The Phospholipid:Diacylglycerol Acyltransferase Lro1 Is Responsible for Hepatitis C Virus Core-Induced Lipid Droplet Formation in a Yeast Model System

Chronic infection with the hepatitis C virus frequently induces steatosis, which is a significant risk factor for liver pathogenesis. Steatosis is characterized by the accumulation of lipid droplets in hepatocytes. The structural protein core of the virus induces lipid droplet formation and localizes on the surface of the lipid droplets. However, the precise molecular mechanisms for the core-induced formation of lipid droplets remain elusive. Recently, we showed that the expression of the core protein in yeast as a model system could induce lipid droplet formation. In this study, we probed the cellular factors responsible for the formation of core-induced lipid-droplets in yeast cells. We demonstrated that one of the enzymes responsible for triglyceride synthesis, a phospholipid:diacylglycerol acyltransferase (Lro1), is required for the core-induced lipid droplet formation. While core proteins inhibit Lro1 degradation and alter Lro1 localization, the characteristic localization of Lro1 adjacent to the lipid droplets appeared to be responsible for the core-induced lipid droplet formation. RNA virus genomes have evolved using high mutation rates to maintain their ability to replicate. Our observations suggest a functional relationship between the core protein with hepatocytes and yeast cells. The possible interactions between core proteins and the endoplasmic reticulum membrane affect the mobilization of specific proteins.


Introduction
Nearly 3% of the world's population is chronically infected with the hepatitis C virus (HCV) [1], which is a major risk factor for liver cirrhosis and hepatocellular carcinoma [2].Chronic infection with HCV causes the abnormal accumulation of significant levels of liver lipids (steatosis).Steatosis has been significantly associated with hepatocellular carcinomas in HCV-infected patients [3].Steatosis is characterized by the accumulation of liver lipid droplets (LDs), which are essential for RNA replication and HCV particle formation [4,5].
The positive-stranded RNA genome of HCV encodes an approximately 3,000 amino acid polyprotein.Following translation, the amino-terminal region of the polyprotein is cleaved by a host signal peptidase to yield three structural proteins (core, E1 and E2).The other regions of the polyprotein are processed by viral proteases to yield seven nonstructural proteins [6].The C-terminal region of the core protein is further processed by a host signal peptidase to generate a mature core (amino acid 1 to 177, hereafter called core), which then enables the translocation to LDs.The core consists of two domains (D1 and D2).The D1 domain is responsible for RNA binding, whereas the D2 domain anchors to the surfaces of the ER and LDs [7,8].The core is believed to have an important role in pathogenesis, as shown by the formation of steatosis and hepatocellular carcinomas in core-transgenic mice [9].Furthermore, interactions between the core and various factors involved in lipogenesis have been reported [10,11].A triglyceride (TAG) synthetic enzyme, acyl-CoA:diacylglycerol O-acyltransferase (DGAT), is required for the core to translocate from the ER to LDs.However, the activity of DGAT1 is not affected by the core [10].The presence of the core on the surface of LDs impairs LD turnover by inhibiting the activity of adipose triglyceride lipase (ATGL) [12].It has been suggested that the core enhances the disequilibrium of ATGL and its cofactor and that the core can alter the intracellular vesicular trafficking of these factors [11].However, the mechanism underlying the effect of the core on these factors, such as the interactions between the core and the endoplasmic reticulum (ER) membrane that lead to the altered translocation of these factors, remains obscure.
Yeast (Saccharomyces cerevisiae) is a viable model system to study neutral lipid homeostasis for higher eukaryotes [13,14].Recently, we showed that the expression of the core protein in yeast reproduced several characteristics of the core protein that are observed in mammalian cells [15].The core protein is localized to the cytoplasmic side of the ER and enhances LD formation in yeast cells.Our observations suggested a functional analogy of the core between hepatocytes and yeast cells, namely in the intrinsic characteristics of the core.
LDs form when sterol ester (SE) and/or TAG accumulate and are surrounded by phospholipid monolayers.The syntheses of cellular SE and TAG are catalyzed by the acyltransferase family of proteins.TAG is synthesized by DGAT in mammalian cells.In yeast, TAG is synthesized by a DGAT homologue, encoded by the gene DGA1, and a phospholipid:diacylglycerol O-acyltransferase (PDAT), encoded by the gene LRO1 [16].Additionally, two acyl-CoA:cholesterol O-acyltransferase (ACAT)-related enzymes (Are1 and Are2) are responsible for SE synthesis [17].Cells with these enzymes (Dga1, Lro1, Are1 and Are2) disrupted failed to form LDs, whereas the presence of at least one of these enzymes permits LD formation [18].
In this work, we attempted to elucidate the genes responsible for the effect of the core protein on lipid droplet formation in yeast cells.We demonstrated that Lro1 was required for core-induced LD formation.The stability of Lro1 was extended, and the localization of Lro1 was altered in response to core expression.We further discussed a possible mechanism underlying the core-induced ER changes that may lead to LD expansion.

