Introduction

Cardiolipin (CL) is a unique dimeric glycerolphospholipid localized almost exclusively to mitochondria in mammalian cells and the yeast Saccharomyces cerevisiae. CL plays critical roles in mitochondrial functions, such as oxidative phosphorylation, and regulation of apoptosis and mitophagy1,2,3, and involved in the biogenesis of mitochondria through the regulation of protein import4,5,6,7 into and fusion8 of mitochondria. This phospholipid exhibits a propensity for the formation of non-bilayer, inverted hexagonal (HII) phase structures9, and thus seems to engage in the formation of local non-bilayer structures within mitochondrial membranes, which may be involved in membrane dynamics, including membrane assembly, transmembrane movement of proteins, and dynamic formation of membrane contact sites.

Mitochondria of the yeast S. cerevisiae contain a set of enzymes required for CL biosynthesis from phosphatidic acid (PA), a common intermediate for phospholipid biosynthesis. The CL biosynthetic pathway in the yeast was shown in Fig. 1. Tam41 catalyzes the first step of the pathway, namely, the synthesis of CDP-diacylglycerol (CDP-DG) from PA and CTP10. The second step is the formation of phosphatidylglycerol phosphate (PGP) from CDP-DG and glycerol 3-phosphate, which is catalyzed by PGP synthase (Pgs1)11,12. The third step is dephosphorylation of PGP by Gep4, resulting in the formation of phosphatidylglycerol (PG)13. In the final step, a cardiolipin synthase, Crd1, catalyzes the CL formation from PG and CDP-DG14,15,16. These enzymes of the CL biosynthetic pathway are associated with the matrix face of or inserted into the mitochondrial inner membrane (MIM) (Fig. 1). Mitochondria are incapable of synthesizing PA de novo, and therefore PA produced in the extramitochondrial space such as the endoplasmic reticulum and lipid droplets should be transported to the MIM for CL biosynthesis. So far, Ups1-Mdm35, a protein complex localized to the mitochondrial intermembrane space (IMS), has been shown to mediate PA transfer from the mitochondrial outer membrane (MOM) to the MIM17,18,19,20.

Figure 1
figure 1

Mitochondrial synthesis of cardiolipin and phosphatidylethanolamine in yeast. CDP-DG, CDP-diacylglycerol; CL, cardiolipin; G3P, glycerol 3-phosphate; MIM, mitochondrial inner membrane; MOM, mitochondrial outer membrane; PA, phosphatidic acid; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; PGP, phosphatidylglycerol phosphate; PS, phosphatidylserine.

Full size image

The Ups1-Mdm35-mediated PA transfer is an important process for CL biosynthesis, because deletion of UPS1 (ups1∆) leads to a ~80% decrease in the cellular CL level21,22. However, CL synthesis occurs even in yeast cells lacking Ups1 and the CL synthesis in ups1∆ cells is strikingly enhanced by depletion of Ups221, which forms a protein complex with Mdm3523,24 and mediates phosphatidylserine (PS) transfer from the MOM to the MIM for phosphatidylethanolamine (PE) synthesis by a PS decarboxylase, Psd125,26. Therefore, a Ups1-independent PA transfer pathway to the MIM for CL biosynthesis exists in the yeast; elucidation of that pathway is important for clarification of the whole picture of CL metabolism. In this study, we show that the CL accumulation through the Ups1-independent CL biosynthetic pathway is enhanced by deletion of not only UPS2, but also PSD1 and CHO1 encoding a PS synthase, both of which are required for mitochondrial PE synthesis, suggesting that low PE levels are relevant to the enhancement of CL accumulation through the Ups1-independent pathway, and that the Ups1-independent and low-level PE-enhanced CL accumulation depends on the functions of FMP30, MDM31, and MDM32, which have been shown to be involved in CL metabolism22,27,28.

