Heterogeneity at the 5’ End of Rat Acetyl-coenzyme A Carboxylase mRNA LIPOGENIC CONDITIONS ENHANCE SYNTHESIS OF A UNIQUE mRNA IN LIVER*

Multiple forms of acetyl-coA carboxylase mRNA were previously detected in the mammary gland (Lo-pez-Casillas, F., Luo, X., Kong, I.-S., and Kim, K.-H. (1989) Gene, in press). We have now established that the rat liver also contains heterogeneous acetyl-coA carboxylase mRNA populations that differ in the 5’-untranslated region. In addition, the liver contains a unique form of acetyl-coA carboxylase mRNA in which the 5’-nontranslated end differs from the species in mammary gland. The 5’ end of this unique species was characterized using a procedure for cloning min-ute amounts of primer extension products (pAU clones). This procedure should also be useful for ob-taining full length clones of other mRNAs. The DNA sequence of pAU clones indicates that this liver-specific acetyl-coA carboxylase mRNA has a 315-base long untranslated region. The first 242 nucleotides replace the 5’ end of the predominant acetyl- CoA carboxylase mRNA found in the mammary gland (FL56 type). Under lipogenic conditions the unique liver acetyl-coA carboxylase mRNA increases and is the major species of acetyl-coA carboxylase mRNA. Livers from rats fed a normal diet and the mammary glands of lactating rats do not contain detectable amounts of the pAU type mRNA. On the other hand, the epididymal adipose tissue from these animals con- tains mainly the pAU type

Acetyl-coA carboxylase catalyzes the rate-limiting reaction in the biosynthesis of long chain fatty acids: the ATP-dependent carboxylation of acetyl-coA into malonyl-CoA (1,2). In order to elucidate the control mechanisms that regulate acetyl-coA carboxylase at the gene level, we have recently purified acetyl-coA carboxylase mRNA from lactating rat mammary gland (3), cloned its cDNA (4), and sequenced its coding region (5). Acetyl-coA carboxylase mRNA is encoded * This research was supported by Grant DK12865 from the National Institutes of Health. This is Journal Paper No. 11,832 from the Purdue University Agriculture Experiment Station. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) 504735.
$. To whom correspondence should be addressed.
in a single gene copy per haploid chromosome set (5). The amount of acetyl-coA carboxylase mRNA is very small even during lactation, a condition that demands a high rate of fatty acid synthesis (3). Acetyl-coA carboxylase mRNA with an apparent size of 10 kilobases (3, 4) contains an open reading frame of 7,035 nucleotides that encodes a 265,220-dalton polypeptide of 2,345 amino acids (5).
The lactating rat mammary gland contains multiple forms of acetyl-coA carboxylase mRNA that differ in the 5'-nontranslated end.' This tissue produces two types of acetyl-coA carboxylase mRNA that differ solely with respect to the presence (FL63) or absence (FL56) of a 61-base insertion in the middle of the 5"untranslated end. The regions upstream of the 61-base insert have a high (C + G) content, a high frequency of the CpG dinucleotide, and originate from the first exon(s) in the acetyl-coA carboxylase transcriptional unit. ' While characterizing the 5' end of acetyl-coA carboxylase mRNA, we found a novel form of rat liver acetyl-coA carboxylase mRNA that is not detected in the mammary gland. We herein describe the 5' end of hepatic acetyl-coA carboxylase mRNA (the pAU type), which becomes the major form of acetyl-coA carboxylase mRNA under lipogenic conditions. This novel acetyl-coA carboxylase mRNA is also present in the epididymal adipose tissue of rats given a normal diet. This pAU type of acetyl-coA carboxylase mRNA contains a different 242-nucleotide sequence in place of the (C + GI-rich region of the FL56 type. Our data suggest the existence of an alternative promoter in the acetyl-coA carboxylase gene, which is active in liver and adipose tissue but inactive in the mammary gland.