Construction of plasmids
A multi-copy plasmid (2 μm Ori, URA3 marker) with the GAL1 promoter (pKT10-GAL) was used to express the core proteins.A cDNA region of the core (1-177 aa) was isolated by PCR from HCV genotype 1b (GenBank: AY045702) [23] and cloned under the GAL1 promoter of pKT10-GAL (pKT10-GAL-core) [20].To construct pRS315, a CEN-based low copy plasmid with a LEU2 marker, [24] carrying DGA1-mCherry and LRO1-Myc, we isolated the genomic fragments of the fusion gene loci from CWY5135 and CWY3773 [14] by PCR and cloned them into pRS315.The core D2 region (the D2 domain and the residual C-terminus, aa 118-177) was fused to the C-terminus of DsRed-Monomer (Clontech, Takara Bio Company, Kyoto, Japan) and GFP and cloned as above to generate pKT10-GAL-DsRed-core and pKT10-GAL-GFP-core, respectively.

The detection and quantitation of LDs and fluorescence microscopy
Live yeast cells were stained with 4,4-difluoro-1,3,5,7,8-pentamethyl-4-bora-3a,4a-diaza-sindacene (BODIPY 493/503, Invitrogen, ThermoFisher Scientific, Waltham, MA, USA) as previously described [14].The fluorescent images of the LDs and the fluorescent protein fusions of the core, Hmg2, Lro1 and Dga1 were captured using a fluorescence microscope, as described previously [14].The number of LDs was counted in the images produced by the maximal projection of ten Z-optical sections (spaced at a thickness of 0.56 μm).We used the same capture condition for the fluorescent images for each of the experiments.There were no differences in contrast enhancement (BODIPY fluorescence) between various strains in each of the experiments.

Lipid analysis
Phospholipid extraction was performed using the Folch procedure [26].Briefly, 20 OD 600 units of cells were collected and treated with 10 mM NaN 3 for 10 min.The cells were pelleted and resuspended in approximately 20 μl of the original medium.The cells were lysed by vortex for 10 min in 330 μl of methanol with 100 μl of glass beads, and then, 660 μl of chloroform was added.The supernatants from the centrifugation were supplemented with 200 μl of 0.9% sodium chloride.After another centrifugation, the lower organic phase was dried using a Speedvac and then dissolved in chloroform:methanol (2:1).The lipids were analyzed by thin layer chromatography as previously described [27].To quantify TAG, we further performed liquid chromatography-mass spectrometry (LC-MS) analysis using an Orbitrap mass spectrometer (Orbitrap Elite, Thermo Fisher Scientific, Bremen, Germany) coupled with an UHPLC system (ACQUITY UPLC, Waters, Milford, MA).Solvent A was 10 mM ammonium acetate, pH 5.0 and 40% acetonitrile as the aqueous phase, and solvent B is 10 mM ammonium acetate, pH 5.0 and 10% acetonitrile in isopropyl alcohol as the mobile phase.The lipids were separated with a CSH C18 column (1.8 μm, 2.1 mm x 100 mm, Waters, Milford, MA) at a flow rate of 450 μl/min using a gradient of 40-99.5% solvent B over 0-10 minute.For TAG profiling, the mass spectrometer was operated in positive ion mode, and the full FT-MS scan data (m/z 100-1200, Res = 30,000) were analyzed and quantified with the Xcalibur software (Thermo Scientific, Waltham, MA).Linoleic acid-d4 (Cayman Chemical, MI) was added to the lipid extraction step to serve as an internal control for quantification.

