A Novel Cytosolic Dual Specificity Phosphatase, Interacting with Glucokinase, Increases Glucose Phosphorylation Rate*

A novel protein was cloned from a rat liver cDNA library by interaction with the liver glucokinase. This protein contained 339 residues and possessed a canonical consensus sequence for a dual specificity phosphatase. The recombinant protein was able to dephosphorylate phosphotyrosyl and phosphoseryl/threonyl substrates. We called this protein the glucokinase-associated phosphatase (GKAP). The GKAP partially dephosphorylated the recombinant glucokinase previously phosphorylated,in vitro, by protein kinase A. The GKAP fused with green fluorescent protein was located in the cytosol, where glucokinase phosphorylates glucose, and not in the nucleus where the glucokinase is retained inactive by the glucokinase regulatory protein. More importantly, the GKAP accelerated the glucokinase activity in a dose-dependent manner and with a stoichiometry compatible with a physiological mechanism. This strongly suggested that the interaction between GKAP and glucokinase had a functional significance. The cloning of this novel protein with a dual specificity phosphatase activity allows the description of a possible new regulatory step in controlling the glycolysis flux.

The glucokinase (ATP:D-hexose 6-phosphotransferase; EC 2.7.1.1), catalyzes the glucose phosphorylation and is the ratelimiting reaction for glycolysis in hepatocytes and pancreatic ␤ cell. In that respect, the regulation of glucokinase activity is involved in glucose homeostasis.
The glucokinase belongs to the hexokinase family of proteins (1). The hexokinases activities are known to be regulated by interactions with proteins. These proteins locate hexokinases in the proper cell compartment, creating a metabolic microcompartmentation, as reported for hexokinase I. The hexokinase I is bound to porin at the outer mitochondrial membrane. The enzyme has thus a direct access to the source of ATP generated within the mitochondria (2). In this way, the enzyme activity is greatly improved (3). In the nucleus, glucokinase is anchored to the glucokinase regulatory protein (4). On the other hand, the glucokinase (hexokinase IV) can be found either in the nucleus or in the cytoplasm of liver cells. The translocation of glucokinase from the nucleus to the cytoplasm, where the glycolysis takes place, depends upon its substrate concentrations in the extracellular medium (5)(6)(7)(8). Glucokinase is found predominantly into the nucleus of hepatocytes from starved rats, and after 1 and 2 h of refeeding, glucokinase is translocated into the cytosol (8). The role of the glucokinase regulatory protein is thus to sequester the enzyme in a compartment where it might be inactive (5,9). Nevertheless, at the present time, it is not known if a protein is able to retain glucokinase in the cytoplasm either at the plasma membrane or at the mitochondria, close to the glucose or ATP sources. Moreover, post-translational protein modifications might also alter the activity of glucokinase.
The aim of this study was to identify proteins expressed in the liver implicated in the regulation of glucokinase activity after interaction with glucokinase. We performed a two-hybrid screen using rat liver glucokinase as a bait, and showed that glucokinase was interacting not only with the glucokinase regulatory protein, but also with a newly identified dual specificity phosphatase called GKAP. 1 The function of this interaction was analyzed.

