The e subunit gene of murine F1F0-ATP synthase. Genomic sequence, chromosomal mapping, and diet regulation.

Genomic sequences encoding murine Lfm1, whose predicted protein sequence is 96% and 98% similar to bovine and rat F1F0-ATP synthase e subunits (respectively), have been amplified from BALB/cByJ DNA, cloned, and sequenced. The 1.1-kilobase gene has 3 introns and 4 exons, and its coding sequence differs by two nucleotides compared to the previously published BALB/cHnn Lfm1 cDNA sequence. A PstI restriction site polymorphism in intron 2 between C57BL/6J and Mus spretus was used to map this gene to Chromosome 5 near D5Mit9. Related sequences were mapped on Chromosomes 8, 11, and 2 unlinked loci on Chromosome 2 using Southern blot analyses with the 1.1-kilobase gene as probe. Previous studies from this laboratory indicated that the Lfm1/e subunit was regulated by the level of dietary fat and carbohydrate. Northern hybridization analyses demonstrated that e subunit mRNA abundance showed statistically significant differences (p < 0.025) between hearts of BALB/c mice fed 3% and those fed 20% corn oil for 2 weeks and in liver (p < 0.05) from the same animals. Significant differences were also observed in hepatic and heart mRNA expression at different times after eating in animals subjected to a fast/refeed regimen. The implications of the high degree of sequence similarity to the e subunit for rat and bovine F1F0-ATP synthase and its regulation by diet are discussed.

One of the ultimate outcomes of the metabolism of nutrients is the establishment of a proton gradient across the inner mitochondrial membrane. F 1 F 0 -ATP synthase, located on the inner surface of the inner membrane, converts energy derived from the gradient into ATP (reviewed in References 1 and 2). Its activity is partly regulated by the availability of substrates and the proton flux through the membrane (1,2). The synthase complex consists of a soluble F 1 catalytic unit with the F 0 portion composed of a joining stalk and inner membrane proteins (1)(2)(3)(4)(5)(6)(7). Proton flux through F 0 membrane proteins has been postulated to cause conformational changes, which may pass to the catalytic F 1 subcomplex through the stalk (2). Three proteins, designated e, f, and g, co-purify in a 1:1 stoichiometry with F 1 F 0 subunits (4 -7) but are peripherally associated with the membrane or stalk (2). Their functions are unknown. In a screen for genes regulated by the level of dietary lipids and carbohydrates (8), we isolated and characterized low fat mammary 1 (Lfm1), 1 whose predicted protein sequence is 96% and 98% similar to e subunit sequences in bovine heart (4) and rat liver (7). Since the latter two proteins are 94% similar, we suggested that Lfm1 is the murine F 1 F 0 -ATP synthase e subunit gene (8). Northern hybridization analyses revealed higher Lfm1/␤-actin mRNA ratios in livers, kidneys, and mammary glands of BALB/cHnn mice fed semipurified diets containing 3% instead of 20% corn oil. The association of the e subunit to a region of F 0 postulated to transmit conformational energy to the F 1 catalytic subunits (2) and the regulation of its mRNA abundance by concentrations of dietary constituents (see Ref. 8 and this report) suggest that the e subunit may play a role in regulating F 1 F 0 -ATP synthase activity.
We report here the sequence of a 1.1-kb e subunit gene, amplified from BALB/cByJ genomic DNA with primers derived from the Lfm1 cDNA sequence (8). The genomic clone encodes 3 introns and 4 exons with two nucleotide differences in the coding sequence compared to the published BALB/cHnn Lfm1 cDNA sequence (8). A 370-bp "spliced" pseudogene was also amplified from genomic DNA. The 1.1-kb e subunit gene was mapped to Chromosome 5 (Chr 5) near D5Mit9 by scoring a PstI RFLP in PCR-amplified (C57BL/6J ϫ Mus spretus)F1 ϫ C57BL/6J (BSB) segregants of The Jackson Laboratory Backcross DNA panel (9). Four related sequences mapped to loci on Chromosomes 11 and 8, and at two unlinked positions on Chromosome 2 by Southern hybridization analyses of TaqIdigested (C57BL/6J ϫ SPRET/Ei)F1 ϫ SPRET/Ei (BSS) segregants (9). We have also analyzed the relative abundance and temporal regulation of e subunit/glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA levels in heart tissue of BALB/cHnn mice during a fast/refeed regimen following their 2-week ad libitum consumption of diets containing 3%, 10%, and 20% corn oil. Statistically significant differences in main effects were observed between e subunit mRNA levels when there were differences in dietary corn oil concentrations and in times after eating. Similar results were obtained with liver.

