Glucokinase Mutations Associated with Non-Insulin-Dependent

Proc. Natl. Acad. Sci. USA Vol. 90, pp. 1932-1936, March 1993 Biochemustry Glucokinase mutations associated with non-insulin-dependent (type 2) diabetes mellitus have decreased enzymatic activity: Implications for structure/function relationships (gene mutations/enzyme structure/frnction) M. GIDH-JAIN*, J. TAKEDAt, L. Z. Xu*, A. J. LANGE*, N. VIONNETt, M. STOFFELt, P. FROGUEL0, G. VELHOt, F. SUNt, D. COHENt, P. PATEL§, Y.-M. D. Lo¶, A. T.- HATrERSLEY§, H. LUTHMANII, A. WEDELLII, R. ST. CHARLES**, R. W. HARRISON**, I. T. WEBER**, G. I. BELLt, AND S. J. PILKIS* *Department of Physiology and Biophysics, State University of New York, Stony Brook, NY 11794; tHoward Hughes Medical Institute and Departments of Biochemistry and Molecular Biology and Medicine, University of Chicago, Chicago, IL 60637; $Centre d'Etude du Polymorphisme Humain, 27 rue Juliette Dodu, 75010 Paris, France; IDepartment of Haematology, John Radcliffe Hospital, and Diabetes Research Laboratories, Radcliffe Infirmary, Oxford OX3 9DU, United Kingdom; lDepartment of Pathology and Bacteriology, John Radcliffe Hospital, Oxford OX3 9DU, United Kingdom; IIRolf Luft Center for Diabetes Research, Department of Clinical Genetics, Karolinska Hospital, S-104 01 Stockholm, Sweden; and **Department of Pharmacology, Jefferson Cancer Institute, Philadelphia, PA 19107 Communicated by William J. Lennarz, December 12, 1992 (receivedfor review October 5, 1992)