RNA Preparations-Total RNA was prepared from the mammary glands of lactating rats on the 5th-7th day postpartum, using a slight modification (3) of the protocol of Cathala et al. (6). The guanidinium thiocyanate method described by MacDonald et al. (7) was employed to prepare total RNA from livers and epididymal fat pads.
tides 218 (5'-ATGGTTCATCCATTATTCT-3') and 263 (5'-Primers, Probes, and Extended Primer Synthesis-Oligonucleo-3') were synthesized in an Applied Biosystems 380A DNA synthesizer and purified by polyacrylamide gel electrophoresis (8). The synthesis of primer 263/H2/Taq (108 nucleotides) will be used to describe the strategy employed in the synthesis of all the "extended" 32P-oligonucleotides (primers and probes) utilized in this paper. The primers were named to reflect the way they were synthesized. Oligonucleotide 263 was labeled at its 5' end by using [y3'P]ATP and T4 polynucleotide kinase as described by Taylor et al. (9). An aliquot of the kinase reaction mixture containing -30 pmol of the 32P-oligonucleotide was mixed with 50 pg of DNA from genomic clone H2 (the complementary M13 single-stranded DNA subclone) in 100 pl of 7 mM Tris-HC1, pH 7.5, containing 7 mM MgC12 and 50 mM NaC1. The nucleic acids were annealed by heating the mixture to 95 "C for 5 min and slowly cooling to room temperature. The annealed mixture was made up to 250 p~ each of dATP, dCTP, dGTP, dTTP, 5 mM DTT and 20 units/ml Klenow fragment in a final volume of 200 pl. The mixture was incubated at 37 "C for 30 min. Following extraction of the reaction mixture with phenol and chloroform (lo), the nucleic acids were recovered by ethanol precipitation (11). The nucleic acids were resuspended in 10 mM Tris-HC1, pH 7.5, and digested with the indicated restriction endonuclease, in the case of primer 263/H2/Taq with TaqI. The extended primers were then separated from their M13 templates by electrophoresis in 6% polyacrylamide, 8 M urea gels and recovered from the gel by electroelution using a Unidirectional Electroeluter Apparatus (International Biotechnologies Inc.). Following the addition of 25 pg of tRNA as carrier, electroeluted primers were precipitated by ethanol and further purified by resuspension, phenol/ chloroform extraction, and ethanol precipitation. The purified primers were resuspended in 150 pl of DEP-treated Hz0 and stored at -70 'C.
Primer Extension and RNA Sequencing-Primer extension was performed as described by Bertholet et al. (12). Briefly, 100-200 pg of extended primer (-15,000-30,000 cpm) was hybridized with 100-200 pg of total RNA in 50pl of 0.04 M Pipes, pH 6.4, containing 80% formamide, 0.4 M NaCl, and 1 mM EDTA by heating to 70 "C for 15 min and then incubating at 43 "C for 4-5 h. Annealed primer/RNA mixture was rapidly precipitated with ethanol and resuspended in 40 p1 of 50 mM Tris-HC1, pH 8.3 (measured at 42 "C), containing 50 mM KCl, 6 mM MgC12,lO mM DTT, 0.4 mM each of dATP, dCTP, dGTP, dTTP, and 25-35 units of AMV reverse transcriptase. The reaction mixture was incubated at 40 "C for 60 min. The reaction was stopped by adding 60 pl of 0.3 M NaOH and incubated for an additional 60 min at 40 "C. Following neutralization, the mixture was extracted with phenol/chloroform. The extension products in the aqueous phase were precipitated by ethanol. The final products were resuspended in 10 pl of gel loading solution (90% formamide, 30 mM NaOH, 1 mM EDTA, 0.1% xylene cyanol, and 0.1% bromphenol blue), denatured by heating at 95 "C for 5 min, and subjected to 6% polyacrylamide, 8 M urea gel electrophoresis. Extension products were detected by autoradiography using Kodak XAR5 x-ray film.
To modify the primer extension protocol for RNA sequencing, the initial reverse transcriptase reaction was carried out for 15 min in the presence of either 25 p~ ddATP, 25 p~ ddCTP, 25 p~ ddGTP, or 12 pM ddTTP in addition to 12 p M of each dNTP. The concentration of dNTPs was then raised to 150 p~, and incubation was continued at 40 "C for another 45 min, before the reaction was stopped as described above.
SI Nuclease Analysis-cDNA probes were annealed to RNA as described above for primer extension. The DNA/RNA hybrids were resuspended in 300 pl of the S1 digestion mixture (13) containing 280 mM NaCl, 30 mM CH3COONa, pH 4.4,4.5 mM (CH&OO)2Zn, 20 pg/ ml denatured calf thymus DNA, and 200 units/ml S1 nuclease. The by the addition of 75 p1 of 2.5 M CH3COONH,, 50 mM EDTA, and mixture was incubated at 37 "C for 30 min. Reactions were stopped 400 p1 of isopropyl alcohol. Precipitated nucleic acids were resuspended in 100 pl of 150 mM NaOH and processed like the stopped AMV reverse transcriptase reaction described above.
