Knockout of the Golgi stacking proteins GRASP55 and GRASP65 impairs Golgi structure and function

Abstract

Golgi reassembly stacking protein of 65 kDa (GRASP65) and Golgi reassembly stacking protein of 55 kDa (GRASP55) were originally identified as Golgi stacking proteins; however, subsequent GRASP knockdown experiments yielded inconsistent results with respect to the Golgi structure, indicating a limitation of RNAi-based depletion. In this study, we have applied the recently developed clustered regularly interspaced short palindromic repeats (CRISPR)/Cas9 technology to knock out GRASP55 and GRASP65, individually or in combination, in HeLa and HEK293 cells. We show that double knockout of GRASP proteins disperses the Golgi stack into single cisternae and tubulovesicular structures, accelerates protein trafficking, and impairs accurate glycosylation of proteins and lipids. These results demonstrate a critical role for GRASPs in maintaining the stacked structure of the Golgi, which is required for accurate posttranslational modifications in the Golgi. Additionally, the GRASP knockout cell lines developed in this study will be useful tools for studying the role of GRASP proteins in other important cellular processes.

FIGURE 1:

Construction of GRASP55 and GRASP65 single-knockout cells. (A) sgRNAs designed to target GRASP55 and GRASP65 using CRISPR/Cas9-mediated gene deletion. The translation initiation ATG codon is indicated and referred to as coding sequence 1 (for A); exons are indicated as boxes and introns indicated by a line, with the number of nucleotide at the splicing borders indicated. sgRNAs sequences and relative locations are indicated as lines above the exons of the gene. (B) Immunofluorescence images of cell populations transfected with sgRNAs to GRASP55 (left panels) or GRASP65 (right panels). HeLa cells were transfected with GFP-Cas9 plasmids containing sgRNAs against either GRASP55 or GRASP65 and selected for GFP expression by flow cytometry. The Golgi morphology of GRASP knockout cells was assessed by immunofluorescence microscopy for either GRASP55 or GRASP65 costained with GM130. Scale bars are 10 µm. (C) Quantification of Golgi morphology in GRASP-positive (arrows) and GRASP-negative (asterisks) cells in B. Blinded determination of the Golgi morphology of 300 cells from each sample were quantified across three biological replicates. Error bars represent SEM (Cheeseman et al., 2012). A Student’s t test was performed to determine statistical significance. *p < 0.05.

FIGURE 2:

GRASP55 deletion has minor effects on the Golgi structure. (A) Western blots of Golgi proteins in GRASP55 knockout HeLa cells. Wild-type and representative GRASP55 knockout clones from three separate sgRNAs (T1, T2, and T3) were lysed and blotted for GRASP55/65, Golgin-45, and GM130. (B) Quantification of A for the relative levels of GRASP65, Golgin-45, and GM130 in GRASP55 knockout cells. Error bars represent SEM. (C) Immunofluorescence of GRASP55 knockout clones stained for GM130 and TGN46. The lower three rows are increased magnifications of the Golgi in a single cell. Scale bars are 10 µm. (D) Colocalization of GM130 and TGN46 quantified by the Pearson’s correlation coefficient of z-stacks from GRASP55 knockout clones from C. Error bars represent SEM. (E) Quantification of Golgi fragmentation in GRASP55 knockout clones in C. Blinded determination of the Golgi morphology of 300 cells from each sample were quantified across three biological replicates. Error bars represent SEM. A Student’s t test was performed to determine statistical significance. *p < 0.05.

FIGURE 3:

GRASP65 deletion does not cause Golgi ribbon unlinking. (A) Western blots of Golgi proteins in GRASP65 knockout HeLa cells. Wild-type and representative GRASP65 knockout clones from two separate sgRNAs (T1 and T2) were analyzed by Western blot for GRASP55/65, Golgin-45, and GM130. (B) Quantification of A for the relative levels of GRASP55, Golgin-45, and GM130 in GRASP65 knockout cells. Error bars represent SEM. (C) Immunofluorescence microscopy of GRASP65 knockout clones stained for GM130 and TGN46. The lower three rows are increased magnifications of a single cell’s Golgi. Scale bars are 10 µm. (D) Colocalization of GM130 and TGN46 quantified by the Pearson’s correlation coefficient of z-stacks from GRASP65 knockout clones from C. Error bars represent SEM. (E) Quantification of Golgi fragmentation in GRASP65 knockout clones in C. Blinded determination of the Golgi morphology of 300 cells from each sample were quantified across three biological replicates. Error bars represent SEM. A Student’s t test was performed to determine statistical significance. *p < 0.05.

