Accumulation of glucosylceramide in the absence of the beta-glucosidase GBA2 alters cytoskeletal dynamics

Abstract

Glycosphingolipids are key elements of cellular membranes, thereby, controlling a variety of cellular functions. Accumulation of the simple glycosphingolipid glucosylceramide results in life-threatening lipid storage-diseases or in male infertility. How glucosylceramide regulates cellular processes is ill defined. Here, we reveal that glucosylceramide accumulation in GBA2 knockout-mice alters cytoskeletal dynamics due to a more ordered lipid organization in the plasma membrane. In dermal fibroblasts, accumulation of glucosylceramide augments actin polymerization and promotes microtubules persistence, resulting in a higher number of filopodia and lamellipodia and longer microtubules. Similar cytoskeletal defects were observed in male germ and Sertoli cells from GBA2 knockout-mice. In particular, the organization of F-actin structures in the ectoplasmic specialization and microtubules in the sperm manchette is affected. Thus, glucosylceramide regulates cytoskeletal dynamics, providing mechanistic insights into how glucosylceramide controls signaling pathways not only during sperm development, but also in other cell types.

Fig 1. GBA2 is expressed in Sertoli cells.

(A) GBA2 expression in testis. Total protein lysates were probed with a GBA2-specific antibody (2F8) on a Western blot. Beta-tubulin was used as a loading control. +/+: wild-type;-/-: GBA2 knockout. (B) GBA2 expression in mouse sperm. Total protein lysates were probed with a GBA2-specific antibody (2F8) on a Western blot. Heterologously expressed HA-tagged GBA2 was used as a positive control, beta-tubulin as a loading control. (C) Immunohistochemical analysis of GBA2 in testis. Testis cross-sections were labeled with a GBA2-specific antibody (pcGBA2, red) and an anti-beta tubulin III antibody (green) as a marker for Sertoli cells. DAPI was used to label the DNA. Scale bars are indicated. (D) GBA2 expression in P7 Sertoli cells. See (B). (E-G) GBA2 expression during development. (E) See (C) for testis cross-sections from P7, P21, and adult wild-type mice. (F) Representative Western blot using protein lysates from wild-type testis at P7, P21, and adult mice (ad). Heterologously expressed HA-tagged GBA2 has been used as a positive control and beta-tubulin III as a Sertoli-cell marker (TubIII). Protein lysates from wild-type (+/+) and GBA2 knockout-brain (-/-) are shown as controls. (G) Quantification of GBA2 protein expression. Expression levels of GBA2 have been normalized to beta-tubulin III. Data are presented as mean ± S.D.; n numbers and p values calculated using One-Way ANOVA are indicated. (H) GBA2 activity during sperm development. Beta-glucosidase activity was measured in protein lysates from wild-type testis at pH 6 using 1.67 mM of the artificial substrate 4-methylumbelliferyl-beta-D-glucopyranoside. Data are presented as mean ± S.D.; n numbers and p values calculated using One-Way ANOVA are indicated.

Fig 2. Lack of GBA2 results in accumulation of GlcCer.

(A) Quantitative analysis of neutral sphingolipids in adult testis. +/+: wild-type;-/-: GBA2 knockout. LCB: long-chain bases, Cer: ceramides, HexCer: hexosylceramides, Spm: sphingomyelins. (B) See (A) for P7 Sertoli cells. (C) See (A) for sperm. (D) Quantitative analysis of HexCer in adult testis. Lipids are classified according to their acyl chain-length. (E) See (D) for sperm. (F) See (D) for very long chain fatty acids. All data are presented as mean ± S.D.; n numbers and p values calculated using One-Way ANOVA are indicated.

Fig 3. Lack of GBA2 causes globozoospermia and cytoskeletal defects in the testis.

