Oleosin of plant seed oil bodies is correctly targeted to the lipid bodies in transformed yeast.

Yeast (Saccharomyces cerevisiae) has been used extensively as a heterologous eukaryotic system to study the intracellular targeting of proteins to different organelles. The lipid bodies in yeast have not been previously subjected to such studies. These organelles are functionally equivalent to the subcellular storage oil bodies in plant seeds. A plant oil body has a matrix of oils (triacylglycerols) surrounded by a layer of phospholipids embedded with abundant structural proteins called oleosins. We tested whether plant oleosin could be correctly targeted to the lipid bodies in transformed yeast. The coding region of a maize (Zea mays L.) oleosin gene was incorporated into yeast high copy and low copy number plasmids in which its expression was under the control of GAL1 promoter. Yeast strains transformed with these plasmids produced oleosin when grown in a medium containing galactose but not glucose. The oleosin produced in yeast had a molecular mass slightly higher than that of the native protein in maize. Oleosin accumulated concomitantly with the storage lipids during growth of the transformed yeast, and it was not secreted. Subcellular fractionation of the cell extracts obtained by two different cell breakage procedures revealed that the oleosin was largely restricted to the lipid bodies. Oleosin apparently did not affect the lipid contents and composition of the transformed yeast lipid bodies but replaced some of the native proteins associated with the organelles. Immunocytochemistry of the transformed yeast cells showed that the oleosin was present mostly on the periphery of the lipid bodies. Oleosin isolated from maize or transformed yeast strain, alone or in the presence of phospholipids or SDS, did not bind to the yeast lipid bodies in vitro We conclude that plant oleosin is correctly targeted to the lipid bodies in transformed yeast and that yeast may be used as a heterologous system to dissect the intracellular targeting signals in the oleosin.

Diverse organisms store lipids in subcellular particles as food reserves that will be mobilized during a forthcoming period of active metabolism. These lipid particles can be found in seeds, pollens, spores, and vegetative organs of plants (1). They are also present in the brown adipose (2) and other tissues of mammals (3), the eggs of some nematodes (4), and unicellular organisms such as yeast (5,6), Euglena (7), and algae (8). Of all these subcellular storage lipid particles, those from plant seeds have been studied most extensively.
The seeds of many plant species store triacylglycerols (TAGs) 1 as food reserves for germination and growth of the seedlings (1,3). The TAGs constitute about 5-40% of the total seed dry weight. They are present in small discrete subcellular organelles called oil bodies (lipid bodies, oleosomes, and spherosomes). The spherical organelles have diameters of about 0.6 -2.0 m, depending on the plant species in which they occur. Each oil body contains a TAG matrix surrounded by a layer of phospholipids (PL) embedded with proteins termed oleosins (9). Oleosins essentially cover the whole surface of the oil body and represent 1-4% of the total mass of the oil body; this percentage is related to the size of the organelle. The M r of oleosins range from 15,000 to 26,000, depending on the isoforms and plant species in which they occur. Each oleosin molecule has the following three structural portions (10 -13): (a) an N-terminal amphipathic stretch (20 -60 residues) of undefined secondary structures residing on the organelle surface; (b) a central hydrophobic domain (72 residues) of long antiparallel ␤-structures penetrating into the matrix; and (c) a C-terminal amphipathic ␣-helix (30 -40 residues) locating on the organelle surface, interacting with the PL layer. The three structural portions enable the protein to interact with other molecules on the surface of an oil body. Oleosins form a steric barrier on the surface of an oil body, preventing the PL layers of adjacent oil bodies from contacting and coalescing. Maintenance of the oil bodies as small entities provides a large surface area per unit TAG and would facilitate lipase binding and lipolysis during germination.
During seed maturation, oil bodies are synthesized on the rough ER by a budding process, which has been postulated on the basis of experimental evidence and consideration of the thermodynamics involved (1). TAGs are synthesized by enzymes in the ER and are sequestered, because of their hydrophobicity, between the two PL layers of the ER membrane, thereby forming a bud. PL synthesized by enzymes in the ER also diffuse to the surface of this budding TAG particle. Simultaneously, oleosin is synthesized on polyribosomes bound to the ER without appreciable co-or posttranslational processing. The newly synthesized oleosin moves to the budding TAG particle, presumably guided by the central hydrophobic domain. Alternatively, the oleosins being synthesized on polyribosomes are inserted directly into the budding particle. The budding particle, which has a TAG matrix surrounded by a layer of PL and oleosins, is released into the cytosol as a mature oil body.
