THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1991 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 266, No. 13, Issue of May 5, pp. 8086-8091,1991 Printed in U. S. A.

Isolation and Characterization of the Transcriptionally Regulated Mouse Liver (B-type) Phosphofructokinase Gene and Its Promoter*

(Received for publication, April 3, 1990)

Pornpimol Rongnoparut, Carl P. Verdon, Stephen C. Gehnrich, and Hei Sook SulS From the Department of Nutrition, Harvard School of Public Health, Boston, Massachusetts 021 15

We have isolated and characterized a mouse gene encoding liver (B-type) phosphofructokinase, a key regulatory enzyme in glycolysis. The gene spans ap- proximately 21.5 kilobase pairs and consists of 22 exons. Compared with the muscle (A-type) phospho- fructokinase gene, the sizes of the introns are different although exon lengths are highly conserved. Two tran- scription start sites 10 bases apart were determined by primer extension experiments. The immediate 5‘ se- quence does not possess a TATA or CCAAT box but contains multiple GC boxes (positions -10, -43, -50, -62, and +28 in the 5’-untranslated region) which may be Spl-binding sites. An unusual feature of 200 base stretches of CT repeats is present at position -480 to -693. In addition, direct repeats of CTCGAAGGAG are found at positions -447 and -478. DNase I foot- printing showed five regions where liver nuclear pro- teins may interact. Two proximal 5”flanking regions spanning -1 to -20 and -30 to -70, which contain GC boxes. Also protected was a region spanning -70 to -90, which contains an AP-1 like sequence (TCAGTCA). The consensus AP-1 sequence, however, did not inhibit footprinting, indicating involvement of a distinct protein. Two distal regions spanning from -450 to -470 and from -500 to -520 were also pro- tected. The former is positioned between the direct repeats and the latter is at the start of the CT repeats. The rate of transcription of the liver phosphofructo- kinase gene, as measured by run-on assays, increased &fold in livers of previously starved mice fed a high carbohydrate diet compared to starved controls. Ad- ministration of dibutyryl cAMP blocked the increase in transcription caused by refeeding. Functional analy- sis of the promoter region of the gene will be necessary to elucidate the mechanisms of transcriptional regula- tion by fastinglrefeeding and by CAMP. These results provide a useful system for the study of regulatory elements in liver phosphofructokinase gene transcrip- tion.

Phosphofructokinase (EC 2.7.1.11) catalyzes the phos- phorylation of fructose 6-phosphate to fructose 1,6-bisphos- phate and is a key regulatory enzyme in the glycolytic path- way. Phosphofructokinase activity is under allosteric regula- tion by various metabolites (1, 2). The smallest active form

*This work was supported in part by a National Institutes of Health Grant DK36264. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in thispaper has been submitted to the GenBankTM/EMBL Data Bank with accession number($ M61210.

$ To whom correspondence should be addressed.

of the enzyme is a tetramer with a subunit M, of 80,000. Three types of subunits have been identified, i.e. muscle (A- type), liver (B-type), and brain (C-type) (3, 4). The genes encoding the three types of subunits are not known to be linked. Variable expression of these loci in different tissues leads to tissue-specific patterns of isozymes with differing physicochemical and regulatory properties (5-7). The phos- phofructokinase isozymes isolated from muscle, liver, and brain show different sensitivity to ATP and citrate inhibition (8). In addition, the muscle isozyme has been reported to be better suited for binding to F-actin (8).

We have previously isolated cDNA sequences to both liver and muscle type phosphofructokinase (10, 11). Using these cDNA clones as probes in Northern blot analysis, we have observed that the muscle type phosphofructokinase expres- sion was largely restricted to muscle tissue whereas liver-type phosphofructokinase was expressed not only in liver but also, although at lower levels, in tissues, such as brain, kidney, and lung (10, 11). Comparison of the deduced amino acid se- quences of mouse liver and muscle phosphofructokinase showed 67% homology, while mouse muscle phosphofructo- kinase showed 90% homology compared with the known rabbit muscle isozyme sequence, indicating higher conserva- tion of amino acid sequence of isozymes between mammalian species than among different isozymes in the same species (10-12). We have also reported that the liver phosphofructo- kinase mRNA level was under hormonal and nutritional control (10). The liver phosphofructokinase mRNA level in- creased 4-fold when previously starved mice were refed a high carbohydrate, fat-free diet. Further, this increase in mRNA level was blocked by 50% by the administration of dibutyryl cAMP (11).

