Specific degradation of phosphatidylglycerol is necessary for proper mitochondrial morphology and function

Abstract

In yeast, phosphatidylglycerol (PG) is a minor phospholipid under standard conditions; it can be utilized for cardiolipin (CL) biosynthesis by CL synthase, Crd1p, or alternatively degraded by the phospholipase Pgc1p. The Saccharomyces cerevisiae deletion mutants crd1Δ and pgc1Δ both accumulate PG. Based on analyses of the phospholipid content of pgc1Δ and crd1Δ yeast, we revealed that in yeast mitochondria, two separate pools of PG are present, which differ in their fatty acid composition and accessibility for Pgc1p-catalyzed degradation. In contrast to CL-deficient crd1Δ yeast, the pgc1Δ mutant contains normal levels of CL. This makes the pgc1Δ strain a suitable model to study the effect of accumulation of PG per se. Using fluorescence microscopy, we show that accumulation of PG with normal levels of CL resulted in increased fragmentation of mitochondria, while in the absence of CL, accumulation of PG led to the formation of large mitochondrial sheets. We also show that pgc1Δ mitochondria exhibited increased respiration rates due to increased activity of cytochrome c oxidase. Taken together, our results indicate that not only a lack of anionic phospholipids, but also excess PG, or unbalanced ratios of anionic phospholipids in mitochondrial membranes, have harmful consequences on mitochondrial morphology and function.

Keywords: Mitochondria; Morphology; Phosphatidylglycerol; Respiration; Yeast.

Figure 1. Pgc1p buffers PG levels in yeast

S. cerevisiae cells were grown in SMDGE I− or I+ medium in the presence of [³²P] orthophosphoric acid for five to six generations. Steady-state phospholipid analysis showing: A, PG as a percentage of [³²P] orthophosphoric acid labeled cellular phospholipids and B, CL as a percentage of [³²P] orthophosphoric acid labeled cellular phospholipids. Data represent mean values of three experiments ±SEM. Statistically significant differences between mutant strains and WT and between I− and I+ media are marked.^*, p < 0.05; ^**, p < 0.01; ^***, p < 0.001.

Figure 2. Deletion of PGC1 and CRD1 induces accumulation of PG, which differ in fatty acid composition and accessibility to degradation by Pgc1p

A, Steady-state analysis of total cellular phospholipids. Yeast cells were grown in SMDGE I− URA− medium with [³²P] orthophosphoric acid for five to six generations. PG is shown as a percentage of [³²P] orthophosphoric acid incorporated into cellular phospholipids. Data represent mean values of three independent experiments ±SEM. Statistically significant differences between mutant strains and WT and between I− and I+ media are marked. *p < 0.05; ^**, p < 0.01; ^***, p < 0.001. B, Gas chromatography analysis of PG fatty acids. Yeast cells were grown in SMD I− URA− media. Relative amounts of individual fatty acids derived from PG extracted from crude mitochondrial fractions are shown. Data represent mean values from two independent experiments ±SEM. C, In vitro analysis of Pgc1p activity. Conversion of [¹⁴C] PG originated from either pgc1Δ or crd1Δ cells to DAG is shown. [¹⁴C] PG was isolated from either pgc1Δ or crd1Δ cells that were grown in SMD I− in the presence of [¹⁴C] acetate for six generations. Description of the in vitro Pgc1p degradation assay could be found in [7]. Data represent mean values of three independent experiments ±SEM. Statistically significant differences between pgc1Δ with EV and PGC1 are marked. ^**, p < 0.01. EV – multicopy empty vector, PGC1 – multicopy vector containing PGC1 controlled by its own promotor.

Figure 3. Expression of PGC1 and Pgc1p enzymatic activity is independent of inositol

A, Analysis of PGC1 transcription. PGC1 mRNA abundance was analyzed by qRT-PCR following growth of WT cells in SMD I+ or I− media to indicated growth phase. The results were normalized to ACT1 and IPP1 mRNA, the mRNA levels in WT I+ cells in log phase of growth were set to 1. Data represent mean values of four independent experiments ±SEM. Statistically significant differences between different phases of growth are marked. ^*, p < 0.05; ^**, p < 0.01; ^***, p < 0.001. B, In vitro enzymatic activity of Pgs1p. Yeast cells were grown in SMD I+ or I− media. PGP synthase activity was determined in mitochondrial fractions. A unit of enzymatic activity is defined as the amount of enzyme that catalyzes the formation of 1 nmol of product/min under the assay conditions described in [7]. Data represent mean values from three independent experiments ±SEM. Statistically significant differences between I− and I+ media are marked. ^**, p < 0.01; ^***, p < 0.001.

