Oxysterol-binding protein homologs mediate sterol transport from the endoplasmic reticulum to mitochondria in yeast

Sterols are present in eukaryotic membranes and significantly affect membrane fluidity, permeability, and microdomain formation. They are synthesized in the endoplasmic reticulum (ER) and transported to other organelles and the plasma membrane. Sterols play important roles in the biogenesis and maintenance of mitochondrial membranes. However, the mechanisms underlying ER-to-mitochondrion sterol transport remain to be identified. Here, using purified yeast membrane fractions enriched in ER and mitochondria, we show that the oxysterol-binding protein homologs encoded by the OSH genes in the yeast Saccharomyces cerevisiae mediate sterol transport from the ER to mitochondria. Combined depletion of all seven Osh proteins impaired sterol transport from the ER to mitochondria in vitro; however, sterol transport was recovered at different levels upon adding one of the Osh proteins. Of note, the sterol content in the mitochondrial fraction was significantly decreased in vivo after Osh4 inactivation in a genetic background in which all the other OSH genes were deleted. We also found that Osh5–Osh7 bind cholesterol in vitro. We propose a model in which Osh proteins share a common function to transport sterols between membranes, with varying contributions by these proteins, depending on the target membranes. In summary, we have developed an in vitro system to examine intracellular sterol transport and provide evidence for involvement of Osh proteins in sterol transport from the ER to mitochondria in yeast.

Sterols are found in the eukaryotic organellar membranes and plasma membrane at distinct concentrations and have significant effects on membrane fluidity, permeability, and microdomain formation (1,2). They are either de novo synthesized in the endoplasmic reticulum (ER) 3 or taken up from out-side the cells, making intermembrane sterol transport critical for membrane homeostasis by the maintenance of membrane sterol levels (3). Intermembrane sterol transport is considered to be carried out by vesicular and nonvesicular mechanisms. Nonvesicular sterol transport (4 -7) has been proposed to be mediated by lipid transfer proteins and/or through membrane contact sites; however, the mechanism remains elusive (8). Oxysterol-binding protein (OSBP)-related proteins (ORPs), including OSBP homologs (Osh1-Osh7) of Saccharomyces cerevisiae, are candidates for lipid transfer proteins that transport sterols (9). Mammalian OSBP, ORP9, and the OSBP-related domain (ORD) of ORP5, as well as yeast Osh4 have been shown to transport sterols between liposomes in vitro (10 -12). In contrast, yeast Osh6 and Osh7 and mammalian ORP5 and ORP8 have been reported to transport phosphatidylserine (PS) from the ER to plasma membrane (13,14). In addition, roles of ORPs as lipid sensors or regulatory proteins in various cellular processes have also been proposed (9). Therefore, the precise molecular functions of ORPs are still debated. In S. cerevisiae, none of the seven OSH genes are essential for cell viability; however, their combined deletion cause lethality, indicating a shared and overlapping essential function (15), which is yet undetermined.
Sterols play important roles in the biogenesis and maintenance of mitochondrial membranes (16,17), and the synthesis of ergosterol, a yeast major sterol, is crucial for mitochondrial morphogenesis in S. cerevisiae (18). In mammals, cholesterol is transported to mitochondria and used for the synthesis of steroid hormones, oxysterols, and hepatic bile acids (17,19). However, the molecular mechanisms underlying sterol transport from the ER to mitochondria remain unclear. In this study, we constructed a system to evaluate sterol transport from the ER to mitochondria in vitro using membrane fractions from yeast and analyzed the molecular mechanism. Our results suggest that Osh proteins mediate sterol transport from the ER to mitochondria.

In vitro sterol transport from the ER to mitochondria
We have previously constructed a system to analyze sterol transport from the ER to mitochondria in vivo in S. cerevisiae (20). In this system, sterol transport was evaluated by measuring the content of steryl ester synthesized by a mitochondrially targeted recombinant sterol acyltransferase, SatA, from Aeromonas salmonicida (mito-SatA-EGFP) in a yeast mutant deleted for ARE1 and ARE2 encoding acyl-CoA:sterol acyltransferases (are1⌬are2⌬). SatA is a homolog of glycerophospholipid:cholesterol acyltransferases, and it is thus considered to use phospholipids as acyl donors (21). Here, to elucidate the mechanism of the transport of sterol synthesized in the ER to mitochondria, we constructed a system to evaluate the sterol transport in vitro using isolated yeast organelle membranes (Fig. 1). In this system, 3 H-labeled sterol was first synthesized in the donor ER membrane prepared from the are1⌬are2⌬ strain (22). The ER membrane was recovered and subsequently incubated with the mitochondrial membrane prepared from the are1⌬are2⌬ strain expressing mito-SatA-EGFP (20). If the 3 H-labeled sterol is transported from the ER to mitochondria, it is predicted to be esterified by mito-SatA-EGFP.
In S. cerevisiae, the mitochondria are fully developed in medium containing nonfermentable carbon sources, and we used the mitochondrial fraction prepared from the are1⌬are2⌬ strain expressing mito-SatA-EGFP cultured in medium containing lactate for in vitro sterol transport reaction. A significant amount of steryl ester was detected in the mitochondrial fraction of the are1⌬are2⌬ strain expressing mito-SatA-EGFP, suggesting that mito-SatA-EGFP is active in mitochondria ( Fig. 2A). The ER and cytosolic fractions were prepared from the are1⌬are2⌬ strain cultured in medium containing glucose (Fig. 2B).
When the ER fraction containing 3 H-labeled sterol was incubated with the mitochondria and cytosolic fractions, 3 H-labeled sterol was esterified to steryl ester in a time-dependent manner (Fig. 2B), suggesting that sterol synthesized in the ER was transported to the mitochondria in vitro. Sterol esterification was also observed in the absence of the cytosolic fraction, but its efficiency was slightly lower than that in its presence.
Sterol esterification in the absence of cytosol could be in part due to sterol transport mediated by Osh proteins associated with membranes (see below). Addition of ATP, EDTA, or divalent cations to the reaction did not show significant effect on the sterol esterification (Fig. 2, C and D). The mitochondrial membrane appeared to be intact after the reaction because Cox2 was resistant to proteinase K treatment (Fig.  2E).