ABSTRACT The glycolytic enzyme glucokinase plays an glucose concentration (3, 9, 10). By contrast, glucokinase important role in the regulation of insuln secretion and recent provides about 80% of the total glucose phosphorylating studies have shown that mutations in the human glucokinas activity in mammalian liver. In this tissue, the expression of gene are a common cause of an autosomal domint form of a glucose phosphorylating activity that has a high Km for non-inssulin-dependent (type 2) diabetes meilitus (NIIDDM) that glucose and relative lack ofinhibition by glucose 6-phosphate has an onset often during childhood. The majority of the serves to maintain a gradient for glucose entry into the mutations that have been identified are missense mutations that hepatocyte, which is particularly important following a meal result in the syntheis ofa glucokinas molecule with an altered when plasma glucose levels are elevated (4, 7). amino acid sequence. To characterize the effect of these We have recently shown that mutations in the glucokinase mutations on the catalytic properties of human 3-cell glucoki- gene can cause an autosomal dominant form of non-insulin- nase, we have expressed native and mutant forms of this dependent diabetes mellitus (NIDDM) with onset usually protein in Eschenchia coli. AU of the missense mutations show before 25 years of age (11-13). Mutations in the glucokinase chnges in enzyme activity including a decrease in V.. and/or gene have been found in about 60%o of French families with increase in K. for glucose. Using a model for the three- early-onset NIDDM (refs. 11 and 13; P.F., unpublished data). dimensional structure of human glucokinas based on the They have also been found in British (12), Swedish (H.L., crystal structure of the related enzyme yeast hexokinae B, the unpublished data), and African-American (G.I.B., unpub- mutations map primarily to two regions of the protein. One lished data) patients with NIDDM and thus are not restricted group of mutations is lcated in the active site cleft separating to a particular ethnic or racial group. Although many muta- the two domains of the enzyme as wefl as in surface loops tions in glucokinase have now been identified, the conse- leading into this cleft. These mutations usually result in large quences of these mutations on enzyme activity are not yet reductions in enzyme activity. The second group of mutations known. The object of this study was to express various is located far from the active site in a region that is predicted mutant forms of human 13-cell glucokinase in bacteria and to to undergo a substrate-induced conformational change that compare their kinetic properties with those of the native results in closure of the active site cleft. These mutations show protein. The results are consistent with the hypothesis that a smail -2-fold reduction in V., and a 5- to 10-fold increase gene dosage is the mechanism for glucokinase gene mutation- in K. for glucose. The characterization of mutations in glu- induced glucose intolerance in this form of NIDDM. cokinase that are asted with a distinct and readily recog- nizable form of NIDDM has led to the identification of key EXPERIMENTAL PROCEDURES amino acids involved in glucokinas catalysis and locaized Materiais. Glucosamine-Sepharose was prepared as de- functionaUly important regions of the glucokinas molecule. scribed by Miwa et al. (14). Construction ofpET Expresso Plasmids for Human (-Ceil The phosphorylation of glucose in mammalian tissues is me- and Liver Glucokinase. A 2.6-kb human pancreatic (-cell diated by a family ofhexose phosphotransferases (1-3). These glucokinase cDNA clone, phGK-20 (15), was used to gener- include glucokinase and hexokinases I, II, and III. Glucoki- ate the construct pEhgk-WT. An Nde I site was generated at nase is expressed in insulin-secreting pancreatic 13 cells and the 5' end using PCR with an oligonucleotide that had a 1-bp hepatocytes and is characterized by a high Km and specificity mismatch (GGC TGG TGT GCA TAT GCT GGA CGA for glucose and a relative lack ofproduct inhibition by glucose CAG). The insert in the pET 3a expression construct in- 6-phosphate compared to the hexokinases (4-6). It plays an cluded the protein coding region of the cDNA as well as the important role in regulating and integrating glucose metabo- 3' untranslated region. lism in both tissues (3, 6-8). In pancreatic 1 cells, glucokinase, The sequences ofhuman 1-cell and liver glucokinase differ which represents about 20%o of the total glucose phosphory- at their NH2 termini (13, 16). The human liver glucokinase lating activity, and the GLUT2 glucose transporter are be- expression plasmid pEhGK was constructed from an Pst lieved to act in concert as the "glucose-sensing apparatus" I-EcoRI fragment of the 13-cell glucokinase cDNA (encoding modulating insulin release in response to changes in plasma amino acids 16-466) and two pairs of complementary oligo- nucleotide linkers. The 5' linkers (Nde I/Pst I) encoding the 16 N-terminal amino acids and the translation initiation site The publication costs ofthis article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. Abbreviation: NIDDM, non-insulin-dependent diabetes mellitus. 