Although it is well documented that oleosin is synthesized by polyribosomes bound to the ER, no cleavable N-terminal signal sequence in the protein has been identified (1). Apparently, oleosin does not enter the lumen of the ER, otherwise budding would occur at the luminal side of the ER and the newly formed oil body would enter the intracellular secretary pathway. The latter scenario occurs in the secretion of lipoproteins in mammals (14). The targeting signal in oleosin for the ER or budding oil body is unclear. At the N-terminal region of oleosin where a targeting signal could occur, there are no appreciable similarities in the length, the amino acid residues, and the charge distribution among oleosins of diverse plant species. Oleosin has a conserved central hydrophobic domain that could act as the targeting signal for the hydrophobic matrix of the oil bodies. This targeting signal would be equivalent to the internal hydrophobic targeting sequences in many membrane proteins of diverse organisms (15,16). Whatever the targeting signal in the oleosin is, it appears to be universal, since maize (a monocotyledonous species) oleosin is correctly targeted to seed oil bodies in transformed Brassica (a dicotyledonous species) (17).
Yeast has been used extensively as a heterologous system to study the intracellular targeting of proteins (review, Ref. 18). Proteins from other eukaryotes have been tested to see whether they are correctly targeted to the appropriate subcellular locations in transformed yeast. These locations include the nuclei, mitochondria, peroxisomes, vacuoles, plasma membranes, and extracellular medium, etc. An obvious omission from this list of subcellular locations is the cytoplasmic lipid bodies. In yeast, the lipid bodies, of about 0.2-0.5 m in diameter, contain about 50% TAGs and 50% steroid esters, which are accumulated as food reserves during the log phase and the late stage of growth (5,6). These lipid bodies are similar to the seed oil bodies in function; whether they are also similar in structure is unknown. In this report, we show that plant seed oleosin is correctly targeted to the lipid bodies in yeast transformed with the plant oleosin gene.

EXPERIMENTAL PROCEDURES
Yeast Strains and Plasmids-Saccharomyces cerevisiae DFY1-1A (MATa, ura3-53, Lys) and the yeast/Escherichia coli shuttle vectors pQC5 and pQC6 (19) were kindly provided to us by Drs. Q. Chao and M. E. Etzler, Section of Molecular and Cellular Biology, University of California, Davis, CA. Plasmid pT7⍀ (20) was obtained from Dr. D. R. Gallie, Department of Biochemistry, University of California, Riverside, CA.
Construction of Yeast Vectors Containing Plant Oleosin Gene-A 594-base pair fragment in a cDNA clone, pL2Ϯ (21), containing the complete coding sequence and 30 base pairs of the 5Ј-untranslated sequence of maize 18-kDa oleosin, was amplified by polymerase chain reaction. Using primers containing sequences for creating specific restriction sites, EcoRI and BamHI sites were added to the 5Ј and the 3Ј end of this fragment, respectively. The blunt-ended EcoRI-BamHI fragment was then subcloned into SmaI-cleaved plasmid pT7⍀ (20) to create plasmid pL2T02 (Fig. 1).
The maize oleosin gene from plasmid pL2T02 was cut with EcoRI and ligated into the EcoRI site of the low copy number centromeric plasmid pQC5 (19) and the high copy number 2-m plasmid pQC6 (19).
Yeast Culture Condition and Transformation-Yeast strain was maintained on a agar medium of 1% yeast extract, 2% peptone, and 2% glucose. Transformation was performed by the lithium acetate method (22). Transformants were selected for uracil auxotrophy on uracil-omitted synthetic complete medium (SCM uraϪ) (22). In the induction of the GAL1 promoter in the transformed yeast strains, glucose (2%) in SCM uraϪ medium was replaced by galactose (2%).
Yeast growth was monitored by reading the A 600 nm of the undiluted culture in a spectrophotometer. The reading reached a maximum of 2.3 after 50 h of growth (to be indicated as 100% in Fig. 4). The reading 2.3 was an underestimated value, since the spectrophotometric reading became nonlinear beyond 1.0. When the 50-h culture was diluted with culture medium such that the reading of A 600 nm was at the range of 0.5-1.0, the calculated A 600 nm of the 50-h culture was about 4 -6. All the A 600 nm readings described in the subsequent sections of "Experimental Procedures" indicated those of the undiluted cultures.
Northern Hybridization-Yeast RNA was prepared from mid-log phase cells (A 600 nm ϭ 1.0) by a method using glass beads for cell breakage (23). Electrophoresis was carried out with a 1.5% (w/v) agarose, 2.2 M formaldehyde gel (24), and 10 g of RNA was applied to each lane. Following electrophoresis, the RNA was transferred to a nylon membrane and hybridized with a maize 18-kDa oleosin cDNA probe. The membrane was washed as described (25). Plasmid pL2T01 containing the complete coding sequence and 30 base pairs of the 5Ј-untranslated sequence of the maize oleosin gene was used as a template to make the DNA probe by a method using random primers (24).
