INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

IN MAMMALIAN SPECIES, de novo synthesis of fatty acids in the liver is stimulated by high-carbohydrate (CHO), low-fat diets, especially if the animals are first fasted for an extended period of time (12). Monosaccharides, particularly fructose, have shown a greater lipogenic potential than complex CHO, the prolonged feeding of which results in obesity and/or hypertriglyceridemia (2, 15, 25). Diet-induced lipogenesis is caused primarily by induction of hepatic lipogenic enzymes, especially the rate-limiting enzyme, fatty acid synthase (FAS) (12, 14, 19). The primary mechanism for FAS induction is transcriptional activation increasing FAS mRNA abundance in the liver (14, 18, 19, 26), although an increase in mRNA stability may also play a role (29). FAS transcription is stimulated by insulin and thyroid hormone (3,5,3'-triiodothyronine) and inhibited by glucagon and catecholamines (16, 19, 26). However, fructose ingestion, which causes a smaller insulin response than does glucose or complex CHO, results in a higher FAS induction (3). Furthermore, fructose feeding has been shown to upregulate FAS in diabetic rats, indicating that mechanisms other than insulin may play a role (19).

The insulin responsive sequence (IRS) located at position 71/50 of the FAS promoter has been shown to play an important role in hepatic FAS regulation by insulin at a physiological concentration (23). The upstream stimulatory factors (USF; USF1, 43 kD and USF2, 44 kD), members of the ubiquitous basic helix-loop-helix transcription factor family, may bind to the E-box motif (CANNTG) found in the promoters of FAS and other lipogenic enzyme DNA to confer the transcriptional regulation (10, 21, 31, 33). The critical role of IRS in the hormonal and dietary induction of FAS has been recently confirmed with a transgenic mice model (32). In addition to IRS, CHO response elements (ChoRE) located in the first intron (+283/303) of the FAS gene, as well as the promoters of several other lipogenic enzymes, also contain E-box sequences that have been shown to interact with USFs and confer glucose responsiveness (10, 30). However, little is known about its significance in FAS regulation in response to nutritional or physiological interventions in vivo.

Endurance exercise is known to decrease plasma insulin and increase glucagon and catecholamine levels, all of which promote CHO and fat utilization and inhibit lipogenesis (6, 20). Using a fasting-refeeding rat model, we have previously demonstrated that an acute bout of prolonged exercise can suppress hepatic FAS activity and mRNA abundance by 50-70% (14). These changes were accompanied by decreased plasma insulin and elevated glucagon levels in the exercised rats. Endurance trained rats meal fed the same diet showed as much as a 50% reduction in FAS activity (9). A greater reduction in body fat was found in rats fed a high-fructose diet (25% decrease) than a high-fat diet (12% decrease) (13). Thus we hypothesize that physical exercise may downregulate CHO-induced FAS at the gene level. The purpose of this study was to examine the role of nuclear protein binding to IRS and ChoRE and the impact of these bindings to FAS expression in response to diet and exercise.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Animal care and tissue preparation. Eighteen female Sprague-Dawley rats were randomly divided into three groups. One group (n = 6) was fasted for 48 h and killed at rest (F). The second and third groups were fasted (48 h), refed a high-fructose diet for 6 h, and killed either after an acute bout of exercise (E) or at rest (R). Compositions of this diet containing 50% fructose, 20% protein, and 5% fat have been reported (14). Previous work has indicated that maximal FAS mRNA transcription occurs after 4-6 h of refeeding (19). Exercise was performed on a rodent treadmill at 27 m/min and 5% grade (~60% maximal O₂ consumption) until exhaustion, defined as the time when a rat was unable to right itself after being lain on its side. Endurance time of the E rats was 88 ± 7 min. R rats rested without food for the same amount of time before being killed. The experiments were scheduled such that an E rat and an R rat were always killed sequentially with no more than a 15-min interval. After the rats were killed by decapitation, the abdominal cavity was immediately opened, and the liver was quickly excised and frozen in liquid N₂. The liver samples were either stored in liquid N₂ or at 80°C until processing and assay.