Linked primer extension/Sl nuclease digestion experiments were carried out in the same manner as for the primer extension. However,

TGATCTCATCTTCTGAGTTGTCTTCAGACACGGACCCAATT-
before stopping the AMV reverse transcriptase reaction, the reaction mixture was divided one-half was used to examine the primer extension products after the reaction was stopped with NaOH, and the other half was diluted with 280 pl of the S I digestion mixture and used for the evaluation of S1 nuclease protection analysis.
Cloning of Primer Extension Products-The products of the primer extension experiment were selectively cloned by the following procedure. For cloning purposes, the primer extension experiment was scaled up to accommodate 4 mg of total RNA as starting material. After the AMV reverse transcriptase reaction, the RNA was hydrolyzed with 250 mM NaOH for 30 min at 65 "C. Following neutralization with an equivalent amount of HC1 and 10 mM Tris-HC1, pH 7.5, the mixture was extracted with phenol and chloroform (lo), and the cDNA products were precipitated with ethanol (11). The cDNA products were resuspended in 100 pl of denaturing gel loading solution (8), denatured by heating to 95 "C for 5 min, quickly cooled at 0 "C, and fractionated by electrophoresis in a 6% polyacrylamide, 8 M urea gel. The primer extension products were identified by autoradiography at -20 "C for 2 h and eluted by electroelution using a Unidirectional Electroeluter Apparatus (International Biotechnologies Inc.). In order to ensure the maximal recovery of cDNA in each of the following manipulations, 25 pg of glycogen was added as carrier. The electroeluted primer extension product was further purified by precipitation with ethanol, resuspension in 10 mM Tris-HC1, pH 7.5, extraction with phenol and chloroform, and reprecipitation with ethanol. The cDNA was then subjected to homopolymeric tailing in 100 pl of 100 mM potassium cacodylate, pH 7.2, containing 2 mM CoC12,0.2 mM DTT, 50 p~ dATP, and 60 units of terminal transferase at 37 "C for 30 min (14). The tailing reaction was stopped by the addition of 3 pl of 0.1 M EDTA, and the reaction products were purified by extraction with phenol and chloroform and precipitation with ethanol. In order to synthesize its complementary strand, the poly(A) tailed primer extension product was resuspended in 30 pl of 10 mM Tris-HC1, pH 7.5, containing 100 ng of oligo(dT) (16-20-mer) and 150 mM KC1. This mixture was heated to 65 "C for 10 min, slowly cooled to room temperature, and then adjusted to 100 mM KCl, 30 mM Tris-HC1, pH 7.5, 15 mM MgC12, 2 mM DTT, 85 pM of each dNTP, and 10 units/ml Klenow fragment in a total volume of 100 p1. The reaction mixture was incubated at room temperature for 45 min, extracted with phenol and chloroform, and the reaction products present in the aqueous phase were precipitated with ethanol. In order to generate flush ends in the double-stranded cDNAs using T4 DNA polymerase (15), the cDNAs were resuspended in 60 p1 of 33 mM Tris acetate, pH 7.9, containing 66 mM potassium acetate, 10 mM magnesium acetate, 150 p~ of each dNTP, and 250 units/ml T4 DNA polymerase and incubated at room temperature for 20 min. The reaction mixture was extracted with phenol and chloroform following the addition of 7 pl of 0.1 M EDTA. The double-stranded cDNAs in the aqueous phase were precipitated by ethanol. Cloning vector pUC18 was prepared for ligation by digestion with SmaI and dephosphorylation with calf intestinal alkaline phosphatase as described by Maniatis et al. (16). Ligation of cloning vector and double-stranded primer extension product cDNA was carried out in 20 pl of 70 mM Tris-HC1, pH 8.0, containing 10 mM MgC4, 50 pg/ml bovine serum albumin, 10 mM DTT, 2.5 mM ATP, and 2 units of T4 DNA ligase at 15 'C for 24 h. The ligation mixture was used to transform competent Escherichia coli cells of strain JM83 (17). Transformants lacking 8-galactosidase activity were analyzed for the presence of acetyl-coA carboxylase mRNA sequences by colony hybridization (18) using end-labeled oligonucleotide 218.
