A dynamic actin cytoskeleton is required to prevent constitutive VDAC-dependent MAPK signalling and aberrant lipid homeostasis

Abstract

The dynamic nature of the actin cytoskeleton is required to coordinate many cellular processes, and a loss of its plasticity has been linked to accelerated cell aging and attenuation of adaptive response mechanisms. Cofilin is an actin-binding protein that controls actin dynamics and has been linked to mitochondrial signaling pathways that control drug resistance and cell death. Here we show that cofilin-driven chronic depolarization of the actin cytoskeleton activates cell wall integrity mitogen-activated protein kinase (MAPK) signalling and disrupts lipid homeostasis in a voltage-dependent anion channel (VDAC)-dependent manner. Expression of the cof1-5 mutation, which reduces the dynamic nature of actin, triggers loss of cell wall integrity, vacuole fragmentation, disruption of lipid homeostasis, lipid droplet (LD) accumulation, and the promotion of cell death. The integrity of the actin cytoskeleton is therefore essential to maintain the fidelity of MAPK signaling, lipid homeostasis, and cell health in S. cerevisiae.

Keywords: Biological sciences; Cell biology.

Graphical abstract

Figure 1

cof1-5-induced actin depolarisation but not mitochondrial fragmentation are VDAC dependent (A and B) Actin phalloidin (red) and DAPI (blue) staining at exponential growth phase (6h) reveals depolarised actin cytoskeleton in the cof1-5 mutant, which is rescued by additional POR1 deletion. Representative microscopy pictures are shown in (A) and cells with polarised actin were quantified in (B). (C) cof1-5 cells have an increased mean cell diameter (as determined with a CASY cell counter). (D) Fluorescence microscopy pictures at exponential growth phase of wt and cof1-5 with and without additional POR1 deletion expressing mitochondrial GFP from a plasmid (pVT100U-mt GFP). (E and F) cof1-5 mutation leads to Por1-dependent vacuole fragmentation as visualised by FM4-64 staining. Representative microscopy images are shown in (E) and a quantification of cells containing multi-lobular vacuoles is depicted in (F). Statistical significance in (B) (C) and (F) was assessed using ordinary one-way ANOVA. See also Figure S1. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 2

cof1-5 mutation triggers growth defect, flocculation, and cell wall alterations, which depend on POR1 (A) Growth performance in liquid culture is reduced in cof1-5 as compared to wild type (Wt) but is restored by additional POR1 deletion. (B) Cultures bearing the cof1-5 mutation sediment quickly when shaking is stopped (flocculation phenotype). Additional POR1 deletion prevents flocculation in cof1-5. (C) Calcofluor white staining detecting chitin exposure at the cell wall confirms flocculation phenotype of cof1-5 cells. (D–F) The cell wall was analyzed by electron microscopy. cof1-5 mutation is associated with a thicker inner cell wall and thinner outer cell wall. EM micrographs are shown in D and quantifications of the inner and outer cell wall are plotted in E and F, respectively. ICW, inner cell wall; OCW, outer cell wall; LD, lipid droplet. Statistical significance in (E) was assessed using Kruskal-Wallis test and in (F) Welch ANOVA was performed. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0 .0001.

Figure 3

Transcriptional changes in the actin mutant act1-159 suggest involvement of MAPK signaling, flocculation and lipid metabolism (A and B) Transcriptional changes of act1-159 vs. ACT1 cells grown to log phase in YPD media were investigated by microarray and plotted in a volcano plot. Gene ontology analysis for the GO-term PROCESS was completed using the GO SLIM mapper function available on the Saccharomyces cerevisiae genome database and upregulated genes clustered within enriched cellular processes are depicted in panel (B). (C) Congo red sensitivity of wild type and act1-159 mutant cells was assessed by a spotting assay using a 10-fold serial dilution series from a starting cell number of 2 x 10⁵. See also Table S1.

Figure 4

cof1-5 mutation triggers activation of the CWI pathway (A and B) Immunoblots detecting Slt2 phosphorylation when grown on glucose and glycerol containing media are shown in A and B, respectively. (C) Pkc1 inhibition by cercosporamide (CSA) administration prevented Slt2 phosphorylation. (D) Porin deletion and overexpression reveal dependence of Slt2 phosphorylation on Por1. (E) cof1-5 mutation triggers loss of plasma membrane integrity, as assessed flow cytometrically with PI positivity at 48 h after inoculation. PI positivity is exacerbated by additional Pkc1 inhibition using the Pkc1-inhibitor cercosporamide or applying additional cell wall stress with calcofluor white (CFW). Combined treatment with CSA and CFW at the same time shows additive effects. (F) POR1 deletion rescues from cof1-5-dependent loss of viability and loss of plasma membrane integrity, whereas SLT2 deletion sensitises to cell death. (G–J) Chronological aging analysis reveals shortening of chronological lifespan in cof1-5 cells which depends on POR1. Colony forming unit formation based on clonogenic survival is depicted in G and H, whereas PI positivity is shown in I and J. Statistical significance in E, F, H, and J was assessed using Brown-Forsythe and Welch-ANOVA test. Asterisks indicate significance based on p-levels of the comparisons to the respective control strains. See also Figure S2. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 5

