Biochemical Basis of Glucokinase Activation and the Regulation by Glucokinase Regulatory Protein in Naturally Occurring Mutations*

Glucokinase (GK) has several known polymorphic activating mutations that increase the enzyme activity by enhancing glucose binding affinity and/or by alleviating the inhibition of glucokinase regulatory protein (GKRP), a key regulator of GK activity in the liver. Kinetic studies were undertaken to better understand the effect of these mutations on the enzyme mechanism of GK activation and GKRP regulation and to relate the enzyme properties to the associated clinical phenotype of hypoglycemia. Similar to wild type GK, the transient kinetics of glucose binding for activating mutations follows a general two-step mechanism, the formation of an enzyme-glucose complex followed by an enzyme conformational change. However, the kinetics for each step differed from wild type GK and could be grouped into specific types of kinetic changes. Mutations T65I, Y214C, and A456V accelerate glucose binding to the apoenzyme form, whereas W99R, Y214C, and V455M facilitate enzyme isomerization to the active form. Mutations that significantly enhance the glucose binding to the apoenzyme also disrupt the protein-protein interaction with GKRP to a large extent, suggesting these mutations may adopt a more compact conformation in the apoenzyme favorable for glucose binding. Y214C is the most active mutation (11-fold increase in \batchmode \documentclass[fleqn,10pt,legalpaper]{article} \usepackage{amssymb} \usepackage{amsfonts} \usepackage{amsmath} \pagestyle{empty} \begin{document} \(k_{\mathrm{cat}}{/}K_{0.5}^{h}\) \end{document}) and exhibits the most severe clinical effects of hypoglycemia. In contrast, moderate activating mutation A456V nearly abolishes the GKRP inhibition (76-fold increase in Ki) but causes only mild hypoglycemia. This suggests that the alteration in GK enzyme activity may have a more profound biological impact than the alleviation of GKRP inhibition.

Glucokinase (GK) 2 plays a central role in maintaining glucose homeostasis (1)(2)(3). It serves as a glucose sensor due to its specific kinetic properties that include low affinity and positive cooperativity for glucose and a lack of inhibition by its product glucose-6-phosphate (4 -6). In pancreatic ␤-cells, GK regulates glucose-dependent insulin secretion by modulation of the glycolytic pathway and subsequently the ATP/ADP ratio. In the liver, GK stimulates glucose disposal by converting glucose to glycogen for storage. Hepatic GK is tightly regulated by the 68-kDa glucokinase regulatory protein (GKRP) through the formation of a GKRP-GK complex followed by sequestration in the nucleus (7). Physiologically, the interaction between GK and GKRP is promoted by fructose-6-phosphate and suppressed by fructose-1-phosphate. Upon increasing glucose concentrations, GK dissociates from GKRP and translocates from the nucleus to the cytoplasm resulting in an increase in GK activity.
Alteration in GK activity and its regulation is associated with abnormal glycemia as evidenced in naturally occurring mutations. More than 190 inactivating GK mutations have been identified in patients with MODY2 (maturity onset diabetes of the young 2). In contrast, five activating GK mutations (T65I, W99R, Y214C, V455M, and A456V) lead to hyperinsulinemic hypoglycemia (8 -11). The degree of hyperinsulinemic hypoglycemia in the affected patients is variable. The mutation Y214C causes severe and possibly fatal hypoglycemia, whereas other mutations are associated with mild hypoglycemia and are in some cases asymptomatic. Interestingly, all of these activating mutations cluster in an allosteric site of GK where small molecule activators bind, suggesting a critical role of the allosteric site in the regulation of GK activity. Both GK activators and activating mutations increase enzyme activity by enhancing the affinity for glucose as described by a decrease in K 0.5 (8 -12). In vivo, GK activators have been shown to lower glucose levels and stimulate insulin release in animal models (12)(13)(14). The potential efficacy of GK activators in humans has drawn great interest as a promising therapeutic treatment for Type 2 diabetes. Because the in vitro effects of small molecule GK activators are similar to the activating mutations in GK, a better understanding of the biochemical basis for the activation of GK activity by these mutations may provide insight for the design of GK activators as effective therapeutics.
