Biochemical, Molecular, and Genetic Mechanisms
Identification of the Regulatory Region of the L-Type Pyruvate Kinase Gene in Mouse Liver by Hydrodynamics-Based Gene Transfection1 Takayuki Suzuki, Masanobu Kawamoto, Atsushi Murai,2 and Tatsuo Muramatsu Department of Applied Molecular Biosciences, Graduate School of Bioagricultural Sciences, Nagoya University, Nagoya, Japan ABSTRACT Expression of L-type pyruvate kinase (L-PK) is upregulated in the liver by dietary carbohydrate. Previously, 3 carbohydrate/insulin response elements were identified in the 59-flanking region of the L-PK gene up to bp 170. Studies of the 59-flanking region beyond bp 183 in transgenic mice suggested that other regulatory elements may be present upstream of bp 183, but the positions of these elements were uncertain. In the present study, the existence of regulatory regions of the L-PK gene responding to stimulation by feeding was examined using in vivo hydrodynamics-based gene transfection (HT) in mouse liver. The firefly-luciferase (FL) gene, fused with various lengths of the 59-flanking region of the L-PK gene, was introduced into mouse liver by HT. The mice had free access to a high-carbohydrate diet. In liver homogenate, luciferase activity of pL-PK(1467)-FL (which included the 59-flanking region from bp 1467 to 117), was markedly stimulated by feeding. 59-Deletion up to bp 1065 caused only minor changes in luciferase activity, but further deletion up to bp 690 and bp 203 caused significant, gradual decreases in activity. Further analyses utilizing 59-deletion mutants indicated the existence of positive regulatory regions that respond to stimulation by feeding between bp 1065 and 945, and between 300 and 203 on the L-PK gene. These results suggest that unidentified cis-acting DNA elements exist in the upstream region of the L-PK gene, and that HT is a useful approach for detecting regulatory regions of genes expressed in the liver. J. Nutr. 136: 16–20, 2006. KEY WORDS: L-type pyruvate kinase mouse hydrodynamics-based gene transfection luciferase
Pyruvate kinase (PK)3 is a glycolytic enzyme that catalyzes the formation of pyruvate and ATP from phosphoenolpyruvate and ADP. PK has several isozymes that are expressed in a tissuespecific manner (1–3). L-type PK (L-PK) is an isozyme that is expressed mainly in liver, and to a lesser degree in kidney and intestine (1). Expression of the rat L-PK gene is upregulated by both dietary glucose and fructose (4–7); this carbohydratedependent regulation is achieved by interactions between cisacting DNA elements and trans-acting proteins (8–10). To date, several cis-acting DNA elements responding to glucose and fructose were identified in vitro, and their biological significance in liver was confirmed in vivo (11-16). Yamada et al. (16) reported that in transgenic mice, sequences ;3000 bp upstream of the rat L-PK gene contain all of the cis-acting DNA elements necessary for transcriptional activation after stimulation by a high-carbohydrate diet. In cultured hepatocytes, the 59-flanking region of the L-PK gene was analyzed by deletion and mutation analyses based on reporter gene assays. Compared with promoters that included the 59-flanking region up to bp 189 from the transcription 1 Supported by a Grant-in-Aid for Scientific Research (B) (2) (Grant No. 15380192) from the Japan Society for the Promotion of Science. 2 To whom correspondence should be addressed. E-mail: atsushi@agr. nagoya-u.ac.jp 3 Abbreviations used: FL, firefly-luciferase; HT, hydrodynamics-based gene transfection; L-PK, L-type pyruvate kinase; RL, renilla-luciferase.
start site, 59-deletions from bp 3200 to 190 of the L-PK gene promoter caused no progressive reduction of transcriptional activity (15). These results suggest the existence of cisacting DNA elements up to bp 189. Other investigators showed that 3 elements, located on bp 170 to 76 of the L-PK gene relative to the transcription start site, are involved in transcriptional activation stimulated by glucose, fructose, and insulin (11,13,14). These elements, termed PKL-I (94 to 76), PKL-II (149 to 125), and PKL-III (170 to 150), play an important role in transcriptional activation stimulated by dietary carbohydrate; cell type–specific expression of the L-PK gene is another important regulatory mechanism. Noguchi et al. (12) confirmed that in transgenic mice, the 59-flanking region of the L-PK gene from bp 189 to 137 contains regulatory elements responding to tissue-specific expression of L-PK, and that the region also contains the element responding to dietary glucose and fructose. However, the responsiveness to dietary glucose and fructose of the 59flanking region of the L-PK gene from bp 189 to 137 varied depending on the line of transgenic mice used, raising the possibility that other elements modulating L-PK gene expression may be present upstream of bp 189. On the basis of analyses of reporter gene assays in transgenic mice, Cuif et al. (17) suggested that regulatory elements may be present upstream of bp 183, but the positions of these elements were not identified.