Results

The PE level is relevant to the accumulation of CL in and growth of ups1∆ cells

The defect in CL accumulation in ups1∆ cells has been shown to be suppressed by deletion of UPS2 21, which leads to a decrease in the PE level in mitochondria21,22,25,28. Therefore, we examined the effect of deletion of PSD1 and CHO1, both of which are required for the mitochondrial PE synthesis, on the accumulation of CL in ups1∆ cells. Figure 2a shows thin layer chromatography (TLC) analyses of the total cellular [32P]phospholipids of various mutant cells metabolically labeled for 24 hours with [32P]Pi. The CL level in ups1∆ cells was about 20% of that in wild-type cells, and increased to about 75% of that in wild-type cells after introduction of the ups2∆ mutation (Fig. 2a,b), this being consistent with previous studies21. To our surprise, deletion of PSD1 and CHO1 as well as UPS2 in ups1∆ cells resulted in an increase in the CL level, as shown in Fig. 2a,b. The CL levels in ups1∆psd1∆ and ups1∆cho1∆ double mutant cells, respectively, were 52 and 58% of that in wild-type cells and slightly lower than that in psd1∆ and cho1∆ single mutant cells (Fig. 2a,b). The PE levels in ups1∆ups2∆, ups1∆psd1∆, and ups1∆cho1∆ double mutant cells, respectively, were about 70, 60, and 25% of that in wild-type cells, and similar to those in ups2∆, psd1∆, and cho1∆ single mutant cells (Fig. 2a,b). Furthermore, concomitant overexpression of PSD1 and CHO1 (2OE or PSD1↑CHO1↑) under the control of a strong glyceraldehyde-3-phosphate dehydrogenase (GPD) promoter in ups1∆ups2∆ cells increased the PE level and abolished the enhancement of CL accumulation in ups1∆ cells caused by the ups2∆ mutation (Fig. 2a,b).

Figure 2
figure 2

The PE level is relevant to the accumulation of CL in and growth of ups1∆ cells. (a and b) Yeast cells, as indicated, were grown at 30 °C to saturation in YPAD medium. The cells were then diluted to an OD600 of 0.05 in 500 µl of YPAD containing [32P]Pi (10 µCi/ml) and cultured at 30 °C for 24 hours. Total cellular phospholipids were extracted, separated by TLC and then analyzed with an imaging analyzer. (a) Typical chromatogram on phospholipid analysis by TLC. PI, phosphatidylinositol; PC, phosphatidylcholine. (b) The percentages of CL and PE relative to total major phospholipids (CL, PA, PE, PS, PI, and PC). Values are means ± SD (n = 3~15). (c) Growth of various mutant cells. Yeast cells, as indicated, were spotted onto YPAD plates in ten-fold serial dilutions starting with a density of 0.7 OD600 units/ml, and then incubated at 30 °C for 24 hours. *1, p = 0.015; *2, p = 0.040; *3, p = 0.010; *4, p = 0.022; *5, p = 0.0022; *6, p = 0.0027; *7, p = 0.0095. N.S., not significant.

Full size image

We also examined the growth rates of the various mutants used for the above experiments by means of a spot test, and found correlation of the cellular CL levels and growth rates, as shown in Fig. 1c. These results suggest that a decrease in PE to under the threshold level enhances CL accumulation in ups1∆ cells and suppresses the growth defect of ups1∆ cells.

FMP30 is required for the UPS1-independent and low-level PE-enhanced CL accumulation

FMP30 encodes a mitochondrial inner membrane protein having a large domain exposed to the intermembrane space, which exhibits strong homology with mammalian N-acylPE (NAPE)-specific phospholipase Ds (NAPE-PLDs)27,29. We have shown that FMP30 is required for the maintenance of a normal CL level in psd1∆ cells, and that deletion of FMP30 is synthetically lethal with the ups1∆ mutation27. Therefore, we decided to examine whether FMP30 was involved in the UPS1-independent and low-level PE-enhanced CL accumulation. For this purpose we constructed yeast strains carrying the FMP30 gene under the control of a tetracycline-regulatable TEToff promoter30 (tet-FMP30) with or without various mutations including ups1∆. Wild-type and tet-FMP30 cells were able to grow well equally in the presence or absence of a tetracycline analogue, doxycycline (Dox) (Fig. 3a). However, tet-FMP30 ups1∆ cells showed a partial growth defect in the absence of Dox and a further strong growth defect in the presence of Dox, as expected (Fig. 3a), indicating that Dox depressed tet-FMP30 gene expression.

Figure 3
figure 3

FMP30 is required for the UPS1-independent and low-level PE-enhanced accumulation of CL. (a) Growth of various mutant cells. Yeast cells, as indicated, were grown at 30 °C to saturation in YPAD medium with or without 10 μg/ml Dox, as indicated. The yeast cells were then spotted onto YPAD plates with or without 10 μg/ml Dox, as indicated, in ten-fold serial dilutions starting with a density of 0.7 OD600 units/ml, and then incubated at 30 °C for 24 hours. (b and c) Yeast cells, as indicated, were grown at 30 °C to saturation in YPAD medium with or without 10 μg/ml Dox, as indicated. The cells were then diluted to an OD600 of 0.05 in 500 µl of YPAD containing [32P]Pi (10 µCi/ml), supplemented with or without 10 μg/ml Dox and cultured at 30 °C for 24 hours. Total cellular phospholipids were extracted, separated by TLC, and then analyzed with an imaging analyzer. (b) Typical chromatogram on phospholipid analysis by TLC. (c) The percentages of CL and PE relative to total major phospholipids (CL, PA, PE, PS, PI, and PC). Values are means ± SD (n = 3~15). *1, p = 0.035; *2, p = 0.0044; *3, p = 0.0050.