Protein structural analyses have shown that substantial conformational changes in GK occur during glucose binding (15). Recently, we investigated the GK structural changes induced by glucose binding through transient kinetic studies and demonstrated that the slow enzyme conformational change contributed to the positive cooperativity and the low binding affinity for glucose (16). We also showed that the most clinically severe, activating GK mutation Y214C facilitated the slow enzyme con-formational changes to the active form by a factor of 7.5-fold. However, it is unknown whether all of the activating mutations at the allosteric site follow the same activation mechanism, how the mutations affect GKRP inhibitory regulation, and how the variation in their biochemical properties relate to the clinical phenotype.
In this paper, we have extended our studies to include published naturally activating mutations. In addition, two engineered activating mutants, Y214A and Y215A, at the allosteric site were included in the studies to further investigate the structure-function relationship of the allosteric site involved in the protein conformational change. The GK activation and GKRP binding for the GK-activating mutations were characterized in vitro. We show that the structural perturbation at the allosteric site by single point mutations alters the kinetic basis of enzyme activation as well as the protein-protein interactions with GKRP.
The recombinant GK WT and mutations with a hexa-His tag at the N terminus were expressed in Escherichia coli and purified using the protocol described previously. A complete protease inhibitor mixture (Roche Applied Science) was added to each step of the purification for the A456V mutation to reduce the proteolysis of the protein. Purity of the enzyme was verified to be Ͼ95% by SDS-PAGE and quadrupole time-of-flight mass spectrometry. The molecular mass of His-tagged GK was confirmed to be 55 kDa by quadrupole time-of-flight mass spectrometry. Enzyme concentration was determined by the Bradford method (18). Aliquots of WT and mutated GK were stored at Ϫ80°C in a pH 7.5 buffer containing 25 mM HEPES, 50 mM NaCl, 5 mM DTT, and 5% glycerol.
The cloning of GKRP has been described previously (19). GKRP with a C-terminal FLAG tag was expressed in E. coli DH10B cells using a pFLAG-CTC expression vector. Expression was confirmed by Western analysis using an anti-GKRP monoclonal antibody obtained from Santa Cruz Biotechnology. Bacterial cells were grown in a 10-liter fermentation with Terric Broth medium initially at 37°C, containing 100 g/ml ampicillin and inducted with 0.2 mM isopropyl 1-thio-␤-D-galactopyranoside upon reaching an A 600 of 0.6. After overnight expression at 23°C, cells were harvested. Cell pellets were resuspended in 50 mM Tris-HCl, pH 7.4, 150 mM NaCl. The cells were subsequently lysed by microfluidization on ice. Purification was achieved by batch absorption using an anti-FLAG monoclonal antibody Mab2 affinity gel (Sigma). Resin was then loaded onto a chromatography column where FLAG-tagged GKRP was eluted from the resin by competition using 100 g/ml FLAG peptide (Sigma). Pooled fractions were then loaded onto a size exclusion column where the peak at 72 kDa was collected. Purity of the enzyme was verified to be Ͼ95% by SDS-PAGE. Protein concentrations were determined by the Bradford method (18). GKRP was stored in aliquots at Ϫ80°C in a pH 7.4 buffer containing 25 mM Tris-HCl, 150 mM NaCl, 20% glycerol, and 5 mM DTT.