(A) Immunofluorescent labeling of wild-type (+/+) and GBA2 knockout-sperm (-/-). Fluorescent peanut lectin (green), a mitotracker (red), and DAPI (blue) was used to label the acrosome, the mitochondria in the sperm flagellum, and the DNA, respectively. Scale bars are indicated. (B) Immunofluorescent labeling of the cytoskeleton in adult wild-type (+/+) and GBA2 knockout-testis (-/-). Microtubules were labeled using an anti-beta tubulin antibody (red), F-actin using Alexa Fluor 488 Phalloidin (green), and the DNA using DAPI (blue). Defects in the F-actin structure are highlighted with arrows. Scale bars are indicated. (C) Immunofluorescent labeling of germ and Sertoli cells isolated from adult wild-type (+/+) and GBA2 knockout-testis (-/-). Cells were labeled with an anti-beta tubulin III antibody (red) as a marker for Sertoli cells, Alexa Fluor 488 Phalloidin (green) to label F-actin, and DAPI to label the DNA (blue). Scale bars are indicated. (D) See (C) for P7 Sertoli cells.

Fig 4. Cytoskeletal defects in GBA2 knockout-testis already occur during the first spermatogenic wave.

(A) Immunofluorescent labeling of the cytoskeleton in wild-type (+/+) and GBA2 knockout-testis (-/-) at P7. Microtubules have been labeled using an anti-beta tubulin III antibody (red), F-actin using Alexa Fluor Phalloidin 488 (green), and the DNA using DAPI (blue). Scale bars are indicated. (B) See (A) for P21. (C) See (A) for P23. (D) See (A) for P34. (E) Development of the manchette in spermatids. The manchette was stained with beta-tubulin (red), DNA was labeled with DAPI (blue). Different developmental stages are indicated. (F) Manchette length. The manchette length of spermatids from wild-type (+/+) and GBA2 knockout-mice (-/-) was determined using ImageJ. At least 30 cells have been analyzed per genotype. Data are shown as mean ± S.D.; n numbers and p values calculated using One-Way ANOVA are indicated.

Fig 5. Acrosome development.

Immunofluorescent labeling of the acrosome using fluorescent peanut lectin (green) in P21, P23, and P34 testes. Microtubules were labeled with an anti-beta tubulin antibody (red), DAPI (blue) has been used to stain the DNA. Scale bars are indicated.

Fig 6. Dermal fibroblasts from GBA2 knockout-mice also display cytoskeletal defects.

(A) GBA2 expression in dermal fibroblasts from adult mice. Total protein lysates were probed with a GBA2-specific antibody (2F8) on a Western blot. Heterologously expressed HA-tagged GBA2 was used as a positive control, calnexin (Clnx) as a loading control. (B) Accumulation of GlcCer in GBA2 knockout-fibroblasts. Thin layer chromatography (TLC) analyzing glycosphingolipids from wild-type (+/+) and GBA2 knockout-fibroblasts (-/-). Representative TLC analysis for neutral sphingolipids. GlcCer: glucosylceramide, LacCer: lactosylceramide, Spm: sphingomyelin. GlcCer levels were quantified by densitometry and are presented as mean ± S.D. The fold change in GlcCer levels in GBA2 knockout-fibroblasts was calculated. (C) Fluorescent labeling of the cytoskeleton in dermal fibroblasts from wild-type (+/+) and GBA2 knockout-mice (-/-). Cells were transfected with lifeact (green) to visualize F-actin and with EB3-cherry to visualize microtubules. Scale bars are indicated. (D) Fluorescent labeling of F-actin in dermal fibroblasts from wild-type (+/+) and GBA2 knockout-mice (-/-). Cells were seeded on CYTOO chips with micropatterns that are coated with fluorescently-labeled fibronectin (purple). F-actin was stained using Alexa Fluor Phalloidin 488 (green) and the DNA was stained with DAPI (blue). Scale bars are indicated. (E) Analysis of cytoskeletal structures. Cells were seeded on the crossbow shape. The number of cells containing filopodia or lamellipodia (left) and the average number of filopodia or lamellipodia per cell (right) were determined. (F) Gene expression-analysis. The mRNA expression level of Cdc42, Rac1, and Rho was analyzed by qRT-PCR. (G) Protein expression-analysis. Total protein lysates were probed with a GBA2- (2F8), a Cdc42-, and a Rac1-specific antibody on a Western blot. Calnexin (Clnx) was used as a loading control. (H) Quantification of protein expression based on (G). (I) Quantification of actin turnover in dermal fibroblasts. Expression levels of G- and F-actin in wild-type (+/+) and GBA2 knockout-fibroblasts (-/-) were determined using Western blot-analysis. Ratio of F-actin/G-actin for wild-type and GBA2 knockout-fibroblasts is expressed relative to the control. (J) See (I) for testis. (K-M) Analysis of microtubule dynamics in dermal fibroblasts from wild-type (+/+) and GBA2 knockout-mice (-/-). (K) Expression of EB3-cherry in dermal fibroblasts. Cells were transfected with EB3-cherry and microtubule dynamics were analyzed. Representative tracks of growing microtubule plus-ends are indicated with white lines. (L) Microtubule growth rate. Wild-type (+/+) and GBA2 knockout-fibroblasts (-/-) were transfected with EB3-cherry and the growth rate of growing plus-ends was analyzed. Per genotype, n = 3 animals with a minimum of 7 cells and 10 tracks per cell were analyzed. Data are presented as mean ± S.D. (M) see (L) for microtubule persistence. For all bar graphs, data are shown as mean ± S.D.; n numbers and p values calculated using One-Way ANOVA are indicated.