Full size image

Figure 3b shows TLC analysis of the total cellular [32P]phospholipids of wild-type and various tet-FMP30-carrying mutant cells metabolically labeled with [32P]Pi for 24 hours in the presence or absence of Dox. The CL and PE levels in tet-FMP30 cells were not significantly affected by the addition of Dox to the growth medium, as was the case with those in wild-type cells, and similar to those in wild-type cells cultivated with or without Dox (Fig. 3b,c). The CL level in tet-FMP30 ups1∆ cells was about 20% of that in tet-FMP30 cells in the medium without Dox, which is similar to that in ups1∆ cells with the native FMP30 gene (Figs 2b and 3c). Although the addition of Dox to the medium significantly affected the growth of tet-FMP30 ups1∆ cells (Fig. 3a), the Dox addition did not lead to a significant further decrease in the CL level in tet-FMP30 ups1∆ cells (Fig. 3b,c), implying that Fmp30 was not involved in the accumulation of the residual amount of CL in ups1∆ cells. In contrast, the UPS1-dependent and low-level PE-enhanced CL accumulation, namely, the CL accumulation in ups1∆ups2∆, ups1∆psd1∆, and ups1∆cho1∆ double mutant cells, were strikingly reduced by repression of tet-FMP30 expression by Dox (Fig. 3b,c). In addition, the growth of ups1∆ups2∆, ups1∆psd1∆, and ups1∆cho1∆ cells carrying tet-FMP30 was significantly impaired by the addition of Dox to the medium, as shown in Fig. 3a. These results suggest that FMP30 is required for the UPS1-independent and low-level PE-enhanced accumulation of CL.

As control experiments, we subjected tet-FMP30 cells having the ups2∆, psd1∆, or cho1∆ mutation, but not the ups1∆ mutation to phospholipid analyses, obtaining interesting results. Although the CL level in tet-FMP30 cells was not significantly affected by the addition of Dox to the medium, those in ups2∆, psd1∆, and cho1∆ cells carrying tet-FMP30 were decreased by about 40, 60, and 40%, respectively, by the Dox addition (Fig. 4b). These results implied that the ups2∆, psd1∆, and cho1∆ mutations changed the FMP30-dependent pathway for CL accumulation from a minor one to a major one.

Figure 4
figure 4

ups2∆, psd1∆, and cho1∆ mutations change the FMP30-dependent pathway from a minor to a major one for accumulation of CL. (a and b) See the legend to Fig. 3b,c. (a) Typical chromatogram on phospholipid analysis by TLC. (b) The percentages of CL and PE relative to total major phospholipids (CL, PA, PE, PS, PI, and PC). Values are means ± SD (n = 3~15). *1, p = 0.0022; *2, p = 0.0019; *3, p = 0.013.

Full size image

MDM31 and MDM32 are required for the UPS1-independent and low-level PE-enhanced CL accumulation

We have identified FMP30 as a gene whose deletion in psd1∆ cells causes a severe growth defect27. During the course of such a study of genetic interactions, we found also that deletion of MDM31 is synthetically lethal with the psd1∆ mutation27. Mdm31 and its homolog Mdm32 are mitochondrial inner membrane proteins with two membrane-spanning regions at the C-terminus and near the N-terminus, respectively, and a middle region exposed to the intermembrane space, and have been shown to be involved in CL metabolism22,28. As shown in Fig. 5, deletion of MDM32 as well as MDM31 was shown to be synthetically lethal with the psd1∆ mutation. Therefore, because of their similarities to FMP30 in submitochondrial localization and genetic interaction with PSD1, MDM31 and MDM32 might be involved in the UPS1-independent and low-level PE-enhanced accumulation of CL. In fact, overexpression of MDM31 partially suppresses the defect in growth of and CL accumulation in ups1∆ cells28. To address the functions of MDM31 and MDM32 in ups1∆ cells, we constructed yeast strains carrying the MDM31 or MDM32 gene under the control of a tetracycline-regulatable TEToff promoter (tet-MDM31 or tet-MDM32) with or without the ups1∆ and/or ups2∆ mutation(s). Wild-type, tet-MDM31, and tet-MDM32 cells were able to grow well equally in the presence or absence of Dox (Fig. 6a). However, the growth of tet-MDM31 and tet-MDM32 cells carrying the ups1∆ or ups2∆ mutation, and both the ups1∆ and ups2∆ mutations was impaired by the addition of Dox to the medium (Fig. 6a).

