Regulation of Rho guanine nucleotide exchange factor 3 by phosphorylation in the PH domain

Abstract

Although most members of the Rho guanine nucleotide exchange factor (RhoGEF) family are found to be phosphorylated, how phosphorylation regulates RhoGEF function is poorly understood. Here we report the discovery of a mechanism of RhoGEF regulation by phosphorylation. We find that ARHGEF3 is phosphorylated in the pleckstrin homology (PH) domain in a protein kinase C (PKC)-dependent manner. This phosphorylation inhibits ARHGEF3 activation of RhoA and actin stress fiber formation in the cells, and it also disrupts ARHGEF3 binding to PI(3,5)P₂ but not to PI(4,5)P₂. Guided by molecular dynamics simulation, a mutation in the PH domain is identified to uncouple ARHGEF3 binding to the two lipids, which is used to rule out a role of PI(3,5)P₂ in regulating GEF activity. Results of in vitro GEF assays suggest that PH domain phosphorylation diminishes ARHGEF3 catalytic activity, likely through an allosteric mechanism. Our findings reveal a previously unknown type of regulatory mechanism for the family of RhoGEFs.

Keywords: Biochemistry; In silico biology; Molecular biology.

Graphical abstract

Figure 1

ARHGEF3 is phosphorylated on the PH domain in a PKC-dependent manner (A) HEK293 cells were transfected with GFP-ARHGEF3 for 24 h, serum-starved overnight, and then stimulated with PMA (100 nM), insulin (100 nM), or fetal bovine serum (10%) for 30 min. Cell lysates were collected and run on a Phos-tag gel (top panel) as well as regular SDS-PAGE, followed by western blotting. Recombinant ARHGEF3 proteins were detected by anti-GFP. (B) Cells were transfected and starved as in (A) and stimulated with 100 nM PMA for 30 min, followed by western analysis as in (A). BIM-1 at increasing concentrations were added to the cells 30 min before PMA addition. (C) Domain structure of ARHGEF3. The amino acid numbers at domain boundaries were predicted by AlphaFold 3. (D) Cells were transfected with Myc-tagged fragments of ARHGEF3 and subsequently treated as in (B) except that 1 μM BIM-I was used, and the lysates were analyzed on a Phos-tag gel. Molecular weight markers (kDa) are indicated on the right side of blots.

Figure 2

ARHGEF3 is phosphorylated in S399 by nPKCs (A) HEK293 cells were transfected with Myc-tagged ARHGEF3 PH domain (WT or S399A) for 24 h, serum-starved overnight, and then stimulated with 100 nM PMA for 30 min. Lysates were collected and run on a Phos-tag gel, followed by anti-Myc western blotting. (B) HEK293 and C2C12 cells were transfected with WT or S399A GFP-ARHGEF3 and then treated with 100 nM PMA and 1 μM BIM-I for 30 min. Lysates were analyzed by western blotting. (C) HEK293 cells were serum-starved overnight and then stimulated with 100 nM PMA for 30 or 90 min, with 1 μM BIM-I added 30 min before PMA addition. Endogenous pS399-ARHGEF3 and total ARHGEF3 were analyzed by western blotting. (D) HEK293 cells were transfected with GFP-ARHGEF3 for 24 h, pretreated with inhibitors for 10 min, and then stimulated with 100 nM PMA for 30 min, followed by cell lysis and western blotting. Inhibitor-treated bands are shown with their respective positive and negative controls from the same membrane even if they are presented separately; original images of the full blots are shown in Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6and S5. All blots shown are representative images of three or more independent experiments with similar outcomes. Molecular weight markers (kDa) are indicated on the right side of blots.