MATERIALS AND METHODS
Plasmid Constructions-All manipulations were carried out by standard techniques and plasmid structures verified by DNA sequencing. The entire coding sequence of rat hepatic glucokinase was recovered (BamHI/AvrII fragment) from a plasmid (10) and inserted in frame at the BamHI site of the yeast expression vector pLex9 (pLex-GK) and of the bacterial expression vector pQE-32 (pQE-32-GK). pBSKS-GKRP contains the cDNA of glucokinase regulatory protein (GKRP) (a gift from E. Van Schaftingen). Then, GKRP was subcloned in frame into pLex9. Hexokinase II was similarly constructed in pLex9, in fusion with the yeast LexA DNA binding domain (pLex-HKII). The cDNA encoding GKAP was excised (EcoRI/XbaI fragment) from the pGAD plasmid of the two-hybrid screen and subcloned into pBSSK to produce pBSSK-GKAP. An EcoRI/NotI fragment containing the GKAP insert was then isolated from pBSSK-GKAP and inserted into pGEX-4T2 vector (Amersham Pharmacia Biotech), in frame with GST coding sequence (pGEX-4T2-GKAP) and into an pEGFPc vector (CLONTECH), in frame with green fluorescent protein (pEGFP-GKAP). cDNA Library Screening-A yeast two-hybrid screen was performed as described previously (11), using the entire coding sequence of glucokinase as the bait. The rat liver cDNA library was constructed in pGAD3S2X plasmid. The positive plasmids were tested for specificity using pLex-HKII. Then, their cDNA inserts were sequenced using an Applied Biosystems sequencer (PerkinElmer Life Sciences). The BLAST program was used to search for sequence homology in the GenBank data base.
Production and Purification of Proteins Expressed in Bacteria-Recombinant GKAP was expressed as a glutathione S-transferase (GST) fusion protein, using the plasmid pGEX-4T2-GKAP to transform Escherichia coli BL21. These cells were grown in LB medium containing 0.1 mg/ml ampicillin until the absorbance at 600 nm reached 0.7. The inducer isopropyl-1-thio-D-␤-galactopyranoside was then added to a final concentration of 0.1 mM, and the cultures were shaken at 30°C for 2 h 30 min. Cell extract, purification and elution were performed as described by the manufacturer (Amersham Pharmacia Biotech). For bacterial expression of glucokinase, pQE-32-GK plasmid was used to produce a N-terminal His-tagged fusion protein. Expression of His-GK was induced by incubating transformed E. coli with 0.4 mM isopropyl-1-thio-D-␤-galactopyranoside at 22°C for 22 h. The protein was affinity purified with nickel-nitrilotriacetic acid-agarose under native conditions according to the manufacturer's instructions (Qiagen) and eluted with 250 mM imidazole. The specific activity of recombinant glucokinase was 5 units/mg of protein as measured by the glucose-6-phosphate dehydrogenase-coupled assay (described below). Protein concentrations were determined by the method of Bradford using BSA as a standard and the integrity of the fusion proteins were verified by SDS-PAGE.
Phosphorylation of Glucokinase-Bacterially expressed His-tagged GK (2 g) was incubated with 10 units of the catalytic subunit of protein kinase A (Sigma) for 1 or 2 h at 30°C in a 25-l reaction mixture containing 20 mM HEPES, pH 7.4, 10 mM MgCl 2 , 8 mM ␤-mercaptoethanol, 50 M ATP, and 20 Ci of [␥-32 P]ATP (5000 Ci/mmol, Amersham Pharmacia Biotech). The reaction was stopped by the addition of 25 l of 10 mM sodium phosphate, 10 mM sodium pyrophosphate, pH 8.0, and the phosphorylated glucokinase was affinity-purified on nickel-nitrilotriacetic acid-agarose beads (Qiagen). Bound protein was eluted with 30 l of 250 mM imidazole, pH 7.0, and analyzed by SDS-PAGE and autoradiography.