In order to study the molecular mechanisms underlying the regulation of liver phosphofructokinase gene expression by nutrients and hormones, we have isolated genomic clones coding for liver phosphofructokinase and analyzed the gene structure by defining the intron-exon positions and junction sequences. We have obtained the promoter and adjacent DNA sequences which contain a number of elements that may play roles in liver phosphofructokinase gene regulation. Several regions interact with liver nuclear proteins as shown by DNase I footprinting. We also show the transcriptional reg- ulation of the mouse liver phosphofructokinase gene by fast- ing/refeeding and by CAMP. It will now be possible to define cis-acting elements responsible for the regulated expression of the liver phosphofructokinase gene under various physio- logical conditions.

EXPERIMENTAL PROCEDURES

Isolation and Characterization of Liver Phosphofructokinase Ge-

partial EcoRI and HaeIIIIAluI digests of BALB/C sperm DNA (ob- nomic Clones-A Charon 4A mouse genomic library constructed from

tained from Dr. L. Hood, California Institute of Technology) and a BALB/C embryo library constructed from partial HaeIII digests

8086

Liver Phosphofructokinase Gene and Its Promoter 8087

(obtained from Dr. P. Leder, Harvard Medical School) were screened at a density of 30,000 plaques/l50-mm plate with a 2.7-kbp' full- length cDNA sequence or restriction fragments of the genomic or cDNA clones labeled by the random primer method (13). Four inde- pendent clones were isolated and characterized further. Restriction fragments of phage DNAs of these genomic clones were subcloned into pBluescript SK+ (Stratagene) or M13 mp18 vectors. Nucleotide sequence analysis was carried out on alkali-denatured double- stranded or single-stranded templates by the chain termination method using Sequenase (USB Biochemicals) and universal primers or oligonucleotides synthesized by standard phosphoamidite tech- niques on a BioSearch synthesizer according to the previously deter- mined nucleotide sequence. Polymerase chain amplification was uti- lized to determine the positions and sizes of the exons and introns in the gene. For intron/exon boundaries, oligonucleotides corresponding to exon sequences according to our cDNA sequences and known rabbit muscle phosphofructokinase gene structure were synthesized and used for amplification by polymerase chain reaction utilizing the genomic and subcloned fragments as templates to verify the intron sizes. Polymerase chain reaction was carried out across not only one but in most cases additional intron sequences. Amplification products were analyzed by agarose gel electrophoresis.

Primer Extension Analysis-A 26-residue oligonucleotide corre- sponding to nucleotides +62 to +87 of the noncoding strand of the mouse liver phosphofructokinase cDNA (11) was synthesized and was '"P-labeled with T4 polynucleotide kinase. The oligonucleotide was hybridized with 20 pg of poly(A)' RNA prepared from mouse liver in 80% formamide, 40 mM PIPES, pH 6.4, 400 mM NaC1, and 1 mM EDTA at 30 "C for 16 h. The hybrid was extended with avian myeloblastosis virus reverse transcriptase in the presence of 0.5 mM each of unlabeled nucleotide triphosphates for 90 min at 42 "C. Following phenol/chloroform extraction and ethanol precipitation, the products were analyzed by electrophoresis on 6% denaturing polyacrylamide gel and exposed to Kodak XAR-5 film at -70 "C with an intensifying screen.

Isolation of Liver Nuclei and Transcription Run-on Analysis- Livers from three mice which were fasted for 48 h, or from fasted mice refed a high carbohydrate, fat-free diet were used to isolate nuclei. Dibutyryl CAMP (6 mg/lOO g of body weight) and theophylline (3 mg/100 g) were given intraperitoneally at the start of refeeding. Nuclei were prepared as previously described by centrifugation through a 2.1 M sucrose cushion at 20,000 rpm for 60 min in a Beckman SW 28 rotor (14). Run-on transcription was carried out, as described previously, at 25 "C for 45 min in a reaction mixture containing 5 X lo7 nuclei, 125 pCi of [w3'P]UTP (3000 Ci/mmol), and three unlabeled triphosphate nucleotides (14). Labeled RNA was extracted with phenol and sodium dodecyl sulfate at 65 "C, ethanol- precipitated, and subjected to a Sephadex G-50 spun column. Five pg of plasmid pPFK-2 was denatured by treating with 0.3 M NaOH, fixed on nitrocellulose filters, and hybridized for 48 h at 65 "C with lo7 cpm in 10 mM Tris, pH 7.4, 0.2% sodium dodecyl sulfate, 10 mM EDTA, 0.3 M NaCl, 1 X Denhardt's, and 250 pg/ml yeast tRNA. The filters were washed twice in 0.1% sodium dodecyl sulfate, 0.1 X SSC at 60 "C. An autoradiogram was obtained by exposing the washed filters to x-ray film with an intensifying screen and was subjected to a quantitative densitometric scanning.