Sterol transport from the ER to mitochondria mediated by Osh4
We next investigated the involvement of Osh proteins in sterol transport from the ER to mitochondria. Osh1-Osh3, Osh6, and Osh7 were reported to localize to the ER (23,24). In addition, although Osh4 was known to localize in the cytosol and to Golgi (25), significant amounts of Osh4 tagged with 3ϫFLAG (Osh4-FLAG) expressed at physiological level, confirmed to be functional (Fig. 3A), were recovered in the ER-enriched fraction (Fig. 3B). Trace amounts of Osh4-FLAG were found in the mitochondrially enriched fraction, which could be due to the contamination by ER or other membranes. In contrast, a cytosolic marker, Pgk1, was exclusively recovered in the cytosolic fraction (Fig. 3B).
To exclude the effect of Osh proteins associated with membranes, sterol transport was examined using membrane and cytosolic fractions prepared from the strains having genetic backgrounds that bear the wildtype (WT) OSH4 or the temperature-sensitive allele osh4-1 along with the deletions of the other six OSH genes (osh⌬ OSH4 and osh⌬ osh4-1) (26) (Fig.  4A). First, sterol was synthesized in vitro using the ER fractions at permissive temperature (25°C). Similar amounts of 3 H-labeled sterol were obtained in the ER fractions from both strains (Fig. 4B). The ER fractions with 3 H-labeled sterol were then incubated with the mitochondrial fractions from the same background strains expressing mito-SatA-EGFP in the presence of the cytosolic fractions at nonpermissive temperature (37°C) (Fig. 4C). The conversion of sterol to steryl ester was found to be lower in the reaction using the fractions from the osh⌬ osh4-1 strain compared with that using the fractions from the osh⌬ OSH4 strain. Next, the effect of Osh4 inactivation was examined using the cytosolic fraction of the WT, osh⌬ OSH4, or osh⌬ osh4-1 strain with the ER and mitochondrial fractions of the osh⌬ osh4-1 strain. The ER with cytosolic fractions and the mitochondrial fraction were separately preincubated at 25 or 37°C for 30 min after which these fractions were mixed and incubated for 1 h (Fig. 4D). At 25°C, lower sterol esterification was observed in the presence or absence of the cytosolic fraction of osh⌬ OSH4 or osh⌬ osh4-1 strain compared with that in the presence of the cytosolic fraction of the WT strain. At 37°C, sterol esterification was similarly lower in the presence of the osh⌬ OSH4 cytosolic fraction than in the presence of the WT cytosolic fraction. In addition, the sterol esterification was significantly lower in the reaction containing the osh⌬ osh4-1 cytosolic fraction than that with the osh⌬ OSH4 cytosolic fraction. These results suggest the involvement of OSH genes in the transport of sterol in vitro. Inactivation of Osh4 did not affect the enzymatic activity of mito-SatA-EGFP (Fig. 4E). To further verify the sterol transport function of Osh4, bacterially

