Acetyl-Coenzyme
A Carboxylase
mRNA Metabolism
Fernando Lbpez-Casillas, M. Verhica
Ponce-Castatieda,
in the Rat Liver
and Ki-Han Kim
The acetyl-coentyme A carboxylase (ACC) gene contains two promoters (PI and PII), both of which are active in the liver. Various physiological stimuli affect one, or both of the promoters of the ACC gene, and result in the generation of two classes of ACC mRNAs which differ in the composition of their 5’ untranslated regions (5’ UTR). We have analyzed the amounts of the two major mRNAs species that are generated from each of these promoters in order to examine the regulation of ACC gene activity in the liver under different physiological conditions. Our findings can be summarized as follows: (1) In liver from normal animals, fed a complete laboratory chow ad libitum, the level of class 2 ACC mRNA species generated by PII is very low. These mRNA species disappear on starvation. Refeeding starved animals with a fat-free diet stimulates both PI and PII with different time courses of induction: PII responds quickly and PII gene products accumulate to maximum levels within 18 hours, while the PI response, as measured by the accumulation of class 1 mRNAs, shows a lag period of 8 hours before reaching maximal levels at the end of a 24.hour refeeding period. The half-lives estimated from the induction kinetics were 4.4 hours for class 2 mRNAs and 11.8 hours for class 1 mRNAs. Reinstatement of starvation causes an almost instantaneous disappearance of class 1 mRNA species, as compared with class 2 mRNA species. This rapid decay of PI transcripts suggests that factors stabilizing this class of ACC mRNAs contribute to the steady-state levels reached after the dietary induction. (2) In starved streptozotocin-diabetic rats, refeeding with a fat-free diet causes activation of both promoters, whereas refeeding with the standard diet has no effect. This indicates that activation of the promoters occurs in the absence of insulin. However, administration of insulin to these animals further activates both promoters, regardless of the type of diet given to the starved animals. There is no insulin effect in diabetic animals that are fed normally. These observations indicate that in the liver two ACC gene promoters are selectively regulated under different physiological conditions. Insulin is one of the factors that affect the ACC gene. However, the factor(s) that is responsible for the activation of the ACC gene under the lipogenic conditions created by administration of a fat-free diet is not insulin. Copyright 0 1992 by W.B. Saunders Company
A
CETYL-coenzyme A carboxylase (ACC) catalyzes the rate-limiting step in the biosynthesis of fatty acids.‘.’ Characterization of the 5’ end of the ACC gene has shown that the ACC gene contains two promoters, designated PI and PII, which can promote the transcription of two distinct classes of ACC mRNA, class 1 and class 2, respectively.“’ PI is located upstream to the first exon and PI1 is located upstream to the second exon of the ACC gene. These two exons, exon 1 and 2, together with exons 3, 4, and 5, compose the 5’ region of the gene. The translation initiation AUG codon of the ACC open-reading frame (ORF) is located in exon 5. Transcription, followed by alternative splicing of the first five exons of the ACC gene generates several forms of ACC mRNA, which can be classified into class 1 or class 2 depending on whether they contain exon 1 or 2, which is determined by whether PI or PI1 is used. The major ACC mRNA species in class 1 is ACC[ 1:4:5]mRNA, which contains exon 1, 4, and 5 sequences at its 5’ end, whereas the major class 2 species is ACC[2:4:5]mRNA, which contains sequences for exons 2,4, and 5 at the 5’ end. The current nomenclature denotes the composition of the exons at the 5’ end of the mRNA. Important differences among these ACC mRNAs appears to be restricted to their 5’ untranslated regions (UTR). All species of ACC mRNA contain exon 5 in their 5’ UTR. Prior to the demonstration of the 5’ end heterogeneity of ACC mRNA, ACC gene activity under different physiological conditions had been investigated by measuring the total amount of ACC mRNA through dot blot or Northern analysis using cDNA probes from the coding or 3’ UTR of the mRNA.8-13Although the results of such studies reflect the patterns of total ACC mRNA metabolism, they cannot distinguish the true nature of ACC gene activity, nor the nature of ACC mRNAs. Because the two ACC gene promoters found in the liver respond differently to particuMetabolism,
Vol41, No 2 (February), 1992: pp 201.207
lar physiological conditions, meaningful studies of hepatic ACC mRNA require the use of procedures that identify the activities of PI and PII. In the present report, we have used primer extension analysis on total RNA using a specifically designed primer. With this approach, we have been able to measure the relative levels of the transcriptional products generated by PI and PI1 in the total ACC mRNA population. This has allowed us to investigate changes in the content of specific ACC mRNAs and the ACC promoter activities that occur in rat liver under different experimental conditions. MATERIALS
AND METHODS
Materials
Biochemicals were purchased as follows: [r-“P] adenosine triphosphate (ATP) (6,000 Ci/mmol) from DuPont-New England Nuclear, Boston, MA, avian myeloblastosis virus (AMV) reverse transcriptase from Life Sciences; T4 polynucleotide kinase from United States Biochemicals; porcine zinc insulin from CalbiochemBehring; streptozotocin from Sigma Chemical, St Louis, MO; Sl nuclease from Pharmacia LKB Biotechnology, Uppsala, Sweden; guanidinium thiocyanate from Fluka; XARS x-ray films from Kodak, Rochester, NY. The llO-mer and the 108-mer (263/H2/
From the Biochemistry Department, West Lafayette, IN. Supported by National Institutes of Health Grant No. CA 46882 and DK 12865. This is Journal Paper No. 12808from the Agricultural Experimentation Station, Purdue University. Dr L6pez-Casilla’s present address is Howard Hughes Medical Institute and Cell Biology and Genetics Program, Memorial SloanKettering Cancer Center, New York, NY10021. Address reprint requests to &Han Kim, PhD, Biochemistry Department, Purdue University, West Lafayette, IN 47907. Copyright 0 1992 by W.B. Saunders Company 00260495/92/4IO2-0020$03.00/O 201
LOPEZ-CASILLAS, PONCE-CASTAfJEDA,
202
Taq) primers used in the primer extension analysis were synthesized in an Applied Biosystems 380A DNA synthesizer, and were purified by urea/polyacryiamide gel eiectrophoresis (PAGE) as described.4 All other chemicals were reagent grade. Animals Male Wistar rats (1.50 to 250 g) were from our departmental colonies. The complete standard chow, Purina’s Rodent Laboratory Chow 5001 (Lafayette, IN), contains 4.5% of fat. The fat-free test diet from United States Biochemicais (Cleveland, OH) contains ceiiufii, 16.45%; salt mixture, 4%; sucrose, 58.45%; and casein, 21.1%, supplemented with vitamins. Rats were made diabetic by the administration of streptozotocin (65 mg/kg, subcutaneously) as described previously.” RNA Preparations Ail RNA samples were prepared from frozen tissues following the guanidinium-thiocyanate method described by MacDonald et ai.l4 RNA amounts were determined by UV absorbance at 260 nm (a 40+g/mL RNA aqueous solution providing 1 optical density unit at 260 nm).
