Proc. Natl. Acad. Sci. USAVol. 86, pp. 4838-4842, July 1989Biochemistry

Rat glucokinase gene: Structure and regulation by insulinMARK A. MAGNUSON, TERESA L. ANDREONE*, RICHARD L. PRINTZ, STEVE KOCH, AND DARYL K. GRANNERtDepartment of Molecular Physiology and Biophysics, Vanderbilt University Medical School, Nashville, TN 37232

Communicated by Charles R. Park, March 20, 1989

ABSTRACT The glucokinase gene is 15.5-kilobases long,appears to be present as a single copy, and contains 10 exonsthat range in size from 96 to 977 base pairs. The transcriptionstart site was located 127 nucleotides upstream from thetranslation initiation codon. The 5' flanking DNA containsseveral regions similar to dermed promoter elements. Theseinclude a probable "TATA box," an Spl binding site, andseveral elements related to liver-specific gene expression. Inaddition, we determined that transcription of the glucokinasegene increased at least 20-fold when diabetic rats were treatedwith insulin for 2 hr.

Glucokinase (ATP:D-hexose 6-phosphotransferase, EC2.7.1.1) plays a key role in the regulation of glucose homeo-stasis by catalyzing the first step in glycolysis (1). Expressionof the enzyme is limited to hepatocytes and pancreatic ,B cells(2, 3), and it is regulated differently in these two tissues. Thehepatic enzyme is induced by insulin and repressed by cAMP(4) whereas in the p cell glucokinase activity is increased byglucose (5). The glucokinase gene is, therefore, of interestboth because of its tissue-specific expression and because ofthe several regulatory processes that can be analyzed. Beforethe cis-acting DNA elements responsible for the tissue-specific expression and hormonal regulation of this enzymecan be identified and studied, the structure ofthe glucokinasegene, the transcription unit in each tissue, and the sequenceof its 5' flanking DNA must be determined.Glucokinase is thought to be a member of a family of

hexokinases that have a common evolutionary origin (6). Thisconcept was based on indirect evidence because none of thecomplete structures of the mammalian hexokinases wereavailable. We recently deduced the structure of rat liverglucokinase and found that it shares 33% and 53% amino acidsequence identity with yeast hexokinase and the carboxyl-terminal portion of rat brain hexokinase I, respectively (7).We now have determined the structure of the glucokinasegenet as an initial step toward answering how the varioushexokinase isozymes are related to each other.

EXPERIMENTAL PROCEDURESGeneral Techniques. Standard procedures were used for

screening phage libraries, DNA labeling, restriction enzymemapping, subcloning, isolation of genomic DNA, and South-ern transfers (8). RNA was isolated from the livers ofdiabeticrats by the method of Chirgwin et al. (9). Genomic DNAfragments were subcloned into pEMBL (10) or BluescriptM13+ (Stratagene) plasmid vectors for restriction map anal-ysis and DNA sequencing using the dideoxynucleotidemethod (11). An amplified rat genomic library in EMBL3 wasobtained from G. Scherer (German Cancer Research Center,Heidelberg, F.R.G.). A second, unamplified rat genomiclibrary was constructed in EMBL3 using DNA isolated fromthe liver of a male Sprague-Dawley rat.

Primer-Extension Analysis. A 36-base oligonucleotide (5'-ATGTTCCTGACTCCTGAGGCCACCTGTTGCAGGTGA-3') complementary to sequences near the 5' end of theglucokinase mRNA was synthesized and 5'-end-labeled with[_y-32P]ATP (>5000 Ci/mmol; 1 Ci = 37 GBq) and T4 poly-nucleotide kinase. The primer (3 x 10s cpm) was annealed ina total volume of 20 gl to 20 pug to poly(A)+ RNA; theannealing buffer contained 20 mM Tris-HCl (pH 7.5), 250 mMNaCl, and 1 mM EDTA. After hybridization for 1 hr at 60TCthe reaction mixtures, containing the annealed primer andRNA, were diluted with 130 1A ofa solution containing 50mMTris-HCl (pH 7.5), 40 mM KCl, 10 mM dithiothreitol, 3 mMMgCl2, actinomycin D (75 ,ug/ml), deoxyribonucleotides at0.5 mM each, and 2 units of avian myeloma virus reversetranscriptase (Promega Biotec) and then incubated at 37TC for1 hr. The products of the reactions were size-fractionated ona 5% polyacrylamide/7 M urea gel and visualized by auto-radiography.S1 Nuclease Protection Analysis. Single-stranded DNA,