Statistical analysis
All experiments were repeated at least three times, and multiple independent replicates (more than n = 3, indicated in the figure legends) were performed for each experiment.The data are presented as the mean of the replicates with the standard error of the mean (SEM).

Lro1 is responsible for core-induced LD formation
We determined the LD levels in yeast cells using BODIPY 493/503, a lipophilic dye that specifically stains neutral lipids.As we recently reported [15], the induction of the expression of the core protein for 3 h significantly enhanced LD levels in the fluorescent images (Fig 1A, Wild type, Core).As shown in Fig 1B, the countable number of LDs per cell was significantly increased by the expression of the core (Wild type, +).To identify the gene responsible for the core-dependent increase in LD formation, we examined several mutants that lack the genes for neutral lipid syntheses.ARE1 and ARE2 are responsible for SE synthesis, and DGA1 and LRO1 are responsible for TAG synthesis.The expression levels of the core in each of the disrupted mutant cells were similar (S1 Fig) .As shown in Fig 1A and 1B, the yeast strain lacking all four genes (a quadruple mutant, 4Δ), which was previously shown to be defective in forming LDs [18], also failed to form LDs by the core expression.Interestingly, the simultaneous loss of Dga1 and Lro1, but not Are1 and Are2, impaired the core-dependent induction of LD formation.We observed an increase in the punctate fluorescence of LDs by the core expression in wild-type and are1Δ are2Δ cells, but not in the quadruple mutant (4Δ) or dga1Δ lro1Δ cells (Fig 1A and 1B).Furthermore, the LRO1 single knockout had a markedly reduced level of the coreinduced LD formation, whereas no changes were observed in the DGA1 single knockout (Fig 1A and 1B).
It should be noted that morphological alterations in the cell surrounding LDs were observed in wild-type cells, are1Δ are2Δ cells and dga1Δ cells when the core was induced (Fig 1A, DIC images), but not in 4Δ cells, dga1Δ lro1Δ cells and lro1Δ cells.Thus, this morphological alteration is likely to be Lro1-dependent.
Neutral-lipid homeostasis is affected by the balance between lipid supply and consumption, and thus, carbon source availability and cell growth [13].We examined the effect of core expression on yeast growth in the culture condition used above.As shown in Fig 1C, the growth of wild-type cells and lro1Δ cells were not affected by the 3 h induction of core expression.A prolonged induction of the core expression (6 h) showed a mild growth retardation for wild-type cells and lro1Δ cells (S2B Fig) .These results revealed that the Lro1-dependent enhancement of LD levels by core expression was independent of cell growth.
As shown in Fig 2A, lipid analyses demonstrated that the TAG level was enhanced in wildtype and dga1Δ cells by core expression, whereas the levels of phospholipids (PLs), ergosterol (ERG) and SE were unaffected.Lipid profiles using the liquid chromatography also showed that the levels of TAG species, but not the levels of lysophospholipids, PLs and diacylglycerol (DAG), were enhanced in wild-type and dga1Δ cells by core expression (S3 Fig) .Most of the TAG species that were upregulated by the core expression were mainly Lro1-dependent (Fig 2B).These results suggested that Lro1 was the enzyme most responsible for the core-dependent induction of the upregulation of TAG levels, resulting in the induction of LDs.

