Evidence that glucokinase regulatory protein is ... - Wiley Online Library

Journal of Neurochemistry, 2002, 80, 45±53

Evidence that glucokinase regulatory protein is expressed and interacts with glucokinase in rat brain Elvira Alvarez, Isabel Roncero, Julie A. Chowen, Patricia VaÂzquez and Enrique BlaÂzquez Department of Biochemistry and Molecular Biology, Faculty of Medicine, Complutense University, Spain

Abstract Our previous description of functional glucokinase isoforms in the rat brain has opened new questions concerning the presence of glucokinase regulatory protein in the brain and the functional role of its interactions with glucokinase. In this study, we found glucokinase regulatory protein mRNA in rat brain, pancreatic islets and liver. In addition, we found two other variant splicing isoforms, both identi®ed in hypothalamus, pancreatic islets and liver. In situ hybridization studies revealed the presence of glucokinase regulatory protein mRNA, the highest number of positive cells being found in the paraventricular nucleus of the hypothalamus. Glucokinase regulatory protein gene expression gave rise to a protein of 69 kDa mainly in nuclear and soluble cell fractions. Glutathi-

one S-transferase protein fused either to rat liver or human pancreatic islet glucokinase were able to precipitate glucokinase regulatory protein from liver or hypothalamic extracts in the presence of fructose-6-phosphate, the amount of protein co-precipitated being decreased with fructose-1-phosphate. These ®ndings suggest that the presence of glucokinase and glucokinase regulatory protein in the rat brain would facilitate the adaptation of this organ to ¯uctuations in blood glucose concentrations, and both proteins may participate in glucosesensing and metabolic regulation in the central nervous system. Keywords: brain, glucokinase, glucokinase regulatory protein, interactions, rat. J. Neurochem. (2002) 80, 45±53.

Glucokinase is a member of the hexokinase family (ATP: D-hexose 6-phosphotransferase, EC 2.7.1.l) that catalyses the phosphorylation of glucose to glucose 6-phosphate. Hexokinases I, II and III have a high affinity for glucose, with low Km values in the micromolar range and a molecular size of 100 kDa. Glucokinase or hexokinase type IV has a low affinity for glucose; it is not inhibited by physiological concentrations of glucose 6-phosphate, and has a molecular 1 mass of about 50 kDa (Iynedjian 1993; Randle 1993; Pilkis et al. 1994). Glucokinase activity is expressed in the liver, the pancreatic islets of Langerhans, jejunal enterocytes and brain. Longterm regulation of glucokinase takes place at transcriptional and post-transcriptional level. The presence of tissue-specific promoters allows differential regulation. The upstream promoter is functional in b-cells and in the brain (Magnuson and Shelton 1989; Liang et al. 1991; Roncero et al. 2000), while the downstream promoter is used only in liver (Magnuson and Shelton 1989). Glucokinase levels in b-cells appear to be controlled by glucose, probably through a post-transcriptional 2 mechanism (Iynedjian 1993; Matschinsky et al. 1993; Pilkis et al. 1994). In contrast, insulin appears to be the major positive effector of glucokinase activity in the liver, insulin

increasing and cAMP decreasing gene transcription (Pilkis 3 and Granner 1992; Iynedjian 1993; Pilkis et al. 1994). The short-term regulation of glucokinase activity involves several mechanisms: (i) Long-chain fatty acyl-CoAs have been shown to be allosteric competitive inhibitors in vitro of the liver 4 enzyme (Tippet and Neet 1982) and human b-cell glucokinase (Veiga da Cunha et al. 1996; Moukil et al. 2000). (ii) Glucose increases the activity of glucokinase through a ÔmnemonicÕ mechanism, which increases the activity of liver glucokinase in the presence of high levels of glucose and decreases it when the glucose level is low (Cornish-Bowden and Stoer 1986). Received July 3, 2001; revised manuscript received Spetember 18, 2001; accepted September 22, 2001. Address correspondence and reprint requests to Elvira Alvarez, Departamento de BioquõÂmica y BiologõÂa Molecular, Facultad de Medicina, Universidad Complutense, 28040-Madrid, Spain. E-mail: [email protected] Abbreviations used: F1P, fructose 1-phosphate; F6P, fructose 6-phosphate; GK, glucokinase; GKRP, glucokinase regulatory protein.

Ó 2002 International Society for Neurochemistry, Journal of Neurochemistry, 80, 45±53