diet to the starved animals induces an increase in ACC of up to 60-fold.” In Fig 1, the results of primer extension analysis of total liver RNA preparations from animals under different nutritional conditions are shown. The extension products identified by the arrows are the major ACC mRNA species generated as a result of PI and PII activity. The 415nt long extension product represents the ACC[l:4:5]mRNA species (closed arrow) and the 269- to 264-nt long extension products (open arrow) represent the ACC [2:4:5]mRNA species in the RNA preparations. RNA samples from normal animals fed a complete standard chow ad libitum, contain no PI promoter products (ACC[l:4: S]mRNA) and have extremely low basal levels of PI1 products (ACC[2:4:5]mRNA). The left side of Fig 1 shows the time course of decay of the basal hepatic levels of ACC mRNA, during starvation. The Iane designated “0” corresponds to the basal levels found before starvation in animals fed a standard laboratory chow ad libitum. The only primer extension products found under these condiREFEEDING
Primer Extension Analysis
ST~~IVITKIY
The primer extension procedure of Berthoiet et al” and the linked primer extension/S1 nuclease protection experiment were performed as described previousiyP Primers were labeled at their 5’ ends with the use of T4 poiynucieotide kinase and [y-‘Zp]ATP as described.16 The two primers used in these experiments were a llO-mer and a lob-mer (263/H2/Taq). A description of the 108-mer and the characterization of the primer extension products it generates have been reported in detail.‘,’ These primers recognize and anneal to ail the species of ACC mRNA because the sequences to which they are complementary derive from exon 5, which exists in all the forms of ACC mRNA characterized to date. The llO-mer has the same sequence as the original 263/H2/Taq (lob-mer), except that it has two more bases at its 3’ end. These extra bases do not affect hybridization to ail forms of ACC mRNA. Also, the sizes of the primer extension products generated from either primer are indistinguishable, and the sizes are solely determined by the type of ACC mRNA to which they hybridize. For example, when these primers hybridize to and are extended on the ACC[1:4:5]mRNA, the are elongated to a length of 415 nucieotide (nt).” On the other hand, when the primers hybridize to and extend on the ACC[2:4:5]mRNA, the extension products are 269 to 264 nt in length. Thus, the sizes of the primer extension are diagnostic of the type of ACC mRNAs present in the analyzed sample. In our experiments, we have used a molar excess of the primers so that ail the ACC mRNAs in the sample could be extended. When the products of the reverse transcriptase reaction are resolved in urea/PAGE and autoradiographed, the autoradiograms show not only the specific species of ACC mRNAs in the sample, but also the relative amounts of each species. Quantitation of the signals derived from the autoradiograms was performed by densitometric analysis using a LKB 2400 Geiscan XL software package. Statistical calculations were performed with the System for Elementary Statistical Analysis Package from the SAS Institute, Gary, NC. RESULTS
Effects of Starvation and Refeeding on ACC mRNA Metabolism in Rat Liver
When rats are fasted, fatty acid synthesis is diminished and ACC virtually disappears.’ However, feeding a fat-free
AND KIM
m
404-
0
309-
a
242-
!ad’z
hrr
243@4@72 ---_m--------
0 12 __--lFNklFNF
hrs
24
48
0 ; 0
i@O- * a
1io-
c
-mu.
Fig 1. Hepatlc ACC mRNA metabolism during starvation and refeeding. Total RNA was prepared from livers of rats subjected to the indicated treatments and 200 pg was analyzed for ACC mRNA by primer extension. The starvation tlme zero aample corresponds to rats fed e complete stendard chow ad Ilbitum. Starvation (72 hours) was followed by refeeding of e complete standard chow (lanes N) or a fat-free diet (lanes F). The animals were killed at different times during the starvation/refeeding period as Indlcated at the top of the figure. PI (solid arrow) indkatr the primer extension product originated from ACC[1:4:5JmRNA (415 nt). P2 (open arrow) Indicates the primer extension products that originated from ACC[2:4:S]mRNA (269 to 264 nt). The migration of the unextended 1gGmer used in the experiment is indicated by the label “primer” at the bottom of the flgure. Lanes m show the migration of the molecular weight markers (5’ end ryP]labeled Hmll fragments of the pWW22). whose sizes are indicated in nucleotldes at the left margin.
REGULATION OF ACElYL-CoA
203
CARBOXYLASE GENE EXPRESSION
tions are trace amounts of extension products of about 269 to 264 nt (P2, open arrow), which originate from ACC[2:4: S]mRNA. The complete absence of 415nt long primer extension products indicates that ACC[1:4:5]mRNA is not present in this sample. During starvation, the extremely low level of ACC[2:4:5]mRNA decreases until it is undetectable. Refeeding starved animals causes an increase in the levels of both classes of ACC mRNA (right side of Fig 1). The composition of the diet used for refeeding affects this response. Full induction of both species of ACC mRNAs only occurs when the starved animals are fed a fat-free diet. The effects of fat-free diet on the activation of PI are particularly striking (F lanes in Fig 1). Refeeding a standard diet also causes an increase in the amount of ACC mRNAs (lanes N in Fig 1); this increase is mainly due to activation of PII. The results in Fig 1 signify the importance of the present approach to examining gene activities as opposed to Northern or dot blot analyses of mRNA populations. These latter analytical methods would not show the underlying gene activity that is involved in the induction of ACC under these experimental conditions. The absolute ratio of the levels of ACC[2:4:5]mRNA and ACC[1:4:5]mRNA reached after 24 hours on the fat-free diet varies because of the high variability in the response of the PI promoter that occurs from animal to animal. However, we have observed that although some rats produce stronger PI responses than others, on the average, the ratio of PI1 products to PI products is about 3:2 after 24 hours on the fat-free diet. Kinetic Analysis of the Dietary Induction of the ACC Gene in Rat Liver Each experimental point in Fig 1 was obtained from total RNA preparations from three animals. To examine the kinetics of mRNA induction and decay in a more reliable manner, the same experiments were performed three times. The responses of PI and PI1 were analyzed and standardized using the values at 24 hours of refeeding. A summary of these experiments is shown in Fig 2. Substantially different kinetics for the induction of the activities of the ACC gene promoters are found when starved animals are refed with a fat-free diet. PII is activated immediately when refeeding begins; by the sixth hour, almost 50% of the maximum response is observed, and within 18 hours, it has reached its maximal level, which is followed by a slight decrease in the levels of the PI1 products, even when the animals are kept on a fat-free diet. On the other hand, the response of the PI promoter shows a lag period of about 6 hours. During the first 6 hours of the refeeding period, PI remains inactive, reaching the 50% and 100% induction levels only after 18 and 24 hours, respectively. If the animals are kept in a fat-free diet, PI products remain at this maximal level for the next 24 hours. From the kinetics of induction elicited by refeeding with the fat-free diet, we estimated the half-life of ACC[2:4: S]mRNA as 4.4 hours and the, half-life of ACC[1:4: S]mRNA as 11.8 hours.” Removing food after the 24th hour causes an almost immediate disappearance of the PI transcriptional products. ACC[1:4:5]mRNA decreases to
q Pn wx![2:4:51mRNAl
T
T
T
12
16
24
0
PI (Acc[I:4:5lmRNAl
OL 0
6
REFEEDING
30
48
@ml
Fig 2. Time course of the dietary induction of hepatic ACC mRNAs. The relative levels of each one of the major products from PI or PII during the refeeding of a fat-free diet were plotted. The upper half of the standard deviation (from four determinations) for each point is shown as lines parting from the plotted average value. Time zero corresponds to initiation of refaeding after 72 hours of starvation. Squares show the relative amounts of ACC[2:4:5jmRNA. Circles show the relative amounts of ACC[1:4:5]mRNA. The values obtained from animals that were returned to starvation after 24 hours of refeeding ara indicated by solid symbols and broken lines.
approximately one tenth of the previous levels within the first 6 hours of the restarvation period (Fig 2). On the other hand, the levels of the PI1 products require longer times to show a comparable decrease. After 24 hours of restarvation, the levels of ACC[2:4:5]mRNA have decreased to approximately one quarter of the previous levels. It is clear that the PI products are more sensitive to the reinstatement of starvation, since they disappear quickly during starvation. This suggests that the accumulation of class 1 ACC mRNAs observed during refeeding with a fat-free diet is the result of two simultaneous processes: an increase in the transcriptional activity of PI and an increase in the stability of the PI products. On the other hand, it appears that the major factor contributing to the new higher steady-state levels of class 2 ACC mRNAs during the same treatment is the increased transcriptional activity of the PI1 ACC gene promoter. Insulin Effects on ACC mRNA Metabolism in the Livers of Diabetic Rats The transcriptional effects on several hepatic genes elicited by a fat-free diet appear to be mediated by the increased levels of insulin provoked by this dietary regimen.‘3~‘sTherefore, we have examined the role played by insulin in the induction of the ACC gene when starved rats are refed with a fat-free diet. Insulin was administered to streptozotocin-diabetic rats and ACC mRNAs were analyzed by primer extension. The hepatic levels of ACC mRNAs, which are primarily represented by class 2 mRNA species (lanes 1 and 2, Fig 3A) under diabetic conditions, did not significantly increase in response to insulin (lanes 3-8, Fig 3A). The diabetic rat exhibits basal levels of
204
LCPEZ-CASILLAS, PONCE-CASTAI;IEDA, AND KIM
LIVER Diabetic INSULIN (-)
---w
1.5
INSULIN
Rata
P(-)
hrs
3
6
1.5 3 --
(hrsj
- 6
527 -
527-
m
Fig 3. Responses of the ACC gene in diabetic rats elicited by insulin. Insulin was administered to diabetic rats (1.6 hepatic ACC mRNA response (A) and ACC mRNA response in the epidydimsl fat tissues (B) were analyzed in 250 extension. Samples from animals not given insulin are shown in lanes 1 and 2. Following injection of insulin, three indicated times. The samole in lane g came from normal animals subjected to the starvation/fat-free diet refeeding arrow), and P2 (open arrow) are as in Fig 1.