Core suppresses ER-associated degradation (ERAD) of Lro1 but increases LDs in an ERAD-independent manner
To examine whether core expression affected the Lro1 protein level, we monitored the timedependent changes in the protein levels of Lro1 and Dga1 after the induction of core expression (Fig 3A) in yeast cells (CWY3773) that expressed both Myc-tagged Lro1 (Lro1-Myc) and Myctagged Dga1 (Dga1-Myc) under their own promoters of the respective gene loci [14].We confirmed that the expression of the core in CWY3773 conferred the induction of LD formation (Fig 3B).Furthermore, the quadruple mutant that expressed only Lro1-myc successfully enhanced LD formation in response to core expression (S4 Fig), suggesting that the 13xMyc tag at the C-terminal end of Lro1 exhibited a core-dependent phenotype.Morphological alterations were also observed in these Lro1-Myc expressing cells (Fig 3B and S4 Fig).Using Lro1myc, we monitored the effect of the core expression on the Lro1 level.The Lro1-Myc level slightly increased in response to core expression (Fig 3C and 3D).In contrast, the Dga1-Myc level was not affected by the core expression (Fig 3C and 3E).
Next, we determined the degradation rates of the proteins Lro1-Myc and Dga1-Myc in coreexpressing cells after protein synthesis had been inhibited by cycloheximide (Fig 4A).As shown in Fig 4B -4D, the cycloheximide-chase experiment suggested that the degradation rate of Lro1-Myc was significantly higher than that of Dga1-Myc in cells without core expression.However, the expression of the core protein inhibited the degradation of Lro1-Myc.Previous reports have indicated that Dga1 is a short-lived protein, whereas Lro1 is relatively stable during the late exponential growth phase in glucose medium [14].This discrepancy suggests that the fates of Dga1 and Lro1 may be dependent on different carbon sources and culture conditions.Lipid analysis for the core-expressing yeast cells.(A) TLC analyses of the neutral lipids were performed as described previously [27].The positions of ergosterol (ERG), sterol ester (SE) and triacylglycerol (TAG) are indicated.Lipids species found in wild-type cells, dga1Δ cells and lro1Δ cells carrying the pKT10-GAL-core plasmid before (raffinose) and after 3 h of culture with galactose (Galactose) compared by TLC analysis.(B) TAG species By extending the Lro1 half-life, which was induced by the core protein, we may increase LD formation.Hrd1 ubiquitin ligase is required for the degradation of proteins embedded in the ER membrane [28]