Gel mobility shift assays. Liver nuclear extracts (NE) were prepared according to the method of Dignam et al. (7) with modifications by Andrews and Faller (1). The following single-stranded oligonucleotides were purchased from GIBCO Life Technologies (Gaithersburg, MD): FAS-IRS-A (71/50): 5'-TCAGCCCATGTGGCGTGGCCGC-3', 3'-AGTCGGGTACACCGCACCGGCG-5'; FAS-ChoRE (+283/+303): 5'-GGCCGCTGTCACGTGGGCGCC-3', 3'-CGGCGACAGTGCACCCGCGG-5'; liver-type pyruvate kinase (LPK)-ChoRE (172/141): 5'-ATGGGCGCACGGGGCACTCCCGTGGTTCCTAC-3', 5'-TACCCGCGTGCCCCGTGAGGGCACCAAGGATG-5'; S14-ChoRE (1443/1423): 5'-GCCAGTTCTCACGTGGTGGCC-3', 5'-CGGTCAAGAGTGCACCACCGG-3'; nuclear factor (NF)-B (consensus sequence): 5'-AGTTGAGGGGACTTTCCCAGGC-3', 3'-TCAACTCCCCTGAAAGGGTCCG-5'.

Ultraviolet (UV) cross-linking. After the gel shift experiments proceeded for 30 min as described above, the reaction mixture was exposed to UV light in an UV Cross-Linker (Strategene, La Jolla, CA) for 20 min. Thereafter, the reactions were subjected to DNase I digestion, thus degrading unbound regions of the oligonucleotide. An equal volume of denaturing loading buffer (0.5 M Tris · HCl, pH 6.6-6.8, 10% SDS, 10% glycerol, 5% -mercaptoethanol) was added to the reaction followed by 1 min of boiling. SDS-PAGE was used to determine the approximate size of the DNA-bound proteins. The gel was fixed with 10% acetic acid-10% methanol and then exposed to X-ray films with an intensifying screen.

Northern blot. Total RNA was isolated from frozen livers by the method of Chomczynski and Sacchi (4) with Trizol reagent (GIBCO Life Technologies) as previously described (14). The cDNA probes for FAS and 18S were labeled by using random primer extension (8) with a labeling kit that used [-³²P]dCTP (Megaprime, Amersham, Arlington Heights, IL). Hybridization solution consisted of dextran sulfate added to a level of 10%, and radiolabeled probes were added at a level of 10⁶ counts · min¹ · ml¹ and allowed to hybridize overnight. The stringency washes consisted of two 20-min washes with 1× saline sodium citrate (SSC) and 0.5% SDS at 45°C, two 20-min washes with 0.5× SSC and 0.5% SDS at 50°C, and one 20-min wash with 0.1× SSC and 0.5% SDS at 60°C. After autoradiography, the probe was removed from the filters with a solution of 50% formamide and 2× sodium chloride-sodium phosphate-EDTA at 65°C for 60 min. Quantification of the FAS signals was achieved by use of a scanning densitometer (model GS-670, Bio-Rad, Richmond, CA). FAS mRNA abundance was expressed relative to the density of the respective 18S values.

Nuclear run-on assay. A nuclear run-on assay was performed according to the method described by Paulauskis and Sul (26). Liver nuclei were isolated via the sucrose gradient technique cited above except that nuclei were left intact, frozen on dry ice, and stored at 80°C until assay. Run-on transcription was performed by incubating nuclei with ³²P-labeled UTP and unlabeled NTPs in a solution containing 25% glycerol, 2.5 mM MgCl₂, 0.05 mM EDTA, 75 mM HEPES, pH 7.5, 100 mM KCl, 4 mM DTT, 0.04 mg/ml creatine kinase, and 8.8 mM creatine phosphate. Labeled nascent transcripts were incubated with RNase-free DNase for 10 min at 37°C in the presence of 5 mM MgCl₂, followed by digestion with 90 µg/ml proteinase K, 0.5% SDS, and 5 mM EDTA at 37°C. Lipogenic enzyme cDNAs were immobilized on nylon membranes (23).

Enzyme activity. Maximal activity of FAS was measured in liver cytosol according to Nepokroeff et al. (24) as previously described. Protein content was determined by the Bradford method with BSA as the standard.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Specificity of nuclear protein binding to IRS and ChoRE. Two major DNA-protein complexes were detected with labeled FAS-IRS by gel mobility shift assay, along with several minor bands (Fig. 1, left). The major complexes were competed away with increasing concentrations (5-, 10-, and 50-fold) of unlabeled IRS oligonucleotides (Fig. 1, left, lanes 2-4). However, adding equal concentrations of a consensus NF-B oligonucleotide probe did not affect binding to the labeled IRS (Fig. 1, right), indicating that the bands detected were specific. Gel mobility shift assays with FAS-ChoRE revealed a single major DNA-protein complex (Fig. 2, left). This band was effectively competed away with excess unlabeled ChoRE oligonucleotides for LPK and S14 (Fig. 2, middle), but not with NF-B. It was interesting to find that much higher concentrations were required for LPK-ChoRE and S14-ChoRE to effectively compete with FAS-ChoRE. Similarly, unlabeled FAS-IRS could compete with labeled FAS-ChoRE, but a 50-fold excess was required (not shown). These results indicate that, although both FAS-IRS and FAS-ChoRE contain an E-box region and have similar sequences, the binding complexes are not identical and protein-DNA binding involves more than the E-box region.