ACC[2:4:5]mRNA, and shows no ACC[ 1:4:5]mRNA, as is the case with normal animals. The experimental results shown in Fig 3A demonstrate that insulin alone failed to activate either ACC gene promoter. The levels of ACC[2:4: S]mRNA and ACC[1:4:5]mRNA were not significantly affected by the administration of insulin. Even when insulin treatment is continued for a period of 24 hours, the ACC mRNA levels remained unaffected (not shown). This lack of response in the liver is in sharp contrast to the quick responses exhibited by white adipose tissue in the same animals. Analysis of the ACC mRNAs in epididymal fat pads of the rats shown in Fig 3B indicates that class 1 ACC mRNAs predominate in fat tissues, and that ACC[1:4: S]mRNA increases to about three times the basal level approximately 6 hours after the administration of insulin. The adipose tissue responsiveness in these animals rules out the possibility that the lack of response found in liver indicates a generalized failure of our experimental animals to respond to hormone treatment. Our liver studies appear to be in disagreement with the data of Katsurada et al,” which show a twofold increase in the levels of hepatic ACC mRNA after insulin was administered to diabetic rats. The main difference between our experiments and those of Katsurada et al resides in the of their diabetic rats to a high glucose “adaptation” (fat-free) diet before insulin was administered. In our studies, diabetic rats were kept on a complete standard chow before insulin treatment. The possibility that feeding
U subcutaneously) and the pg of total RNA by primer animals were killed at the scheme. Lanes m, PI (solid
a fat-free diet causes a substantial difference in the activation of ACC promoters was substantiated by our starvationrefeeding experiments with diabetic rats, as shown in Fig 4. When diabetic rats are subjected to the starvation/ refeeding protocol without insulin treatment, refeeding with a fat-free diet causes the activation of both ACC promoters as shown by the presence of both classes of ACC mRNAs (lane 4, Fig 4). In contrast, refeeding starved diabetic animals with standard chow does not activate either of the ACC promoters (lane 3, Fig 4). Although this activation is quantitatively smaller than the one observed in normal animals (lane 5, Fig 4) it is significant in that the ACC promoters in the livers of diabetic rats are activated by a fat-free, carbohydrate-rich diet in the absence of insulin. On the other hand, the observation that both promoters are activated in the livers of normal animals which have been subjected to starvation and refed with the standard diet (Fig l), while this does not occur in diabetic animals, suggests that insulin is one of the factors needed for the activation of the promoters. Indeed, one can demonstrate the effect of insulin on the accumulation of both forms of ACC mRNA species in diabetic animals as shown in Fig 5. The administration of insulin to diabetic animals which had been subjected to starvation and refeeding with the standard diet clearly shows the activation of both PI and PI1 and establishes that insulin is one of the factors involved in activation. The administration of insulin to diabetic animals
REGULATION
OF ACETYL-CoA
CARBOXYLASE
205
GENE EXPRESSION
4Pl
cJp2
tion and what physiological purpose they could serve in the cell. To address the first question, we have analyzed, by means of primer extension experiments, how the levels of the major hepatic ACC mRNA species are affected by two important physiological conditions that affect lipogenesis: nutritional status and diabetes. The primer extension experiment was designed so that in a single experiment, we could unambiguously distinguish and quantitate each one of the diverse ACC mRNA forms and thus the activities of the two promoters. This technique directly assesses the activity of each one of the two ACC gene promoters, a task that cannot be accomplished by Northern or dot blot analysis. Analyzing ACC gene activity by means of primer extension showed that the promoters of the ACC gene are under tight control. Very low basal levels of ACC[2:4:5]mRNA are always observed, confirming the “housekeeping” role that was suspected for PI1 on the basis of its structure.5.6 However, it should be noted that PI1 contains various cis elements and is highly regulated by various hormones.‘y.2o In this sense, PI1 is involved in the regulation of ACC gene expression other than merely housekeeping functions. The basal steady-state levels for ACC mRNA in the liver have been estimated to be approximately 4 to 15 molecules per cell. Such a low copy number indicates that only the mRNA for RNA polymerase II would be less abundant than ACC mRNA among about 40 enzymes that were examined.*’ The rarity of ACC mRNA is clearly manifested in the primer extension experiment shown in Fig 1. A very faint signal ACC GENE EXPRESION
(percent of hepatic response)
(-1
(-1
Fig 4. Response of the ACC gene in diabetic rats elicited by dietary induction. Primer extension was used to analyze ACC mRNAs in total liver RNA (256 pg/assay) prepared from animals treated as follows: diabetic rats subjected to no treatment (lane 1, D), diabetic rats starved for 72 hours (lane 2, D/S), diabetic rats starved and refed a complete standard chow (lane 3, D/S/N), diabetic rats starved and refed a fat-free diet (lane 4, D/S/F) for 24 hours, normal rats starved and refed fat-free diet (lane 5, H/S/F). The sample in lane 6 is transfer RNA. Pl (solid arrow) and P2 (open arrow) are as in Fig 1.
that had
been starved and then refed with fat-free diet further augments the activation of PI. DISCUSSION
The gene for ACC contains two promoters whose activities are differentially affected under different physiological conditions. The simultaneous activity of both promoters gives rise to several species of ACC mRNAs which show sequence heterogeneity in their 5’ UTR.4S5The presence of these isoforms of ACC mRNA raised questions concerning how different physiological conditions affect their forma-
Eti
izin
NF @ insulin
Fig 5. Effects of diet and insulin on the ACC gene response of diabetic rats. ACC mRNAs in total liver RNA preparations (200 pg/assay) from animals treated as follows were analyzed: All rats were starved for 46 hours and then refed for 24 hours with a standard laboratory chow [lanes NF) or with a fat-free diet (lanes FFD). Half of the diabetic rats were also treated with insulin (subcutaneous injections of 1.6 U every 6 hours) during the refeeding period. Densitometric analysis of autoradiograms were used to quantitate the ACC gene responses. PI responses were calculated from the relative levels of ACC[l:4:5]mRNA (pAU) end PII responses were estimated from the levels of ACC[2:4:5]mRNA (FL56). The values were then standardized, using the responses seen in the fat-free diet plus insulin treatment as 160%. Illustrated in the graph are the average values for the PI response (shadowed bars) and for the PII response (open bars) with their standard deviations (lines et the top of the bars). NF, rats refed standard chow; FFD, rats refed fat-free diet; (+) rats given insulin; (-) rats not given insulin; n, number of samples for each group.