The Lro1 distribution changed in response to core expression
Because the quantitative enhancement of Lro1 by the core protein did not seem to be directly responsible for the core-dependent increase in LD levels, we speculated that the distribution of Lro1 might be affected by core expression.We examined the effect of core expression on the distribution of Dga1-mCherry and Lro1-mCherry, both of which are functional in terms of TAG synthesis [14].The distribution of Dga1-mCherry (Fig 6A ) in yeast cells growing in raffinose/galactose was perinuclear and cortical shaped and also in vacuoles, and the distribution was unaltered by core expression (Fig 6A).Interestingly, as shown in Fig 6B and 6C, mCherrylabeled Lro1 was also perinuclear and cortical shaped, especially hrd1Δ cells, in which the fluorescent level of Lro1-mCherry was enhanced (Fig 6B).The Lro1-mCherry fluorescence was shown as punctuated and laminar structures in response to core expression in the proximity of the LDs.Although a portion of Lro1-mCherry also localized to the vacuole, the vacuole distribution of Lro1-mCherry was unaltered in response to the core expression (Fig 6B and 6C).
To examine whether the presence of LDs triggered the redistribution of Lro1-mCherry, even in the absence of the core protein, we observed Lro1-mCherry in cells in which LDs were induced by tunicamycin treatment.As an ER-stress inducer, tunicamycin can increase LD formation in yeast [29] and in human hepatoma cells [30].As shown in Fig 6C, tunicamycin treatment in the absence of the core protein successfully induced LD formation but had no apparent effect on the Lro1-mCherry distribution.These results suggested that the relocalization of Lro1-mCherry did not seem to be determined by the formation of LDs.
The core D2 region is responsible for LD formation and its localization close to LDs and Lro1 As in mammalian cells [7,8], the D2 region, (the D2 domain and the residual C-terminus), has the ability to localize on the periphery of the nucleus [20].We examined whether the core D2  The same experiments in (A) were performed with hrd1Δ cells expressing Lro1-mCherry using an endogenous promoter of the chromosomal locus (hrd1Δ LRO1-mCherry: BY4742 hrd1Δ::KAN LRO1-mCherry::LEU2). (C) The expression of the core, but not of tunicamycin, altered the distribution of Lro1-mCherry in wild-type cells expressing Lro1-mCherry using an endogenous promoter of the chromosomal locus (LRO1-mCherry: BY4742 region could co-localize with LDs, as observed in mammalian cells [8].As shown in Fig 7A, DsRed-core (core amino acids 118-177 fused with DsRed) accumulated as punctate structures and localized close to the BODIPY 493/503-stained LDs.Next, we investigated whether the core protein punctate structures adjacent to the LDs were Lro1-dependent.To examine this issue, we induced LD formations using a tunicamycin treatment with the simultaneous induction of DsRed-core in lro1Δ cells.As shown in Fig 7B, the DsRed-core colocalized with LDs, which was induced by tunicamycin in lro1Δ cells.Thus, we concluded that Lro1 might be unnecessary for the punctate localization of DsRed-core on the adjacent surface of LDs.
The above results suggested that both the core protein and Lro1 might have adjacently colocalized on the surface of LDs.We found that the GFP-core was partially colocalized with Lro1-mCherry.Fig 8A shows that Lro1-mCherry was localized in larger patches and the GFPcore surrounded Lro1 patches in punctate structures.
Our previous results suggested that the core protein may affect the ER membrane [20].To examine whether the punctate localization of core protein affected the ER structure, we examined the localization of the GFP-core in cells expressing the ER membrane protein Hmg2-mCherry.As shown in Fig 8B, the GFP-core expression also altered the localization of Hmg2-mCherry.The Hmg2-mCherry fluorescence surrounding the ER membrane was altered from the laminar structures to the punctate structures by the GFP-core expression.In contrast to Lro1-mCherry, the GFP-core did not colocalized with Hmg2-mCherry (Fig 8B).Collectively, the core protein might affect the ER membrane and affect the localization of the ER LRO1-mCherry::LEU2).The cells carrying the empty vector were treated with tunicamycin (4 μg/ml) in SRM + Gal medium for 3 h (the lowest panels).Scale bars: 5 μm.doi:10.1371/journal.pone.0159324.g006proteins (Hmg2 and Lro1).However, the adjacent localization with a punctate structure of core protein was specific to Lro1, but not to Hmg2 and Dga1.