generated from a 1300-base-pair (bp) (BamHI-BamHI) frag-ment from AGK5, cloned into pEMBL19, was annealed to thesame 36-base primer used in the primer-extension analysisand extended in the presence of [a-32P]dATP and the Klenowfragment ofDNA polymerase I. The DNA was digested withBgl I, the strands were separated on an alkaline agarose gel,and the desired DNA fragment was isolated. The resulting32P-labeled, single-stranded 365-nucleotide DNA fragment(1.2 x 105 cpm) was hybridized to 20 gg of poly(A)+ RNA in30 tkl of a buffer containing 80% (vol/vol) formamide, 40 mMPipes (pH 7.0), 0.4 M NaCl, and 1 mM EDTA. The mixtureswere heated at 75TC for 15 min and incubated at 500Covernight. After the hybridization, 300 p.l of a solutioncontaining 300 mM NaCl, 3 mM ZnSO4, 60 mM sodiumacetate (pH 4.5), and 200 units of S1 nuclease was added andthe reaction mixture was incubated at 37°C for 1 hr. Afterphenol extraction and ethanol precipitation, the DNA wassize-fractionated on a 5% polyacrylamide/7 M urea gel andan autoradiograph was obtained.

Transcription Run-On Analysis. Male Sprague-Dawleyrats weighing 125-150 g (Harlan Animal Supply, Indianapo-lis) were made diabetic as described (12). Rat liver nucleiwere isolated by the method ofSchibler et al. (13), except thatthe suspension buffer was that of Sasaki et al. (14). Theelongation of nascent RNA transcripts was quantitated es-sentially as described by McKnight and Palmiter (15) andmodified by Sasaki et al. (14). The plasmid DNAs bound tothe filters to measure transcription were, for glucokinase,pGK.Z1; for the glucokinase control, Bluescript M13+; forphosphoenolpyruvate carboxykinase (PEPCK), PC116 (16);and for the PEPCK control, pBR322.

Abbreviation: PEPCK, phosphoenolpyruvate carboxykinase.*Present address: Chicago Medical School, North Chicago, IL60064.tTo whom reprint requests should be addressed at: 607 Light Hall,Vanderbilt University, Nashville, TN 37232.tThe sequences reported in this paper have been deposited in theGenBank data base (accession no. M24943 to M24952).

4838

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement"in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 27

, 202

0

Proc. Natl. Acad. Sci. USA 86 (1989) 4839

m CEC) 0o g

H H

-23 Kilobases

- 9.4-6.6

-44

- 2.3-2.0

I

0

FIG. 1. Southern blot analysis of rat genomic DNA using aglucokinase cDNA fragment. High molecular weight genomic DNAwas isolated from a rat liver, digested with EcoRl, HindIII, orBamHI, as indicated, size-fractionated on a 1.0% agarose gel, andtransferred to a nylon membrane. The hybridization probe was a389-bp BstEH-EcoRI 32P-labeled DNA fragment isolated from theGK1 cDNA (7). Size markers were electrophoresed in an adjacentlane.