MATERIALS AND METHODS
DNA Probes-DNA probes for Southern and Northern hybridization analyses were prepared by random priming (Amersham Corp.) of pLMC-1, a cDNA clone encoding BALB/cHnn Lfm1 (murine e subunit gene; Ref. 8) in Bluescript SKϩ containing 216 nucleotides of coding sequence, 39 nucleotides from the 5Ј-untranslated region (UTR), and 56 from the 3Ј-UTR (8). Primer sequences and cDNA are referred to as Lfm1, while genomic sequences and mRNA are designated as the e subunit. In some experiments, e subunit probes were prepared by PCR with [␣-32 P]dCTP (ICN, Irvine, CA). Primers were either e subunitspecific (see Fig. 2) or T7 and SP6 (Life Technologies, Inc.). PCR reaction conditions were as described below except that the final dCTP concentration was 2.5 M. The radiolabeled probes emitted at least 1-10 ϫ 108 dpm/g.
Published (8) and unpublished sequence data about the mouse cDNA Lfm1 gene were used to design 4 oligonucleotides for amplifying e subunit sequences from genomic DNA. These are shown in Fig. 2. One g of BALB/c genomic DNA was amplified with oligonucleotides complementary to ϩ36 3 ϩ61 and ϩ222 3 ϩ198 (referred to as internal primers) in 60 mM Tris-HCl (pH 8.5-9.5), 15 mM NH 4 SO 4 , 2 mM MgCl 2 (Buffers B, F, or J of the Invitrogen Optimizer kit; Invitrogen, La Jolla, CA), 0.5 mM in each dNTP in a total volume of 20 l. Reactions were heated to 95°C for 2 min followed by addition of 0.25 l of Taq DNA polymerase (Life Technologies, Inc.). The reaction conditions were 3 min at 95°C, 35 cycles of 95°C for 1 min, 1 min at 50°C or 55°C, and 3 min at 72°C. Products were directly cloned into pCRII (Invitrogen, La Jolla, CA) according to manufacturer's specifications (Invitrogen pC-RII kit). A second amplification was done with oligonucleotides complementary to Ϫ29 3 Ϫ4 and ϩ233 3 ϩ208 (external primers) of the Lfm1 cDNA sequence, and the fragments were cloned in pCRII. Cloned DNAs were sequenced with double-stranded sequencing methods with Sequenase II kits (US Biochemical Corp.) following conditions suggested by the manufacturer. GenBank (National Center for Biological Information, National Library of Medicine, NIH) was searched for similar sequences. Comparisons between published sequences were done with ALIGN and AALIGN programs (DNASTAR, Madison, WI).
Mapping of the 1.1-kb e Subunit Gene-Two hundred ng of M. spretus, C57BL/6J, and the BSB backcross panel DNAs were substrates for DNA amplification in 10 l with Buffer J as described above. The primers were Lfm1 (e subunit)-specific encoding nucleotides ϩ36 3 ϩ61 and sequences complementary to ϩ198 3 ϩ222. Following amplification, the reactions were adjusted to 50 l with 10 mM Tris-HCl (pH 8), 1 mM EDTA (TE buffer) and the DNA was precipitated with 2 volumes of cold 100% EtOH. DNA precipitates were pelleted for 10 min in a microcentrifuge, resuspended in 10 l of TE buffer, and digested with PstI (Life Technologies, Inc.) in a final volume of 20 ml. Digestion products were resolved on a 1% agarose Tris acetate/EDTA gel (10) with ethidium bromide. BSS Backcross Panel DNA blots were hybridized and washed with Church hybridization buffers at 65°C (11) containing 32 P-labeled 1.1-kb e subunit sequences. Allele typings were compared to all previously mapped loci in the appropriate Jackson interspecific backcrosses and matched to their best-fit map positions by minimizing double crossovers and inferring phenotypes for untyped noncrossover animals.