DNA Sequencing-DNA inserts were subcloned into M13mp19 by using standard techniques (16). DNA was sequenced in both directions by the dideoxy chain termination method (19).

RESULTS
Heterogeneity in the 5' End of Rat Liver Acetyl-coA Carboxylase mRNA-Multiple products were obtained by primer extension of the 108-nucleotide fragment 263/H2/Taq ( Fig.  1) with liver RNA prepared from rats that were subjected to the starvation/refeeding regimen (Fig. 2, lune 1 ). Two of these species are the same size as products a and b that are generated by primer extension of the predominant form of rat mammary gland acetyl-coA carboxylase mRNA.' However, four additional primer extension products were detected. These additional cDNAs are indicated as bands A, B, C, and indicates the location of the first AUG codon of ACC coding region. The asterisk locates the 5' end limit of the exon present in genomic clone H2. D in Fig. 2 and have estimated sizes of 415,370,355, and 325 nucleotides, respectively. None of these cDNAs resulted from the primer extension of liver RNA prepared from rats that were starved for 3 days prior to death.' Under starvation conditions, there is no detectable acetyl-coA carboxylase mRNA in the liver (20).
The cDNA species shown in Fig. 2, lane 1, are the result of specific priming as indicated by the fact that they are S1 nuclease-resistant (Fig. 2,lane 2). In this experiment an aliquot of the primer extension reaction was subjected to S1 nuclease digestion. Only specific primer extension products will have full length base pairing with their RNA template and will be protected from S1 nuclease digestion. Further proof for the specificity of the primer extension reaction was obtained by directly sequencing the RNA template used for primer extension. The partial sequence obtained from total liver RNA using primer 263/H2/Taq matches the sequence at the 5' end of the H2 exon (segment IV, Fig. l ) , the 250-base long exon containing the AUG initiation codon of the acetyl-coA carboxylase coding region' (5) (Fig. 3A). Extended exposure of the same sequencing gel allowed an additional sequence CTCTTATAATTCTTATTG to be read, and this sequence matched the sequence of the FL56 type of mammary gland acetyl-coA carboxylase mRNA' (Fig. 3B). These results established that our primer extension experiments transcribe specific products that reflect the high degree of 5' end acetyl-coA carboxylase mRNA heterogeneity in the rat liver.
Additional Forms of Acetyl-coA Carboxylase mRNA-Primer extension experiments and RNA sequencing have revealed the existence of acetyl-coA carboxylase mRNA forms in the liver that resemble the FL56 and FL63 types of acetyl-coA carboxylase mRNA present in the mammary gland' (Fig. 1). In order to confirm the existence of such forms of acetyl-coA carboxylase mRNA in the liver and to deter- mine how the liver primer extension products  are structurally related to acetyl-coA carboxylase mRNAs in the mammary gland, we carried out S1 nuclease analysis. For this experiment two 5' end-labeled cDNA-specific probes were synthesized. Probe 263/FL56/Hind (283 nucleotides) and probe 263/FL63/Hind (360 nucleotides) can discriminate between the FL56 and FL63 types of acetyl-coA carboxylase mRNA (Fig. 4). These probes also contain 54 bases of the M13mp19 polylinker region at their 3' ends. When these probes are used for the S1 nuclease analysis, these 54 bases will be digested from the hybrids, and the fragments that are protected by acetyl-coA carboxylase mRNA can be unequivocally distinguished from the undigestedprobe. In Fig. 5 bands  mRNA 7179 FIG. 3. Direct sequence (lunes A, C, G, and T ) of rat liver total RNA using primer 263/H2/Taq (open arrow). The results of two independently prepared RNA preparations are shown. A shows a 9-hr exposure (-70 "C) and B shows a 7-day exposure (-70 "C) of the same sequencing gel. The heavy black arrow points to the 5' end of the exon present in genomic clone H2 (6). Lane F shows the primer extension reaction. Lane m shows the size markers from 32P-labeled HpaII fragments of pBR322. f and g identify the original size of the probes, whereas bands a-e correspond to the fragments protected by the liver acetyl-CoA carboxylase mRNA (lanes 2 and 4 ) . None of the species a-e are protected when liver RNA from the fasted rat is used in the S1 nuclease analysis (not shown). In the experimental results shown in Fig. 5, the fragments protected by mammary gland RNA are included for comparison (lanes 1 and 3 ) because they can be used to interpret the results obtained with the liver RNA. How these specific primers are protected by different forms of acetyl-coA carboxylase mRNA is schematically summarized in Fig. 4. For example, the FL56 type of acetyl-coA carboxylase mRNA would completely protect the cDNA segment present in the FL56 probe (bund c, 227 nucleotides) and only the first 173 bases of the FL63 probe  (bund b, lanes 3 and 4 ) . On the other hand, acetyl-coA carboxylase mRNA of the FL63 type would protect only the first 173 bases of the FL56 probe (bund b in lanes 1 and 2) and the full length cDNA included in the FL63 probe (bund e, 303 nucleotides). In other words, the protection of 227 nucleotides of FL56 probe (bund c) and 173 nucleotides of the FL63 probe (bund b ) is due to the presence of the FL56 type of acetyl-coA carboxylase mRNA. The protection of band e with FL63 probe and band b with FL56 probe is due to the presence of the FL63 type of acetyl-coA carboxylase mRNA. These experiments allow us to conclude that liver RNA contains the FL56 type of acetyl-coA carboxylase mRNA and only trace amounts, if any, of the FL63 type of acetyl-coA carboxylase mRNA.