Preparation of Nuclear Extracts-Nuclear extracts were prepared from mouse liver by the methods of Hennighausen and Lubon (15). Protein concentrations were measured by the Bradford method using bovine serum albumin as the standard (16).

DNase I Footprinting Analysis-The DNA probes used in the footprinting assays were prepared by end-labeling specific DNA frag- ments prepared by polymerase chain reaction using synthetic oligo- nucleotides with added restriction sites and a PstI fragment (as marked in Fig. 1) subcloned into Bluescript SK+. DNA was end labeled using T4 polynucleotide kinase and [y3*P]ATP, digested with a restriction enzyme. The fragments separated by low melting agarose gel electrophoresis were purified on an Elutip-d column (Schleicher & Schuell). Binding assays were performed in the total volume of 50 pl containing 50,000 cpm (2 ng) of the probe, 1 pg of poly d(I.C):poly d(I.C), 50 mM NaC1, 100 mM EDTA, 20 mM Hepes, pH 7.9, 0.5 mM dithiothreitol, 10% glycerol, and varying concentra- tion of nuclear extracts. For the competition both sense and antisense strands were synthesized according to the sequence, heated at 95 "C

' The abbreviations used are: kbp, kilobase pair; PIPES, 1,4-piper- azinediethanesulfonic acid; Hepes, 4-(2-hydroxyethyl)-l-piperazine- ethanesulfonic acid PFK, phosphofructokinase.

for 5 min, and slowly cooled to room temperature for 2 h in 10 mM Tris, pH 7.9, 1 mM EDTA, and 100 mM NaC1. The resulting double- stranded oligonucleotides were added for the preincubation. After 20 min at room temperature, 3 pl of 100 mM MgC1, and 20 mM CaCl, solution was added followed immediately by the addition of 2 p1 of 0.025 units of DNase I solution. Digestion was carried out for 30 s at room temperature followed by addition of 15 pl of 100 mM EDTA, pH 8.0. Samples were extracted with phenol/chloroform, ethanol- precipitated, resuspended in 90% formamide containing xylene cyano1 and bromphenol blue, and separated on 8 M urea, 6% polyacrylamide sequencing gels. Gels were dried and autoradiographed.

RESULTS AND DISCUSSION

Isolation and Ch.aracterization of the Mouse Liver Phospho- fructokinase Gene--We initially screened Charon 4A mouse genomic libraries with a previously cloned 2.7-kilobase full- length PFK cDNA sequence (10). Clone XLPFK-1 contained an 18.5-kbp insert. of mouse DNA encompassing most of the liver phosphofructokinase gene. An overlapping clone (XLPFK-2) with an insert size of 10.8 kbp contained 130 bp of the 5"flanking region of the liver phosphofructokinase gene. We further screened the genomic libraries with the most 5' 1.4-kbp fragment of XLPFK-2, and isolated a genomic clone with an insert size of 16.2 kilobase (XLPFK-3) which contained the first, two introns and approximately 10 kilobase of 5'-flanking region. Since the 3' most cDNA sequences were not contained in these genomic clones, the genomic libraries were also screened with the 730-bp PstI fragment from the 3' most region of the cDNA clone, pPFK-2. Positive clone ALPFK-4 (14.6 kbp) contained the three most 3' exons and the 3'-flanking region of the liver phosphofructokinase gene.