206
LbPEZ-CASILLAS,
that can be assigned to ACC[2:4:5]mRNA, is the only trace of ACC mRNAs found under normal conditions. The level can be significantly increased by refeeding a fat-free diet to starved animals undergoing active lipogenesis. Importantly, this increase in ACC gene products include ACC[1:4: S]mRNA, whose synthesis requires the activity of PI. ACC[1:4:S]mRNA can reach levels of approximately 25 molecules per cell under these conditions. If we take into account the other ACC mRNA species generated by PI1 which are also activated by this dietary manipulation, we can estimate an overall increase in the ACC mRNA pool of at least one or two orders of magnitude, as compared with that of the basal level. Kinetic analysis of this induction is also very revealing. The quick response exhibited by PI1 contrasts markedly with the lag period exhibited by the PI response. This difference is further manifested by the sustained activity of PI, as opposed to the decreased activity of PII, as long as the animals are kept on a fat-free diet. This evidence supports the hypothesis that different mechanisms are responsible for the activation of each promoter. It is clear that PI is normally repressed, and that to express its full activity, it is necessary to subject the animal to dietary manipulations which stimulate lipogenesis, such as refeeding starved rats with a fat-free diet. Refeeding with a complete diet containing 4.5% fat diminishes this response. The kinetics of deinduction (reinstatement of starvation, after the fat-free diet refeeding) showed the dramatic dependency that the ACC[1:4:5]mRNA has on the fat-free diet. The rapid disappearance of ACC[1:4:5]mRNA, as compared with ACC[2:4:5]mRNA, suggests the existence of factors that stabilize this mRNA and that such factors contribute to the increased amount of the mRNA observed after refeeding. Thus, it seems that a fat-free diet can affect the level of class 1 ACC mRNAs in two ways: at the transcriptional level, by activating PI, and at the posttranscriptional level, by stabilizing the PI transcriptional products (class 1 ACC mRNAs). The nature of the factors provided by the fat-free diet remains unknown. However, the studies we have performed with diabetic animals seem to rule out the assumption that insulin is the only factor involved.‘“,‘” Since the injection of insulin does not activate either promoter of the ACC gene and since refeeding a fat-free diet in the absence of insulin is enough to elicit a complete ACC gene response in starved diabetic animals, it is
PONCE-CASTAfJEDA,
AND KIM
reasonable to conclude that at least two factors are mediating the dietary induction of ACC in the rat liver. One factor is insulin, and the other is an unidentified entity, which is presumably provided by or induced when these animals are fed a fat-free diet. Activation of the PI promoter seems to be absolutely dependent on the presence of these two factors. However, the need for both factors appears to be peculiar to hepatic tissue; epididymal white adipose tissue is fully responsive to insulin in the absence of the factor provided by the fat-free diet. This tissue specificity may explain why the PI promoter is constitutively expressed in white adipose tissue while it is stringently “inducible” in the liver. The discovery of the complex regulation of hepatic ACC mRNA species gives rise to many questions, such as why the regulation of hepatic ACC requires two classes of ACC mRNA. Our in vitro studies provide suggestive evidence concerning the roles played by structurally different ACC mRNAs. In the rabbit reticulocyte translation system, ACC[1:4:5]mRNA is translated up to about 10 times more efficiently than ACC[2:4:5]mRNA and produces peptides with extended N-termini (F. Gpez-Casillas and K.-H. Kim, submitted). Translational efficiency and stringent control of the PI promoter could have a significant impact on the amount of carboxylase in the liver during the fat-free diet refeeding. Correlations between the time course of induction of ACC mRNAs and the induction of the enzyme suggest that the differences in translational efficiency measured in vitro also occur in vivo. The de novo synthesis of carboxylase parallels the time course for the appearance of class 1 ACC transcripts; that is, after refeeding of a fat-free diet, there is a lag period of at least 6 hours for both the ACC[l:4:5]mRNA and the de novo enzyme. In contrast, the activation of PI1 does not exhibit such a lag phase; within 6 hours of refeeding, the amount of ACC[2:4: S]mRNA is already at 50% of the maximal level. This lack of temporal correlation indicates that even when class 2 ACC transcripts are already available for the synthesis of new ACC molecules, they are not immediately utilized. Whether this is the result of their intrinsically poor translational efficiency, or a reflection of the lack of essential factors required for their translation, remains to be determined. However, these findings strongly suggest that some form of translational control resides in and accounts for the heterogeneity of the 5’ ends of hepatic ACC mRNAs.
REFERENCES 1. Wakil SJ, Stoops JK, Joshi VC: Fatty acid synthesis and its regulation. Ann Rev Biochem 52:537-579, 1983 2. Numa S, Tanabe T: Fatty Acid Metabolism and Its Regulation. Amsterdam, The Netherlands, Elsevier, 1984, pp l-27 3. Kim K-H, tipez-Casillas F, Bai D-H, et al: Role of reversible phosphoxylation of ace@-CoA carboxylase in long-chain fatty acid synthesis. FASEB J 3:2250-2256,1989 4. tipez-Casillas F, Kim K-H: Heterogeneity at the 5’ end of rat acetyl-CoA carboxylase mRNA. J Biol Chem 264:7176-7184, 1989 5. tipez-Casillas F, Luo X, Kong I-S, et al: Characterization of different forms of rat mammary gland acetyl-CoA carboxylase
mRNA: Analysis of heterogeneity in the 5’ end. Gene 83:311-319, 1989 6. Luo X, Park K, tipez-Casillas F, et al: Structural features of the acetyl-CoA carboxylase gene: Mechanisms for the generation of mRNA with 5’ end heterogeneity. Proc Nat1 Acad Sci USA 86:4042-4046, 1989 7. tipez-Casillas F, Bai D-H, Luo X. et al: Structure of the coding sequence and primary amino acid sequence of acetyl-CoA carboxylase. Proc Nat1 Acad Sci USA 85:5784-5788, 1988 8. Pape ME, Mpez-Casillas F, Kim K-H: Physiological regulation of acetyl-CoA carboxylase gene expression: Effect of diet.
REGULATION
OF ACETYL-CoA
CARBOXYLASE
GENE EXPRESSION
diabetes, and lactation on acetyl-CoA carboxylase mRNA. Arch Biochem Biophys 267:104-109,1988 9. Pape ME, Kim K-H: Effect of tumor necrosis factor on acetyl-CoA carboxylase gene expression and preadipocyte differentiation. Mol Endocrinol2:395-403,1988 10. Pape ME, Kim K-H: Transcriptional regulation of acetylCoA carboxylase gene expression by tumor necrosis factor in 3OA5 preadipocytes. Mol Cell Biol9:974-982,1989 11. Batenburg JJ, Whitsett JA: Levels of mRNAs coding for lipogenic enzymes in rat lung upon fasting and refeeding and during perinatal development. Biochim Biophys Acta 1006:329334,1989 12. Coupe C, Perdereau D, Ferre P, et al: Lipogenic enzyme activities and mRNA in rat adipose tissue during weaning: Role of the diet. Am J Physiol258:E126-E133,1990 13. Katsurada A, Iritani N, Fukuda H, et al: Effects of nutrients and hormones on transcriptional and post-transcriptional regulation of acetyl-CoA carboxylase in rat liver. Eur J Biochem 190:435441,199o 14. MacDonald RJ, Swift GH, Przybyla AE, et al: Isolation of RNA using guanidinium salts. Methods Enzymol 152:219-227, 1987
207