Since the specific activities of these probes and the amounts of RNA used in these experiments were the same, the autoradiographic intensities of the protected fragments reflect the relative amount of the specific mRNA in the RNA preparation. Thus, in the case of the FL56 type acetyl-coA carboxylase mRNA, the most abundant species in the mammary gland, a strong band b with the FL63 probe (Fig. 5, lane 3 ) should be accompanied by an equally strong band c with the FL56 probe and a much dimmer band b with the same probe. The FL56 type of acetyl-coA carboxylase mRNA is obviously present in the liver, since the same set of bands is protected by liver RNA. However, the relative intensities of the bands composing this set (fainter band c and stronger band b with the FL56 probe) indicate that this form of acetyl-coA carboxylase mRNA is not the predominant species in liver. The fact that band b stands out as the best protected fragment with both the FL56 and the FL63 probes indicates that the major acetyl-coA carboxylase mRNA species in liver has portions of its nucleotide sequence that are identical to the FL56 and FL63 types of acetyl-coA carboxylase mRNA. These common sequences extend to the insertion point of the 61-base insert peculiar to FL63; they do not include either the 61-base insert ( Fig. 1, segment ZZ) or the (C + G)-rich region (Fig. 1, segment   I A ) of the FL56 type of acetyl-coA carboxylase mRNA. Therefore, the major species of liver acetyl-coA carboxylase mRNA contain their own specific sequences at their 5' ends.
Cloning of the Liver Primer Extension Product A: pAU Clones-In order to characterize the structure of the major forms of acetyl-coA carboxylase mRNA in the liver, we have cloned primer extension product A (Fig. 2), which is generated by the largest and most abundant species of acetyl-coA carboxylase mRNA in the liver. A cloning strategy was devised that would selectively enrich a size-selected primer extension product, as described under "Experimental Procedures." A total of 23 positive clones (pAU clones) was detected among the 440 recombinant clones screened. Restriction endonuclease analysis indicated that all of these clones had the same -440-base pair insert.

/ / / / / / / / / / / / / / / /~/ / / / / / / / / / / / / / / / A band B mRNA
The DNA sequences of the inserts that are present in the pAU clones (Fig. 6) showed all the features that were expected from the primer extension product and our cloning strategy. Specifically, their 3' ends finished at the first residue of primer 263 and they contained poly(T) tails at their 5' ends. These independent clones, differing in the lengths of the poly(T) tails, have identical cDNA nucleotide sequences. The sequence alignment of pAU clones with XFL56 and XFL63 clones confirms that their regions of sequence identity are exactly confined to the 173 nucleotides located between the start of primer 263 and the insertion point of XFL63's 61-base insert. The 242 nucleotides upstream to this segment constitute the novel 5' end leader of band A acetyl-coA carboxylase mRNA.