45

46 E. Alvarez et al.

(iii) Glucokinase activity in the liver may also be regulated by the presence of glucokinase regulatory protein (GKRP), which in the presence of fructose 6-phosphate binds to glucokinase and inhibits its activity 5 (Van Schaftingen 1989; Vandercammen and Van Schaftingen 1993), whereas fructose 1-phosphate prevents the formation of the complex. In pancreatic islets, glucokinase activity may also be regulated by a protein, in a similar way to that described in hepatocytes (Malaisse et al. 1990). GKRP was first identified in rat liver after the observation that fructose stimulates the phosphorylation of glucose in rat 6 hepatocytes (Van Schaftingen and Vandercammen 1989). This effect is mediated by fructose 1-phosphate, which releases the inhibition exerted by GKRP on glucokinase (Van Schaftingen et al. 1994). GKRP inhibits glucokinase competitively with glucose (Vandercammen and Van Schaftingen 1991), fructose 6-phosphate reinforcing and fructose 1-phosphate antagonizing the association of the enzymeinhibitor complex (Van Schaftingen et al. 1994). When glycogenolysis and/or gluconeogenesis are activated, the concentration of fructose 6-phosphate increases in the liver and produces the inhibition of glucokinase, which facilitates the release of glucose by the liver. When carbohydrates are present in the diet, fructose is phosphorylated to fructose-1 phosphate, which favours glucose utilization by the liver. In the liver, the subcellular translocation of glucokinase regulates the enzyme activity in accordance with the metabolic needs of the cells. Thus, GKRP is present mainly in the nucleus (Toyoda et al. 1995; De la Iglesia et al. 1999; Shiota et al. 1999). The locations of both proteins change as a function of the metabolic status. In the basal state, both glucokinase and GKRP are bound in the nucleus, but in the post-prandial state when glucose and fructose circulating levels rise, glucokinase is released from GKRP and remains free in the cytoplasm ready to phosphorylate glucose (Agius and Peak 1993; Agius et al. 1995; Toyoda et al. 1995; De la Iglesia et al. 1999; Shiota et al. 1999). These findings indicate that GKRP functions as an allosteric inhibitor of glucokinase and as a metabolic sensor, as well as a nuclear chaperone that binds and transports glucokinase to the nucleus. Studies carried out in GKRP-deficient mice suggest a regulatory and a stabilizing role for this protein in maintaining adequate glucokinase levels in the liver (Farrelly et al. 1999; Grimsby et al. 2000). Tissue-specific differences between glucokinase regulation in liver and pancreatic b-cells have been observed. Recent studies have provided evidence for the existence of a protein factor with a glucokinase regulatory function in insulinsecreting cells and with characteristics different from liver GKRP (Tiedge et al. 1999). Our recent report describing the presence of functional glucokinase isoforms in rat brain (Roncero et al. 2000) opens

new avenues concerning the existence of a brain GKRP and its functional interaction with glucokinase. Accordingly, we used RT-PCR and in situ hybridization histochemistry procedures to check the presence of GKRP mRNA transcripts in rat brain and pancreatic islets. Here we also present experimental evidence of the GKRP subcellular distribution in hypothalamus and of its functional interactions with glucokinase.

Experimental procedures Experimental animals Male Wistar rats weighing 200±250 g were fed ad libitum with a standard pellet diet and housed at a constant temperature (21°C) on a 12-h light±dark cycle with lights on at 08.00 h. Rats were killed by decapitation, and the brain and liver were rapidly removed, frozen in liquid nitrogen, and then stored at )80°C. All procedures were carried out according to the European Union ethical regulations for animal research. RNA isolation Total cellular RNA was isolated by the acid±guanidium isothiocyanate method (Chomczynski and Sacchi 1987). cDNA synthesis, PCR ampli®cations and Southern blot analysis Sequence-speci®c primers were designed to amplify GKRP mRNA. The antisense priming oligonucleotide GKR-2 (5¢-TCACCTTTTCCTTCTCGTGG-3¢) corresponding to nucleotide bases 1712±1731, GKR-3 (5¢-GTTAAACATGTCACTGTGGTC-3¢) corresponding to nucleotides 1117±1140, GKR-6 (5¢-CACAGAGATGCCTATGACGACC-3¢) corresponding to nucleotides 540±561 and upstream oligonucletide GKR-1 (5¢-TTGGTGACCACAGTGACATG-3¢) corresponding to nucleotides 1115±1134, GKR-4 (5¢-CATGCCAGGCACCAAACGATATC-3¢) corresponding to nucleotides 20±43 and GKR-5 (5¢-GTCGTCATAGGCATCTCTGTGG-3¢) corresponding to nucleotides 541±562 of rat liver GKRP (Detheux et al. 1993) were used. RT-PCR ampli®cation was carried out using ÔThe TitanTM One tube RT-PCR SystemÕ (Roche Molecular Biochemicals, Barcelona, Spain). First-strand cDNA was synthesized at 53°C for 1 h, and ampli®cation of GKRP cDNA was performed at an annealing temperature of 65°C and an extension temperature of 68°C for 30 cycles. The ampli®ed DNA using oligonucleotide pairs GKR-1/GKR-2, GKR-4/GKR-6 and GKR-5/GKR-7 were ligated into PCRII vector (Invitrogen Co., San Diego, CA, USA). They were sequenced on one strand using an automated LKB ALF DNA sequencer (Pharmacia, Barcelona, Spain). RT-PCR ampli®cation products using oligonucleotide pairs (GKR-1/GKR-2) and (GKR-5/GKR-7) were size-fractionated in 5% acrylamide gel and transferred to a nylon membrane. Blots were probed under high stringency with a fragment corresponding to nucleotides 1115±1731 or 20±1140 of rat liver GKRP labelled with digoxigenin by random priming using the DIG DNA labelling kit (Roche Molecular Biochemicals, Barcelona, Spain).


Glucokinase regulatory protein in rat brain 47

RT-PCR ampli®cation was also carried out using oligonucleotide pairs GKR-7 (5¢ CTGGTTGGTGAGTTCATCCTTC 3¢) corresponding to nucleotide bases 1141±1164 and GKR-8 (5¢ GTGTTGTTGCCTCCTTAGCC3¢) that recognizes a sequence between nucleotides 771±772 (Fig. 1) hitherto undescribed (Detheux et al. 1993).