Steady-state Kinetics of GK-activating Mutations-Steadystate kinetic parameters of GK WT and activating mutations were determined as described previously using a pyruvate kinase/lactate dehydrogenase-coupled assay (16). Experiments were performed with 0 -50 mM glucose at a saturating concentration of 5 mM ATP in a buffer containing 50 mM HEPES, pH 8.0, 6 mM MgCl 2 , 25 mM KCl, 0.7 mM NADH, 2 mM DTT, 4 mM phosphoenolpyruvate, and 1 unit/ml of pyruvate kinase/lactate dehydrogenase at 25°C. The saturating concentration of MgCl 2 was used in the assay to ensure the majority of ATP (Ͼ98%) was present as MgATP. Control experiments were carried out to demonstrate the coupling enzymes were in excess and the initial rate was linearly dependent on GK concentrations. To determine whether activating mutations affect K m for ATP, initial rates were measured by varying ATP (0 -5 mM) at a saturating concentration of 50 mM glucose at 25°C using pH 8.0 buffer. The kinetic parameters for substrate analog 2-deoxyglucose were determined at 25°C using pH 8.0 buffer by varying 2-deoxyglucose concentrations from 0 -400 mM at 5 mM ATP. The steady-state kinetic parameters were obtained by fitting the initial rates to the Hill equation (Equation 1) for sugar substrates due to the sigmoidal kinetics and to the Michaelis-Menten equation (Equation 2) for ATP with hyperbolic kinetics using nonlinear regression analysis (Prism, Graph Pad Inc.) as follows, where V max is the maximal activity of GK, S is the sugar concentration, K 0.5 is the sugar concentration at half-maximal activity, h is the Hill number, and v is the activity at a given sugar concentration S. Glucose Binding to GK WT and Activating Mutations-Binding of glucose to the GK enzymes caused a change in the intrin-sic enzyme fluorescence that was monitored with excitation at 290 nm. The equilibrium binding affinity of glucose was determined as described previously (16). Briefly, glucose binding to GK WT and the mutations was monitored at 330 nm by fluorescence spectrometry. The K D values were determined by fitting data to a binding isotherm using non-linear regression analysis.
The transient kinetics study of glucose binding to GK using a stopped flow spectrophotometer has been described in detail elsewhere (16). Final enzyme concentrations of 5, 10, or 20 M were utilized for these measurements. Briefly, presteady state kinetics were monitored using a 320-nm cutoff filter upon excitation at 290 nm by the rapid mixing of glucose and enzyme in separate syringes. The time traces were fit to the double exponential equation to obtain the rate constants (k obs1 and k obs2 ), the amplitudes (A 1 and A 2 ), and the fitting variable C.
Inhibition of GK WT and the Activating Mutations by GKRP-GKRP binds to GK and inhibits the enzyme activity in a competitive manner with respect to glucose (20). Kinetic analyses were performed to determine the inhibition constant of GKRP with GK WT and the mutations by varying the glucose concentrations as described below and GKRP concentrations (0 -500 nM). The range of glucose concentrations was adjusted to reflect the differences in glucose affinity of the mutations (0 -50 mM for GK WT, W99R, V455M, and Y214A and 0 -10 mM for T65I, Y214C, A456V, and Y215A). GK and GKRP were mixed in a buffer containing 50 mM HEPES, pH 8.0, various concentrations of glucose, 6 mM MgCl 2 , 0.7 mM NADH, 2 mM DTT, 1 unit/ml pyruvate kinase/lactate dehydrogenase, 4 mM phosphoenolpyruvate, 25 mM KCl, and 0.2 mM sorbitol-6-phosphate at 25°C. Sorbitol-6-phosphate was used in the assay, because it is a more potent analog than fructose-6-phosphate in promoting the association of GKRP with GK. The assay mix was allowed to equilibrate for 10 min before the reaction was initiated by the addition of 5 mM ATP. The assay was conducted in a 96-well plate with a final volume of 100 l. The apparent inhibition constant of GKRP, K i app , at a given glucose concentration was determined by fitting initial rates to the Morrison and where v i /v o is the fractional velocity of GK activity in the presence of GKRP relative to that in the absence of GKRP, E t is the total enzyme concentration, and I is the GKRP concentration. The inhibition constant K i was determined by plotting K i app against varying sugar substrate concentrations S. The experiments were performed in triplicates.
The K i values of GKRP were also determined using the noncooperative substrate analog 2-deoxyglucose (0 -300 mM). Kinetic analyses were performed as outlined above.

Steady-state Kinetics of GK-activating Mutations-
The central clinical feature of glucose levels in the range of 1-3 mM compared with the normal 4 -5.4 mM (8 -11). To correlate the effects of activating mutations to the clinical consequence, steady-state kinetic properties of the recombinant activating mutations were characterized and compared with GK WT ( Table 1). All of the activating mutations lowered the K 0.5 for glucose and reduced the positive cooperativity of glucose as described by a decrease in the Hill number (h). As a result, the catalytic efficiency with regards to glucose, k cat /K 0.5 h , is enhanced for the activating mutations. The K m values for ATP changed by 0.7-2.5-fold for the mutations.