15. Bertholet C, Van Meir E, Heggeler-Bordier B, et al: Vaccinia virus produces late mRNAs by discontinuous synthesis. Cell 50:153-162,1987 16. Taylor WL, Collier KJ, Deschenes RJ, et al: Sequence analysis of a cDNA coding for a pancreatic precursor to somatostatin. Proc Nat1 Acad Sci USA 78:6694-6698, 1981 17. Berger FG, Loose D, Meisner H, et al: Androgen induction of messenger RNA concentrations in mouse kidney is posttranscriptional. Biochemistry 25:1170-1175,1986 18. Nasrin N, Ercolani L, Denaro M, et al: An insulin response element in the glyceraldehyde-3-phosphate dehydrogenase binds a nuclear protein induced by insulin in cultured cells and nutritional manipulations in vivo. Proc Nat1 Acad Sci USA 87:5273-5277,199O 19. Luo X, Kim K-H: An enhancer element in the housekeeping promoter for acetyl-CoAcarboxylase gene. Nucl Acids Res 18:32493254,199O 20. Park K, Kim K-H: Regulation of acetyl-CoA carboxylase gene expression: insulin induction of acetyl-CoA carboxylase and differentiation of 30A5 preadipocytes requires prior CAMP action on the gene. J Biol Chem 266:12249-12256,199l 21. Galau GA, Klein WH, Britten RJ, et al: Significance of rare mRNA species in liver. Arch Biochem Biophys 179:584-599,1977
A Carboxylase
mRNA Metabolism
Fernando Lbpez-Casillas, M. Verhica
Ponce-Castatieda,
in the Rat Liver
and Ki-Han Kim
The acetyl-coentyme A carboxylase (ACC) gene contains two promoters (PI and PII), both of which are active in the liver. Various physiological stimuli affect one, or both of the promoters of the ACC gene, and result in the generation of two classes of ACC mRNAs which differ in the composition of their 5’ untranslated regions (5’ UTR). We have analyzed the amounts of the two major mRNAs species that are generated from each of these promoters in order to examine the regulation of ACC gene activity in the liver under different physiological conditions. Our findings can be summarized as follows: (1) In liver from normal animals, fed a complete laboratory chow ad libitum, the level of class 2 ACC mRNA species generated by PII is very low. These mRNA species disappear on starvation. Refeeding starved animals with a fat-free diet stimulates both PI and PII with different time courses of induction: PII responds quickly and PII gene products accumulate to maximum levels within 18 hours, while the PI response, as measured by the accumulation of class 1 mRNAs, shows a lag period of 8 hours before reaching maximal levels at the end of a 24.hour refeeding period. The half-lives estimated from the induction kinetics were 4.4 hours for class 2 mRNAs and 11.8 hours for class 1 mRNAs. Reinstatement of starvation causes an almost instantaneous disappearance of class 1 mRNA species, as compared with class 2 mRNA species. This rapid decay of PI transcripts suggests that factors stabilizing this class of ACC mRNAs contribute to the steady-state levels reached after the dietary induction. (2) In starved streptozotocin-diabetic rats, refeeding with a fat-free diet causes activation of both promoters, whereas refeeding with the standard diet has no effect. This indicates that activation of the promoters occurs in the absence of insulin. However, administration of insulin to these animals further activates both promoters, regardless of the type of diet given to the starved animals. There is no insulin effect in diabetic animals that are fed normally. These observations indicate that in the liver two ACC gene promoters are selectively regulated under different physiological conditions. Insulin is one of the factors that affect the ACC gene. However, the factor(s) that is responsible for the activation of the ACC gene under the lipogenic conditions created by administration of a fat-free diet is not insulin. Copyright 0 1992 by W.B. Saunders Company
A
CETYL-coenzyme A carboxylase (ACC) catalyzes the rate-limiting step in the biosynthesis of fatty acids.‘.’ Characterization of the 5’ end of the ACC gene has shown that the ACC gene contains two promoters, designated PI and PII, which can promote the transcription of two distinct classes of ACC mRNA, class 1 and class 2, respectively.“’ PI is located upstream to the first exon and PI1 is located upstream to the second exon of the ACC gene. These two exons, exon 1 and 2, together with exons 3, 4, and 5, compose the 5’ region of the gene. The translation initiation AUG codon of the ACC open-reading frame (ORF) is located in exon 5. Transcription, followed by alternative splicing of the first five exons of the ACC gene generates several forms of ACC mRNA, which can be classified into class 1 or class 2 depending on whether they contain exon 1 or 2, which is determined by whether PI or PI1 is used. The major ACC mRNA species in class 1 is ACC[ 1:4:5]mRNA, which contains exon 1, 4, and 5 sequences at its 5’ end, whereas the major class 2 species is ACC[2:4:5]mRNA, which contains sequences for exons 2,4, and 5 at the 5’ end. The current nomenclature denotes the composition of the exons at the 5’ end of the mRNA. Important differences among these ACC mRNAs appears to be restricted to their 5’ untranslated regions (UTR). All species of ACC mRNA contain exon 5 in their 5’ UTR. Prior to the demonstration of the 5’ end heterogeneity of ACC mRNA, ACC gene activity under different physiological conditions had been investigated by measuring the total amount of ACC mRNA through dot blot or Northern analysis using cDNA probes from the coding or 3’ UTR of the mRNA.8-13Although the results of such studies reflect the patterns of total ACC mRNA metabolism, they cannot distinguish the true nature of ACC gene activity, nor the nature of ACC mRNAs. Because the two ACC gene promoters found in the liver respond differently to particuMetabolism,
Vol41, No 2 (February), 1992: pp 201.207
lar physiological conditions, meaningful studies of hepatic ACC mRNA require the use of procedures that identify the activities of PI and PII. In the present report, we have used primer extension analysis on total RNA using a specifically designed primer. With this approach, we have been able to measure the relative levels of the transcriptional products generated by PI and PI1 in the total ACC mRNA population. This has allowed us to investigate changes in the content of specific ACC mRNAs and the ACC promoter activities that occur in rat liver under different experimental conditions. MATERIALS
AND METHODS
Materials
Biochemicals were purchased as follows: [r-“P] adenosine triphosphate (ATP) (6,000 Ci/mmol) from DuPont-New England Nuclear, Boston, MA, avian myeloblastosis virus (AMV) reverse transcriptase from Life Sciences; T4 polynucleotide kinase from United States Biochemicals; porcine zinc insulin from CalbiochemBehring; streptozotocin from Sigma Chemical, St Louis, MO; Sl nuclease from Pharmacia LKB Biotechnology, Uppsala, Sweden; guanidinium thiocyanate from Fluka; XARS x-ray films from Kodak, Rochester, NY. The llO-mer and the 108-mer (263/H2/
From the Biochemistry Department, West Lafayette, IN. Supported by National Institutes of Health Grant No. CA 46882 and DK 12865. This is Journal Paper No. 12808from the Agricultural Experimentation Station, Purdue University. Dr L6pez-Casilla’s present address is Howard Hughes Medical Institute and Cell Biology and Genetics Program, Memorial SloanKettering Cancer Center, New York, NY10021. Address reprint requests to &Han Kim, PhD, Biochemistry Department, Purdue University, West Lafayette, IN 47907. Copyright 0 1992 by W.B. Saunders Company 00260495/92/4IO2-0020$03.00/O 201
LOPEZ-CASILLAS, PONCE-CASTAfJEDA,
202
Taq) primers used in the primer extension analysis were synthesized in an Applied Biosystems 380A DNA synthesizer, and were purified by urea/polyacryiamide gel eiectrophoresis (PAGE) as described.4 All other chemicals were reagent grade. Animals Male Wistar rats (1.50 to 250 g) were from our departmental colonies. The complete standard chow, Purina’s Rodent Laboratory Chow 5001 (Lafayette, IN), contains 4.5% of fat. The fat-free test diet from United States Biochemicais (Cleveland, OH) contains ceiiufii, 16.45%; salt mixture, 4%; sucrose, 58.45%; and casein, 21.1%, supplemented with vitamins. Rats were made diabetic by the administration of streptozotocin (65 mg/kg, subcutaneously) as described previously.” RNA Preparations Ail RNA samples were prepared from frozen tissues following the guanidinium-thiocyanate method described by MacDonald et ai.l4 RNA amounts were determined by UV absorbance at 260 nm (a 40+g/mL RNA aqueous solution providing 1 optical density unit at 260 nm).