RESULTSSouthern Analysis of Genomic DNA. A Southern blot hy-

bridization of rat genomic DNA was performed to determinewhether there were multiple glucokinase genes or otherclosely related genes. A fragment of the liver glucokinasecDNA was used to probe EcoRI, BamHI, and HindIIIrestriction digests of rat genomic DNA. As shown in Fig. 1,a single DNA species was seen for each restriction digest.This result suggests that the glucokinase gene is present in therat genome as a single copy.

Screening for the Rat Glucokinase Gene. A glucokinasegenomic DNA fragment, AGK2, was identified by screeninga rat genomic DNA library in EMBL3 with the partial-lengthglucokinase cDNA GK1 (7). The restriction map of the insertin AGK2 (Fig. 2) predicted EcoRI, BamHI, and HindIlIfragments of the lengths seen in the genomic Southern blotexperiment (Fig. 1), indicating that AGK2 contained somepart of the glucokinase gene. A DNA fragment from near the5' end of AGK2 was isolated and used to probe additionalrecombinant bacteriophage from a different, unamplified ratgenomic library to identify DNA segments that would overlapAGK2 but extend farther in the 5' direction. Two additionalclones (AGK5 and AGK9) that overlapped AGK2 and ex-

tended further in the 5' direction were isolated (Fig. 2). AGK5extended 13.5 kilobases (kb) farther than AGK9 so it wassubjected to additional analysis. Thus genomic DNA wasfound to contain 10 exons and the entire hepatic transcriptionunit of the glucokinase gene.Intron-Exon Organization of the Gene. The location of

individual exons within the glucokinase gene was determinedby a combination of Southern blot analysis and DNA se-quencing (Fig. 3). The intron-exon junctions were deter-mined by sequencing the genomic DNA with exon-specificprimers and comparing this sequence with the cDNA se-quence. All of the exons of the glucokinase gene weresequenced (Fig. 3) and these data were compared to thecDNA sequence (7). Three single-base differences betweenthe cDNAs and genomic DNAs were identified§; however,none changed the predicted amino acid sequence of theprotein (7). Splice sites, located at points of cDNA andgenomic DNA homology divergence, were determined byobtaining the best fit to the splice consensus sequence (17).The putative intron-exon splice site sequences are shown inTable 1. The exons ranged in size from 96 bp to 977 bp.

Identification of the Transcription Initiation Sites. The S1nuclease protection assay was used to identify the transcrip-tion initiation site(s) of the glucokinase gene. After hybrid-ization of the S1 probe to poly(A)+ RNA isolated from thelivers of diabetic and insulin-treated diabetic rats, the hy-bridization products were treated with S1 nuclease, and thenthe protected DNA fragments were size-fractionated andvisualized by autoradiography. A portion of the autoradio-graph obtained is shown in Fig. 4A. Lanes A-D show DNAsize markers from a dideoxynucleotide sequencing reaction.S1 nuclease-resistant fragments were obtained when RNAfrom the insulin-treated diabetic rats was used (lane E) butnot when the RNA was isolated from the untreated diabeticrats (lane F). The protected DNA fragments varied in sizefrom 83 to 89 bases but the strongest bands were 87-89 baseslong, centered around an adenine.A primer-extension assay was also used to locate the

transcription initiation site. A portion of the autoradiographobtained from this experiment is shown in Fig. 4B. Lane Eshows the DNA products of a primer-extension reactionusing poly(A)+ RNA from the livers of insulin-treated dia-betic rats and lane F shows the products from poly(A)+ RNAfrom the livers of diabetic rats. The following results corre-spond with the S1 nuclease protection experiment: (i) Prod-ucts derived from glucokinase mRNA were seen only in lane

§The sites at which the glucokinase cDNA and genomic DNAsequences differ are described in the sequence data submitted toGenBank.

