Identification of G protein alpha subunit-palmitoylating enzyme

Abstract

The heterotrimeric G protein alpha subunit (Galpha) is targeted to the cytoplasmic face of the plasma membrane through reversible lipid palmitoylation and relays signals from G-protein-coupled receptors (GPCRs) to its effectors. By screening 23 DHHC motif (Asp-His-His-Cys) palmitoyl acyl-transferases, we identified DHHC3 and DHHC7 as Galpha palmitoylating enzymes. DHHC3 and DHHC7 robustly palmitoylated Galpha(q), Galpha(s), and Galpha(i2) in HEK293T cells. Knockdown of DHHC3 and DHHC7 decreased Galpha(q/11) palmitoylation and relocalized it from the plasma membrane into the cytoplasm. Photoconversion analysis revealed that Galpha(q) rapidly shuttles between the plasma membrane and the Golgi apparatus, where DHHC3 specifically localizes. Fluorescence recovery after photobleaching studies showed that DHHC3 and DHHC7 are necessary for this continuous Galpha(q) shuttling. Furthermore, DHHC3 and DHHC7 knockdown blocked the alpha(1A)-adrenergic receptor/Galpha(q/11)-mediated signaling pathway. Together, our findings revealed that DHHC3 and DHHC7 regulate GPCR-mediated signal transduction by controlling Galpha localization to the plasma membrane.

FIG. 1.

Dynamic palmitate turnover on Gα_q/11. (A) Gα_q shuttles between the PM and the Golgi apparatus. When Gα_q-Dendra2, Gβ₁, and Gγ₂ were coexpressed in HeLa cells, Gα_q-Dendra2 (green) was localized at the PM and the endomembrane. Gα_q-Dendra2 in the endomembranes (upper) and a part of PM (lower) within the white regions was photoconverted by 405-nm laser. Converted Gα_q-Dendra2 (gray scale) was monitored for 20 min. Cells were then immunostained with anti-GM130 (magenta) antibody (right). Scale bar, 10 μm. (B) Inhibition of palmitoylation causes detachment of Gα_q/11 from the PM. HeLa cells were treated with 2-BP (100 μM) or CHX (20 μg/ml) for 4 h. The cells were then doubly stained with anti-Gα_q/11 (green) and anti-β-catenin (red) antibodies. Scale bar, 20 μm. (C) 2-BP blocks palmitoylation of Gα_q/11. HeLa cells were metabolically labeled with [³H]palmitate ([³H]palm) for 4 h either in the presence or absence of 2-BP. Gα_q/11 was immunoisolated and subjected to fluorography (upper) or Western blotting (lower). IB, immunoblotting. (D) HeLa cells were treated with 2-BP for 4 h. Then, 2-BP was washed out, and protein synthesis was inhibited by 20 μg/ml CHX for 4 h. The cells were stained with anti-Gα_q/11 antibody at the indicated times. The inhibition of palmitoylation for 4 h caused delocalization of Gα_q/11 from the PM. The Gα_q/11 dispersed by 2-BP came back to the PM again within 2 h after removal of 2-BP in the presence of CHX, indicating that the relocalization of Gα_q/11 depends on palmitoylation and depalmitoylation. Scale bar, 20 μm. (E) Pulse-chase analysis of Gα_q/11 palmitoylation. HeLa cells were labeled with [³H]palmitate or [³⁵S]methionine-cysteine ([³⁵S]Met/Cys) for 4 h. After incubation with chase medium for 0, 1, 2, 4, and 6 h, cells were lysed and subjected to IP with anti-Gα_q/11 antibody. Immunoprecipitates were separated by SDS-PAGE, followed by fluorography. The ratio of [³H]palmitate to [³⁵S]methionine-cysteine -labeled Gα_q/11 was plotted in the graph. Error bars show ±SD (n = 3). IgG, control immunoglobulin G.

FIG. 2.

Screening of potential Gα_q palmitoylating enzymes. (A) Individual HA-DHHC clones (0.5 μg plasmid) were transfected with Gα_q-GFP (0.5 μg) into HEK293T cells. After metabolic labeling with [³H]palmitate, proteins were separated by SDS-PAGE, followed by fluorography and Western blotting with anti-GFP antibody for Gα_q-GFP and anti-HA antibody for DHHC proteins. An arrow indicates the position of Gα_q-GFP. White asterisks indicate autopalmitoylation of expressed DHHC proteins. Coexpression of DHHC3 or -7 robustly and specifically increased Gα_q palmitoylation. Note that several DHHC proteins, such as DHHC16, -20, and -21, express at lower levels than DHHC3 and -7. M, molecular mass. (B) HEK293T cells were transfected with the indicated amount of DHHC proteins with Gα_q-GFP and were labeled with [³H]palmitate. DHHC16 and -20 did not increase Gα_q palmitoylation, whereas limited DHHC3 expression by 10 ng of plasmids (showing expression levels similar to DHHC16 and -20) still enhanced Gα_q palmitoylation. An arrow indicates the position of DHHC21. (C) Although DHHC21 expressed at a lower level than DHHC3, DHHC21 has apparent PAT activity toward Lck. An arrow indicates the position of DHHC21. (D) Treatment of labeled cell lysates with 0.5 M hydroxylamine (NH₂OH) but not 0.5 M Tris-HCl (−) released DHHC3- or DHHC7-mediated [³H]palmitate incorporated into Gα_q-HA, indicating that DHHC3- or DHHC7-induced palmitoylation is mediated by a thioester bond. (E) The DHHC3(C157S) and DHHC7(C160S) mutations (shown as CS) abolished the palmitoylating activity. The mutations cysteines 9 and 10 in Gα_q (CS) abolished its palmitoylation. Asterisks indicate the autopalmitoylation of DHHCs. (F) HEK293T cells cotransfected with indicated DHHC clones (2, 3, 7, 21, and 15) and Gα-GFP subfamily (α_q, α_s, and α_i2) were metabolically labeled with [³H]palmitate. All Gα members were palmitoylated by DHHC3 and DHHC7. Gα_i2 was also palmitoylated by DHHC2 and DHHC21 to a lesser extent. WT, wild type; IB, immunoblotting.