View larger version (68K): [in this window] [in a new window]   Fig. 1.   Gel mobility shift assay with fatty acid synthase (FAS) insulin response sequence (IRS). Each reaction contained 15 µg of liver nuclear extract protein, 1× gel shift reaction buffer (see EXPERIMENTAL PROCEDURES), 0.5 µg poly(dI-dC), and ~0.1 ng of 32P-labeled FAS-IRS probe (20,000 counts/min) with total volume of 30 µl. Lanes 1-5: addition of 0, 5-, 10-, 50-, and 100-fold concentration of unlabeled FAS-IRS or nuclear factor-B (NF-B) probes.

View larger version (68K): [in this window] [in a new window]   Fig. 2.   Gel mobility shift assay with FAS carbohydrate response element (ChoRE). Assay conditions are same as in Fig. 1, except cold ChoRE for liver-type pyruvate kinase (LPK) and S14 oligonucleotides were included. Lanes 1-6: addition of 0, 5-, 10-, 50-, 100-, and 500-fold concentration of unlabeled ChoRE or NF-B probes.

FAS-IRS and FAS-ChoRE bindings affected by diet and exercise. To determine whether the binding of transcription factors to FAS-IRS and ChoRE was affected by nutritional and metabolic factors, liver NE prepared from F, R, and E rats were analyzed with gel mobility shift assays. As shown in Fig. 3, refeeding previously fasted rats resulted in a marked increase in the relative binding for both FAS-IRS (Fig. 3A, left) and FAS-ChoRE (right). When the samples were analyzed individually, mean binding intensity was increased by 90% (P < 0.05) and 60% (P < 0.05) with refeeding for FAS-IRS and FAS-ChoRE, respectively (Fig. 3B). However, the refeeding-induced nuclear protein binding was completely abolished in E rats vs. R rats with both oligonucleotides.

View larger version (34K): [in this window] [in a new window]   Fig. 3.   Gel mobility shift assay (A) and relative binding (B) of IRS and ChoRE in liver nuclear extracts. F, rats fasted for 48 h; R, rats fasted and then refed high-fructose diet for 6 h; E, rats fasted, refed, and then exercised (Ex) to exhaustion. Data are expressed as means ± SE for n = 6. * Significantly different from F and E conditions, P < 0.05.

Major nuclear protein binding complexes containing USF1 and/or USF2. Both FAS-IRS and FAS-ChoRE are known to contain E-box regions that can bind USF (30, 31, 33). To determine whether the DNA-protein complexes observed in the gel mobility shift assays contain USF1 and/or USF2, gel mobility supershift assays were performed with antibodies to USF1 and USF2. As shown in Fig. 4A, left, the addition of anti-USF1 resulted in a supershift of the major band of FAS-IRS without affecting the second band (lane 2). Similarly, the antibody to USF2 shifted the major band for IRS (lane 3). The addition of both USF1 and USF2 antibodies simultaneously produced a further supershift of the supershifted bands (lanes 4 and 5). Furthermore, adding antibody to the reaction with labeled FAS-ChoRE resulted in similarly supershifted bands (Fig. 4A, right). These results indicate that USF1 and USF2 are constituents of the DNA-binding complexes of both FAS-IRS and FAS-ChoRE as shown previously (30, 31, 33).

View larger version (49K): [in this window] [in a new window]   Fig. 4.   A: gel supershift assay with IRS and ChoRE probes in F and R rats. Each reaction (30 µl) contained 15 µg of live nuclear extracts, 1× gel shift reaction buffer, 0.1 µg poly(dI-dC), ~0.1 ng 32P-labeled oligonucleotide probe, and 1 µg anti-upstream stimulatory factor 1 (USF1) and/or 1 µg anti-USF2 antibodies. B: gel supershift assay with IRS and anti-USF1 in various treatment groups. + and , With and without specified antibody.