Preparation of Yeast Crude Extract-Yeast cells in 20 ml of culture medium were grown to an A 600 nm of 2.0 and harvested by centrifugation at 3000 ϫ g for 5 min. Cells were resuspended in TS buffer (50 mM Tris-HCl, pH 7.5, 1 M sorbitol) and heated at 100°C for 10 min. The heated cells were centrifuged again and resuspended in new TS buffer. Acid-washed glass beads (425-600 microns) were added to the cell suspension. The cells were lysed by vortexing the suspension at a maximum speed for 10 min. The supernatant was collected as crude extract after the beads settled.
SDS-PAGE and Immunoblotting-Urea (8 M) and acrylamide (12.5%) SDS-PAGE and the pretreatment of samples were performed as described (26). Immunoblotting and the preparation of chicken antibodies raised against maize 18-kDa oleosin were as described (27).
Subcellular Fractionation-Yeast cells harboring plasmid pQC6-ole were grown in 400 ml of SCM uraϪ medium containing 2% galactose at 30°C to an A 600 nm of 2.0 (at the late log phase). Cells were collected by centrifugation (3000 ϫ g for 5 min), and the leftover medium was saved for further analysis. The pelleted cells were resuspended in 60 ml of TS buffer containing 3 mg/ml lyticase (ICN, Costa Mesa, CA) and incubated at 30°C for 30 min. The resulting spheroplasts were harvested by centrifugation (3000 ϫ g for 5 min), washed in TS buffer, and resuspended in 60 ml of 50 mM Tris-HCl, pH 7.5, 2 mM PMSF. The resuspension in a 250-ml flask was shaken vigorously by hand and homogenized in an ice-cold glass tissue homogenizer with 25 strokes, using a tight-fitting pestle. The preparation was centrifuged at 1000 ϫ g for 5 min to remove cell debris. The supernatant was designated as the crude homogenate. An aliquot of the crude homogenate was saved for further use. Sorbitol was added to the crude homogenate to give a final concentration of 1 M. The homogenate (15 ml each in two tubes) was overlaid with 15 ml of TPS buffer (50 mM Tris-HCl, pH 7.5, 2 mM PMSF, 0.5 M sorbitol), and centrifuged at 100,000 ϫ g for 2 h. A floating layer consisting of lipid bodies was collected from the top of the tube with a pipette. The leftover culture medium, the 100,000 ϫ g supernatant, and the crude homogenate were subjected to 80% ammonium sulfate precipitation to concentrate the proteins in the fractions. Each precipitated sample was resuspended in 300 l of Tricine buffer (0.15 M Tricine-KOH, pH 7.5, 1 mM MgCl 2 , 10 mM KCl, 1 mM EDTA, 2 mM dithiothreitol, 0.25 M sucrose). The 100,000 ϫ g pellet was resuspended in 100 l of Tricine buffer. The lipid body preparation was microcentrifuged at 14,000 ϫ g for 15 min. The floating layer was left undisturbed, and the solution below the layer was removed with a fine needle. The microcentrifugation was performed several times to concentrate the lipid bodies to a final volume of 100 l.
An alternative method of subcellular fractionation was carried out. Yeast cells collected by centrifugation (3,000 ϫ g for 5 min) were mixed with cold, acid-washed glass beads and homogenized with a mortar and pestle in a cold room. TPS buffer was added to make a suspension. The suspension was centrifuged (1,000 ϫ g for 5 min) to yield a crude homogenate. The crude homogenate was subjected to a similar procedure of subcellular fractionation by centrifugation, as described in the preceding paragraph.
Preparation of Oleosin for in Vitro Binding Test-Maize oleosin, in a 1:1:2 (w/w) mixture of three isoforms of 18, 17, and 16 kDa, was prepared from oil bodies isolated from maize kernel (9). The oil body fraction, in 0.15 M Tricine-KOH, pH 7.5, 1 mM MgCl 2 , 10 mM KCl, 1 mM EDTA, 2 mM dithiothreitol, 0.25 M sucrose, was mixed with an equal volume of diethyl ether to remove the neutral lipids. The ether extraction was repeated twice. The remaining suspension of oleosin and PL was sonicated and used (as preparations a and b under "Results"). An aliquot of the suspension was subjected to a procedure for PL removal using chloroform and methanol as described earlier (9). After removal of the PL, the remaining oleosin fraction was resuspended in either water or a solution containing 0.125 M Tris-HCl, pH 8.0, 1 mM EDTA, and 0.1% SDS, and the water suspension was sonicated and used (as preparations c and d, respectively, under "Results"). Sonication of each of the above suspensions was performed with a 4-mm diameter probe in a Braun-Sonic 2,000 ultrasonic generator (Freeport, IL) with a digital meter reading of 100 for three 20-s periods.