0022-3166/06 $8.00 Ó 2006 American Society for Nutrition. Manuscript received 19 July 2005. Initial review completed 2 August 2005. Revision accepted 30 September 2005. 16
REGULATORY REGION OF THE L-TYPE PYRUVATE KINASE GENE
Transient nonviral gene transfer to tissues of live animals is now feasible, allowing functional cis-acting DNA elements to be investigated more directly in vivo than previously. Nonviral gene transfer methods require chemical manipulations (e.g., lipofection), and physical manipulations such as laser poration, electroporation, sonoporation, use of gene guns, and direct injection (18,19). When the liver is the primary target, direct injection by hydrodynamics-based gene transfection (HT) is more powerful and convenient than other nonviral gene transfer methods. Liu et al. (18) demonstrated that liver exhibits the strongest gene expression among tissues after HT, and that the intensity of expression is sufficient for use in reporter gene assays. Thus, HT may be a powerful approach for investigating functional cis-acting DNA elements on the 59-flanking region of the L-PK gene in liver, although to date, no practical applications of HT have been published. In the present study, HT was used to transfer a reporter gene fused to various lengths of the 59-flanking region of the rat L-PK gene into mice, and the existence of regulatory regions of the L-PK gene responding to stimulation by feeding was examined.
MATERIALS AND METHODS Plasmid DNA construction. pLcat3200 containing the 59-flanking region of the L-PK gene from bp 23200 to 137 (16), and pL-PK(2203)-FL containing the 59-flanking region of the L-PK gene from bp 2203 to 117 with a frame of pGL3basic [including the fireflyluciferase (FL) gene; Promega], were generously donated by Dr. T. Noguchi, Nagoya University, Nagoya, Japan. DNA fragments containing the 59-flanking region of the rat L-PK gene from bp 21467 to 117 were obtained by digestion of pLcat3200 using KpnI and HindIII. The DNA fragment was subcloned into the KpnI/HindIII site of pGL3basic; this plasmid DNA was designated pL-PK(21467)-FL. A series of 59-deletion mutants of the L-PK gene were constructed from pL-PK(21467)-FL using PCR-based mutagenesis (20). DNA fragments were amplified with the Expand High Fidelity PCR System (Roche) using the following primers: common lower, 59-GGT ACC GAA ATG TTC TGG CAC CTG CAC-39; pL-PK(21065)-FL upper, 59-GGT GGT CAA CAA GGC AAG GTC-39; pL-PK(2945)-FL upper, 59-ACG GCT GGT CAA CAA CAA TA-39; pL-PK(2860)FL upper, 59-TGC GAG AAA CTG AGA GAC CCT-39; pL-PK(2690)-FL upper, 59-AAT GAT GAG CCA ACA ATG TGA-39; pL-PK(2490)-FL upper, 59-CCA GTC TGC CGT TCT TG39; pL-PK(2300)-FL upper, 59-GAA TCA GCG TTG AGA GAT GGA-39. Self-ligation was achieved using a BKL kit (Takara Bio). Animals. Female ICR strain mice 4–5 wk old were maintained in a clean room at 23 6 18C with a 12-h light:dark cycle, and had free access to a commercial diet (MR Stock, Nosan) and water. The mice were cared for in accordance with the Guideline of Animal Experimentation as specified by the Committee of Experimental Animal Care, Nagoya University, Nagoya, Japan. Diets. The mice were acclimated to a ground form of the commercial diet for 2 d before the experiments. After HT, half of the mice had access to a high-carbohydrate diet overnight. This diet was composed as follows (g/kg diet): dextrin, 700 (Fujistar 5V, Nihon Syokuhin Kako); casein, 100; sucrose, 100; cellulose, 50; AIN-93 mineral mixture (21), 35; AIN-93 vitamin mixture (21), 10; L-cystine, 2.5; and choline bitartrate, 2.5. Optimization of the experimental design. To investigate changes in reporter gene expression with time, pRL-SV40 [including the renilla-luciferase (RL) gene; Promega] was transfected into mouse liver by HT as described below. The mice had access to the commercial diet until sample collection. On d 1, 2, 3, and 5 after HT, the mice were overanesthetized with diethyl ether, and RL activity was measured in the liver samples. To investigate differences in the response of endogenous L-PK mRNA in mice fed either the commercial diet or the highcarbohydrate diet, lactated Ringer’s solution was injected into the