Phosphatase Assays-Hydrolysis of pNPP by GKAP was carried out in a reaction volume of 800 l containing 50 mM imidazole, pH 7.0, 0.1% ␤-mercaptoethanol, and 20 mM pNPP at 37°C for 2 h. The reaction was stopped by addition of 200 l of 1 M NaOH and the absorbance at 405 nm was measured. For peptide assays, raytide was labeled at its tyrosine residues using [␥ -32 P]ATP and c-Src tyrosine kinase according to the manufacturer's instructions (Oncogene Science), except that the reaction was incubated for 14 -16 h. Phosphorylation of kemptide (Sigma) at its serine residues was conducted in a manner similar to that employed for His-GK. The radiolabeled peptides were separated from free [␥ -32 P]ATP using 1 ϫ 1-cm P-81 sheets of phosphocellulose paper essentially as described in Ref. 12. Dephosphorylation with GST-GKAP was performed for the indicated time at 30°C in 50 l of assay mixture containing 50 mM imidazole, pH 7.0, 0.1% ␤-mercaptoethanol, purified enzyme, and phosphorylated peptide (2-5 ϫ 10 5 cpm). The reactions were stopped by the addition of 0.75 ml of acidic charcoal mixture (0.9 M HCl, 90 mM sodium pyrophosphate, 2 mM NaH 2 PO 4 , 4% (v/v) Norit A). The radiolabeled substrates bound to the charcoal were removed by centrifugation, and released phosphate was measured by scintillation counting of the supernatant. The ability of GKAP to dephosphorylate glucokinase was tested by incubating 10 l of His-GK elution (see above) with GST-GKAP in the presence of 0.1% ␤-mercaptoethanol. The reaction was carried out at 30°C for 10 min, stopped by the addition of SDS sample buffer, and then analyzed by SDS-PAGE and autoradiography. The dephosphorylation study was performed using GST-GKAP purified from bacteria, GKAP, and GKRP generated using an in vitro transcription/translation system (Promega). 10 l of the translation mix was then used for the phosphatase assay.
Tissue-specific Expression of GKAP-Total RNA was purified from rat tissues using the method of Chomczynski and Sacchi (13). Northern blot analysis was performed as described previously (11) using as probe the full-length GKAP cDNA. The presence of GKAP in ␤ versus other pancreatic cells was assessed by RT-PCR analysis. The rat pancreatic islets were isolated after collagenase digestion of the pancreas, and resuspended in PBS-EGTA supplemented with 0.1 mg/ml trypsin. The ␤ cells were sorted by the autofluorescence for FAD and the light scatter parameter (cell size) using a FACStar plus (Becton Dickinson) with an argon laser (163, Spectra Physics) at 488 nm. The quality of the sorting was verified by subsequent FACS analysis of samples of the purified cells. RNA from ␤ cells were about 40 -50-fold enriched in specific mRNA when compared with RNA from total pancreas (data not shown). Reverse transcription was performed using 5 g of RNA from the sorted ␤ cells (and white adipose tissue) using a random primer, and then PCR was achieved using two oligonucleotides: ccgtacttaccacaggg and gtttgtgatcccagcgc, specific for GKAP leading to a band of 250 bp.
Transfection of Fluorescent GKAP-The mhAT3F hepatoma cell line was derived from transgenic mice synthesizing the SV40 large T and small t antigens under the control of antithrombin III promoter (14). Cells were grown on glass four-chamber slides (Falcon, CultureSlides) in Dulbecco's modified Eagle's medium/Ham's F12, Glutamax (Life Technologies, Inc.) supplemented with penicillin, streptomycin, 0.1 M insulin, 1 M dexamethasone, 1 M triiodothyronine, and 5% fetal calf serum. mhAT3F cells were plated and transfected by using the calcium phosphate-DNA precipitation method (15). The cells were transfected with pEGFP and pEGFP-GKAP vectors. After 2 days, cells were washed three times with PBS, fixed for 30 min in 4% paraformaldehyde in PBS, and quenched for 10 min in 50 mM NH 4 Cl. The nuclei were stained in bright blue using Hoechst 33258 (0.5 g/ml) for 5 min. After extensive washes in PBS the coverslips were mounted in Mowiol (Hoechst). The fluorescence microscopy was performed using an Olympus IMT2 inverted microscope, fluorescence immersion lens (40ϫ) and fluorescein isothiocyanate filter. Photographs (Kodak Panther P1600 film) were taken with an Olympus MO-4Ti camera adapted to the microscope.