XGK9XGK5

XGK2

H I4c o

ay 0n

pGK.E3 pGK.E2

H Ha0E a(n cn

pGK.EI

I I I 1 * * * _2 3 4 56 7 8 9 10

FIG. 2. Structure of the rat glucokinase gene.The structure of the rat glucokinase gene is illus-trated. Shown, from top to bottom, are: (i) theoverlapping rat glucokinase genomic DNA clonesin the bacteriophage A vector EMBL3; (it) therestriction map of the cloned DNA using theenzymes EcoRI, HindIlI, BamHI, Bgl II, Kpn I,Sac I, and Sma I; (iiM) plasmid subclones of thegene; and (iv) the location of the 10 exons asdepicted by the solid boxes. The scales in kilo-bases is indicated below.

EcoR IBamH IHind mOthers

P CL

pGK.HE

1kb

I

I

a --I

I I LL- I

I --- I I

Biochemistry: Magnuson et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 27

, 202

0

4840 Biochemistry: Magnuson et al.

ECD co _ I

i -H- - H zmi\-H i-

I * _al

l --

I ,

* I

100 bpF-H

He

Z CL

CL x-vsmv

a.

36gz~

Ur 0ur cw 4 I

_& a

sDX 7

I 8 a.

. I

E 'a CL

., I

obtain it is shown in Fig. 3. As can be seen by examining thissequence, the cap site is located 127 bases upstream from theATG initiation codon. Eleven additional nucleotides notpresent in the GK.Z2 cDNA (7) were placed in the first exonof the gene by this analysis.

Regulation of Glucokinase Gene Transcription by Insulin.As an initial step in the effort to understand how insulin exertspositive and negative effects on transcription in the same cell,we quantitated transcription of the PEPCK and glucokinasegenes in the same aliquots of nuclei isolated from the liversof diabetic rats. Transcription of the glucokinase gene was atthe limit of detection in the absence of insulin (Fig. 6) butincreased to 15 ppm at 0.5 hr, to 309 ppm at 1 hr, and to 619ppm at 2 hr after insulin treatment. Transcription of thePEPCK gene decreased from 6954 ppm in the untreateddiabetic animal to 1666 ppm at 0.5 hr, to 754 ppm at 1 hr, andto 1027 ppm after 2 hr of insulin treatment. A somewhatdifferent transcription run-on experiment was performed tovalidate the magnitude of the induction of glucokinase genetranscription by insulin. Nuclei were isolated from untreateddiabetic rats and from diabetic rats treated with insulin for 2hr. In this experiment the hybridization filters were exposedto film for 2 weeks and the induction was quantitated bydensitometry. Transcription of the glucokinase gene in-creased at least 20-fold (data not shown).

_w x

'0 'PCIDm E

//s-I

FIG. 3. Exon locations and sequencing strategy. The strategyused to sequence the exons and 5' flanking DNA of the glucokinasegene is indicated. The sizes of the exons are, for exons 1 to 10,respectively, 172, 163, 155, 120, 96, 100, 184, 156, 234, and 977 bp.Each exon is shown as a hatched box and the sequencing strategy isindicated below.

E, which contains RNA isolated from the insulin-treated rats.(it) Bands of 85-88 bases long were seen. (iii) Transcriptioninitiation was localized over a range of 4 bases. (iv) Thestrongest band, 88 nucleotides long, corresponds to initiationat an adenine on the glucokinase gene. Since this base is nearthe center of the size range seen for the S1 nuclease protec-tion experiment (compare Fig. 4 A and B), it was designatedas position + 1 for numbering the nucleotides in the glucoki-nase gene.

Glucokinase Promoter Sequence. The sequence of the DNAon both sides of the transcription initiation site was deter-mined. A Bgl II site located 1448 bp upstream of adenine + 1was the limit for sequence determination in the 5' direction.This sequence is shown in Fig. 5 and the strategy used to

DISCUSSIONStructure of the Rat Glucokinase Gene. The 15.5-kb glu-

cokinase gene consists of 10 exons separated by 9 introns. Itappears to be present as a single copy. The location of thestart site of transcription (127 bases upstream from thetranslation initiation codon) predicts that the mRNA pro-duced from this gene in the liver is 2357 bases long, excludingthe poly(A) tail. The putative translation initiation codon islocated within exon 1, and the translation termination signalis located in exon 10. The portion of the protein homologouswith the ATP binding domains in a variety of proteins [aminoacids 78-102 (7)] is encoded entirely by DNA located withinexon 3. The region that constitutes the core of the glucose-binding domain [amino acids 144-171 (7)] is encoded by partsof exons 4 and 5.