Figure 5
figure 5

Deletion of MDM31 and MDM32 are synthetically lethal with deletion of PSD1. A yeast strain, psd1∆ α (MATα psd1∆::kanMX4 can1∆::STE2pr-Sp_his5), was mated with another strain, mdm31∆ (MATa mdm31∆::hphNT1 CAN1 WT) or mdm32∆ (MATa mdm32∆::natNT2 CAN1 WT). The resulting heterozygous diploid strains ((PSD1/psd1∆::kanMX4 MDM31/mdm31∆::hphNT1 CAN1/can1∆::STE2pr-Sp_his5) and (PSD1/psd1∆::kanMX4 MDM31/mdm31∆::hphNT1 CAN1/can1∆::STE2pr-Sp_his5)) were allowed to sporulate, plated on the various selection media indicated, and then incubated at 30 °C for 3 days. All the media contained Canavanine but not histidine for the selection of MATa haploid strains.

Full size image
Figure 6
figure 6

MDM31 and MDM32 are required for the UPS1-independent and low-level PE-enhanced accumulation of CL. (a) Growth of various mutant cells. See the legend to Fig. 3a. (b and c) See the legend to Fig. 3b,c. (b) Typical chromatogram on phospholipid analysis by TLC. (c) The percentages of CL and PE relative to total major phospholipids (CL, PA, PE, PS, PI, and PC). Values are means ± SD (n = 3~15). *1, p = 0.0092; *2, p = 0.0012; *3, p = 0.0015; *4, p = 0.035; *5, p = 0.0027; *6, p = 0.0057.

Full size image

Next, we examined the CL levels in these mutant cells cultivated in the presence or absence of Dox by metabolic labeling with [32P]Pi for 24 hours, followed by TLC analysis. The CL level in tet-MDM31 cells was about 95% of that in wild-type cells in the medium without Dox, and decreased by 7% on the addition of Dox to the medium (Fig. 6b,c). The CL level in tet-MDM31 ups1∆ cells cultivated without Dox was 48% of that in wild-type cells and about 2.5-fold that in ups1∆ cells (Figs 2a,b and 6b,c). This elevation of the CL level in ups1∆ cells on replacement of the native MDM31 promoter with the TEToff promoter might have resulted from an increase in the expression level of the gene, because overexpression of MDM31 had been shown to partially suppress the defect in CL accumulation in ups1∆ cells28. The addition of Dox to the medium reduced the CL level in tet-MDM31 ups1∆ cells to a level similar to that in ups1∆ cells (Figs 2a,b and 6b,c), indicating that MDM31 is involved in the UPS1-independent CL accumulation. However, because there was no significant difference between the CL levels in ups1∆ cells and tet-MDM31 ups1∆ cells cultivated with Dox (Figs 2a,b and 6b,c), MDM31 might not be involved in the accumulation of the residual amount of CL in ups1∆ cells. In the medium without Dox, the CL level in tet-MDM31 ups1∆ ups2∆ cells was about 80% of that in wild-type cells and about 1.7-fold that in tet-MDM31 ups1∆ cells (Fig. 6b,c), indicating that the UPS1-independent and low-level PE-enhanced CL accumulation had occurred. In contrast to the CL level in tet-MDM31 cells, that in tet-MDM31 ups1∆ ups2∆ cells was strikingly reduced by depression of tet-MDM31 expression by Dox, as shown in Fig. 6b,c, suggesting that MDM31 was required for the UPS1-independent and low-level PE-enhanced CL accumulation.

As shown in Fig. 6b,c, the CL levels in tet-MDM32, tet-MDM32 ups1∆, and tet-MDM32 ups1∆ups2∆ cells cultivated with or without Dox, respectively, were similar to those in tet-MDM31, tet-MDM31 ups1∆, and tet-MDM31 ups1∆ups2∆ cells cultivated with or without Dox. It is therefore likely that MDM32 is also required for the UPS1-independent and low-level PE-enhanced CL accumulation.

Although the accumulation of CL in and growth of tet-MDM31 and tet-MDM32 cells were not largely affected by the addition of Dox to the medium, those of ups2∆ cells carrying tet-MDM31 or tet-MDM32 were remarkably impaired by the Dox addition (Fig. 6b,c). These results implied that the ups2∆ mutation changed the main CL biosynthetic pathway from an MDM31- and MDM32-independent pathway to an MDM31- and MDM32-dependent pathway.