The k cat values for some of the mutations determined here were 2-3-fold lower than the reported results in earlier studies (8 -11). The underlying cause of the discrepancy is unknown but may be attributable to the experimental conditions, such as temperature and pH, or the GK construct, because we used a hexahistidine tag, whereas previous reports utilized a glutathione S-transferase tag (8 -11). Because of the large size of the glutathione S-transferase tag (26 kDa), we chose a His 6 tag to minimize the potential disturbance of the native conformation of GK proteins. Identical kinetic properties were observed between the non-tagged and hexa-His-tagged GK enzymes (data not shown).
Effects of GK-activating Mutations on Glucose Binding-Previous transient kinetic studies of glucose binding by stopped flow spectrophotometry demonstrated that glucose binding to GK WT followed biphasic kinetics that fit best to a two-step reversible mechanism (16). The first kinetic phase was interpreted as a bimolecular event where glucose loosely binds to the apoenzyme form E*, followed by a second kinetic phase of enzyme conformational change from E*⅐S to E⅐S. The kinetic scheme was reported as follows. All of the activating mutations retained the two-step mechanism of glucose binding but altered the individual microscopic rate constants (Table 2). Some mutations enhanced the first step of E*⅐glucose formation, as evidenced by either increasing k 1 (observed in T65I and A456V) or decreasing k Ϫ1 (observed in Y215A). Some mutations favor E*⅐glucose to E⅐glucose transformation by increasing k 2 (observed in W99R, Y214A, and V455M) or decreasing k Ϫ2 (observed in Y214C). The ratio of k 1 /k Ϫ1 and k 2 /k Ϫ2 represent the association constant of glucose binding to the apoenzyme E* and the degree of conversion of E*⅐S to E⅐S, respectively. The overall equilibrium binding affinity of glucose can be calculated based on the individual microscopic rate constants of the two-step glucose binding using the following equation (22).
The results showed a good agreement with the experimentally determined K D ( Table 2). Based on the measured K D , T65I and A456V increased glucose binding the most (Ͼ10-fold), whereas W99R and V455M only increased the binding moderately (2-3-fold).
Effects of GK-activating Mutations on GKRP Inhibition-As shown in Fig. 1, the naturally occurring GK-activating mutations have different responses to GKRP inhibition at the physiological concentration of 5 mM glucose. At various glucose concentrations, the GK WT and mutations showed a glucosedependent linear increase in IC 50 value of GKRP, indicative of competitive inhibition of GKRP with respect to glucose (see supplemental data). It has been shown that GKRP is a competitive inhibitor for GK WT (20). To quantify the effects of the mutations on the protein-protein interaction with GKRP, the inhibition constant (K i ) of GKRP was determined for each mutation as shown in Table 3. W99R and Y214A retained similar binding interactions with GKRP compared with WT, whereas T65I, Y215A and A456V nearly abolished the GKRP interaction.
Because the GK WT and some of the mutations displayed sigmoidal kinetics for glucose that may complicate the inhibition data interpretation, the substrate analog 2-deoxyglucose was also utilized in the determination of K i for GKRP, because it exhibits little or no cooperativity for GK WT and the activating mutations. The K i values obtained using both glucose and 2-deoxyglucose as the substrate were in close agreement with each other, consistent with a similar competitive inhibition of GKRP by both sugar substrates ( Table 3). The K i value using glucose as the substrate was used for the rest of discussion.

DISCUSSION
The naturally occurring GK-activating mutations T65I, W99R, Y214C, V455M, and A456V were studied together with the engineered mutations Y214A and Y215A because of previous reports implicating the latter two residues as novel activat-

TABLE 2 The microscopic rate constants of transient kinetics and the equilibrium constants of glucose binding for GK WT and activating mutations
The microscopic rate constants were determined from k obs1 and k obs2 as described previously (16).  ing mutations (23,24). All of the activating mutations display an increased catalytic efficiency that is contributed mainly from the decrease in K 0.5 for glucose. The 0.75-2.5-fold changes in the K m value of ATP for the activating mutations may not have significant biological consequence, because the cellular ATP concentration is consistently maintained at mM ranges where the enzyme is nearly saturated by ATP. Because of its low affinity with glucose, GK does not operate at its maximal activity (k cat ) under physiological conditions; therefore, the catalytic efficiency, k cat /K 0.5 h , may represent the most relevant rate constant of the activity of the enzyme.