Discussion
Yeast is a good model system to study neutral lipid homeostasis [13].We utilized yeast cells as a model system to study the interaction between an animal virus protein (HCV core) and a eukaryotic cell system [20,31].In this study, based on the findings that the core protein had the ability to interact with the ER [20] and LDs (this study) and could induce LD formation in yeast (Fig 1 ), we performed genetic analyses and identified Lro1 as a factor required for the core-induced LD formation in yeast.We showed that the expression of the core protein might cause changes in the localization of the ER surface protein Lro1 to a position adjacent to LDs (Fig 6B).Despite the Lro1 changes, the distribution of another protein responsible for TAG synthesis, Dga1, was unaffected (Fig 6A).Additionally, the localization of the punctate core was not affected by the loss of Lro1 (Fig 7B).Thus, we speculated that the punctate expression of the core protein on the ER surface might guide the accumulation of Lro1 as patches, which results in LD accumulation (Fig 8C).The mechanism by which Lro1 accumulation occurs remains unknown.We failed to detect the core protein in anti-Myc immunoprecipitates from the lysate of Lro1-Myc expressing yeast (see S6 Fig).This suggested that there might be no direct interaction between Lro1 and the core protein.It should be noted that morphological alterations occurred, and the surrounding punctate LDs were observed in response to core expression in cells in the presence of Lro1.The regulatory mechanism for TAG synthesis via the regulation of Lro1 and Dga1 remains elusive.Nevertheless, our present data suggest that the mobilization of Lro1 may be regulated by a specific mechanism.
Our previous results indicate that upon the induction of LDs, the core is unable to induce the unfolded-protein response [15], suggesting that the core accumulation on the cytoplasmic side of the ER membrane has potential to affect the ER membrane integrity without influencing intra-luminal protein homeostasis.This idea is supported by the fact that the core protein inhibits the degradation of the ER surface protein Lro1, which is likely to be governed by the Hrd1-dependent ERAD system in cells (Fig 5).Thus, an investigation of the retardation of ERAD may be another interesting study on core-dependent events in yeast.Wang and Lee [14] reported that a ubiquitin-like domain containing the Ubx2 protein regulated LD homeostasis by controlling Lro1.Future studies may be interested in determining whether the core protein has an effect on the Ubx2-related system.Collectively, our results suggested that the core accumulation on the cytoplasmic side of the ER membrane changes various aspects of ER homeostasis, namely, the inhibition of ERAD, the alteration of the distribution of ER membrane proteins and the Lro1-dependent accumulation of TAGs (LDs) (see Fig 8C).
Although Lro1 is a homologue of mammalian phosphatidylcholine:cholesterol acyltransferase and lecithin:cholesterol acyltransferase, Lro1 is functionally characterized as a PDAT [32,33].DGAT, which uses acyl-CoA as an acyl donor and diacylglycerol as an acceptor to synthesize TAGs, exists in yeast and mammalian cells.In contrast, PDAT, which uses phospholipids as acyl donors, exists in yeasts and plants [13].Thus, the Lro1-related mechanism identified of Lro1-mCherry that was not affected in the absence of GFP-core expression.(B) The expression of GFP-core induced the redistribution of Hmg2-mCherry.Hrd1Δ HMG2-mCherry (BY4742 hrd1Δ::KAN HMG2-mCherry::LEU2) cells carrying the pKT10-GAL-GFP plasmid or the pKT10-GAL-GFP-core plasmid were cultured in SRM + GAL medium at 30°C.Because Hmg2 is also a substrate for the Hrd1-dependent ERAD pathway [28], we used the hrd1Δ genotype for this experiment.Fluorescent images of GFP and Hmg2-mCherry were taken.Scale bars: 5 μm.(C) A model for core-induced LD formation.The expression of core accumulated at a specific locus on the ER, which inhibits ERAD and triggers Lro1 accumulation.The accumulation of Lro1 induced LD formation.doi:10.1371/journal.pone.0159324.g008here may not be fully applicable to the core-dependent LD accumulation in mammalian cells.Previous reports indicated that the core protein appeared to inhibit LD turnover by inhibiting adipose triglyceride lipase (ATGL)-mediated lipolysis in mammalian cells by inhibiting the localization of ATGL and its cofactor [11].Additional investigations to determine the mechanism responsible for the changes in Lro1 localization by the ER-core interaction in this yeast model system may provide important insights into understanding how the core protein affects the homeostasis of ER surface proteins.
RNA virus genomes evolve quickly because of their high mutation rates to adapt to different circumstances to maintain effective replication.Our observations suggested a functional relationship between the core proteins of viruses with hepatocyte and yeast cells, which in turn implicates an intrinsic interaction of the core D2 region with the ER membrane and possible alterations of the ER membrane.S6 Fig. Immunoprecipitation of Lro1.Wild-type yeast cells carrying the empty (Lro1-myc, -) vector or pRS315-Lro1-myc (Lro1-myc, +) with the empty vector pKT10-GAL (Core, -) or the pKT10-GAL-core (Core, +) were cultured as described above.The cells were collected and resuspended in lysis buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA, 150 mM NaCl and 0.1% NP-40) containing 2 mM phenylmethanesulfonyl fluoride, 1 μg/ml leupeptin and 1 μg/ml pepstatin.The cells were then frozen in liquid nitrogen and disrupted by shaking at 2,000 rpm for 30 s with a multi-bead shocker (Yasui Kikai Corporation, Osaka, Japan).The whole cell extract (WCE; 400 μg) was mixed with anti-Myc-Tag agarose (MBL) at 4°C for 3 h.After the beads were extensively washed, the bound proteins were eluted from the beads with 90 μl of sample buffer (50 mM Tris-HCl, pH 6.8, 2% SDS, 0.1% bromophenol blue, and 10% glycerol) with 50 mM DTT.The immunoprecipitates were analyzed by SDS-PAGE (10% and 15% acrylamide gel for Lro1-Myc and Core, respectively) and immunoblotted using anti-Myc rabbit polyclonal antibodies (upper panel) and anti-core mouse monoclonal antibodies (lower panel).The positions of the immunoglobulin heavy chain (HC) and light chain (LC) are indicated.(PDF) S1 Table .Yeast strain used in this study.(PDF)

