Liver Fatty Acid-binding Protein Targets Fatty Acids to the Nucleus

Although unesterified long chain fatty acids interact with peroxisome proliferator-activated receptors to initiate transcription within the nucleus, almost nothing is known regarding factors regulating long chain fatty acid distribution to the nucleus of living cells. The possibility that the liver fatty acid-binding protein (L-FABP) may function in this role was addressed in transfected L-cell fibroblasts overexpressing L-FABP using a series of fluorescent fatty acids differing in chain length and unsaturation. After 30 min of incubation, oxidation of BODIPY-, NBD-, and cis-parinaric acids was undetectable in L-cells. Likewise, L-cells very poorly esterified these fluorescent fatty acids in the following order: 0% BODIPY-C5, NBD-C6 (short chain length) < 0–3% NBD-C18, BODIPY-C16, cis-parinaric acid (long chain length) < 11% BODIPY-C12 (medium chain length). Real time confocal and multiphoton laser scanning microscopy (CLSM and MPLSM) showed that these fluorescent fatty acids were generally taken up in the following order: long chain (BODIPY-C16, NBD-C18) > medium chain (BODIPY-C12) ≫ short chain (BODIPY-C5, NBD-C6). The fluorescent fatty acids were imaged in the nucleus, primarily associated with the nuclear envelope, at levels about 2–3-fold lower than outside the nucleus. CLSM and MPLSM showed that L-FABP expression enhanced by 2–4-fold the initial rate and/or average maximal uptake of the long and medium chain but not the short chain fluorescent fatty acids in living cells. Furthermore, L-FABP expression increased the targeting of long and medium but not short chain fluorescent fatty acids to the nucleus by 2.9–4.4-fold and increased the proportion (i.e.nuclear:cytoplasm ratio) of medium and long chain but not short chain fatty acids by 2–3.6-fold. In summary, these results showed for the first time the presence of unesterified fatty acids in the nucleus of living cells and demonstrated that expression of a fatty acid-binding protein, L-FABP, specifically enhanced uptake and intracellular targeting of long and medium chain fatty acids to the nucleus.

Because the nuclear peroxisome proliferator-activated receptors (PPAR) 1 regulate expression of genes involved in fatty acid oxidation (PPAR␣) and adiposity (PPAR␥), it has been postulated that PPARs play important roles in diabetes, atherosclerosis, and obesity (reviewed in Refs. [1][2][3][4][5][6]. In support of this hypothesis, fatty acid oxidation is diminished in PPAR␣ geneablated mice (7)(8)(9) accompanied by fat accumulation (heart (10), centrilobular regions of the liver (7,8), adipose tissue (11)), hypercholesterolemia (12), and hypoglycemia (10). Although there is considerable debate regarding the endogenous ligands activating PPARs (reviewed in Ref. 1), a recent report utilized a direct fluorescent binding assay and parinaric acid (naturally occurring fluorescent fatty acid) to show very high affinity of PPAR␣ (i.e. 5-17 nM K d values) for long chain fatty acids (LCFAs) (13). Likewise, the parinaric acid binding assay (14) and the crystal structure of PPAR␥ containing bound LCFA (i.e. eicosapentaenoic acid) (15) both demonstrated that PPAR␥ also binds LCFAs (16,17), albeit with much lower affinity (K d values of 0.6 -3.1 M) (14) than PPAR␣ (i.e. 5-17 nM K d values) (13). These data strongly support the possibility that LCFAs may be natural ligands for PPARs and that LCFAs may be more specific ligands for PPAR␣ as compared with other PPARs. Unfortunately, physiological significance of these findings is not yet clear since it is not known whether unesterified, unbound LCFAs are present in the cell and that their concentration in the nucleus is at least in the nanomolar range.
A variety of studies suggest that unesterified LCFAs are present within the cell at levels physiologically significant for regulating PPAR␣ activity. Based on the Michaelis constant of long chain fatty acyl-CoA synthetase, the total cellular unesterified LCFA concentration has been estimated near 20 M (18). However, because of the presence of intracellular fatty acid-binding proteins (FABPs) (reviewed in Ref. 19), the intracellular concentration of unbound, unesterified LCFAs is thought to be nearly 3 orders of magnitude lower, near 7-50 nM in liver, heart, and intestine (20,21). Whether nuclear levels of unbound LCFAs are also in the nanomolar range is not known.
Another key question is the origin of nuclear unesterified LCFAs. Nuclei do not synthesize LCFAs but must import them from the cytosol (22)(23)(24) by as yet unresolved transport systems. Nearly 20 and 4% of the plasma-derived unesterified LCFAs appear in nuclei isolated from hepatocytes (23) and cerebral cortex or leukemic lymphocytes (22, 24 -26), respectively. Despite data suggesting the presence of high (up to 40% of nuclear lipids) levels of unesterified LCFAs in isolated nuclei (27), subcellular fractionation studies may be complicated by LCFAs derived from lipolytic release, transfer from other intracellular sites, or cross-contamination with other mem-branes. Furthermore, analysis of isolated nuclei does not discriminate between unesterified LCFAs present in the nuclear envelope versus the nucleoplasm.
Although mechanism(s) regulating the distribution of LCFAs to the nucleus have not yet been elucidated, several factors suggest that the liver fatty acid-binding protein (L-FABP) may be a candidate to mediate LCFA targeting to the nucleus (28). First, L-FABP is present in the highest concentration in the cytoplasm (19). Second, L-FABP increases the cellular uptake of naturally occurring LCFAs (palmitic acid (29,30), oleic acid (19), and cis-parinaric acid (19,(31)(32)(33)) as well as synthetic fluorescent fatty acid (NBD-stearic acid (34)). Third, L-FABP increases the intracellular transport/diffusion of LCFAs such as NBD-stearic acid (19, 34 -37). Fourth, L-FABP has 5-50-fold (depending on the specific fatty acids examined) lower affinity for LCFAs as compared with the affinity of PPAR␣ for LCFAs and may thereby donate LCFAs to the nucleus and/or to PPAR within the nucleus (28, 38 -44). Despite these observations, there is as yet no evidence showing that L-FABP expression enhances the targeting/transport of unesterified LCFAs to the nucleus of living cells. In fact, since the majority of L-FABP is cytoplasmic, an alternate possibility is that the near millimolar concentration of L-FABP in the cytoplasm could compete with the nucleus for binding LCFAs and thereby potentially inhibit LCFA uptake to/into the nucleus.
The objectives of the present paper are as follows. (i) Develop a series of poorly or non-metabolizable fluorescent fatty acids essential for examining the intracellular distribution of unesterified LCFAs in living cells. Both synthetic (i.e. BODIPY-or NBD-tagged) and naturally occurring (cis-parinaric acid) fluorescent fatty acids were shown to be poorly or not metabolized. (ii) Visualize and record in real time not only the cellular uptake of these fluorescent fatty acids but determine whether they are present in the nucleus of living cells. Synthetic (i.e. BODIPY-or NBD-tagged) fluorescent fatty acids were visualized by laser scanning confocal microscopy (LCSM), whereas naturally occurring cis-parinaric acid was visualized by multiphoton laser scanning microscopy (MPLSM). All were detected in nuclei. (iii) Examine if L-FABP expression in L-cell fibroblasts altered not only cellular uptake of these fatty acids but also their uptake into nuclei of living cells. Find out if L-FABP expression enhanced targeting of medium and long chain fatty acids to the nuclei. (iv) Determine if L-FABP targeting of fluorescent LCFAs to nuclei is specific. L-FABP expression did not enhance targeting of short chain fatty acids (not bound by membrane or intracellular FABPs) to nuclei of living cells. (v) Show whether L-FABP expression altered the relative intracellular distribution of fatty acids to the nucleus as determined by the nuclear/cytoplasmic ratio. L-FABP expression increased the nuclear/cytoplasmic ratios of medium and long chain fluorescent fatty acids.
