Disruption of the Saccharomyces cerevisiae Homologue to the Murine Fatty Acid Transport Protein Impairs Uptake and Growth on Long-chain Fatty Acids*

The yeast Saccharomyces cerevisiae is able to utilize exogenous fatty acids for a variety of cellular processes including (cid:98) -oxidation, phospholipid biosynthesis, and protein modification. The molecular mechanisms that govern the uptake of these compounds in S. cerevisiae have not been described. We report the characterization of FAT1 , a gene that encodes a putative membrane-bound long-chain fatty acid transport protein (Fat1p). Fat1p contains 623 amino acid residues that are 33% identical and 54% with similar chemical properties as compared with the fatty acid transport protein FATP described in 3T3-L1 adipocytes (Schaffer and Lodish (1994) Cell 79, 427–436), suggesting a similar function. Disruption of FAT1 results in 1) an impaired growth in YPD medium containing 25 (cid:109) M cerulenin and 500 (cid:109) M fatty acid (myristate (C 14:0 ), palmitate (C 16:0 ), or oleate (C 18:1 )); 2) a marked decrease in the uptake of the fluo- rescent long-chain fatty acid analogue boron dipyrromethene difluoride dodecanoic acid (BODIPY-3823); 3) a reduced rate of exogenous oleate incorporation

Exogenous long-chain fatty acids represent an important class of hydrophobic compounds that serve as substrates for lipid biosynthesis, protein modification, and ␤-oxidation. While the mechanism that facilitates the uptake of these compounds into eukaryotic cells is not completely understood, information gleaned over the past 15 years is consistent with a facilitated, protein-mediated process. In higher eukaryotic cells, three general classes of membrane-bound fatty acid transport proteins have been described. The first, identified by Abumrad and co-workers (1-4) is fatty acid translocase (FAT). 1 This protein represents the adipocyte homologue of the glycoprotein CD36 and appears to act in concert with intracellular fatty acidbinding proteins to mediate fatty acid uptake (5,6). The second, identified by Berk and Stremmel is a membrane-bound fatty acid-binding protein (FABPpm) that is identical to mitochondrial aspartate aminotransferase (7)(8)(9)(10)(11)(12). Trotter et al. (13) have demonstrated that pretreatment with anti-FABPpm sera does not block the uptake of long-chain fatty acids in human intestinal cells that normally express FABPpm. These data raise the question of whether the mitochondrial aspartate aminotransferase actually represents a membrane-bound fatty acid transport protein. The third, described by Schaffer and Lodish (14) is FATP (fatty acid transport protein). FATP is differentially expressed in 3T3-L1 adipocytes and is predicted to function in the transport of exogenous long-chain fatty acids. The expression of FATP is apparently negatively regulated by insulin at the level of transcription in cultured adipocytes (15). Studies in yeast and in bacteria also demonstrate that longchain fatty acid transport is a facilitated process (16 -19). While the proteins that mediate uptake in yeast have not been described, considerable evidence has been accumulated describing the fatty acid transporter FadL in Escherichia coli (20 -23). Exogenous long-chain fatty acids traverse the cell envelope by a high affinity transport process that minimally requires the outer membrane-bound fatty acid transport protein FadL and the inner membrane-associated acyl-CoA synthetase. The fatty acid transporter FadL, while mechanistically similar, is quite distinct on the basis of amino acid similarities, to the eukaryotic fatty acid transporters described above.
The notion of a membrane-bound fatty acid transporter is not universally accepted. One school of thought is that fatty acids are merely transported by diffusion across the membrane. In support of this, it has been demonstrated that when fatty acids are in the nonionized form, uptake into artificial membrane vesicles and 3T3-L1 cells occurs by simple diffusion, obviating the need for a membrane-bound fatty acid transporter (24 -26). Therefore, it appears that in 3T3-L1 cells, long-chain fatty acid transport can occur by both diffusional and protein-mediated processes. The need for a protein-mediated process must lie in the requirement to regulate the entry of these compounds. A long-chain fatty acid transporter may be required for cells that specifically use these compounds for the production of metabolic energy or for the synthesis of triglycerides. These transporters are also required by microbes such as E. coli that have a hydrophilic cell envelope that is refractory to hydrophobic compounds.