In situ hybridization histochemistry In situ hybridization histochemistry was performed as previously described (Chowen et al. 1993). Antisense or sense 33P-labelled cRNA probes were generated with SP6 or T7 polymerase in pCRIIGKRP, using a standard transcription reaction containing 10 lM [33P]UTP. This resulted in probes with speci®c activities of approximately 1.8 ´ 109 dpm/lg. Probes were hydrolysed in bicarbonate buffer to an average length of 150 bases. The sense cRNA probe was used as a speci®city control and under identical conditions showed no detectable labelling. The slides were dipped in LM-1 photographic emulsion and exposed for 3 weeks, after which they were developed and coverslipped. Isolation of subcellular fractions from hypothalamus and liver Immediately after the rats had been killed, the brains and livers were promptly removed. The brain was dissected using the stereotaxic atlas of Paxinos and Watson (1986) as reference to isolate the hypothalami. Subcelular fractions were isolated as described by Dignam et al. (1983). Both hypothalamus and liver were homogenized in ice-cold medium [10 mM HEPES, 1.5 mM MgCI2, 10 mM KCI (pH 7.9) supplemented with 0.5 mM dithiothreitol DTT), 1 mM phenylmethylsulfonyl ¯uoride (PMSF) and 10 lM leupeptin] using a glass Dounce homogenizer. Homogenates were centrifuged at 4°C for 10 min at 500 g. The resulting supernatant was used for subsequent separation of membrane and soluble fractions, while the pellets thus obtained were further centrifuged at 25 000 g for 20 min at 4°C. Thereafter, the supernatants were removed (residual cytoplasmic components) and the pellets (nuclei) were resuspended in 20 mM HEPES, 25% glycerol, 1.5 mM MgCl2, 0.42 M NaCl, 0.2 mM EDTA (pH 7.9) supplemented with 0.5 mM DTT, 1 mM PMSF and 10 lM leupeptin, and then homogenized in a glass Dounce homogenizer; the resulting suspension was stirred for 30 min and centrifuged at 25 000 g for 30 min. The supernatant (nuclear extracts) was dialysed against 50 volumes of 20 mM HEPES buffer, 20% glycerol, 0.1 M KCI, 0.2 mM EDTA (pH 7.9) supplemented with 0.5 mM DTT and 1 mM PMSF overnight at 4°C (nuclear fraction).

Fig. 1 Schematic representation of the different putative glucokinase regulatory protein cDNAs. Position of the oligonucleotide primers used for PCR ampli®cation of the hypothalamic and pancreatic islet glucokinase regulatory protein cDNAs.

The supernatant from the low-speed 500-g centrifugation was mixed with 0.11 volumes of 0.3 M HEPES buffer 0.03 M MgCI2, 1.4 M KCI (pH 7.9) supplemented with 0.5 mM DTT, 1 mM PMSF and 10 lM leupeptin and centrifuged at 100 000 g for 1 h. The resulting supernatant from the last step was dialysed against 50 volumes of 20 mM HEPES buffer, 20% glycerol, 0.1 M KCI, 0.2 mM EDTA (pH 7.9) supplemented with 0.5 mM DTT and 1 mM PMSF overnight at 4°C (soluble fraction). The pellet was resuspended in 20 mM HEPES buffer, 20% glycerol, 0. 1 M KCI, 0.2 mM EDTA (pH 7.9) containing 0.5 mM DTT and 1 mM PMSF (membrane fraction). Detection of proteins by western blot Soluble, nuclear or membrane fractions from liver (20 lg) or hypothalamus (200 lg) were resolved by electrophoresis through a sodium dodecyl sulfate (SDS)-polyacrylamide gel and electrotransferred onto a nitrocellulose ®lter. After blocking overnight in Trisbuffered saline [TBS; 20 mM Tris, 159 mM NaCI (pH 7.4)] containing 0.2% NP-40, 5% non-fat dry milk at 4°C, the ®lters were incubated with a polyclonal sheep antiserum (1 : 2000) against glutathione S-transferase glucokinase (GST-GK) fusion protein (a generous gift from Dr M. A. Magnuson, Tennessee, USA) for 1 h at room temperature. After washing off the excess antibody, the ®lters were then reblocked in TBS with 5% non-fat dry milk and 0.2% NP-40 for 1 h at room temperature and incubated with an anti-sheep IgG conjugated to horseradish peroxidase for 1 h at room temperature. Chemiluminescence detection was carried out in the presence of ECL reagents (Amersham Pharmacia Biotech, Barcelona, Spain). The identi®cation of GKRP was determined by western blot using a rabbit polyclonal antibody against rat liver GKRP (a generous gift from Dr E. Van Schaftingen, Bruxelles, Belgium). After blocking in TBS [20 mM Tris, 150 mM NaCI (pH 7.4)], 0.1% Tween-20, 5% non-fat dry milk for 1 h at room temperature, the membranes were incubated at 4°C for 16 h with a polyclonal antibody against GKRP (1: 1000) in TBS, 1% bovine serum albumin (BSA) and 0.1% Tween-20. Then, membranes were incubated at room temperature for 1 h with horseradish peroxidasecoupled goat anti-rabbit Ig G (Upstate Biotechnology, Lake Placid, NY, USA) 1 : 3000 in TBS, 1% BSA, 0.1% Tween-20. Chemiluminescence detection was carried out in the presence of ECL (Amersham Pharmacia Biotech, Barcelona, Spain). Plasmid construction of GST fusion vectors and puri®cation of proteins expressed in bacteria Glutathione S-transferase (GST) fusion proteins were constructed in the vector pGEX-5X-2 (Amersham Pharmacia Biotech). The coding region of both human pancreatic beta-cell glucokinase and rat liver glucokinase were ampli®ed, using ÔThe TitanTM One tube RT-PCR SystemÕ from Roche Molecular Biochemicals, with oligonucleotide primers (5¢-GATGCTGGACGACAGAGCCAG-3¢) and (5¢-CTCACTGGCCCAGCATACAGG-3¢), using total RNA from human pancreatic islet template and (5¢-GATGGCTATGGATACTACAAG-3¢) and (5¢-TCACTGGGCCAGCATGCAAGC-3¢), using rat liver total RNA template. The ampli®ed DNA was cloned in the PGEMT vector (Promega, Madison, WI, USA). The glucokinase cDNA fragments EcoRI-EcoRI were then ligated into pGEX-5X-2 (pGEX5X hGKi) and (pGEX-5X rGKI). Escherichia coli BL21 strain were transformed with the indicated plasmids and grown at 27°C to mid-log phase. The expression of fusion proteins was induced