Animal Feeding-Animals, diet composition, 2 and liver RNA isolation for the experiments reported briefly here are described by Paisley et al. 3 BALB/cHnn mice (Harlan-Sprague-Dawley, Indianapolis, IN) were fed a semipurified diet containing 3% corn oil for 1 week and then randomly assigned for 2 weeks to the same 3% corn oil diets or otherwise identical diets, with 10% or 20% corn oil replacing carbohydrates isocalorically. Room lighting was on a 12-h light/dark cycle. Groups of animals were killed at predesignated times at the end of the 2-week feeding period as follows; at the end of a 12-h light cycle (ad libitum fed), after 12 h without food (Ϫ2 h, fasted), and at the end of a 12-h fasting phase followed by a 2-h period with food provided (0 h). All uneaten food was removed from the cages at 0 h, and three other groups were killed at 1, 2, and 6 h after the end of the 2-h period with food available. The quantity of food placed in the cages at the beginning of the 2-h feeding period was 80% of the average daily food consumption during the proceeding 2 weeks. 3 RNA Hybridization Analyses-Total RNA was extracted from ground livers and hearts by the Ultra-SpecII RNA isolation system (Biotecx, Houston, TX). Twenty g of total heart RNA were resolved by electrophoresis in gels containing 1.2% agarose and 2.2 M formaldehyde at 110 V for 2.5 h. Sizes were estimated by comparison to a 0.24 -9.5-kb ladder of synthetic RNA (Life Technologies, Inc.). Gels were stained with 0.5 mg/liter ethidium bromide, photographed, and blotted onto nylon membranes (U. S. Biochemical Corp.) following standard protocols (10). Hybridizations were in 7% SDS, 0.5 M NaPO 4 (pH7.2), 1 mM EDTA, and 1% SDS at 65°C with one 4% SDS, 40 mM NaPO 4 (pH 7.2), 1 mM EDTA wash at 65°C for 15 min and three 15-min washes with 1% SDS, 40 mM NaPO 4 (pH 7.2), 1 mM EDTA at 65°C (11). Hybridization signals were quantified using a Molecular Dynamics (Sunnyvale, CA) 425S Phosphor-Imager, and autoradiograms were obtained by placing blots next to preflashed film (0.25-s flash through a diffusion plate to give A 540 nm ϭ 0.2). The e subunit mRNAs were normalized to GAPDH mRNA levels because two actin mRNAs are expressed in heart tissues. The e subunit/ GAPDH mRNA ratios from individual hearts were normalized to the ratio of e subunit/GAPDH in identical aliquots (approximately 20 g) of a control RNA analyzed on each of the five blots. The control RNA, isolated from 5 pooled hearts of BALB/c mice fed 10% corn oil diets ad libitum, was used to control for blot to blot variation. Based upon the specific activity of the probes, PhosphorImager counts and efficiency, and the duration of exposure, there were approximately 10 7 molecules of e subunit mRNA/20 g of total RNA. Liver RNA hybridizations and analyses were as described previously (8). Hepatic e subunit mRNA was normalized to ␤-actin mRNA as a control for loading and expressed as a ratio of e subunit/␤-actin mRNAs. The data were normalized to e subunit/␤-actin mRNAs in hepatic total RNA from mice fed 3% corn oil. Effects of sacrifice time and linear trends with dietary fat level were evaluated in two-way analyses of variance, followed by examination of trends with fat level at individual sacrifice times. The SAS GLM program (12) was employed for statistical computing, with p Ͻ 0.05 employed as nominal criterion of statistical significance. Bonferroni's method (13) was used to adjust for testing of fat trend at five separate times, requiring that an observed p Ͻ (0.05/5 ϭ 0.01) be the standard of statistical significance for fat trend at any single sacrifice time.