Fig 1 .
Fig 1.The core-induced LD level is mediated by the genes responsible for TAG synthesis.(A) LD levels in yeast mutants with defects in neutral lipid synthesis.The genotypes of the yeast strains are indicated in the figure.The wild type (BY4742) and its isogenic mutants, dga1Δ lro1Δ (BY4742 dga1Δ::KAN lro1Δ::HYG), are1Δ are2Δ (BY4742 are1Δ::LEU2 are2Δ::HIS3), dga1Δ (BY4742 dga1Δ::KAN) and lro1Δ (BY4742 lro1Δ::HYG).Δ4 indicates a quadruple disruption mutant (CWY3768: BY4742 are1Δ::LEU2 are2Δ::HIS3 dga1Δ::KAN lro1Δ::HYG), which lacks all four neutral lipid synthesizing genes.Cells carrying the empty vector pKT10-GAL (upper panels) or the pKT10-GAL-core plasmid (lower panels, designated as "Core") were cultured in raffinose/galactose-containing medium (SRM + Gal) for 3 h.The LDs in live yeast cells were stained with BODIPY 493/503 and analyzed by fluorescent microscopy.The differential interference contrast images (DIC) and BODIPY 493/ 503 fluorescent images (BODIPY) are shown.Scale bars: 5 μm.The expression level of core in each of disruption mutant is not significantly different (S1 Fig).(B) The number of LDs/cell was counted in images produced by the maximal projection of ten z-sections in 5 μm thickness (S2 Fig).The data are represented as the mean ± SEM (N = 5, five countings of LD number in 10 cells: total 50 cells)."*" indicates significance (P < 0.01), and other P values are indicated.No LDs were detected in Δ4 cells (n.d.).(C) The growth rate of wild-type cells and lro1Δ cells carrying the empty vector pKT10-GAL (-) or the pKT10-GAL-core (+) upon the induction of core expression.Galactose (3%) was added to exponentially growing yeast cell cultures in SRM (OD 600 = 0.1).The cultures were further cultured for 3 h.The growth levels after the addition of galactose are shown.The P values are indicated.doi:10.1371/journal.pone.0159324.g001