diet to the starved animals induces an increase in ACC of up to 60-fold.” In Fig 1, the results of primer extension analysis of total liver RNA preparations from animals under different nutritional conditions are shown. The extension products identified by the arrows are the major ACC mRNA species generated as a result of PI and PII activity. The 415nt long extension product represents the ACC[l:4:5]mRNA species (closed arrow) and the 269- to 264-nt long extension products (open arrow) represent the ACC [2:4:5]mRNA species in the RNA preparations. RNA samples from normal animals fed a complete standard chow ad libitum, contain no PI promoter products (ACC[l:4: S]mRNA) and have extremely low basal levels of PI1 products (ACC[2:4:5]mRNA). The left side of Fig 1 shows the time course of decay of the basal hepatic levels of ACC mRNA, during starvation. The Iane designated “0” corresponds to the basal levels found before starvation in animals fed a standard laboratory chow ad libitum. The only primer extension products found under these condiREFEEDING
Primer Extension Analysis
ST~~IVITKIY
The primer extension procedure of Berthoiet et al” and the linked primer extension/S1 nuclease protection experiment were performed as described previousiyP Primers were labeled at their 5’ ends with the use of T4 poiynucieotide kinase and [y-‘Zp]ATP as described.16 The two primers used in these experiments were a llO-mer and a lob-mer (263/H2/Taq). A description of the 108-mer and the characterization of the primer extension products it generates have been reported in detail.‘,’ These primers recognize and anneal to ail the species of ACC mRNA because the sequences to which they are complementary derive from exon 5, which exists in all the forms of ACC mRNA characterized to date. The llO-mer has the same sequence as the original 263/H2/Taq (lob-mer), except that it has two more bases at its 3’ end. These extra bases do not affect hybridization to ail forms of ACC mRNA. Also, the sizes of the primer extension products generated from either primer are indistinguishable, and the sizes are solely determined by the type of ACC mRNA to which they hybridize. For example, when these primers hybridize to and are extended on the ACC[1:4:5]mRNA, the are elongated to a length of 415 nucieotide (nt).” On the other hand, when the primers hybridize to and extend on the ACC[2:4:5]mRNA, the extension products are 269 to 264 nt in length. Thus, the sizes of the primer extension are diagnostic of the type of ACC mRNAs present in the analyzed sample. In our experiments, we have used a molar excess of the primers so that ail the ACC mRNAs in the sample could be extended. When the products of the reverse transcriptase reaction are resolved in urea/PAGE and autoradiographed, the autoradiograms show not only the specific species of ACC mRNAs in the sample, but also the relative amounts of each species. Quantitation of the signals derived from the autoradiograms was performed by densitometric analysis using a LKB 2400 Geiscan XL software package. Statistical calculations were performed with the System for Elementary Statistical Analysis Package from the SAS Institute, Gary, NC. RESULTS
Effects of Starvation and Refeeding on ACC mRNA Metabolism in Rat Liver
When rats are fasted, fatty acid synthesis is diminished and ACC virtually disappears.’ However, feeding a fat-free
AND KIM
m
404-
0
309-
a
242-
!ad’z
hrr
243@4@72 ---_m--------
0 12 __--lFNklFNF
hrs
24
48
0 ; 0
i@O- * a
1io-
c
-mu.
Fig 1. Hepatlc ACC mRNA metabolism during starvation and refeeding. Total RNA was prepared from livers of rats subjected to the indicated treatments and 200 pg was analyzed for ACC mRNA by primer extension. The starvation tlme zero aample corresponds to rats fed e complete stendard chow ad Ilbitum. Starvation (72 hours) was followed by refeeding of e complete standard chow (lanes N) or a fat-free diet (lanes F). The animals were killed at different times during the starvation/refeeding period as Indlcated at the top of the figure. PI (solid arrow) indkatr the primer extension product originated from ACC[1:4:5JmRNA (415 nt). P2 (open arrow) Indicates the primer extension products that originated from ACC[2:4:S]mRNA (269 to 264 nt). The migration of the unextended 1gGmer used in the experiment is indicated by the label “primer” at the bottom of the flgure. Lanes m show the migration of the molecular weight markers (5’ end ryP]labeled Hmll fragments of the pWW22). whose sizes are indicated in nucleotldes at the left margin.
REGULATION OF ACElYL-CoA
203
CARBOXYLASE GENE EXPRESSION
tions are trace amounts of extension products of about 269 to 264 nt (P2, open arrow), which originate from ACC[2:4: S]mRNA. The complete absence of 415nt long primer extension products indicates that ACC[1:4:5]mRNA is not present in this sample. During starvation, the extremely low level of ACC[2:4:5]mRNA decreases until it is undetectable. Refeeding starved animals causes an increase in the levels of both classes of ACC mRNA (right side of Fig 1). The composition of the diet used for refeeding affects this response. Full induction of both species of ACC mRNAs only occurs when the starved animals are fed a fat-free diet. The effects of fat-free diet on the activation of PI are particularly striking (F lanes in Fig 1). Refeeding a standard diet also causes an increase in the amount of ACC mRNAs (lanes N in Fig 1); this increase is mainly due to activation of PII. The results in Fig 1 signify the importance of the present approach to examining gene activities as opposed to Northern or dot blot analyses of mRNA populations. These latter analytical methods would not show the underlying gene activity that is involved in the induction of ACC under these experimental conditions. The absolute ratio of the levels of ACC[2:4:5]mRNA and ACC[1:4:5]mRNA reached after 24 hours on the fat-free diet varies because of the high variability in the response of the PI promoter that occurs from animal to animal. However, we have observed that although some rats produce stronger PI responses than others, on the average, the ratio of PI1 products to PI products is about 3:2 after 24 hours on the fat-free diet. Kinetic Analysis of the Dietary Induction of the ACC Gene in Rat Liver Each experimental point in Fig 1 was obtained from total RNA preparations from three animals. To examine the kinetics of mRNA induction and decay in a more reliable manner, the same experiments were performed three times. The responses of PI and PI1 were analyzed and standardized using the values at 24 hours of refeeding. A summary of these experiments is shown in Fig 2. Substantially different kinetics for the induction of the activities of the ACC gene promoters are found when starved animals are refed with a fat-free diet. PII is activated immediately when refeeding begins; by the sixth hour, almost 50% of the maximum response is observed, and within 18 hours, it has reached its maximal level, which is followed by a slight decrease in the levels of the PI1 products, even when the animals are kept on a fat-free diet. On the other hand, the response of the PI promoter shows a lag period of about 6 hours. During the first 6 hours of the refeeding period, PI remains inactive, reaching the 50% and 100% induction levels only after 18 and 24 hours, respectively. If the animals are kept in a fat-free diet, PI products remain at this maximal level for the next 24 hours. From the kinetics of induction elicited by refeeding with the fat-free diet, we estimated the half-life of ACC[2:4: S]mRNA as 4.4 hours and the, half-life of ACC[1:4: S]mRNA as 11.8 hours.” Removing food after the 24th hour causes an almost immediate disappearance of the PI transcriptional products. ACC[1:4:5]mRNA decreases to
q Pn wx![2:4:51mRNAl
T
T
T
12
16
24
0
PI (Acc[I:4:5lmRNAl
OL 0
6
REFEEDING
30
48
@ml
Fig 2. Time course of the dietary induction of hepatic ACC mRNAs. The relative levels of each one of the major products from PI or PII during the refeeding of a fat-free diet were plotted. The upper half of the standard deviation (from four determinations) for each point is shown as lines parting from the plotted average value. Time zero corresponds to initiation of refaeding after 72 hours of starvation. Squares show the relative amounts of ACC[2:4:5jmRNA. Circles show the relative amounts of ACC[1:4:5]mRNA. The values obtained from animals that were returned to starvation after 24 hours of refeeding ara indicated by solid symbols and broken lines.
approximately one tenth of the previous levels within the first 6 hours of the restarvation period (Fig 2). On the other hand, the levels of the PI1 products require longer times to show a comparable decrease. After 24 hours of restarvation, the levels of ACC[2:4:5]mRNA have decreased to approximately one quarter of the previous levels. It is clear that the PI products are more sensitive to the reinstatement of starvation, since they disappear quickly during starvation. This suggests that the accumulation of class 1 ACC mRNAs observed during refeeding with a fat-free diet is the result of two simultaneous processes: an increase in the transcriptional activity of PI and an increase in the stability of the PI products. On the other hand, it appears that the major factor contributing to the new higher steady-state levels of class 2 ACC mRNAs during the same treatment is the increased transcriptional activity of the PI1 ACC gene promoter. Insulin Effects on ACC mRNA Metabolism in the Livers of Diabetic Rats The transcriptional effects on several hepatic genes elicited by a fat-free diet appear to be mediated by the increased levels of insulin provoked by this dietary regimen.‘3~‘sTherefore, we have examined the role played by insulin in the induction of the ACC gene when starved rats are refed with a fat-free diet. Insulin was administered to streptozotocin-diabetic rats and ACC mRNAs were analyzed by primer extension. The hepatic levels of ACC mRNAs, which are primarily represented by class 2 mRNA species (lanes 1 and 2, Fig 3A) under diabetic conditions, did not significantly increase in response to insulin (lanes 3-8, Fig 3A). The diabetic rat exhibits basal levels of
204
LCPEZ-CASILLAS, PONCE-CASTAI;IEDA, AND KIM
LIVER Diabetic INSULIN (-)
---w
1.5
INSULIN
Rata
P(-)
hrs
3
6
1.5 3 --
(hrsj
- 6
527 -
527-
m
Fig 3. Responses of the ACC gene in diabetic rats elicited by insulin. Insulin was administered to diabetic rats (1.6 hepatic ACC mRNA response (A) and ACC mRNA response in the epidydimsl fat tissues (B) were analyzed in 250 extension. Samples from animals not given insulin are shown in lanes 1 and 2. Following injection of insulin, three indicated times. The samole in lane g came from normal animals subjected to the starvation/fat-free diet refeeding arrow), and P2 (open arrow) are as in Fig 1.