Cell Culture of Stock Control and L-FABP Overexpressing L-cells-Murine L cells (L arpt Ϫ tk Ϫ ) transfected with cDNA encoding L-FABP were obtained and cultured as described earlier (45). Western blotting showed that L-FABP was absent in untransfected and mock-transfected L-cells but represented 0.4% of the total cytosolic protein in L-cells transfected with cDNA encoding L-FABP, consistent with earlier findings on this cell line (45). Control (mock-transfected) and L-FABP expressing transfected L-cells were cultured with Higuchi medium containing 10% fetal bovine serum (Invitrogen) in 75-cm 2 Falcon flasks and grown to confluence at 37°C with 5% CO 2 in a humidified incubator.
Cell Culture for Real Time Imaging of Fluorescent Fatty Acids in Intact Cells-For real time imaging, cells were seeded onto Lab-Tek chambered cover glass slides (Nunc, Naperville, IL or VWR Scientific, Sugarland, TX) at a density of 25,000 -50,000 cells/chamber. The cells were then cultured for 36 -48 h at 37°C with 5% CO 2 in Higuchi medium containing 10% fetal bovine serum (Invitrogen). To determine fluorescent fatty acid uptake and intracellular distribution, cells were allowed to equilibrate to room temperature basically as described (34,(45)(46)(47). Briefly, culture medium was removed from cells grown on coverslips followed by washing with PBS (Dulbecco). An aliquot of the stock solution of fluorescent fatty acid was mixed separately with uptake solution, and 1-2 ml of this solution was added to the cells. A variety of uptake solutions was tested with and without fluorescent fatty acid including PBS Ϯ fatty acid-free albumin, Higuchi medium Ϯ fatty acid-free albumin, or Higuchi medium Ϯ serum. The type of uptake solution did not affect the intracellular distribution of fluorescent fatty acids. Therefore, PBS was used as the buffer for measuring intracellular distribution of the fluorescent fatty acids. After testing a range of concentrations for each fluorescent fatty acid, it was determined that for each a concentration of 100 nM (i) was on the linear portion of the uptake concentration curves, and (ii) yielded a bright fluorescence signal with minimal photobleaching under instrumental LSCM or MPLSM conditions used over the time course of the experiment. The final solvent concentration in the assay did not exceed 0.1%, a level at which cell growth was normal and no cytotoxicity was observed.
Real Time Visualization of Synthetic BODIPY-and NBD-tagged Fluorescent Fatty Acids in Living Cells, Laser Scanning Confocal Microscopy (LSCM)-LSCM studies were performed with ϫ63 Plan-Fluor oil immersion objective, N.A.1.45, with a Axiovert 135 microscope (Zeiss, Carl Zeiss Inc., Thornwood, NY) and MRC-1024 fluorescence imaging system (Bio-Rad) equipped with three independent low noise photomultipliers. BODIPY-or NBD-labeled fatty acids were excited at 488 nm with a krypton-argon laser (5 milliwatts, all lines, measured at the microscope stage) (Coherent, Sunnyvale, CA) set at 1-3% scan strength of laser power (100 milliwatts). Excitation was passed through a 1% neutral density filter prior to entering the sample. The microscope stage was maintained at room temperature. Emission from BODIPY-or NBD-labeled fatty acids was recorded by a photomultiplier after passing through a 525-565 nm bandpass filter under manual gain and black level control.
Real Time Visualization of Naturally Occurring Fluorescent cis-Parinaric Acid Uptake in Living Cells, MPLSM-Although LSCM was optimal for imaging fluorophores excited in the visible wavelength region, this could not be accomplished with the naturally occurring fluorescent cis-parinaric acid, a C18 fatty acid whose conjugated tetraene chromophore absorbs in the deep ultraviolet region (290 -340 nm range). Deep UV excitation of cis-parinaric acid by LSCM requires quartz optics, almost complete photobleaches the probe, and/or is phototoxic within seconds. These problems were avoided by imaging cis-parinaric acid in living cells with MPLSM which utilizes multiphoton excitation (high intensity, femtosecond pulsed infrared laser) at 680 or 930 nm to provide 2-or 3-photon excitation, respectively (equivalent to single photon excitation of cis-parinaric acid at 340 or 310 nm, respectively). However, 2-photon excitation at 680 nm was less desirable since it excited the red edge of cis-parinaric acid absorbance spectrum, yielded only weak cis-parinaric emission, and simultaneously excited NADH resulting in high autofluorescence. In contrast, 3-photon excitation at 930 nm was in the main part of cis-parinaric acid absorbance and did not excite NADH. Because MPLSM excited cis-parinaric acid only at the objective focal volume (about 0.01 cu ) (versus LSCM excitation which occurs throughout the entire cell), more sensitive external detectors were used not requiring a pinhole as is the case for the internal detectors used in LSCM. Together, these features of MPLSM resulted in much less photobleaching such that cis-parinaric emission intensity was stable for up to 1 h. All MPLSM instrumental procedures were basically the same as for LSCM except that the excitation source was a femtosecond pulsed Mira 900 Ti-Sapphire laser pumped at 8 watts with a SABRE 35-watt argon-ion laser (Coherent, Palo Alto, CA). The Tisapphire laser was tuned to emit at 930 nm and focused to the Axiovert 135 microscope stage (Carl Zeiss Inc., Thornwood, NY) via a modified epiluminescence light path. The fluorescence signal, collected by the ϫ63 objective, was transmitted through a dichroic mirror and 450 SP or 550 LP barrier filters (CVI Laser Corp., Albuquerque, NM) and refocused on an external, low noise, photon counting R5600-P photomultiplier tube (Hamamatsu, Bridgewater, NJ).
Image Acquisition by LSCM and MPLSM-Cells were first visualized by transmission microscopy to ensure normal morphology, and the objective was focused to acquire images through a median section (i.e. slicing through the cell and the nucleus) of multiple cells in the field, and LSCM or MPLSM images of background fluorescence were obtained prior to incubation with medium containing fluorescent fatty acid. For LSCM single 0.3-m confocal slices or z series were acquired, whereas similar MPLSM images were obtained from the focal point of excitation. Several cells from each image were analyzed, and reported intensities represented an average of multiple cells from 5-6 individual chamber slides, repeated 3-4 times with the same batch of cells to limit variability due to cell counts, and with different batches to check repeatability. Fluorescence intensity from 10 regions of interest per field were collected every 10 s. Cells were excited for 0.1-s intervals, regulated by computer-controlled shutter and LaserSharp software (Bio-Rad).