Sterol transport by OSBP homologs from ER to mitochondria
expressed and purified His 6 -tagged Osh4 (His 6 -Osh4) (Fig. 5A) was added to the transport reaction having the membrane fractions from the osh⌬ osh4-1 strain at 37°C. The amount of Osh4 added to the reaction was calculated according to the reported numbers of Osh4 molecules per cell (27). Although the effect of addition of an equivalent amount of His 6 -Osh4 to the cytosolic fraction was less significant, the addition of 10 -50 times more Osh4 facilitated the sterol esterification (Fig. 5B), suggesting that sterol transport from the ER to mitochondria is mediated by Osh4. It has been shown that methyl-␤-cyclodextrin can transfer cholesterol between liposomes (28). Methyl-␤-cyclodextrin was suggested to transfer sterol from the ER to mito- Recovery of the ER and mitochondrial fractions was confirmed by Western blotting using antibodies against Kar2 and Cox2. Lower panel, the ER fraction from the are1⌬are2⌬ strain and the mitochondrial fraction from the are1⌬are2⌬ strain expressing mito-SatA-EGFP were incubated in the presence or absence of the cytosolic fractions from the are1⌬are2⌬ strain at 30°C. Data are the mean of three independent experiments. Error bars represent S.E. C, requirement of ATP for in vitro sterol transport. Sterol transport was measured in vitro using the fractions of the are1⌬are2⌬ strain in the presence or absence of 1 or 5 mM ATP or 5 mM EDTA at 30°C for 1 h. Data are the mean of triplicates. Error bars represent S.E. D, requirement of divalent cations for in vitro sterol transport. Sterol transport was measured in vitro using the fractions of the are1⌬are2⌬ strain in the presence or absence of 5 mM Mg 2ϩ and/or Ca 2ϩ at 30°C for 1 h. Data are the mean of triplicates. Error bars represent S.E. E, integrity of the mitochondrial membranes during the in vitro sterol transport reaction. After the sterol transport reaction, samples were treated with proteinase K in the absence or presence of 0.2% Triton X-100. Cox2 was detected by Western blotting using antibody against Cox2. WCE, whole-cell extract.

Sterol transport by OSBP homologs from ER to mitochondria
chondria in our in vitro system more efficiently than Osh4 (Fig. 5C).

Involvement of Osh4 in the maintenance of mitochondrial sterol level
We next examined the sterol transport in vivo in the OSH mutants expressing mito-SatA-EGFP as described previously (20). Although similar amounts of sterol were synthesized in these strains at 25°C, sterol synthesis was extremely lower in the osh⌬ osh4-1 strain compared with the osh⌬ OSH4 strain at 37°C (Fig. 6A), and accurate evaluation of the sterol transport was difficult. Therefore, the ER and mitochondrial fractions were prepared from the osh⌬ OSH4 and the osh⌬ osh4-1 strains cultured at permissive or nonpermissive temperature (Fig. 6B), and their levels of ergosterol were quantified (Fig. 6B). Sterol contents normalized by the amounts of phosphorus derived from phospholipids in the whole-cell extracts, ER-enriched fractions, and mitochondrially enriched fractions were similar between the osh⌬ OSH4 and osh⌬ osh4-1 strains cultured at 25°C. In contrast, the sterol content in the mitochondrial fraction of the osh⌬ osh4-1 strain cultured at 37°C was significantly lower than that of the osh⌬ OSH4 strain cultured under the same condition, whereas that in the ER fraction of the osh⌬ osh4-1 was similar to that of the osh⌬ OSH4 strain. These in vivo results indicate the involvement of OSH4 in the maintenance of sterol content in the mitochondria. The sterol content of the whole-cell extract was lower in the osh⌬ osh4-1 strain cultured at 37°C than in the osh⌬ OSH4 strain in agreement with the decreased sterol synthesis in the osh⌬ osh4-1 strain at 37°C (Fig. 6A). Sterol synthesis may be inhibited in the osh⌬ osh4-1 strain at 37°C due to the defect in the sterol export from the ER to other membranes (see below). The mitochondrial protein Cox2 was detected in the ER-enriched fractions from the osh⌬ osh4-1 strain, but not much Cox2 was detected in those from the osh⌬ OSH4 strain. The temperature-sensitive mutation of osh4-1 might affect separation of the ER and mitochondria.

