Rapid degradation of GRASP55 and GRASP65 reveals their immediate impact on the Golgi structure

Abstract

GRASP55 and GRASP65 have been implicated in stacking of Golgi cisternae and lateral linking of stacks within the Golgi ribbon. However, RNAi or gene knockout approaches to dissect their respective roles have often resulted in conflicting conclusions. Here, we gene-edited GRASP55 and/or GRASP65 with a degron tag in human fibroblasts, allowing for induced rapid degradation by the proteasome. We show that acute depletion of either GRASP55 or GRASP65 does not affect the Golgi ribbon, while chronic degradation of GRASP55 disrupts lateral connectivity of the ribbon. Acute double depletion of both GRASPs coincides with the loss of the vesicle tethering proteins GM130, p115, and Golgin-45 from the Golgi and compromises ribbon linking. Furthermore, GRASP55 and/or GRASP65 is not required for maintaining stacks or de novo assembly of stacked cisternae at the end of mitosis. These results demonstrate that both GRASPs are dispensable for Golgi stacking but are involved in maintaining the integrity of the Golgi ribbon together with GM130 and Golgin-45.

Figure 1.

Generation of cell lines expressing endogenous GRASP55 and/or 65 tagged with mAID. (A) Scheme for tagging the endogenous loci of GRASP55 or 65 with 3xFlag and mAID. The DNA sequence encoding 3xFlag-mAID separated by a self-cleaving T2A peptide from the resistance gene (puromycin [Puro] or hygromycin [Hygro]) was inserted by CRISPR-Cas9 gene editing in front of the stop codon of GRASP55 or 65. The amino acids between the insert location and the stop codon were scarlessly repaired by adding the sequence in front of 3xFlag-mAID of the repair donor plasmid. (B) Scheme of the primer sets used for genotyping gene-edited cell lines. The cell line with GRASP55 scarlessly tagged with 3xFlag-mAID and puromycin is referred to as RC55, GRASP65 scarlessly tagged with 3xFlag-mAID and hygromycin as RC65, and both GRASP55 and 65 tagged with 3xFlag-mAID and indicated resistance genes as RC65+55. (C and D) Genomic PCR to genotype heterozygous (−/+) and homozygous (+/+) RC55, RC65, and RC65+55 cell lines for GRASP55 tagged with 3xFlag-mAID or GRASP65-3xFlag-mAID. Parental SV589 cells are shown as negative control (WT). (E) Immunoblot analysis of whole cell lysates of heterozygous (−/+) and homozygous (+/+) RC55, RC65, and RC65+55 cells using antibodies against GRASP55, GRASP65, the degron mAID tag, and GAPDH. (F) SV589 cells and RC55 cells expressing GRASP55-3xFlag-mAID were immunostained for mAID (red) and GRASP55 (green) and labeled for DNA (blue). Scale bar, 10 μm. (G) SV589 cells and RC65 cells stably expressing GRASP65-3xFlag-mAID were immunostained for mAID (red) and GRASP65 (green) and labeled for DNA (blue). Scale bar, 10 μm.

Figure S1.

Long-term TIR1 expression in the absence of auxin leads to partial loss of GRASP55 and 65. (A) RC55, RC65, and RC65+55 cells were treated with doxycycline to induce TIR1 expression for 20 h (long-term dox) before IAA was added for a further 2 h. Cell lysates were subjected to immunoblotting with indicated antibodies. (B) RC65+55 cells stably expressing the Golgi enzyme NAGTI-GFP were treated with doxycycline for 6 h and IAA for an additional 2 h. Fixed cells were immunostained for GRASP55 (red) and GRASP65 (green) and labeled for DNA (blue). Scale bar, 10 µm. (C) RC55, RC65, and RC65+55 cells were treated with doxycycline for 6 h and with IAA for a further 2 h as in B and immunolabeled for mAID (red) and Golgin-84 (green) and stained for DNA (blue). Scale bar, 10 µm. (D) Quantification of the fluorescence signal of mAID in the Golgi region marked by Golgin-84 from C. n = 3 independent experiments with >50 cells analyzed per experiment and condition. ** P < 0.01; *** P < 0.001. Error bars represent mean ± SD. (E) Quantitation of the Golgi area marked by Golgin-84 from C. n = 3 independent experiments with an average number of cells analyzed per experiment: RC55 cells: control: n = 58 cells, +IAA: n = 50; RC65 cells: control: n = 66, +IAA: n = 84; RC65+55 cells: control: n = 74, IAA: n = 58. ns, not significant. Error bars represent mean ± SD. (F) Quantitation of the number of Golgi elements labeled by Golgin-84 from C. n = 3 independent experiments with an average cell number analyzed per experiment: RC55 cells: control: n = 60 cells, +IAA: n = 53; RC65 cells: control: n = 86, +IAA: n = 89; RC65+55 cells: control: n = 64, IAA: n = 70. * P < 0.05; ns, not significant. Error bars represent mean ± SD.