Figure 3

S399 phosphorylation reduces ARHGEF3 activity (A) HEK293 cells were transfected with GFP-ARHGEF3 (or GFP as control) for 24 h and then stimulated with 100 nM PMA for 1 h. Cell lysates were subjected to active RhoA pull-down by GST-Rhotekin beads and analyzed by western blotting. Western signals were quantified by densitometry, and relative RhoA activity was expressed as the ratio of active RhoA (pulldown) versus total RhoA (lysate), normalized to control as 1. Data are presented as mean ± SEM (n = 6). One-sample or paired t test was performed, and p values for significant differences (p < 0.05) are indicated on the graph. Representative blots are shown. pS399-GFP-ARHGEF3 western blot is also shown. Molecular weight markers (kDa) are indicated on the right side of blots. (B) Cells were transfected with WT-, S399A-, or S399D-GFP-ARHGEF3 (or GFP as control), followed by active RhoA pulldown and quantification as in (A) (n ≥ 5). Linear mixed-model analysis followed by a Tukey test was performed for pairwise comparison, and p values for significant differences (p < 0.05) are indicated on the graph. Representative blots are shown. (C) HeLa cells were transfected with WT-, S399A-, or S399D-GFP-ARHGEF3 (or GFP as control) followed by fixation and phalloidin/DAPI staining. Representative images of three independent experiments are shown (scale bar, 2 μm). (D) Fluorescence intensities of stress fibers were quantified for the experiments in (C). Data are shown as mean ± SEM (n = 3 experiments, >20 cells per experiment). Paired t test was performed, and p values for significant differences (p < 0.05) are indicated on the graph. (E) HeLa cells were transfected with WT- or S399A-GFP-ARHGEF3 for 24 h and then stimulated with 100 nM PMA for 1 h, followed by fixation and phalloidin/DAPI staining. Representative images of three independent experiments are shown (scale bar, 2 μm). (F) Fluorescence intensities of stress fibers were quantified for the experiments in (E). Data are shown as mean ± SEM (n = 3 experiments, >25 cells per experiment). Two-way ANOVA was performed followed by Tukey test. Significant effect was found with PMA treatment but not with mutation, and p values for significant differences (p < 0.05) are indicated on the graph.

Figure 4

S399 phosphorylation impairs ARHGEF3 binding to PI(3,5)P₂ HEK293 cells were transfected with WT-, S399A-, or S399D-GFP-ARHGEF3 for 24 h, and cell lysates were subjected to lipid SiMPull assays using small unilamellar vesicles (SUVs) containing 5% PI(3,5)P₂ or PI(4,5)P₂. Data shown are mean ± SEM for one experiment (n ≥ 10 images). Three independent experiments were performed with similar outcomes. The threshold of 100 GFP counts (pulled down by SUVs), previously determined as a detectable interaction, is indicated by dotted lines.

Figure 5

H427 distinguishes PH domain interaction with PI(3,5)P₂ and PI(4,5)P₂ (A) MD simulations were performed with the ARHGEF3 PH domain bound to membrane bilayers containing PI(4,5)P₂ or PI(3,5)P₂. The histograms show normalized ensemble-averaged number of contacts formed between residues in the PH domain and the 3- and 4-phosphate groups (3P and 4P) of PI(3,5)P₂ and PI(4,5)P₂, respectively. Data were averaged over all simulation replicas for each lipid composition. (B) ARHGEF3 PH domain, and highlighted on the ribbon structure are the residues with the highest frequency of interaction with 3P or 4P. (C) HEK293 cells were transfected with WT- or H427D-GFP-ARHGEF3 for 24 h, and cell lysates were subjected to lipid SiMPull assays as in Figure 4. Data shown are mean ± SEM for one experiment (n ≥ 10 images). Three independent experiments were performed with similar outcomes. Previously determined threshold for binding (100 GFP counts) is indicated by dotted lines. (D) HEK293 cells were transfected with WT or H427D-GFP-ARHGEF3 (or GFP as control) for 24 h. Cell lysates were subjected to active RhoA pulldown by GST-Rhotekin beads and analyzed by western blotting. Western signals were quantified by densitometry, and relative RhoA activity was expressed as the ratio of active RhoA (pulldown) versus total RhoA (lysate), normalized to control as 1. Data are presented as mean ± SEM (n = 9). One-sample or paired t test was performed, and p values for significant differences (p < 0.05) are indicated on the graph. Representative blots are shown. Molecular weight markers (kDa) are indicated on the right side of blots.

Figure 6

S399D reduces the catalytic activity of ARHGEF3 (A) Purified WT- and S399D-GST-ARHGEF3 were subjected to in vitro RhoA guanine nucleotide exchange assays. EDTA: positive control for complete nucleotide dissociation. Data from five independent experiments with similar outcomes were fit as a single exponential decay curve (left panel). RhoA nucleotide exchange activity at t = 45 min is shown on the bar graph (right panel), normalized to EDTA as 100%. Data are shown as mean ± SEM (n = 5). Student’s t test was performed. (B) A model of ARHGEF3 regulation. Left panel: ARHGEF3 activates RhoA at the plasma membrane, requiring binding to PI(4,5)P₂. PKC-dependent phosphorylation in the PH domain allosterically inhibits the GEF activity of ARHGEF3, dampening RhoA activation and actin stress fiber formation. Right panel: the same phosphorylation also disrupts ARHGEF3 binding to PI(3,5)P₂ (presumably in the late endosome), the role of which is yet to be determined. The graphics were created using BioRender.