Glucokinase Activity-The glucokinase activity was determined by following the accumulation of NADH at 340 nm using a spectrophotometer. The glucokinase phosphorylates glucose into glucose 6-phosphate (G-6-P), this substrate is transformed into 6-phosphogluconate and NADH by the G-6-P dehydrogenase (from Leuconostoc mesenteriode). This last reaction depends directly on the glucokinase activity. The glucokinase activity was measured in the presence of 0.1 M glucose, 50 mM Hepes, pH 7.4, 0.1 M KCl, 2.5 mM dithiothreitol, 0.5 mM NAD, 2 units of G-6-P dehydrogenase, and 1 or 2 g of recombinant GK, during 15 min at 30°C. The base line was recorded for 2 min, and then the reaction was initialized by the addition of ATP-Mg 2ϩ (5 mM). The glucose phosphorylating activity was also tested without recombinant glucokinase, without glucose, without ATP, or with 1 mM glucose or GKAP only. In these conditions, no activity was detected. 10 l of recombinant glucokinase was phosphorylated directly after elution from nickel beads by incubation during 30 min at 30°C in 90 l of medium containing protein kinase A buffer (1ϫ Biolabs), 8 mM ␤-mercaptoethanol, 200 M ATP-Mg 2ϩ , 50 mM NaCl, 5% glycerol, and in absence or presence of 25 units of the protein kinase A catalytic subunit (Biolabs). We verified by autoradiography that the recombinant glucokinase was phosphorylated by protein kinase A in presence of 10 Ci of [␥-32 P]ATP. Then, the glucokinase activity was measured on 100 l of the phosphorylation reaction. The GKAP effect on glucokinase activity was recorded after the addition of various amounts of the fusion protein 10 min after the beginning of the reaction. The results were expressed as enzymatic units per milligram of protein.

Identification of Proteins Interacting with Hepatic
Glucokinase-To identify novel proteins that interact with hepatic glucokinase, we have used the two-hybrid system (16). A fusion between the LexA DNA-binding domain and the full-length glucokinase was used as a bait to screen a library of rat liver cDNA fused with the Gal4 activation domain. Approximately 5 ϫ 10 6 yeast transformants were tested, and 70 clones were classified as positive, since they interacted strongly with glucokinase and not with an unrelated protein such as lamin.
Restriction mapping and sequence analysis revealed five different groups of positive clones. One of these groups corresponded to the glucokinase regulatory protein, whose specific interaction with glucokinase is known. Another group contains the cDNA of an unknown protein, which was studied here.
The deduced amino acid sequence of this new protein possesses the canonical motif, HCXXGXXR(S/T), common to all protein-tyrosine phosphatases (PTPs) (see below), suggesting that it may be a novel protein phosphatase. We called it GKAP (glucokinase-associated phosphatase). GKAP cDNA extends 1314 bp with a possible translation start site (17) at bp 19, followed by a 1017-bp coding region (Fig. 1). Searches of missing sequence upstream of this initiating methionine, using the 5Ј-rapid amplification of cDNA ends technique (CLONTECH), failed to recover a longer fragment than the insert of cDNA isolated during the yeast two-hybrid screen. Thus, we have considered the position 19 -21 bp as the first ATG codon. The GKAP was shown to interact specifically with glucokinase but not with another isoform of mammalian hexokinase family, hexokinase II (Table I).
The GKAP amino acid sequence predicted a polypeptide of 339 residues with an estimated molecular mass of 37 kDa. Comparison of this sequence with the GenBank data base revealed that GKAP is related to members of a dual specificity protein-tyrosine phosphatases (dsPTP) family, which hydrolyze phosphate from Ser/Thr as well as Tyr residues. Throughout its entire length, GKAP protein exhibits the greatest similarity to the yeast dsPTP yVH1. The two proteins are 28% identical (42% similar), the greater homology being within C-terminal regions, and both of them contain the PTP domain at the N-terminal half ( Fig. 2A). With regard to mammalian members of dsPTP whose PTP signature motif is located on the C-terminal portion, GKAP also shows significant identity to hVH3 (35%), rVH6 (33%), CL100 (30%), rVH2 (29%), and PAC1 (27%), over a stretch of 141 residues throughout the active site consensus sequence. The alignment of these regions between GKAP and these other family members is illustrated in Fig. 2B.