RESULTS
e Subunit Gene Amplification-Four prominent DNA fragments A (900 bp), B (800 bp), C (350 bp), and D (200 bp) from BALB/cJ (not shown), C57BL/6J, and M. spretus genomic DNA were amplified by PCR using the internal Lfm1-specific primers, ϩ36 3 ϩ61 and ϩ222 3 ϩ198 (Fig. 1, lanes 1 and 5). Fragment A encoded an e subunit gene beginning at nucleotide ϩ36 of the cDNA sequence (with ϩ1 the A of the initiating ATG) and ending at ϩ216. This gene contained 3 exons and 2 introns and the sequence of the exons differed by 1 base pair (Fig. 2, nucleotide 611) from the previously reported Lfm1 cDNA sequence (8) but lacked the expected coding nucleotides ϩ1 to ϩ35. A second PCR reaction was performed with oligonucleotides complementary to Ϫ29 3 Ϫ4, the untranslated region of the cDNA sequence, and ϩ233 3 ϩ208, which overlaps the TAG termination codon (ϩ214 3 ϩ216). Three prominent DNA fragments of 1.1 kb (AЈ), 900 bp (BЈ), and 250 bp (DЈ) were produced (data not shown). 2 All diets contained a base of 20 g of alcohol-extracted casein, 5 g cellulose, 1 g of AIN-76A vitamin mix (25,26), 0.2 g of choline bitartrate, and 0.1 g of dye. In addition, the 3% corn oil diet contained 3 g of corn oil, 40 g of corn starch, 14.5 g of sucrose, 0.3 g of DL-methionine, 2.5 g of AIN-76A mineral mix (25,26) for a total of 100.1 g. The 10% corn oil diet contained the base plus 9.8 g of corn oil, 28.1 g of corn starch, 11.2 g of sucrose, 0.3 g of DL-methionine, and 3.5 g of mineral mix for a total of 91.7 g. The 20% corn oil diet contained the base plus 16.6 g of corn oil, 16.1 g of corn starch, 7.9 g of sucrose, 0.4 g of DL-methionine, and 4.5 g of mineral mix for a total of 83.1 g. 3 Paisley, E. A., Park, E. I., Swartz, D. A., Mangian, H. J., Visek, W. J., and Kaput, J. (1996) J. Nutr., in press. The 1.1-kb fragment (AЈ) encoded the 900-bp A fragment obtained with the internal primers plus a 5Ј exon and an additional intron. The sequences of the 4 exons match the cDNA sequence exactly (Fig. 2) except for two substitutions. Cytosine at 310 of the genomic clone is replaced by thymidine in the cDNA. This base overlapped the ϩ36 3 ϩ61 primer used for the first amplification (Fig. 2). Guanosine at 611 is replaced by an adenosine in the cDNA. The 1.1-kb gene sequences, compared to sequences in GenBank 4 identified the same sequences as Lfm1 cDNA: bovine (4) and rat (7) F 1 F 0 -ATP synthase e subunit sequences, a DNA damage-inducible gene from Chinese hamster cells (14, 15), and a human gene with similarity to the e subunit, which is unpublished. With conservative amino acid substitutions considered, murine Lfm1 is 98% similar to the protein sequence of rat liver e subunit and 96% similar to bovine heart e subunit.
Fragment C (370 bp) amplified by the internal set of primers encoded nucleotides ϩ36 to ϩ218 of the e subunit cDNA sequence joined by 4 nucleotides of unknown origin to nucleotides ϩ38 to ϩ222 of a second e subunit cDNA sequence (data not shown). Some of the nucleotide substitutions (5 total) in the first half of this gene created translation termination codons. The second half had seven substitutions with several termination codons and a dinucleotide deletion. This pseudogene was not found in over 200 clones of BALB/cJ, C57BL/6J, and M. spretus genomic DNAs amplified with oligonucleotides hybridizing to noncoding sequences (external primers) of the e subunit cDNA gene indicating that flanking sequences differed from those flanking the 1.1-kb gene. No e subunit sequences other than primer sequences were found within 390 and 320 base pairs of either end of fragment BЈ. This sequence matched no sequence in GenBank and was not further analyzed. Fragment D was not sequenced.