Different DNA-binding proteins binding FAS-IRS. To further characterize the DNA-binding proteins interacting with the gene regulatory sequences, we performed UV cross-linking experiments followed by denaturing SDS-PAGE to separate the bound proteins according to size in the FAS-IRS binding experiments. Two visible bands were found corresponding to 42- to 44-kDa and 70-kDa proteins, corrected for probe size (Fig. 5, lane 1). Competition with unlabeled oligonucleotides FAS-IRS, LPK-ChoRE, and S14-ChoRE in 10- to 100-fold molar excess almost eliminated both bands (lanes 2-7); however, a 100-fold excess NF-B could compete away only the 70-kDa protein, but the 42- to 44-kDa protein largely remained (lanes 8 and 9). These protein bands were largely absent in F rats but clearly visible in R and E rats (data not shown). Thus refeeding caused FAS-IRS to bind two liver nuclear proteins, of which the 42- to 44-kDa protein was specific to IRS binding. Furthermore, LPK-ChoRE and S14-ChoRE seemed to share some of the binding characteristics of FAS-IRS.

View larger version (72K): [in this window] [in a new window]   Fig. 5.   Ultraviolet cross-linking of nuclear extracts to 32P-labeled FAS-IRS. Each reaction (20 µl) contained 8 µg of protein, 1× gel shift reaction buffer, 0.4 µg poly(dI-dC), 32P-labeled oligonucleotide probe (20,000 counts/min), and cold nucleotide competitors at 10- or 100-fold concentration.

Exercise-attenuated transcription rate of FAS gene. To examine whether the observed changes in IRS and ChoRE binding with nuclear proteins could directly influence FAS mRNA transcription rate, we performed nuclear run-on experiments. Liver NE from F, R, or E rats were pooled (n = 3-4), and experiments were repeated three times. As shown in Fig. 6, the transcription rate was increased by 49-fold in R vs. F rats, which agrees with data reported by Paulauskis and Sul (26). Exercise suppressed the transcription rate by 78% in E vs. R rats. The changes in FAS gene transcription were specific as the transcription rate of -actin was unaltered by treatments (not shown).

View larger version (18K): [in this window] [in a new window]   Fig. 6.   Nuclear run-on assay. Liver nuclei were incubated with 32P-labeled UTP and unlabeled NTPs in solution containing 25% glycerol, 2.5 mM MgCl2, 0.05 mM EDTA, 75 mM HEPES (pH 7.5), 100 mM KCl, 4 mM dithiothreitol, 0.04 mg/ml creatine kinase, and 8.8 mM creatine phosphate for 10 min at 37°C. Each bar represents pooled liver samples from 3-4 rats randomly selected from each group and repeated 3 times.

Dietary and exercise effects on FAS mRNA abundance and FAS activity. Major transcripts for FAS mRNA detected with Northern blot were consistent with those previously reported (5, 9, 14, 26) (Fig. 7A, top). FAS mRNA abundance was nondetectable in the F rats but showed a dramatic increase in R rats (Fig. 7B). FAS mRNA levels appeared lower in E vs. R rats, but the change was not statistically significant. FAS activity was 170% higher (P < 0.05) in R vs. F rats (Fig. 7C). No significant difference in FAS activity was detected between E and R rats.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Northern blot (A), relative intensity of FAS mRNA signals using 18S as an internal standard (B), and FAS activity (C) in rat liver. Data are expressed as means ± SE for n = 6. ND, nondetectable. * P < 0.05, R vs. F.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Regulation of FAS occurs primarily at the transcriptional level, possibly mediated by cis-acting sequences of the 5'-flanking region of the FAS gene (16). Both IRS and ChoRE have been identified within the FAS gene, conferring insulin and CHO induction of FAS, respectively, although the role of the form gene regulatory sequence is much better established (10, 23, 30, 33). The major finding of the present study was that an acute bout of exhaustive exercise almost completely abolished the fructose-induced transcription factor binding to IRS and ChoRE in rat liver NE. The supershift assays demonstrated that both USF1 and USF2 were components of the major complexes resulting from FAS-IRS and FAS-ChoRE binding. Our data were consistent with the findings that both IRS and ChoRE contain an E-box to which USFs are known to bind (31) and that enhanced USF binding to IRS may account for the refeeding-induced FAS in rat liver (33). Recent work by Soncini et al. (32) using transgenic mice provided further evidence that IRS plays a functional role in FAS gene regulation by diet and hormones. Our present data showed that the FAS gene transcription rate was increased by nearly 50-fold from the fasted to the refed state, whereas this increase was severely suppressed by an acute bout of exercise (Fig. 6). Although a causal relationship cannot be established at present, we postulate that the altered nuclear protein binding to FAS-IRS might be a primary reason for the observed FAS upregulation with refeeding and downregulation with exercise. However, characteristics of the binding protein for IRS and ChoRE are still largely unknown. Among the two major complexes found in the FAS-IRS gel shift experiment, only one supershifted with the addition of USF antibodies (Fig. 4). When the proteins were UV cross-linked and separated according to size with SDS-PAGE, we observed that the size of one major band was 42-44 kDa, consistent with the molecular weight of the USFs. Another larger bound protein (~70 kDa) was probably nonspecific as it disappeared in the gel mobility assays with excessive NF-B followed by UV cross-link and SDS-PAGE (Fig. 5).

