VAMP4 is required to maintain the ribbon structure of the Golgi apparatus

Abstract

The Golgi apparatus forms a twisted ribbon-like network in the juxtanuclear region of vertebrate cells. Vesicle-associated membrane protein 4 (VAMP4), a v-SNARE protein expressed exclusively in the vertebrate trans-Golgi network (TGN), plays a role in retrograde trafficking from the early endosome to the TGN, although its precise function within the Golgi apparatus remains unclear. To determine whether VAMP4 plays a functional role in maintaining the structure of the Golgi apparatus, we depleted VAMP4 gene expression using RNA interference technology. Depletion of VAMP4 from HeLa cells led to fragmentation of the Golgi ribbon. These fragments were not uniformly distributed throughout the cytoplasm, but remained in the juxtanuclear area. Electron microscopy and immunohistochemistry showed that in the absence of VAMP4, the length of the Golgi stack was shortened, but Golgi stacking was normal. Anterograde trafficking was not impaired in VAMP4-depleted cells, which contained intact microtubule arrays. Depletion of the cognate SNARE partners of VAMP4, syntaxin 6, syntaxin 16, and Vti1a also disrupted the Golgi ribbon structure. Our findings suggested that the maintenance of Golgi ribbon structure requires normal retrograde trafficking from the early endosome to the TGN, which is likely to be mediated by the formation of VAMP4-containing SNARE complexes.

Fig. 1

Knockdown of VAMP4 using RNAi. a HeLa cells were treated with a control siRNA or two separate siRNAs against VAMP4. Total cell extracts were analyzed by Western blotting with an anti-VAMP4 antibody. β-Actin was used as a loading control. b Knockdown of VAMP4 by VAMP4 siRNA-1 was assessed by indirect immunofluorescence using an anti-VAMP4 antibody. Immunoreactivity of the VAMP4 protein is absent in siRNA-knockdown cells. Bar 50 μm

Fig. 2

Depletion of VAMP4 causes the Golgi apparatus to fragment. a HeLa cells were treated with a control siRNA (top) or VAMP4 siRNA-1 (bottom) and then double-stained with a rabbit monoclonal anti-GM130 antibody (left, green) and an anti-VAMP4 antibody (middle, red). The nuclei were counterstained with Hoechst-33342 (blue). Images shown in the z-plane are all maximum-intensity projections of confocal image stacks through the entire cell. Bar, 50 μm. b HeLa cells were treated with a control siRNA (top) or VAMP4 siRNA-1 (bottom) and then double-stained with an anti-Golgin-97 antibody (left, green) and an anti-VAMP4 antibody (middle, red). Images shown in the z-plane are all maximum-intensity projections of confocal image stacks through the entire cell. Bar, 50 μm. c The number of distinct Golgi objects per cell was estimated using fixed acquisition parameters and image thresholding, followed by counting using the “Analyze Particles” function in ImageJ. Values represent the mean ± SE from three independent experiments (n = 300 cells in each case) (*p < 0.01). (Color figure online)

Fig. 3

VAMP4 depletion disrupts the connections between Golgi stacks. a HeLa cells expressing GalNAcT2-GFP were treated with a control siRNA (top) or VAMP4 siRNA-1 (bottom). After 72 h, the area of the Golgi indicated by the arrow was bleached. Fluorescence recovery was followed for 160 s. Representative images at the indicated times are shown. b Fluorescence recovery was quantified by dividing GFP fluorescence within the bleached spot by the fluorescence of a nearby unbleached region in the same Golgi object. Each curve was normalized by the values between a minimum of 0 and a maximum of 1. The rates of recovery in control siRNA-treated cells (black circles) and VAMP4 siRNA-treated cells (white circles) are plotted. The results show the mean fluorescence recovery ± SE from five to six cells from three independent experiments. Bar, 10 μm

Fig. 4

Depletion of VAMP4 converts the Golgi into ministacks. a HeLa cells were treated with a control siRNA (top) or VAMP4 siRNA-1 (bottom) and then double-stained with a rabbit monoclonal anti-GM130 antibody (green) and an anti-Golgin-97 antibody (red). Images shown in the z-plane are all maximum-intensity projections of confocal image stacks through the entire cell. Bar, 50 μm. b An enlarged selection from each of the images above (white squares). Bar, 5 μm. c Transmission electron micrographs of the juxtanuclear region in control siRNA (right) or VAMP4 siRNA-1 (left)-treated HeLa cells. Bar, 1 μm. (Color figure online)

Fig. 5

Anterograde transport in VAMP4-depleted cells. a, b HeLa cells were transfected with control (top) or VAMP4-1 (bottom) siRNAs. Two days after siRNA transfection, the cells were transfected with ts045 VSV-G-GFP and incubated at 40 °C overnight. Cells were shifted to 32 °C for 0 min (a) or 60 min (b), transferred to ice, and exposed to anti-VSV-G before fixation. Total VSV-G was measured using GFP fluorescence (middle), whereas surface VSV-G was measured by staining live cells with a monoclonal antibody against the lumenal domain of VSV-G (right). Depletion of VAMP4 was assessed by staining with an anti-VAMP4 antibody (left). Bar, 50 μm. c Surface and total fluorescence values of the z-stacked images were quantified and expressed as ratios to indicate the relative amount of VSV-G on the plasma membrane. Data represent the mean ± SE from three to four independent experiments (n = 50–100 cells in each case)

Fig. 6

Microtubule integrity is maintained in VAMP4-depleted cells. Microtubules were examined in HeLa cells 72 h after treatment with the control siRNA (top) or VAMP4 siRNA-1 (bottom) and then double-stained with antibodies against GM130 (left, green) and β-tubulin (middle, red). Images shown are all maximum-intensity projections of confocal image stacks in the z-plane through the entire cell. Bar, 50 μm. (Color figure online)

Fig. 7

Golgi fragmentation by depletion of the VAMP4-cognate binding partners. HeLa cells were treated with control, syntaxin 6, syntaxin 16, or Vti1a siRNAs. a Total cell extracts were analyzed by immunoblotting using antibodies against syntaxin 6 (left), syntaxin 16 (middle), or Vti1a (right). β-actin was used as a loading control. b, c, d Golgi structures in siRNA-treated cells were analyzed by immunostaining using antibodies against GM130 (b–d, left), syntaxin 6 (b, middle), syntaxin 16 (c, middle), or Vti1a (d, middle). Images shown in the z-plane are all maximum-intensity projections of confocal image stacks through the entire cell. Bar, 50 μm. e The number of distinct Golgi objects per cell was estimated using fixed acquisition parameters and image thresholding, followed by counting using the “Analyze Particles” function in ImageJ. Values represent the mean ± SE from three independent experiments, n = 100–200 cells in each case. (*p < 0.01)