adding isopropyl-b-D-thiogalactoside (IPTG) at a ®nal concentration of 0.2 mM and the cultures were incubated in a orbital shaker for 3 h at 27°C. Collection of cell extracts and puri®cation of fusion proteins was carried out according to the manufacturer's instructions. Cells were harvested by centrifugation and washed with phosphate-buffered saline (PBS; 140 mM NaCI, 2.7 mM KCI, 10 mM Na2HP04, 1.8 mM KH2PO4, pH 7.4) supplemented with 10 mM EDTA, 1 mM PMSF. Lysozyme 100 lg/mL was added and the cells were lysed by mild sonication. Triton X-100 2% was then added, mixed gently for 30 min, and centrifuged at 12 000 g for 10 min at 4°C. Supernatants were incubated with 0.5 mL of glutagarose bead slurry equilibrated with PBS. After 2 h incubation with gentle mixing at room temperature, they were centrifuged at 500 g for 5 min. The pellets were washed twice with PBS and ®nally the fusion proteins were eluted with 10 mM glutathione in 100 mM Tris, 120 mM NaCI, pH 8.0, and designated GST-hGKi (human pancreatic beta cell glucokinase) and GST-rGKl (rat liver glucokinase), respectively. Co-precipitation assays GST-hGKi or GST-rGKI (20±30 lg) were incubated in 25 mM HEPES, pH 7.2, 1 mM MgCI2, 1 mM DTT, 5 mM glucose containing 0.5 mM F1P or F6P and either liver or hypothalamus extracts. After 20 min with gentle mixing at room temperature, 0.03 mL of glut-agarose bead slurry were added and mixed for 20 min. Beads were pelleted in a microfuge and washed twice with 1 mL of 25 mM HEPES, 1 mM MgCI2, pH 7.2, 1 mM DTT, and 5 mM glucose containing 0.5 mM F1P or F6P. Controls without tissue extracts were included in each experiment. Proteins retained in the glut-agarose beads were processed by SDS-PAGE and electrotransferred onto nitrocellulose membranes. The identi®cation of GKRP was determined by western blotting.

Results

PCR ampli®cations and Southern blot analysis To verify the presence in rat brain and pancreatic islets of GKRP mRNAs similar to that expressed in liver, first-strand cDNA was prepared using total RNA from hypothalamus,

pancreatic islets and liver. Subsequent PCR amplification using oligonucleotide pairs GKR-1/GKR-2, GKR-4/GKR-6 and GKR-7/GKR-5 (see Fig. 1 for location of primers) revealed products of the predicted size in hypothalamus, pancreatic islets and liver. Southern blot analysis of the PCR products (Fig. 2) confirmed the presence of GKRP mRNAs in rat hypothalamus, pancreatic islets and liver. PCR products amplified from hypothalamus and pancreatic islets were cloned and sequenced on one strand. The sequence of these cDNAs was compared with known rat liver glucokinase regulatory protein sequence (Detheux et al. 1993) and some differences were identified. In pancreatic islets, at position 127 (Detheux et al. 1993) we found a thymine instead of cytosine in three different clones that would change Pro36 to Ser36. A clone from hypothalamus was found to contain an insertion of 17 bp between positions 771 and 772 (Detheux et al. 1993). Recently, the organization of the human glucokinase regulatory protein gene has been reported (Hayward et al. 1998) and, taking into account the information from both rat liver glucokinase regulatory protein and human glucokinase regulatory protein gene sequences, it is known that the above insertion corresponds to an alternative splicing in the junction between exons 9 and 10, which alters the reading frame and would produce a protein of 265 amino acids (Fig. 3a). To investigate the presence of such an isoform in pancreatic islets and liver, we designed a primer that hybridizes in that sequence (GKR-8). RT-PCR using GKR-7/GKR-8 amplified a fragment of 412 bp in hypothalamus, pancreatic islets and liver. Cloning and sequencing of those PCR products confirmed that mRNAs carrying those 17 bp were also present in pancreatic islets and liver. Southern blot analysis of the PCR products is shown in Fig. 2. In pancreatic islets, we detected a new cDNA with an insertion of 67 bp between positions 1593 and 1594 (Detheux et al. 1993) corresponding to an alternative splicing in the junction of exons 17 and 18 (Fig. 3b), which predicts a 523 amino acid protein. Southern blot analysis of

Fig. 2 RT-PCR analysis of glucokinase regulatory protein mRNA transcripts in hypothalamus, pancreatic islets and liver. Total RNA from hypothalamus, pancreatic islets and liver was reverse-transcribed and ampli®ed by PCR. Southern blot analysis of the ampli®ed DNA using three different oligonucleotide pairs: GKR-5/GKR-7, GKR-1/GKR-2 and GKR-8/GKR-7. L.E., longer exposure of the Southern blot RT-PCR products ampli®ed with pair GKR-1/ GKR-2.