To define the positions and boundaries of the exon blocks, the restriction fragments of the four clones were characterized by Southern blot. analysis (Fig. 1). The EcoRI and XhoI fragments were subcloned into the plasmid Bluescript SK+. All of the intronlexon junctions were sequenced using oligo- nucleotides synthesized based on the cDNA sequences. The sequences of the splice junctions, as well as the lengths of the introns and exons are shown in Table I. All of the splice donor and accept.or sites conform to the known consensus sequences, except. the donor site of exon number 21 with somewhat diverged sequence (17). Twelve of the 21 interven- ing sequences which are located in coding sequences occur between codons, while nine interrupt codons. The liver phos- phofructokinase gene spans 21.5 kbp and contains 22 exons. When compared with the previously reported rabbit muscle phosphofructokinase gene, the exon-intron organization of the mouse liver and rabbit muscle phosphofructokinase genes is similar with respect to the exon sizes and positions with the exception of exon number 13, which is three nucleotides (1 amino acid residue) shorter in the mouse liver gene (11, 12). In addition, we have previously observed an overall 67% homology between mouse liver and rabbit muscle phospho- fructokinase mRNA (11). However, the intron sizes of the two genes encoding the muscle (12) and liver isozymes of phosphofructokinase, as shown in Table I, are not conserved. Nonetheless, their overall structural similarity in coding se- quences and the exon-intron junctions described above sup- ports our conclusion that liver and muscle phosphofructoki- nase evolved from a common ancestral gene (11).

Analysis of 5'-E'lanking Region of the Mouse Liver Phospho- fructokinase Gene-In order to determine the transcription start site, an oligonucleotide complementary to a sequence starting six base pairs upstream from the initiator methionine was synthesized. The oligonucleotide was labeled with 32P, hybridized to mouse liver RNA, and extended by reverse transcriptase (Fig. 2). Poly(A)+ RNA from mouse liver di- rected the synthesis of two extension products of 87 and 77

8088 Liver Phosphofructokinase Gene and Its Promoter

A 6 8 1 0 1 2 14 16 18 20 22 24 26 28 30 3? 34 3bbb

FIG. 1. Structure of the mouse liver phosphofructokinase gene. A, the four overlapping genomic clones and the restriction sites which were used to characterize the gene. E, EcoRI; X , XhoI , P, PstI; H, HindIII. The hatched box indicates the location of the liver phos- phofructokinase structural gene. B, the organization of the liver phosphofructo- kinase gene and the comparison with the rabbit muscle phosphofructokinase gene. The solid boxes show the locations of the exons.

6

0 9 10 I1 I 2 I1 ‘ 4 15 ‘ 6 7 1 ’ 9 .:I .: E L

TABLE I Exon-intron organization of the mouse liver phosphofructokinase gene

The nucleotide sequences of the exon-intron junctions were determined as described under “Experimental Procedures.” The sizes of the all introns except intron 2 were verified by polymerase chain reaction.

Exon Size 5‘ donor Intron Size 3’ acceutor

1 2 3 4 5 6 7 8 9

10 11 12 13 14 15 16 17 18 19 20 21 22

base pairs 152 74 78

190 166 45

109 96 93

126 65 64

147 71 88

153 165 62

112 100 106 145

CGCAAGgtgatgggac TATGAGgtaagtgtac CAGCTGgtgaggctcg AGGAAGgtgggtccgt CCAAAGgtaagtgagg CTGCGGgtgaggaggt GGCGAGgtgagtggcc AAGGATgtgagtatgg ATCCTGgtatgtttat CAAGTGgacaagtcct TGGCAGgtgaggaggg GAGAAGgtaggtgtcg GGTCAGgtcagtagct CAAGAGgtgaggcttt TTTGAGgtgaggtatc ATGGAGgtgagcccag TTAAAGgtgaggctcc GCTCCGgtacagctct CAGCAGgtaggtgggt GTAAAGagtaagtcct TTTTGAgcttactgtc

1 2 3 4 5 6 7 7 9

10 11 12 13 14 15 16 17 18 19 20 21

kbP 2.30 4.70 0.53 0.47 0.50 0.37 2.24 1.11 0.97 1.11 0.32 0.78 0.30 0.46 1.06 0.30 0.21 0.17 0.91 0.13 0.21

tccctcacagGTATGA tcgcttgcagGGCTAC ttccatgcagGGTGGC tccgacccagGCAAGA ttctccacagTCACCA gctcttccagGTACCT tctctgatagACTCGG tgacaaatagCTGGTC tcctgaacagAGTAGC tctctttaagACAAAG tcctttccagGAGCTT ctttctgaagTCCAAC tcttcctgagGTGCAA ctaaccacagGACACT tgcactgcagGCCTAC atctctccagAGTTGT tccttcccagGCCAAT gtctccacagGAACGA ctgtttccagGGTGGT ttccataacaGGCGGG acctgcgcagGCACCG

nucleotides. Both of the transcription initiation sites, which are 66 and 56 nucleotides upstream from the AUG translation initiator codon, were utilized with equal frequency as shown by the similar intensity of the two extension products shown in Fig. 2.

