 |
INTRODUCTION |
The transport of long-chain fatty acids across cell membranes is a
multifaceted process, which involves delivery to the membrane, transmembrane movement, and abstraction. The concentrations of free
fatty acids in the extracellular milieu and within cells are
particularly low as a consequence of their relative insolubility under
aqueous conditions. To compensate for their poor solubility, specific
mechanisms have evolved to regulate the movement, storage, and
utilization of fatty acids between and within cells. The biochemical mechanisms governing the transport of fatty acids across biological membranes are poorly understood. Unlike sugars, amino acids, and nucleotides, long-chain fatty acids are apolar compounds and readily partition into the membrane, which adds complexity to the process (1-4). Two opposing models have been proposed to explain the biochemical mechanisms that underpin long-chain fatty acid transport. The first suggests that transport is primarily diffusional and is
independent of membrane-bound fatty acid transporters (5-7). In this
scenario, fatty acids bind to and flip (diffuse) through the lipid
bilayer; transport is driven by concentration gradients generated by
either intracellular utilization (import) or extracellular fatty
acid-binding proteins (export). The second model suggests that
membrane-bound and membrane-associated proteins mediate transport (8-11). Several lines of evidence favor the involvement of proteins in
the transport mechanism, including studies showing that this is a
regulated process. In addition, cell types with high levels of fatty
acid metabolism (either degradation or storage) transport exogenous
fatty acids at higher rates when compared with those with low levels of
lipid metabolism (8-13). In a number of cell types, fatty acid
transport is inducible and commensurate with the expression of specific
sets of proteins thought to participate in this process (14-16).
Insulin regulates fatty acid uptake in several tissues, including white
adipose tissue and skeletal muscle (17). Activation of
hormone-sensitive lipase catalyzes triacylglyceride hydrolysis and
initiates fatty acid release from the cell. The process of fatty acid
transport is protease-sensitive and can be blocked through protein
modification and the use of specific antibodies (9, 12, 18). Whereas
the biophysical studies showing that fatty acid transport is a
diffusive process are compelling, equally compelling are studies
showing protein involvement. Thus, a third model, which we favor,
suggests that both protein-dependent and diffusive
components contribute to the mechanism governing long-chain fatty acid
transport. The present challenge is to understand how these processes
work together in the regulated import of long-chain fatty acids.
Using a variety of different experimental approaches, three protein
families have been identified as fatty acid transporters in mammals.
These include CD36, FABPpm, and FATP (reviewed in Ref. 18). A
complication in understanding the roles played by these proteins in
fatty acid transport is that each has been shown to have other
biochemical activities distinct from their role in transport. CD36 is a
multiligand receptor and, in addition to binding fatty acids, is able
to bind thrombospondin-1, modified low density lipoprotein, anionic
phospholipids, and collagens I and IV (19-23). In addition, there is
evidence that CD36 specifically binds erythrocytes infected with
Plasmodium falciparum (24, 25). FABPpm is also mitochondrial
aspartate aminotransferase, which plays a central role in cellular
energetics (26-28). FATP has intrinsic very long-chain fatty acyl-CoA
synthetase activity suggestive of a role in intracellular fatty acid
homeostasis (29-32). Whereas experimental evidence is strong showing
that each of these proteins participates in fatty acid import and fatty
acid trafficking, there is little direct evidence showing that any one
functions as a bona fide fatty acid transporter,
mechanistically similar to classical transporters of hydrophilic compounds.
Our approach to investigate the process of fatty acid transport and
subsequent intracellular trafficking is to use genetically tractable
systems so that individual components can be evaluated using both
molecular-genetic and biochemical tools. In the yeast Saccharomyces cerevisiae, fatty acid transport across the
membrane is complex and emulates the conditions present in mammalian
cells. Work from our laboratory and others is consistent with the
hypothesis that several distinct proteins are involved in regulated
fatty acid import, activation, and subsequent intracellular trafficking (29, 33-35). The principal players in this process include the membrane-bound protein Fat1p, the yeast orthologue of murine FATP, and
long-chain fatty acyl-CoA synthetase (primarily Faa1p). Emerging evidence suggests that Fat1p and Faa1p work in concert and are each
required for fatty acid import and targeting to specific intracellular
pools and organelles (29). Yeast strains containing a deletion in the
structural gene for Fat1p (fat1
) are distinct from the
wild type cells on the basis of a number of growth and biochemical
phenotypes (29, 33, 34). These strains 1) are compromised in their
ability to grow on media containing the fatty acid synthesis inhibitor
cerulenin and long-chain fatty acids; 2) show reduced uptake of
radioactively labeled long-chain fatty acids; 3) fail to accumulate the
fluorescent long-chain fatty acid analogue
4,4-difluoro-5-methyl-4-bora-3a,4a-diaza-S-indacene-3-dodecanoic acid (C1-BODIPY-C12), and 4) are defective in
intracellular trafficking of exogenous fatty acids (29, 34).
Recombinant DNA clones encoding either native yeast Fat1p or murine
FATP1 alleviate all growth and biochemical abnormalities when expressed
in mutant fat1 strains (33).