Analysis of LSCM and MPLSM Images-Images were analyzed using LaserSharp (Bio-Rad) and MetaMorph Image Analysis (Advanced Scientific Imaging, Meraux, LA) software. Fluorescence intensity of the medium was set to 0, and an increase above 0 was recorded as arbitrary units. Autofluorescence was very low, at the level of background. Fluorescence intensities from low to high were represented by pseudocolors; black was the lowest and white was the highest. Most experiments were performed in the frame acquisition mode with fast scan images 512 pixels ϫ 512 lines obtained in 1 s. Frame mode images were recorded and stored over 180 -300 s every 10 s for later analysis. Fluorescent intensity in images of single living cells was expressed as the mean fluorescence intensity in gray scale units expressed Ϯ S.E. per unit area (average pixel intensity). The regions of interest measured were the nucleus (including the nucleoplasm and the nuclear envelope) and the cytoplasm (including all other intracellular membranes). Although images of fluorescent fatty acid uptake were generated for up to 1-72 h of incubation, most studies were completed in Ͻ30 min. The increase in intensity over the first few minutes (up to 15) of the linear incorporation curve was calculated by linear regression analysis to obtain the initial rate. The maximal intensities were averaged. The analysis of the time course of fluorescent fatty acid uptake and intracellular distribution utilized Time Course Software (Bio-Rad), Scion Image and SigmaPlot software (SPSS Inc., Chicago, IL). Statistical analysis was performed using Student's t test.
Mass of Fluorescent Fatty Acids Taken Up, Lipid Extraction, and Chemical Analysis-To find out if the fluorescent fatty acids were metabolized inside the cell, L-cells were incubated with fluorescent fatty acids (1 M in PBS, 5 ml, at room temperature) on 10-cm plates for 5 or 30 min, followed by lipid extraction and chemical analysis basically as described earlier (47). Total fluorescent fatty acid uptake was quantitated by comparison of the fluorescence intensity of an aliquot of the extracted lipids to a standard curve (prepared by adding increasing amounts of fluorescent fatty acids to control L-cell homogenates before lipid extraction). Total fluorescence was determined after spotting on TLC plates, acquiring and analyzing fluorescence images with a Chemi-Imager System and FluorChem version 2.0 software (Alpha Innotech Corp., San Leandro, CA). Fluorescence was converted to mass of fluorescent fatty acid taken up by comparing to standards on the same plate and expressed per mg of protein. To determine whether fluorescent fatty acids were esterified, an aliquot of the lipid extracts was resolved by TLC into free fatty acids, phospholipids, triacylglycerides, diacylglycerides, monoacylglycerides, cholesteryl esters, and cholesterol. Three different solvent systems were required to resolve adequately the unesterified as well as esterified individual lipid classes containing the respective fluorescent fatty acids (BODIPY-C16, -C12, -C5; NBD-C18, -C6). First, to determine whether BODIPY-C16, -C12, -C5, as well as NBD-C18 were esterified into cholesteryl esters or triacylglycerides, an aliquot of each lipid extract was resolved with petroleum ether/diethyl ether/methanol/glacial acetic acid (180:14:4:1, v/v). With this solvent system the remaining more polar (unesterified fatty acid, phospholipid) fluorescent lipid fractions were unresolved near the origin. Second, to resolve the more polar (unesterified fatty acid, phospholipid) lipid fractions containing fluorescent fatty acids (BODIPY-C16, -C12, -C5, or NBD-C18), another aliquot of lipid extract was resolved with petroleum ether/diethyl ether/glacial acetic acid (70:30:1, v/v). Third, although the preceding solvent systems were adequate for resolving BODIPY-C16, -C12, -C5, as well as NBD-C18 unesterified and esterified fluorescent lipid fractions, they did not resolve NBD-C6. Because NBD-C6 was the most polar fatty acid tested, the more polar solvent system petroleum ether/diethyl ether/glacial acetic acid (50:50:1, v/v) was required.
For each type of solvent system, aliquots of the lipid samples were dried under nitrogen, resuspended in 100 l of CHCl 3 , and then spotted in separated lanes on Silica Gel G TLC plates (pre-baked at 100°C for 1 h). The TLC tank was pre-equilibrated with the respective solvent system (see above) for 1 h, and samples were resolved by TLC using the same running solvent system. After the solvent front reached the top of the TLC plates, they were removed from the tank and allowed to air dry. Fluorescence images of the TLC plates were taken and analyzed with a ChemiImager System and FluorChem version 2.0 software (Alpha Innotech). Fluorescence intensities were converted to mass of fluorescent fatty acid in each lipid fraction by comparing the intensity to a series of standards on the same plate. The standards, prepared as described in the preceding section, were spotted on the same TLC plate as the unknowns to construct a standard curve.
Distribution of Fluorescent Fatty Acids within the L-cell Phospholipid Fraction, Individual Phospholipid Classes-The phospholipid spots from the preceding TLC systems were scraped into separate acidwashed test tubes, and 4 ml of chloroform/methanol/concentrated hydrochloric acid (100:50:0.375, v/v) was added. After vortexing and centrifugation at 1,000 rpm for 10 min, the supernatants were transferred to separate acid-washed glass test tubes, and sedimented silica was re-extracted with another 4-ml aliquot of the above solvent. Thereafter, 4 ml of pure water was added to the 8-ml solvent-containing sample, which was then vortexed and centrifuged at 1,500 rpm for 10 min. The bottom phase was transferred to separate acid-washed glass test tubes, dried under N 2 , resuspended in 100 l of chloroform, and spotted on Silica Gel 60 TLC plates that had been preactivated at 100°C. The phospholipid fractions were then resolved with chloroform/methanol/ double distilled water/glacial acetic acid (150:112.5:10.5:6, v/v). Fluorescence images of the air dried TLC plates were then obtained and quantitated with the ChemiImager System and FluorChem version 2.0 software (Alpha Innotech).
Distribution of Fluorescent Fatty Acids into the Fatty Acyl-CoA Pool-After removal of culture medium from L-cells on 10-cm culture plates, cells were washed with 2 ml of PBS and incubated with fluorescent fatty acid (1 M in PBS, 5 ml) at 24°C for 5 or 30 min. Incubation medium was then removed, and cells were washed 3ϫ with 2 ml of PBS. After removal of the last PBS wash the culture plates were floated on liquid nitrogen for 10 s, and 2 ml of 2-propanol was added. The frozen cells were quickly scraped from the plate surface, transferred to acidwashed glass test tubes with 2 ml of KH 2 PO 4 (100 mM, pH 5.3), and saturated (NH 4 ) 2 SO 4 (0.25 ml) and acetonitrile (4 ml) were added, and the resulting emulsion was vortexed and centrifuged at 1,900 ϫ g for 5 min. The supernatant was diluted with 10 ml of KH 2 PO 4 (100 mM, pH 5.3) and fatty acyl-CoA esters were extracted by a solid phase extraction procedure (48) using oligonucleotide purification cartridges (Applied Biosystems, Foster City, CA). Each cartridge was prewashed with 5 ml of acetonitrile, dried by gently flushing with air, washed with 2 ml of KH 2 PO 4 buffer (25 mM, pH 5.3), and again flushed gently with air. The sample pool was loaded with a polypropylene syringe into the cartridge and pushed completely through the syringe slowly. The cartridge was flushed with KH 2 PO 4 buffer (5 ml, 25 mM, pH 5.3), dried by flushing with air twice, and eluted slowly with 60% CH 3 CN in 100 mM KH 2 PO 4 (0.3 ml). The acyl-CoA fraction was eluted in the last 0.25 ml. The samples were dried and resuspended in 100 l of CHCl 3 /MeOH, 2:1, and spotted onto Silica Gel 60 TLC plates that had been preactivated at 100°C before use. The TLC plates were then run with the solvent system chloroform/methanol/double distilled water/glacial acetic acid (150:112.5:10.5:6, v/v) until the solvent front reached the top plate marker. The TLC plates were removed from the tank and allowed to briefly dry. Fluorescent images of the TLC plates were acquired and analyzed with the ChemiImager System and FluorChem version 2.0 software (Alpha Innotech).