Fig. 3 Use of an alternate splicing in glucokinase regulatory protein cDNA. Clones were obtained by PCR ampli®cation of ®rst-strand cDNA. The sequences were compared with that of known rat glucokinase regulatory protein from liver reported by Detheux et al. (1993). Two alternate splicing events were identi®ed: (a) an alternate splicing in the junction of exons 9 and 10 contains an insertion of 17 bp and (b) an alternate splice site in the junction of exons 17 and 18, containing an insertion of 67 bp. The top lines in (a and b) (Detheux et al. (1993) show the nucleotide and predicted amino acid sequences of the cDNA

isolated from rat liver, in the junctions of exons 9 and 10 (a, 1) and exons 17 and 18 (b, 1). The lower lines in (a and b) show the nucleotide sequence of the cDNA from hypothalamus or pancreatic islets in the junctions of exons 9 and 10 (a, 2) and exons 17 and 18 (b, 2). The underlined sequences are as reported by Detheux et al. (1993). The frame-shift introduced by use of the alternative splicing leads to early termination as show the predicted amino acid sequence of exons 9 and 10 (a, 2) and exons 17 and 18 (b, 2).

PCR products amplified with the oligonucleotide pair GKR2/GKR-1 showed that the isoform carrying an insertion of 67 bp in the junction of exons 17 and 18 was also present as a minor isoform in hypothalamus and liver but it was a major form in pancreatic islets (Fig. 2).

GST-GK/GKRP co-precipitation assay Based in previous observations related to the interactions of glucokinase and GKRP induced by F6P and the inhibition of the formation of this complex by F1P in rat liver, experiments were designed to analyse the interactions of glucokinase and GKRP from liver and hypothalamus and to investigate whether the interactions between GKRP with liver or pancreatic islet glucokinases were similar in the presence of fructose phosphate esters. The GST fusion gene with rat liver glucokinase (pGEX5X-rGKI) or human islet glucokinase (pGEX-5X-hGKi) were expressed in E. coli and produced GST-rGKI 75 kDa and GST-hGKi 76 kDa purified proteins. A Coomassie bluestained gel is shown in Fig. 6(a). Both proteins were able to phosphorylate glucose (data not shown). Co-precipitation of GKRP was performed as described in Experimental procedures. Both GST-hGKi and GST-rGKI were able to precipitate GKRP from liver and hypothalamus in the presence of F6P, but the amount of GKRP co-precipitated decreased in the presence of F1P (Fig. 6b).

In situ hybridization histochemistry Labelled cells were seen throughout the brain but mainly in the paraventricular nucleus, using the antisense probe (Fig. 4), whereas no labelling was found with the sense probe (data not shown). Subcellular localization of GK and GKRP proteins in liver and hypothalamic cells To investigate whether GKRP mRNAs are translated in the hypothalamus as well as the subcellular localization of GKRP and glucokinase, protein blots containing soluble, nuclear or membrane protein fractions from hypothalamus and liver were incubated with polyclonal antibodies against GST-GK fusion protein or GKRP. A protein of 52 kDa was detected with anti-GST-GK antibody in liver and hypothalamus (Fig. 5). A band of 110 kDa was also detected, mainly in hypothalamus membrane and soluble fractions. When the anti-GKRP antibody was used, a protein of 69 kDa was detected in liver soluble and nuclei fractions, while in hypothalamus it was found mainly in nuclear and soluble fractions (Fig. 5).

Discussion

Several hypothalamic areas are involved in metabolic regulation and in the control of feeding behaviour and some of the neurones located in such nuclei are glucose-responsive and



Fig. 4 In situ hybridization histochemistry of GKRP mRNA. (a) Dark ®eld photomicrograph of labelled cells of the third ventricle wall and paraventricular nucleus. Cluster of white grains represents cells positive for GKRP mRNA. (b) Light-®eld photomicrograph in the third ventricle and paraventricular hypothalamic nucleus. Black grains represent cells positive for GKRP mRNA.

sensitive to the changes in glucose concentrations in the extracellular space (Yang et al. 1999). High Km glucose phophorylating activity has been described in human cerebral cortex (Bachelard 1967). Also, new questions about its functional properties and interactions with GKRP are raised by the identification of glucokinase mRNA in certain neural

cells (Liang et al. 1991; Navarro et al. 1996), as well as our own recent report describing that the mRNA that codes for the major pancreatic islet isoform (B1) is also the most abundant in rat hypothalamus, and that the activity of the glucokinase in this area was 25±40% of the total glucose phosphorylating activity (Roncero et al. 2000). GKRP was first identified in parenchymal liver cells and seems to function as a metabolic sensor that binds or releases glucokinase and then inhibits or activates glucokinase, respectively, depending on the metabolic requirements of the cell. Here, we have developed an experimental design with which, using several different technical approaches, we found the expression of GKRP mRNA and protein in rat brain and its interaction with glucokinase. Thus, in situ hybridization histochemistry studies revealed GKRP mRNAs in cells throughout the brain, with the highest number of positive cells found in the paraventricular hypothalamic area, in the same locations where glucokinase mRNAs were found. Also, RT-PCR analysis of total RNA from hypothalamus, pancreatic islets and liver and subsequent Southern blot analysis of the PCR products disclosed the presence of GKRP mRNAs in all three tissues. Sequencing of the amplified GKRP cDNA from hypothalamus and pancreatic islets indicated that one of the mRNA transcripts was identical to the cDNA of GKRP cloned from liver (Detheux et al. 1993). The presence at position 127 of a thymine instead of a cytosine in pancreatic islet cDNA, modifying Pro36 to Serine in the amino acid sequence, might be introduced during PCR amplification. We also found two other variant splicing isoforms. One of them was in the junction between exons 9 and 10, due to the insertion of 17 bp, which would change the reading frame and give rise to a 265 amino acid protein, and was identified in hypothalamus, pancreatic islets and liver. A second cDNA with an insertion of 67 bp in the junction of exons 17 and 18 was detected as a major isoform in pancreatic islets but as a minor isoform in hypothalamus and liver. In previous studies, the expression of GKRP in liver was reported ± with no presence of GKRP mRNA (Detheux et al.