GKAP Tissue Distribution-We have examined the expression pattern of GKAP in various rat tissues by Northern blot analysis (Fig. 3). A single 2.2-kilobase pair transcript was observed in all the tissues tested. Relatively higher levels of expression were seen in lung, brain, and pancreas, whereas moderate levels were found in heart, liver, intestine, skeletal muscle, and brown adipose tissue. Lower levels of GKAP mRNA were also detected in kidney and white adipose tissue. Among pancreatic cells, we verified by RT-PCR analysis that GKAP was present in ␤ cells, pancreatic cells that express glucokinase (Fig. 3).
Location of the Fusion GFP-GKAP in mhAT3F Cells-Hepatoma cells, mhAT3F, were transiently tranfected with pEGFP-GKAP. The subcellular location of the fusion proteins was studied using the green fluorescence of GFP (Fig. 4). The nuclei are visualized by bright blue fluorescence (Fig. 4). The GFP is distributed uniformly in mhAT3F cells, whereas the GFP-GKAP is shown to be located in the cytoplasm and not in the nucleus. A specific subcellular location is thus observed for the GFP-GKAP fusion protein.
Protein Phosphatase Activity of GKAP-To investigate whether GKAP encodes an active phosphatase, recombinant GKAP was expressed, purified, and assayed for enzymatic activity. As shown in Fig. 5A, the fusion protein GST-GKAP hydrolyzed the artificial substrate p-nitrophenyl phosphate (pNPP). Hydrolysis increased linearly as a function of GKAP concentration, and this catalytic activity was effectively abolished by sodium orthovanadate, an inhibitor of the proteintyrosine phosphatases. We then tested the ability of GKAP to dephosphorylate two peptide substrates, raytide and kemptide, which contain respectively phosphotyrosine and phosphoserine. Results with raytide were similar to those with pNPP. Purified GST-GKAP hydrolyzed the phosphate monoester from raytide and this dephosphorylation was blocked by sodium orthovanadate (Fig. 5B). However, under similar conditions, we could not detect phosphatase activity using phosphoseryl kemptide as substrate (data not shown). This could be an index of a substrate selectivity of GKAP. It remains possible that in vivo GKAP dephosphorylates phosphoserine/threonine residues within its natural substrates.
Since we have observed an association of glucokinase with GKAP, we explored the possibility that glucokinase might be the physiological substrate of GKAP. Glucokinase was first phosphorylated by protein kinase A (Fig. 6A), using the puri-

TABLE I Specificity of GKAP interaction with GK in the double-hybrid system
The GKAP was not able to interact with another isoform of hexokinase, hexokinase II in yeast. The known interaction between glucokinase and glucokinase regulatory protein was observed. The number of ϩ indicated the intensity of the blue coloration obtained after ␤-galactosidase filter assay and compared with a positive control (Ras/Raf interaction: ϩϩϩ) on the same filter. The absence of blue coloration was indicated by Ϫ, comparable to the negative control (Lamin/Raf: Ϫ) on the same filter. These cotransfections of yeast were reproduced independently two times. fied glucokinase expressed in E. coli as a His-tagged fusion protein (Fig. 6B). Then, labeled glucokinase was incubated with GST-GKAP. Recombinant GST-GKAP (Fig. 6D) dephosphorylated glucokinase in a concentration-dependent manner, as shown in Fig. 6C (lanes 1-4). The dephosphorylation of labeled glucokinase by GST-GKAP was not complete even when the GST-GKAP treatment was greater than 2 h (data not shown). Moreover, neither E. coli lysate (data not shown) nor GST alone purified in the same conditions as GST-GKAP (Fig.  6C, lane 5) were able to dephosphorylate glucokinase. Thus, proteins copurified with GST-GKAP are unlikely to be responsible for the glucokinase dephosphorylation observed. To fur-ther assess the specificity of the reaction, the GKAP was produced by using an in vitro transcription/translation system (Fig. 6F). This GKAP generated an identical dephosphorylation of glucokinase (Fig. 6E, lane 2). The glucokinase regulatory protein, produced by reticulocyte lysate (Fig. 6F, lane 3), was unable to dephosphorylate glucokinase (Fig. 6E, lane 4). This eliminated the possible dephosphorylating activity of proteins present in the reticulocyte lysate. Taken together these experiments suggest that glucokinase dephosphorylation was due to GKAP. We also showed that the dephosphorylating activity was sensitive to inhibition by a phosphatase inhibitor, sodium orthovanadate (2 mM) (Fig. 6E, lane 1).