We have sequenced and analyzed 700 base pair upstream from the transcription initiation site (Fig. 3). A computer- assisted search of the 5’-flanking region of the mouse liver phosphofructokinase gene reveals elements that have poten- tial roles in its transcription. The canonical polymerase 11 transcription elements TATA and CCAAT are lacking in the liver phosphofructokinase gene promoter. There are clusters of GC-rich domains with the consensus sequence for the Spl- binding site at positions -10, -43, -50, and -62 as well as at +28 in the 5”untranslated region of the liver phosphofruc- tokinase gene (Fig. 3). These characteristics, i.e. absence of TATA and CCAAT and presence of multiple GC boxes, have been reported for the promoters of several genes which encode enzymes involved in metabolic reactions carried out in many

cell types, and are thus named housekeeping genes (18-21). We have previously shown that whereas the muscle phospho- fructokinase mRNA was present only in muscle tissue, mRNA for the liver isozyme was detected in all tissues examined, including liver, brain, lung, and kidney (11). Universal expres- sion of liver phosphofructokinase suggests a housekeeping role, while muscle and brain isozymes may be expressed in a tissue-specific manner according to glycolytic needs. However, unlike other housekeeping genes, the liver phosphofructoki- nase mRNA is inducible in liver by refeeding of a high carbohydrate diet to previously fasted mice. This induction is blocked by 50% by the administration of dibutyryl CAMP at the start of refeeding (11). An unusual feature of the promoter region of the phosphofructokinase gene is the extended stretches at positions -480 to -693 of pyrimidine-rich (dC- dT) tracts, which have been associated with H form DNA (22, 23). Palindromes with a motif CCTC(N),CTCC are found in polyoma and SV40; in SV40 this sequence serves as a large T antigen-binding site. It has been reported that the alternating

Liver Phosphofructokinase Gene and Its Promoter 8089

FIG. 2. Primer extension analysis of the liver phosphofruc- tokinase mRNA. 20 pg of poly(A)+ RNA from liver was hybridized to the oligonucleotide and extended by reverse transcriptase as de- scribed under "Experimental Procedures." The last four lanes (G, A, T, and C) represent sequence of the genomic fragment by dideoxy chain termination using the same primer as for the RNA analysis. Lane 1 shows the extension products with arrows indicating them. The start sites are also underlined in the adjacent nucleotide se- quence. Essentially identical results were obtained from three sepa- rate experiments.

-689

-850

-800

-660

-600

-460

-400

-360

-300

-260

-200

-160

-100

-60

*l

+ 6 1

+lo1

+161

+201

+261

* s o 1

FIG. 3. The DNA sequence of the 5"flanking region of the liver phosphofructokinase gene. The sequences of the first exon are represented by capital letters. The sequence is numbered in relation to the first transcription initiation site. There are two tran- scription start sites a t +1 and +11. The translation initiator is ouerlined. The GC boxes are boxed. The AP-1 like sequences are underlined with heavy lines. The two light-dotted lines show the direct repeat sequences. The numbered heavy dotted lines with arrows in- dicate the oligonucleotides used in primer extension and for the probe synthesis in DNase I footprinting. The underlined sequences with letters indicate the regions protected by liver nuclear extracts from DNase I as shown in Fig. 4.

C and T residues present in the promoters of heat shock genes and histone genes bind to nuclear proteins and adopt triple helical structures, and thus may be responsible for maintain- ing a potentially inducible promoter structure (24). Also pres- ent in the liver phosphofructokinase promoter is a direct repeat sequence, CTCAAGGAG, a t position -447 and at position -478 with a one base (G) insertion, which has no homology with any known regulatory element (Fig. 3).