Oleosin synthesized in yeast strain transformed with pQC6-ole was prepared as follows. The transformed yeast cells were homogenized with glass beads, and the lipid bodies were isolated as described in the preceding section. The lipid body fraction was treated with diethyl ether to remove the neutral lipids as mentioned in the preceding paragraph. The resulting suspension of the 18-kDa oleosin, other proteins, and PL was sonicated and used (as preparation e under "Results"). An aliquot of the lipid body fraction was subjected to SDS-PAGE. After electrophoresis, the gel containing the oleosin was cut, and the protein in the gel was eluted into a solution containing 0.125 M Tris-HCl, pH 8.0, 1 mM EDTA, and 0.1% SDS. This suspension was used (as preparation f under "Results").
In Vitro Binding Test-The oleosin preparations were assayed for their protein contents using the Bradford method (28). The amounts of 18-kDa oleosin in the preparations were estimated by the amount of this oleosin in comparison with those of other proteins of the preparations in an SDS-PAGE gel. Each oleosin preparation was incubated with the homogenate (1:50 v/v) of nontransformed yeast strain (prepared as described under "Subcellular Fractionation") by shaking in a horizontal shaker at 200 rpm for 30 min in a cold room. The mixture was then subjected to subcellular fractionation (described under "Subcellular Fractionation"). In each mixture, the proportion of 18-kDa oleosin to homogenate was similar to that in the homogenate of yeast strain transformed with pQC6-ole (to be described in Fig. 5), except in oleosin preparation b, in which the oleosin:homogenate ratio was five times higher.
Determination of Lipids-Yeast cells were collected by centrifugation from the culture at different stages of growth. The volume of the pelleted cells was considered as 1 volume. The sample was mixed with 5 volumes of TS buffer containing 2 g/l lyticase and incubated at 30°C for 2 h. The resulting spheroplasts were harvested and lysed by osmotic breakage as described in the preceding section. The lysate was extracted immediately with an equal volume of diethyl ether. The ether fraction was collected, and the ether was evaporated under a stream of nitrogen gas. The acyl esters in the residual lipids were quantitated (29).
Isolated yeast lipid bodies were extracted with diethyl ether three times. The acyl esters of the extracted lipids were quantitated similarly (29). The lipids were resolved by TLC, using a plate coated with silica gel 60A (Whatman, Maidstone, United Kingdom). The plate was developed in hexane/diethyl ether/acetic acid (80:20:2, v/v/v), dried, and allowed to react with iodine for color development.
Immunocytochemistry-Yeast strains nontransformed and transformed with pQC5-ole and pQC6-ole were grown to an A 600 nm of 1.9. The culture was centrifuged at 3,000 ϫ g for 5 min to pellet the cells. The pelleted cells were rinsed twice with water and resuspended in half-strength Karnovsky's solution at 4°C (30). After 0.5 h and mild agitation, the solution was replaced with half-strength Karnovsky's solution, and the mixture was incubated for 12 h at 4°C. The cells were washed twice with cold, 50 mM NaP buffer (pH 6.8). The preparation was rapidly dehydrated (in 1 h) in a series of ethanol solutions to 95% ethanol. The cells were infiltrated with L.R. White resin over 12 h and allowed to polymerize at 60°C for 30 h.
Grids were poststained for 15 min in 1% aqueous uranyl acetate and for 2 min in Reynold's lead citrate. Photographs were taken using Kodak 4489 Electron Microscope film (Eastman Kodak Co.) on a Philips 400 transmission electron microscope at 100 kV.

Maize Oleosin Gene Was Expressed in Transformed Yeast-
The coding region of a maize seed oleosin gene was incorporated into two yeast plasmids in which its expression was under the regulation of GAL1 promoter (Fig. 1). A low copy number centromeric plasmid (pQC5-ole) and a high copy number 2-m plasmid (pQC6-ole) were constructed.
Untransformed yeast strain and yeast strains transformed with the above two plasmids and with the same plasmids but harboring no maize oleosin gene were grown in a medium containing galactose. Total RNAs were extracted from these yeast strains at the mid-log phase of growth and subjected to Northern blot hybridization using a 32 P-labeled maize oleosin cDNA probe (Fig. 2). Oleosin mRNA, of about 0.8 kilobase pairs, was present in yeast strains transformed with pQC5-ole and pQC6-ole and was substantially more abundant in the latter strain. Yeast strain that was not transformed or transformed with pQC5 or pQC6 (plasmid without oleosin gene) did not contain oleosin mRNA. When yeast strain transformed with pQC6-ole was grown in glucose instead of galactose, it did not contain oleosin mRNA. These findings indicate that the oleosin gene in yeast strain transformed with pQC6-ole was properly expressed under the control of the GAL1 promoter.