Conclusive evidence that the clone pAU represents the full length cDNA of the liver extension product A was obtained through specific primer extension and cDNA S1 nuclease protection experiments. Three primers covering different extends of the pAU clone were synthesized primers 263/H2Taq, 263/FL56/RV, and 263/pAU/RI are 108,155, and 275 nucleotides long, respectively (Fig. 1). Their 5' ends are aligned to the same point in acetyl-coA carboxylase mRNA but their 3' ends span into distinct regions upstream of the acetyl-coA carboxylase mRNA (Fig. 6). The result of the extension of these primers on rat liver RNA is shown in Fig. 7A. Primer 263/H2/Taq (lane 1 ) gave the complete set of primer extension products: bands A-D and a and b, while the extension of the more template selective primer 263/FL56/RV (lane 2 ) did not include a product corresponding to band B. The mRNA selectivity of primer 263/FL56/RV indicates that the template producing band B differs from the template producing bands A, C, and D by the exclusion of the bases belonging to segment I11 (Fig. 1). The extension of the pAU-specific primer 2631 pAU/RI produced bands A, C, and D, confirming that the pAU clones were derived from primer extension product A and that primer extension products C and D are shorter versions of A.
Further information about the structures of bands A-D was obtained from cDNA probe S1 nuclease analysis. The pAU cDNA-specific probe 263/AU/Bam was synthesized for these experiments (Fig. 4); this probe is 458 bases long and includes seven nucleotides from the cloning vector and 36 bases from the cloning poly(T) tail of pAU. These 43 extra bases are important in the recognition of bona fide S1 nuclease-protected fragments. The position of migration of this probe is marked as band e in Fig. 7B, while the fragments that are protected from S1 nuclease by liver RNA from refed animals are identified as bands a (-128 nucleotides), b (173 nucleotides), c (355 nucleotides), and d (415 nucleotides). For comparison, the extension products generated by using primer 263/H2/Taq are also shows in Fig. 7B (lane 1 ). The correspondence of band sizes and intensities between the primer extension and the S1 nuclease experiments allows us to conclude that the acetyl-coA carboxylase mRNA template producing primer extension product A and the S1 nucleaseresistant band d is fully cloned in pAU clones. Therefore, the acetyl-coA carboxylase mRNA form producing the S1 nuclease-resistant band c is a shorter version of the full length pAU type and is very likely the one responsible for primer extension product C. The structure of the acetyl-coA carboxylase mRNA species producing the many minor primer extension products in the range of 309 nucleotides (including band D) cannot be conclusively identified by these experiments. However, it is very likely that they are the result of incomplete extension of the full length pAU type acetyl-coA carboxylase mRNA and not the result of full length extension of independent shorter forms, because the S1 nuclease experiment did not reveal any protected fragments matching their sizes.
On the other hand, we can speculate about the structure of the acetyl-coA carboxylase mRNA producing band B as being an acetyl-coA carboxylase mRNA species in which the first 242 bases of the pAU type of acetyl-coA carboxylase mRNA are spliced together with exon H2. This hypothesis is based on the following observations: 1) band B is approximately 47 nucleotides shorter than band A, the same length as the fragment linking exon H2 and the 242-base long pAU-specific leader segment; 2) the inclusion of sequences belonging to this linking fragment in primer 263/FL56/RV results in the disappearance of band B from the set of this primer extension product; 3) the presence of an S1 nuclease-resistant fragment (band a in Fig. 7B) of -128 nucleotides in length; and 4) the overlapping faint RNA sequencing ladder in the background  1 and 3 ) or to rat liver total RNA from refed animals (lunes 2 and 4 ) and subjected to S1 nuclease digestion as described under "Experimental Procedures." Protected fragments ae are discussed in the text. Size markers (3ZP-labeled HpaII fragments of pBR322 indicated in nucleotides) are shown in the right lune.
of the main ladder in Fig. 3B. This background sequence starts 3 bases upstream to the end of exon H2 (arrow, Fig.  3B) and reads GGGAGGCCACTG (residues in common with the main FL56 type of sequenceare underlined). This background sequence represents the 3' end of the 5' leader portion of the novel form of acetyl-coA carboxylase mRNA detected as band A in the primer extension experiment (Fig. 2) and cloned in pAU (Fig. 6).