e Subunit Gene Mapping- Fig. 1 (lanes 3 and 7) shows polymorphism in amplification products between C57BL/6J and M. spretus parental DNAs after digestion with PstI ( Fig. 2 at  nucleotdies 426 3 432). XhoI control digestions (site at 419 3 424 in Fig. 2) produced similar 650 bp products (Fig. 1, lanes 4  and 8) for DNA amplified from these strains. The 1.1-kb e subunit gene was mapped by amplifying 94 segregants in the BSB Backcross Panel DNAs from The Jackson Laboratory with the internal Lfm1 cDNA primers followed by PstI digestion (data not shown). All 94 animals were typed. The gene maps to Chr 5 (left chromosome in Fig. 3), 2.13 Ϯ 1.49 cM distal from D5Mit9 (two crossovers out of 94 animals; Ref. 16). We designated this locus Atp5k 5 in consultation with the Mouse Genome 4 Genome Data Base (GDB) accession numbers for genes in this report: Lfm1, murine F 1 F 0 -ATP synthase e subunit gene is GDB no. S52977; bovine e subunit is GDB nos. M64751 and J05330; rat is GDB no. D13121; DNA damage-inducible gene from Chinese hamster cells is GDB no. G191230; and a human gene with similarity to the e subunit, which is unpublished, is GDB no. H29426. The 370-bp pseudogene is U59282, and the 1094-bp e subunit gene (Atp5k) on Chr5 is U59283. 5 Mouse Genome Data Base (MGD) accession numbers are as follows.  1 and 5, respectively), purified by Qiagen columns (lanes 2 and 6), and digested with PstI (lanes 3 and 7) or XhoI (lanes 4 and 8). ϩ and Ϫ refer to the addition of enzyme. PstI and XhoI sites that produce the 650-bp fragment are identified in Fig. 2. The XhoI site appears to be in the same position in M. spretus e subunit DNA, but additional or different PstI sites produce a 450-bp fragment. Lane 9 is e subunit cDNA (8)  Internal primers yielded an 800-bp fragment that started at ϩ36 of the cDNA sequence and ended at ϩ233. Nucleotide sequence was determined by sequencing both strands. The second amplification with external primers yielded a 1094-bp fragment encoding 3 introns and 4 exons. Exons are in uppercase letters and the introns in lowercase. The PstI site (426 -432) used to map the gene to Chr 5 also is found in C57BL/6J, but a different or additional site(s) is present in M. spretus.
Data base and Nomenclature Committees (Jackson Laboratory, Bar Harbor, ME).
Radiolabeled e subunit sequences were also hybridized to Southern blots of TaqI-digested Jackson Laboratory BSS Backcross Panel DNAs (Fig. 4). The 1.1-kb e subunit probe detected sequences on Chr 5 (5.1-and 0.9-kb bands, equivalent to the locus mapped in the BSB cross by PCR), Chr 8 (3.5 kb), Chr 11 (ϳ7 kb), and two sites on Chr 2 (5.8 and 0.7 kb) (Fig. 5). Similarly to the BSB PCR mapping data above, the Chr 5 e subunit-related sequences (Atp5k) map 2.13 Ϯ 1.49 cM distal from D5Mit9 in the BSS cross if the 13 missing noncrossover typings are inferred from surrounding data (Fig. 3, right chr; Ref. 16). Interpretation based upon combining the PCR/RFLP and Southern hybridization mapping data places the e subunit 2.13 Ϯ 1.05 cM distal to D5Mit9 (four crossovers out of 188 animals).
A 0.7-kb C57BL/6J band on the same BSS TaqI-digested Southern blots (Fig. 4), which could be scored in only 27 animals, cosegregated with D2Mit61 and D2Bir9 (16) and was designated Atp5k-rs1, for related sequence-1. A 5.8-kb band was scored in 85 animals, inferred from the remaining 9 animals, and co-segregated with Gnas (17) on Chr 2 but unlinked to Atp5k-rs1. This distal Chr 2 locus has been named Atp5k-rs2. From these same Southern blots, 82 animals were also typed for a 2.9-kb C57BL/6J band, which cosegregated with Gcdh (18) on Chr 8. This locus is designated Atp5k-rs3. Since all crossover animals were typed, the remaining phenotypes can be inferred to give 0 crossovers/94 animals between Gcdh and Atp5k-rs3. This locus had previously been identified in mapping studies with mice from CXB, CXH, BXA, and AXB recombinant inbred (RI) strains (19). PhosphorImager analyses indicated that the Chr 8 locus had about 1.5 times as many counts as other e subunit-related sequences, suggesting that it might encode the 370-bp pseudogene previously described. A 7.0-kb C57BL/6J band, scored in 72 of the animals cosegregated with D11Mit36 and D11Bir11 (16). Since none of the untyped animals were crossovers, typings in the remaining animals can be inferred in calculating the map position of this locus, designated Atp5k-rs4. Two of the three related loci are likely to encode another e subunit pseudogene and a cDNA-like gene which we have isolated from genomic libraries (not shown). Establishing the sequences at each locus will require additional studies.