Members of the FATP family contain amino acid sequence
elements common to all adenylate-forming enzymes including the fatty acyl-CoA synthetases (36-38). On the basis of these observations in
the initial investigations on Fat1p, we suggested that this protein was
likely to have enzymatic activity (37). Indeed, subsequent studies from
our laboratory and from others have shown that Fat1p, mmFATP1, and
mmFATP4 have intrinsic very long-chain fatty acyl-CoA synthetase
activity in addition to playing a role in long-chain fatty acid import
(29-32). These data pose a dilemma, since very long-chain fatty acids
(>C22) in yeast are the product of de novo synthesis and
are generally not transported into the cell in response to a specific
metabolic requirement. In addition, there is an explicit requirement
for fatty acyl-CoA synthetase (Faa1p or Faap4p) in addition to Fat1p in
the transport of exogenous long-chain fatty acids (29). Therefore, we
propose that Fat1p is a bifunctional protein, which plays central roles
in fatty acid trafficking at the level of long-chain fatty acid
transport and very long-chain fatty acid activation. If this is true,
the two activities should be distinguishable by the selection of
specific mutant derivatives defective in one activity but not both.
Using a combination of directed mutagenesis of the FAT1 gene
and a series of biochemical studies to distinguish these two functions,
we present evidence that whereas these activities are generally linked, four FAT1 mutant alleles distinguish the fatty acid
transport and the very long-chain fatty acid activation functions.
| |
EXPERIMENTAL PROCEDURES |
Yeast Strains, Plasmids, and Growth Conditions--
The yeast
strains used in this study were derived from YB332 (MATa ura3
leu2-3 his3 ade2 lys2) or W303a (MATa leu2 ura3 trp1 ade2
his3). The fat1::KnR mutation
in YB332 was introduced by transformation of the strain of interest
with linear DNA generated by amplification of the kanamycin resistance
cassette (KnR) using oligonucleotides complementary to both
FAT1 and the cassette as described (39). The oligonucleotide
for the coding strand was
5'-CACTGTCAAGAAGGGCAAGAAGGCAGCAGTATGGCTTGGGCATAGGCCACTAGTGGATCTG-3', and the oligonucleotide for the template strand was
5'-CCACTGGATCATTCGTAAGTGATCCTGAAACAAACCATTCAGCAGCTGAAGCTTCGTACGC-3'. Chromosomal replacement of the native gene was confirmed by
Southern analysis of chromosomal DNA from the disruptants by
comparison with DNA obtained from the parental strain. One
fat1::KnR strain was selected and
designated LS2020. The fat1::HIS3 mutation constructed in W303a (LS2296) has been previously described (34). Yeast
strains were transformed using lithium acetate (40), whereas bacterial
strains were transformed using standard methods (41).
For growth of yeast, YPDA (1% yeast extract, 2% peptone, 2%
dextrose, and 20 mg/liter adenine-hemisulfate) or
yeast-supplemented minimal media contained 0.67% yeast nitrogen
base (YNB),1 2% dextrose,
adenine (20 mg/liter), uracil (20 mg/liter), and amino acids as
required (arginine, tryptophan, methionine, histidine, and tyrosine (20 mg/liter); lysine (30 mg/liter); and leucine (100 mg/liter) were used.
To assess growth when fatty acid synthase was inhibited, cells were
grown on YNBD plates supplemented with 45 µM cerulenin,
100 µM oleate (YNBD-CER-FA) unless otherwise indicated.
Growth in liquid culture (optical density monitored at
A600) and on plates was at 30 °C.
Bacteria were grown using standard methods in LB or minimal medium E
(42). When required to maintain plasmids, antibiotics were added to 100 µg/ml ampicillin, 40 µg/ml kanamycin, 10 µg/ml tetracycline, and
40 µg/ml chloramphenicol. Growth of bacterial cultures was routinely
monitored using a Klett-SummersonTM colorimeter equipped with a blue filter.
Oligonucleotide-directed
Mutagenesis--
Oligonucleotide-directed mutations within the
FAT1 gene were generated using the Altered
SitesTM mutagenesis system from Promega as previously
described (37) using single-stranded pDB106 as a template. Table I
lists both the mutation generated and the corresponding mutagenic
oligonucleotide. Specific mutations were confirmed using automated
dideoxy sequencing and FAT1-specific oligonucleotides (34).
Once the mutation was confirmed, restriction fragments containing the
different FAT1 mutant alleles were purified and ligated into
the centromeric plasmid YCpRS316 and the 2µ plasmid YEplac181 to
generate the plasmids listed in Table II. All plasmid constructs
retained the native FAT1 promoter. The resultant
FAT1 plasmid derivatives were transformed into the host
strains LS2020 (YB332fat1::KnR) and
LS2296 (W303a fat1::HIS3) for
phenotypic and biochemical analyses.
Phenotypic Analyses of FAT1 Mutant Alleles--
Following
transformation, individual transformants harboring the different mutant
fat1 alleles were evaluated for growth on agar plates of
YNBD, YNBD-CER, and YNBD-CER-FA (see above). Transformants harboring
centromere-derived plasmids were initially grown in YNBD to
midlog phase (A600 = 0.6-0.8), harvested by
centrifugation, washed with YNB, and resuspended in YNB to
A600 = 2.0 (6 × 107 cells/ml).