The crude extracts of the above yeast strains were analyzed for their contents of oleosin protein by SDS-PAGE and immunoblotting, using antibodies against the maize oleosin (Fig. 3). The preimmune IgY and antibody-containing IgY did not recognize any protein of the yeast strains not transformed or transformed with control plasmids (pQC5 or pQC6). The antibodies recognized a protein of 19 kDa in yeast strains transformed with pQC5-ole and pQC6-ole, but they did not recognize any other protein on the blot. This 19-kDa protein was not recognized by the preimmune IgY. Although the 19-kDa protein in yeast strain containing pQC6-ole was easily detectable, its presence in yeast strain containing pQC5-ole was barely observable by the eye but not observable after photography (Fig. 3). This 19-kDa protein should be the oleosin derived from the maize gene in the plasmids. Apparently, both oleosin mRNA and protein were stable in yeast strains transformed with pQC5-ole and pQC6-ole. FIG. 1. Construction of plasmids pQC5-ole and pQC6-ole, which contain the coding region of a maize oleosin gene. The coding region of a maize 18-kDa oleosin was cut by EcoRI from pL2TO2 and inserted into the EcoRI site of pQC5 or pQC6 (19). The final plasmids, pQC5-ole and pQC6-ole, contained the oleosin gene in the correct orientation between the GAl1 promoter and the cyc1 transcription termination sequence (cyc1 TTS).
Oleosin present in the two transformed yeast strains had an M r of about 19,000, and no apparent breakdown products of oleosin of lower M r were detected by immunoblotting (Fig. 3). In the preparation of crude extracts (shown in Fig. 3), the yeast cells were heated to 100°C to prevent proteolysis and then homogenized. Breakdown products of oleosin were observed when the cells were treated with a commercial lyticase preparation and lysed by osmotic shock (to be described). Oleosin in transformed strains (M r about 19,000) was larger than the native oleosin in maize (M r about 18,000) (Fig. 3). The M r of the maize oleosin deduced from the gene coding sequence was 18,615 (21). In maize, as well as in other plant species, oleosin synthesized in vivo does not undergo a co-or posttranslational processing that leads to an appreciable modification of the M r . Nevertheless, oleosins isolated from different plant species all had their N termini blocked, as revealed in several reported failures of direct sequencing of the N termini (1). These findings show that in vivo modification of the protein in plant seeds did occur but did not lead to an appreciable change in M r . Apparently, processing of newly synthesized oleosin mRNA or protein in transformed yeast is different from that in plants.
The yeast strain transformed with pQC6-ole produced a higher amount of oleosin, and we used this yeast strain to explore biochemically the expression of the oleosin gene and the subcellular location of the oleosin protein.
Oleosin Accumulated Concomitantly with Lipids in Transformed Yeast during Growth-Yeast accumulates lipids (TAGs and steroid esters) as food reserves during growth. In the current study, yeast strains nontransformed and transformed with pQC6-ole or pQC6 grew in galactose and accumulated lipids by an indistinguishable pattern (data not shown). The accumulation of lipids in these strains lagged behind the increase in cell culture density (Fig. 4); this delay is expected for the accumulation of a food reserve. In the yeast strain transformed with pQC6-ole, the amount of oleosin per culture volume increased with time and appeared to follow that of the lipids (Fig. 4); the low sensitivity of the immunoblot assays prevents us from making a definite statement. Supporting evidence comes from the finding that the amount of oleosin per equal amounts of total cell proteins also increased with time ( Fig. 4). One possibility is that oleosin was produced constitutively but did not accumulate (i.e. was degraded) in the absence of lipids. Further studies are required to examine this possibility. In all of the above yeast strains, cell growth was faster in a medium containing glucose than in one containing galactose (data not shown).

Subcellular Fractionation Revealed That Oleosin in Transformed Yeast Was Present Mostly in the Lipid Bodies-We
explored the subcellular location of oleosin in the yeast strain transformed with pQC6-ole by fractionating the cell homogenate into various subcellular fractions and analyzing their oleosin contents. Cells at the late log phase (45 h, see Fig. 4) were harvested by low speed centrifugation. They were lysed by treatments with a lyticase preparation and followed by an osmotic shock. After cell breakage, the homogenate was centrifuged to yield fractions of 100,000 ϫ g supernatant, 100,000 ϫ g pellet, and floated lipid bodies. The crude homogenate, the culture medium (containing excreted proteins), and the various subcellular fractions were subjected to SDS-PAGE and immu-  (26). Antibodies against the 18-kDa oleosin (used in this report) recognized the 17-kDa oleosin almost to the same extent but barely recognized the 16-kDa oleosin. Because there was substantially less oleosin in the yeast crude extracts than in the maize oil bodies, the gel for immunoassay (lower panel), in comparison with the gel for protein staining (upper panel), was loaded with the same amounts of yeast crude extracts but 10 times less of the maize oil bodies. Positions of the three oleosins (shown with their molecular mass), as well as position of the oleosin synthesized in yeast (arrow), are shown on the right. noblotting with antibodies against the maize oleosin (Fig. 5). About 80% of the oleosin was recovered in the lipid body fraction, which contained substantially less than 1% of the proteins in the homogenate. The remaining oleosin was present in the 100,000 ϫ g supernatant, and no oleosin was detected in the 100,000 ϫ g pellet. Oleosin was absent in the medium, and thus the protein was not secreted. In this experiment, the oleosin was present as an intact form (M r 19,000) and a degraded form (M r 18,000). The degraded form apparently was produced not in vivo but in vitro during cell lysis (by ingredients in the lyticase preparation or internal proteases), because it was absent when we boiled the cells before cell breakage (Figs. 3 and  4) or lysed the cells rapidly using mechanical force (following paragraph).