Selection of Fluorescent Fatty
Acids-To determine in real time the presence of unesterified fatty acids in the nucleus and the effects of L-FABP expression thereon, three groups of fatty acids were chosen, Group 1 was composed of fluorescent fatty acids with short fatty acyl carbon chain length (ՅC6) (BODIPY-C5 (Fig. 1A), NBD-C6 (Fig. 1B)). These fatty acids were chosen to demonstrate the specificity of L-FABP effects on fatty acid targeting to the nucleus because they are very aqueous soluble, readily pass through membranes independent of specific transporters that translocate medium and long chain fatty acids across membranes (49,50), and are not bound by L-FABP (51)(52)(53). Group 2 consists of a medium chain length (C12) fluorescent fatty acid (BODIPY-C12 (Fig. 1A)) that is relatively lipophilic, is utilized by translocases to cross membranes (49), and is bound by L-FABP, albeit with 2-fold lower affinity than longer chain fatty acids (52,53). Group 3 is composed of synthetic (BODIPY-C16 (Fig. 1A), NBD-C18 (Fig. 1C)) and naturally occurring (cis-parinaric acid (Fig. 1C)) long chain (ՆC16) fluorescent fatty acids. Long chain fatty acids are highly lipophilic, are utilized by translocases to cross membranes (54,55), and are bound (19,49,50,56) by L-FABP with highest affinity (52,53). The following sections individually address the uptake, intracellular distribution, and metabolism (esterification) of each group of fluorescent fatty acids and examine the effects of L-FABP expression on these parameters. Cellular Uptake, Intracellular Distribution, and Esterification of Short Chain Fluorescent Fatty Acids (BODIPY-C5, NBD-C6)-The uptake of short chain fatty acids (BODIPY-C5, NBD-C6) was determined in control L-cells by chemical analysis and by real time LSCM in living cells. L-cells took up only low quantities of short chain fatty acids. When L-cells were incubated for 5-30 min with the respective short chain fatty acids and lipids extracted, uptake was very low, barely detectable by 5 min for either BODIPY-C5 or NBD-C6 (Fig. 2), and clearance from the medium was not significant (not shown). Even by 30 min uptake of BODIPY-C5 (0.024 nmol/mg protein) represented only 1.1 Ϯ 0.1% of total cleared from the medium, A, BODIPY-fatty acids (n ϭ 5 for hexanoic acid, n ϭ 11 for dodecanoic acid, and n ϭ 15 in hexadecanoic acid). The BODIPY group is located at the alkyl end of each molecule. B, NBD-fatty acids (n ϭ 6 for hexanoic acid and n ϭ 18 for stearic acid). The NBD group is located at carbons 6 and 12 of the NBD-C6 and NBD-C12, respectively. C, cis-parinaric acid. a level that was 28-fold lower than with BODIPY-C16, a long chain fatty acid probe (Fig. 2). Similarly, when L-cells were incubated for 5-30 min with NBD-C6, uptake was also very low, barely detectable by 5 or 30 min (Fig. 2), and clearance from the medium was not significant. When these experiments were repeated with the shortest commercially available BODIPY and NBD reagent (i.e. BODIPY-C3 and NBD-Cl), uptake was also extremely low. BODIPY-C3 uptake was less than 3% that of BODIPY-C16, and NBD-Cl uptake was undetectable.
Because lipid extraction and chemical analysis of the watersoluble, short chain fatty acids may not entirely account completely for uptake due to some loss in aqueous fractions of the extract, uptake of BODIPY-C5 and NBD-C6 was examined by LSCM of intact, living cells. LSCM showed only a low intensity signal (Ͻ50 arbitrary units) for BODIPY-C5 in the cells during the time course of incubation ( Fig. 3A and Fig. 4A). The uptake of BODIPY-C5 was essentially maximal within 1 min of incubation and did not increase at longer incubation, from 30 min to 24 h (data not shown). The initial rate (0.73 average fluorescence units/s) and average maximal extent for total uptake (43 average maximal fluorescence units) of BODIPY-C5 into Lcells were 6-and 4-fold lower, respectively, than for long chain fluorescent fatty acids such as BODIPY-C16 (Table I). NBD-C6 was even more weakly taken up than BODIPY-C5, and intensity levels were barely detectable (not shown).
Determination of the intracellular distribution of the short chain fatty acids in real time by LSCM of intact L-cells showed that BODIPY-C5 was not only weakly taken up but within the cell was distributed primarily into cytoplasm. BODIPY-C5 appeared in the cytoplasm (small red Ͼ yellow Ͼ white pixels) and less so in the nuclei, either close to the nuclear envelope or inside nucleoplasm (Fig. 3A). Analysis of a three-dimensional reconstruction of a z series of sections through the cells (not shown) suggested that the bright (white) spots of BODIPY-C5 localized at protrusions of the nuclear envelope represent invaginations of the nuclear envelope into the nucleoplasm. The initial rate (0.59 average intensity/s) and maximal uptake (26.1 Ϯ 1.8 average maximal intensity) of BODIPY-C5 appear-ing in the cytoplasm were about 3-and 2.6-fold faster, respectively, as compared with those appearing in the nucleus (0.2 and 10.4 Ϯ 1.7) (Table I). Because the average maximal intensities in cytoplasm (Fig. 4A) and nuclei (Fig. 4B) remained constant over time after 1 min, this suggested that once taken  up the BODIPY-C5 was rapidly (by 1 min) equilibrated within the cells. Furthermore, these data showed that the average maximum pixel intensities of BODIPY-C5 in cytoplasm were 2.6-fold higher than in the nucleus (Table I) and did not significantly change with time after the first min of uptake (Fig. 4, A  versus B). Finally, the ratio of average pixel intensities/unit area of BODIPY-C5 in the nucleoplasm/cytoplasm did not increase with time, suggesting that this probe equilibrated within 1 min in both compartments, albeit remaining higher outside the nucleus (Fig. 4C). Similar data showing poor or undetectable uptake were obtained for the BODIPY-C3 and NBD-Cl reagent alone (not shown).