1932 Downloaded by guest on September 28, 2021 Biochemistry: Gidh-Jain et al. Proc. Natl. Acad. Sci. USA 90 (1993) 1933 were phosphorylated, annealed, and ligated to the Nde I site 50-fold after 2 hr. The native and mutant forms of p-cell of the pET 3a vector and the Pst I site of the cDNA. glucokinase were expressed at similar levels with the excep- Site-Directed Mutagenesis of Human (3-Cell Glucoinase. In tion ofthe truncated form ofthe enzyme, which has a Glu-279 vitro mutagenesis ofhuman 3-cell glucokinase was carried out -- amber mutation and lacks residues 279-465. We were using the Altered Sites in vitro mutagenesis system (Promega). unable to detect expression ofthis nonsense protein presum- All mutations were confirmed by DNA sequencing. ably because it was unstable and rapidly degraded. Like the Bacterial Expression and Purification of Native Human native enzyme, all ofthe mutant forms were recovered in the «3Cell and Liver Glucokinases and 3-Cell Human Glucokinase soluble fraction and could be readily purified to homogeneity. Mutants. Native and mutant human glucokinases were ex- Effects of Missense Mutations on (-Cell Activity. The mu- pressed in Escherichia coli essentially as described (17, 18). tations showed a wide range of effects on Vn,,, and Km for Glucokinase was purified 20-fold from extracts of E. coli in glucose with no detectable effect on affinity for ATP (Table four steps, including (NH4)2SO4 precipitation (45-65%), gel 1). Since the kinetic parameters of the p-cell and liver filtration on a Sephadex G-100 column, chromatography on isoforms are very similar (Table 1), we expect that these a glucosamine-Sepharose column, and FPLC Mono Q-Seph- mutations will have the same effect on the activity of liver arose chromatography. The latter step was necessary to glucokinase. However, this issue needs to be addressed remove a small amount of low Km hexokinase activity in the directly. The higher efficiency ofbacterial expression ofliver cell extract that could confound kinetic studies. The purified glucokinase compared to the p-cell isoform has been reported protein was homogeneous as judged by SDS/PAGE. Native by others (25) for the corresponding rat isoforms and suggests (-cell glucokinase had a specific activity of 100 units/mg, that the first 15 NH2-terminal amino acids affect stability which is similar to that reported for the purified rat liver and/or proper folding of glucokinase in E. coli cytosol. enzyme (19). The purified recombinant protein was also Structure/Function Analysis ofGlucoinase Mutations. We subjected to NH2-terminal sequence analysis. The protein, have used the three-dimensional structure of the related analyzed through 12 cycles, gave the expected sequence: protein yeast hexokinase B to develop a model of human Met-Leu-Asp-Asp-Arg-Ala-Arg-Met-Glu-Ala-Ala-Lys-. p-cell glucokinase (13, 20-23). As shown in the alignment in Modeling ofHuman P-Cell Glucokinase Structure. A simple Fig. 3, there are 140 identical residues (-30% identity) model of human p-cell glucokinase was generated using the between human p-cell glucokinase and yeast hexokinase B, known crystal structure of yeast hexokinase (20-23). and we predict that the human p-cell glucokinase will have the same arrangement of a-helices and p-strands as observed RESULTS in the crystal structure of yeast hexokinase. The locations of Expression of Native and Mutant Forms of Human Glucoki- the 11 missense mutations are indicated in Fig. 4, which nase in E. coUl. Mutations in the glucokinase gene associated represents a model of the open form of human p-cell glu- with early-onset NIDDM have been found in 9 ofthe 12 exons cokinase. Glucokinase, like hexokinase, is predicted to fold (Fig. 1) and include splicing, nonsense, and missense muta- into a small domain and large domain, which are separated by tions, the latter of which provide an opportunity to address a deep cleft (21). When glucose binds at the bottom of the structure/function relationships in this important regulatory cleft, there is a large conformational change that closes the enzyme ofglycolysis. To date, there has been only one study cleft between the two domains and is essential for yeast employing site-directed mutagenesis to investigate structure/ hexokinase B catalysis. In the op)en form of glucokinase, function relationships of glucokinase (18). In that study, Asn-204, Asp-205, Asn-235, Glu-256, and Glu-290 are pre- mutation of Asp-205 -* Ala in the rat liver enzyme resulted dicted to form hydrogen-bond interactions with hydroxyls of in a 500-fold decrease in the kcat of the reaction with no bound glucose (Fig. 4 and ref. 23). change in the Km for glucose or ATP. This finding is consis- As shown in Fig. 4, many of the mutations map to the tent with this carboxylate group hydrogen bonding to the region of the active site cleft or to surface loops leading into phosphoryl acceptor and functioning as a general base cat- the cleft. Three ofthe mutations involve amino acids that are alyst (20-23). predicted to be near the active site, Val-203, Thr-228, and The human (3-cell form of glucokinase was readily ex- Glu-256 (Figs. 3 and 4; Table 1), and may form part of the pressed in E. coli (Fig. 2A) using the T7 RNA polymerase- glucose binding site. Val-203 is next to Asn-204 and Asp-205, based expression system (24); easily detectable levels of both of which are predicted to contact glucose. Mutation of glucokinase activity were found in the soluble fraction ofcell Val-203 -- Ala caused a 200-fold decrease in V,. Since this extracts after 15 min of induction of expression with isopro- residue is Ile in yeast hexokinase B (Fig. 3), the hydrophobic pyl P-D-thiogalactopyranoside and activity increased nearly character of the amino acid in this position may be very Exon la lb lc 2 3 4 5 6 7 8 9 10 cL1