We also studied the subcellular location of oleosin in the transformed yeast cells using an alternative procedure of cell breakage in which the oleosin was maintained intact. After being harvested by a low speed centrifugation, the cells were homogenized with glass beads using a mortar and pestle. The homogenate was subjected to a similar procedure of subcellular fractionation by centrifugation. As much as 80 -90% of the oleosin could be recovered in the lipid body fraction, and the remaining oleosin was present in the supernatant and the pellet (Fig. 5). In this procedure of cell breakage, the percentage of oleosin recovered in the lipid body fraction varied, from 50 to 90%, apparently depending on the severity of the homog-enization with the pestle. The advantage of using this procedure was that it generated no degraded oleosin.
The results obtained from subcellular fractionation using the two different procedures of cell breakage clearly show that oleosin in the cells of the yeast strain transformed with pQC6ole was localized in the lipid bodies. Collaborative evidence comes from immunocytochemistry results (to be described).
Oleosin Did Not Alter the Lipids of the Transformed Yeast Lipid Bodies but Replaced Some of the Native Proteins Associated with the Organelles-Lipid bodies isolated from yeast strains nontransformed and transformed with pQC6-ole were subjected to lipid extraction, and their lipid constituents were separated by TLC (Fig. 6). Two major neutral lipid constituents were resolved; they were tentatively identified to be steroid esters and TAGs, as reported earlier (5,6). The lipid composition of the lipid bodies from both yeast strains as resolved by TLC were indistinguishable (Fig. 6). This finding, together with the identical patterns of cell growth and lipid accumulation during culturing of the two yeast strains, shows that the presence of oleosin in the yeast cell did not affect appreciably the synthesis and accumulation of lipids in the lipid bodies. Nevertheless, oleosin did affect the native yeast proteins in the lipid bodies. The proteins in the lipid body fractions from the two yeast strains were resolved by SDS-PAGE (Fig. 6). Equal amounts (acyl ester bonds) of lipid bodies from the two yeast strains were loaded onto the gel for a direct comparison. The

FIG. 5. SDS-PAGE (upper panel) and immunoblotting (lower panel) of proteins of the crude homogenate and subcellular fractions of cells of the yeast strain transformed with pQC6-ole.
Yeast cells at the late log phase (45 h, see Fig. 4) were lysed by either lyticase and osmotic shock (left panels) or grinding (right panels), and the lysates were subjected to centrifugation. The 100,000 ϫ g supernatant, 100,000 ϫ g pellet, lipid body fraction, and the leftover culture medium were used, with loading samples representing amounts derived from an equal quantity of the crude homogenate. The crude homogenate was loaded with an unspecified amount, which was half of that used to produce the various subcellular fractions and the leftover culture medium. Immunoblotting was performed with antibodies against maize 18-kDa oleosin. Only the portion of the immunoblot containing the oleosin is shown. Isolated maize oil bodies were used as a control (see legend of Fig. 3 for explanation). The samples obtained by lysing the cells with mechanical force did not contain degraded oleosin, whereas those obtained by treating the cells with lyticase and osmotic shock possessed intact oleosin (19 kDa) as well as degraded oleosin (18 kDa). Positions of the three oleosins (shown with their molecular mass), as well as position of the intact oleosin synthesized in yeast (arrow), are shown on the right. proteins from these two samples resolved in the gel would represent those from the same number of lipid bodies, since the lipid bodies in both yeast strains appeared to be of a similar size (next section). The organelles from the nontransformed yeast strain contained several proteins, of M r in the range of 30,000 -75,000 (Fig. 6). This pattern of proteins associated with the lipid body fraction of yeast was similar to that reported earlier (6). In the transformed yeast strain, although the native proteins were still present, their amounts were reduced to about half of those in the nontransformed yeast strain. The loss of the native proteins apparently was compensated by the gain of a similar amount of oleosin, and the protein-to-lipid ratios in the lipid bodies from the two yeast strains remained about the same. Thus, replacement of half of the native proteins with an equal amount of oleosin did not affect appreciably the integrity and synthesis of the organelles.