Transport of sterol mediated by other Osh proteins
Furthermore, we examined whether other Osh proteins have the ability to transport sterol. The addition of Osh5, Osh6, and Osh7 to the in vitro sterol transport reaction having the membrane fractions from the osh⌬ osh4-1 strain at 37°C elevated the sterol esterification to higher levels than did addition of Osh4 (Figs. 5A and 7A). Using GC-MS, it was determined that  Error bars represent S.E. C, in vitro sterol transport using the ER, mitochondrial, and cytosolic fractions of the OSH mutant strains. The fractions from the osh⌬ OSH4 strain (OSH4) and osh⌬ osh4-1 strain (osh4-1) were mixed and incubated at 37°C. Data are the mean of three independent experiments. Error bars represent S.E. D, in vitro sterol transport using the ER and mitochondrial fractions from the osh⌬ osh4-1 strain with the cytosolic fraction of the WT, osh⌬ OSH4, or osh⌬ osh4-1 strain. All fractions were preincubated at 25 or 37°C for 30 min after which they were mixed and incubated at the same temperature for 1 h. Data are the mean of triplicates of two independent experiments. Error bars represent S.E. *, p Ͻ 0.05; **, p Ͻ 0.01; ***, p Ͻ 0.001 (unpaired t test, twotailed). E, sterol acyltransferase activities of the mitochondrial fractions of the osh⌬ OSH4 and osh⌬ osh4-1 strains. Sterol acyltransferase activities in the mitochondrial fractions from the osh⌬ OSH4 and osh⌬ osh4-1 strains were measured at 37°C. Data are the mean of triplicates. Error bars represent S.E. WCE, whole-cell extract. 4.0 nmol of ergosterol/reaction was included in the ER fraction used in the in vitro sterol transport assay. Thus, rates of sterol transport by these Osh proteins are calculated to be 0.092, 0.18, 0.17, and 0.15 ergosterol molecules/Osh protein/min for Osh4, Osh5, Osh6, and Osh7, respectively. Although Osh6 and Osh7 were reported to transport PS from the ER to plasma mem- Figure 5. Sterol transport mediated by Osh4 from the ER to mitochondria in vitro. A, purification of His 6 -tagged Osh proteins. His 6 -tagged proteins were expressed in E. coli and purified. The purity of each protein was confirmed by Coomassie Blue staining. B, in vitro sterol transport using recombinant His 6 -Osh4 protein. Sterol transport was measured in vitro using the ER and mitochondrial fractions from the osh⌬ osh4-1 strain in the presence or absence of 200 g of the cytosolic fraction of the osh⌬ OSH4 (osh⌬ OSH4 cytosol); 5.2, 52, or 260 pmol of His 6 -Osh4 (Osh4 ϫ 1, ϫ 10, or ϫ 50, respectively); or 12.9 g of ␥-globulin (same weight as 260 pmol of His 6 -Osh4) at 37°C after preincubation at 37°C for 30 min. Data are the mean of triplicates of two independent experiments. Error bars represent S.E. **, p Ͻ 0.01; ***, p Ͻ 0.001 (unpaired t test, two-tailed). C, in vitro sterol transport using methyl-␤cyclodextrin. Sterol transport was measured in vitro using the ER and mitochondrial fractions from the osh⌬ osh4-1 strain in the presence or absence of 52 pmol (ϫ 10) His 6 -Osh4 or 5.2 or 52 pmol of methyl-␤-cyclodextrin (m␤CD) (ϫ 1 or ϫ 10, respectively) at 37°C for 1 h after 30-min preincubation. Lipids were extracted and counted. Data are the mean of the four experiments. Error bars represent S.E. **, p Ͻ 0.01; ***, p Ͻ 0.001 (unpaired t test, two-tailed).  brane (13), the results of our sterol binding assay indicate that Osh6, Osh7, and Osh5 bind to sterol with significant affinities (Fig. 7B) respectively. Because Osh1-Osh3 proteins could not be purified using the bacterial expression system, we used the cytosolic fraction of the yeast strain that has OSH1, OSH2, or OSH3 expressed under the native promoter with deletions of the other OSH genes. The addition of these cytosolic fractions was found to elevate the sterol esterification in the sterol transport reaction (Fig. 7C), suggesting that Osh1-Osh3 proteins also have the ability to mediate transport of sterol. Because the numbers of Osh1 and Osh2 molecules in the cell were reported to be much lower than that of Osh4 (27), they may be able to transport sterol more efficiently than Osh4. It was reported that the number of Osh3 molecules per cell was not significantly different from that of Osh1 (27), and lower sterol transport activity of Osh3 could be due to its lower affinity to sterol molecules (29).

Discussion
Because sterols are synthesized in the ER, intermembrane sterol transport is critical for maintaining sterol contents in other organelles and the plasma membrane. Mitochondria are independent of vesicle transport and thus provide an excellent model to study nonvesicular sterol transport. Here, we established a system to evaluate sterol transport from the ER to mitochondria using yeast membrane fractions in vitro and analyzed the mechanism. Our results suggest that Osh proteins are involved in sterol transport from the ER to mitochondria. Because Osh4 -Osh7 bound cholesterol in vitro (Fig. 7B), it is possible that Osh proteins directly transport sterol from the ER to mitochondria.
ORPs are widely conserved in eukaryotes, and multiple paralogs of ORP genes are encoded in their genomes. It has been suggested that OSH genes of S. cerevisiae share an overlapping essential function (15). Although it has been shown that a subset of ORPs in mammals and yeast transport sterols between liposomes in vitro (10 -12), yeast Osh6 and Osh7 and mammalian ORP5 and ORP8 have been reported to mediate transport of PS from the ER to plasma membrane (13,14). It was reported that Osh6, Osh7, and ORP5 do not bind to sterols (13). However, Osh6 and Osh7 were suggested to transport cholesterol between liposomes in vitro, albeit more slowly than Osh4 and Osh5 (23). In addition, ORP5 and ORP8 were suggested to bind or transport sterols in vitro (12,30). Therefore, it is plausible that Osh6 and Osh7 transport sterols between biological membranes. The possibility that binding to phosphatidylinositol 4-phosphate (PI4P) could be a unifying feature of all ORPs and that only a subset of ORP family members also bind sterols has been proposed (9). Furthermore, ORPs have been proposed to function as lipid sensors or regulators in exocytosis (31), plasma membrane sterol organization (32), ceramide transport (33,34), phosphoinositide metabolism (35), TORC1 signaling and nitrogen sensing (36), phospholipid synthesis (37), extracellular signal-regulated kinase (ERK) signaling (38), STAT3 signaling (39), and regulation of the ABC transporter (40). Therefore, common molecular functions of ORPs have remained unclear.
Because all Osh proteins mediated sterol transport in our system (Fig. 7, A and C), one essential function shared by Osh proteins may be intermembrane sterol transport.
Osh4 -Osh7 transported 0.09 -0.18 ergosterol molecules/ Osh protein/min in our in vitro sterol transport reaction. When mito-SatA-EGFP was expressed in the are1⌬are2⌬ strain, sterols were esterified at a much higher rate than in the in vitro reaction (Fig. 2B) (20). Therefore, these in vitro transport rates could be slower than those of in vivo transport.