17
tail vein (sham HT treatment), and the mice were divided into 3 groups. One group had free access to the ground form of the commercial diet, and another had free access to the high-carbohydrate diet. The third group was deprived of food overnight. Liver samples were collected from the 3 groups and L-PK mRNA expression was quantified. Quantification of L-PK mRNA expression. Expression of L-PK mRNA after HT was estimated using the RT-PCR assay. Briefly, total RNA was isolated from the liver samples and cDNAs were obtained using the RT reaction. PCR was performed with the cDNAs using mouse L-PK and b-actin–specific primers as follows: L-PK upper, 59GCA GAA TCC ATC GCC AAC-39; L-PK lower, 59-TCC TCG TGC CCA AGA TAC-39; b-actin upper, 59-CGG GAC CTG ACA GAC TAC CTC-39; b-actin lower, 59-GGG CAT CGG AAC CGC TCG TTG-39. Then, 33 thermal cycles were performed, with each cycle consisting of 948C for 1 min, 558C for 1 min, and 728C for 2 min. The amplifications were electrophoresed, and the intensities of the bands were measured. The abundance of L-PK mRNAs was normalized to the density of the b-actin bands. Identification of the regulatory regions of the L-PK gene. HT was performed as described by Liu et al. (18). Briefly, the mice were anesthetized using diethyl ether; 25 mg of pL-PK-FL (fused to a portion of the 59-flanking region of the L-PK gene) and 5 mg pRL-SV40 in lactated Ringer’s solution (0.1 mL/g body weight) were injected within 10 s into the tail vein using a 27-gauge needle. pRL-SV40 expresses RL under the control of the SV40 promoter, and promoter activity is not affected by diet. Thus, pRL-SV40 was introduced into the liver for the purpose of correcting variations in transfection efficiency. The pGL3 control vector (including the FL gene under the control of the SV40 promoter; Promega) was also injected with the pRL-SV40 as a negative control. After HT, the mice were allowed to recover from the procedure for 24 h; then they were divided into 2 groups according to their body weight. One group had free access to the high-carbohydrate diet, whereas the other group was deprived of food overnight; both groups of mice were then overanesthetized with diethyl ether. The left lobe of each mouse liver was removed and stored at 2808C until analysis. The regulatory region of the L-PK gene was surveyed in 3 individual experiments; Expt. 1 (bp 21467 to bp 2203), Expt. 2 (bp 21065 to bp 690), and Expt. 3 (bp 2690 to bp 2203). In Expt. 4, the response of pL-PK(2203)-FL to stimulation by feeding was reexamined, and the data of Expt. 4 were pooled with those of Expts. 1 and 3 to obtain a larger data set. Luciferase assay. Reporter gene assays were performed using the Dual-Luciferase Reporter Assay System kit (Promega). Briefly, the liver samples collected were homogenized and sonicated with 10 volumes (wt/v) of lysis buffer. After centrifugation (10,000 3 g for 15 min), the supernatants were collected, and a portion of each supernatant was mixed with the reaction buffer and appropriate substrate. FL and RL activities were measured using a luminometer (AutoLumat LB953; EG&G Berthold). Relative FL activity was calculated by dividing FL activity by RL activity. Statistical analysis. Data were analyzed by 1-way (optimization of experimental design and Expt. 4) or 2-way (Expts. 1, 2, and 3) ANOVA after common logarithmic transformation to stabilize error variance. The interaction between L-PK deletion and feeding state was significant in Expts. 1, 2, and 3; therefore, individual means were compared by Fisher’s Protected Least Significant Difference test. Statistical procedures were performed using a commercially available statistical package (Statview version 5.0, SAS Institute). A value of P , 0.05 was considered statistically significant.