To examine the potential regulatory sequences present in

the liver phosphofructokinase gene promoter, we probed liver nuclear extracts for the presence of proteins capable of bind- ing to the regulatory elements defined above using three fragments spanning nucleotides +87 to -169, -101 to -400, and -351 to -699. Fig. 4 shows the results of the experiments in which an increasing amounts of liver nuclear extract were used for DNase I footprinting of the liver phosphofructokinase gene promoter. Liver nuclear proteins were found to interact with several regions of the liver phosphofructokinase gene promoter in vitro. When the DNA fragment spanning +87 to -169 was labeled on the antisense strand and used for DNase I footprinting (Fig. 4 4 left panel), two regions which contain GC boxes in the proximal 5"flanking sequence of the gene, corresponding to nucleotides +5 to -16 (box A ) and -34 to -59 (box B ) , were protected from DNase I digestion. When the same fragment was labeled on the sense strand (Fig. 4a, right panel A ) , the footprints were shown in the regions extending from -4 to -16 (box A ) and -37 to -67 (box B ) . These regions are most likely the binding sites for Spl. Immediately upstream of the Spl footprints was a DNA region protected by the nuclear extract extending from -72 to -83 (left panel, box C) on the antisense strand and from -69 to -83 (right panel, box C) on the sense strand. The footprint region contained an AP-1-like sequence (TCAGTCA), which was previously identified in the 12-0-tetradecanoylphorbol- 13-acetate-inducible interleukin 2 gene promoter (25). The AP-1-binding sequence, in addition to mediating the respon- siveness to phorbol esters, has been shown to be involved in mediating cAMP response in several genes (26). Since liver PFK mRNA is negatively regulated by cAMP as shown by inhibition of the induction caused by fastinglrefeeding, we tested if AP-1 binds to liver PFK promoter. The addition of a 60-fold excess (120 ng) of the oligonucleotide (-65 to -97) in the preincubation step of DNase I digestion reactions, the footprint extending from -69 to -83 disappeared (Fig. 4a, right panel B) . However, the oligonucleotide synthesized ac- cording to the AP-1 consensus sequence (27) did not compete as shown in Fig. 4a, right panel C, indicating involvement of distinct protein in the liver PFK promoter. Surprisingly, occupation of this domain helps with the binding of nuclear protein, presumably Spl to the proximal GC box (box A), since the latter was only protected when the AP-1-like se- quence was occupied. Binding of nuclear proteins to their respective Spl and Ap-1 like sequences may be cooperative. Further studies are necessary to understand the mechanism.

When the DNA fragment spanning -101 to -400 was used in the footprinting assay, no appreciable protection was de- tected in this middle 5'-flanking sequence of the gene. We could not detect any protection around another copy of an AP-1-like sequence present in this region at -273. In the distal 5'-flanking sequence of the gene, two regions of nu- clease protection were shown a t nucleotides -453 to -469 and -502 to -519 (Fig. 4b, panel A, boxes D and E ) , when DNA fragment spanning -351 to -699 was labeled on noncoding strand and used for footprinting. The competition experi- ments carried out with the 60-fold excess of oligonucleotides synthesized according to the sequences from -445 to -475 and from -485 to -522 eliminated the footprints to the respective regions completely (Fig. 4b, panels B and C). The former region is positioned between the direct repeat se- quences described above. The latter region is the sequence at the start of the CT repeat sequences. The CT repeat sequences themselves did not show protection with the addition of nuclear extracts. In addition, binding of liver nuclear proteins produced distinct hypersensitive sites adjacent to the pro- tected regions. In all cases, increasing amounts of liver nuclear extract caused higher degrees of protection. When we initially

8090

A 0 10 20 40 -

A

I - L

-30

Liver Phosphofructokinase Gene and Its Promoter

(b)