Figure 2.

The auxin IAA induces rapid depletion of mAID-tagged GRASP55 and 65. (A) Homozygous RC55, RC65, and RC65+55 cells inducibly expressing TIR1-2xMyc were treated with doxycycline for 6 h (short-term dox) and then with the auxin IAA for a further 2 h to degrade GRASP55 and/or 65. Cell lysates were immunoblotted with the indicated antibodies. (B) RC55, RC65, and RC65+55 cells were treated with doxycycline for 6 h and with IAA for a further 2 h as in A. Cells were then immunolabeled for GRASP55 (red) and GRASP65 (green) and stained for DNA (blue). Scale bar, 10 µm. (C) Quantitation of the fluorescence signal of GRASP55 and 65 in the Golgi region from B for RC55, RC65, and (S1B) for RC65+55. n = 3 independent experiments with >50 cells analyzed per experiment and condition. ** P < 0.01; *** P < 0.001; ns, not significant. Error bars represent mean ± SD. (D and E) GRASP55 does not compensate for the loss of GRASP65 to stabilize GM130. RC65 cells were treated with doxycycline for 6 h before adding IAA for 2 d (long-term IAA) to deplete GRASP65. GM130 (D) and GRASP55 (E) were then immunoprecipitated from the cell lysates and analyzed by Western blotting. * denotes unspecific band. (F and G) Simultaneous degradation of GRASP55 and 65 causes a delayed reduction of GM130 levels. RC65+55 cells expressing TIR1 for 6 h were treated with IAA. Cell lysates were collected at the indicated time points and subjected to Western blotting with the indicated antibodies. * denotes unspecific band. The intensities of each band are shown in G.

Figure 3.

Acute simultaneous depletion of GRASP55 and 65, but not separate depletion, partially dislocates GM130 from the Golgi. (A) GM130 is partially displaced from the Golgi and relocated to the nucleus after degradation of GRASP55 together with GRASP65, but not after individual depletion of GRASP55 or 65. RC55, RC65, or RC65+55 cells were treated with doxycycline for 6 h and then with IAA for a further 2 h. The cells were then fixed and immunolabeled for GM130 (red) and the cis-Golgi protein Golgin-84 (green) and stained for DNA (blue). The brightness of the GM130 image of RC65+55 cells in the bottom panel was increased to better visualize the redistribution of GM130. The arrow indicates the nuclear signal of GM130. White lines show the position of the line-scan used to measure the fluorescence intensity of GM130 (red) and Golgin-84 (green) across the Golgi and nucleus as shown in the graphs. Scale bar, 10 µm. (B) Quantitation of the GM130 fluorescence signal on the Golgi (marked by Golgin-84) from A. n = 3 independent experiments with >50 cells analyzed per experiment and condition. ** P < 0.01; ns, not significant. Error bars represent mean ± SD. (C) Degradation of GRASP55 and 65 partially delocalizes p115 together with GM130 from the Golgi. RC55, RC65, and NAGTI-GFP–expressing RC65+55 cells were treated as in A and immunostained for p115 (red) and GM130 (green) and labeled for DNA (blue). The Golgi is marked by the GFP fluorescence signal of the Golgi enzyme NAGTI-GFP. White lines show the position of the line-scan used to measure the fluorescence intensity of p115 (red), GM130 (green), and NAGTI-GFP across the Golgi as shown in the graphs. Scale bars, 10 µm. (D) Quantitation of the p115 fluorescence signal on the Golgi marked by GM130 or NAGTI-GFP from C. n = 3 independent experiments with >50 cells analyzed per experiment and condition. ** P < 0.01; ns, not significant. Error bars represent mean ± SD. (E) GM130 is partially degraded by the proteasome after acute double depletion of GRASP55 and 65. RC65+55 cells were treated with doxycycline for 6 h, followed by IAA or IAA plus 10 µM MG132. Cell lysates were collected at the indicated time points and subjected to immunoblotting with the indicated antibodies.