In the yeast S. cerevisiae, the gene products involved in uptake of exogenous long-chain fatty acids are unknown. Kohlwein and Paltauf (16) demonstrated that fatty acid uptake in Saccharomyces uvarum and Saccharomycopsis lipolytica occurs via a saturable process. The same observations were made in Candida tropicalis supporting the hypothesis that fatty acid transport is a facilitated process (17). Knoll et al. (19) have shown the uptake of exogenous long-chain fatty acids in S. cerevisiae is saturable and that uptake and activation to CoA thioesters are separable. On the basis of these studies, it is reasonable to predict that a membrane-bound long-chain fatty acid transport protein exists in yeast. S. cerevisiae requires supplementation of oleic acid (C 18:1 ) to remain viable under anaerobic growth conditions due to the suppression of fatty acid desaturation (27). Supplementation of exogenous longchain fatty acids is also required for cell viability when fatty acid synthase is blocked by the antibiotic cerulenin (28,29). Under both conditions, transport and activation of exogenous long-chain fatty acids are required to overcome growth inhibition. The acyl-CoA synthetases encoded by FAA1 and FAA4 (Faa1p and Faa4p, respectively) have been shown to be responsible for activation of imported fatty acids in S. cerevisiae (30). The overexpression of either Faa1p or Faa4p in faa1⌬ faa4⌬ cells does not increase the levels of fatty acid uptake. The finding that long-chain fatty acid uptake in yeast is saturable and occurs in the absence of Faap activity supports the proposal for a specific plasma membrane-bound fatty acid transporter.
We report the characterization of FAT1 in S. cerevisiae that encodes a 623-amino acid residue protein that shares 33% amino acid identity and 54% amino acids with similar chemical properties to FATP described in 3T3-L1 adipocytes (14). Disruption of FAT1 results in 1) an impaired growth in YPD medium containing 25 M cerulenin and 500 M fatty acid (myristate (C 14:0 ), palmitate (C 16:0 ), or oleate (C 18:1 )); 2) a decrease in the uptake of the fluorescent long-chain fatty acid analogue BODIPY-3823; 3) a reduced rate of exogenous oleate incorporation into phospholipids; and 4) a 2-3-fold decrease in the rates of oleate uptake. These data support the hypothesis that Fat1p is a long-chain fatty acid transport protein.

EXPERIMENTAL PROCEDURES
Materials-Yeast extract, yeast peptone, agar, and yeast nitrogen base were obtained from Difco. Fatty acids and cerulenin were obtained from Sigma. 3 H-Labeled fatty acids were from DuPont NEN. BODIPY-3823 was purchased from Molecular Probes. Enzymes required for all DNA manipulations were obtained from Epicenter, Promega, New England Biolabs, U.S. Biochemical Corp., or Boehringer Mannheim. The oligonucleotides for DNA amplification were synthesized on a Gene Assembler Plus (Pharmacia Biotech Inc.). All other chemicals were obtained from standard suppliers and were of reagent grade.
Data Base Analysis and Alignments-Homology searches were performed using Biological Sequence Comparative Analysis Node (Bio-SCAN) at the University of North Carolina. The nonredundant data base Genpept (Release 93) was used. Alignments of the primary structures of FATP homologies were generated with BESTFIT (Genetics Computer Group) (31).
Disruption of FAT1-Yeast genomic DNA was purified from strain W303a as detailed in Kaiser et al. (32). For disruption, the FAT1 gene was amplified by thermocycling from genomic DNA using sets of specific primers; 5Ј-ACCCGGATCCAGAAATCCTGGGTTATCT-3Ј (upstream) and 5Ј-CAACGGATCCTCTACTTCAGTAGTGGAAAC-3Ј (downstream), both containing BamHI sites (underlined). The HIS3 gene was amplified from pHB3 (obtained from David Nelson, University of Tennessee, Memphis) using the upstream primer 5Ј-AG-CGGAATTCCCTTTTAAACCACGACGCTTTGTC-3Ј and the downstream primer 5Ј-TACGGAATTCCACTTGCCACCTATCACCACAA-3Ј, both containing EcoRI sites (underlined). The 1549-bp fragment containing the coding sequence of FAT1 was cloned into the BamHI-site of pACYC177 to generate pNJF1. The 480-bp EcoRI fragment within FAT1 was replaced by the 1346-bp EcoRI-fragment containing the HIS3 gene to generate pNJF2. The strain W303a was rendered competent using lithium acetate using standard procedures and transformed with linearized pNJF2 according to Kaiser et al. (32). Clones were selected on minimal plates lacking histidine supplementation. His ϩ isolates were selected and colony-purified on minimal places without histidine. Several isolates were obtained. The disruption of the FAT1 gene with HIS3 was confirmed using DNA amplification of genomic DNA. One such isolate shown to contain a disruption in FAT1, designated W303a-fat1⌬-1 (fat1⌬) was selected for all further studies.