Individual samples were diluted 10
1 to 104
in YNB, and 2-µl aliquots were applied to the three different plates
(YNBD-CER-FA, YNBD-CER, and YNBD). Plates were incubated at 30 °C,
and growth was scored at 12-h intervals for 7 days. Using the
102 dilution on the YNBD-CER-FA plates as a guide, cells
that grew at a rate equivalent to the wild type showed growth after
48 h and were scored as 1. Those that showed growth at 72 h
were scored as 2, whereas those with growth after 72 h were scored
as 3. Cells that did not grow after 96 h were scored as 4.
Antibody Production, SDS-Polyacrylamide Gel Electrophoresis, and
Western Blotting--
Fat1p-specific antisera were generated using the
purified C-terminal 125-amino acid residue peptide of Fat1p as an
antigen. The coding sequence corresponding to the C-terminal 125 amino acids of Fat1p was amplified using PCR, and the product was ligated into pRSET-B (Invitrogen) to generate pDB212. This construct encodes a
hexameric histidine-tagged peptide and is under the control of a T7 RNA
polymerase-responsive promoter. Bacterial cells (BL21(
DE3)/pLysS) were transformed with pDB212, and expression of the His-tagged Fat1p
peptide (His-C125-Fat1p) was induced using
isopropyl-1-thio-
-D-galactopyranoside as previously
described (37). Following induction, cells were lysed, and
His-C125-Fat1p was purified to homogeneity using Ni2+
chelation chromatography (44). The purified peptide was used as the
antigen for the production of polyclonal anti-Fat1p serum using a
commercial vendor (BioWorld, Dublin, OH). Methods for SDS-polyacrylamide gel electrophoresis and Western blotting are standard and have been previously described (45).
Fatty Acid Import Using Fluorescent Fatty Acid--
Fatty acid
import was assessed using confocal laser-scanning microscopy to
detect the accumulation of the fluorescent long-chain fatty acid
analogue C1-BODIPY-C12 (Molecular
Probes, Inc., Eugene, OR) as previously described (33, 34). Cells were
harvested, washed with phosphate-buffered saline (PBS), and resuspended
to a final cell density of 3 × 108 cells/ml. All
steps were performed at room temperature. Washed cells were incubated
with 25 µM C1-BODIPY-C12 for 3 min; washed once in PBS containing 50 µM fatty acid-free
bovine serum albumin and once in PBS alone; and finally resuspended in
the original volume of PBS. C1-BODIPY-C12
incorporation was visualized on a NORAN-OZ confocal laser-scanning
microscope, interfaced with a Nikon Diaphot 200 inverted microscope
equipped with a PlanApo ×60, 1.4 NA oil immersion objective lens. The
instrument settings for brightness, contrast, laser power, and slit
size were optimized for the wild type strain within each experiment to
ensure that the confocal laser-scanning microscope was set for its full
dynamic range. The same settings were used for all subsequent image
collections. The average fluorescence from two or three different
determinations was calculated using these settings and a field of
50-75 cells as a semiquantitative measurement of fatty acid import.
Fatty Acid Import Using Radiolabeled Fatty Acid--
Rates of
fatty acid transport were determined essentially as described by
DiRusso et al. (33). Cells were grown in YNBD containing the
appropriate supplements at 30 °C to midlog phase (3 × 107 cells/ml), collected by centrifugation, washed once in
PBS, and resuspended to a final cell density of 8 × 108 cells/ml in PBS. 500 µl of cells (4 × 108 cells) were preincubated for 10 min at 30 °C in PBS,
and the assay was initiated by the addition of
[9,10-3H]oleate at fatty acid/bovine serum albumin ratios
of 2:1, 1:1 and 0.5:1 (holding the bovine serum albumin concentration
constant at 173 µM). The specific activity of the
[9,10-3H]oleate used in all experiments was 30 mCi/mmol.
At the defined time points (0, 2, 4, and 8 min), 100-µl duplicate
samples were diluted into 5 ml of PBS containing 0.5% Brij 58 (PBS-B)
and then were immediately filtered through a prewetted Whatman Gf/B
filter. The filters were washed two times with PBS-B. The uptake
reactions were linear between 0 and 5 min following the initiation of
the reaction. Rates were defined within the linear range. All wash steps were carried out at room temperature. Filters were air-dried, and
the amount of cell-associated radioactivity was subsequently determined
by scintillation counting. Background counts, less than 1% of the
total fatty acid, estimated as the amount of radioactivity on control
filters in the absence of cells, were subtracted from the experimental
samples. The final data were expressed in pmol of fatty acid
transported/min/1 × 108 cells. All data presented
represents the mean ± S.E. from at least three independent experiments.