E70K E98X G175R V203A T228M G299R K414E S418-1G-*C V182M E256K E3000 R186X G261R E300K K161+2Gdel5 E279X L309P E279Q G227-2A-T

FIG. 1. Schematic structure of the human glucokinase gene and localization of mutations identified in patients with NIDDM. The (-cell isoform of glucokinase is encoded by exons la and 2-10 (13). Two different glucokinase transcripts have been identified in human liver mRNA; the major transcript is encoded by exons lb and 2-10, and the minor transcript is encoded by exons lb, lc, and 2-10. Exons la, lb, and lc encode the 15, 16, and 14 NH2-terminal residues ofthe different glucokinase isoforms. Except for these amino acids, the sequences ofthe three isoforms are otherwise identical. The (-cell, liverlb, and liverlc transcripts encode proteins of 465, 466, and 464 amino acids, respectively. In this report amino acids are designated relative to the sequence of human (-cell glucokinase. The single-letter abbreviations for amino acids are used. "X" indicates a nonsense mutation. The splicing mutations are as follows: K262+2Gdell5 is a 15-bp deletion that removes the T of the GT in the splice donor site of intron 4 and the following 14 bp resulting in a GG; G227-2A-*T is a mutation of the splice acceptor site in intron 6 from AG-. TG; and S418-1G-*C is a mutation of the splice acceptor site in intron 9 from AG -+ AC. The mutations that were expressed in E. coli

-- -- and characterized kinetically (Table 1) include those from French [Gly-175 -. Arg, Val-182 Met, Val-203 -. Ala, Thr-228 Met, Gly-261

Arg, Glu-279 amber, Gly-300 Lys, Glu-300 Gln, Leu-309 Pro (refs. 11 and 12; P.F., unpublished data)], Swedish [Glu-256 -. Lys

(H.L., unpublished data)], British [Gly-299 -. Arg (12)], and African-American [Glu-279 Gln (G.I.B., unpublished data)] patients. Downloaded by guest on September 28, 2021 1934 Biochemistry: Gidh-Jain et al. Proc. Natl. Acad ScL USA 90 (1993) A B wIT G261R E300Q E300K T228 v203A G175R v182m 200- 200 -

116 ...... 116 - 97- 6- 0O

...... nS,,S ,? .. G K- 66 45s

..j.:::::?':......