Figure 4. CL-deficient mutants exhibit flat Mitotracker-stained sheets

Abnormal mitochondria were visualized by Mitotracker Red CMX-Ros staining in cells grown in SMDGE medium for 24 hours. A, Three examples of pgc1Δcrd1Δ cells containing Mitotracker-stained large flat sheets are presented as mean projections of five consecutive confocal sections from the cell cortex with axial spacing of 370 nm. Flat sheets detected in other analyzed strains including WT were morphologically identical at the resolution of fluorescence microscopy. B, isosurface projections of full 3D stacks, encompassing the whole cell. Bar: 2μm. C, Statistical evaluation of sheet occurrence in the analyzed strains. Fraction of cells showing this morphological feature was determined in WT, pgc1Δ, crd1Δ and pgc1Δcrd1Δ strain cultures grown in SMDGE I+ or I-media. Data represent mean values of three independent experiments ±SEM. At least 300 cells were evaluated in each experiment. Statistically significant differences between mutant strains and WT and between I− and I+ media are marked. ^***, p < 0.001.

Figure 5. Flat Mitotracker-stained sheets observed in CL-deficient cells correspond to morphologically abnormal mitochondria

A–C, pgc1Δcrd1Δ cells expressing the mitochondrial marker PDA1-GFP or D–E, ER marker ss-GFP-HDEL were grown in SMDGE I− medium for 24 hours, stained with Mitotracker Red CMX-Ros and observed. Mean projections of five consecutive confocal sections from the cell cortex with axial spacing of 370 nm (A–D), and a single transversal confocal section (E) are presented. Bar: 5μm.

Figure 6. Yeast cells accumulating PG exhibit fragmented mitochondria

Yeast cell were grown in SMDGE media for 24 hours. A, Mitochondrial networks visualized by Mitotracker staining. Mean projections of five consecutive confocal sections with axial spacing of 370 nm encompassing the cell cortex are shown in cultures grown in presence or B, absence of inositol in the growth medium. Bar: 5μm. C, Statistical analysis of the mitochondrial length distributions in the analyzed strains. Median (central line), first and third quartile (lower and upper box borders, respectively) and 10 and 90 percentiles (error bars) are depicted in a box plot. Data were collected in three independent experiments. At least 250 cells were analyzed in each experiment.

Figure 7. Mitochondrial respiration is affected by PGC1 deletion

Yeast cells were cultivated in SMDGE I+ or I− media for 24 hours. Isolated mitochondria were used to measure of oxygen consumption. NADH was used as a respiratory substrate. Mitochondrial respiration in the presence of A, NADH (LEAK respiration); B, NADH and ADP (OXPHOS capacity) and C, protonophore CCCP (Electron Transfer System capacity). D, Respiratory Control Index as a measure of OXPHOS coupling. Data represent mean values of 3–6 independent experiments ±SEM. Statistically significant differences between mutant strains and WT and between I− and I+ media are marked. ^*, p < 0.05; ^**, p < 0.01; ^***, p < 0.001.

Figure 8. Activity of Complex IV is increased in yeast strains with PGC1 gene deletion

A, Immunoblot of mitochondrial lysates separated by blue native-PAGE using anti-Rip1p (subunit of Complex III) and anti-Cox2p (Complex IV) antibodies. Bottom panels represent immunoblots following SDS-PAGE and serve as loading controls. B, In-gel staining for Complex V activity. C, in vitro activities of Complex III, Complex IV and Complex V. Data in C represent mean values from 4–5 independent experiments ±SEM. Statistically significant differences between mutant strains and WT and between I− and I+ media are marked. ^*, p < 0.05; ^**, p < 0.01; ^***, p < 0.001.

Figure 9. The amount of Cox4p is not affected by PGC1 deletion

Immunoblot analysis of Cox4p. Upper panel with Porin serves as a loading control. For loading, 5 (right) and 10 μg of mitochondrial protein (left) was used.