Lignoceryl CoA Synthetase Activities--
Cells were inoculated
from overnight cultures into 50 ml of YNBD to an
A600 of 0.1 and grown under selective
conditions to an A600 of 1.0. Following growth,
cells were harvested by centrifugation, washed twice with PBS, and
resuspended to a density of 1.2 × 109 cells/ml in 200 mM Tris-HCl, pH 8.0, 4 mM EDTA, 5 mM 2-mercaptoethanol, 10% glycerol, 0.5 mM
phenylmethylsulfonyl fluoride, 4 µM pepstatin A, and 8 µM leupeptin. The cells were lysed by vigorously
vortexing the cell suspension containing glass beads for 1 min five
times at 0 °C. Samples were clarified by centrifugation (1500 × g, 2 min), and supernatants were used to assess acyl-CoA
synthetase activities. Acyl-CoA synthetase activities were determined
in cell extracts as described (29). The reaction mixtures contained 100 mM Tris-HCl, pH 8.0, 10 mM ATP, 20 mM MgCl2, 0.01% Triton X-100,
[14C]lignoceric acid dissolved in 10 mg/ml
-cyclodextrin (final concentration of fatty acid was 20 µM), 0.2 mM coenzyme A, and cell extract in a
total volume of 0.5 ml. The specific activity of the
[14C]lignoceric acid was 5 mCi/mmol. The reactions were
initiated by the addition of coenzyme A, incubated at 30 °C for 20 min, and terminated by the addition of 2.5 ml of isopropyl alcohol, n-heptane, 1 M H2SO4
(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 (46). The values reported represent the
average from at least three independent experiments and were analyzed
using analysis of variance and paired t tests.
Fatty Acid Analysis--
Cells were grown overnight in YNBD
supplemented with amino acids as required and diluted to an
A600 of 0.1 in 100 ml of YNBD. When the cell
density reached an A600 of 0.8-1.0, cell growth and metabolism were stopped by the addition of 
volume of
6.6 M perchloric acid. Cells were harvested by
centrifugation for 10 min at 5000 × g (4 °C). The
pellet was resuspended in 10 mM perchloric acid, and the
sample was transferred to a 15-ml glass tube. Acid-washed cells were
again pelleted by centrifugation, and the supernatant was removed by
aspiration. To monitor recovery of fatty acids, 100 µg of
heptadecanoic acid (in hexane) was added prior to lipid extraction. The
total volume was adjusted to 0.8 ml by the addition of H2O.
Lipids were extracted following the addition of 3 ml of
chloroform/methanol (2:1), cell breakage using glass beads (2 g,
425-600 µm; vigorously shaken for 1 h at 4 °C), and the
addition of 1 ml of chloroform and 1 ml of H2O. The samples were vortexed for 30 s and centrifuged for 10 min at 1500 × g (4 °C), and the upper aqueous phase was discarded. The
lower organic phase was gently transferred to a dram glass and dried
under nitrogen, and lipids were resuspended in 0.5 ml of NaOH in
methanol (20 g/liter). The samples were incubated at 100 °C for
14 h, after which time 0.5 ml of BF3 (20% in
methanol) was added, and the samples were incubated for an additional
1 h. The fatty acid methyl esters were extracted twice with 0.5 ml
of hexane, hexane phases were combined, and methyl esters were
dried under nitrogen. The fatty acid methyl esters were analyzed by gas
chromatography/mass spectrometry using an Agilent 6890 series gas
chromatograph equipped with a 5873 mass-selective detector.
Materials--
Yeast extract, yeast peptone, and yeast nitrogen
base were obtained from Difco. Fatty acids and cerulenin were obtained
from Sigma. [3H]oleic acid and
[14C]lignoceric acid were from PerkinElmer Life Sciences
and American Radiolabeled Chemicals, respectively.
C1-BODIPY-C12 was purchased from Molecular
Probes. DNA oligonucleotides were obtained from Integrated DNA
Technologies. Enzymes required for all DNA manipulations were from
Promega, New England Biolabs, U. S. Biochemical Corp., Invitrogen,
PerkinElmer Life Sciences, or Roche Molecular Biochemicals. All other
chemicals were obtained from standard suppliers and were of reagent grade.
| |
RESULTS |
Generation of Oligonucleotide-directed Mutations within
FAT1--
Fat1p was identified as the yeast ortholog of the murine
FATP1 (34). More recently, additional members of the FATP family have
been identified experimentally or by sequence comparisons in other
nonmammalian systems including Caenorhabditis elegans, Drosophila melanogaster, and Mycobacterium
tuberculosis (47). These proteins are characterized by 1) a highly
conserved sequence common to all adenylate-forming enzymes called the
ATP/AMP signature motif and 2) a highly conserved sequence restricted
to this family of proteins called the FATP/VLACS signature motif (Fig.
1). In addition to playing a central role
in fatty acid import, several members of this family also have
intrinsic fatty acyl-CoA synthetase activity, suggesting an additional
intracellular role in fatty acid trafficking into different metabolic
pools. In an effort to further define the functional activities
associated with Fat1p found in yeast, 16 oligonucleotide-directed
mutant alleles of FAT1 were constructed (Fig. 1; Tables
I and II).
In the present study, we chose to focus our experiments on dissecting
regions within Fat1p that were highly conserved to other members of the FATP family in both the ATP/AMP and FATP/VLACS signature motifs.
View larger version (52K):
[in this window]
[in a new window]
|
Fig. 1.