RESULTS Optimization of the experimental design. Time course changes of reporter gene expression in liver were examined to confirm the validity of the transfected plasmids (Fig. 1). The RL activity was highest 1 d after HT; after 5 d, RL activity had decreased to background levels. Thus, reporter gene assays using this system should be completed by d 3 after HT.
18
SUZUKI ET AL.
FIGURE 1 Changes in RL gene expression in livers of mice after HT (optimization experiment). Values are means 6 SEM, n ¼ 4 or 5. Means without a common letter differ, P , 0.05.
To maximize the L-PK response to stimulation by feeding, endogenous L-PK mRNA levels in mice fed the commercial diet or the high-carbohydrate diet were compared. The commercial diet produced a 140% increase in L-PK mRNA levels compared with L-PK mRNA levels in the food-deprived group. However, induction of L-PK mRNA by the high carbohydrate diet was ;500% higher than induction by the commercial diet (data not shown). Thus, in the present study, the high-carbohydrate diet was used to stimulate the L-PK response after HT. Identification of the regulatory regions of the L-PK gene. In Expt. 1, regulatory regions were surveyed in the sequence from bp 21467 to 2203 (Fig. 2A). The FL activity of pLPK(21467)-FL was markedly stimulated by feeding mice the high carbohydrate diet. 59-Deletion up to bp 21065 caused only minor changes in FL activity, but further deletion up to bp 2690 significantly decreased FL activity. Furthermore, 59-deletion up to bp 2203 caused progressive reduction of FL activity. In contrast, FL activity was not generally affected in the food-deprived groups. The FL activity of the pGL3 control vector was also not influenced by feeding (data not shown), suggesting that the increases in FL activity of pL-PK-FL constructs were unique responses limited to the L-PK gene. These results suggest the existence of multiple positive regulatory regions responding to stimulation by feeding, from bp 21065 to 2690 and 2690 to 2203. Therefore, we analyzed the L-PK gene by dividing the sequence into two regions: 21065 to 2690, and 2690 to 2203. In Expt. 2, two additional 59-deletion mutants were constructed (up to bp 2945, and up to 2860), and the region from bp 21065 to 2690 was investigated. The 59-deletion up to bp 2945 caused progressive reduction of FL activity compared with mutants deleted up to bp 21065, but further deletion up to bp 2690 did not affect FL activity (Fig. 2B). These results indicate that the positive regulatory region responding to a carbohydrate diet exists on the L-PK gene from bp 21065 to 2945. In Expt. 3, two further 59-deletion mutants were constructed (up to bp 2490, and up to bp 2300), and the region from bp 2690 to 2203 was investigated. FL activity did not differ among 59-deletion mutants up to bp 2690 and 59-deletion mutants up to bp 2300 (Fig. 2C). However, 59-deletion up to
FIGURE 2 Expression of 59-deletion mutants of the L-PK/FL fusion gene in livers of fed and food-deprived mice. Values are means 6 SEM, n ¼ 7-12 (Expt. 1: A), 7-12 (Expt. 2: B), and 4 (Expt. 3: C). Means in a panel without a common letter differ, P , 0.05.
bp 2203 significantly decreased FL activity, suggesting the existence of a positive regulatory region from bp 2300 to 2203. In these experiments, the 59-flanking region up to bp 2203, which includes regions previously shown to be responsive to carbohydrate intake (11–14), did not respond in a reproducible fashion to stimulation by feeding. In Expt. 1, stimulation by feeding did not affect the promoter activity of pL-PK(2203)FL (Fig. 2A), whereas in Expt. 3, stimulation increased FL activity 190% (Fig. 2C). Therefore, Expt. 4 was performed to obtain a larger data set for evaluating the response of pL-PK(2203)-FL to stimulation by feeding. When data from the 3 experiments were pooled, FL activity in fed mice (176 6 30%, n 5 21) differed (P , 0.05) from that in food-deprived mice (100 6 15%, n 5 23). Thus, the present study partly substantiates earlier findings (11–14) that regulatory regions involving carbohydrate metabolism are located up to bp 2203, but the magnitude of the response observed is not consistent with previous findings.
DISCUSSION We report here the first survey of the regulatory regions of the L-PK gene by in vivo transient gene transfer. New positive regulatory regions responding to stimulation by feeding, residing from bp 1065 to 945 and from 300 to 203 of the L-PK gene, were identified by functional analysis using HT. Inclusion of these regions in the L-PK gene resulted in transcriptional activation in the liver of mice fed a high-carbohydrate diet.