The mechanism responsible for the exercise attenuation of IRS and ChoRE binding is still elusive. Prolonged exercise decreases plasma insulin and increases plasma glucagon and catecholamine concentration, but how the changed hormonal milieu results in decreased liver nuclear protein binding is largely unknown. There are several possibilities. One possibility is that prolonged exercise caused enhanced hepatic proteolysis, thereby decreasing the availability of USFs and other nuclear proteins to bind FAS-IRS. Although we did not perform quantitative assessment on USFs in the present study, such a mechanism appears plausible, because liver weight and protein content were indeed decreased after an exhaustive exercise bout (14). A second possibility is that elevated plasma glucagon during exercise increased liver cAMP levels that inhibited FAS induction via a cAMP-adenylate cyclase cascade (12, 14). Many lipogenic enzyme genes, including FAS, possess cAMP response elements, which are recognized by specific phosphoprotein transcription factors (11, 27). Enhanced cAMP response element binding might inhibit IRS and ChoRE binding by transcriptional factors, such as USFs. Finally, exercise may result in increased phosphorylation of certain nuclear proteins, thereby affecting their affinity with cis-acting sequences on certain genes. USF binding to IRS and/or ChoRE may possibly be attenuated due to an altered phosphorylation state during exercise.

Despite decreased protein binding to IRS-ChoRE and the transcription rate of FAS mRNA in the liver NE, E rats showed no significant decrease in fructose-induced FAS mRNA levels or FAS activity. This finding was not entirely surprising because of the following reasons. 1) The 6-h refeeding time was chosen mainly to manifest altered nuclear protein binding before induction of mRNA and enzyme synthesis. Time course studies have revealed that peak mRNA levels are reached 9-16 h after the start of refeeding, whereas FAS activity is not fully expressed until after 16 h (14, 19). Indeed, FAS mRNA levels at 6 h were about one-half of those at 8 h, and FAS activity in R rats was only one-third of that in 24-h refed rats (9). Thus it is plausible that FAS mRNA abundance and enzyme activity measured at 6 h were still too low to show a significant exercise inhibition. 2) The exercise intensity used in the present study was higher than that in our previous study in which rats ran at 20 m/min and 5% grade, resulting in a much longer duration (162 and 214 min for 8- and 12-h refed rats, respectively; cf. Ref. 6). A longer exercise time could suppress insulin and mobilize glucagon secretion to a greater extent, thereby decreasing FAS mRNA and activity more effectively. 3) In addition to transcriptional control, FAS gene expression is also influenced by mRNA stability protected by high concentrations of glycolytic metabolites resulting from CHO feeding (28, 29). The longer exercise duration in our previous studies might have decreased liver glycolytic metabolites such as pyruvate and glucose 6-phosphate to a greater extent than that in the present study, resulting in lower FAS mRNA stability (14).

Exercise-induced downregulation of FAS gene expression may have significant biological implications. In addition to increasing fat oxidation, inhibition of de novo fat synthesis provides an additional metabolic pathway to reduce body fat deposit. In animals, reduced lipogenesis due to exercise ensures that energy is directed to more important metabolic functions. Although humans do not synthesize large amount of fat, which should be plentiful in normal Western diets, present dietary trends of reducing fat and increasing CHO (especially fructose) consumption have made de novo lipogenesis a valid concern (17). Newly synthesized fat is primarily saturated fatty acids that could displace polyunsaturated fatty acids and alter lipid composition in cell membranes (9). Exercise appears to offer some merit in counteracting this adverse effect.