ATP/AMP and VLACS/FATP signature motifs
common to Fat1p (from S. cerevisiae) and the mouse
isoforms of the FATP family. The plus
symbols denote the residues within Fat1p selected for amino
acid substitution in the present study.
|
|
Phenotypic Characteristics and Expression Patterns of fat1 Mutant
Alleles--
Following construction, the different FAT1
mutant alleles were subcloned into the yeast centromeric vector
YCpRS316 for phenotypic analysis (see Table II for specific
designations). The yeast strain YB332
fat1::KnR was transformed with the
collection of centromeric plasmids containing the different
fat1 alleles, a plasmid harboring native FAT1
(YCpDB102; positive control), or the plasmid vector (YCpRS316; negative
control). Cells were grown to midlog phase under selective conditions
and spotted to YNBD-CER-FA, YNBD-CER, and YNBD plates as detailed under
"Experimental Procedures," and growth was scored at 12-h intervals
for 7 days. In addition, whole cell lysates were prepared and monitored
for Fat1p (or mutant Fat1p) expression using Western blots and
anti-Fat1p sera (data not shown). The expressions of the wild type and
mutant forms of Fat1p were equivalent. As shown in Table
III, the majority of the FAT1
mutant alleles were phenotypically distinct from the wild type on
YNBD-CER-FA plates and were diminished in growth. Only one allele
(fat1T538A) was indistinguishable from the wild
type. The majority of the fat1 mutants had growth rates
intermediate to wild type and the fat1 on YNBD-CER-FA
(showing growth between 48 and 72 h at 30 °C). Several mutant
alleles (fat1S258A,
fat1D508A, fat1Y519A,
fat1S528A, fat1R523A, and
fat1L669R) were unable to restore growth on the
fatty acid cerulenin plates.
View this table:
[in this window]
[in a new window]
|
Table III
Growth phenotypes on YNBD-FA-CER plates of the fat1 strain
transformed with centromeric plasmids encoding the collection of
fat1 site-directed mutant alleles
|
|
Each of the mutant alleles with reduced activity produced a protein
that when expressed from a 2µ plasmid could be detected in whole cell
lysates using anti-Fat1p sera, suggesting that the altered phenotypic
patterns observed were not the consequence of an unstable protein being
degraded (Fig. 2). The finding that wild
type and mutant Fat1p protein levels were comparable was not
unexpected, because the targeted amino acid was changed to alanine,
which is not expected to disrupt protein architecture. In addition,
both the 2µ and centromeric constructs harboring wild type
FAT1 and the fat1 alleles were constructed in a
manner to maintain the native FAT1 promoter elements.
View larger version (32K):
[in this window]
[in a new window]
|
Fig. 2.
Western blots of total cell lysates
(10-12 µg of protein) from wild type
FAT1 and the different fat1 alleles
probed with anti-HisC125-Fat1p showing expression of wild type
(WT) and mutant forms of Fat1p.
|
|
Patterns of Fatty Acid Import Using
C1-BODIPY-C12 and
[3H]Oleate--
As a gauge of fatty acid import, we
monitored the import of the fluorescent long-chain fatty acid
C1-BODIPY C12 into the fat1 strain harboring the different centromeric plasmids containing the
different FAT1 mutant alleles using confocal microscopy. The length of C1-BODIPY-C12 closely approximates
the length of C16:0 and thus was an appropriate
fluorescently labeled long-chain fatty acid analogue for these studies.
The positive control for these experiments was the fat1
strain transformed with YCpDB102 (FAT1), whereas the
negative control was the same strain transformed with the plasmid
vector (YCpRS316). As shown in Fig. 4, the ability to transport
C1-BODIPY-C12 varied considerably and was
dependent upon the specific fat1 allele. Indeed, there were
several mutations, including fat1R523A,
fat1S536A, and fat1L669R,
that resulted in accumulation levels that were more comparable with the
fat1 strain transformed with the vector alone. The levels of C1-BODIPY-C12 accumulation were generally
consistent with the growth patterns, with the exception of
fat1D508A, which was unable to grow on
YNBD-CER-FA but had fairly high levels of
C1-BODIPY-C12 accumulation (Fig.
3). The average fluorescence for a field
of cells was calculated as detailed under "Experimental Procedures"
in order to provide a semiquantitative measurement of fatty acid
transport. These averaged values ranged from 137.01 ± 13.4 for
the wild type FAT1 strain to 62.32 ± 4.9 for the
fat1 strain. In an effort to quantify these observations
further, fatty acid import was monitored using [3H]oleate
in a selected subset of FAT1 mutant alleles (Fig.
4). These data were generally consistent
with the C1-BODIPY-C12 accumulation data; it
was of interest to note, however, that the
C1-BODIPY-C12 import experiments seemed to show
more subtle differences in the patterns of fatty acid import. For
example, fat1Y504A had a high apparent rate of
[3H]oleate import when compared with the
C1-BODIPY-C12 accumulation. For the other
mutant alleles, there were comparable patterns of uptake using both
methods. Of particular note was the intermediate to low levels of
transport observed using each method for
fat1S258A. In contrast, the plasmid-encoded
fat1R523A allele and the vector control had
barely detectable levels of activity measured using both methods.
View larger version (58K):
[in this window]
[in a new window]
|
Fig. 3.
Patterns of
C1-BODIPY-C12 import in the
fat1 strain transformed with YCpRS316 or
centromeric plasmids encoding wild type FAT1 or
site-directed fat1 mutant alleles. Wild type and
mutant forms designated +++ had an average arbitrary relative
fluorescence value of 135-150; those slightly reduced compared
with wild type were designated ++ and had an average fluorescence of
110-130; those greatly diminished compared with wild type were
designated + and had an average fluorescence of 90-105; those with a
signal only slightly visible were designated +/ and had an average
fluorescence of 70-85; those with essentially no signal and equivalent
to the fat1 strain were designated and had an
average fluorescence of 60-65.
|
|
View larger version (62K):
[in this window]
[in a new window]
|
Fig. 4.