REGULATORY REGION OF THE L-TYPE PYRUVATE KINASE GENE
These in vivo results may be more biologically important than assays using cultured cells. In cultured hepatocytes, the 59-flanking region up to bp 189 was sufficient to confer tissue-specific glucose-, fructose-, and insulin-regulated expression of the reporter gene (11,13,14). Studies on transgenic mice supported these results by showing that the 59-flanking region up to bp 189 conferred carbohydrate feeding–induced reporter gene expression and tissue-specific expression (12). However, in the 59-flanking region up to bp 189, responsiveness to carbohydrate intake varied depending on the line of transgenic mice used, suggesting the existence of other regulatory regions functioning in vivo (12). Cuif et al. (17) also demonstrated that the region between bp 392 and 189 had an additional effect on promoter activity in the livers of transgenic mice, and suggested the existence of cis-acting DNA elements in this region. The present results support this assumption, and indicate the existence of a positive regulatory region between bp 300 and 203 that responds to stimulation by feeding (Fig. 2C). Thus, it is likely that unidentified cis-acting DNA elements are present between bp 300 and 203. We also present evidence for a novel positive regulatory region between bp 1065 and 945 that responds to stimulation by feeding (Fig. 2B). Presently, it is unclear which nutritional or hormonal factors are involved in feeding-induced transcriptional activation in these regulatory regions. To predict potential cis-acting DNA elements and trans-acting proteins, the regions from bp 1065 to 945 and 300 to 203 were analyzed by searching a database of transcriptional factors (22). The region from bp 1065 to 945 includes potential sequences on which several trans-acting proteins, including the CCAAT displacement protein cut repeat, GATA-1, and nuclear factor-kap, are bound. The region
FIGURE 3 Potential regulatory regions of L-PK gene identified in the study. The potential regulatory regions are indicated by shaded rectangles. Open rectangles represent well-established cis-acting DNA elements responding to carbohydrate/insulin stimulation. The panel includes well-established trans-acting proteins and E box-like sequences in the PKL-III. Potential cis-acting DNA elements and trans-acting proteins were analyzed using a database for transcription factors (22). The position of each regulatory region is indicated by counts of bp relative to the transcription start site. CDP CR, CCAAT displacement protein cut repeat; ChoRE, carbohydrate response element; HNF, hepatocyte nuclear factor; NF, nuclear factor; USF, upstream stimulatory factor.
19
from bp 300 to 203 includes potential sequences on which GATA-1, -2 and -3 are bound. These sequences and proteins may participate in the regulation of L-PK gene expression via stimulation by feeding (Fig. 3). The use of transgenic mice is an attractive methodology for surveying regulatory regions of genes in vivo; to date, it has been the method of choice for examining functional regulatory regions of the L-PK gene. However, in the present study, transient nonviral gene transfer technology, HT, was employed. This approach has 3 advantages. First, the transfection procedure is simple, requiring only the injection of a DNA solution into the mouse tail vein. Second, HT can introduce exogenous plasmid DNA efficiently into the liver, and gene expression is adequate for the reporter gene assay used. Third, HT experiments are faster and more economical than experiments using transgenic mice. Using HT, 2 new positive regulatory regions of the L-PK gene were detected, supporting our assertion that HT is an effective methodology for surveying potential regulatory regions of genes in vivo. In conclusion, unidentified cis-acting DNA elements that respond to feeding stimulation are present in the 59-flanking region of the L-PK gene between bp 300 and 203, 1065 and 945. In addition, HT is useful for detecting cis-acting DNA elements in various genes expressed in the liver. ACKNOWLEDGMENTS We gratefully acknowledge Dr. Tamio Noguchi for the gifts of pLcat3200 and pL-PK(203)-FL.