Sterol transport by OSBP homologs from ER to mitochondria
The conditions of the in vitro sterol transport reaction might not be optimized.
It has been reported that a subset of ORPs transport sterols and PS from the ER to the Golgi and plasma membrane through counterexchange for PI4P. Furthermore, the transport of sterols and PS is coupled with the synthesis and degradation of PI4P (14,(41)(42)(43). In our in vitro sterol transport reaction, significant sterol transport was observed in the absence of ATP or the presence of EDTA (Fig. 2C), suggesting that synthesis of phosphoinositides is not required for in vitro sterol transport from the ER to mitochondria, although the presence of phosphoinositides in the mitochondrial membrane needs to be validated. It has been proposed that Osh4 transports sterol against a pre-existing sterol concentration gradient through sterol/ PI4P counterexchange using the energy of PI4P hydrolysis (44). Because ergosterol content in the mitochondrial outer membrane is lower than that in the ER in yeast (2), the energy generated by PI4P degradation might not be required for sterol transport from the ER to mitochondria. Further in vitro sterol transport analysis using membrane and cytosolic fractions of mutant strains defective in phosphatidylinositol 4-kinases will contribute to elucidating the involvement of PI4P in sterol transport from the ER to mitochondria.
When sterol transport was examined using the ER and mitochondrial fractions from the WT strains (are1⌬are2⌬), the effect of addition of the cytosolic fraction on sterol transport was less significant (Fig. 2B). If Osh4 directly transports sterols to mitochondria, then addition of the cytosolic fraction should increase the transport rate. Therefore, it cannot be denied that Osh4 indirectly affects sterol transport to mitochondria. Osh proteins might mediate sterol transport through forming contact sites between the ER and mitochondria because it was reported that ORDs of Osh proteins simultaneously bind two membranes (23). When the ER and mitochondrial fractions from the osh⌬ osh4-1 strain were used, addition of the cytosolic fraction from the WT or osh⌬ OSH4 strain significantly facilitated sterol esterification at 37°C (Fig. 4D). In Fig. 3B, it was suggested that a fraction of Osh4 localizes to the ER. In addition, it has also been reported that Osh1-Osh3, Osh6, and Osh7 localize to the ER (23,24). Therefore, sterol transport in the absence of the cytosol could in part be mediated by Osh proteins associated with the ER membranes.
None of the Osh proteins have been shown to be localized to mitochondria. In addition, unambiguous localization of Osh4-FLAG to the mitochondrial membrane was not observed in this study (Fig. 3B). Osh proteins may, therefore, be transiently targeted to mitochondria. In mammals, ORP5 and ORP8 are localized to ER-mitochondria contacts in addition to ER-plasma membrane contact sites (45). Because the ORD of ORP5 transports sterol between liposomes in vitro (12) and Osh6 and Osh7, which are orthologs of ORP5 and ORP8, transport sterol in vitro (23) (Fig. 7A), these ORPs could mediate sterol transport from the ER to mitochondria in vivo.
Esterification of radiolabeled sterol was observed in the absence of Osh proteins in vitro (Figs. 4D; 5B; and 7, A and B). Osh-independent sterol transport could be mediated by cytosolic protein(s) associated with membranes or membrane-anchored protein(s). In S. cerevisiae, Lam6/Ltc1, a steroidogenic acute regulatory protein-related lipid transfer (StART) domain-containing protein with a transmembrane domain, is localized to the ER-mitochondria contact site, and its cytosolic domain transported sterol between liposomes in vitro (46,47). Thus, Osh-independent sterol transport from the ER to mitochondria could be mediated by a StART domain-containing protein(s). ER-mitochondria contact sites formed by the ER-mitochondria encounter structure complex (ERMES) (48) and the conserved ER membrane protein complex (EMC) (49) have been suggested to be involved in phospholipid transport between the ER and mitochondria in S. cerevisiae. A contact site, named the vacuole and mitochondria patch (vCLAMP), between the vacuole and mitochondria (50,51) was also proposed to serve as a route for phospholipid transport. The role of StART domaincontaining proteins and ER-mitochondria contact sites in sterol transport can be determined by analyzing sterol transport using our system.
Sterols play important roles in the structures and functions of various membranes. A system in which sterol acyltransferase is targeted to a membrane of interest in the are1⌬are2⌬ strain and the sterol transport is assessed by determining sterol esterification can be applied to study the transport of sterol synthesized in the ER to other membrane compartments.
Plasmids used in this study are listed in Table 2. Primers used for plasmid construction are listed in Table 3. The plasmid YCp33-Osh4-FLAG was constructed as follows. A DNA fragment encoding native promoter and ORF of OSH4 was amplified with primers BamHI-OSH4 and OSH4-SalI using total DNA of W303-1A as a template, and the amplified fragment was digested with BamHI and SalI and cloned into BamHI-SalI Sterol transport by OSBP homologs from ER to mitochondria sites of YCplac111 (54). A DNA fragment encoding 3ϫFLAG was amplified with primers SalI-FLAG-f and FLAG-TDH3-r using p3ϫFLAG-Myc-CMV TM -26 (Sigma) as a template. TDH3 terminator was amplified with primers FLAG-TDH3-f and TDH3-HindIII-r using pMito-SatA (20) as template. A DNA fragment encoding 3ϫFLAG with TDH3 terminator was obtained by fusion PCR. The obtained fragment was digested with SalI and HindIII and cloned into SalI-HindIII sites of YCp111-OSH4, resulting in YCp111-Osh4-FLAG. YCp111-Osh4-FLAG was digested with BamHI and HindIII, and a DNA fragment containing the promoter and ORF of OSH4, 3ϫFLAG, and TDH3 terminator was cloned into BamHI-Hin-dIII sites of YCplac33 (54).