Physical interaction of Fmp30 with Mdm31 and Mdm32

We showed that depletion of any one of the factors Fmp30, Mdm31, and Mdm32 almost completely prevented CL synthesis under the mitochondrial-PE-reduced and Ups1-defective conditions (Figs 3 and 6), suggesting that these three factors act in the same pathway. To obtain further evidence for this, we performed immunoprecipitation experiments. Mitochondria were isolated from the yeast cells expressing genomically-3xHA-tagged FMP30 and carrying the plasmid pRS426, pRS426-FLAG-MDM31, or pRS426-FLAG-MDM32, and then subjected to immunoprecipitation with anti-FLAG agarose beads. As shown in Fig. 7a and S1, Fmp30-3xHA was co-immunoprecipitated with both FLAG-Mdm31 and FLAG-Mdm32, although Fmp30-3xHA was detected more strongly in the immunoprecipitate fraction of FLAG-Mdm32 than that of FLAG-Mdm31. In contrast, another inner membrane protein, Tim23, was detected in neither the immunoprecipitate fraction of FLAG-Mdm31 nor that of FLAG-Mdm32. To confirm the interactions of Fmp30 with Mdm31 and Mdm32, we performed immunoprecipitation experiments using the yeast cells overproducing Fmp30-HA under the control of a strong glyceraldehyde-3-phosphate dehydrogenase (GPD) promoter and carrying the plasmid encoding FLAG-Mdm31 or FLAG-Mdm32. As shown in Fig. 7b and S1, bands of Fmp30-3xHA were clearly detected in the immunoprecipitate fraction of FLAG-Mdm31 and that of FLAG-Mdm32. Thus, the immunoprecipitation experiments showed that Fmp30 physically interacted with both Mdm31 and Mdm32. These results, taken together, indicate that Fmp30, Mdm31 and Mdm32 cooperatively act in the same pathway, which is essential for CL synthesis under the mitochondrial-PE-reduced and Ups1-defective conditions.

Figure 7
figure 7

Fmp30 physically interacts with Mdm31 and Mdm32. (a) Mitochondria (2 mg) from the yeast cells expressing genomically-3xHA-tagged FMP30, and carrying the plasmid pRS426 (Empty), pRS426-FLAG-MDM31 (FLAG-MDM31), or pRS426-FLAG-MDM32 (FLAG-MDM32) were solubilized with 1% digitonin and then subjected to immunoprecipitation with anti-FLAG agarose beads. The immunoprecipitates were eluted from the beads with 2% SDS and then analyzed by Western blotting with anti-HA, anti-FLAG, and anti-Tim23 antibodies. The asterisk indicates the IgG from the anti-FLAG agarose, which migrates similarly to FLAG-Mdm31. (b) Mitochondria (2 mg) from wild-type or the yeast cells expressing genomically-3xHA-tagged FMP30 under GPD promotor and carrying the plasmid pRS424 (Empty), pRS424-FLAG-Mdm31 (FLAG-MDM31), or pRS424-FLAG-Mdm32 (FLAG-MDM32) were analyzed as described in (a). *indidates the IgG from the anti-FLAG agarose, which migrates similarly to FLAG-Mdm31. **indicates the potential unprocessed/precursor forms of Fmp30-3xHA, which may accumulate due to the overproduction of Fmp30. Note that the fastest-migrating form of Fmp30-3xHA, which appeared to be mature form, was co-precipitated with FLAG-Mdm31 and FLAG-Mdm32. Multiple exposures of full-length blots are presented in Supplementary Figure S1.