Fig. 5 Immunoblot detection of glucokinase and glucokinase regulatory protein. Nuclear (EN), soluble (S100) and membrane (Memb) fractions from liver (20 lg) and hypothalamus (200 lg) were fractionated on a 12% SDS-polyacrylamide gel. Glucokinase or GKRP immunoreactivities were detected in blots with an anti-GST-GK or against rat liver GKRP antibodies as described in Experimental procedures. Size standards are shown in kDa. The position of immunoreactive glucokinase (GK), low-Km hexokinases (HK) and glucokinase regulatory protein (GKRP) being indicated by the arrow.



Fig. 6 Co-precipitation of GST-glucokinase and glucokinase regulatory protein from liver and hypothalamus. GST fusion proteins were incubated with liver or hypothalamus extracts. The co-precipitated proteins were analysed by polyacrylamide gel electrophoresis and immunoblot as described in Experimental procedures. (a) Puri®ed GST-fusion proteins containing human islet glucokinase GST-hGKi or rat liver glucokinase GST-rGKI were processed by SDS-PAGE and visualized by Coomassie blue staining. Size standards are shown on the left side in kDa. (b) Immunoblot detection of GKRP in proteins co-precipitated by GST-rGKl (lines 1, 2 and 3) and GST-hGKi (lines 4, 5, 6, 7, 8, 9 and 10) from extracts of liver (lines 1, 2, 3, 4, 5 and 6) and hypothalamus (lines 8, 9 and 10) in the absence (C) (lines 1, 4 and 8) or presence of fructose 1-phosphate (F1P) (lines 2, 5 and 9), fructose 6-phosphate (F6P) (lines 3, 6 and 10). Line 7 shows a control that was processed without tissue extract (b).

1993; Tiedge et al. 1999) or protein (Vandercammen and Van Schaftingen 1993) ± in extrahepatic tissues, probably because the GKRP mRNA content is lower in the central nervous system and is not uniformly distributed in all brain areas, as confirmed by our in situ hybridization studies. Later on, reports related to GKRP in pancreatic islets were contradictory. In 1990, Malaisse et al. (1990) suggested that glucokinase in pancreatic islets would be regulated by a protein sensitive to fructose esters, similar to what occurs in the liver, whereas Tiedge et al. (1999) reported evidence for the presence in pancreatic islets of a glucokinase regulatory protein different from liver GKRP. In addition, the results of 7 Hayward et al. (1998), who compared the size of products obtained by PCR from cDNAs from pancreatic islets and liver using primers in exons 1±14, indicated no evidence for alternative splicing isoforms in either tissue. Our data

indicate that the most abundant GKRP mRNA transcript in pancreatic islets corresponds to an isoform with an alternative splicing in the junction of exons 17 and 18, which would produce a protein of 523 amino acids. Next, a further step was to search whether GKRP interacts functionally with glucokinase in the central nervous system, as happens in the liver. For this purpose, we studied the effects of fructose esters on the behaviour of the putative glucokinase-GKRP (GK-GKRP) complexes in liver and hypothalamus. In previous studies (Roncero et al. 2000), using glycerol density gradients to characterize hypothalamic glucokinase, we found that the profiles of both glucokinase activity and immunoreactivity were distributed in two peaks. That profile could be explained if glucokinase was present as a individual (52 kDa) component and also forming a GK-GKRP complex (120 kDa). In liver extracts the presence of fructose-6-phosphate increased the putative peak of GK-GKRP complexes, whereas fructose-1-phosphate decreased the formation of the complexes (data not shown), confirming previous data by others (Vandercammen 8 and Van Schaftingen 1990; Van Schaftingen et al. 1994). When these experiments were performed using hypothalamus extracts, the sedimentation profiles were slightly different in the presence or absence of fructose-6-phosphate or fructose-1-phosphate (data not shown). Several explanations could be advanced to account for these results: the amounts of glucokinase and GKRP are much lower in the hypothalamus than in the liver and it is possible that the effect of fructose esters can only be detected with high concentrations of these proteins. It is also possible that the behaviour of brain proteins in glycerol gradients is different and may be that brain GK-GKRP complexes might not respond to fructose esters. Further experimental approaches were developed to elucidate the functional interactions between glucokinase and GKRP as well as whether pancreatic islet and liver glucokinase interactions with GKRP are modulated in the same way by fructose esters. For this purpose we used a glutathione S-transferase protein fused either to rat liver glucokinase (GST-rGKl) or to human pancreatic islet glucokinase (GST-hGKi). Both proteins were incubated with liver or hypothalamic extracts, and then the fused proteins ±bound or not to GKRP ± were isolated with glutagarose. Both GST-rGKl and GST-hGKi were able to precipitate GKRP from liver and hypothalamus in the presence of fructose-6-phosphate, but the amount of GKRP coprecipitated was decreased in the presence of fructose1-phosphate. Our results are in agreement with the data reported by Mookhtiar et al. (1996) using human hepatic recombinant glucokinase and GKRP, and indicate that glucokinase and GKRP also interact in rat brain and may respond to fructose esters. GKRP gene expression in rat brain gave rise to a protein of 69 kDa, as determined by western blotting, with