Potential Regulatory Motifs. It is likely that multiple tran-scription factors, some of which may be liver-specific, inter-act with DNA elements in the glucokinase promoter. Thesequence TATTT, located at positions -29 to -25, may bea "TATA box," although it is not a good consensus sequence(17). The TATA box is generally considered to be importantfor the precise positioning of transcription initiation (17). Therather poor TATA homology may explain the observationthat the glucokinase gene initiates over several bases. Thesequence CCCCCGCCCC (at positions -442 to -433) is a

Table 1. Identification of intron-exon boundariesIntronnumber 5' intron junction 3' intron junction

1 TTGTTGACTCTG/gtaagggccatt cccccgcacag/GTCGAGCAGATC2 CAGAAGGCTCAG/gtaccgcaggtt ctotgcctgcag/AAGTCGGAGACT3 ACTGCCGAGATG/gtgagcagcctg tgcaactcctag/CTCTTTGACTAC4 GACCTAGACAAG/gtgagccgggtg cctaccttacag/GGCATCCTCCTC5 AAGAGGAGAGGG/gtgagcacagcg acttcttggcag/GACTTTGAGATG6 GCATGATTGTGG/gtaagggcttot ccctccctctag/GCACTGGCTGCA7 CGGTCAGCAGCT/gtaaggatgctc ttctgtatccag/GTACGAGAAGAT8 ACAAGTGGAGAG/gtgcctgcaggg tcctgcccgcag/CGACTCCGGGGA9 GCTGCACCCGAG/gtcagcttccac cctcctgttcag/CTTCAAGGAGCG

The consensus sequences determined for the splice donor and acceptor sites of exons from class IIgenes are MAG/gtragt and yyyyyynyag/GK, respectively (17), where M is A or C, R is G or A, Y isT or C, K is G or T, and N is any base. Exon sequences are designated by uppercase letters, intronsequences are designated by lowercase letters, and splice sites are designated by the slashes (/).

I

Proc. Natl. Acad. Sci. USA 86 (1989)

755 CID 6

or iffm--G±S.L-0

rCL

OL(f)

a.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 27

, 202

0

Proc. Natl. Acad. Sci. USA 86 (1989) 4841

A B C D E_ _ _~' F

I~~~~~~~_m s_ T

T: _ A.4-A

G_-T_ cJ

_f-i GJ

Si Nuclease Protection

B A B C D E

100 - o_

F

t ,

70- _-

Primer Extension

FIG. 4. Identification of the transcription initiation site of the hepatic glucokinase gene. (A) An S1 nuclease protection assay was performed.Lanes A-D show A, C, G, and T dideoxynucleotide sequencing reactions used to define nucleotide positions. The S1 nuclease protectionexperiment was performed using 20 ,ug of poly(A)+ RNA isolated from livers of diabetic rats treated with insulin for 4 hr (lane E) or using 20Ag of poly(A)+ RNA isolated from livers of untreated diabetic rats Oane F). The length of the DNA fragments protected is indicated by the sizescale to the left. The nucleotides surrounding the transcription initiation site are shown to the right. (B) A primer-extension assay was performed.Lanes A-D show A, C, G, and T dideoxynucleotide sequencing reactions used to define the nucleotide positions. The primer-extensionexperiment was performed using 20 j&g of poly(A)+ RNA isolated from livers of diabetic rats treated with insulin for 4 hr (lane E) or using 20,ug of poly(A)' RNA isolated from livers of untreated diabetic rats (lane F). The length of the DNA fragments generated is indicated by the sizescale to the left. The nucleotides surrounding the transcription initiation site are shown to the right.

