Localization of G_iα in the centrosomes. (A, top) Confocal images of live HeLa cells transfected with rat G_iα1-YFP, RGS14-CFP, and merged image. (bottom) Cells expressing G_iα1-YFP, RGS14-CFP, or YFP vector control. (B) HeLa cells expressing rat G_iα1-YFP were stained with anti-RGS14, anti–γ-tubulin, or ninein antibody, followed by Alexa Fluor 568–conjugated secondary antibodies. The right column shows merged images of G_iα1-YFP and endogenous centrosome protein staining. (C) G_iα2-YFP and G_iα3-YFP also localize in the centrosomes. HeLa cells transfected with human G_iα1-YFP, G_iα2-YFP, or G_iα3-YFP were immunostained with anti-pericentrin antibody, followed by Alexa Fluor 568–conjugated anti–rabbit antibody. The right column shows merged images of G_iα-YFP proteins and endogenous pericentrin staining. (D) HeLa cells transfected with EE-G_iα2 were immunostained with anti-EE monoclonal and Alexa Fluor 568–conjugated anti–mouse antibodies. The control cells were incubated only with the Alexa Fluor 568–conjugated secondary antibody, omitting the anti-EE antibody. Cells were then stained with anti-ninein and Alexa Fluor 488–conjugated anti–rabbit antibodies. The right column shows merged images of EE-G_iα2, ninein, and Hoechst 33342 DNA staining. The exact same confocal settings were used to acquire the images shown. None of the Alexa Fluor–conjugated secondary antibodies used (1:1,000 dilution; 45 min to 1 h of incubation time) resulted in any substantial staining of the cells (not depicted) excluding the possibility of nonspecific staining by secondary antibodies. The arrows indicate colocalization. Bar, 10 μm.

Localization of endogenous G_iα in the centrosomes and midbody. (A) NIH3T3 and HeLa cells were immunostained with anti-pericentrin/Alexa Fluor 488–conjugated secondary antibodies, and then with anti-G_iα1 (or -G_iα2) monoclonal/Alexa Fluor 568–conjugated secondary antibodies. The cell cycle phase is indicated as M (mitotic), G1, S, or G2. For HeLa cells, only G2 phase cells are shown. (B) NIH3T3 cells were immunostained as described in A. Cells were also immunostained with mouse anti–γ-tubulin antibody, followed by Alexa Fluor 488 (for G_iα3 double staining) or 568 (for RGS14 double staining) conjugated anti–mouse antibody. Cells were subsequently stained with rabbit anti-G_iα3 or -RGS14 antibody, followed by Alexa Fluor 568– or 488–conjugated anti–rabbit antibody. The arrowheads indicate centrosomes. Certain centrosomes appear larger because they are overexposed to capture the much less prominent midbody staining. Twofold magnification of the midbody is shown in insets. The merge shows merged images of Alexa Fluor 488, and Hoechst 33342 DNA staining. Bar, 10 μm.

Interaction between RGS14-CFP and G_iα1-YFP in the centrosomes. (A) RGS14-CFP fusions containing various deletions and point mutations were generated. RGS domain (aa 67–184), two Raf-like Ras binding domains (RBDs; aa 303–374 and 376–445), Rap- interacting domain (RID; aa 300–427), and GoLoco domain (Loco; aa 500–522) are indicated as filled boxes. Wild-type RGS14-CFP construct is designated as HC30. A mutant lacking N-terminal 66 amino acids and the RGS domain is named as HC31. HC32 construct contains only the RID domain. HC33 lacks a part of the GoLoco domain and C-terminal 25 amino acids. The point mutants defective in GAP, GDI, and both activities are named as HC34, HC35, and HC36, respectively. Asterisks indicate that point mutations were introduced. (B) Acceptor photobleaching was performed with wild-type rat G_iα1-YFP (or the QL mutant) and various RGS14-CFP constructs. NB* and cyt** stand for nonbleached centrosomes and cytoplasmic bleaching, respectively. The FRET efficiencies were averaged for comparison, and the FRET result was accepted as truly positive only when the average FRET efficiency was considerably higher than that of the negative control. (C) Confocal images of live HeLa cells expressing the HC32 construct (RID-CFP) and G_iα1-YFP. The merged image shows colocalization of RID-CFP and G_iα1-YFP in the centrosomes. Bar, 10 μm.