Sterol transport by OSBP homologs from ER to mitochondria
liters of YPD medium, and the CBY926are1⌬are2⌬ strain was cultured in 8 liters of YPD medium. Spheroplasts were formed in spheroplast formation buffer containing 1 ⁄ 2 YPD medium. ER fractions were prepared in two or four batches, respectively, and mixed. The mitochondrial fractions were prepared from the WT and are1⌬are2⌬ strains containing vector and are1⌬are2⌬, CBY924are1⌬are2⌬, and CBY926are1⌬are2⌬ strains containing pMito-SatA (20) according to the method of Zinser and Daum (58) with slight modifications. The yeast strains were precultured in 10 ml of SD medium containing methionine for 1 day and then transferred to 50 ml of semisynthetic lactate medium in 250-ml-scale flasks. After 1-day cultivation, cells were transferred to 1 liter of semisynthetic lactate medium in 5-liter-scale flasks and incubated until A 600 reached 2.0 -2.5. Spheroplasts were formed in spheroplast formation buffer containing 1 ⁄ 2 semisynthetic lactate medium. The spheroplasts were suspended in 20 ml of MES lysis buffer (20 mM MES-KOH, pH 6.0, 0.6 M sorbitol, 1 mM DTT, 1% (v/v) protease inhibitor mixture) and mildly disrupted using a French press at 1,500 p.s.i. Unbroken cells and cell debris were removed by centrifugation at 1,500 ϫ g for 5 min at 4°C, and the supernatant was collected. The pellets were resuspended in 10 ml of MES lysis buffer and centrifuged at 1,500 ϫ g for 5 min at 4°C, and the supernatant was collected. This washing process was repeated.
The supernatants were mixed and used as the whole-cell extracts. The whole-cell extracts were centrifuged at 13,000 ϫ g for 10 min at 4°C, and the pellets were resuspended in 2 ml of MES lysis buffer (P13 fraction). Two milliliters of the P13 fraction was loaded on 11 ml of sucrose step gradient containing 5.5 ml of 30% (w/v) sucrose in MES lysis buffer and 5.5 ml of 60% (w/v) sucrose in MES lysis buffer and centrifuged in the swing rotor at 28,000 rpm for 3 h at 4°C. The membranes at the 30%/60% interface were collected, washed once with MES lysis buffer, and resuspended in buffer containing 25 mM HEPES, pH 7.5, 100 mM KCl, 10% (w/v) glycerol, 1 mM DTT, 1% (v/v) protease inhibitor mixture. The WT and are1⌬are2⌬ strains containing vector and the are1⌬are2⌬ strain containing pMito-SatA were cultured in 1 liter of semisynthetic lactate medium for steryl ester detection. The are1⌬are2⌬ and CBY924are1⌬are2⌬ strains were cultured in 4 liters of semisynthetic lactate medium, and the CBY926are1⌬are2⌬ strain was cultured in 12 liters of semisynthetic lactate medium in total. Mitochondrial fractions were prepared in four or six batches, respectively, and mixed. The purity of ER and mitochondrial fractions was analyzed by immunoblotting using anti-Kar2 (59) and anti-Cox2 (Invitrogen) antibodies as described previously (20).