Effect of GKAP on GK Activity-We further investigated the role of GKAP by studying its effect on the glucokinase activity. The basal recombinant glucokinase activity was about 1-4 units/mg of protein, in our experimental conditions, as de-scribed under "Materials and Methods." No hexokinase or glucokinase activity could be detected in GST-GKAP or in GST preparations (data not shown). Adding GST-GKAP (6 g) to the active glucokinase led to a 2-fold augmented production of NADH quantified by recording the optical density at 340 nm (Fig. 7A). This was not observed when GST alone was added. A second addition of 6 g of GST-GKAP similarly induced an increase in GK activity (Fig. 7A). The 2.5 stimulatory effect of GKAP on GK activity was dependent on the quantity of GKAP added and it was saturated over 20 g of added GKAP (Fig.  7B). The glucokinase activity was significantly decreased by 31 Ϯ 9% (n ϭ 5) when glucokinase was previously phosphorylated by protein kinase A (Fig. 7C). The addition of 3 g of GKAP on phosphorylated GK induced a 73 Ϯ 26% (n ϭ 8) stimulation of its activity, whereas a similar amount of GKAP (3 g) increased only by 50 Ϯ 19% the unphosphorylated glucokinase activity.

DISCUSSION
The association between glucokinase and a new protein (GKAP) was observed by using the two-hybrid system. GKAP mRNA was detected in all the rat tissues tested. Although these results does not correlate with the limited distribution pattern of glucokinase, GKAP mRNA appears in all tissues expressing glucokinase, suggesting that GKAP may have a specific role in these tissues. Furthermore, GKAP specifically interacts with GK and not with the closely related protein HKII. Because a dual specificity phosphatase consensus sequence was revealed after computer analysis of the cloned  3 and 4, 1 and 3 g). C, a representative autoradiograph of dephosphorylation of glucokinase by recombinant GKAP. Increasing amounts of purified recombinant GST-GKAP (0, 1, 5, and 10 g) were assayed for their ability to dephosphorylate radiolabeled His-glucokinase (lanes 1, 2, 3, and 4). sequence, the function of this new protein was studied.
It has been shown that purified glucokinase from rat liver is phosphorylated by PKA on serine residue(s) and that the phosphorylation reduced (50%) the enzyme activity (18). In our hands, recombinant glucokinase activity was also diminished by PKA phosphorylation. The yeast hexokinase HK2, the isoform homologous to glucokinase, has two phosphorylation sites. The serine 15, belonging to a protein kinase A consensus phosphorylation sequence, is phosphorylated after a shift to a medium with low glucose concentration (19). This phosphorylation would affect the oligomeric state of HK2 (20) and is essential for glucose signaling (21). On serine 158, a transitory phosphorylation might occur during the transfer of ␥-phosphate from ATP to glucose (22). In human ␤ cell glucokinase, serine 151 corresponding to serine 158 of yeast hexokinase 2 is also phosphorylated, as a product of ATP hydrolysis (23). Thus, a phosphorylated-dephosphorylated state of glucokinase is observed. Nevertheless, the identity of the amino acid residues bearing the phosphate, transitorily or not, is still an open question (23)(24)(25).