The expression of FAT1 was evaluated in strain W303a grown in YPD containing 500 M oleate. Total RNA was purified using the RNeasy kit as recommended by the supplier (Qiagen). Reverse transcriptase amplification was performed on 1 g of total RNA using oligo(dT) primers and Moloney murine leukemia virus reverse transcriptase as described by the manufacturer (Boehringer-Mannheim). Controls received no Moloney murine leukemia virus reverse transcriptase. First-strand cDNA was amplified using primers derived from the sequences corresponding to the C termini of Fat1p or histone H4 (used as an internal control).
Fatty Acid Transport-Fatty acid transport was performed essentially as described by Knoll et al. (19). Cells (W303a and W303a-fat1⌬-1) were grown in SMM at 24 or 30°C to midlog phase (A 600 ϭ 1.0), collected by centrifugation, washed once in phosphate-buffered saline (PBS), and resuspended in 1 ⁄10 of the original volume in PBS. 200 l of cells (1 ϫ 10 8 cells) were preincubated for 10 min at 24 or 30°C in SMM, and the assay was initiated by the addition of [9,10-3 H]myristate, [9 -10-3 H]palmitate, or [9 -10-3 H]oleate at the concentrations indicated. All fatty acids were prepared as ethanolic stocks. At the defined time points, the reactions were terminated by the addition of 10 ml of ice-cold PBS. The cells were immediately filtered through a Whatman Gf/Bfilter, washed three times with ice-cold PBS, and air-dried. The amount of cell-associated radioactivity was determined by scintillation counting. Background counts were subtracted from the experimental samples by evaluating the amount of radioactivity on control filters with no cells. The final data were expressed in nmol of cell-associated fatty acid/min/1 ϫ 10 8 cells. All transport data were analyzed using En-zymeKinetics software (version 1.0.4; Trinity Software). The data presented represents the mean from at least three independent experiments.
Metabolic Labeling of Cellular Lipids-W303a and W303a-fat1⌬-1 cells were grown in SMM at 30°C to midlog phase (A 600 ϭ 1.0). One ml of cells were transferred to a tube containing 50 M [9,10-3 H]oleate (specific activity of 1 Ci/mmol, 50 Ci/ml). At the time points indicated, 1 volume of 10% ice-cold trichloroacetic acid was added to stop yeast fatty acid metabolism. The cells were immediately collected by centrifugation (15,000 ϫ g) and resuspended in 100 mM Tris-HCl, pH 7.5. After neutralization with 1 M KOH, cells were washed four times in 100 mM Tris-HCl (pH 7.5) containing 100 M fatty acid-free bovine serum albumin. Lipids were extracted according to Bligh and Dyer (33). Samples were analyzed using high performance thin layer chromatography as described by Knoll et al. (19).
Analysis of Acyl-CoA Synthetase Activities-Yeast cells (FAT1 and fat1⌬) as noted for transport, harvested by centrifugation, washed twice with yeast nitrogen base, resuspended to a density of 1.2 ϫ 10 9 cells/ml in 10 mM Tris-HCl, pH 7.5, and lysed by three cycles of sonication at 0°C. Acyl-CoA synthetase activities were determined in sonicated cell extracts as described by Kameda  H]myristate, 0.5 mM coenzyme A, and cell extract in a total volume of 0.5 ml. The reactions were initiated with the addition of coenzyme A, incubated at 35°C for 10 min, and terminated by the addition of 2.5 ml of isopropyl alcohol:n-heptane:1MH 2 SO 4 (40:10:1). The radioactive fatty acid was removed by organic extraction using n-heptane. Acyl-CoA formed during the reaction remained in the aqueous fraction and was quantified by scintillation counting. Protein concentrations in the enzyme extracts and purified enzyme samples were determined using the Bradford assay and bovine serum albumin as a standard (35). The values presented represent the average from at least three independent experiments. All experiments were analyzed using analysis variance (StatView; Abacus Concepts, Inc.) BODIPY-Labeled Fatty Acid Uptake-Cells were grown in SMM at 24°C and prepared as described above for fatty acid transport. The fluorescent long-chain fatty acid analogue BODIPY-3823 was added from a 5 mM ethanolic stock solution to 1 ml of cells to a final concentration of 50 M, and cells were incubated for 60 s. Cells were centrifuged and washed three times with 50 M fatty acid-free bovine serum albumin in PBS. Finally, cells were resuspended in 1 ml of PBS, and the labeled cells were analyzed in a Olympus BH2 microscope using a Zeiss 63 ϫ oil planapo objective, and in an MRC 1024 laser Sharp confocal microscope (Bio-Rad) on an Olympus BX50 with a ϫ 60 objective.