ACC[2:4:5]mRNA, and shows no ACC[ 1:4:5]mRNA, as is the case with normal animals. The experimental results shown in Fig 3A demonstrate that insulin alone failed to activate either ACC gene promoter. The levels of ACC[2:4: S]mRNA and ACC[1:4:5]mRNA were not significantly affected by the administration of insulin. Even when insulin treatment is continued for a period of 24 hours, the ACC mRNA levels remained unaffected (not shown). This lack of response in the liver is in sharp contrast to the quick responses exhibited by white adipose tissue in the same animals. Analysis of the ACC mRNAs in epididymal fat pads of the rats shown in Fig 3B indicates that class 1 ACC mRNAs predominate in fat tissues, and that ACC[1:4: S]mRNA increases to about three times the basal level approximately 6 hours after the administration of insulin. The adipose tissue responsiveness in these animals rules out the possibility that the lack of response found in liver indicates a generalized failure of our experimental animals to respond to hormone treatment. Our liver studies appear to be in disagreement with the data of Katsurada et al,” which show a twofold increase in the levels of hepatic ACC mRNA after insulin was administered to diabetic rats. The main difference between our experiments and those of Katsurada et al resides in the of their diabetic rats to a high glucose “adaptation” (fat-free) diet before insulin was administered. In our studies, diabetic rats were kept on a complete standard chow before insulin treatment. The possibility that feeding
U subcutaneously) and the pg of total RNA by primer animals were killed at the scheme. Lanes m, PI (solid
a fat-free diet causes a substantial difference in the activation of ACC promoters was substantiated by our starvationrefeeding experiments with diabetic rats, as shown in Fig 4. When diabetic rats are subjected to the starvation/ refeeding protocol without insulin treatment, refeeding with a fat-free diet causes the activation of both ACC promoters as shown by the presence of both classes of ACC mRNAs (lane 4, Fig 4). In contrast, refeeding starved diabetic animals with standard chow does not activate either of the ACC promoters (lane 3, Fig 4). Although this activation is quantitatively smaller than the one observed in normal animals (lane 5, Fig 4) it is significant in that the ACC promoters in the livers of diabetic rats are activated by a fat-free, carbohydrate-rich diet in the absence of insulin. On the other hand, the observation that both promoters are activated in the livers of normal animals which have been subjected to starvation and refed with the standard diet (Fig l), while this does not occur in diabetic animals, suggests that insulin is one of the factors needed for the activation of the promoters. Indeed, one can demonstrate the effect of insulin on the accumulation of both forms of ACC mRNA species in diabetic animals as shown in Fig 5. The administration of insulin to diabetic animals which had been subjected to starvation and refeeding with the standard diet clearly shows the activation of both PI and PI1 and establishes that insulin is one of the factors involved in activation. The administration of insulin to diabetic animals
REGULATION
OF ACETYL-CoA
CARBOXYLASE
205
GENE EXPRESSION
4Pl
cJp2
tion and what physiological purpose they could serve in the cell. To address the first question, we have analyzed, by means of primer extension experiments, how the levels of the major hepatic ACC mRNA species are affected by two important physiological conditions that affect lipogenesis: nutritional status and diabetes. The primer extension experiment was designed so that in a single experiment, we could unambiguously distinguish and quantitate each one of the diverse ACC mRNA forms and thus the activities of the two promoters. This technique directly assesses the activity of each one of the two ACC gene promoters, a task that cannot be accomplished by Northern or dot blot analysis. Analyzing ACC gene activity by means of primer extension showed that the promoters of the ACC gene are under tight control. Very low basal levels of ACC[2:4:5]mRNA are always observed, confirming the “housekeeping” role that was suspected for PI1 on the basis of its structure.5.6 However, it should be noted that PI1 contains various cis elements and is highly regulated by various hormones.‘y.2o In this sense, PI1 is involved in the regulation of ACC gene expression other than merely housekeeping functions. The basal steady-state levels for ACC mRNA in the liver have been estimated to be approximately 4 to 15 molecules per cell. Such a low copy number indicates that only the mRNA for RNA polymerase II would be less abundant than ACC mRNA among about 40 enzymes that were examined.*’ The rarity of ACC mRNA is clearly manifested in the primer extension experiment shown in Fig 1. A very faint signal ACC GENE EXPRESION
(percent of hepatic response)
(-1
(-1
Fig 4. Response of the ACC gene in diabetic rats elicited by dietary induction. Primer extension was used to analyze ACC mRNAs in total liver RNA (256 pg/assay) prepared from animals treated as follows: diabetic rats subjected to no treatment (lane 1, D), diabetic rats starved for 72 hours (lane 2, D/S), diabetic rats starved and refed a complete standard chow (lane 3, D/S/N), diabetic rats starved and refed a fat-free diet (lane 4, D/S/F) for 24 hours, normal rats starved and refed fat-free diet (lane 5, H/S/F). The sample in lane 6 is transfer RNA. Pl (solid arrow) and P2 (open arrow) are as in Fig 1.
that had
been starved and then refed with fat-free diet further augments the activation of PI. DISCUSSION
The gene for ACC contains two promoters whose activities are differentially affected under different physiological conditions. The simultaneous activity of both promoters gives rise to several species of ACC mRNAs which show sequence heterogeneity in their 5’ UTR.4S5The presence of these isoforms of ACC mRNA raised questions concerning how different physiological conditions affect their forma-
Eti
izin
NF @ insulin
Fig 5. Effects of diet and insulin on the ACC gene response of diabetic rats. ACC mRNAs in total liver RNA preparations (200 pg/assay) from animals treated as follows were analyzed: All rats were starved for 46 hours and then refed for 24 hours with a standard laboratory chow [lanes NF) or with a fat-free diet (lanes FFD). Half of the diabetic rats were also treated with insulin (subcutaneous injections of 1.6 U every 6 hours) during the refeeding period. Densitometric analysis of autoradiograms were used to quantitate the ACC gene responses. PI responses were calculated from the relative levels of ACC[l:4:5]mRNA (pAU) end PII responses were estimated from the levels of ACC[2:4:5]mRNA (FL56). The values were then standardized, using the responses seen in the fat-free diet plus insulin treatment as 160%. Illustrated in the graph are the average values for the PI response (shadowed bars) and for the PII response (open bars) with their standard deviations (lines et the top of the bars). NF, rats refed standard chow; FFD, rats refed fat-free diet; (+) rats given insulin; (-) rats not given insulin; n, number of samples for each group.