FIGURE 4:

Double deletion of GRASP55 and GRASP65 results in Golgi fragmentation. (A) Western blots of Golgi proteins in GRASP55 and GRASP65 double-knockout cells. Wild-type and two representative GRASP55 and GRASP65 double-knockout clones (DKO-C1 and DKO-C2) were analyzed by Western blot for GRASP55/65, Golgin-45, and GM130. (B) Quantification of the relative levels of Golgin-45, and GM130 in GRASP double-knockout cells in A. Error bars represent SEM. (C) Immunofluorescence microscopy of GRASP55/65 double-knockout cells stained for GM130 and TGN46. The lower three rows are increased magnifications of a single cell’s Golgi. Scale bars are 10 µm. (D) Colocalization of GM130 and TGN46 quantified by the Pearson’s correlation coefficient of z-stacks from GRASP doubl-knockout clones from C. Error bars represent SEM. (E) Quantification of cells with fragmented Golgi from GRASP double-knockout clones in C. Blinded determination of the Golgi morphology of 300 cells from each sample were quantified across three biological replicates. Error bars represent SEM. A Student’s t test was performed to determine statistical significance. *p < 0.05; **p < 0.01; ***p < 0.001.

FIGURE 5:

Double deletion of GRASP55 and GRASP65 proteins impairs Golgi stack formation. (A) High-resolution AiryScan confocal immunofluorescence images of GM130 and TGN46 in HeLa wild-type and GRASP single- and double-knockout clones. The lower three rows are increased magnifications of a single cell’s Golgi. Scale bar is 10 µm. (B) High-resolution AiryScan confocal immunofluorescence images for GM130 and TGN46 in HeLa wild-type and GRASP single- and double-knockout clones after 4 h treatment with 100 ng/ml nocodazole. The lower three rows are increased magnifications of a single cell’s Golgi. Scale bar is 5 µm. (C) Quantification of GM130 and TGN46 colocalization in B. Pearson’s correlation coefficient in wild-type and GRASP55 and GRASP65 double-knockout cells from 20 cells across three biological replicates and quantified using the ImageJ2 coloc2 plug-in with z-stacks. Error bars represent SEM. A Student’s t test was performed to determine statistical significance. (D) Electron micrographs from wild-type and GRASP single- and double-knockout HeLa clones. Note the reduced number of cisternae in GRASP single-knockout cells and the disorganized Golgi membranes in double-knockout cells. Scale bar is 200 nm. (E) Quantification of the proportion of cells exhibiting distinguishable Golgi stacks vs. disorganized membranes in D. (F) Quantification of well-organized Golgi stacks per cell in wild-type and GRASP single- and double-knockout clones from D. (G) Quantification of the length of well-organized Golgi stacks in wild-type and GRASP single-knockout clones from D. Double-knockout cells were not quantified due to the lack of well-organized stacks. (H) Quantification of the number of cisternae per stack in wild-type and GRASP single-knockout clones from D. Double-knockout cells were not quantified due to the lack of well-organized stacks. In all EM pictures, E–H, at least 20 cells across three biological replicates were quantified. Error bars represent SEM. A Student’s t test was performed to determine statistical significance. ***p < 0.001.

FIGURE 6:

GRASP deletion accelerates protein trafficking but causes glycosylation defects. (A) GRASP deletion accelerates VSV-G trafficking. HeLa wild-type and GRASP knockout cells were infected with VSV-G-ts045-GFP adenovirus and incubated at 40.5°C for 16 h followed by cyclohexamide treatment for 1 h. Cells were then shifted to 32°C to permit trafficking of VSV-G from the ER to the plasma membrane through the Golgi. Cells were collected at the indicated time points, treated with endoglycosydase H (Endo-H), and analyzed by Western blot for GFP. Note that GRASP deletion increased VSV-G Endo-H resistance by the 15-min time point. (B) Quantification of the percentage of Endo-H resistant (upper band) VSV-G from A. (C) GRASP deletion reduces the amount of sialic acid modifications on the cell surface. Wild-type and indicated GRASP knockout HeLa cells were stained with WGA or MAA lectin without permeabilization. Note the reduced WGA and MAA intensity on GRASP knockout cells. (D) Flow cytometric analysis of WT and GRASP knockout HeLa cells stained with WGA and MAA. The fluorescence intensity of 10,000 cells was determined by flow cytometry across three biological replicates. Error bars represent SEM. A Student’s t test was used to determine statistical significance. (E) GRASP deletion reduces glycosylation of Lamp1 and Lamp2 glycoproteins. HeLa wild-type and GRASP knockout cells were analyzed by Western blots for Lamp1 and Lamp2; note their increased migration shift on the gel when GRASPs are deleted. (F) GRASP deletion impacts glycolipids at the cell surface. Wild-type and indicated GRASP knockout HeLa cells were stained with Shiga toxin B (Shtx) or cholera toxin B (CtxB) without permeabilization. Note the reduced ShTx intensity but increased CtxB intensity in GRASP knockout cells. (G) Quantitation of cell-surface ShTx intensity from cells in F. (H) Quantification of cell-surface cholera toxin from cells in F. For G and H, the mean intensity of 200 cells from each condition was quantified in maximum projections across three biological replicates. Error bars represent SEM. A Student’s t test was performed to determine statistical significance. *p < 0.05; ***p < 0.001.