RESULTS
Identification of FAT1-While the processes that mediate fatty acid uptake in yeast are not defined, it was reasonable to predict that a homologue of one or more of the three putative fatty acid transport proteins from higher eukaryotic cells may be required. We compared the sequence of FAT, mitochondrial aspartate aminotransferase (plasma membrane FABPpm), and FATP to open reading frames within the Saccharomyces cerevisiae data base. We were unable to find open reading frames within the yeast data base that had significant homology to FAT. While seven different reading frames on five different chromosomes were found to have homology to mitochondrial aspartate aminotransferase, it was difficult to assess which if any of these open reading frames represented the FABPpm homologue in yeast. Of these comparisons, the FATP identified in 3T3-L1 adipocytes appeared to be the most promising candidate. We identified an open reading frame on yeast chromosome II encoding a 623-amino acid protein with 33% sequence identity and 54% similarity to FATP (Fig. 1A). We have designated this open reading frame Fat1p and the structural gene encoding this protein as FAT1.
The sequence similarities and identities exist throughout the lengths of Fat1p and FATP and thus suggest they may have similar structures and functions. Three segments of these two proteins are nearly identical. The first segment, between amino acids 255 and 268 of Fat1p, contains a consensus sequence (LIYSGTTGLPK) common to members of the AMP-binding protein family including the family of acyl-CoA synthetases (36,37). The second and third regions of high sequence similarity include amino acid residues 324 -399 and 491-544, respectively, and are restricted to Fat1p and FATP. Using the algorithms of Kyte and Doolittle (38), we compared the hydropathy profiles of Fat1p and FATP (Fig. 1B). These analyses demonstrated that both proteins have comparable profiles and in particular predicted that Fat1p, like FATP, contains at least four potential membrane-spanning segments. Fat1p also has four potential N-linked glycosylation sites. One of these glycosylation sites (at amino acid residue 534) is identical to that predicted for FATP. On the basis of these comparisons, we propose that Fat1p represents the yeast homologue of the murine long-chain fatty acid transport protein FATP.
Examination of the DNA flanking the Fat1p coding sequence identified two potential TATA boxes. The first is found 301 bases upstream from the presumptive translational initiation codon, and the second is 119 bases upstream. Of the two, the first (TATATAA) is predicted to be a very strong binding site for TFIID (39). There were also two potential polyadenylation sites, AATAAAN 14/22 CA (14 -22 nucleotides without consensus features), found 286 and 294 bases 3Ј relative to the presumptive translational termination codon. Examination of the 5Ј upstream region indicated that expression of the gene may be controlled by at least three regulators. There are three potential CCAAT elements located beginning at Ϫ846, Ϫ882, and Ϫ982 and two potential overcoming glucose repression ele- FIG. 1. A, comparison of the amino acid sequence of Fat1p from S. cerevisiae with FATP from murine adipocytes. The lines indicate amino acid identity, while the double and single dots indicate high or moderate conservation in amino acids, respectively. The identity between these two proteins is 33.4%, while the overall degree of similarity is 53.4%. The potential N-linked glycosylation sites are noted by the asterisks above the Fat1p sequence. B, hydropathy profile of Fat1p and FATP using the algorithm of Kyte and Doolittle and an averaging len of 19 (38). ments (consensus (A/C)(A/G)GAAT) beginning at Ϫ130 and ϩ88 bp from the predicted start of translation (40). Since our prediction is that Fat1p is involved in fatty acid transport and metabolism, we searched for potential regulatory sites identified particularly in genes encoding other yeast proteins required for these functions. There were no matches to the consensus sequence for UAS INO , which is CATGTGAAAT (41). These elements are found in many genes whose product is required for phospholipid and fatty acid biosynthesis. Several genes encoding yeast peroxisomal proteins and the ⌬9-acyl-CoA desaturase are activated or repressed, respectively, after growth in media containing oleate (42). Several laboratories have identified DNA elements important for response to oleate (43)(44)(45). Each contains the minimal consensus sequence CGGN 15/18 CCG. There were no elements of this type identified in FAT1. A second oleate response element was identified in the peroxisomal trifunctional fatty acid oxidizing enzyme gene of C. tropicalis when expressed in S. cerevisiae (40). The element has the consensus sequence YGTTRTT(A/C/G). We identified four regions upstream of FAT1 at Ϫ369, Ϫ520, Ϫ727, and Ϫ835, which conform to this consensus sequence. It is not known at this time whether any of these DNA segments contribute to the expression or regulation of FAT1.