206
LbPEZ-CASILLAS,
that can be assigned to ACC[2:4:5]mRNA, is the only trace of ACC mRNAs found under normal conditions. The level can be significantly increased by refeeding a fat-free diet to starved animals undergoing active lipogenesis. Importantly, this increase in ACC gene products include ACC[1:4: S]mRNA, whose synthesis requires the activity of PI. ACC[1:4:S]mRNA can reach levels of approximately 25 molecules per cell under these conditions. If we take into account the other ACC mRNA species generated by PI1 which are also activated by this dietary manipulation, we can estimate an overall increase in the ACC mRNA pool of at least one or two orders of magnitude, as compared with that of the basal level. Kinetic analysis of this induction is also very revealing. The quick response exhibited by PI1 contrasts markedly with the lag period exhibited by the PI response. This difference is further manifested by the sustained activity of PI, as opposed to the decreased activity of PII, as long as the animals are kept on a fat-free diet. This evidence supports the hypothesis that different mechanisms are responsible for the activation of each promoter. It is clear that PI is normally repressed, and that to express its full activity, it is necessary to subject the animal to dietary manipulations which stimulate lipogenesis, such as refeeding starved rats with a fat-free diet. Refeeding with a complete diet containing 4.5% fat diminishes this response. The kinetics of deinduction (reinstatement of starvation, after the fat-free diet refeeding) showed the dramatic dependency that the ACC[1:4:5]mRNA has on the fat-free diet. The rapid disappearance of ACC[1:4:5]mRNA, as compared with ACC[2:4:5]mRNA, suggests the existence of factors that stabilize this mRNA and that such factors contribute to the increased amount of the mRNA observed after refeeding. Thus, it seems that a fat-free diet can affect the level of class 1 ACC mRNAs in two ways: at the transcriptional level, by activating PI, and at the posttranscriptional level, by stabilizing the PI transcriptional products (class 1 ACC mRNAs). The nature of the factors provided by the fat-free diet remains unknown. However, the studies we have performed with diabetic animals seem to rule out the assumption that insulin is the only factor involved.‘“,‘” Since the injection of insulin does not activate either promoter of the ACC gene and since refeeding a fat-free diet in the absence of insulin is enough to elicit a complete ACC gene response in starved diabetic animals, it is
PONCE-CASTAfJEDA,
AND KIM
reasonable to conclude that at least two factors are mediating the dietary induction of ACC in the rat liver. One factor is insulin, and the other is an unidentified entity, which is presumably provided by or induced when these animals are fed a fat-free diet. Activation of the PI promoter seems to be absolutely dependent on the presence of these two factors. However, the need for both factors appears to be peculiar to hepatic tissue; epididymal white adipose tissue is fully responsive to insulin in the absence of the factor provided by the fat-free diet. This tissue specificity may explain why the PI promoter is constitutively expressed in white adipose tissue while it is stringently “inducible” in the liver. The discovery of the complex regulation of hepatic ACC mRNA species gives rise to many questions, such as why the regulation of hepatic ACC requires two classes of ACC mRNA. Our in vitro studies provide suggestive evidence concerning the roles played by structurally different ACC mRNAs. In the rabbit reticulocyte translation system, ACC[1:4:5]mRNA is translated up to about 10 times more efficiently than ACC[2:4:5]mRNA and produces peptides with extended N-termini (F. Gpez-Casillas and K.-H. Kim, submitted). Translational efficiency and stringent control of the PI promoter could have a significant impact on the amount of carboxylase in the liver during the fat-free diet refeeding. Correlations between the time course of induction of ACC mRNAs and the induction of the enzyme suggest that the differences in translational efficiency measured in vitro also occur in vivo. The de novo synthesis of carboxylase parallels the time course for the appearance of class 1 ACC transcripts; that is, after refeeding of a fat-free diet, there is a lag period of at least 6 hours for both the ACC[l:4:5]mRNA and the de novo enzyme. In contrast, the activation of PI1 does not exhibit such a lag phase; within 6 hours of refeeding, the amount of ACC[2:4: S]mRNA is already at 50% of the maximal level. This lack of temporal correlation indicates that even when class 2 ACC transcripts are already available for the synthesis of new ACC molecules, they are not immediately utilized. Whether this is the result of their intrinsically poor translational efficiency, or a reflection of the lack of essential factors required for their translation, remains to be determined. However, these findings strongly suggest that some form of translational control resides in and accounts for the heterogeneity of the 5’ ends of hepatic ACC mRNAs.
REFERENCES 1. Wakil SJ, Stoops JK, Joshi VC: Fatty acid synthesis and its regulation. Ann Rev Biochem 52:537-579, 1983 2. Numa S, Tanabe T: Fatty Acid Metabolism and Its Regulation. Amsterdam, The Netherlands, Elsevier, 1984, pp l-27 3. Kim K-H, tipez-Casillas F, Bai D-H, et al: Role of reversible phosphoxylation of ace@-CoA carboxylase in long-chain fatty acid synthesis. FASEB J 3:2250-2256,1989 4. tipez-Casillas F, Kim K-H: Heterogeneity at the 5’ end of rat acetyl-CoA carboxylase mRNA. J Biol Chem 264:7176-7184, 1989 5. tipez-Casillas F, Luo X, Kong I-S, et al: Characterization of different forms of rat mammary gland acetyl-CoA carboxylase
mRNA: Analysis of heterogeneity in the 5’ end. Gene 83:311-319, 1989 6. Luo X, Park K, tipez-Casillas F, et al: Structural features of the acetyl-CoA carboxylase gene: Mechanisms for the generation of mRNA with 5’ end heterogeneity. Proc Nat1 Acad Sci USA 86:4042-4046, 1989 7. tipez-Casillas F, Bai D-H, Luo X. et al: Structure of the coding sequence and primary amino acid sequence of acetyl-CoA carboxylase. Proc Nat1 Acad Sci USA 85:5784-5788, 1988 8. Pape ME, Mpez-Casillas F, Kim K-H: Physiological regulation of acetyl-CoA carboxylase gene expression: Effect of diet.
REGULATION
OF ACETYL-CoA
CARBOXYLASE
GENE EXPRESSION
diabetes, and lactation on acetyl-CoA carboxylase mRNA. Arch Biochem Biophys 267:104-109,1988 9. Pape ME, Kim K-H: Effect of tumor necrosis factor on acetyl-CoA carboxylase gene expression and preadipocyte differentiation. Mol Endocrinol2:395-403,1988 10. Pape ME, Kim K-H: Transcriptional regulation of acetylCoA carboxylase gene expression by tumor necrosis factor in 3OA5 preadipocytes. Mol Cell Biol9:974-982,1989 11. Batenburg JJ, Whitsett JA: Levels of mRNAs coding for lipogenic enzymes in rat lung upon fasting and refeeding and during perinatal development. Biochim Biophys Acta 1006:329334,1989 12. Coupe C, Perdereau D, Ferre P, et al: Lipogenic enzyme activities and mRNA in rat adipose tissue during weaning: Role of the diet. Am J Physiol258:E126-E133,1990 13. Katsurada A, Iritani N, Fukuda H, et al: Effects of nutrients and hormones on transcriptional and post-transcriptional regulation of acetyl-CoA carboxylase in rat liver. Eur J Biochem 190:435441,199o 14. MacDonald RJ, Swift GH, Przybyla AE, et al: Isolation of RNA using guanidinium salts. Methods Enzymol 152:219-227, 1987
207
15. Bertholet C, Van Meir E, Heggeler-Bordier B, et al: Vaccinia virus produces late mRNAs by discontinuous synthesis. Cell 50:153-162,1987 16. Taylor WL, Collier KJ, Deschenes RJ, et al: Sequence analysis of a cDNA coding for a pancreatic precursor to somatostatin. Proc Nat1 Acad Sci USA 78:6694-6698, 1981 17. Berger FG, Loose D, Meisner H, et al: Androgen induction of messenger RNA concentrations in mouse kidney is posttranscriptional. Biochemistry 25:1170-1175,1986 18. Nasrin N, Ercolani L, Denaro M, et al: An insulin response element in the glyceraldehyde-3-phosphate dehydrogenase binds a nuclear protein induced by insulin in cultured cells and nutritional manipulations in vivo. Proc Nat1 Acad Sci USA 87:5273-5277,199O 19. Luo X, Kim K-H: An enhancer element in the housekeeping promoter for acetyl-CoAcarboxylase gene. Nucl Acids Res 18:32493254,199O 20. Park K, Kim K-H: Regulation of acetyl-CoA carboxylase gene expression: insulin induction of acetyl-CoA carboxylase and differentiation of 30A5 preadipocytes requires prior CAMP action on the gene. J Biol Chem 266:12249-12256,199l 21. Galau GA, Klein WH, Britten RJ, et al: Significance of rare mRNA species in liver. Arch Biochem Biophys 179:584-599,1977