To determine whether the above uptake and intracellular distribution data for short chain fatty acids (BODIPY-C5 or NBD-C6) reflected the unesterified BODIPY-C5 and NBD-C6, L-cells were incubated with BODIPY-C5 or NBD-C6 for 5 and 30 min followed by lipid extraction and analysis by TLC. The fluorescent BODIPY-C5 and NBD-C6 appeared as unesterified fatty acids, with no detectable esterified forms in any other lipid fraction (Table II).
In summary, short chain fluorescent fatty acids (BODIPY-C5, NBD-C6) were rapidly taken up in low amounts distributed within L-cells primarily in the cytoplasm and less so in the nucleus. All uptake of BODIPY-C5 and NBD-C6 as well as their LSCM fluorescence images reflected the unesterified forms of these fatty acids, since neither BODIPY-C5 nor NBD-C6 was detected esterified to fatty acyl CoAs, cholesteryl esters, or glycerides.
Effect of L-FABP Expression on Uptake and Intracellular Distribution of Short chain Fatty Acids (BODIPY-C5, NBD-C6)-Because L-FABP does not bind short chain fatty acids (51), the BODIPY-C5 and NBD-C6 served as important controls for determining the specificity of L-FABP effects on fluorescent fatty acid uptake and intracellular distribution. L-FABP expressing L-cells took up BODIPY-C5 (Fig. 3B) more rapidly than NBD-C6 (Fig. 3C). Although BODIPY-C5 fluorescence intensity was maximal within a few minutes, the observed fluorescence signals were very low (Ͻ50 arbitrary units) (Fig. 4D). L-FABP expression did not significantly increase either the time course (Fig. 4, A and B), initial rate of uptake (0.71 versus 0.73 average intensity/s, Table I), or average maximal extent of uptake (36.9 Ϯ 2.4 versus 43.2 Ϯ 1.8 average maximal intensity, Table I) of total BODIPY-C5 uptake into the cells. NBD-C6 uptake in L-FABP expressing cells was also very low (Fig. 3C) and did not change significantly after 5 min of uptake (Fig. 4D). Because NBD-C6 uptake was barely detectable in both L-FABP expressing and control cells, it was not possible to make accurate quantitative comparisons.
Within the L-FABP expressing L-cells, the initial rates of uptake and average maximal uptake of BODIPY-C5 into cytoplasm were about 4.5-and 3-fold, respectively, faster than into nuclei (Table I). However, L-FABP expression did not differentially alter the distribution of BODIPY-C5 into nucleus versus cytoplasm. The initial rate (0.59 versus 0.5 average intensity/s) and maximal average uptake (32.7 Ϯ 0.1 versus 26.1 Ϯ 1.8 average maximal intensity) of BODIPY-C5 into cytoplasm did not differ significantly from those of control L-cells (Table I). Likewise, L-FABP expression did not appear to increase significantly the initial rate (0.13 versus 0.2 average intensity/s) or average maximal uptake of BODIPY-C5 into nuclei (10.8 Ϯ 0.6 versus 10.4 Ϯ 1.7 average maximal intensity, Table I). This was confirmed, at least at early time points, by the ratio of average BODIPY-C5 pixel intensities/unit area in the nucleus/cytoplasm (Fig. 4C). Likewise, NBD-C6 fluorescence was barely detectable in the nucleus of L-FABP expressing cells (Figs. 3C and 4D) and did not differ from controls (not shown). Basically, similar data were obtained with BODIPY-C3 and NBD-Cl reagent alone (not shown).
Taken together, LSCM imaging of BODIPY-C5 and NBD-C6 in L-FABP expressing cells indicated that L-FABP did not significantly alter the uptake parameters (initial rate, average maximal extent) into whole cells (i.e. total uptake), the cytoplasm, or nuclei. Thus, L-FABP expression did not alter the pattern and intracellular distribution of the short chain fatty acids, consistent with the inability of L-FABP to bind short chain fatty acids.

Effect of L-FABP Expression on Cellular Uptake, Intracellular Distribution, and Esterification of Medium Chain Fluorescent Fatty Acid (BODIPY-C12)-Although
BODIPY-C12 appeared 12-fold more rapidly in L-cell lipid extracts than the short chain BODIPY-C5 (Fig. 2), at 5 and 30 min this represented clearance of only 7.4 Ϯ 1.4 and 12.4 Ϯ 1.6% of BODIPY-C12 in the medium, respectively. The more extensive uptake of BODIPY-C12 was confirmed by LSCM of individual cells (Fig.  6, A and B, versus Fig. 4, A and B) and kinetic analysis ( Table  I). The majority of medium chain BODIPY-C12 was localized diffusely in the cytoplasm, more intensely in perinuclear regions (endoplasmic reticulum or lipid droplets), and by 3 min within the nucleoplasm as few but discrete and brightly fluorescing areas (not shown). A three-dimensional reconstruction (not shown) suggested the latter structures were not part of the nuclear membrane but were distinct nucleoplasmic structures (e.g. nucleoli). Finally, the ratio of average pixel intensity for BODIPY-C12 in nucleus/cytoplasm increased from 0.3 at initial time points to 0.5 by 20 min (Fig. 6C), about 2.3-fold higher  a Cells were incubated for 5 or 30 min with the respective fluorescent fatty acids (1 M in PBS) at room temperature; lipids were extracted and analyzed. Data were presented as mean Ϯ S.E., n ϭ 3-8. ND refers to not detected. FFA, free fatty acids; CE, cholesteryl esters; TG, triglycerides; MG, monoglycerides; DG, diglycerides; PL, phospholipids. b cis-Parinaric acid data were taken from Heyliger et al. (60).
than that of the short chain BODIPY-C5 at the same time (Fig.  4C). Because the majority (i.e. 92%) of the fluorescent BODIPY-C12 taken up appeared as unesterified fatty acid after 5 min of incubation (Table II), it was estimated that at 1 min Ͼ98% of BODIPY-C12 was unesterified. In contrast, nonfluorescent fatty acids (e.g. oleic acid) were Ͻ5% unesterified in the same time frame (57,58). This would suggest that after 1 min of incubation BODIPY-C12 detected in the nucleus (Fig. 6B) represented primarily unesterified BODIPY-C12. Although the qualitative distribution of BODIPY-C12 in L-FABP expressing cells (Fig. 5B) was similar to that of controls (Fig. 5A), quantitative analysis of many cells revealed that L-FABP expression increased the initial rate and average maximal extent for total uptake of BODIPY-C12 into L-cells by 2.5and 3-fold, respectively (Table I). Within the cell, L-FABP differentially enhanced BODIPY-C12 uptake into cytoplasm (Table I and Fig. 6A) versus nucleus (Table I and Fig. 6B) such that the relative distribution of this medium chain fatty acid in nuclei/cytoplasm was increased as much as 3.6-fold (Fig. 6C). Thus, L-FABP expression significantly enhanced uptake of medium chain BODIPY-C12 and differentially targeted BODIPY-C12 into nuclei, in marked contrast to the lack of effect of L-FABP expression on the intracellular distribution of the short chain fluorescent fatty acids (Fig. 4).