Fig 3 .
Fig 3. Lro1 level is enhanced by core.(A) The scheme for sample preparation.Galactose was added at time 0 to induce core expression.The cells were collected and fixed at the indicated time (0, 1.5 and 3 h).(B) LDs were successfully induced in CWY3773 (BY4742 LRO1-13xMyc::LEU2 DGA1-13xMyc::HIS3) cells expressing both Lro1-Myc and Dga1-Myc with core expression.CWY3773 cells were transformed with the empty vector pKT10-GAL (Control) or the pKT10-GAL-core plasmid (Core), and cultured for 3 h with galactose.(C) The Western blot of Lro1-Myc and Dga1-Myc.See S5 Fig for the original data.(D) and (E) The intensity was normalized using the actin intensity.The relative levels of Lro1-Myc (D) and Dga1-Myc (E) are indicated (N = 3).doi:10.1371/journal.pone.0159324.g003

Fig 4 .
Fig 4. Degradation of Lro1 is inhibited by core.(A) The scheme for sample preparation.The core expression was induced by the addition of galactose for 3 h.Then, protein synthesis was stopped by adding 0.3 mM cycloheximide (CHX) at the time 0. The cells were fixed at the indicated time (0, 0.75 and 1.5 h).(B) The CHX chase Western blot of Lro1-Myc and Dga1-Myc.CWY3773 (BY4742 LRO1-13xMyc::LEU2 DGA1-13xMyc::HIS3) cells expressing both Lro1-Myc and Dga1-Myc were transformed with the empty vector pKT10-GAL (Control) or the pKT10-GAL-core plasmid (Core).See S5 Fig for the original data.(C) and (D) The intensity was normalized using the actin intensity.The relative levels of intensity of Lro1-Myc (C) and Dga1-Myc (D) are shown (N = 3).doi:10.1371/journal.pone.0159324.g004

Fig 5 .
Fig 5. Core-induced LD formation is not mediated by inhibition of Hrd1-dependent Lro1 degradation.(A) The scheme for sample preparation is the same as Fig 4A.(B) The CHX chase western blotting of Lro1-Myc and Dga1-Myc in wild-type and hrd1Δ cells.BY4742 and BY4742 hrd1Δ cells were transformed with the pRS315-Lro1-Myc plasmid and the empty vector pKT10-GAL (Control) or pKT10-GAL-core (Core).pRS315 is a CEN-based plasmid containing LEU2 as a selection marker [24], and these vectors each contained their own promoter for LRO1 and DGA1.See S5 Fig for the original data.(C) and (D) The intensity was normalized using the actin intensity.The relative levels of the intensity of wild-type cells (C) and hrd1Δ cells (D) are shown (N = 3).(E) The LD levels in hrd1Δ cells were determined as described in the Fig 1 legend.doi:10.1371/journal.pone.0159324.g005

Fig 6 .
Fig 6.Changes in the distribution of Lro1-mCherry, but not Dga1-mCherry, in response to the core.(A) The distribution of Dga1 in cells expressing the core.Wild-type cells carrying pRS315-Dga1-mCherry with the empty vector pKT10-GAL (Control) or the pKT10-GAL-core plasmid (Core) were stained with BODIPY 493/503 after the cells were cultured in SRM + Gal medium for 3 h.The fluorescent images of BODIPY 493/503-stained LDs and Dga1-mCherry were examined using fluorescent microscopy.Images of mCherry, BODIPY 493/503, and the overlay of mCherry (red) and BODIPY (green), and DIC are shown.Representative images were magnified (right panels).Scale bars: 5 μm.(B) Changes in the distribution of Lro1 by the core expression.The same experiments in (A) were performed with hrd1Δ cells expressing Lro1-mCherry using an endogenous promoter of the chromosomal locus (hrd1Δ LRO1-mCherry: BY4742 hrd1Δ::KAN LRO1-mCherry::LEU2). (C) The expression of the core, but not of tunicamycin, altered the distribution of Lro1-mCherry in wild-type cells expressing Lro1-mCherry using an endogenous promoter of the chromosomal locus (LRO1-mCherry: BY4742