Silencing of G_iα1, G_iα2, and G_iα3 in HeLa cells using siRNA. (A) HeLa cells transfected with control siRNA (siC) or G_iα1-3 siRNA (siG) were immunoblotted with anti-G_iα1, -G_iα2, -G_iα3 or -actin antibody. The actin was used as loading control. (B) HeLa cells transfected with control siRNA (si-control) or G_iα1-3 siRNA (si-Gi1-3) were stained with anti-G_iα1 or -G_iα2 antibody, followed by Alexa Fluor 568–conjugated anti–mouse secondary antibody. Cells were subsequently stained with anti-ninein antibody, followed by Alexa Fluor 488–conjugated anti–rabbit secondary antibody. DNA was stained with Hoechst 33342. The arrows indicate centrosomes. Three-fold magnification of the centrosomes is shown in the insets. Merged images are of Alexa Fluor 488, Alexa Fluor 568, and Hoechst 33342. The confocal microscope settings were kept constant, and images were equally treated within each series of images for fair comparison between control siRNA and G_iα1-3 siRNA. Bars, 10 μm.

Reduced RGS14 expression causes mainly cytokinesis defects in HeLa cells. (A) A semiquantitative RT-PCR was performed to examine the mRNA level of RGS14 in different cell lines such as B lymphocyte cell line, HS-Sultan (HSS), HeLa, and human embryonic kidney 293 (HEK). RNAs were isolated from these cell lines, as well as from HeLa cells transfected with control siRNA (siC) or RGS14 siRNA (siR). Con, control RT-PCR performed in the absence of cDNA. Human β-actin primers were used to monitor cDNA synthesis. (B) Time-lapse videomicroscopy was performed on HeLa cells transfected with RGS14 siRNA as described earlier. Eight snapshots of DIC images chosen from 24 h time-lapse images for si-R14 a and si-R14b are shown. Time shown indicates the hours and minutes lapsed from the beginning of recording. Two representative snapshots of cytokinesis defects resulting from reduced expression of RGS14 are shown (si-R14a and b; Videos 9 and 10). (C) MTs of si-control- or si-RGS14–transfected cells were stained with anti–α-tubulin/Alexa Fluor 488–conjugated secondary antibodies and Hoechst 33342. In addition to a representative image of an intercellular bridge MT staining in control cells, three representative images of defective MT staining in si-RGS14–transfected cells are shown (si-R14a–c) as merged images of Alexa Fluor 488 and DNA staining. The si-R14 images shown are the projection images of 8–12 0.2-μm-thick optical sections, which are generated by the Leica maximal projection module. (bottom) Sixfold magnification of the midbody area. Three independent siRNA knockdown experiments were performed showing the same results. Videos 9 and 10 are available at http://www.jcb.org/cgi/content/full/jcb.200604114/DC1. Bar, 10 μm.

Epifluorescence images of HeLa cells expressing YFP fusions of G_iα proteins during cytokinesis. HeLa cells transfected with vector-YFP, human G_iα1-YFP, G_iα2-YFP, or G_iα3-YFP were subjected to videomicroscopy. Representative snapshots showing expression of these proteins during cytokinesis were taken from the 481 images of time-lapse videomicroscopy (3-min intervals for 24 h). To ensure that these cells underwent cytokinesis, we followed cell division starting from rounding up of cells (indication of cells entering mitosis) to completion of cytokinesis (reattachment of cells). Compared with YFP expressed from the control vector, all three G_iα-YFP fusions showed very strong expression at the junction between two daughter cells (arrows).

FRET assays. (A) HeLa cells expressing RGS14-CFP and rat G_iα1-YFP were subjected to acceptor photobleaching. A representative image from >20 experiments is shown. Only one (arrow) of the two centrosomes expressing both CFP and YFP fusions was photobleached. The nonphotobleached centrosome (arrowhead) did not show any substantial changes in fluorescence intensity. Pre and post stand for before and after photobleaching. (bottom) Twofold magnification. Fluorescence intensities of Pre (open bar) and Post (hatched bar) are shown in the graph. (B) Ratio imaging of RGS14-CFP and G_iα1-YFP was obtained in live HeLa cells using Leica sensitized emission routine. A representative FRET image of three experiments is shown. Twofold magnification of the centrosomes is shown in the insets. FRET intensities are encoded by using the color bar scale shown on the right. Colors range between blue (lowest FRET) and red (highest FRET). Bar, 10 μm.