Uptake, Intracellular Distribution, and Esterification of Long Chain Fluorescent Fatty Acids (BODIPY-C16, NBD-C18, cis-Parinaric Acid)-The uptake, intracellular distribution, and esterification of synthetic, long chain, fluorescent fatty acids (BODIPY-C16 (Fig. 1A, n ϭ 16), NBD-C18 (Fig. 1B)) were examined in control L-cells. Lipid extraction and quantitation showed that by 5 min of incubation, the uptake of the long chain BODIPY-C16 was about the same as that of medium chain BODIPY-C12 (Fig. 2, solid bars). However, by 30 min of incubation 2.3-fold more BODIPY-C16 was taken up. At 5 and 30 min this represented clearance of 5.6 Ϯ 1.2 and 26.4 Ϯ 2.8% of BODIPY-C16 in the medium, respectively. This indicated that the maximal uptake for BODIPY-fatty acid was in the order BODIPY-C16 Ͼ BODIPY-C12 Ͼ Ͼ BODIPY-C5. In comparison to BODIPY-C16 and BODIPY-C12 uptake into the lipid extract, by 5 min of incubation NBD-C18 was taken up 3.4-and 3-fold more rapidly and was already maximal (Fig. 2). Consequently, by 30 min of incubation the NBD-C18 uptake was less than BODIPY-C16 and BODIPY-C12 by 40% and not significantly different from BODIPY-C12 (Fig. 2). It should be noted that the relative fluorescence intensity of NBD-C18 (as well as other NBD-fatty acids) was low compared with BODIPY fatty acids. This was due to differences in the intrinsic properties of NBD versus BODIPY labels. For example, the quantum yield of BODIPY in hydrophobic environments is near 0.9, whereas that of NBD is at least 3-fold lower (Molecular Probes Inc., product information MP 03792). Furthermore, in contrast to BODIPY the quantum yield of NBD strongly depends on the dielectric constant of the environment (59). As evidenced by clearance of 22.6 Ϯ 4.6 and 20.4 Ϯ 5.6% of NBD-C18 from the medium at 5 and 30 min of incubation, respectively, the low relative fluorescence intensity of NBD-C18 did not reflect low uptake as compared with BODIPY-C16 or BODIPY-C12.
These observations on the total cellular uptake of long chain fatty acids (BODIPY-C16 and NBD-C18) obtained by extraction and quantitation were in general supported by the LSCM of BODIPY-C16 total uptake in individual cells ( Fig. 6 and Table I). The initial rate and average maximal extent for total uptake of BODIPY-C16 in L-cells were nearly 2-fold higher than for BODIPY-C12 and about 4-fold higher than for BODIPY-C5 (Table I). The maximal extent for total uptake of NBD-C18 into L-cells was 62% lower than for BODIPY-C16, while being 30% lower than for BODIPY-C12 (Table I).
The intracellular distribution of long chain (BODIPY-C16, NBD-C18) fluorescent fatty acids determined by LSCM was basically similar to that of the medium chain BODIPY-C12 for cytoplasm but not for the nucleus. LSCM detected BODIPY-C16 and NBD-C18 localized with highest intensities in the cytoplasm and lipid droplets therein, BODIPY-C16 (Fig. 5A) and NBD-C18 (Fig. 5C). The maximal average pixel intensity of BODIPY-C16 (Fig. 6D) and NBD-C18 (Fig. 6G) in cytoplasmic areas was achieved by 5 min. However, differences in uptake kinetics, relative quantum yields, and/or the environment wherein they were localized resulted in the maximal average pixel intensity of the BODIPY-C16 being 2.5-fold greater than that of NBD-C18 in the cytoplasmic region (Table I). Nevertheless, both were present at nearly 5-and 2-fold higher levels in cytoplasm as compared with the short chain BODIPY-C5 (Table I) and NBD-C6 (not shown).
With regards to nuclei, BODIPY-C16 and NBD-C18 accumulation and distribution were also dependent of the specific long chain fatty acid. The distribution patterns of the long chain fluorescent fatty acids (BODIPY-C16, NBD-C18) differed significantly. For example, BODIPY-C16 was localized in the nuclear envelope region as well as being diffusely distributed within the nucleoplasm (Fig. 5A), whereas NBD-C18 was present in the nuclear envelope region as well as in areas of high intensity within the nucleoplasm (Fig. 5C). BODIPY-C16 and NBD-C18 accumulated to average maximal levels in nuclei that were about 20% higher and 50% lower, respectively, than for the medium chain BODIPY-C12. Nevertheless, both long chain fatty acids (BODIPY-C16, NBD-C18) achieved average maximal levels in nuclei that were 4.5-and 1.8-fold higher than the short chain BODIPY-C5 (Table I) and NBD-C6 (not shown).
The maximal average pixel intensity of BODIPY-C16 (Fig. 6E) and NBD-C18 (Fig. 6H) in nuclear areas was obtained by 5 min for both of these fluorescent fatty acids. Because the relative intensities of the respective probes varied over a similar range in cytoplasm and nuclei, the ratio of the average pixel intensi-  cytoplasm (A, D, and G) and nucleus (B, E, and H) are the average relative intensities/unit area, whereas the nuclei/cytoplasm ratio (C, F, and I) represents the (average relative intensity/unit area in nucleus)/(average relative intensity/unit area in cytoplasm). ties/unit area of these fluorescent long chain fatty acids in the nucleus/cytoplasm was very similar, near 0.4.
The fluorescent LCFAs were very poorly (about 2% for BODIPY-C16) or not (NBD-C18) esterified into lipids after incubation of each fatty acid with L-cells for 5-30 min (Table  II). At 5 min of incubation only 1.9% of BODIPY-C16 was esterified, appearing exclusively in phospholipids, while at 30 min incubation 0.3 and 2.8% of BODIPY-C16 appeared as fatty acyl-CoAs and phospholipids (primarily into the phosphatidylcholine class).
In summary, the synthetic fluorescent LCFAs were taken up most extensively and distributed in cytoplasm (highly so in bright structures resembling lipid droplets) as well as less so in nuclei. The maximal uptake of these LCFAs into both the cytoplasm and the nucleus was in the order BODIPY-C16 Ͼ BODIPY-C12, NBD-C18 (Table I). These LCFAs were very poorly (ϳ2% for BODIPY-C16) or not (NBD-stearic acid) esterified in the time frame of the LSCM imaging experiments. The distribution of the synthetic fatty acids to the nucleus and cytoplasm accurately reflected that of naturally occurring fatty acids, because cis-parinaric acid (a naturally occurring fluorescent fatty acid) was also 97% unesterified (Table II) and MPLSM imaging showed its distribution to cytoplasm and nucleus of L-cells to be similar to the synthetic fluorescent LCFAs (not shown). Thus, these data for the first time detected the rapid appearance of unesterified LCFAs in the nucleus of living cells.

L-FABP Expression Stimulates Uptake and Nuclear Distribution of Long Chain Fluorescent Fatty Acids (BODIPY-C16, NBD-C18, cis-Parinaric Acid)-
To determine whether expression of L-FABP enhanced the uptake and/or increased the distribution of long chain fatty acid to the nucleus, the above experiments were repeated with L-FABP expressing cells. Although the two long chain fluorescent fatty acids BODIPY-C16 and NBD-C18 differed somewhat in their basal uptake kinetics, LSCM imaging showed that L-FABP expression enhanced the initial rate of BODIPY-C16 and NBD-C18 by 1.5-and 3.4-fold, respectively (Table I). Likewise, L-FABP expression increased the average maximal intensities by 1.5-and 3.1-fold, respectively (Table I).