A 0 C 0 15 30 45 0 - -

4

" - 1

-50

-70

-90

A B C 0 15 30 45 0 0 15 30 45 0 0 15 30 4 5 0

,530

- .-

-510

-490

470

__._ 'E _" " -. . -= -450

.. - FIG. 4. Binding of nuclear proteins to the mouse liver phosphofructokinase promoter region. DNA

fragments containing the mouse liver phosphofructokinase promoter regions, +87 to -169 for section ( a ) and -351 to -699 for section ( b ) were utilized for footprinting. The binding assays were carried out with probes made by labeling noncoding or coding strands by T 4 polynucleotide kinase, restriction digestion, low melting gel electro- phoresis, and elution. Fifty thousand cpm of each probe were incubated with indicated amounts of liver nuclear extracts. DNase I footprinting reactions were performed as described under "Experimental Procedures" and analyzed on a 6% polyacrylamide/urea gel. The regions of protection from DNase I digestion are demarcated by boxes. Section a: left panel A; DNA fragment spanning +87 to -169 was labeled on noncoding strand; right panel A; DNA fragment spanning +87 to -169 was labeled on coding strand was employed for footprinting and indicated amounts (pg) of nuclear extract with no competing oligonucleotides were added, B; 60-fold excess of oligonucleotides (-65 to -97) added during preincubation, C; 60-fold excess of AP-I consensus sequence (CGTGACTCAGCGCGC) added during preincubation. Section b: DNA fragment spanning -361 to -699 was labeled on noncoding strand and was employed for footprinting. A, no competing oligonucleotides were added B, oligonucleotides (-445 to 475) added during preincubation; C, oligonucleotides (-485 to 522) added during preincubation. The regions of protection (A-E) are demarcated by the boxes on the side and are underlined with the designating letters in Fig. 3.

employed labeled DNA fragment spanning nucleotides +87 to -698 in the footprinting, similar results were obtained. At the present time, it is not known what functional roles the nuclease protected regions may play. However, these se- quences may be important in isozyme-specific expression of the phosphofructokinase gene. We have cloned and sequenced the promoter region of the mouse muscle phosphofructokinase gene (data to be reported elsewhere). When compared with the promoter region of the mouse muscle phosphofructokinase gene, there is no sequence homology in the 5"flanking regions of these two genes encoding liver and muscle isozymes.

Transcriptional Regulation of the Liver Phosphofructokinase Gene-We have carried out transcription run-on assays using nuclei prepared from the livers of previously fasted mice which were refed a high carbohydrate diet. As shown in Fig. 5, the transcription of the liver phosphofructokinase gene was barely detectable in fasted mouse liver. However, after 6 h of refeed- ing, the transcription rate increased approximately &fold. The increase in transcription rate in the livers of refed mice was completely abolished by the administration of CAMP. The actin gene transcription did not change appreciably dur- ing fasting, refeeding, and refeeding with CAMP treatment

Liver Phosphofructokina

DGEM

DLPFK

1 2 3

FIG. 5. Effect of fastinglrefeeding and cAMP on liver phos- phofructokinase gene transcription. Mice were fasted for 48 h and refed a high carbohydrate, fat-free diet for 6 h. Dibutyryl cAMP was administered at the start of refeeding. The nuclei were isolated from pooled livers of three mice for each group and used for run-on transcription analysis as described under “Experimental Procedures.” Lane 1, fasted; lane 2, refed; lane 3, refed/dibutyryl CAMP. Five pg of either the pGEM vector or pPFK-2 was used. A representative autoradiogram of two independent experiments is shown.

(data not shown). We have previously reported that the liver phosphofructokinase mRNA level is low in fasted mouse liver and increased 4-fold when the previously fasted mice were refed with a high carbohydrate diet for 12 h. In addition, administration of cAMP at the time of refeeding inhibited the induction by approximately 50%. The present results show for the first time transcriptional regulation of the liver phos- phofructokinase gene. Moreover, the increase in liver phos- phofructokinase gene transcription brought about by fasting/ refeeding is approximately the same degree as noted for the change in its mRNA level. The effect of cAMP on liver phosphofructokinase gene transcription was more marked than that on steady-state mRNA level, possibly due to its half-life. The liver phosphofructokinase gene promoter con- tains two copies of an AP-1 like sequence (TCAGTCA), which may be involved in mediating cAMP response in several genes (26). However, as shown in Fig. 4, the AP-1 consensus se- quence did not compete the protected proximal AP-1-like sequence in the footprinting experiment shown in Fig. 4, indicating involvement of a distinct protein. The electropho- resis mobility shift assay showed two complexes and the complex formation was not inhibited by AP-1 consensus sequence.* Unlike the positive regulation by cAMP via the cAMP response element TGACGTCA, the negative regula- tion of gene transcription by cAMP has not been examined. Certainly, functional analysis of the 5‘-flanking region as well as a detailed protein-DNA interaction analysis in uitro will be necessary to understand the mechanisms underlying the positive transcriptional regulation by fastinglrefeeding and the negative regulation by CAMP.