Disruption and Expression of the Gene Encoding Fat1p in S. cerevisiae-On the basis of the sequence similarities noted above, we predicted that Fat1p, like FATP, was involved in the uptake of long-chain fatty acids. To determine the function of Fat1p in this process, we disrupted FAT1 in S. cerevisiae by the replacement of a 480-bp internal EcoRI fragment (within the open reading frame of FAT1) with a 1.346-kilobase pair fragment encoding the yeast HIS3 gene. This construction was crossed onto the chromosome in the yeast strain W303a (his3) as detailed under "Experimental Procedures," and several transformants complementing the his3 mutation were identified. One such isolate, designated W303a-fat1⌬-1, was selected and shown to contain a disruption in FAT1 by DNA amplification using thermocycling. Genomic DNA purified from both the FAT1 and fat1⌬ strains were amplified using oligonucleotides specific to the 3Ј-and 5Ј-ends of the open reading frame within FAT1. The DNA amplification product from the fat1⌬ strain was approximately 0.9 kilobase pair larger than the product obtained from the parental strain, confirming that the FAT1::HIS3 fragment was integrated into the chromosome. Reverse transcription and amplification of total mRNA isolated from the FAT1 strain W303a grown in YPD containing 500 M oleate demonstrated that this gene is expressed (data not shown).
Disruption of FAT1 Impairs Growth on YPD Containing Cerulenin and Oleate-We initially tested whether the fat1⌬ strain was phenotypically distinct from the parental FAT1 strain. FAT1 and fat1⌬ cells were grown to midlog phase and diluted in yeast nitrogen base, and 2 ϫ 10 3 cells were plated on cerulenin-containing media supplemented with oleate, palmitate, or myristate. The cultures were incubated at 24, 30, and 37°C for 72 h. As predicted, both strains were unable to grow on YPD plates containing 25 M cerulenin (YPD/CER) due to the inhibition of fatty acid synthesis. Growth of the FAT1 strain could be rescued at all temperatures by the addition of 500 M myristate, palmitate, or oleate to the YPD/CER plates (YPD/CER/MYR, YPD/CER/PAL, or YPD/CER/OLE, respectively). The fat1⌬ strain was viable on both YPD/CER/MYR and YPD/CER/PAL, although there was an apparent decrease in growth when compared with the wild type. The growth of the fat1⌬ strain on YPD/CER/OLE was reduced dramatically when grown at 24, 30, and 37°C when compared with the wild type.