29- 6 i i -' ~~~~~~~~~ ~ ~~~~~ ~~~~~~~~~~~~~~~...... L ~ -: i_

29 -

HOURS: 0 2 8 24 IPTG:;, + - . d + - + + - + FiG. 2. Expression of native and mutant forms of human p-cell glucokinase in E. coli. (A) Coomassie blue-stained SDS/PAGE showing the induction of native human p-cell glucokinase. A growing culture of BL21(DE3)pLysS cells carrying pEhGK-WT (20) was induced by addition of0.4 mM isopropyl P-D-thiogalactoside (IPTG) atA6w and allowed to grow for 0, 2, 8, and 24 hrat 22C. A portion ofcells (100 AI) was dissolved in SDS/PAGE loading buffer and subjected to SDS/PAGE. Glucokinase (GK) has a molecular size of 50 kDa and is indicated. (B) Coomassie blue-stained SDS/PAGE of wild-type and mutant glucokinases (see text). Cells were treated with 0.4 mM IPTG for 24 hr. The native (WT) and the mutant forms of p-cell glucokinase shown on this SDS/PAGE are indicated with an arrow. The native and mutant forms ofglucokinase were recovered in the soluble fraction to a similar extent. This was also true of the other expressed mutants (data not shown). The single-letter abbreviations for amino acids are used. important and substitution with a less hydrophobic amino resulted in a large decrease in maximal velocity that could not acid such as Ala may alter the conformation ofthe active site. be overcome by increasing the ATP/Mg concentration to 20 Thr-228 is a conserved amino acid in all hexokinases and mM (data not shown). Glu-256 is predicted to hydrogen bond glucokinases and its mutation was predicted to affect affinity with glucose and its mutation to Lys resulted in a large for glucose and/or ATP (13). The Thr-228 -- Met mutation decrease in maximal velocity but with no significant change in Km for glucose. Mutation of Glu-256 to a basic amino acid Table 1. Enzymatic properties of native and mutant forms of would remove this hydrogen bond and thus reduce the human P-cell glucokinase affinity for glucose and/or interfere with the open/closed Level of Km, mM equilibrium of the enzyme. In support of this hypothesis, the expression, Vmax, Gibbs free energy (AAG) estimated from the relative ratios of Mutation mg/liter units/mg Glucose ATP Vmax/Km of the Glu-256 -+ Lys mutation and the native Native liver enzyme is about 3.0 kcal/mol (1 kcal = 4.18 kJ), which is isozyme 80 98 ± 9 6.8 ± 1.3 0.23 ± 0.19 consistent with the energy of one hydrogen bond of 3.5-4.5 Native p-cell kcal/mol (26). However, Km is not simply correlated with isozyme 20 100 ± 8 8 ± 2 0.15 ± 0.18 binding affinity, especially when a significant conformational Gly-175 Arg 20 51 ±* 39 ± 12* 0.10 ± 0.03 change occurs during catalysis, so that the effect ofmutation Val-182 Met 10 49 ± 6* 70 ± 9* 0.20 ± 0.15 ofa binding residue such as Glu-256, or perhaps Thr-228, may Val-203 Alat 15 0.5 ± 0.04* 100 ± 20* 0.20 ± 0.05 be on V. rather than Km, which is what was observed. Thr-228 Mett 10 0.4 ± 0.03* 10 ± 2 0.20 ± 0.17 Other mutations affect residues that are conserved be- Glu-256 Lyst 8 0.25 ± 0.02* 2.4 ± 2.0 0.20 ± 0.13 tween human 3-cell glucokinase and yeast hexokinase B Gly-261 Argt 8 c0.46* 2.5 ± 2.1 0.20 ± 0.10 (Val-182, Glu-279, Gly-299, and Glu-300), implying that they Glu-279 amber § are functionally important. For example, Glu-300 is predicted Glu-279 Gln 15 55 ± 6* 41 ± 1* 0.20 ± 0.2 to form a salt bridge with Arg-303, a residue that is also Gly-299 Argt 5 _0.32* 3.1 ± 2.1 0.15 ± 0.1 conserved (Fig. 3), and mutation of Glu-300 to Lys or Asn Glu-300 Lys 12 33 ± 3* 25 ± 4* 0.14 ± 0.03 would eliminate this interaction, which may be required for Glu-300 Gln 20 100 ± 8 20 ± 1.2* 0.19 ± 0.04 glucose binding. Consistent with this, the Lys and Asn Leu-309 -. Pro* 5 s0.98* 2.2 ± 1.9 0.13 ± 0.10 mutants had a lower affinity for glucose (Table 1). Several Glucokinase activity was determined as described (17, 18). The mutations such as Gly-261 -> Arg, Gly-299 -) Arg, and values shown are the mean ± SEM of three to five separate Leu-309 -- Pro are predicted to alter the structure of the expression experiments. The kinetic analysis was done with purified protein, and these mutations are associated with large de- native and mutant enzyme forms. However, the same changes in creases in V.. Ofthe missense mutations described in Table kinetic properties were observed after partial purification of the 1, all but two, Gly-175 -. Arg and Val-182 -* Met, involve enzyme by (NH4)2S04 fractionation. amino acids located in the active site cleft or in surface loops *V. and Km values that are significantly differentfrom native p-cell enzyme. leading into this cleft. Gly-175 and Val-182 map to a region tResidues located in the cleft near glucose. that is on the surface of the enzyme and far from the active tResidues whose mutation predicts a distortion in glucokinase struc- site. However, this region is part of the smaller of the two ture. domains of glucokinase and is predicted to participate in the No active enzyme. substrate-induced conformational change that results in cleft Downloaded by guest on September 28, 2021 Biochemistry: Gidh-Jain et al. Proc. Natl. Acad. Sci. USA 90 (1993) 1935 $ 1 5 10 15 20 25 30 35 40 45 Hugk MLDDRARMEAAKKEKVUQI LAEEOL KK MRM Q K I R L Yshkb M V H L G P K K POARKGSM DPEW.:T TETEO A TIPF I SK 1 5 10 1s 20 25 30 35 40 45 50 55 cx1 ae2