a characteristic subcellular distribution. Also, the presence of glucokinase and GKRP in similar subcellular fractions from liver and brain suggests that both proteins may have the same biological effects in brain as in liver. Analysis of the presence of glucokinase and GKRP by western blot in subcellular fractions from hypothalamus showed that the glucokinase was mainly located in soluble fractions and that GKRP was present in the nuclear and soluble fractions. In hepatocytes, GKRP is mainly located in the nucleus (Brown et al. 1997), while glucokinase moves between nucleus and cytoplasm, depending on the requirements of cellular metabolism. In this intracellular glucokinase movement, nuclear import of the enzyme depends upon GKRP, whereas export is due to a nuclear export signal sequence present in glucokinase. Thus, it has been proposed that GKRP seems to function as both a nuclear chaperone and a metabolic sensor. In our experiments, a high percentage of the GKRP was found in the soluble fractions, which must be related to a transitory location of the protein in the cytoplasm as a consequence of its chaperone activity. Additionally, the possibility that the GKRP present in the soluble fractions might be the result, at least in part, of contamination of the cell fractions cannot be ruled out. Our studies identified GKRP mRNA transcripts and protein in hypothalamus and pancreatic islets. Glucokinase may play an important role in both tissues. In pancreatic islets, where glucokinase acts as a glucose sensor, it modulates glucosedependent insulin secretion, while in the hypothalamus, glucose-responsive neurones may sense glucose using a mechanism similar to that of pancreatic b-cells (Yang et al. 1999). In both tissues the interactions between glucokinase and GKRP could be necessary for the above effects to be exerted. Glucose metabolism in the central nervous system deserves special interest because normal functioning of the brain requires a constant supply of glucose, as well as defense mechanisms against hypoglycaemia and other mechanisms able to recognize fluctuations in circulating glucose levels. However, the kinetic properties of the most abundant glucose phosphorylating enzymes of the brain, the low Km hexokinases, are saturated at 1 mM glucose and are unable to detect changes at higher physiological glucose concentrations. By contrast, glucokinase may recognize fluctuations in plasma glucose levels, and its interactions with GKRP may facilitate the functioning of glucose-sensing sites located in selected areas of the brain. Thus, interactions between glucokinase and GKRP could regulate key mechanisms in the brain. In addition, the coexpression of glucokinase and GKRP with GLUT-2 in the same hypothalamic neurones, as previously reported by us (Alvarez et al. 1996; Navarro et al. 1996) adds a further suggestion concerning the cooperative actions of these molecules in glucose-sensing and metabolic regulation in the central nervous system. Further experiments to confirm the physiological role of GKRP in the hypothalamus and pancreatic islets may be of great interest.

Acknowledgements The authors are indebted to Dr M. A. Magnuson for the gift of GK antiserun, as well as Dr E. Van Schaftingen and Dr M. Veiga da Cunha for the gift of GKRP antiserum. This work was supported by grants from the DireccioÂn General de InvestigacioÂn CientõÂfica y TeÂcnica, the Fondo de InvestigacioÂn Sanitaria, and the Comunidad de Madrid, Spain.

References Agius L. and Peak M. (1993) lntracellular binding of glucokinase in hepatocytes and transolocation by glucose, fructose and insulin. Biochem. J. 296, 785±796. Agius L., Peak M. and Van Schaftingen E. (1995) The regulatory protein of glucokinase binds to the hepatocyte matrix, but, unlike glucokinase, does not translocate during substrate stimulation. Biochem. J. 399, 711±713. Alvarez E., Roncero I., Chowen J. A., Thorens B. and BlaÂzquez E. (1996) Expression of the glucagon-like peptide-1 receptor gene in rat brain. J. Neurochem. 66, 920±927. Bachelard H. S. (1967) Glucose-phosphorylating enzyme with high Km in human brain. Nature 215, 959±960. Brown K. S., Kalinowski S. S., Megill J. R., Durham S. K. and Mookhtiar K. A. (1997) Glucokinase regulatory protein may interact with glucokinase in the hepatocyte nucleus. Diabetes 46, 179±186. Chomczynski P. and Sacchi N. (1987) Single-step method of RNA isolation by acid guanidinium thiocyanate±phenol±chloroform extraction. Anal. Chem. 162, 156±159. Chowen J. A., Argente J., Gonzalez-Parra S. and Garcia Segura L. M. (1993) Differential effects of the neonatal and adult sex steroid environments on the organization and activation of hypothalamic growth hormone-releasing hormone and somatostatin neurons. Endocrinology 133, 2792±2802. Cornish-Bowden A. and Stoer A. C. (1986) Mechanistic origin of the sigmoidal rate behaviour of rat liver hexokinase D (ÔglucokinaseÕ). Biochem. J. 240, 293±296. De la Iglesia N., Veiga da Cunha M., Van Schaftingen E., Guinovart J. J. and Ferrer J. C. (1999) Glucokinase regulatory protein is essential for the proper subcellular localization of liver glucokinase. FEBS Lett. 456, 332±338. Detheux M., Vandekerckhove J. and Van Shaftingen E. (1993) Cloning and sequencing of rat liver cDNAs encoding the regulatory protein of glucokinase. FEBS. Lett. 321, 111±115. Dignam J. D., Lebovitz R. M. and Roeder R. G. (1983) Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei. Nucleic Acids Res. 11, 1475±1489. Farrelly D., Brown K. S., Tieman A., Ren J., Lira S. A., Hagan D., Gregg R. and Mookhtiar K. A. (1999) Mice mutant for glucokinase regulatory protein exhibit decreased liver glucokinase: a sequestration mechanism in metabolic regulation. Proc. Natl. Acad. Sci. USA 96, 14511±14516. Grimsby J., Coffey J. W., Dvorozniak M. T., Magram J., Li G., Matschinsky F. M., Shiota C., Kaur S., Magnuson M. A. and Grippo J. F. (2000) Characterization of glucokinase regulatory protein-de®cient mice. J. Biol. Chem. 275, 7826±7831. Hayward B. E., Dunlop N., Intody S., Leek J. P., Markham A. T., Warner J. P. and Bothron D. T. (1998) Organization of the human glucokinase regulatory gene. Genomics 49, 137±142. Iynedjian P. B. (1993) Mammalian glucokinase and its gene. Biochem. J. 293, 1±13. Liang Y., Jetton T. L., Zimmerman E. C., Naja® H., Matschinsky F. M. and Magnuson M. A. (1991) Effects of alternate RNA splicing on