9/10 match for the reverse complement of the consensusbinding site (G/T)GGGCGG(G/A)(G/A)(C/T) of the Spltranscription factor (18, 19).The 5' flanking DNA sequence ofthe glucokinase gene was

searched for the presence of elements found in severalliver-specific genes (20, 21). Hepatocyte nuclear factor Ibinds to the DNA sequence ATTAAC in a region of thea1-antitrypsin gene demonstrated to be essential for tissue-specific expression (20). Elements containing the same AT-TAAC core sequence are located in the promoters of the a-

and P-fibrinogen genes and appear to bind the same protein(20). The sequence ATTAAC is located in the promoterregion of the glucokinase gene (positions -171 to -166).Liver factor Al, another protein that may be important forliver-specific gene expression, binds to the DNA sequenceTG(G/A)(A/C)CC (21). This motif is present in the regula-tory region of the human a1-antitrypsin, apolipoprotein Al,and haptoglobin-related genes (21). Several copies of thisDNA sequence are located in the 5' flanking region of theglucokinase gene. Single copies of the sequence lie at posi-tions -795 to -790 and -684 to -679, while three tandemrepeats of the sequence, TGGCCC, occur between positions-70 and -53.

Insulin Regulates Glucokinase mRNA Synthesis in the Liver.Our results indicate that the increase of glucokinase mRNAin response to insulin treatment of diabetic rats occurs

-1448

-1400 cAGTACT-1300_

-1200 GGAGGACACTT ATACA

-1100Ac_-1000

900 GGcC&GCTrrGCTCTATr!TGGGMAGcGATCMT.TGTCAAGGGGTGAcC&TACTGCcCAATCrrGGGAGMcCTGGC&TrccX&CAGGC&ccCYA.G- 800A

R-S- 700 AGGAAGGTcTGCTCTAMGCTCCTGGTrGCGCMT- XAAT- 500- 500

(b)- 400 COGGCAGGTGECC>(ATCCCACCGGGTCCCCA

- 300 TATTCACAATTAAGCCT CAC~:IG&CC~:GCATGAAGGTG

(C ) ----_ --

100

+ 1

+ 101

FIG. 5. Sequence of the glucokinasepromoter region. This figure shows thesequence of the 1448 bases of DNA thatflank the 5' side of the transcription ini-tiation site and the 130 bases of sequenceon the 3' side of the transcription initia-tion site. Putative regulatory motifs, de-scribed in the Discussion, are underlinedand designated as: (a), TATA box; (b),Spl binding site; (c), hepatocyte nuclearfactor I binding site; and (d), liver factor-Al binding site. The translation initiationcodon, ATG, is located at position +128and is designated as (e).

primarily at a transcriptional level. The magnitude and ki-netics of the increase, at least a 20-fold change within 2 hr,correspond well with data reported (22). The increase intranscription is rapid; it occurs within 1 hr of injection of theanimal with insulin. The rapidity of insulin's action on thetranscription of the glucokinase gene suggests that a directeffect is involved, as is the case in the suppression oftranscription of the PEPCK gene by insulin (23). Transcrip-tion of the PEPCK gene was inhibited while the glucokinasegene was stimulated by insulin treatment, but the kinetics ofthe changes in the transcription rates for the two genes wereremarkably similar. In the diabetic animal the PEPCK geneis maximally stimulated (almost 7000 ppm) while the glucoki-nase gene is virtually inactive. Thus, these promoters havedramatically different basal activities in absence of insulin,and in the presence of insulin the transcription of one gene isstimulated while that of the other is inhibited.