Tissue Specificity and Inducibility of thepA U Type of Acetyl-Coa Carboxylase-The existence of two unique 5' end leader regions in acetyl-coA carboxylase mRNA (segments IA and IB in Fig. 1) suggests the existence of alternative usage of promoters in the acetyl-coA carboxylase gene. Such mechanisms are very often tissue-specific and provide the cell with an additional control mechanism for gene expression in response to changes in the environment (21-23). In order to examine whether or not the occurrence of pAU and FL56 types of acetyl-coA carboxylase mRNA is subject to tissue specificity and/or inducibility, we carried out primer extension experiments using RNA prepared from different rat tissues under various dietary conditions. Fig. 8 shows the migration patterns of primer extension products and S1 nuclease-protected fragments from the linked primer extension/Sl nuclease digestion (lanes s) using primer 263/H2/Taq. Mammary gland RNA (lanes 2 ) gave the FL56derived primer extension products a and b (open arrows), while liver RNA from animals that have been fasted and refed a fat-free diet generated, as expected, products A-D from the pAU types of acetyl-coA carboxylase mRNA (black arrows, denote band A) in addition to those from the FL56 type (lanes 3 ) . RNA prepared from livers of rats fed a normal diet produced only bands a and b, which are derived from the FL56 type, in barely detectable amounts (lunes 5). On the other hand, the RNA prepared from adipose tissue of the same rats resulted in large amounts of pAU-derived bands A-C and only traces amounts of FL56's bands a and b (lunes 6). The tRNA, which served as a negative control (lanes 1 ), and starved liver RNA (lunes 4), which has negligible amounts of acetyl-coA carboxylase mRNA, produced no acetyl-coA carboxylase-specific primer extension products. These results indicate that the physiological state plays an important role in controlling the tissue-specific levels of different acetyl-coA carboxylase mRNAs.

DISCUSSION
The present studies revealed that the diversity of structural motifs exhibited by the 5' end untranslated region of rat acetyl-coA carboxylase mRNA is more profuse in liver than in mammary gland. Among the several approaches available to disclose the structure of these liver forms of acetyl-coA carboxylase mRNA, cloning of the primer extension products is most likely to lead to unambiguous results. Direct sequencing of any mixture of mRNAs poses not only technical difficulties but also results in useless sequence data composed of multiple overlapping ladders. Even in the event that one obtains a partial sequence from which oligonucleotide probe could be designed, the need to generate an appropriate cDNA library from which the rest of the mRNA sequence could be established still remains. Therefore, we directly cloned the primer extension product detected as band A in Fig. 2, following the procedure describes under "Experimental Procedures." A number of positive clones (pAU clones) resulted from this approach. DNA sequencing of two of these clones, pAU3 and pAU4, showed that after offsetting the size of the cloning poly(T) tails they had an identical 415-base pair long cDNA insert, i.e., the expected size for a full length copy of primer extension product A. Supporting experimental evidence that the cDNA in pAU4 is indeed derived from primer extension product A was provided by the specific primer extension and cDNA S1 nuclease protection experiments of Fig. 7.
The DNA sequences of pAU clones made clear the extent of their identity with mammary gland clones XFL56 and XFL63. These three acetyl-coA carboxylase cDNAs share the same nucleotide sequence from the first base of primer 263 to the point where the (C + G)-rich region of XFL56 starts. This segment comprises 173 bases and is the one protecting band b in the S1 nuclease analysis of the FL56 and FL63 probes (Fig. 5). The slightly larger size of band b derived from the FL56 probe (175 nucleotides) can be accounted for by the fact that this probe includes 2 extra matching bases (3' end bases of the (C + G)-rich region of XFL56) that are made unavailable for S1 nuclease protection by the presence of the 61-base insert characteristic of XFL63.