Whereas the overall qualitative pattern of BODIPY-C16 (Fig. 5B) and NBD-C18 (Fig. 5D) distribution in the cytoplasm of L-FABP expressing cells was basically similar to the corresponding controls, quantitative analysis revealed important differences. L-FABP expression significantly increased the initial rate of BODIPY-C16 and NBD-C18 uptake by 1.5-and 3.1-fold, respectively (Table I). The time course of BODIPY-C16 (Fig. 6D) and NBD-C18 (Fig. 6G) uptake into cytoplasm of L-FABP expressing cells did not achieve maximal values until 15 min, at least 3-fold longer than for control cells. Nevertheless, at that time L-FABP expression increased the average maximal uptake into cytoplasm by 1.4-and 2.6-fold, respectively (Table I).
L-FABP expression also increased BODIPY-C16 and NBD-C18 incorporation into the nucleus. The initial rate of BODIPY-C16 and NBD-C18 uptake was increased by 1.6-and 1.4-fold, respectively, in L-FABP expressing cells. Furthermore, L-FABP also altered the time course for BODIPY-C16 (Fig. 6E) and NBD-C18 (Fig. 6H) uptake into nuclei such that maximal uptake was not achieved until 10 -15 min, 3-fold longer than for control cells. However, L-FABP increased the average maximal uptake of BODIPY-C16 and NBD-C18 into nuclei by 1.7and 4.4-fold, respectively (Table I). Finally, L-FABP expression increased the relative distribution of BODIPY-C16 and NBD-C18 into the nucleus as compared with cytoplasm. The significantly higher ratio of average pixel intensities/unit area for BODIPY-C16 (Fig. 6F) and NBD-C18 (Fig. 6I) in the nucleus/ cytoplasm in L-FABP expressing cells was detectable at the earliest time points examined (1 and 5 min, respectively), and at later time points L-FABP expression increased the distribution of the BODIPY-C16 and NBD-C18 as much as 2.1-and 1.9-fold higher into nuclei as compared with cytoplasm.
L-FABP expression also increased the initial rates and maximal extent of the naturally occurring cis-parinaric acid total uptake into cells measured by 1.2-and 1.6-fold, respectively (Table I). In cytoplasm, low levels of cis-parinaric acid were distributed diffusely as well as in bright structures resembling lipid droplets (not shown). L-FABP expression increased the initial rate and average maximal extent of cis-parinaric acid incorporation into cytoplasm by 1.6-and 1.6-fold, respectively. (Table I). Within nuclei cis-parinaric acid was present also in low amounts and distributed in a similar pattern in both control and L-FABP expressing L-cells (not shown). L-FABP expression also increased the maximal extent of cis-parinaric acid uptake into nuclei by 1.8-fold (Table I). At time points Ͼ10 min, L-FABP expression significantly increased the ratio of average pixel intensity per unit area in nuclei/cytoplasm by as much as 1.6-fold (not shown).
Thus, L-FABP expression significantly enhanced the total uptake of both synthetic (BODIPY-C16, NBD-C18) and naturally occurring (cis-parinaric acid) long chain fluorescent fatty acids. Uptake was increased not only into cytoplasm but also into nuclei. Finally, L-FABP expression enhanced preferential targeting of these long chain fatty acids to nuclei. This preferential targeting reflected unesterified fatty acid, not only for BODIPY-C16 (98% unesterified) and cis-parinaric acid (97% unesterified), but especially for NBD-C18 (no detectable esterification during the time frame of the experiment). Furthermore, cis-parinaric acid is an 18-carbon chain length, polyunsaturated LCFA whose cis-double bond orientation at C-9 makes its structure closely similar to other naturally occurring non-fluorescent fatty acids (Fig. 1). Thus, these data show for the first time that L-FABP expression enhanced the uptake and targeting of unsaturated (cis-parinaric acid) as well as saturated (BODIPY-C12, BODIPY-C16, NBD-C18) LCFAs into nuclei of living cells. DISCUSSION A fundamental problem in understanding the potential role of LCFAs as natural ligands that activate the transcriptional activity of PPAR␣ in the nucleus is our lack of knowledge regarding the presence of unesterified LCFAs in the nuclei of living cells. Almost nothing is known regarding potential factors regulating the distribution and targeting of unesterified LCFAs to the nucleus. The data presented herein make several new contributions addressing these issues.
Second, the fluorescent fatty acids were taken up in an acyl chain length-dependent manner similar to that exhibited by non-fluorescent fatty acids, long chain (BODIPY-C16, NBD-C18) Ͼ medium chain (BODIPY-C12) Ͼ Ͼ short chain (BODIPY-C5, NBD-C6). Cells take up long and medium chain, but not short chain, fatty acids by specific plasma membrane transporters (19,49) and are bound by cytoplasmic proteins such as L-FABP (50 -53). Third, imaging microscopy (LSCM and MPLSM) showed that poorly or non-metabolizable (i.e. unesterified) fluorescent LCFAs were rapidly (1-5 min) taken up and distributed primarily into the cytoplasm and/or cytoplasmic structures (e.g. lipid droplets) of living L-cells. The high distribution to the fluorescent LCFAs to cytoplasm reflected the predominant localization of soluble proteins that bind LCFAs, i.e. FABPs (42,44,63). Mock-transfected control L-cells contain an endogenous FABP immunologically distinct from L-FABP (45). Finally, the high distribution of fluorescent LCFAs to lipid droplets in the cytoplasm reflects the fact that a lipid droplet-specific protein, adipocyte differentiation related protein, has high affinity for fatty acids such as NBD-C18 (47,64,65). Thus, the intracellular distribution of the poorly or non-metabolizable fluorescent fatty acids primarily to cytoplasm and lipid droplets reflects that of proteins that bind fatty acids.
Fourth, real time imaging (LSCM and MPLSM) revealed for the first time that both saturated (i.e. BODIPY-or NBDtagged) and unsaturated (cis-parinaric acid) fluorescent unesterified LCFAs rapidly (1-5 min) distributed not only into the cytoplasm but also to nuclei, especially the nuclear envelope, of living cells. In contrast, short chain (BODIPY-C5, NBD-C6) fatty acids were diffusely localized in both the nucleoplasm as well as cytoplasm but not the nuclear envelope. Finally, the nuclear/cytoplasmic ratio of average pixel intensities for these fluorescent fatty acids was in the range of 0.4 -0.5, indicating that 28 -33% these fatty acids appeared localized to nuclei. Thus, LSCM and MPLSM of fluorescent fatty acids in living cells for the first time supported earlier suggestions regarding the presence of unesterified LCFAs in nuclei (isolated by subcellular fractionation) were not simply due to artifacts of nuclear isolation or hydrolytic cleavage of esterified lipids to release LCFAs (23,27,66,67).