Perturbed G_iα expression or function leads to defective cell division. (A) Time-lapse videomicroscopy was performed with HeLa cells expressing YFP vector control or various human G_iα-YFP fusions. At least two independent transfections were performed for each construct. Recordings of 20 dividing cells were analyzed for each experiment. Epifluorescence images were captured at 3-min intervals for 24 h. The time shown indicates the hours and minutes lapsed from the beginning of recording. Eight snapshots chosen from the 481 time-lapse images are shown for YFP vector, Giα2QL-YFP, Giα3-YFP, and Giα1QL-YFP (Videos 1–4). (B) NIH3T3 cells treated with 100 ng/ml of PTX for 3 h were subjected to time-lapse videomicroscopy. DIC images were acquired at 3-min intervals for 24 h. (top row) Three different dividing cells that are representative of the ∼100 untreated control cell divisions observed. The red lines are drawn to show dividing cells aligned at a nearly 180° angle. (bottom row) Three different PTX-treated cells undergoing cytokinesis. Approximately 10% of PTX-treated cells readily showed intercellular bridge defects (Videos 5 and 6). (C) MTs of NIH3T3 control and PTX-treated cells were stained with anti–α-tubulin/Alexa Fluor 488–conjugated secondary antibodies and Hoechst 33324. The merge shows DIC, Alexa Fluor 488, and DNA staining. The arrow in the DIC image indicates midbody. Red lines are drawn to show MT alignment in the intercellular bridge. (D) NIH3T3 control and PTX-treated cells were stained as described in C. A representative image of multinucleated cells treated with PTX is shown. Confocal images of stained cells were obtained from two independent experiments (n = 194 and 200 for control; n = 334 and 251 for PTX-treated cells) to quantitate the number of multinucleated cells as shown in the histogram. Error bars represent the SD. Bars, 10 μm. Videos 1–6 are available at http://www.jcb.org/cgi/content/full/jcb.200604114/DC1.

Reduced G_iα expression causes mainly cytokinesis defects in HeLa cells. (A) Time-lapse videomicroscopy was performed on HeLa cells transfected with control siRNA or G_i1-3 siRNA, as described in Materials and methods. Two independent siRNA transfections were performed. Recordings of 40–50 dividing cells were analyzed for each experiment. Eight snapshots of DIC images chosen from 24-h time-lapse images are shown for control, si-G_i1-3a, and si-G_i1-3b. The time shown indicates the hours and minutes lapsed from the beginning of the recording. Two representative snapshots of cytokinesis defects resulting from reduced expression of the G_iα isoforms are shown (si-G_i1-3a and si-G_i1-3b; Videos 7 and 8). (B) MTs of si-control– or si-G_i1-3–transfected cells were stained with anti–α-tubulin/Alexa Fluor 568–conjugated secondary antibodies and Hoechst 33324. Three representative images of defective intercellular bridge are shown (si-Gi1-3a–c). The merge shows DIC, Alexa Fluor 568, and DNA staining. The red arrow in the DIC image indicates midbody. (C) A representative image of a multinucleated cell and a cell with micronuclei (indicated by an arrow) is shown. The multinucleated cells from two independent siRNA transfections were counted from confocal images of stained cells (n = 470 and 405 for control; n = 467 and 559 for Gi1-3 knockdown). The average percentage of multinucleated cells is shown in the bar graph. (D) Flow cytometric analysis of si-control– or si-G_i1-3–transfected cells. Three independent siRNA transfections were performed, followed by flow cytometry analysis of Hoechst 33342–stained live cells. A representative result from three experiments is shown. The data are plotted with DNA area against DNA width as measured by Hoechst 33342 incorporation (both on linear scales) to separate the population of multinucleated cells and possible doublets (marked with the large red box) from that of normal cells (marked with the small red box). The percentage of multinucleated cells was calculated by FlowJo (v.6.2.1). The average percentage of multinucleated cells from three experiments is shown in the bar graph. Error bars represent the SD. Videos 7 and 8 are available at http://www.jcb.org/cgi/content/full/jcb.200604114/DC1. Bars, 10 μm.