50 55 60 65 70 75 80 85 90 95 100 105 Hugk EEE K ELEAE T SiE VE F$ E G E E W VE KH Yshkb KG- -G NP. G WM D FETKSA L VLE- - -K R TF D TQ 60 65 70 75 80 85 90 95 100 105 (31 (2 (3 (4

115 110 120 125 130 135 140 145 150 160 155 Hugk 0 M fl..1 A flG TM - A j.$ C S D KH* EM K HK K V R DU H Yshkb SK R fl MR TON PSDE AMRWT aNADP-SK M E W?EOFP G S E PASO 115 110 120 125 130 135 140 145 150 155 165 160 a 3 (5 (36

165 170 175 180 185 190 195 200 205 210 215 Hugk K LN K AGA G N: G K G D FEG GFVVEAK A C*EI Yshkb E YshkbQ E_QR RD I .P.NDPI I gN ~~~~~~H.ND PM~EQQQK E T.~N.QKT...... HET . ....EVSE...... TO_TGMNH-I 170 175 180 185 190 195 200 205 210 215 220 (37 a4 (8 a5 t 220 225 230 235 240 245 250 255 260 265 Hugk ....EVMlMVCACAME MON. ELV - - E G E C Yshkb -E-TK.MEVUF V GAAY--V CSS.UK KLSi13IPPSAPaA-*C*Y*$*DNELGP 225 230 235 240 245 250 255 260 265 270 275 280 (39 (10 11

270 275 280 285 290 295 300 305 310 315 320 325 Hugk L L E RLV*SEA N Y GEK V;VINVDLRL,LLEGG A S E a L R T R Yshkb R TK T E R _ T T F: S SEY AD KD 285 290 295 300 305 310 315 320 325 330 335 a6 a 8

330 335 340 345 350 355 360 365 370 375 380 Hugk G A F RTVS SE - -G D R K I YN LSTLS F TT C D ACSVST Vshk FVMD S.YPA. EEEPFELET EFTKLF.L...... G...... SE 340 345 350 355 360 365 370 375 380 385 390 395 a9 aCoa0 11

385 390 395 400 405 410 415 420 425 430 Hugk ARLHEGV I S,MES.EDVMR T.G Y K L HIS_ F H A$Yl RLTP - SC Yshkb C sA- C AA STR YG A N Y GK T H L DY 400 405 410 415 420 425 430 435 440 445 (12 a12