Full size image

Discussion

CL and PE levels in and growth rates on YPAD medium of various mutant cells used in this study are summarized in Supplementary Table S3. In the present study, the enhancement of CL accumulation by ups2∆ in ups1∆ cells was suggested to be relevant to the decrease in the PE level by the finding that deletion of not only UPS2 but also PSD1 and CHO1, both of which are required for mitochondrial PE synthesis, suppressed the defect in CL accumulation in ups1∆ cells (Fig. 2a,b), and that elevation of the PE level on concomitant overexpression of PSD1 and CHO1 (PSD1↑CHO1↑) in ups1∆ ups2∆ cells abolished the increase in the CL level in ups1∆ cells caused by the ups2∆ mutation (Fig. 2a,b). How does a decrease in the PE level enhance the accumulation of CL in ups1∆ cells? A previous study indicated that decreased PE levels have beneficial effects for maintenance of mitochondrial functions and morphology in yeast cells with a defect in the mitochondrial contact site and cristae organizing system (MICOS)26. Although deletion of MIC10 or MIC60, both of which encode a core subunit of MICOS, leads to a partial defect in respiration growth and enlarged mitochondria with abnormally stacked multilamellar cristae observed on electron microscopy, deletion of UPS2 in mic10∆ and mic60∆ cells improves respiration growth and restores the mitochondrial architecture with a normal cristae morphology26. In addition, deletion of PSD1 in MICOS-deficient cells significantly improves respiratory growth26, though loss of Psd1 in wild-type cells reduces the PE level and impairs respiratory growth. Furthermore, deletion of UPS2 improves cell growth and CL levels in yeast cells lacking the ER-mitochondria encounter structure (ERMES) complex28, which connects mitochondria to the ER and is suggested to facilitate phospholipid exchange between these two organelles31,32. On the other hand, it has been shown that the cristae morphology is disturbed upon inhibition of CL synthesis by the tam41∆ or pgs1∆ mutation but restored when PA transfer to the MIM by Ups1-Mdm35 is impaired17. These observations indicate that moderate alterations in lipid composition can have dramatic effects on mitochondrial functions and morphology. Thus, a decrease in the PE level in ups1∆ cells might have a large effect on mitochondrial morphology, such as the number and extent of contact sites, and therefore enhance UPS1-independent PA supply to the MIM via contact sites for CL synthesis. Alternatively, a decrease in the PE level might nullify the inhibitory effect(s) of PE on Fmp30, Mdm31, and/or Mdm32, which are required for the UPS1-independent and low-level PE-enhanced CL accumulation (See below).

Although deletion of UPS2, PSD1, or CHO1 restored the CL level in ups1∆ cells, depression of tet-FMP30 expression by Dox in ups1∆ups2∆, ups1∆psd1∆, and ups1∆cho1∆ cells decreased the CL level to about 10% of that in wild-type cells (Fig. 3b,c). Similarly, depression of the tet-MDM31 and tet-MDM32 expression by Dox in ups1∆ups2∆ cells decreased the CL level to about 10% or less of that in wild-type cells (Fig. 6b,c). These results suggest that all three genes, FMP30, MDM31, and MDM32, are essential for the UPS1-independent and low-level PE-enhanced CL accumulation, and therefore this CL accumulation is carried out through a single pathway, in which the three genes function cooperatively. Physical interactions of Fmp30 with Mdm31 and Mdm32 (Fig. 7), and positive genetic interactions of FMP30 with MDM31 and MDM32 33 support their cooperative functions in the same pathway. We hereafter call the UPS1-independent low-level PE-enhanced pathway the “FMM (Fmp30-, Mdm31, and Mdm32)-dependent” pathway.

In the wild-type cells, the UPS1-dependent pathway for CL synthesis appeared to be the main pathway for the CL accumulation, because deletion of UPS1 led to an about 80% decrease in the CL level (Fig. 2a,b). However, depression of the tet-FMP30 expression by Dox in ups2∆, psd1∆, and cho1∆ cells with the wild-type allele of UPS1, respectively, led to about 40, 60, and 40% decreases in the CL level (Fig. 4a,b). Furthermore, repression of the tet-MDM31 and tet-MDM32 expression by Dox in ups2∆ cells with the wild-type allele of UPS1, respectively, led to about 75 and 80% decreases in the CL level. These results suggest that in yeast cells with decreased mitochondrial PE levels, irrespective whether they carry a null (ups1∆) or the wild-type allele of UPS1, the main pathway for the CL accumulation appeared to change from the UPS1-dependent one to the FMM-dependent one. The differences of the decreases in the CL level among tet-FMP30 ups2∆, tet-MDM31 ups2∆, and tet-MDM32 ups2∆ cells cultivated in the presence of Dox might be caused by the inability of complete shut-off of gene expression by Dox, and the difference in the minimal effective doses of Fmp30, Mdm31, and Mdm32. On the other hand, the FMM-dependent pathway seemed not to be responsible for the accumulation of the residual amount of CL in ups1∆ cells with a normal mitochondrial PE level, because the CL levels in tet-FMP30 ups1∆, tet-MDM31 ups1∆, and tet-MDM32 ups1∆ cells cultivated with Dox were similar to that in ups1∆ cells (Figs 2a,b, 3b,c and 6b,c).