Patterns of [3H]oleate
transport in the fat1 strain transformed
with YCpRS316 or centromeric plasmids encoding wild type
(WT) FAT1 or selected site-directed
fat1 mutant alleles. Error
bars reflect the S.E. (n = 3).
|
|
Patterns of Lignoceryl CoA Synthetase Activities--
In addition
to its role in transport, Fat1p, like other members of the FATP family,
has intrinsic very long-chain fatty acyl-CoA synthetase activity
(natural substrates are fatty acids including C22:0
and above) (29-32). Fatty acyl-CoA synthetase activity in yeast
strains directed toward very long-chain fatty acid substrates amounts
to less than 1% of total acyl CoA synthetase activity. For example, in
previous work we showed the C24:0 CoA synthetase activity
in the wild type yeast strain is 32.4 ± 4.2 pmol/min/mg of
protein (29). This compares to 3888.9 ± 382.2 pmol/min/mg of
protein C18:1 CoA synthetase activity in the wild type
strain (29). The low activity of lignoceryl CoA synthetase activity within yeast presents a considerable technical challenge when working
to understand the interrelated roles of import and activation. To
overcome the experimental problems imposed by the low level of activity
when in a single gene dose, we chose to monitor C24:0 CoA
synthetase activities in the fat1 strain transformed with 2µ plasmids containing both wild type FAT1 and the
different FAT1 mutant alleles. As noted above, the levels of
the wild type and mutant forms of Fat1p, when expressed from a 2µ
plasmid, were comparable (Fig. 2). Expression of wild type
FAT1 from a 2µ plasmid resulted in a 6-fold increase in
lignoceryl CoA synthetase activity (Table
IV) over what was defined in the
wild-type strain. Using this value as a benchmark (197.5 pmol/min/mg of
protein), the levels of C24:0 CoA synthetase activities in
the collection of fat1 mutants were defined and compared
(Table IV). As with the C1-BODIPY-C12
accumulation data, there was general consensus between the growth
phenotype and the overexpressed lignoceryl CoA synthetase activities.
There were, however, several notable exceptions.
fat1F528A and fat1L669R
both had robust activities (38 and 24% of wild type, respectively) but
were essentially unable to grow on YNBD-CER-FA.
View this table:
[in this window]
[in a new window]
|
Table IV
Lignoceryl CoA Synthetase Activities in the fat1 strain transformed
with 2µ plasmids containing wild type FAT1 and FAT1 mutant
alleles
|
|
Total Fatty Acid Profiles in Selected fat1 Mutants--
A deletion
in FAT1 results in the accumulation of very long-chain fatty
acids (32). This is particularly true for C24:0, where
levels increase nearly 8-fold. In an effort to further define phenotypic alterations associated with mutations in FAT1,
several mutant fat1 alleles were selected (each encoded
within centromeric plasmids), and total fatty acid profiles were
determined using gas chromatography/mass spectrometry (Table
V). Two important observations came from
these experiments. First, the fatty acid profiles in the wild type and
the different mutants tested were essentially the same for
C14:0, C16:0, C16:1,
C18:0, and C18:1. Second, the levels of very
long-chain fatty acids (C22:0-C26:0), especially C24:0, mirrored the lignoceryl CoA synthetase
data. In general, when the levels of lignoceryl CoA synthetase activity were high, there were low (normal) levels of very long-chain fatty acids. In fat1 mutants with low to moderate levels of
lignoceryl CoA synthetase activity, there were correspondingly high to
moderate levels of very long-chain fatty acids, respectively. These
data are consistent with a role for Fat1p in the turnover of very
long-chain fatty acids within yeast cells.
Transport and Very Long-chain Fatty Acyl-CoA Synthetase Activities
Attributable to Fat1p Are Distinguishable--
The data described
above are generally consistent with the notion that defects in growth
phenotype on YNBD-CER-FA for this collection of site-directed
fat1 mutants translate into depressed levels of fatty acid
import, decreased levels of lignoceryl CoA synthetase activity, and
increased pools of very long-chain fatty acids. The fat1
mutants were further classified into Types I, II, and III on the basis
of coordinate changes in these morphological and biochemical
phenotypes. Mutant alleles that did not follow this trend were
classified as Types IV and V (Table VI).
Mutant alleles designated Type I had no major alterations in growth on media containing cerulenin and long-chain fatty acids and demonstrated nearly wild type levels of fatty acid import and C24:0 CoA
synthetase activities. Only one mutant allele
(fat1T538A) fell into this category. Types II
and III mutant alleles were characterized by a coordinate decrease in
growth on YNBD-CER-FA, fatty acid import, and C24:0 CoA
synthetase enzymatic activity. As noted above, the levels of enzymatic
activity were generally inversely correlated with levels of
C24:0. Mutants of Types II and III were only distinguished
in the severity of the phenotypic and biochemical defects.