435 440 445 450 455 460 465 Hugk EflTF[°EI-EsI-_.-. .-.., ..... E...... RV.EVECE AMG. Yshkb P KV PA QPI KSVAEGKA E G SVG GA 450 455 460 465 470 475 480 485 (313 a13 FIG. 3. Comparison ofthe sequences ofhuman ,-celi glucokinase and yeast hexokinase B: Structural alignment ofthe amino acid sequences of human ,-cell glucokinase (Hugk) and yeast hexokinase B (Yshkb) based upon placing insertions and deletions between secondary structural elements in the crystal structure ofyeast hexokinase B. The single-letter abbreviations for the amino acids are used. Identical residues are shown in black boxes, and similar residues are in shaded boxes. The 13 a-helices and ,-strands are indicated below the sequences. The small arrows indicate the positions at which introns interrupt the sequence. closure (23). Our results suggest that even relatively conser- exons, exons 5, 7, and 8 (Figs. 1, 3, and 4), which suggests vative substitutions in this surface loop region may affect this that these exons may encode functionally important domains conformational change. ofthe protein. Residues encoded by exons 7 and 8 form much of the surface of the cleft between the two domains, as well DISCUSSION as surface loops leading into this cleft. However, only one of Although mutations in the glucokinase gene have been iden- the mutations in this region is of a residue (Glu-256) directly tified in nine different exons, the missense mutations appear implicated in the binding ofglucose. The others are predicted to be clustered with 10 of these mutations occurring in three to be in the cleft (Thr-228, Gly-261), in a region leading to the Downloaded by guest on September 28, 2021 1936 Biochemistry: Gidh-Jain et al. Proc. Nati. Acad. Sci. USA 90 (1993) per se on (-cell glucose sensing and/or hepatic glucose metabolism (13). Glucokinase mutations may be the most common genetic cause of NIDDM identified to date, especially among young patients with NIDDM diagnosed before age 25. Studies ofthe effects of these mutations on glucokinase activity in vitro have led to the identification of key amino acid residues involved in determining glucokinase activity. The identifica- tion and functional characterization of further glucokinase mutations should lead to the identification of other key residues. These studies together with determination of the crystal structure of human (-cell glucokinase will provide new insight into the molecular mechanism responsible for the effi'cient transfer of phosphate from ATP to glucose and may lead to therapeutic approaches for treating NIDDM resulting from mutations in this important glycolytic enzyme. We thank Drs. J. S. Wainscoat and R. C. Turner for their encour- agement and support. This research was supported by the Howard Hughes Medical Institute, U.S. Public Health Service Grants DK- 20595, DK-38354, and DK44840, U.S. National Science Foundation Grant DBM-8608989, Juvenile Diabetes Foundation International, Association Francaise contre les Myopathies (through the Genethon program), Assistance Publique-Hopitaux de Paris, the French Min- istry for Research and Technology, the Swedish Medical Research Council (9109), and the Berth von Kantzow Foundation. M.S. was supported by a fellowship from the Deutsche Forschungsgemein- FIG. 4. Model for human glucokinase. A ribbon drawing of the schaft. a-carbon backbone of glucokinase, based on the yeast hexokinase B 1. Grossbard, L. & Schimke, R. T. (1966) J. Biol. Chem. 241, 3546- structure in the open conformation, is shown in blue with exons 5, 3560. 7, and 8 in green. The amino acids in which missense mutations have 2. Ureta, T. (1982) Comp. Biochem. Physiol. B 71, 549-555. been identified in human glucokinase are indicated in red and the 3. Meglasson, M. D. & Matschinsky, F. M. (1984)Am. J. Physiol. 246, glucose molecule is shown in green. The structure of the surface E1-E13. loops of glucokinase is expected to differ from that of yeast hexoki- 4. Weinhouse, S. (1976) Curr. Top. Cel. Regul. 11, 1-24. nase in the region of insertion and deletion as shown in Fig. 3. 5. Walker, D. G. (1990) in Essays in Biochemistry, eds. Campbell, P. N. & Greenville, G. D. (Academic, New York), Vol. 2, pp. cleft (Gly-261, Glu-279, Gly-299, Glu-300), or very far from 33-35. 