Fig 7. The absence of GBA2 affects cell migration.

(A-B) Wound-healing assay to analyze cell migration of dermal fibroblasts from wild-type (+/+) and GBA2 knockout-mice (-/-). (A) Representative images at different time points after initiating the assay. Scale bars are indicated. (B) Analysis of cell migration. The rate of cell migration has been analyzed. Average data points for wild-type (+/+) and GBA2 knockout-fibroblasts (-/-) for different time points are shown; n numbers and p values using One-Way ANOVA are indicated.

Fig 8. NB-DNJ treated fibroblasts also display cytoskeletal defects.

Dermal fibroblasts have been treated for 48 h with 2 μM NB-DNJ. (A) GBA activity at pH 4 (GBA1) and pH 6 (GBA2). Beta-glucosidase activity was measured using 1.67 mM of the artificial substrate 4-methylumbelliferyl-beta-D-glucopyranoside. Data are presented as mean ± S.D. N numbers are indicated. (B) Lipid accumulation. Left: Thin layer chromatography (TLC) analyzing glycosphingolipids from wild-type fibroblasts in the presence (+) and absence of (-) of 2 μM NB-DNJ. Representative TLC analysis for neutral sphingolipids. GlcCer: glucosylceramide, LacCer: lactosylceramide, Spm: sphingomyelin. Right: Quantification of GlcCer. Data are presented as mean ± S.D.; n numbers and p values using One-Way ANOVA are indicated. (CFluorescent labeling of F-actin in dermal fibroblasts. Cells were seeded on CYTOO chips on micropatterns that are coated with fluorescently-labeled fibronectin (purple). F-actin was stained using Alexa Fluor Phalloidin 488 (green) and the DNA was stained with DAPI (blue). Scale bars are indicated. (D) Wound-healing assay to analyze cell migration. Average data points for wild-type dermal fibroblasts at different time points are shown; n numbers and p values using One-Way ANOVA are indicated.

Fig 9. Accumulation of GlcCer alters lipid packaging of the plasma membrane

Analysis of lipid packaging in GPMVs isolated from wild-type (+/+, black line), GBA2 knockout-fibroblasts (-/-, red line), and wild-type fibroblasts treated for 48 h with 2 μM NB-DNJ (blue line). GPMVs were labeled with 5 μM laurdan and analyzed by fluorescence spectroscopy. Representative emission spectra normalized to the maximum are shown. GP index has been calculated and presented as mean ± S.D. (p = 0.001); n numbers are indicated in brackets.