Immunocytochemistry Showed That Oleosin Was Present Mostly on the Periphery of the Lipid Bodies in the Transformed
Yeast-Immunocytochemistry was performed to locate oleosin in the yeast strain transformed with pQC5-ole and pQC6-ole, and the nontransformed yeast strain was used as a control. The lipid bodies in the cells of these yeast strains were electron transparent, of diameters about 0.2-0.5 m (Fig. 7). They were usually present in clusters and apparently did not coalesce. There were interfacial materials on the periphery of the lipid bodies. In the immunodetection of oleosin in the yeast strains, the nontransformed strain showed no immunogold particles in the cytoplasm including the lipid bodies, whereas yeast strains transformed with pQC5-ole and pQC6-ole had numerous immunogold particles, which were concentrated on the lipid bodies. There were more gold particles in the yeast strain transformed with pQC6-ole (high copy number plasmid) than in that with pQC5-ole (low copy number). Enlarged views of the cells showed that most of the immunogold particles were restricted to the periphery of the lipid bodies (Fig. 8). Under the electron microscopy, the vicinity of the lipid body surface was often sectioned, and the tangential areas were exposed and heavily labeled. The immunocytochemical observation complements the biochemical findings that most oleosin was associated with the lipid bodies and, specifically, was present on the periphery of the organelles.
By SDS-PAGE and immunoblotting (Fig. 3), we barely detected the oleosin in yeast strain transformed with the low copy number plasmid pQC5-ole. By immunocytochemistry, we were able to observe in this yeast strain numerous immunogold particles that were concentrated on the lipid bodies (Fig. 8). Apparently, immunocytochemistry was a more sensitive method of detection.
Oleosins Did Not Bind to Yeast Lipid Bodies in Vitro-Whether oleosin synthesized by transformed yeast strains bound to the lipid bodies in vivo merely because of its hydrophobicity was investigated. Maize oleosin was incubated with the homogenate of nontransformed yeast strain, and the mixture was then subjected to subcellular fractionation. Six different preparations of oleosin were used. (a) Oleosins in three isoforms, of 16, 17, and 18-kDa, together with the native PL, were obtained from isolated maize oil bodies and sonicated into a suspension; (b) same as a except the amounts of all components were five times higher; (c) same as a except the PL were removed by chloroform/methanol; (d) same as c except the oleosins were not sonicated but resuspended in 0.1% SDS; (e) yeast-synthesized oleosin, together with other proteins and PL, was obtained from lipid bodies isolated from pQC6-ole transformed yeast strain and sonicated into a suspension; and (f ) yeast-synthesized oleosin was obtained from lipid bodies isolated from pQC6-ole transformed yeast strain by SDS-PAGE and resuspended in 0.1% SDS.
In each of the mixtures, the proportion of 18-kDa oleosin to homogenate was similar to that in the homogenate of yeast strain transformed with pQC6-ole (see Fig. 5), except for oleo-sin preparation b, in which the proportion of oleosin to homogenate was 5 times higher. Fig. 9 shows results of the subcellular fractionation. In each of the six mixtures, the in vitro applied oleosin was not found in the isolated lipid bodies. Instead, it was present in the pellet fraction and, in most mixtures, also in the supernatants. Oleosin in the pellet should represent molecules that had aggregated, presumably because of their hydrophobicity. Oleosin in the supernatant should be in a non-aggregated form, of individual molecules alone or associated with PL or other amphipathic molecules, such that it was not pelleted. Oleosin in this form was present in the mixture during its incubation with the homogenate and did not bind to the lipid bodies in vitro. This observation reiterates that the oleosin present on the lipid bodies in transformed yeast represents the consequence of a specific in vivo targeting event. DISCUSSION Yeast transformed with a plant oleosin gene synthesizes oleosin, which remains stable in the cell. In addition, the oleosin is correctly targeted to the lipid bodies. The targeting ap- Oleosin prepared in different forms was incubated with the homogenate, and the mixture was subjected to subcellular fractionation. The 100,000 ϫ g supernatant, 100,000 ϫ g pellet, and lipid body fraction were subjected to SDS-PAGE. The gels were stained with Coomassie Blue (upper panel) or treated for immunoblotting using antibodies against maize 18-kDa oleosin (lower panel). The loading amounts of the various fractions represented those derived from an equal quantity of the homogenate. Only the portion of the immunoblot containing the oleosin is shown. Isolated maize oil bodies were used as a control (see legend of Fig. 3 for explanation). Positions of the three oleosins (shown with their molecular mass), as well as position of the intact oleosin synthesized in yeast (arrow), are shown on the right. The oleosin preparations included: a, oleosins in three isoforms, of 16, 17 and 18 kDa, together with the native PL, were obtained from isolated maize oil bodies and sonicated into a suspension; b, same as a except the amounts of all components were 5 times higher; c, same as a except the PL were removed by chloroform/methanol; d same as c except the oleosins were not sonicated but resuspended in 0.1% SDS; e, yeast-synthesized