The most interesting and informative observations came from four
mutants (classified Type IV and Type V) that did not have a coordinate
reduction in the biochemical and morphological phenotypes. Type IV
mutants were essentially unable to import exogenous fatty acid and were
severely restricted in growth when the fatty acids synthase inhibitor
and oleate were included in the growth media but had relatively high
levels of lignoceryl CoA synthetase activity (albeit reduced compared
with wild type). This indicates that the very long-chain fatty acyl-CoA
activity attributed to Fat1p alone does not govern the import of
long-chain fatty acids. In contrast, Type V mutants retained the
ability to import fatty acids but were essentially devoid of lignoceryl
CoA synthetase activity. The cells expressing the altered Type V
alleles also had elevated levels of C24:0 fatty acids
compared with the wild type (and more characteristic of the null
mutant). These four Type IV and Type V fat1 mutants are
particularly noteworthy, since they discriminated, for the first time,
the intrinsic activities (transport and activation) attributable to Fat1p.
| |
DISCUSSION |
The current study was undertaken to evaluate the contribution of
selected highly conserved amino acids shared within the FATP protein
family to the two functions attributed to FATP: fatty acid transport
and very long-chain acyl-CoA synthetase activity. The process of
long-chain fatty acid transport is distinct from other transport
processes, which are more classically defined (e.g. see Ref.
48). It is well understood that fatty acids partition into the membrane
and in an uncharged form flip between the two membrane surfaces (6, 7).
One model of fatty acid import suggests that downstream metabolism,
including activation, is the primary driving force behind abstraction
of the fatty acid from the lipid bilayer and, hence, apparent transport
(4, 5). Therefore, the simplest model for FATP function in fatty acid transport, which takes into account both biochemical activities, is the
abstraction of the fatty acid from the membrane concomitant with
thioesterification with coenzyme A. However, in previous work, we have
shown that the activation of exogenous long-chain fatty acids upon
import in yeast occurs through one of two long-chain acyl-CoA
synthetases encoded within the FAA1 and FAA4
genes (29). Therefore, we propose that Fat1p and fatty acyl-CoA
synthetase are each required for fatty acid import. The major finding
coming from the present work is that the long-chain fatty acid import and very long-chain fatty acid activation activities inherent to Fat1p
can be distinguished. These findings suggest that Fat1p is involved in
both the transmembrane movement of fatty acids and in the maintenance
and turnover of very long-chain fatty acids. Evidence emerging from our
laboratory supports both a functional and physical association between
Fat1p and fatty acyl-CoA
synthetase.2 In this regard,
the precise role of Fat1p in governing fatty acid import may simply
involve docking fatty acyl-CoA synthetase to the membrane, where the
latter functions in the abstraction and concomitant esterification of
long-chain fatty acids. The very long-chain fatty acyl-CoA synthetase
activity intrinsic to Fat1p may contribute to intracellular fatty acid homeostasis.
Our data evaluating the morphological and biochemical phenotypes of
this collection of fat1 mutants demonstrated that for the
most part amino acid substitutions that resulted in defective growth
phenotypes were likewise defective in fatty acid import and very
long-chain fatty acid activation. Twelve of the fat1 alleles
characterized followed this trend with varying degrees of severity on
the biochemical and morphological phenotypes evaluated. It was
interesting to note that fat1 alleles containing a
substitution within the ATP/AMP signature motif were generally Type II,
showing partial defects in both biochemical activities. On the basis of previous work from the Schaffer laboratory, this finding was not unexpected (49, 50). Whereas the protein domain encompassing the
ATP/AMP signature motif is predicted to bind ATP, there is no
structural evidence from members of the FATP family supporting this
hypothesis. Rather, the use of directed mutagenesis has provided a clue
as to the functional importance of this region of the protein. In the
present work, we demonstrated that fat1T260A was
moderately defective in fatty acid import and lignoceryl CoA synthetase
activities. This mutant corresponds to the T252G substitution in
mmFATP1, which Shaffer's group has shown is defective in fatty acid
transport and binds nearly 4-fold less azido-ATP, indicating that this
mutant protein is defective in ATP binding (50). If it is presumed that
the enzymatic activity intrinsic to mmFATP1 requires the formation of
an adenylated intermediate (common for all acyl-CoA synthetases), then
this mutant allele of the mouse isoform is catalytically defective. Our
data on the Type II mutants in FAT1, which are primarily
localized to the ATP/AMP signature motif seem to bear this out. Single
amino acid substitutions within this region of the protein depress (but
do no eliminate) both fatty acid import and fatty acid activation, supporting the notion that the binding of ATP is required for both functions.
The Type III mutants identified in this study were phenotypically
distinct from Type II by being more severely affected in each phenotype
evaluated. Strains expressing these fat1 alleles were unable
to grow on YNBD-CER-FA and were defective in both fatty acid import and
very long-chain fatty acid activation. All three Type III
FAT1 mutant alleles had amino acid substitutions within the
FATP/VLACS signature, which is characteristic of this family of
proteins and distinguishes them from other fatty acyl-CoA synthetases
and adenylate-forming enzymes (36-38). Using our past work on the
bacterial fatty acyl-CoA synthetase FadD, which plays a central role in
fatty acid import by vectorial esterification, as a guide, we suspect
that the FATP/VLACS signature is involved in the binding of the fatty
acid ligand (36, 37). The finding that most of the FAT1
mutant alleles with substitutions in the FATP/VLACS signature were
defective for both transport and activation supports the notion that
this region of Fat1p is involved in binding the fatty acid ligand
and/or is required for catalysis. What remains to be resolved is how
the substrate specificity for transport (long-chain fatty acids) and
the substrate specificity for fatty acid activation (very long-chain
fatty acids) are distinguished. We hypothesize that this may be the
result of a functional interaction with different isoforms of fatty
acyl-CoA synthetase, which in yeast (as in bacteria) also play a
pivotal role in fatty acid import (29).
