and one with the MEK inhibitor PD98059 for 30 min before analysis. The right panel shows examples of bleached ERESs with indicated time points. (B) Quantification of mobile fractions from the curves in A (see Materials and methods). Asterisks indicate statistically significant differences (means ± SD; three independent experiments; *, P < 0.01; unpaired t test). (C) HeLa cells expressing GFP-Sec16 were serum starved for 8 h before the FRAP experiment (WT-ut). Afterward, cells were treated with EGF for 10 min, and further FRAP curves were recorded from the same dish (WT+EGF). (D) Quantification of mobile fractions from the curves in C (see Materials and methods; means ± SD; three independent experiments, four curves each; paired t test). (E) HeLa cells expressing YFP-Sec24D were serum starved for 8 h before the FRAP experiment (Sec24D-ut). Afterward, cells were treated with EGF for 10 min, and additional FRAP curves were recorded from the same dish (Sec24D+EGF). (F) Quantification of mobile fractions from the curves in E (see Materials and methods; means ± SD; three independent experiments, four curves each; paired t test). Bar, 1 µm.

Figure 8.

Sec16 is a target of the MAPK pathway. (A) Cells were transfected with GFP-Sec16 cDNA. After 8 h, medium was replaced by complete medium (Cont) or medium without serum (starved). Sec16 was immunoprecipitated with anti-GFP, subjected to SDS-PAGE, and immunoblotted with anti-Sec16 (top). The blot was stripped and reprobed with anti–MAPK substrate (bottom). The numbers above the blot indicate relative amounts determined by densitometric scanning. The intensity of the MAPK substrate band was normalized to the Sec16 band. To facilitate comparisons, the value of the control band was set to 1. Note that starvation reduces phosphorylation of Sec16 by 40% (means ± SD; three independent experiments; P < 0.05). (B) Cells were transfected with GFP-Sec16 and, after 8 h, serum starved. Sec16 was immunoprecipitated with anti-GFP. A parallel cell culture was treated with EGF. Active ERK was immunoprecipitated with agarose-coupled anti–phospho-ERK and incubated with the immunoprecipitated active ERK. Immunoprecipitated GFP-Sec16 and immunoprecipitated phospho-ERK were mixed, and the kinase reaction was performed for 30 min at 30°C (reaction). 50% of the reaction was separated by SDS-PAGE and immunoblotted with anti-Sec16 (top). The blot was stripped and reprobed with anti–MAPK substrate (bottom). (C) Cells were transfected with GFP-Sec16 cDNA (wild type [wt] or T415I). Immunoprecipitated Sec16 (anti-GFP) was coincubated with active ERK2 and ATP for 30 min at 30°C. 50% of the reaction was loaded, subjected to SDS-PAGE, and immunoblotted with anti-Sec16 (top). The blot was stripped and reprobed with anti–MAPK substrate antibody (bottom). The numbers above the blot indicate relative amounts determined by densitometric scanning. The intensity of the MAPK substrate band was normalized to the Sec16 band. To facilitate comparisons, the value of the control band was set to 1. Note that the mutation of T415I reduces the ability of ERK2 to phosphorylate Sec16 by at least 65% (means ± SD; three independent experiments). (D) Phosphorylation of purified GST-tagged Elk1 by immunoprecipitated phospho-ERK1/2 (IP p-ERK) as described in B or recombinant active ERK2 (rec. ERK2). The reaction was performed at 30°C for 30 min either with buffer (−) or with the kinase (+). (E) HeLa cells were transfected with a plasmid encoding GFP-Sec16 lacking the N-terminal 923 aa. Sec16 was immunoprecipitated with anti-GFP and incubated with active ERK2 and ATP for 30 min at 30°C. Samples were subjected to SDS-PAGE and immunoblotting with anti-Sec16 (top). The blot was stripped and reprobed with anti–MAPK substrate antibody (bottom). The arrow indicates the position where a band would be expected in case of corresponding phosphorylation. (F) HeLa cells expressing GFP-tagged truncation mutant of Sec16 lacking the N-terminal 923 aa (delta) were serum starved for 8 h before the FRAP experiment (delta-ut). Afterward, cells were treated with EGF for 10 min, and further FRAP curves were recorded from the same dish (delta+EGF). (G) Quantification of the mobile fractions from the curves in F (see Materials and methods; means ± SD; three independent experiments). (H) Cells were transfected with GFP-Sec16 cDNA (wild type [wt] or T415I). Sec16 was immunoprecipitated with anti-GFP, subjected to SDS-PAGE, and immunoblotted with anti-Sec16 (top). The blot was stripped and reprobed with anti–MAPK substrate (bottom). The numbers above the blot indicate relative amounts determined by densitometric scanning. The intensity of the MAPK substrate band was normalized to the Sec16 band. To facilitate comparisons, the value of the control band was set to 1. Note that the T415I mutation reduces the phosphorylation of Sec16 down to 30% (means ± SD; four independent experiments). (I) HeLa cells expressing GFP-tagged Sec16-T415I were serum starved for 8 h before the FRAP experiment (T415I-ut). Thereafter, the cells were treated with EGF for 10 min, and additional FRAP curves were recorded from the same dish (T415I+EGF). (J) Quantification of the mobile fractions from the curves in I (see Materials and methods; means ± SD; three independent experiments). (A–E and H) Black lines on the blots indicate the position of the 250-kD (A–C), 50-kD (D), 150-kD (E), or 250-kD (H) molecular mass marker.

Figure 9.

Phosphorylation of Sec16 is linked to ERES number. (A–C) Hells cells were transfected with cDNA encoding wild-type GFP-Sec16 (GFP–Sec16-wt; A) or GFP–Sec16-T415I (B), fixed 24 h later, and immunostained for Sec31 to label ERESs or ERGIC-53 to label the ERGIC. (C) The number of ERESs and peripheral ERGIC punctae was counted. The bar graph represents the number of punctae positive for Sec31 (ERES) or ERGIC-53 (ERGIC). The values determined in cells expressing GFP–Sec16-wt were set to 100% (means ± SD; three independent experiments; asterisks indicate statistically significant difference to GFP–Sec16-wt; *, P < 0.05; unpaired t test). (D) Cells were transfected with cDNA for GFP-Sec16 together with either wild-type H-Ras (Ras-wt) or M1–H-RasV12 (RasV12). Sec16 was immunoprecipitated with anti-GFP, subjected to SDS-PAGE, and immunoblotted with anti-Sec16 (top). The blot was stripped and reprobed with anti–MAPK substrate (bottom). The numbers above the blot indicate relative amounts determined by densitometric scanning. The intensity of the MAPK substrate band was normalized to the Sec16 band. To facilitate comparisons, the mean value of the samples transfected with wild-type Ras was set to one (mean ± SD; three independent experiments). The black line on the blot indicates the position of the 250-kD molecular mass marker. (E) Cells were transfected with cDNA for HA-tagged M1-RasV12. After 24 h, cells were fixed and costained for HA (RasV12) and Sec31. The bar graph shows ERESs/cell comparing cells expressing RasV12 or not (control; means ± SD; three independent experiments; asterisk indicates statistically significant difference; *, P < 0.05 from control; unpaired t test). Bars, 10 µm.

Figure 10.

Model of the pathway signaling to ERESs. EGFR, EGF receptor.