The alignment of the nucleotide sequences of XFL56, XFL63, and pAU provides evidence of the unmistakable modular nature of the 5"untranslated region. This modular nature was initially manifested by the cDNA S1 nuclease protection patterns and the selectivity of the primer extension products generated by cDNA specific primers. These landmarks of the 5' end of acetyl-coA carboxylase mRNA permit us to separate it into discrete segments, as shown in Fig. 1. Segment IA designates the (C + G)-rich, 5' end portion of XFL56 and XFL63, the putative first exon of the mammary gland's acetyl-CoA carboxylase transcriptional unit. Segment I1 corresponds to the 61-base long putative shuffling exon present in the FL63 type of acetyl-coA carboxylase mRNA. Segment I11 is the 47-base string located upstream of exon H2; it is present in all of the cloned forms of acetyl-coA carboxylase mRNA. However, as explained under "Results," specific primer 2631 FL56/RV extension experiments (Fig. 7 A ) , together with the appearance of the 128-base long fragment in the cDNA S1 nuclease experiments (Fig. 7B, band a)  There are numerous examples of alternative use of exons and promoters as a mechanism to introduce new modes of control of gene expression and to generate diversity from single genes (24-25). The modular nature of the acetyl-coA carboxylase mRNA 5' end is very suggestive of differential splicing of the acetyl-coA carboxylase transcriptional unit. This suggestion is further reinforced and complicated by the discovery of new forms of acetyl-coA carboxylase mRNA with totally distinct 5' ends. The fact that the acetyl-coA carboxylase mRNA forms present in liver exhibit two mutually exclusive 5' end sequences (segment IA ? I1 or segment IB in Fig. l), while sharing the same set of 3' end segments ( Fig. 1, segments I11 and/or IV), implies that in addition to the necessary alternative usage of exons, there may be an alternative usage of promoters in the transcription of the acetyl-coA carboxylase gene in the rat liver. Since there is only one copy of the acetyl-coA carboxylase gene per haploid chromosomal set (5), one would have to postulate a hitherto unknown "splicing" mechanism for the generation of both pAU and FL types of acetyl-coA carboxylase mRNA if there were only one promoter for the acetyl-coA carboxylase gene. Recently, we have obtained supporting evidence for multiple promoters in the acetyl-coA carboxylase gene.3 When about 1 kilobase pair of genomic DNA flanking either the 5' end of the exons for segments IA or segments IB was placed in front of the chloramphenicol acetyltransferase gene and the con- b. Panel B, S1 nuclease analysis. Lane 2 shows the fragments protected (bands indicated with lower case letters) from S1 nuclease digestion by total liver RNA from refed rats using pAU-specific probe. structs were transfected into 30A-5 preadipocytes, both of these DNAs were able to stimulate the expression of chloramphenicol acetyltransferase activity. Further experiments are being carried out to establish that these two DNA fragments are indeed the promoters that are responsible for the generation of the heterogeneous acetyl-coA carboxylase mRNA. Interestingly, the results in Fig. 8 suggest that such an alternative usage of promoters may be tissue-specific and sensitive to the dietary manipulations that regulate lipogenesis. The pAU forms of acetyl-coA carboxylase mRNA are virtually absent in the mammary gland, even at the peak of the lactation-induced increase of acetyl-coA carboxylase mRNA levels (3), as well as in the liver of rats fed a complete diet. Under these conditions the predominant form of acetyl-CoA carboxylase mRNA corresponds to the FL56 type, presumably a "housekeeping" kind of transcript. On the other hand, the pAU forms of acetyl-coA carboxylase mRNA are the predominant species in the liver upon dietary treatment that maximizes lipogenesis (2) and acetyl-coA carboxylase mRNA levels (3), suggesting that such a transcriptional unit a -160 e -147 @ -122 FIG. 8. Primer extension analysis and linked primer extension/Sl nuclease analysis (lanes s) using primer 263/H2/Taq. 1OO-Ng samples of the following RNAs were tested in each experiment: tRNA (lanes I ) ; total RNA from rat mammary glands at day 6 of lactation (lanes 2); total RNA from livers of rats that before their death underwent either feeding with a standard laboratory animal diet (lanes 5 ) or 3 days of starvation (lanes 4 ) or 3 days of starvation followed by 2 days of refeeding a fat-free diet (lanes 3); and total RNA from epididymal adipose tissue from a rat fed a standard chow (lanes 6 ) . Prominent primer extension products of the FL56 type of acetyl-coA carboxylase mRNA (open arrows) and of the pAU type of acetyl-coA carboxylase mRNA (black arrows) are indicated. Lanes m show the molecular weight markers (32P-labeled HpaII fragments of pBR322) whose sizes are indicated in nucleotides at the margin of the figure. is the one responsive to changes in the demand for fatty acid biosynthesis. Interestingly, adipose tissue from rats fed a complete diet have minimal amounts of the FL56 type of acetyl-coA carboxylase mRNA and substantial amounts of the pAU type of acetyl-coA carboxylase mRNA, implying that tissue-specific controls exist in addition to the mechanisms regulating the hepatic expression of two acetyl-coA carboxylase gene promoters.
Independently of the actual mechanism giving rise to the 5' end heterogeneity of acetyl-coA carboxylase mRNA, the fact that the heterogeneity is confined to the untranslated region of the mRNA opens the possibility that these diverse forms of acetyl-coA carboxylase mRNA are subjected to some form of translational control (22, [26][27][28][29]. Work is in progress in our laboratory that will address the many questions posed by the present studies.