oleosin, together with other proteins and PL, was obtained from lipid bodies isolated from pQC6-ole transformed yeast strain and sonicated into a suspension; and f, yeast-synthesized oleosin was obtained from lipid bodies isolated from pQC6-ole-transformed yeast strain by SDS-PAGE and resuspended in 0.1% SDS. pears to be a specific in vivo event. In a transformed yeast strain that contains a high amount of oleosin, we have demonstrated by biochemical means that the oleosin is present mostly in the lipid bodies. This localization does not appear to be a fortuitous association of the abundant foreign hydrophobic protein with the lipid bodies because we have also shown by immunocytochemistry that a transformed yeast strain containing a minimal amount of oleosin has the oleosin present exclusively in the lipid bodies. Oleosin is present only on the amphipathic surface of the lipid bodies rather than on other amphipathic membranes. Also, oleosin prepared in different forms, including solitary molecules and those associated with PL and SDS, fails to bind to the yeast lipid bodies in vitro. The targeting signal in oleosin is unknown, although apparently it is not the C-terminal ␣-helix (31,32). What is known is that oleosin is synthesized on bound polyribosomes without appreciable co-or posttranslational processing (11). Whether the targeting signal resides on the N terminus or the long central anti-parallel ␤-stranded hydrophobic domain, or both, and what role the ER plays in directing the oleosin to the budding oil body remain to be elucidated. It is possible that the targeting signal and the stability of the oleosin on the lipid body together contribute to its stable association with the organelles.
It is tedious to study the targeting signal in oleosin by transforming plants with numerous modified oleosin genes. The current study offers the opportunity of using yeast as a heterologous system to analyze in detail the targeting signal in oleosin via in vitro mutagenesis of an oleosin gene. Plants and yeast share many intracellular targeting signals in proteins, although yeast does not recognize the targeting signal in plant storage protein for the vacuoles (19,33). One should analyze extensively the targeting signals in the oleosins for the yeast lipid bodies and then test the validity of selected results with the tedious plant transformation system.
In plants, oleosin exists in two isoforms, which apparently occur as a heterodimer or heteromultimer on the surface of the oil bodies (26,27). The current study shows that one oleosin isoform can be present alone on the organelle surface (i.e. the maize isoform 18-kDa oleosin used in the current study is present alone without its counterpart isoform 16-kDa oleosin). Thus, dimerization does not appear to play a role in targeting the oleosin to the yeast lipid bodies.
The plant oil bodies and the yeast lipid bodies are similar in their apparent structure and function. Whether they possess some differences remains to be explored. A plant oil body has a matrix of TAGs surrounded by a layer of PL embedded with unique and abundant oleosins (1). This organization allows the stable association of all the molecules involved, such that numerous oil bodies of small sizes can be maintained in the cytosol. On the basis of this concept, we visualize a yeast lipid body to have the abundant hydrophobic TAGs and steroid esters located in the matrix and the amphipathic PL and proteins (of minimal but sufficient quantities (5,6)) at the periphery. Unlike oleosins in plant oil bodies, the proteins in yeast lipid bodies are of numerous molecular species (Fig. 6); whether these proteins serve as the structural proteins is unknown. Isolated yeast lipid body fractions contained diacylglycerol acyltransferase (34) and sterol ⌬ 24 methyltransferase (6), which catalyze the last steps of TAG and steroid ester synthesis, respectively. The activities of these two enzymes in the lipid body fractions were high in terms of specific activities (on per mg of protein basis) but low in terms of percentage of total cellular activities (because there were minimal amounts of proteins in the lipid body fractions). Whether these enzymes are authentic proteins of the lipid bodies needs to be elucidated. In plants, diacylglycerol acyltransferase of a high specific ac-tivity but of a low percentage of total cellular activity was reported to be present in an isolated oil body fraction; it was subsequently shown to be a contaminant from the ER (1, 3). Thus, we do not know whether in yeast TAGs and steroid esters are synthesized in the ER, as has been shown for the plant TAGs, or directly on the surface of existing lipid bodies, as has been suggested on the basis of the high enzyme specific activities. During utilization of the oil reserves in plants, lipase is newly synthesized and binds to the oil bodies (35). In yeast, steroid ester hydrolase activity was detected in many subcellular fractions (36). Whether the activity belonged to one or more enzymes and which of these enzymes, if any, mobilized the storage steroid esters remain to be studied. Plant seeds contain several acyl hydrolases, which are known not to be responsible for the mobilization of the storage oils (1). Even less is known about the catabolism of TAGs in yeast lipid bodies. The abundant information on the structure, function, and ontogeny of plant oil bodies, both in concept and study techniques, should be utilized for detailed studies of yeast lipid bodies.