FMP30 encodes a mitochondrial inner membrane protein with a large domain exposed to the intermembrane space, which exhibits a strong homology with mammalian NAPE-PLDs27,29. Mammalian NAPE-PLDs catalyze the hydrolysis of NAPE, resulting in the formation of PA and N-acylethanolamime. The substrates of Fmp30 remain to be established, but analyses of Fmp30 with point mutations have shown that the hydrolase activity of Fmp30 is essential for its function27. How does Fmp30 function in the FMM-dependent pathway for the CL accumulation? Given that one product of the enzymatic reaction catalyzed by Fmp30 is PA, a possible explanation for the enhancement is that Fmp30 provides PA used for CL biosynthesis in the MIM. In this regard, it is noteworthy that deletion of FMP30 worsens the defects in cell growth and the CL level in yeast cells lacking the ERMES complex27, which is probably involved in PA transfer from the ER to mitochondria. The second explanation is that because PA is one of the cone shaped phospholipids, so-called “non-bilayer lipids”, the local synthesis of PA by Fmp30 induces membrane rearrangement, such as the formation of contact sites, which facilitates PA transfer from the MOM to the MIM, independently of Ups1-Mdm35. To address these possibilities, we are currently trying to clarify the substrates and products of the enzyme Fmp30.

MDM31 and MDM32 have been identified as genes required for the normal distribution and morphology of mitochondria34, and have been shown to encode homologous mitochondrial inner membrane proteins that have two membrane-spanning regions at the C-terminus and near the N-terminus, respectively, and a middle region exposed to the intermembrane space35, and form distinct protein complexes35. Deletion of MDM31 or MDM32 has been shown to decrease the mitochondrial CL level to about a half of that in wild-type mitochondria22. What might be the roles of Mdm31 and Mdm32 in the FMM-dependent pathway for the CL accumulation? Because Mdm31 and Mdm32 are MIM proteins with the same membrane topology, as described above, and exhibit 16.4% amino acid identity with each other35, these two proteins are supposed to have similar molecular functions. In the absence of Ups1, PA should be supplied to the MIM for CL synthesis in a Ups1-independent manner. In addition, deletion of MDM31 or MDM32 is synthetically lethal with deletion of the ERMES complex subunits involved in PA transfer from the ER to mitochondria35. Therefore, the simplest explanation for the roles of Mdm31 and Mdm32 in the CL accumulation is that these factors are involved in PA-supply to the MIM. The localization and topology of Mdm31 and Mdm32 are consistent with this explanation, but their amino acid sequences predict no domain structures and thus do not suggest their molecular functions. To address the molecular functions of Mdm31 and Mdm32, identification of the subunits of the Mdm31- or Mdm32-containing protein complexes35 would be an effective strategy, which is currently under investigation in this laboratory.

In conclusion, the present work shows that when the mitochondrial PE level is reduced to under the threshold level, Fmp30, Mdm31, and Mdm32 cooperatively function for the maintenance of a proper CL level even in the absence of Ups1-Mdm35-mediated PA-transfer to the MIM. This provides new insight into the CL metabolism and intramitochondrial transfer of phospholipids.

Methods

Yeast strains, genetic methods, media, and plasmids

The yeast strains used in this study are listed in Table S1. Yeast strains TKY705 (wild-type), TKY706 (wild-type α), TKY707 (psd1∆), and TKY709 (psd1∆ α) were obtained by dissection of the asci in sporulated cultures of a diploid strain, TKY12827. Complete disruption, promoter replacement, and tagging of the yeast gene were accomplished by PCR-mediated gene replacement36 with a pair of primers and a template plasmid, as listed in Table S2.

Yeast cells were grown in YPAD (1% yeast extract, 2% peptone, 0.008% adenine, and 2% glucose, pH6.0). In some experiments, YPAD was supplemented with 10 µg/ml of a tetracycline analog, doxycycline (Dox) (Nakarai Chemicals), as indicated. SCAD medium (0.67% yeast nitrogen base without amino acids, 0.2% drop out mix, 0.008% adenine, and 2% glucose, pH6.0) without leucine was used for the selection of leucine-prototrophic transformants. Cells that have the kanMX4 37, nat (natMX4 38 and natNT2 39), and hphNT1 39 genes were selected with 200 µg/ml of G418 sulfate (Nakarai Chemicals), 100 µg/ml of clonNAT (Werner BioAgents), and 300 µg/ml of hygromycin B (Nakarai Chemicals), respectively. SCAD-MSG medium (0.17% yeast nitrogen base without amino acids and ammonium sulfate, 0.1% l-glutamic acid sodium salt, 0.2% drop out mix, 0.008% adenine, and 2% glucose, pH6.0) was used for random spore analysis40.