GKAP is related to a subclass of PTPs commonly referred to as dsPTP (reviewed in Ref. 26), hydrolyzing phosphate from Ser/Thr and Tyr residues. Importantly, all of the amino acids previously shown to participate in the catalytic mechanism underlying dual specificity phosphatase activity are conserved in GKAP, including Cys-114, the catalytic cysteine that functions as the active site nucleophile, and Asp-83 and Ser-121, both involved in hydrolysis of the thiol phosphate intermediate (27,28). This protein has a long C-terminal amino acid domain of unknown function. By comparison, the recently cloned MKP-5 dual specificity phosphatase for p38 possesses a long N-terminal domain of unknown function (29), suggesting that the position of the catalytic domain is not the signature for this family of proteins. We confirmed experimentally that the recombinant protein is indeed a phosphatase because it dephosphorylated a model substrate pNPP and the phosphotyrosine raytide and its activity is blocked by phosphatase inhibitor, vanadate. Since glucokinase is interacting with GKAP and, generally a phosphatase is associated to its own substrate, glucokinase might be the phosphosubstrate for this phosphatase. Indeed, the recombinant glucokinase, phosphorylated by PKA (on serine residues, according to Ref. 18), was dephosphorylated by GKAP produced by two techniques. However, the dephosphorylation was not complete presumably because the in vitro phosphorylation of glucokinase by protein kinase A occurred also on residues outside the specific motif recognized by GKAP. Similarly, it was reported that MKP dephosphorylates only partially the activated MAPK (30,31). Our results supported the notion that GKAP is a dual (phosphoserine/ threonine and phosphotyrosine) phosphatase and may be a glucokinase-specific phosphatase.
GKAP is located in the cytoplasm and not in the nucleus. This is the cell compartment where the glucokinase is catalytically active and where glycolysis occurs. Thus, it can be hypothesized that the interaction between glucokinase and GKAP regulates glucokinase activity. We showed that GKAP stimulated glucokinase activity in a dose-dependent manner. This effect occurred at a stoichiometry of GK/GKAP of 1/3 to 1/5, compatible with a physiological phenomenon. The glucokinase activity, diminished by a previous phosphorylation of the recombinant protein by PKA, was augmented by the presence of the phosphatase GKAP to a higher extent than that obtained with unphosphorylated glucokinase. The mechanism by which glucokinase activity is stimulated by GKAP remains unknown. The dephosphorylation of the glucokinase might help the phosphotransfer, as suggested in the yeast hexokinase where a transitory phosphorylated state exists as the result of the phosphotransferase activity of the enzyme (21). The dephosphorylation of glucokinase by GKAP could be implicated in the proteolysis of glucokinase. Indeed, it has recently been reported that mice bearing an inactivated glucokinase regulatory protein gene exhibit decreased levels of liver glucokinase. Since similar amounts of glucokinase are produced by control and knockout mice, the retention of glucokinase in the nucleus by its interaction with GKRP might protect it from degradation in the cytosol (9).
In conclusion, by interaction with glucokinase, we cloned a novel dual specificity phosphatase (GKAP), which is located into the cytoplasm. We described a regulatory mechanism that FIG. 7. Effect of the addition of GKAP on glucokinase activity. A, the glucokinase activity (NADH appearance at 340 nm recorded during 9 min) was measured in the presence of GST-GKAP (6 g were added twice) or in the presence of GST (10 g) (gray line). This experiment was reproduced four times. B, the effect of increasing amount of GST-GKAP on the recombinant glucokinase activity expressed in units/mg of protein. C, the glucokinase activity was measured on recombinant glucokinase phosphorylated (gray bars) or not (white bars) by protein kinase A, and in absence or presence (hatched bars) of the GST-GKAP, expressed in units/mg of protein.
accelerates the glucose phosphorylation. This new step might control the glycolysis rate, and as a consequence, the glucose uptake by the liver and the insulin secretion by pancreatic ␤ cell.