Insulin response elements (IREs) probably reside up-stream of genes regulated by insulin. In addition to PEPCK(23) and glucokinase, transcription of the growth hormone(24), glyceraldehyde-3-phosphate dehydrogenase (25), gene33 (26, 27), c-fos (28), and amylase (29) genes is regulated byinsulin. A consensus IRE has not been identified; however,when 5' flanking DNA from some of these genes is linked toreporter genes and expressed either in responsive cell lines orin transgenic mice, the expected effect of insulin is conferred

A

100

90

80-

70-

hG&TtrC&TcTAhfGCAL...ii,.L..1IALb,,L'.AA.J.LLi.,L'.Af'".

(d) (d) (d) (a)AGrGlCCTGAGTTCrrCrTTCFGGcATcCTcGGcCGAcAGTCCrGcAC&I ccCATc cTCTACc

MC~~alT

Biochemistry: Magnuson et al.

Dow

nloa

ded

by g

uest

on

Nov

embe

r 27

, 202

0

4842 Biochemistry: Magnuson et al.

G)c0

--

V 0

_.0o~~~~~~~~~~~~~_~~~~~~~~~~~~~~~~

CD

0 30 60 90 120Insulin Treatment Time

(min)

FIG. 6. Effect of insulin on transcription of the glucokinase andPEPCK genes. Rats made diabetic with streptozotocin were treatedwith insulin for various times, then the hepatic nuclei were isolatedand used to quantitate transcription of the PEPCK and glucokinasegenes. Each data point represents the average of duplicate transcrip-tion assays using nuclei isolated from two rats for each treatmentcondition. The scale, in parts per million (ppm) ofRNA transcription,is expanded 10 times for the glucokinase gene (right axis) comparedto that for the PEPCK gene (left axis).

upon the reporter genes (30). It will be interesting to comparethe IREs from a gene stimulated by insulin (e.g., glucokinase)with one inhibited by insulin (e.g., PEPCK) to determinewhether similar or different IREs are involved and whethersimilar or different transcription factors are involved.

Concluding Statement. The availability of the glucokinasegene, including significant 5' flanking DNA, makes possiblethe construction of fusion genes that can be used to identifythe elements that confer tissue-specific and hormone-regulated expression. The characterization of this gene alsorepresents an important step toward elucidating the evolutionof the hexokinase enzyme family.

Note Added in Proof. Two recent papers present evidence in favor ofthe hypothesis that the mammalian hexokinases of -100 kDa evolved

from an -50-kDa ancestral molecule through a process of geneduplication and tandem ligation (31, 32).

We thank Emmanuel Eusebio, Charles Davis, III, Deborah Ca-plenor, and Elizabeth Zimmerman for their assistance. This workwas supported by Grant DK 35107 (D.K.G.) from the NationalInstitutes of Health, by the Vanderbilt Diabetes Research andTraining Center (DK 07061), and grants from the Juvenile DiabetesFoundation (M.A.M. and D.K.G.). M.A.M. is the recipient of aResearch Career Development Award from the American DiabetesAssociation.

1. Weinhouse, S. (1976) Curr. Top. Cell. Regul. 11, 1-50.2. Meglasson, M. D., Burch, P. T., Berner, D. K., Najafi, J.,

Vogin, A. P. & Matschinsky, F. M. (1983) Proc. Natl. Acad.Sci. USA 80, 85-89.

3. lynedjian, P. B., Mobius, G., Seitz, H. J., Wollheim, C. B. &Renold, A. E. (1986) Proc. Nati. Acad. Sci. USA 83, 1998-2001.

4. Sibrowski, W. & Seitz, H. J. (1984) J. Biol. Chem. 259,343-346.

5. Bedoya, F. J., Matschinsky, F. M., Shimizu, T., O'Neill, J. J.& Appel, M. C. (1986) J. Biol. Chem. 261, 10760-10764.

6. Lawrence, G. M. & Trayer, I. A. (1984) Comp. Biochem.Physiol. B 79, 233-238.

7. Andreone, T. L., Printz, R. L., Pilkis, S. J., Magnuson, M. A.& Granner, D. K. (1989) J. Biol. Chem. 264, 363-369.

8. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) MolecularCloning: A Laboratory Manual (Cold Spring Harbor Labs.,Cold Spring Harbor, NY).