Figure 4.

Acute simultaneous degradation of GRASP55 and 65, but not separate degradation, disrupts the lateral linking of the Golgi ribbon. (A) Schematic illustration of the treatment with doxycycline (dox) and IAA or IAA with nocodazole (noc) washout. (B) FRAP analysis of RC55, RC65, and RC65+55 cells stably expressing the Golgi enzyme NAGTI-GFP. Representative images at the indicated time points are shown, with white boxes indicating the photobleached area of the Golgi. Scale bar, 5 µm. (C) Quantitation of the FRAP results. The recovery rate at each time point was calculated as the ratio of the average intensity of the photobleaching area to that of the adjacent area and then normalized to the closest time point before bleaching. Error bars represent mean ± SEM from the indicated number of cells (n) per condition from three to four independent experiments. (D) Long-term degradation of GRASP55, but not GRASP65, disrupts the integrity of the Golgi ribbon. Scheme of the experiment. RC55 and RC65 cells were treated with doxycycline for 6 h and IAA for a further 48 h before FRAP analysis. FRAP results were quantified as in C. Error bars represent mean ± SEM from the indicated number of cells (n) per condition from three independent experiments.

Figure S2.

GRASP55 and 65 can be degraded in nocodazole-treated cells. (A) Scheme of the experiment. RC55, RC65, and RC65+55 cells were treated with doxycycline (dox) for 6 h. IAA was then added to degrade GRASPs and nocodazole (noc) to depolymerize microtubules and to disperse Golgi stacks. (B) 2 h later, the cells were fixed and immunostained for Golgin-84 (red) and mAID (green) and labeled for DNA (blue). Scale bar, 10 µm. (C) Quantification of the mAID fluorescence signal on the Golgi (labeled by Golgin-84). n ≥ 3 independent experiments with >50 cells analyzed per experiment and condition. *** P < 0.001; **** P < 0.0001. Error bars represent mean ± SD. Scale bar, 10 µm.

Figure 5.

Acute degradation of GRASP55 and/or 65 during interphase does not affect the cis to trans polarization of Golgi stacks. (A and B) Cells were treated with doxycycline (dox) for 6 h, IAA was added for an additional 2 h, and cells were prepared for immunofluorescence analysis. IAA treatment did not change the localization of the cis-Golgi marker Golgin-84 (green) and the trans-Golgi protein Golgin-97 (red) to the perinuclear Golgi ribbon. Scale bar, 10 μm. (C and D) Cells incubated with doxycycline were treated with IAA or IAA with nocodazole (noc) for an additional 2 h to depolymerize microtubules and to disperse the Golgi ribbon into individual stacks. Cells were then immunostained for Golgin-84 (green), Golgin-97 (red), and DNA (blue). Insets show a magnified Golgi stack with the white line marking the line scan of the fluorescence intensities shown in the graphs. Scale bar, 10 µm. Inset scale bar, 1 µm. (E and F) Polarized Golgi stacks reassemble in cells depleted of GRASPs. Cells treated with doxycycline were incubated with IAA and nocodazole for 2 h, followed by BFA to fragment the Golgi stacks. BFA was then removed, and nocodazole (plus dox and IAA) was maintained to allow reformation of individual Golgi stacks. The cells were fixed and immunostained for Golgin-84 (green) and Golgin-97 (red) and labeled for DNA (blue). Insets show a magnified Golgi stack with the white line marking the line scan of the fluorescence intensities shown in the graphs. Scale bar, 10 µm. Inset scale bar, 1 µm.