P. Rongnoparut and H. S. Sul, unpublished results.

ae Gene and Its Promoter 8091 REFERENCES

1. Kemp. R. G., and Foe, L. G . (1983) Mol. Cell. Biochem. 5 7 , 147- 154

2. Reinhart, G. D., and Lardy, H. A. (1980) Biochemistry 1 9 , 1477- 1484

3. Tsai, M. Y., and Kemp. R. G. (1972) J. Biol. Chem. 2 4 8 , 785- 792

4. Thrasher, J. R., Cooper, M. D., and Dunaway, G . A. (1981) J. Biol. Chem. 2 5 6 , 7844-7848

5. Vora, S. (1981) Blood 57, 724-731 6. Vora, S., and Francke, U. (1981) Proc. Natl. Acad. Sci. U. S. A.

7. Weil, D., Cottreau, D., Van Cong, N., Rebourcet, R., Foubert, C., Gross, M. S., Dreyfuss, J. C., and Kahn, A. (1980) Annu. Hum. Genet. 4 4 , 11-16

8. Foe, L. G., and Kemp. R. G. (1985) J. Biol. Chem. 260,726-730 9. Choate, G. L., Lan, L., and Mansour T. E. (1985) J. Biol. Chem.

10. Gehnrich, S. C., Gekakis, N., and SUI, H. S. (1988) J. Biol. Chem.

11. Gekakis, N., Gehnrich, S. C., and SUI, H. S. (1989) J. Biol. Chem.

12. Lee, C.-P., Kao, M.-C., French, B. A., Putney, S. D., and Chang,

13. Feinberg, A. P., and Vogelstein, B. (1983) Anal. Biochem. 132 ,

14. Paulauskis, J. D., and SUI, H. S. (1989) J . Biol. Chem. 264,574-

15. Hennighausen, L., and Lubon, L. (1987) Methods Enzymol. 152 ,

16. Bradford, M. (1976) Anal. Biochem. 72, 248-254 17. Breathnach, R., and Chambon, P. (1981) Annu. Rev. Biochem.

18. Valerio, D., Duyvesteyn, M. G. C., Dekker, B. M. M., Weeda, G., Berkvens, T. M., Van der Voorn, L., Van Ormondt, H., and Van derE, A. J. (1985) EMBO J. 4,437-443

19. Melton, D. W., Konecki, D. S., Brennand, J., and Caskey, C. T. (1984) Proc. Natl. Acad. Sci. U. S. A. 8 1 , 2147-2151

20. McGrogan, M. M., Simonsen, C. C., Smouse, D. T., Farnham, P. J., and Schimke, R. T. (1985) J. Biol. Chem. 260,2307-2314

21. Singel-Sam, H., Keith, D. H., Tani, K., Simmer, R. L., Shively, L.. Lindsav. S., Yoshida, A.. and Riggs, A. D. (1984) Gene

78,3738-3742

260,4815-4822

263,11755-11759

264,3658-3661

S. H. (1987) J. Biol. Chem. 262,4195-4199

6-13

577

721-735

50,349-383

(A.mst.) 32,.409-417 . . _ _

22. Mirkin. S. M.. Lvamichev. V. I.. Drushlvak. K. N.. Dobrvnin, V. N., Filippov, S . A., an’d Frank-Kamenetskii, M . D: (1987) Nature 330,495-497

23. Rilev. D. E.. Reeves. R., and Gartler, S. M. (1986) Nucleic Acids R& 14,9407-9423

. .

24. Gilmour. D. S.. Thomas. G. H.. and Elzin. S. C. R. (1989) Science 245. im-1’490

- .

25. Fujita, T., Takaoka, C., Matsui, H., and Taniguchi, T. (1983) Proc. Natl. Acad. Sci. U. S. A. 8 0 , 7437-7441

26. Angel, P., Imagawa, M., Chiu, R., Stein, B., Imbra, R. J., Rahms- dorf, H. J., Jonat, C., Herrlich, P., and Karin, M. (1987) Cell

27. Rauscher, F. J., 111, Sambucetti, L. C., Curran, T., Distel, R. J., 49,729-739

and Spiegelman, B. M. (1988) Cell 52,471-480