FIG. 3.

DHHC3 directly palmitoylates Gα_q in vitro. (A) Immunopurified GFP-tagged wild-type (WT) Gα_q and Gα_q (CS) and FLAG-DHHC3 were stained with Coomassie brilliant blue (CBB). An arrowhead and an arrow indicate the positions of Gα_q-GFP and FLAG-DHHC3, respectively. An asterisk indicates the nonspecific band. M, molecular mass. (B) Purified Gα_q and DHHC3 were incubated in the presence of [³H]palmitoyl-CoA for 10 min at 30°C. Then, radiolabeled Gα_q-GFP was evaluated by fluorography (upper) and immunoblotting (IB) with anti-GFP antibody (lower). DHHC3 palmitoylated wild-type Gα_q but not Gα_q (CS).

FIG. 4.

DHHC3 and -7 are necessary for Gα_q/11 palmitoylation and PM localization. (A) DHHC3 and -7 siRNAs downregulate targeted mRNA expressions. HeLa cells were transfected with a control or DHHC3 or -7 siRNA duplex. At 72 h after transfection, the expression of DHHC3 or -7 mRNA was quantitated by real-time RT-PCR. The expression of DHHC3 and -7 was normalized to that of GAPDH. Error bars show ±SD (n = 3). (B) HEK293T cells were transfected with the indicated siRNAs, and cells were labeled with [³H]palmitate for 4 h. Endogenous Gα_q/11 was then immunoprecipitated, followed by fluorography and Western blotting. (C) HeLa cells were transfected with the indicated siRNAs and doubly stained with anti-Gα_q/11 (green) and anti-β-catenin (red) antibodies. Note that knockdown of DHHC3 and/or DHHC7 specifically impairs PM localization of Gα_q/11. Graphs indicate fluorescent intensity along white lines. Arrows mark the cell-to-cell contact sites along the white line. Scale bar, 20 μm. (D) siRNAs to human-specific DHHC3 and DHHC7 were cotransfected into HeLa cells together with wild-type (WT) GFP-mDHHC3 or catalytically inactive GFP-mDHHC3(C157S) [mDHHC3(CS)]. Cells were immunostained with anti-GFP (blue), anti-Gα_q/11 (green), and anti-β-catenin (red) antibodies. The colors are pseudocolors. Scale bar, 20 μm. Relative fluorescence intensities of PM to cytosol are shown (right graph). Error bars show ±SD (n = 20). **, P < 0.01. IB, immunoblotting.

FIG. 5.

Depletion of DHHC3 expression impairs the PM targeting of Gα_q/11 in hippocampal neurons. Rat hippocampal neurons (DIV8) were transfected with mCherry-miR RNA (to rat DHHC2 or -3) expression vectors (red). At 7 days after transfection, neurons were fixed and stained with anti-Gα_q/11 antibody (green) and Hoechst (blue). Scale bars, 5 μm (high magnification) and 10 μm (low magnification). Graph shows ratio of fluorescence intensities of the PM to the cytosol. Error bars show ±SD (n = 12). **, P < 0.01. miRNA, microRNA.

FIG. 6.

TIRFM imaging of membrane-bound palmitoylated Gα_q. (A) HeLa cells were transfected with GFP-tagged Gα_q (WT), palmitoylation-deficient Gα_q (CS), or the C-terminal sequence of K-Ras including polybasic and prenylation sequences (GFP-CAAX). Cells were observed by epifluorescence microscopy (Epi) and TIRFM before and at 4 h after treatment with 2-BP. Gα_q-GFP was clearly detected by TIRFM, and the intensity was reduced on 2-BP treatment. The intensity of Gα_q (CS) was apparently weaker than that of Gα_q (WT). Scale bar, 20 μm. (B) HeLa cells were transfected with control or DHHC3/DHHC7 siRNAs together with Gα_q-GFP or GFP-CAAX. Scale bar, 20 μm. Relative fluorescence intensities of cell images from TIRFM compared to those of epifluorescence microscopy are indicated in the graph. Note that Gα_q-GFP intensity visualized by TIRFM was reduced significantly in DHHC3 and DHHC7 knocked down cells. Error bars show ±SD (n = 5). **, P < 0.01; *, P < 0.05.