Slt2 localisation to the mitochondrial compartment is enhanced in cof1-5 cells (A) Slt2, which is mostly found in the nucleus in wild-type cells at stationary phase, forms punctate foci in cof1-5, as documented by fluorescence microscopy using chromosomally tagged SLT2-GFP under control of its endogenous promoter. Deconvolved pictures with Hoechst staining for nuclei are shown. (B and C) Cells showing nuclear localisation of Slt2-GFP (B) and foci-forming cells (C) were quantified, plotted and analyzed for significant localisation change as compared to wild type. (D) Representative fluorescence microscopy pictures of Slt2-GFP expressing cells with rhodamine B hexylester staining for colocalization analysis with mitochondria. (E) Colocalisation of Slt2-GFP (green) with the mitochondrial stain rhodamine B hexylester was increased in cof1-5 as shown by significant increase of the Pearson colocalisation coefficient. (F–I) Slt2-GFP expression from a plasmid was used to monitor cellular Slt2 localisation in cof1-5 cells in dependence of POR1. Representative microscopy pictures are shown in (F), Slt-GFP foci-containing cells are quantified in (G), cells with nuclear Slt2 are quantified in (H) and the Pearson coefficient of Slt2-GFP colocalization with mitochondria is visualised in (I). (J and K) Expression of Pkc1-GFP under control of its endogenous promoter reveals increased PKC1-GFP foci formation in cof1-5 (J, K). Representative fluorescence microscopy pictures of individual Pkc1-GFP expressing cells with additional rhodamine-B-hexylester staining for mitochondria and autodot staining for LDs are shown in (J) and cells with Pkc1-GFP foci are quantified in (K). (L) The Pearson coefficient was determined as a measure of colocalisation of Pkc1-GFP with mitochondria (rhodamine-B-hexylester). Statistical significance in (B) and (C) was assessed using Welch's t test, (E, K, L) were analyzed using unpaired t test and (G, H, I) by ordinary one-way ANOVA. See also Figure S3. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 6

Reduced actin dynamics lead to a porin-dependent increase in lipid droplet number that are required for CWI activation (A and B) Increase of LD number in cof1-5 depends on Lro1 and Dga1. Mean LD-numbers per cell are plotted in (A) and representative microscopy pictures are shown in (B). Each dot in (A) represents the mean LD number per cell of a single experiment (n = 6) with at least 119 cells being evaluated per experiment. (C and D) Gene deletions of POR1 or SLT2 are sufficient to prevent LD accumulation in cof1-5. Mean LD numbers per cell were assessed by quantifying fluorescence microscopy pictures using Bodipy staining (C). Representative fluorescence microscopy images are shown in (D). (E) Representative EM micrographs supporting the observation of LD-number-increase in cof1-5. V, vacuole; LD, lipid droplet; N, nucleus. Statistical significance in (A) and (C) was assessed using two-way ANOVA with cof1-5 mutation as first factor and additional KO as second factor. See also Figure S4. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 7

The acyltransferases Lro1 and Dga1 are required for MAPK-related changes in cof1-5 (A) Polarisation of the actin cytoskeleton was assessed using phalloidin (red) and DAPI (blue) staining. The cof1-5 mutant shows loss of actin polarisation, which is not rescued by gene KO of LRO1, DGA1 or a combined double deletion of the ladder. (B and C) The growth defect (B) and flocculation phenotype (C) as observed in cof1-5 are restored by additional deletion of the acyltransferases LRO1, DGA1, or in the double deletion mutant (lro1Δ dga1Δ). (D) Slt2 phosphorylation in cof1-5 depends on Lro1 and Dga1 as the KO of either corresponding gene and the double KO prevents Slt2 phosphorylation. (E) Mean LD abundance per cell was quantified in diverse conditions of stress. 150 cells per condition and experiment were quantified with a total of three independent experiments (n = 3). Statistical significance in (E) was assessed using ordinary one-way ANOVA. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

Figure 8

Lipidomic analysis reveals characteristic Por1-dependent changes in the lipid profile of cof1-5 (A and B) Mass spectrometry-assisted lipidomic quantification of highly abundant (A) and less abundant yeast lipids from total cell extracts (B). (C and D) Sterolesters and free ergosterol were quantified separately as shown in (C) and sphingolipids are depicted in (D). (E–H) Lipidomic changes in the neutral lipid classes TG (E), SE (F), Erg (G) and the index_SE/TG (H) were further verified by additional HPTLC analysis. Statistical significance in (A-G) was assessed using two-way ANOVA with cof1-5 mutation as first factor and additional KO as second factor, except for CL, LPC and LPE in (B), which were analyzed using Kruskal-Wallis test; (H) was analyzed by Brown-Forsythe and Welch ANOVA test. See also Figure S5. Error bars indicate standard error of the mean (SEM) and asterisks indicate significant differences based on p-values, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.