ERK regulates Golgi and centrosome orientation towards the leading edge through GRASP65

Abstract

Directed cell migration requires the orientation of the Golgi and centrosome toward the leading edge. We show that stimulation of interphase cells with the mitogens epidermal growth factor or lysophosphatidic acid activates the extracellular signal-regulated kinase (ERK), which phosphorylates the Golgi structural protein GRASP65 at serine 277. Expression of a GRASP65 Ser277 to alanine mutant or a GRASP65 1-201 truncation mutant, neither of which can be phosphorylated by ERK, prevents Golgi orientation to the leading edge in a wound assay. We show that phosphorylation of GRASP65 with recombinant ERK leads to the loss of GRASP65 oligomerization and causes Golgi cisternal unstacking. Furthermore, preventing Golgi polarization by expressing mutated GRASP65 inhibits centrosome orientation, which is rescued upon disassembly of the Golgi structure by brefeldin A. We conclude that Golgi remodeling, mediated by phosphorylation of GRASP65 by ERK, is critical for the establishment of cell polarity in migrating cells.

Figure 1.

ERK-induced Golgi orientation is inhibited by GRASP65 1–201. (A) Mitogen-induced Golgi orientation depends on ERK signaling. NRK monolayers starved for 24 h were wounded and treated with serum for 90 min or pretreated with U0126 for 30 min before stimulation. The Golgi was labeled with an antibody against GM130 (red) and DNA was stained with Hoechst (blue). Golgi were counted as oriented (+) if the majority lay in a 120° angle facing the wound edge at the top of the image, and not oriented if the majority was outside the angle (−). (B) Quantitation. Serum, EGF, or LPA was added for 90 min with or without preincubation with U0126 to inhibit MEK/ERK with an average of 103 cells counted per condition, per experiment. (C) GRASP65 mutants were used in this study. Green, GRASP oligomerization domain; black, phosphorylation and regulatory domain; red, myristoylation site at Gly2; blue, ERK phosphorylation site at Ser277. The S277A point mutant and GRASP65 1–201 lack the ERK phosphorylation site. The constructs were expressed in NRK cells by microinjection of the plasmids into the nuclei. After 2 h, the cells were fixed and stained for Golgi with anti-GM130 (red), DNA (blue), and the expressed GRASP65 construct (green; GRASP65 1–201 and G2A GRASP 1–201 with against-GFP; FL and S277A with anti-GRASP65). Nonmyristoylated G2A GRASP65 is cytosolic, whereas the other constructs are targeted to the Golgi. (D–F) GRASP65 1–201 inhibits Golgi polarization. (D) Serum-starved cells at the wound edge were microinjected with GFP-tagged GRASP65 1–201 or G2A GRASP65 1–201 cDNA, allowed to express proteins for 2 h, and stimulated with LPA for 90 min. Antibodies against GFP (green) and GM130 (red) were used to visualize the expressed proteins and the Golgi, respectively. Nuclei were stained with Hoechst (blue). Golgi were counted as oriented and marked as in A. (E) Quantitation of Golgi orientation of cells expressing GRASP 1–201 or G2A GRASP 1–201 fixed at 0, 30, 90, or 270 min after LPA addition. Nonexpressing cells were counted as an additional control. An average of 64 cells were counted per condition, per experiment. (F) Cells were treated as in D and Golgi orientation was induced with serum, EGF, or LPA and quantified after 90-min stimulation with an average of 100 cells per condition, per experiment. Bars, 10 μm. Red lines mark basal levels expected for random orientation of 33%. Shown are mean ± SEM, from n ≥ 3 independent experiments for each condition or time point. *, P < 0.05; ***, P < 0.001.

Figure 2.