6. Matschinsky, F. M. (1990) Diabetes 39, 647-652. the active site (Leu-309) and, as described above, may affect 7. Granner, D. K. & Pilkis, S. J. (1990) J. Biol. Chem. 265, 16012- the structure ofthe active site or the open-closed equilibrium 16016. of the enzyme. The mutations occurring in exon 5 are located 8. Magnuson, M. A. (1990) Diabetes 39, 523-527. in a surface loop preceding a-helices 3 and 4 (Fig. 3) in the 9. Newgard, C. B., Quaade, C., Hughes, S. D. & Milburn, J. L. (1990) Biochem. Soc. Trans. 18, 851-853. small domain, which, by analogy with the hexokinase B 10. Unger, R. H. (1991) Science 271, 1200-1205. structure (23), undergoes a conformational change on glucose 11. Vionnet, N., Stoffel, M., Takeda, J., Yasuda, K., Bell, G. I., binding, and mutations in this region may interfere with this Zouali, H., Lesage, S., Velho, G., Iris, F., Passa, Ph., Froguel, P. & Cohen, D. (1992) Nature (London) 356, 721-722. conformational change and thereby affect glucose binding 12. Stoffel, M., Patel, P., Lo, Y.-M. D., Hattersley, A. T., Lucassen, and/or catalysis. A. M., Page, R., Bell, G. I., Turner, R. C. & Wainscoat, J. S. We have previously suggested, based on the identification (1992) Nature Genet. 2, 153-156. of a nonsense mutation that was linked with early-onset 13. Stoffel, M., Froguel, P., Takeda, J., Zouali, H., Vionnet, N., Nishi, S., Weber, I. T., Harrison, R. W., PiLkis, S. J., Lesage, S., Vaxil- NIDDM (11), that mutations in the glucokinase gene lead to laire, M., Velho, G., Sun, F., Iris, F., Passa, Ph., Cohen, D. & Bell, the development of glucose intolerance by a gene-dosage G. I. (1992) Proc. Natl. Acad. Sci. USA 89, 7698-7702. mechanism. The demonstration of decreased enzymatic ac- 14. Miwa, I., Mitseyoma, S., Toyoda, Y., Murata, T. & Okuda, J. tivity of missense mutations in glucokinase that are associ- (1990) Prep. Biochem. 20, 163-178. 15. Nishi, S., Stoffel, M., Xiang, K., Shows, T. B., Bell, G. I. & ated with NIDDM is consistent with this hypothesis. De- Takeda, J. (1992) Diabetologia 35, 743-749. creased cellular levels of glucokinase activity may increase 16. Tanizawa, Y., Koranyi, L. I., Welling, C. M. & Permutt, M. A. the glucose threshold for insulin secretion in the , cell as well (1991) Proc. Natl. Acad. Sci. USA 88, 7294-7297. 17. Chein, C. T., Tauler, A., Lange, A. J., Chan, K., Printz, R. L., as impair hepatic uptake and metabolism ofglucose (3, 6, 7). El-Maghrabi, M. R., Granner, D. K. & Pilkis, S. J. (1989)Biochem. Since diabetic patients have one normal as well as one mutant Biophys. Res. Commun. 165, 817-824. glucokinase allele (11-13), they are not totally deficient but 18. Lange, A. J., Xu, L. Z., Van Poelwijk, F., Granner, D. K. & Pilkis, rather are predicted to have glucokinase levels that are about S. J. (1991) Biochem. J. 277, 159-163. 19. Andreone, T., Prince, R. L., Pilkis, S. J., Magnuson, M. A. & 50%6 of normal. However, this level of activity is apparently Granner, D. K. (1989) J. Biol. Chem. 264, 363-369. not sufficient to ensure normal (-cell and/or hepatocyte 20. Steitz, T. & Bennett, W. S. (1980) J. Mol. Biol. 140, 211-221. function since these individuals have readily detectable de- 21. Anderson, C. M., McDonald, R. C. & Steitz, T. A. (1978) J. Mol. fects in glucose metabolism that include mild fasting hyper- Biol. 123, 1-13. 22. Anderson, C. M., Zucker, F. H. & Steitz, T. A. (1979) Science 204, glycemia, impaired glucose tolerance, as well as overt 375-380. NIDDM (refs. 11-13; P.F., unpublished data). Interestingly, 23. Harrison, R. W. (1985) Ph.D. thesis (Yale Univ., New Haven, CT). a metabolic defect is evident even in individuals having a 24. Studier, F. W. & Moffat, B. A. (1986) J. Mol. Biol. 189, 113-130. mutation such as Glu-300 -- Gln, which has no 25. Quaade, C., Hughes, S. D., Coats, W., Sestak, A. & Newgard, apparent C. B. (1991) FEBS Lett. 280, 47-52. effect on V.. and only a 2- to 3-fold increase in the Km for 26. Fersht, A. R., Shi, J.-P., Knill-Jones, J., Lowe, D. M., Wilkinson, glucose (Table 1). This finding suggests that the product of A. J., Blow, D. M., Brick, P., Carter, P., Waye, M. M. Y. & the mutant allele may in some instances also have an effect Winter, G. (1985) Nature (London) 314, 235-238. Downloaded by guest on September 28, 2021