9. Chirgwin, J. M., Przybyla, A. E., MacDonald, R. J. & Rutter,W. J. (1979) Biochemistry 18, 5294-5299.

10. Dente, L., Cesareni, G. & Cortese, R. (1983) Nucleic AcidsRes. 11, 1645-1655.

11. Sanger, F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl.Acad. Sci. USA 74, 5463-5467.

12. Beale, E., Andreone, T., Koch, S., Granner, M. & Granner,D. K. (1984) Diabetes 33, 328-332.

13. Schibler, U., Hagenbuchle, O., Wellaur, P. K. & Pittet, A. C.(1983) Cell 33, 501-508.

14. Sasaki, K., Cripe, T. P., Koch, S. R., Andreone, T. L., Pe-tersen, D. D., Beale, E. G. & Granner, D. K. (1984) J. Biol.Chem. 259, 15242-15251.

15. McKnight, G. S. & Palmiter, R. D. (1979) J. Biol. Chem. 254,9050-9058.

16. Beale, E. G., Chrapkiewicz, N. B., Scoble, H. A., Metz, R. J.,Quick, D. P., Noble, R. L., Donelson, J. E., Biemann, K. &Granner, D. K. (1985) J. Biol. Chem. 260, 10748-10760.

17. Breathnach, R. & Chambon, P. (1981) Annu. Rev. Biochem. 50,349-383.

18. Kadonaga, J. T., Jones, K. A. & Tjian, R. (1986) TrendsBiochem. 11, 20-23.

19. Dynan, W. S. & Tjian, R. (1985) Nature (London) 316, 774-778.-

20. Courtois, G., Morgan, J. G., Campbell, L. A., Fourel, G. &Crabtree, G. R. (1987) Science 238, 688-692.

21. Hardon, E. M., Frian, M., Paonessa, G. & Cortese, R. (1988)EMBO J. 7, 1711-1719.

22. lynedjian, P. B., Ginovci, A. & Renold, A. E. (1988) J. Biol.Chem. 263, 740-744.

23. Magnuson, M. A. & Granner, D. K. (1988) Recept. Biochem.Methodol. 12B, 189-209.

24. Yamashita, S. & Melmed, S. (1986) J. Clin. Invest. 78, 1008-1014.

25. Alexander, M. C., Lomanto, M., Nasrin, N. & Ramaika, C.(1988) Proc. Natl. Acad. Sci. USA 85, 5092-5096.

26. Messina, J. L., Hamlin, J., Azizkahn, J. & Larner, J. (1985)Biochem. Biophys. Res. Commun. 133, 1168-1174.

27. Chu, D. T., Davis, C. M., Chrapkiewicz, N. B. & Granner,D. K. (1988) J. Biol. Chem. 263, 13007-13011.

28. Stumpo, D. J., Stewart, T. N., Gilman, M. Z. & Blackshear,P. J. (1988) J. Biol. Chem. 263, 1611-1614.

29. Osborne, L., Rosenberg, M. J., Keller, S. A. & Meisler, M. H.(1987) Mol. Cell. Biol. 7, 326-334.

30. Magnuson, M. A., Quinn, P. G. & Granner, D. K. (1987) J.Biol. Chem. 262, 14917-14920.

31. Nishi, S., Seino, S. & Bell, G. I. (1988) Biochem. Biophys. Res.Commun. 157, 937-943.

32. Schwab, D. A. & Wilson, J. E. (1989) Proc. Natl. Acad. Sci.USA 86, 2563-2567.

.2-Ecs

0

a-

%a -

Proc. Natl. Acad Sci. USA 86 (1989)

Dow

nloa

ded

by g

uest

on

Nov

embe

r 27

, 202

0