To evaluate these observations further, growth of the FAT1 and fat1⌬ strains was monitored at 30°C in liquid YPD with and without cerulenin and fatty acid (Fig. 2). These data were in agreement with the phenotypic data noted above. The growth rate of the FAT1 strain while depressed with the addition of cerulenin was able to be rescued by the addition of fatty acid (myristate, palmitate, or oleate). The growth of the fat1⌬ strain was reduced even when supplemented with fatty acid when compared with the isogenic FAT1 strain. This reduction was particularly notable with the oleate supplementation, which paralleled the observations made on YPD agar plates containing oleate and cerulenin. Use of the Fluorescent Fatty Acid BODIPY-3823 to Monitor Fatty Acid Uptake-To visualize in a more direct way that fatty acid uptake was reduced in the fat1⌬ strain of S. cerevisiae, we evaluated the uptake of a fluorophore-labeled long-chain fatty acid analog BODIPY-3823, employing confocal laser scanning microscopy. Pagano et al. (46) demonstrated that the spectral properties of several BODIPY-labeled ceramide analogues are highly dependent upon the concentration of the probe in lipid vesicles as characterized by a shift in the emission maximum from green (515 nm) to red (620 nm) with increasing concentrations. Typical confocal scanning micrographs of FAT1 and fat1⌬ cells labeled with BODIPY-3823 are shown in Fig. 3. When the cells were observed in the green channel only, the wild-type strain appeared highly fluorescent, whereas the disrupted strain showed limited fluorescent labeling (Fig. 3, A and  D, respectively). When the same cells were viewed using the red channel (Fig. 3, B and E), the difference between the FAT1 and fat1⌬ strains was even more dramatic. When labeled cells were viewed in the red and green channels (Ն520 nm), wildtype cells were yellow-orange in color, in contrast to the fat1⌬ cells, which appeared pale yellow-green (Fig. 3, C and F). These results are consistent with the conclusion that a disruption of FAT1 reduces the cell's ability to take up the fluorescent longchain fatty acid BODIPY-3823.
Incorporation of Exogenous Fatty Acids into Phospholipids in the fat1⌬ Strain-To distinguish fatty acid transport from metabolic utilization, we evaluated the distribution of fatty acids when supplied exogenously to the intracellular fatty acid and phospholipid pools. We predicted that overall rate of uptake

FIG. 2. Growth kinetics of FAT1 (wild type; circles) and fat1⌬ (squares) cells in (A) YPD (open circles and squares) or YPD containing cerulenin (YPD/CER; filled circles and squares), YPD/CER containing myristate (B), YPD/CER containing palmitate (C), and YPD/CER containing oleate (D).
and incorporation of exogenous long-chain fatty acids into the phospholipid pool in the fat1⌬ strain would be reduced while the distribution of incorporated long-chain fatty acids among the phospholipid classes would be the same as the wild-type. To test these predictions, we monitored the time-dependent incorporation of exogenous [ 3 H]oleate into cellular lipids using onedimensional high performance TLC as detailed under "Experimental Procedures." The initial uptake and incorporation of [ 3 H]oleate into total lipids was notably reduced in the fat1⌬ strain (Fig. 4). Furthermore, the level of free fatty acid was markedly decreased when compared with the wild-type (Fig.  4A). At the later time points, the differences in incorporation of exogenous oleate were still striking. We noted that while oleate incorporation was reduced in the mutant, the pattern of incorporation into the various classes of phospholipids remained the same, arguing that the enzymatic machinery required for lipid biosynthesis was still intact. These data imply that Fat1p must necessarily operate prior to the incorporation of exogenous oleate into the phospholipid pool. We interpret these data to suggest a defect in the uptake of oleate prior to metabolic utilization.
As noted above, Fat1p shares amino acid sequence similarities with the acyl-CoA synthetases; we therefore evaluated acyl-CoA synthetase profiles in the FAT1 and fat1⌬ strains using oleate, palmitate, and myristate as substrates (Table I). Acyl-CoA synthetase activities using all three fatty acid substrates were comparable, although the fat1⌬ strain had higher levels of oleoyl-CoA synthetase activity. On the basis of these data, we conclude that the observed decrease in the uptake of BODIPY-3823 and the incorporation of exogenous oleate into the phospholipid pool observed for the mutant strain were not the consequence of decreased acyl-CoA synthetase activity.