LITERATURE CITED 1. Imamura K, Tanaka T. Multimolecular forms of pyruvate kinase from rat and other mammalian tissues. I. Electrophoretic studies. J Biochem (Tokyo). 1972;71:1043–51. 2. Tanaka T, Harano Y, Sue F, Morimura H. Crystallization, characterization and metabolic regulation of two types of pyruvate kinase isolated from rat tissues. J Biochem (Tokyo). 1967;62:71–91. 3. Nakashima K, Miwa S, Oda S, Tanaka T, Imamura K. Electrophoretic and kinetic studies of mutant erythrocyte pyruvate kinases. Blood. 1974;43: 537–48. 4. Inoue H, Noguchi T, Tanaka T. Rapid regulation of L-type pyruvate kinase mRNA by fructose in diabetic rat liver. J Biochem (Tokyo). 1984;96:1457–62. 5. Matsuda T, Noguchi T, Takenaka M, Yamada K, Tanaka T. Regulation of L-type pyruvate kinase gene expression by dietary fructose in normal and diabetic rats. J Biochem (Tokyo). 1990;107:655–60. 6. Munnich A, Lyonnet S, Chauvet D, Van Schaftingen E, Kahn A. Differential effects of glucose and fructose on liver L-type pyruvate kinase gene expression in vivo. J Biol Chem. 1987;262:17065–71. 7. Noguchi T, Inoue H, Tanaka T. Transcriptional and post-transcriptional regulation of L-type pyruvate kinase in diabetic rat liver by insulin and dietary fructose. J Biol Chem. 1985;260:14393–7. 8. Dynan WS, Tjian R. Control of eukaryotic messenger RNA synthesis by sequence-specific DNA-binding proteins. Nature. 1985;316:774–8. 9. Grosveld F, van Assendelft GB, Greaves DR, Kollias G. Positionindependent, high-level expression of the human beta-globin gene in transgenic mice. Cell. 1987;51:975–85. 10. Maniatis T, Goodbourn S, Fischer JA. Regulation of inducible and tissuespecific gene expression. Science. 1987;236:1237–45. 11. Liu Z, Thompson KS, Towle HC. Carbohydrate regulation of the rat L-type pyruvate kinase gene requires two nuclear factors: LF-A1 and a member of the c-myc family. J Biol Chem. 1993;268:12787–95. 12. Noguchi T, Okabe M, Wang Z, Yamada K, Imai E, Tanaka T. An enhancer unit of L-type pyruvate kinase gene is responsible for transcriptional stimulation by dietary fructose as well as glucose in transgenic mice. FEBS Lett. 1993;318: 269–72. 13. Shih HM, Towle HC. Definition of the carbohydrate response element of the rat S14 gene. Evidence for a common factor required for carbohydrate regulation of hepatic genes. J Biol Chem. 1992;267:13222–8. 14. Thompson KS, Towle HC. Localization of the carbohydrate response element of the rat L-type pyruvate kinase gene. J Biol Chem. 1991;266:8679–82. 15. Yamada K, Noguchi T, Matsuda T, Takenaka M, Monaci P, Nicosia A, Tanaka T. Identification and characterization of hepatocyte-specific regulatory regions of the rat pyruvate kinase L gene. The synergistic effect of multiple elements. J Biol Chem. 1990;265:19885–91.
20
SUZUKI ET AL.
16. Yamada K, Noguchi T, Miyazaki J, Matsuda T, Takenaka M, Yamamura K, Tanaka T. Tissue-specific expression of rat pyruvate kinase L/chloramphenicol acetyltransferase fusion gene in transgenic mice and its regulation by diet and insulin. Biochem Biophys Res Commun. 1990;171:243–9. 17. Cuif MH, Cognet M, Boquet D, Tremp G, Kahn A, Vaulont S. Elements responsible for hormonal control and tissue specificity of L-type pyruvate kinase gene expression in transgenic mice. Mol Cell Biol. 1992;12:4852–61. 18. Liu F, Song Y, Liu D. Hydrodynamics-based transfection in animals by systemic administration of plasmid DNA. Gene Ther. 1999;6:1258–66.
19. Muramatsu T, Nakamura A, Park HM. In vivo electroporation: a powerful and convenient means of nonviral gene transfer to tissues of living animals [review]. Int J Mol Med. 1998;1:55–62. 20. Zhao LJ, Zhang QX, Padmanabhan R. Polymerase chain reaction-based point mutagenesis protocol. Methods Enzymol. 1993;217:218–27. 21. Reeves PG, Nielsen FH, Fahey GC, Jr. AIN-93 purified diets for laboratory rodents: final report of the American Institute of Nutrition ad hoc writing committee on the reformulation of the AIN-76A rodent diet. J Nutr. 1993;123:1939–51. 22. Akiyama YTFSEARCH: searching transcription factor binding sites http:// mbs.cbrc.jp/research/db/TFSEARCHJ.html 2002 [accessed June 1, 2005]