Regulation of e Subunit mRNA Abundance-Mice fed 3%, 10%, and 20% corn oil containing diets consumed the equivalent of approximately 5 kcal during the 2-h feeding period. 4 The autoradiogram in Fig. 6 shows heart e subunit mRNA hybridization signals at 6 h after eating for mice fed 3% (lanes 2-6), 10% (lanes 7-11), and 20% corn oil (lanes 12-16). Similar analyses were done for mice that had been fed 3%, 10%, and 20% corn oil and killed at Ϫ2 h (fasted) and 0, 1, and 2 h after having food available for a 2-h period (i.e. fed mice). ANOVA showed that the e subunit/GADPH mRNA ratios were greater in all mice fed 20% corn oil compared to those fed 3% corn oil (p Ͻ 0.05). There was also a main effect (p Ͻ 0.0001) for time, showing that the e subunit/GAPDH ratio was significantly Atp5k. This BSB mapping of Atp5k was done by PCR/RFLP, and all 94 segregants were scored. On the right side of the figure is a partial map of the data from The Jackson BSS backcross, where an equivalent locus was mapped by Southern hybridization (see Fig. 4). In this analyses, 82 animals were typed for Atp5k and the phenotypes of the remaining 12 noncrossover animals were inferred. Bars between the two maps join loci mapped in both backcrosses.

FIG. 4. Southern hybridization analyses of TaqI-digested C57BL/6J and SPRET/Ei genomic DNAs. C57BL/6J and SPRET/Ei
DNAs, digested with TaqI, were hybridized with the 1.1-kb e subunit genomic probe. Approximate sizes were determined from molecular weight standards. The locus designation (Locus), chromosome assignment (Chr), and nearby locus (Marker) from The Jackson BSS backcross are given for each mapped band. The estimated fragment size is given in kilobases. Equivalent SPRET/Ei bands were assigned to the loci on the basis of scoring band intensities indicating homozygous or heterozygous backcross segregants (see Fig. 5). lower in fasted (2 h) relative to fed mice at 0, 1, and 6 h. The data were also analyzed at each time point for fat effects using Bonferroni's method (13) for multiple comparisons (Fig. 7). Differences approached significance 6 h after eating (p Ͻ 0.075, after Bonferroni's correction).
Liver tissues from mice in this fast/refeed regimen that were fed 3% and 20% corn oil were analyzed for e subunit abundance in fed mice ad libitum (Ϫ14 h), fasted mice (Ϫ2 h), and fed mice (2 h). A statistically significant (p Ͻ 0.05) difference in hepatic e subunit/␤-actin mRNA ratios was observed between different times during the fast/refeed regimen (Fig. 8). There was also a highly significant interaction (p Ͻ 0.0003) between time and fat level. Analyses by Bonferroni's method (13) showed highly significant (p Ͻ 0.0007, after Bonferroni's correction) differences in e subunit/␤-actin mRNA ratios as a function of fat levels following the 12-h fast (Ϫ2 h) and differences approaching statistical significance (p Ͻ 0.08, after Bonferroni's correction) at the beginning of a 12-h fast (Ad lib, Fig. 8).

DISCUSSION
Sequence analyses of a 1.1-kb genomic BALB/cByJ F 1 F 0 -ATP synthase e subunit gene with 3 introns and 4 exons (fragments A and AЈ) showed two differences in the coding sequence when compared to the BALB/cHnn cDNA reported previously (8). The C 3 T difference at position 251 produces a Phe 3 Ser substitution, which would be expected to alter protein structure, while the other change at 611 (A 3 G) would cause a conservative change (Lys 3 Arg). These differences in coding sequence may be due to substrains of mice (Harlan Sprague-Dawley versus Jackson Laboratory) because at least two reverse transcriptase-PCR amplifications of RNA from liver, heart, and kidneys of BALB/cHnn mice consistently yielded only the BALB/cHnn cDNA sequence in multiple isolates. 6 The structure of the gene, which encodes the same 5Ј UTR as the cDNA, suggests that it would be transcriptionally active. We also isolated a 370 bp (Fragment C) pseudogene from genomic PCR products which apparently arose from an abortive splicing event. The 1.1-kb gene was mapped to Chr 5 in The Jackson BSS and BSB interspecific backcrosses. Related sequences were mapped to Chrs 11, 8, and 2. Preliminary evidence indicated that the 370-bp pseudogene may be the Chr 8 locus for Atp5k-rs3. Further characterizations of e subunit sequences at this and other loci, including analyses of transcriptional activity, await additional studies.