ERK phosphorylates GRASP65 in response to mitogens. (A) Antibodies against phosphorylated Ser277 of GRASP65 (phospho-GRASP65) detect mitotic cells (arrows) but not serum-starved interphase cells. Phospho-GRASP65 is detected in interphase cells upon EGF stimulation and is sensitive to U0126. In contrast, mitotic GRASP65 phosphorylation is resistant to U0126. (B) Phospho-GRASP65 antibodies detect Golgi in cells treated for 10 min with serum, EGF, or LPA, but not in serum-starved cells. No signal was detected when cells were preincubated for 30 min with U0126. Exposure time and image processing are equal. Bars, 10 μm.

Figure 3.

S277A GRASP65 FL inhibits Golgi orientation. (A) GRASP65 FL or S277A GRASP65 FL cDNA were microinjected into serum-starved cells at the wound edge. Proteins were expressed for 2 h before inducing polarization with LPA for 90 min. Cells were then labeled with antibodies against GRASP65 (green) and GM130 (red), and DNA was stained with Hoechst (blue). Injected cells were identified by overexpressed GRASP65. Golgi were counted as oriented and marked as in Fig. 1 A. Bar, 10 μm. (B) Quantitation of A. Shown are mean ± SEM of n = 4 independent experiments with an average of 27 cells per condition, per experiment. *, P < 0.016. Red lines mark basal levels expected for random orientation of 33%.

Figure 4.

GRASP65 phosphorylation by ERK causes loss of oligomerization and Golgi unstacking. (A) Serum-staved NRK cells were treated with serum, EGF, or LPA for 10 min. Proteins were separated by SDS-PAGE and blotted with antibodies against phospho-GRASP65, GRASP65, phospho-ERK1/2, and ERK1. Serum, EGF, and LPA induced phosphorylation of ERK and GRASP65, which was abolished by U0126 (30-min preincubation). (B) RLG membranes were treated with recombinant MEK/ERK and then subjected to Western blotting for phospho-GRASP65 and GRASP65. Kinase treatment induced GRASP65 phosphorylation, which was inhibited by U0126 and the kinase inhibitor staurosporine. (C) RLG membranes were treated with MEK/ERK as in B and analyzed by EM. Bar, 0.5 μm. (D) The percentage of stacked cisternae in RLG membranes was quantified after EM analysis using the intersection method. MEK/ERK treatment significantly reduced the percentage of stacked cisternae. (E and F) Purified recombinant GRASP65 FL or S277A GRASP65 FL protein was cross-linked to magnetic beads. Incubation with interphase cytosol (control) caused the beads to aggregate. Further incubation with MEK/ERK dispersed the GRASP65 FL beads, but had no effect on S277A GRASP65 FL beads. Bar, 50 μm. Mean ± SEM for n ≥ 4 independent experiments are shown in D and F. ***, P < 0.001.

Figure 5.

Expression of GRASP65 1–201 inhibits centrosome orientation. (A) Serum-starved cells expressing GFP-tagged GRASP65 1–201 or G2A GRASP65 1–201 (green) were stimulated with LPA for 90 min. The cells were then stained for centrosomes (anti–γ-tubulin; white), Golgi (anti-GM130; red), and DNA (blue). Injected cells were identified by the GFP signal (green). Centrosomes were counted as oriented (+) if they fell into a 120° angle facing the wound, and not oriented if the majority was outside the angle (−). Bar, 10 μm. (B and C) Fragmentation of the Golgi by BFA rescues centrosome orientation. Cells expressing GRASP65 S277A/GRASP65 FL (B) or GFP-tagged GRASP65 1–201/G2A GRASP65 1–201 (C) were treated with BFA for 30 min to fragment the Golgi or treated with vehicle alone. BFA was washed out and the cells were stimulated with LPA for 90 min. Cells were fixed and stained for GRASP65 (B) or GFP (C) to identify injected cells, γ-tubulin to count centrosome orientation, and Hoechst. Mean percentage of centrosome orientation ± SEM was calculated from greater than or equal to three independent experiments, with averages of 64 (A) and 105 (B) cells counted per condition, per experiment. ** P, < 0.01; ***, P < 0.001. Red lines mark basal levels expected for random orientation of 33%.