Fatty Acid Transport Profiles in the fat1⌬ Strain-The transport of fatty acid (oleate, palmitate, and myristate) was evaluated in FAT1 and fat1⌬ cells following growth in SMM. The data gleaned from these types of assays must be evaluated with caution, since they measure both fatty acid transport and subsequent metabolic utilization. As transport precedes utilization, we routinely measured levels of cell-associated fatty acid for 90 s following the initiation of the reaction and thus interpret the data in terms of uptake. We found that the uptake of oleate was linear for the first 90 s although reduced in the fat1⌬ strain when compared with the wild type. Fig. 5 illustrates the substrate-dependent uptake of oleate at 30 and 24°C in both the FAT1 and fat1⌬ strains. Using the program EnzymeKinetics (Trinity Software), the calculated maximal transport rates FIG. 3. Confocal laser scanning microscopy of FAT1 (A, B, and C) and  fat1⌬ cells (D, E, and F)  at 30°C from these data demonstrated that the fat1⌬ strain transported oleate at 64% of wild-type levels (8.36 nmol/min/ 10 8 cells versus 12.99 nmol/min/10 8 cells). The differences were most pronounced at 24°C (6.54 nmol/min/10 8 cells for the FAT1 strain versus 2.74 nmol/min/10 8 cells for the fat1⌬ strain). From these analyses, the apparent K t values for oleate transport at 24°C were similar for both the fat1⌬ and FAT1 strains (2.3 and 13.7 M, respectively). Likewise at 30°C, both strains had comparable apparent K t values (63.5 M for fat1⌬ and 61.5 M for FAT1). These calculated K t values at 30°C are similar to those previously defined in S. cerevisiae and C. tropicalis (17,19). These data are consistent with the notion that Fat1p functions to maximize oleate uptake. Our initial data indicated that the growth rates on the fat1⌬ strain on YPD/CER with palmitate and myristate were also reduced when compared with the isogenic FAT1 parental strain. We therefore tested the rates of uptake in the FAT1 and fat1⌬ strains using 100 M palmitate or myristate as substrate (Table II). These data demonstrated that the levels of palmitate and myristate uptake were also decreased in the fat1⌬ strain when compared with the wild type (20 and 47% decrease, respectively). Although these data demonstrate that fatty acid uptake is compromised in the fat1⌬ strain, it is clear that an additional transport component remains operational. DISCUSSION The present work describes the identification of the fatty acid transport protein Fat1p in the yeast S. cerevisiae. Disruption of the FAT1 structural gene results in a marked decease in 1) cell growth on YPD containing oleate, palmitate, or myristate and cerulenin, 2) the uptake of the fluorescent long-chain fatty acid BODIPY-3823, 3) the uptake and incorporation of exogenous oleate into the phospholipid pool, and 4) the rates of long-chain fatty acid uptake. We hypothesize that Fat1p represents the yeast homologue of the murine fatty acid transport protein FATP.
Fat1p was identified on the basis of amino acid similarity to the murine fatty acid transport protein FATP and is encoded within a structural gene located on chromosome II. Fat1p and FATP are remarkably similar proteins in that they 1) are of comparable length (623 and 646 amino acid residues, respectively) and calculated molecular mass (71,700 and 71,200 daltons, respectively), 2) have comparable calculated isoelectric points (8.14 and 8.32, respectively), and 3) share 54% amino acid sequence similarity and 33% amino acid sequence identity. It is predicted using the algorithms of Kyte and Doolittle (38) that each protein has four potential membrane-spanning segments. The hydropathy profiles of Fat1p and FATP are remarkably alike, suggesting these two proteins span the membrane in similar ways. Fat1p has four potential glycosylation sites, while FATP has three. Both proteins are members of the family of AMP-binding proteins on the basis of sequences conserved in adenylate-forming enzymes. In addition to Fat1p and FATP, this family of enzymes includes the CoA synthetases from mammals, yeast, and bacteria. On the basis of these conserved sequences, it is tempting to postulate that fatty acid transport proteins FATP and Fat1p and the acyl-CoA synthetase have a common evolutionary lineage.
In the present study, a deletion in the FAT1 gene has been generated to evaluate the role of Fat1p in the uptake and metabolism of fatty acids. Strains carrying a deletion of FAT1 are phenotypically asymptomatic unless the cells are grown on cerulenin, thereby blocking fatty acid biosynthesis. Under these conditions, wild-type cells overcome growth inhibition by supplementation with long-chain fatty acids. However, the fat1⌬ strain was unable to be rescued by fatty acid supplementation. These results indicated that deletion of FAT1 causes a block in one of the steps in fatty acid utilization including fatty acid transport, activation by acyl-CoA synthetase, or synthesis  1 and 2), 60 s (lanes 3 and 4), 120 s (lanes 5 and 6) 300 s (lanes 7 and 8), and 30 min (lanes 9 and 10) by the addition of trichloroacetic acid. Lipids were extracted using the technique of Bligh and Dyer (33) and analyzed by high performance TLC. The samples were standardized per 5 ϫ 10 7 cells. Lanes 1, 3, 5, 7, and 9, samples extracted from the FAT1 stain; lanes 2, 4, 6, 8, and 10, samples extracted from the fat1⌬ strain. FA, fatty acid; PE, phosphatidylethanolamine; PI, phosphatidylinositol; PS, phosphatidylserine; PC, phosphatidylcholine.