Plasmid pCM225 + S4 carrying the primer annealing sequence S4, which was used for promoter replacement with the tetO7 promoter30, was constructed as follows. A DNA fragment containing the tetO7 promoter sequence was amplified from pCM22530 using a forward primer containing a Xho1 site, 5′-AAGCTCCTCGAGTAATTCGC-3′, and a reverse primer containing a SfiI site and the S4 sequence, 5′-ATATGGCCGCATAGGCCCATCGATGAATTCTCTGTCGATAGGCCACTAGTGGATC-3′ (underlining, S4 sequence), digested with XhoI and SfiI, and then ligated with a large DNA fragment of pCM225 digested with XhoI and SfiI.

Plasmid pRS424-FLAG-MDM31 encoding FLAG-tagged Mdm31 was constructed as follows. A DNA fragment encoding Mdm31 tagged with a FLAG epitope at near the N-terminus was constructed by PCR overlap extension recombination, using two pairs of primers (M13 rev primer (5′-CAGGAAACAGCTATGAC-3′) and FLAG-MDM31-R primer (5′-TGTCATCGTCATCCTTGTAATCCTCATTAGAATATGCTCTTAGC-3′), and FLAG-MDM31-F primer (5′-TACAAGGATGACGATGACAAGTCTAAAACTGGAAGGGATG-3′) and MDM31-R2 primer (5′-TCAATTGCGGTAGATCG-3′)), and a template plasmid, pRS424-MDM31, encoding Mdm3128 (Gift from Y. Tamura). The resulting fragment was digested with NotI and HpaI, and then ligated with a large DNA fragment of pRS424-MDM31 digested with NotI and HpaI.

Plasmid pRS424-FLAG-MDM32 encoding FLAG-tagged Mdm32 was constructed as follows. A DNA fragment encoding Mdm32 tagged with a FLAG epitope at near the N-terminus was constructed by PCR overlap extension recombination, using two pairs of primers (M13 rev primer (5′-CAGGAAACAGCTATGAC-3′) and FLAG-MDM32-R primer (5′-TGTCATCGTCATCCTTGTAATCAGCCTTGGTAGTGAAC-3′), and FLAG-MDM32-F primer (5′-TACAAGGATGACGATGACAAGTCCAATATTGAGACTATTTTGC-3′) and MDM32-R2 primer (5′-CCGTGAAATCAAACTTCG-3′)), and a template plasmid, pRS424-MDM32, encoding Mdm3228 (Gift from Y. Tamura). The resulting fragment was digested with NotI and HpaI, and then ligated with a large DNA fragment of pRS424-MDM32 digested with NotI and HpaI.

Analysis of cellular phospholipid compositions

Yeast cells were grown at 30 °C to saturation in YPAD supplemented with or without 10 µg/ml Dox. The cells were then diluted to an OD600 of 0.05 in YPD, and further incubated at 30 °C for 24 hours in the presence of 1 µCi/ml of [32P]Pi, supplemented with or without 10 µg/ml Dox. After incubation, cells were harvested, resuspended in 150 µl of 80% ethanol, heated at 95 °C for 15 min, mixed with 800 µl of chloroform/methanol (1:1, v/v), and then vortexed. 330 µl of 0.1 M HCl/0.1 M KCl was then added to the samples. The organic phase was separated by centrifugation at 3000 × g for 2 min. Samples containing equivalent radioactivity were collected and dried in a centrifugal evaporator and resuspended in chloroform/methanol (1:2, v/v). The samples were then subjected to TLC on a TLC plate (SILGUR-25-C/UV254; MACHEREY-NAGEL), which had been pretreated with 1.8% boric acid41, with the solvent system chloroform/ethanol/water/triethylamine (30:30:5:35, v/v). 32P-labeled phospholipids were detected and quantitated with an imaging analyzer, FLA-5000 (Fuji Photo Film) and MultiGauge software (Fuji Photo Film).

Statistical analysis

The results of all quantitative experiments as means for independent experiments performed multiple times as indicated. The statistical significance of mean differences was assessed by means of two-tailed Student’s t-test.

Immunoprecipitation

2 mg of yeast mitochondria was solubilized with lysis buffer (20 mM HEPES-KOH (pH 7.4), 100 mM KCl, 10% glycerol, 1% digitonin, and complete mini EDTA-free (Roche)). The lysates were incubated with anti-FLAG agarose beads (Sigma-Aldrich) at 4 °C for 2 h. After washing the beads with lysis buffer three times, immunoprecipitates were eluted with 2% SDS. The eluates were then analyzed by Western blotting using antibodies against HA (Santa Cruz), FLAG M2 (Sigma-Aldrich), and Tim23 (A gift from T. Endo, Kyoto Sangyo University, Kyoto, Japan).