glucokinase isoform activities in the pancreatic-islet, liver and pituitary. J. Biol. Chem. 266, 6999±7007. Magnuson M. A. and Shelton K. D. (1989) An alternate promoter in the glucokinase gene is active in the pancreatic b cell. J. Biol. Chem. 10 264, 15936±15942. Magnuson M. A., Andreone T. L., Printz R. L., Koch S. and Granner D. K. (1989) Rat glucokinase gene: structure and regulation by insulin. Proc. Natl. Acad. Sci. USA 86, 4838±4842. Malaisse W. J., Malaisse-Lagae F., Davies D. R., Vandercammen A. and Van Schaftingen E. (1990) Regulation of glucokinase by a fructose 1-phosphate-sensitive protein in pancreatic islets. Eur. J. Biochem. 190, 539±545. Matschinsky F., Liang I., Kesavan P., Wang L., Froguel P., Velho G., Cohen D., Permutt M. A., Tanizawa Y., Jetton T. L., Niswender K. and Magnuson M. A. (1993) Glucokinase as pancreatic b-cell glucose sensor and diabetes gene. J. Clin. Invest. 92, 2092±2098. Mookhtiar K. A., Kalinowski S. S., Brown K. S., Tsay Y. H., SmithMonroy C. and Robinson G. W. (1996) Heterologous expression and characterization of rat liver glucokinase regulatory protein. Diabetes 45, 1670±1677. Moukil M. A. , Veiga da Cunha M. and Van Schaftingen (2000) Study of the regulatory properties of glucokinase by site-directed mutagenesis. Diabetes 49, 195±201. Navarro M., Rodriguez de Fonseca F., Alvarez E., Chowen J. A., Zueco J. A., GoÂmez R., Eng J. and BlaÂzquez E. (1996) Colocalization of glucagon-like peptide-1 (GLP-1) receptors, glucose tansporter GLUT-2 and glucokinase mRNAs in rat hypothalamic cells: evidence for a role of GLP-1 receptor agonists as an inhibitory signal for food and water intake. J. Neurochem. 67, 1982±1991. Paxinos G. and Watson C. (1986). The Rat Brain in Stereotaxic Coordinates, 2nd edn. New York: Academic Press. Pilkis S. J. and Granner D. K. (1992) Molecular physiology of the regulation of hepatic gluconeogenesis and glycolysis. Annu. Rev. Physiol. 54, 885±909. Pilkis S. J., Weber I. T., Harrison R. W. and Bell G. I. (1994) Glucokinase structural analysis of a protein involved in susceptibility to diabetes. J. Biol. Chem. 269, 21925±21928. Randle P. J. (1993) Glucokinase and candidate genes for Type 2 (noninsulin-dependent) diabetes mellitus. Diabetologia 36, 269±275. Roncero I., Alvarez E., VaÂzquez P. and BlaÂzquez E. (2000) Functional glucokinase isoforms are expressed in rat brain. J. Neurochem. 74, 1848±1857.

Shiota C., Coffey J., Grinsby J., Grippo J. F. and Magnuson M. A. (1999) Nuclear import of hepatic glucokinase depends upon glucokinase regulatory protein, whereas export is due to a nuclear export signal sequence in glucokinase. J. Biol. Chem. 274, 37125± 37130. Tiedge M., Steffeck H., Elsner M. and Lenzen S. (1999) Metabolic regulation, activity state, and intracellular binding of glucokinase in insulin-secreting cells. Diabetes 48, 514±523. Tippet P. S. and Neet K. E. (1982) Speci®c inhibition of glucokinase by long-chain acyl coenzymes A below the critical micelle concentration. J. Biol. Chem. 257, 12839±12845. Toyoda Y., Miwa l Satake S., Anai M. and Oka Y. (1995) Nuclear location of the regulatory protein of glucokinase in rat liver and translocation of the regulator to the cytoplasm in response to high glucose. Biochem. Biophys. Res. Commun. 215, 467± 473. Van Schaftingen E. (1989) A protein from rat liver confers to glucokinase the property of being antagonistically regulated by fructose 6-phosphate and fructose 1-phosphate. Eur. J. Biochem. 179, 11 179±184. Van Schaftingen E. and Vandercammen A. (1989) Stimulation of glucose phosphorylation by fructose in isolated rat hepatocytes. Eur. J. Biochem. 179, 173±177. Van Schaftingen E., Detheux M. and Veiga da Cunha M. (1994) Shortterm control of glucokinase activity: role of a regulatory protein. FASEB J. 8, 414±419. Vandercammen A. and Van Schaftingen E. (1990) The mechanism by which rat liver glucokinase is inhibited by the regulatory protein. Eur. J. Biochem. 191, 483±489. Vandercammen A. and Van Schaftingen E. (1991) Competitive inhibition of liver glucokinase by its regulatory protein. Eur. J. Biochem. 200, 545±55118. Vandercammen A. and Van Schaftingen E. (1993) Species and tissue distribution of the regulatory protein of glucokinase. Biochem. J. 294, 551±556. Veiga da Cunha M., Xu L. Z., Lee Y.-H., Marotta D., Pilkis S. I. and Van Schaftingen E. (1996) Effect of mutations on the sensitivity of human beta-cell glucokinase to liver regulatory protein. Diabetologia 39, 1173±1179. Yang X., Kow L. M., Funabashi T. and Mobbs C. V. (1999) Hypothalamic glucose sensor similarities to and differences from pancreatic b-cell mechanisms. Diabetes 48, 1763±1772.