of essential higher lipids including phospholipids. There was no decrease in acyl-CoA synthetase in the fat1⌬ strain compared with the wild-type strain, and while the overall rate of incorporation of fatty acids into lipids was reduced, there was no indication that a specific class of lipid was altered. Therefore, we conclude that fat1⌬ is specifically deficient in fatty acid transport.
Fatty acid transport is a complex process that contains both protein-mediated and diffusional components. The protein-mediated component appears to be operational at low or physiological levels of long-chain fatty acids and thus must play a role in governing the accessibility of exogenous long-chain fatty acids for metabolic utilization. The murine fatty acid transporter, FATP, is induced during adipogenesis and, in addition to fat cells, is expressed at high levels in cardiac and skeletal muscle (14). In this regard, FATP plays a pivotal role in regulating available long-chain fatty acid substrates from exogenous sources in tissues undergoing high levels of ␤-oxidation or triglyceride synthesis. The identification of Fat1p in S. cerevisiae as the homologue to the murine FATP is especially significant, since this model system will allow us to specifically address the mechanism of long-chain fatty acid transport across the plasma membrane in eukaryotic cells. Like FATP, Fat1p functions to mediate the uptake of exogenous long-chain fatty acids and thus may play a pivotal role in regulating accessibility of these hydrophobic compounds prior to metabolic utilization. We hypothesize that Fat1p acts to facilitate longchain fatty acid transport in S. cerevisiae by a saturable, high affinity process. It is clear that a second mechanism for the uptake of exogenous long-chain fatty acids remains operative in fat1⌬ derivatives. The efficiency of uptake using this alternative pathway is severely compromised. We suggest that, under physiological conditions, Fat1p functions primarily when long-chain fatty acids are limiting and required for growth to facilitate the efficient uptake of these hydrophobic compounds into the cell.
The connection between the uptake and metabolism of longchain fatty acids in eukaryotic cells is largely unresolved. How-ever, a murine isoform of acyl-CoA synthetase, when expressed in COS7 cells, increases the rate of oleate uptake, suggesting a role for this enzyme in the transport of long-chain fatty acids (14). In E. coli, the acyl-CoA synthetase must necessarily operate in conjunction with the outer membrane-bound longchain fatty acid transport protein FadL to mediate the efficient uptake of long-chain fatty acids across the bacterial cell envelope (18,47,48). In the context of transport, the activity of acyl-CoA synthetase of E. coli is described as "vectorial acylation" (49). Our present data are consistent with the postulate that Fat1p of S. cerevisiae acts to facilitate the uptake of exogenous long-chain fatty acids and is distinct from the acyl-CoA synthetases. The activity of Fat1p must result in an increase the intracellular pool of long-chain fatty acids. The cell, in turn, may respond to this increased pool by activating these compounds to long-chain acyl-CoA thioesters via acyl-CoA synthetase.
In the context of yeast physiology, Fat1p may play a significant role in the uptake of exogenous unsaturated long-chain fatty acids under anaerobic conditions or conditions were oxygen is limiting. Under these conditions, the desaturase involved in the synthesis of unsaturated fatty acids is dysfunctional, resulting in auxotrophy for unsaturated long-chain fatty acids. While the present study has not evaluated the requirement of Fat1p under anaerobic conditions, the use of cerulenin to block fatty acid biosynthesis coupled with long-chain fatty acid rescue provides compelling evidence suggesting an important role of Fat1p in the uptake of these hydrophobic compounds under conditions where fatty acid synthesis is compromised. This work represents a valuable foundation to further explore the problem of long-chain fatty acid transport in eukaryotic cells. The use of S. cerevisiae as a model system is of great utility as transport can be evaluated at the genetic level in addition to the biochemical and physiological levels.  Fatty Acid Uptake in Yeast