Sterol transport by OSBP homologs from ER to mitochondria
vated to an A 600 0.4 -0.5 in 500 ml of LB medium or to an A 600 of 0.7 in 1,000 ml of 2ϫ YT medium at 25°C. Following 4 -6 h of incubation with 0.1 mM isopropyl ␤-D-thiogalactopyranoside, cells were collected and broken using a French press in breaking buffer (25 mM HEPES-KOH, pH 7.4, 100 mM KCl (for His 6 -Osh4) or 400 mM KCl (for His 6 -Osh5, Osh6, and Osh7), 10% (w/v) glycerol, 1 mM DTT, 0.5% (v/v) protease inhibitor mixture) for use in purification of histidine-tagged proteins (Sigma). Cell lysate was centrifuged at 18,000 ϫ g for 30 min at 4°C. The supernatant was incubated overnight at 4°C with Profinity immobilized metal affinity chromatography nickelcharged resin (Bio-Rad) equilibrated with breaking buffer. The resin was extensively washed first with breaking buffer containing 1% (v/v) Triton X-100, then with breaking buffer without Triton X-100, and finally with breaking buffer containing 20 mM imidazole. Protein was eluted from the resin with breaking buffer containing 500 mM imidazole. The buffer was exchanged using a NAP-5 column (GE Healthcare) to buffer A (25 mM HEPES-KOH, pH 7.4, 100 mM KCl, 10% (w/v) glycerol, 1 mM DTT).

Sterol transport assay in vitro and in vitro
Sterols were synthesized in vitro as described (22) with slight modification as follows. The reaction mixture (50 l) containing the ER fractions (100 g of protein), cytosolic fractions (200 g), 3 mM reduced GSH, 3 mM glucose 6-phosphate, 1 unit of glucose-6-phosphate dehydrogenase, 5 mM ATP, 1 mM NADPH, 1 mM NADP, 1 mM NAD, 2 Ci of L-[methyl-3 H]methionine (PerkinElmer Life Sciences), 25 mM HEPES (pH 7.4), 100 mM KCl, 10% glycerol, 5 mM MgCl 2 , and 2 mM MnCl 2 was incubated at 30 or 25°C. After 1-h incubation, the reaction mixture was centrifuged at 27,000 ϫ g for 11 min at 4°C. The pellets were suspended in buffer containing 25 mM HEPES-KOH, pH 7.4, 100 mM KCl, 10% (w/v) glycerol, 1 mM DTT, and 1% (v/v) protease inhibitor mixture and used as ER fractions for the sterol transport assay. The sterol transport assay was performed as follows. The reaction mixture (50 l) containing the ER fraction obtained as described above and the mitochondrial fractions (150 g of protein) in the presence or absence of the cytosolic fractions (200 g of protein) or purified proteins was incubated at 30 or 37°C. In the reaction with preincubation, the ER and cytosolic fractions or Osh proteins and the mitochondrial fraction were separately preincubated at 25 or 37°C for 30 min after which these fractions and proteins were mixed and incubated at the same temperature for 1 h. The reaction was stopped by addition of chloroform/methanol (1:2, v/v). Lipids were extracted by the method of Bligh and Dyer (60), dissolved in chloroform/methanol (2:1, v/v), and separated by thin layer chromatography (TLC) on Silica Gel 60 (Merck, 5721) with hexane/diethyl ether/acetic acid (80:20:1, v/v/v) as a developing solvent. After spraying with sulfuric acid/acetic acetate (1:1, v/v), plates were heated at 90°C for 10 min. Sterols and steryl esters were scraped off the TLC plate, and their radioactivities were determined by a liquid scintillation counter (Aloka) using Clear-sol I (Nacalai Tesque).
To examine the integrity of the mitochondrial membrane during the in vitro sterol transport reaction, the reaction mixture (50 l) containing the ER fraction (100 g of protein) from the are1⌬are2⌬ strain, the mitochondrial fraction from the are1⌬are2⌬ strain expressing mito-SatA-EGFP (150 g of protein), 25 mM HEPES-KOH, pH 7.4, 100 mM KCl, and 10% (w/v) glycerol was incubated at 30°C for 1 h. After the incubation, samples were treated with 1 g/ml proteinase K (Wako) in the absence or presence of 0.2% Triton X-100 at 4°C for 30 min. Cox2 was detected by Western blotting using antibody against Cox2.
The in vivo sterol transport assay was performed as described previously (20) with modifications as follows. The osh⌬ OSH4 and osh⌬ osh4-1 strains containing pMito-SatA (20) were cultivated at 25°C overnight in SD medium containing methionine. Cells were harvested and washed twice in SD medium. Cells were resuspended in SD medium at 5 A 600 /ml and incubated at 25°C for 1 h and then at 25 or 37°C for 30 min.
L-[methyl- 3 H]Methionine was added at a final concentration of 2 Ci/ml, and cells were further incubated at 25 or 37°C for 5 min. Lipids were extracted, and their radioactivities were determined.