FIG. 7.

Gα_q and DHHC3 colocalize to the Golgi apparatus. (A) HeLa cells were transfected with control or DHHC3 siRNA. The lysates were then immunoblotted with anti-DHHC3 and anti-moesin antibodies, indicating that DHHC3 antibody specifically detects endogenous DHHC3. Molecular weights (in thousands) are shown at the right. (B) HeLa cells transfected with control or DHHC3 siRNA were immunostained with antibodies to DHHC3 (green) and Golgi marker GM130 (red). White boxes are magnified. The Golgi staining by DHHC3 antibody is specific because the signal disappeared in siDHHC3-treated cells. Scale bars, 20 μm (low magnification) and 10 μm (high magnification). (C) HeLa cells were cotransfected with Gα_q-GFP and mCherry-DHHC3. Some population of Gα_q was colocalized with DHHC3 at the Golgi apparatus. Scale bar, 20 μm. (D) Dendra2-DHHC3 (green) at the Golgi apparatus was photoconverted (gray scale). DHHC3 was stably localized at the Golgi apparatus, where GM130 was labeled (magenta). Scale bar, 10 μm.

FIG. 8.

Retrograde PM-Golgi trafficking of Gα_q depends on Gα_q palmitoylation by Golgi apparatus-resident DHHC3 and -7. (A) HeLa cells were transfected with Gα_q-GFP (green and white) and GalT-mCherry (red) expression vectors with or without Gβ₁γ₂. The Golgi region (white circle) marked by GalT-mCherry was bleached, and then fluorescent recovery of Gα_q-GFP was monitored by acquiring images every 10 s. Fluorescence intensities were plotted in graphs (right). (B) HeLa cells expressing Gα_q-GFP (green and white) and GalT-mCherry (red) were treated with or without 2-BP for 30 min before FRAP analysis. The region (white circle) identified with GalT-mCherry was bleached, and then fluorescent recovery of Gα_q-GFP was monitored. Treatment with 2-BP inhibited the recovery of fluorescence. (C) Knockdown (KD) of DHHC3 and -7 inhibited the Golgi and PM targeting of Gα_q-GFP, resulting in the diffuse cytoplasmic localization. (D) Palmitoylation-deficient Gα_q (CS)-GFP showed similar cytoplasmic distribution and rapid recovery both around the GalT-positive region (CS region 1) and at cytoplasmic region (CS region 2). The recovery of palmitoylation deficient Gα_q (CS)-GFP was as rapid as that of DHHC3 and -7 knocked down cells in panel C. Scale bar, 10 μm. Error bars show ±SD (n = 5). WT, wild type.

FIG. 9.

DHHC3 and -7 are necessary for GPCR-mediated signal transduction through the Gα_q palmitoylation. (A) α_1A-AR activation leads to CREB phosphorylation. HeLa cells transiently transfected with α_1A-AR were treated with 50 μM phenylephrine with or without 50 μM prazosin for 5 min. The cell lysates were treated with 10% trichloroacetic acid, and the resulting precipitates were subjected to immunoblotting with antibodies to CREB and Ser133 pCREB. (B to D) HeLa cells were transiently transfected with α_1A-AR and indicated plasmids or siRNAs. (B) α_1A-AR-mediated CREB phosphorylation requires palmitoylated Gα_q. Knockdown (KD) of Gα_q and Gα₁₁ blocked α_1A-AR-induced CREB phosphorylation, and RNAi-resistant Gα_q (WT), but not palmitoylation-deficient Gα_q (CS), rescued α_1A-AR-mediated CREB phosphorylation. (C) Knocked down DHHC3 and DHHC7 (DHHC3/7 KD) inhibited α_1A-AR-induced CREB phosphorylation, indicating that DHHC3 and -7 are essential for the GPCR-signaling pathway through Gα_q palmitoylation. pCREB/CREB ratios are indicated in the graph. Error bars show ±SD (n = 3). **, P < 0.01. (D) HeLa cells expressing α_1A-AR were treated with phenylephrine and stained with anti-pCREB antibody (red). Note that α_1A-AR stimulation induced robust phosphorylation of CREB in the nucleus. The defect of CREB phosphorylation by knocked down human DHHC3 and human DHHC7 (hDHHC3/7 KD) was rescued by GFP-mDHHC3 (WT; green) but not catalytically inactive GFP-mDHHC3 (C157S) [mDHHC3 (CS)]. Scale bar, 20 μm. (E) The HeLa cells cotransfected with α_1A-AR and the indicated siRNAs were treated with phenylephrine for 10 min, followed by IP₃ extraction. Average IP₃ production of three independent experiments is indicated in the graph. Agonist-induced IP₃ production was specifically blocked by knockdown of DHHC3 and -7 (KD 3/7), indicating that DHHC3 and -7 are necessary for the α_1A-AR-mediated GPCR signaling pathway. Error bars show ±SD (n = 3). **, P < 0.01; n.s., not significant (P = 0.20).