Figure S3.

Acute degradation of GRASP55 and/or 65 does not affect Golgi stack reformation after BFA washout. (A) Cells were treated with doxycycline (dox) and then incubated with IAA and nocodazole for an additional 2 h. BFA was then added for 30 min to disassemble Golgi stacks. BFA was removed, and nocodazole (plus dox and IAA) was maintained to allow the reformation of individual Golgi stacks. The cells were fixed at the indicated time points and immunostained for Golgin-97 (red) and Golgin-84 (green) and labeled for DNA (blue). Scale bar, 10 µm. (B) Quantitation of the number of Golgin-97–marked Golgi elements per cell from A. n = 3 with an average of >10 cells per time point. Error bars represent mean ± SD. (C) Quantitation of stack formation from A as the percentage of Golgin-97–labeled elements that associate with Golgin-84. n = 3 with >10 cells per time point per condition per experiment. Error bars represent mean ± SD.

Figure 6.

GRASP55 and 65 are dispensable for Golgi stacking. (A) Experiment scheme. TIR1 expression in RC55, RC65, and RC65+55 cells was induced for 6 h with doxycycline (dox), and IAA was added for 2 h to degrade GRASP55 and/or 65. The cells were then treated for 30 min with BFA to disassemble the Golgi. BFA was removed for 2 h to allow Golgi reformation, and the cells were processed for EM. (B) Representative EM images showing stacked Golgi cisternae in parental SV589 cells and upon Golgi reassembly after BFA washout (+IAA → +BFA → BFA washout). Scale bar, 1 µm. (C) EM images of RC55, RC65, and RC65+55 cells after control treatment, after acute degradation of GRASP55 and/or 65 (+IAA) and after degradation and Golgi reassembly upon BFA washout (+IAA → +BFA → BFA washout). Scale bar, 1 µm. (D) Quantitation of the number of cisternae per stack (RC55 cells: control: n = 15 cells, +IAA: n = 16, +IAA BFA w/o: n = 25; RC65 cells: control: n = 19, +IAA: n = 20, +IAA BFA w/o: n = 30; RC65+55 cells: control: n = 44, +IAA: n = 40, +IAA BFA w/o: n = 44; SV589 cells: control: n = 12, +IAA BFA w/o: n = 10 cells). An average of 4.5 stacks per cell was counted. Error bars represent mean ± SD.

Figure 7.

Golgi stacks reform at the end of mitosis in the absence of GRASP55 and/or 65. (A) Scheme of the experiment. Cells released from double thymidine G1/S block were treated with doxycycline (dox) and arrested in prometaphase with the Eg5 kinesin inhibitor STLC. The mitotic cells were treated with IAA to degrade GRASPs, released from STLC block to allow mitotic progression, and processed for EM analysis. (B) Representative EM images showing a midbody between a pair of daughter cells. Scale bar, 1 µm. (C) Representative EM images of Golgi stacks at the end of mitosis/G1 in control and IAA-treated RC55, RC65, and RC65+55 cells. Scale bar, 1 µm. (D) Quantitation of the number of cisternae per stack (RC55 cells: control: n = 15 pairs of daughter cells, +IAA: n = 17; RC65 cells: control: n = 15, +IAA: n = 15; RC65+55 cells: control: n = 10, +IAA: n = 15). An average of 6.6 stacks per pair of daughter cells was counted. Error bars represent mean ± SD.

Figure S4.

GRASP55 and 65 can be degraded in STLC-arrested mitotic cells. (A) Scheme of the approach. Cells were released from double thymidine G1/S block, treated with doxycycline (dox), and then with STLC to arrest cells at prometaphase. The mitotic cells were then treated with IAA for 2 h to degrade GRASPs followed by STLC washout and the addition of dox and IAA to allow mitotic progression and then fixed at different time points. (B) Cells fixed at t = 0 h after STLC washout were immunostained for Flag-tag (red) and Golgin-84 (green) and labeled for DNA (blue). Cells at later time points were immunostained for α-tubulin (red) and Golgin-84 (green) and labeled for DNA (blue). Scale bar, 10 µm.