Analysis of Osh4-FLAG distribution
The W303Ϫ1A strain containing YCp111-Osh4-FLAG was cultured in 2 liters of SD medium until A 600 reached 1.5-2.0. Cells were collected, and spheroplasts were prepared. The spheroplasts were suspended in 20 ml of MES lysis buffer and mildly disrupted using a French press at 1,500 p.s.i. Unbroken cells and cell debris were removed by centrifugation at 1,000 ϫ g for 10 min at 4°C, and the supernatant was collected. The pellet was suspended in 10 ml of MES lysis buffer and centrifuged at 1,000 ϫ g for 10 min at 4°C, and the supernatant was collected. The supernatant was mixed and used as whole-cell extracts. An aliquot of the whole-cell extracts was used to prepare fractions of mitochondria, and another aliquot of the whole-cell extracts was used to prepare ER and cytosolic fractions as described above. The recovery of the ER and mitochondria in obtained fractions was confirmed as described previously (20). The recovery of Osh4-FLAG and Pgk1 was analyzed by Western blot analysis using anti-FLAG M2 (Sigma) and anti-Pgk1 (Abcam) antibodies. Images were obtained using a LAS-500 (Fujifilm), and band intensity was quantitated with Image Gauge 3.4 (Fujifilm).

Quantification of ergosterol in fractions
The CBY924 and CBY926 strains were cultured in 800 ml of YPD medium to an A 600 of 2.0 at 25°C after which cells were incubated at 25 or 37°C for 80 min. Spheroplasts were formed in 10 ml of spheroplast formation buffer containing 1 ⁄ 2 YPD medium at 25 or 37°C, respectively. The spheroplasts were suspended in MES lysis buffer and mildly disrupted using a French press. Unbroken cells and cell debris were removed by centrifugation at 1,500 ϫ g for 5 min at 4°C, and the supernatant was collected. The pellets were resuspended in 14 ml of MES lysis buffer and centrifuged at 1,500 ϫ g for 5 min at 4°C, and the supernatant was collected. The supernatants were mixed and used as whole-cell extracts. Small aliquots of the whole-cell extracts were separated for the quantification of ergosterol and phospholipids. One-third of the whole-cell extracts was used for the preparation of the ER fractions, and two-thirds of the whole-cell extracts was used for the preparation of the mitochondrial fractions as described above. Lipids in the whole-cell Sterol transport by OSBP homologs from ER to mitochondria extracts, ER fraction, and mitochondrial fraction were extracted as described above. Cholesterol was added to each sample for use as an internal standard. Extracted lipids were dried under nitrogen, dissolved in ethyl acetate, and analyzed by GC-MS suing QP2010 SE (Shimadzu) with a DB5 capillary column (30 m ϫ 0.25 mm; J & W Scientific). Samples were injected at a 280°C injection temperature with helium as a carrier gas. The following temperature program was used: 1 min at 120°C, 75°C/min to 270°C, 3°C/min to 300°C, and 3-min hold at 300°C. Total phospholipid contents of the whole-cell extracts, ER fraction, and mitochondrial fraction were determined by phosphorus assay according to the method of Bartlett (61). The ergosterol content in the ER fraction of the CBY926are1⌬are2⌬ strain used for the in vitro sterol transport assay was measured by GC-MS as described above.

Sterol acyltransferase activity assay
The sterol acyltransferase activity assay was performed as described (62) with slight modifications. Each reaction contained the mitochondrial fractions (150 g) from the osh⌬ OSH4 or osh⌬ osh4-1 strain, 25 mM HEPSE-KOH, pH 7.5, 100 mM KCl, 1 mM DTT, and 0.1 Ci of [4-14 C]cholesterol (American Radiolabeled Chemicals Inc.) in a total volume of 50 l. The reaction was performed at 37°C for 1 h and stopped by the addition of chloroform/methanol (1:2, v/v). Sterols and steryl esters were extracted and separated by TLC. Their radioactivities were measured as described above. [1,[2][3] H]Cholesterol (PerkinElmer Life Sciences) binding by Osh proteins was performed as described (63) with slight modifications as follows. His 6 -tagged Osh proteins (10 pmol) were incubated in 50 l of binding buffer (25 mM HEPES-KOH, pH 7.4, 100 mM KCl, 0.05% Triton X-100) containing radioactive cholesterol in the presence or absence of a 40-fold excess of unlabeled cholesterol. After incubation at 25°C for 1 h, 10 l of nickel-charged resin (Bio-Rad) was added and incubated for 15 min at 25°C. The resin was pelleted by centrifuge at 15,000 rpm for 1 min. After three washes with wash buffer (25 mM HEPES-KOH, pH 7.4, 100 mM KCl), the proteins were eluted from the resin with 710 l of elution buffer (25 mM HEPES-KOH, pH 7.4, 100 mM KCl, 150 mM imidazole). After centrifugation at 15,000 rpm for 10 min, radioactivities in the supernatants were determined as described above. Binding curves were plotted, and K d and B max were determined using GraphPad Prism software.