Heart e subunit/GAPDH mRNA ratios were increased by the level of dietary fat and were also affected by time of eating. The trends observed in e subunit mRNA levels in fasting and 1 and 6 h postprandial, which apparently are due to different levels of 6 D. A. Swartz, unpublished data.  6. Autoradiogram of heart e subunit mRNA at 6 h after eating. Representative Northern hybridization analyses of e subunit/ glyceraldehyde-3-phosphate dehydrogenase mRNA in total RNA from hearts of individual BALB/c mice. Twenty g of total RNA from heart tissue of individual mice 6 h after eating. Lane 1, pooled RNA from five BALB/c mice fed 10% corn oil diet ad libitum was analyzed on each blot as a standard; lanes 2-6, total RNA from mice fed 3% corn oil diets; lanes 7-11, total RNA from mice fed 10% corn oil diets; lanes 12-16, total RNA from mice fed 20% corn oil diets. Differences in intensity between lanes from the same dietary treatment are due to differences in total RNA loaded. GAPDH was used to control for the loading differences between the lanes. Sizes were determined from standards on each blot. corn oil, were not statistically significant, although Bonferroni's correction method of multiple comparisons showed values approaching significance at 6 h. An experiment of greater statistical power would be required to achieve definitive results. The available data suggest that heart e subunit mRNA expression, as measured by its ratio to the housekeeping gene GAPDH, increases with increasing dietary fat (and concomitant decreases in carbohydrate) and the nutritional state of the animals (fasting versus fed). The trend in e subunit/GADPH mRNA ratios as a function of level of dietary oil was not observed immediately following eating (0 h, refed) and 2 h after eating (ϩ2 h, postprandial). These results indicate that e subunit expression is regulated by dietary constituents, their levels and availability, and by other factors. Explanation of these relationships will require additional observations.
Since we had shown previously that hepatic e subunit/␤actin ratios are decreased by high fat diets (8), we also analyzed e subunit mRNA abundance in livers of BALB/cHnn mice fed 3% or 20% ad libitum for 2 weeks before fasting (in light phase), after eating their assigned diets following fasting (in the dark phase), and 2 h after being allowed feed for 2 h. The 3 versus 20% corn oil and fasting versus fed states were the most extreme nutritional conditions imposed. The ad libitum fed mice confirmed our previous results (8) showing that the e subunit/␤-actin mRNA ratios increased significantly with the 3% corn oil diet compared to the 20% corn oil diet. Fasting and eating altered the response to the level of dietary corn oil, since e subunit/␤-actin ratios were increased in mice fed 20% corn oil compared to those fed 3% corn oil following a fast/refeed regimen similar to that observed in heart tissue. Differences in e subunit/␤-actin ratios between ad libitum feeding (mice killed during the light phase) and the fasting/feeding regimen (mice killed during the dark phase) may reflect diurnal variations and/or differences in nutritional status, which were not ad-dressed in our previously published studies (8). Others have shown similar dependence of transcript abundance on time of feeding for certain diet-regulated genes (20,21). The responses of hepatic and heart e subunit mRNA are likely caused by differences in nutrient supply and metabolic changes induced by the 3%, 10%, and 20% corn oil-containing diets. The role of insulin, glucagon, fatty acids (22), carbohydrates (23), and other metabolic factors (24) in regulating e subunit expression by diets differing in fat and carbohydrate content remain to be clarified.
Implications of Diet Regulation and Sequence Similarities-Since F 1 F 0 -ATP synthase has a central role in producing cellular energy, it is likely that multiple factors such as substrate concentrations, reducing equivalents, and cellular energy balance will regulate its activity (1,2). Many of these factors will differ if the primary nutrient source changes from lipids to carbohydrates. In addition, other agents may also modify its activity. Higuti et al. (7) have proposed that the e subunit contains a putative Ca 2ϩ binding site, suggesting that a Ca 2ϩ -e subunit complex participates in regulating e subunit binding or activity (7). Regulating F 1 F 0 -ATP synthase activity by altering transcriptional responses may seem less probable because the required enzyme machinery must be constantly responsive to changes in energy balance and substrate concentration. However, one or more regulatory subunits may be transiently regulated at the transcriptional level by different concentrations of the same or different nutrients. For example, certain carbohydrates and lipids may regulate gene transcription directly as ligands for transcription factors (22,23). Based upon its presumed interaction with the F 0 portion of the complex and the data presented here showing modulation of e subunit expression by the dietary concentration of nutrients and by the time of their consumption, we suggest that the e subunit participates in regulating ATP production. Since our studies did not directly address the function of the e subunit protein or analyze the promotor of the gene, further in vitro and in vivo studies will be required clarify these relationships.