Fifth, it was striking that, despite the detection of nearly one-third of the fluorescent fatty acids appearing in the nucleus, much less appeared in the nucleoplasm. For example, in a separate experiment the proportion of cellular BODIPY-C16 in the nucleus of control and L-FABP expressing cells was estimated to be about 22.3 Ϯ 4.2% (n ϭ 9) and about 46.9 Ϯ 2.8% (n ϭ 7), respectively, representing a 2.1-fold increase in the L-FABP expressing cells (p Ͻ 0.001). However, when the proportion of cellular BODIPY-C16 in the nucleoplasm was estimated (by counting only those pixels within the nucleus and not in the nuclear membrane), only about 7.4 Ϯ 1.5% (n ϭ 9) of BODIPY-C16 was in the nucleoplasm of control L-cells. In contrast, about 17.3 Ϯ 1.8% (n ϭ 7) of BODIPY-C16 was in the nucleoplasm of L-FABP expressing cells, 2.3-fold higher (p Ͻ 0.001). It is important to note that z series scans showed that L-cell nuclei were 6-m thick. To avoid fluorescence from the cytoplasm over the nucleus, all confocal (BODIPY-and NBD-tagged fatty acids) images were taken from slices (0.3 m) through the median section of the nuclei. In multiphoton laser scanning microscopy (cis-parinaric acid) only those fatty acids in the focal volume (0.01 cu ) within the median of the nuclei were excited, and images were collected therefrom. The presence of fluorescent fatty acids in the nucleus was also verified by colocalizing with a DNA-binding stain (not shown).
Sixth, L-FABP expression in L-cell fibroblasts significantly (2.9 -4.4-fold) and preferentially (2-3.6-fold) increased the total uptake of the medium and long chain fluorescent fatty acids into the nuclei of living L-cells. These were L-FABP-specific because of the following. (i) L-FABP expression (0.4% of cytosolic protein) increased the total FABP binding capacity by nearly 4-fold (45). (ii) Although L-cells contain low levels of endogenous FABP (0.12% of cytosolic protein) that does not cross-react with antisera to L-FABP, L-FABP expression did not alter endogenous FABP expression in the transfected cells (45). (iii) As shown, L-FABP expression did not alter the nuclear content of short chain fatty acids (BODIPY-C5, NBD-C6). This was consistent with earlier findings demonstrating that L-FABP binds medium and long chain fatty acids but does not bind short chain fatty acids (51)(52)(53). Although it had been proposed earlier (28) that L-FABP might be involved in trafficking of unesterified LCFAs to nuclei, data show increased LCFAs in the nucleus were lacking. Thus, the results presented herein provide the first direct experimental evidence in support of this hypothesis. Whereas the mechanism whereby L-FABP enhances the trafficking and targeting of medium and long chain fatty acids to the nucleus is not yet completely clear, the data presented herein and in the literature suggest the following. (i) LSCM and MPLSM showed that L-FABP expression enhanced severalfold the total cellular uptake of nonmetabolizable (NBD-C18) and poorly metabolizable (especially BODIPY-C16 and cis-parinaric) fluorescent fatty acids in real time in living cells, consistent with radiolabeled (30,32) and stirred cuvette fluorescence assays (31)(32)(33). By L-FABP increasing total fatty acid uptake, it is likely that more LCFA is available for spontaneous transfer to the nucleus simply by mass action. (ii) Once exogenous LCFAs have entered and translocated across the cell surface membrane, it is thought that intracellular L-FABP may enhance the desorption of LCFAs into the cytoplasm (19). In vitro studies show that L-FABP alters the partitioning of LCFAs from membranes toward the aqueous (68). (iii) It is postulated that intracellular FABPs enhance LCFA diffusion through the cytoplasm (37). Indeed, L-FABP expression enhances NBD-C18 intracellular diffusion in L-cells and hepatocytes by nearly 2-fold (34,34,35,69). (iv) L-FABP enhanced specific targeting of medium and long chain, saturated and unsaturated fluorescent fatty acids to the nuclei by increasing the nucleus/ cytoplasmic ratio by 2-3.6-fold. Whereas processes i-iii enhance transfer of LCFAs to the nucleus, they would not account for L-FABP specifically targeting LCFAs to the nuclei to increase the ratio of average maximal intensities of medium and long chain fluorescent fatty acids in nuclei/ cytoplasm. This implies a more specific interaction of L-FABP with the nucleus. Indeed, L-FABP interacts directly with nuclei to bind a 33-kDa protein in a ligand-dependent manner (70). Although unesterified LCFAs only bind nonspecifically to isolated nuclei, in the presence of L-FABP the unesterified oleic acid is cotransported to isolated nuclei (38). Cotransport may be facilitated by the fact that LCFA binding alters L-FABP conformation (70).
The physiological significance of these findings is that for the first time, the intracellular events of unesterified fatty acid uptake to the nucleus were visualized in real time in living cells. The demonstration that the exogenously derived fluorescent LCFAs accumulated in nuclei of intact, living cells to concentrations only 3-4-fold lower than outside the nucleus was consistent with earlier data showing that nearly 20% of plasma-derived unesterified LCFAs appear in nuclei isolated from hepatocytes (23)(24)(25)(26) and in some studies account for as much as 40% of the lipids in isolated nuclei (23,27,66,67). The imaging data presented herein, however, indicate that most of the fatty acid in nuclei was not present in the nucleoplasm but rather in the nuclear envelope, a location close to transcriptional machinery underlying the nuclear envelope. PPAR␣ has high affinity for LCFAs (K d ϭ 5-17 nM) (13), thereby indicating that the very low LCFA concentration in the nucleoplasm is physiologically significant. The present data showing that L-FABP expression enhanced LCFA uptake and targeting to the nucleus (increased ratio of LCFA average maximum fluorescence in nuclei/ cytoplasm) is especially significant because L-FABP binds the LCFAs (K d ϭ 41-400 nM (71)) with 5-50-fold lower affinity than does PPAR␣ (K d ϭ 5-17 nM (13)). Because low amounts of L-FABP are present within nuclei (hepatocytes (28, 40 -43); transfected embryonic stem cells expressing L-FABP (44)) and L-FABP interacts directly with PPAR␣ (28), it may be postulated that L-FABP serves to shuttle fatty acids to and/or into the nucleus for donating fatty acids to PPAR␣. The fatty acid shuttling function of L-FABP could be similar to that of CRBP, which shares structural homology with L-FABP (70). It is important to note that L-FABP may also provide a specific pool of fatty acids for nuclear utilization to modify the activity of certain nuclear enzymes, including DNA nucleotidase and DNA-dependent RNA polymerase (38) as well as DNA polymerases (72) and DNA topoisomerase II (73,74). Because L-FABP expression is regulated by fatty acid-induced PPAR␣ transcriptional activity (75), L-FABP may also regulate its own transcription by enhancing LCFA targeting to the nucleus. Thus, by altering the distribution of intracellular LCFAs to the nucleus, L-FABP may affect multiple aspects of gene transcription therein.
In summary, the data presented herein show for the first time that unesterified LCFAs are present in the nucleus of intact, living cells and that L-FABP expression significantly enhances the uptake and distribution of medium and long chain fatty acids to the nucleus. These data suggest that by regulating the distribution of fatty acids to the nucleus, L-FABP may have modulated the activities of LCFA-sensitive enzymes and transcription factors therein.