The most notable mutants of FAT1 characterized in the
present study were those that distinguished the fatty acid import and very long-chain fatty acyl-CoA synthetase activities associated with
Fat1p (Types IV and V). Type IV alleles were unable to import fatty
acids, which was consistent with their defective growth phenotypes on
YNBD-CER-FA plates. These mutant fat1 alleles, however, had
significant lignoceryl CoA synthetase activity, albeit reduced compared
with wild type (fat1F528A had 38% wild type
activity, whereas fat1L669R had 24%). The very
long-chain fatty acid profiles were consistent with the intermediate
levels of C24:0 CoA synthetase activity. These results are
of particular significance, since they demonstrate that the enzymatic
activity intrinsic to Fat1p can be unlinked from the fatty acid import
function. A second group of fat1 alleles (Type V) were also
valuable, since they had high levels of fatty acid import relative to
lignoceryl CoA synthetase activities. One interesting feature of the
Type V mutants was the finding that, although able to accumulate
exogenous C1-BODIPY-C12, they were unable to
grow on YNBD-CER-FA. There are two possible interpretations for this
outcome. The first is that the fatty acid, which is accumulated under
these conditions, is not activated for further metabolism as a
consequence of a nonproductive functional linkage between Fat1p and a
cognate fatty acyl-CoA synthetase (Faa1p or Faa4p). In previous work,
we have demonstrated that Faa1p is pivotal in the coupled
import/activation of exogenous long-chain fatty acids (29). Thus, we
imagine, for example, that Fat1pS258A and
Fat1pD508A are subtly altered in a manner that results in
an inability to interact with Faa1p or Faa4p to promote this coupled
import/activation. A second interpretation is that the very long-chain
fatty acyl-CoA synthetase activity of Fat1p is essential for growth on
YNBD-CER-FA.
One fundamental question arising from the observation that FATP family
members have roles in both fatty acid transport and very long-chain
fatty acid activation is whether fatty acid import is driven as a
consequence of this intrinsic acyl-CoA synthetase activity
(i.e. metabolic trapping) or whether these activities are
distinct but performed by a bifunctional enzyme. To assess very
long-chain fatty acyl-CoA synthetase activity accurately and
reproducibly, we found it necessary to express Fat1p from 2µ
plasmids, because the activity contributed by Fat1p in yeast is at the
limits of detection using C24:0 as substrate. On the other
hand, amplification was not necessary to distinguish the differential
levels of growth, transport, and intracellular fatty acid profiles
contributed by the mutant alleles; those analyses were each done using
fat1 cells expressing the Fat1p allele of interest from a
centromeric plasmid. The latter situation more closely emulates the
normal condition in terms of gene dose. However, the role of Fat1p (and
other members of the FATP family) in fatty acid trafficking and
intracellular fatty acid homeostasis cannot be described independently
from the fatty acyl-CoA synthetases. Indeed, in the report that
identified mmFATP1 as a fatty acid transport protein, an essentially
overlooked observation was the identification of a cDNA clone
encoding a fatty acyl-CoA synthetase (51). Like mmFATP1, when expressed
in COS7 cells, this enzyme functions to promote the import of
exogenous fatty acids. When expressed in combination with mmFATP1, an
additive accumulation of exogenous fatty acid was observed, suggesting
that these two proteins may work together (51). Indeed, in these
earlier studies, Shaffer and Lodish compared their observations to
those our laboratory has reported in Gram-negative bacteria: namely a
fatty acid transport/activation-coupled process designated vectorial
esterification (52). More recent studies from the Schaffer laboratory
have shown that one isoform of fatty acyl-CoA synthetase (ACS1) from
mice is also associated with the plasma membrane, where it is
hypothesized to function in concert with mmFATP1 (53). Using the yeast
model, we have clearly demonstrated a specific requirement for fatty
acyl-CoA synthetase (either Faa1p or Faa4p) as well as Fat1p in the
fatty acid import process (29). In the context of intracellular fatty acid homeostasis, Fat1p may function independently of Faa1p (or Faa4p)
and specifically be involved in very long-chain fatty acid metabolism.
The present study has provided fundamental information regarding the
role of Fat1p and the interrelated activities of fatty acid transport
and fatty acid activation. The use of this genetic system has allowed
us to distinguish, for the first time, these two activities and thus
lay a foundation necessary to further understand the role of this
protein in fatty acid metabolism. Additionally, this model system has
allowed us to demonstrate that the mammalian ortholog of Fat1p
(mmFATP1) functions in a similar manner and can substitute for Fat1p in
strains deleted for FAT1 (33). Thus, this work establishes a
foundation for complementary studies using the different mammalian
isoforms. It is expected that these studies will help to distinguish
functional domains and specific amino acid amino acid residues among
the FATP family members required for substrate preference and
additional functions associated with fatty acid transport and activation.
,
TOP
ABSTRACT
INTRODUCTION