Correlation of in Vitro Biochemical Properties with the Clinical Phenotype for GK-activating Mutations-Actions of GK in both glucose-dependent insulin regulation in ␤-cells and glucose disposal in the liver contribute to its role in maintaining glucose homeostasis. Therefore, to evaluate the correlation of the in vitro biochemical properties with the clinical phenotype, both effects of activating mutations on the enzyme activity and GKRP inhibitory interaction need to be considered. Table 4 summarizes the changes in the catalytic efficiency and the GKRP inhibition by the mutations relative to WT and the associated clinical phenotype.
Y214C is the most activating mutation with a catalytic efficiency of 11-fold higher than GK WT, and it also decreases the GKRP inhibitory interaction by 6-fold. In the clinical observations, the Y214C-affected patient has the most severe hypoglycemia and hyperinsulinemia (9). Treatment with diazoxide and partial pancreatectomy did not relieve the symptoms, and eventually the patient's brain was irreversibly damaged. Enlarged and hyperfunctional islets were observed by pancreatic histology. This is consistent with the prediction by mathematical modeling based on in vitro analysis that Y214C has the lowest threshold for glucose-stimulated insulin release (9). The patient also showed signs of accumulation of glycogen in the liver. This could be the result of both GK activation and decreased GKRP inhibitory regulation in the liver.
On the other hand, A456V moderately increases the catalytic efficiency but decreases the GKRP binding by 76-fold. Significant glycogen storage in liver was observed in the patient by glucagon stimulation tests that showed rapid increases in blood glucose from low fasting levels. However, the significant disruption in GKRP inhibitory regulation in the liver does not translate into a severe clinical phenotype for A456V. Patients are either asymptomatic or can be treated by diazoxide, suggesting ␤-cell GK may be more essential in glucose homeostasis (8,9). This is consistent with the results of GK tissue-specific knockout in mice. A ␤-cell-specific knock-out of GK in mice led to death within 1 week due to severe diabetes similar to the global GK knock-out, whereas the liver-specific GK knock-out mice only experienced mild hyperglycemia but displayed pronounced defects in glycogen synthesis (25). Another possibility is that the alleviation of GK inhibition by GKRP in A456V may be offset by the loss of GK protection by GKRP from degradation by sequestering GK in the hepatocyte nucleus. A parallel loss of GK protein and activity has been observed in GKRP knock-out mice (26). These results suggest that the alteration in GK enzyme activity has more profound biological impact than the change in GKRP binding.
The other three naturally occurring mutations, T65I, W99R, and V455M, have a similar catalytic efficiency of 2-3-fold higher than WT. The fasting glucose levels are in the range of 2-3 mM for these patients, and the insulin levels vary (8,10,11). Although the functional characterization of GK-activating mutations may explain some of the phenotype, the clinical symptoms and course of hyperinsulinemic hypoglycemia patients cover a broad spectrum from asymptomatic hypoglycemia to unconsciousness and seizures, even within the same family, implicating a complex mechanism for GK regulation in vivo.
Biochemical Basis for GK Activation and GKRP Regulation for GK-activating Mutations-All of the mutations studied here achieve their activation effects by enhancing overall glucose binding. The transient kinetic studies of glucose binding reveal that the mutations have a different biochemical basis for the increase in glucose binding affinity. The T65I, Y215A and A456V mutations mainly accelerate the initial bimolecular interaction of glucose binding to the apoenzyme E* to form E*⅐S by an increase in the k 1 /k Ϫ1 ratio of 6 -9-fold, whereas W99R, Y214A, and V455M mainly enhance the enzyme conformational changes from E*⅐S to the catalytically competent E⅐S form by a moderate increase in the k 2 /k Ϫ2 ratio of 1.7-2.5-fold. Y214C has the most significant effect on k 2 /k Ϫ2 (7.5-fold relative to WT) and also moderate effect on k 1 /k Ϫ1 (3-fold relative to WT).
The different effects of the mutations on the two steps of glucose binding also reflect the protein-protein interaction with GKRP. One of the interesting observations here is the correlation between the k 1 /k Ϫ1 ratio in the transient kinetics of glucose binding and the K i for GKRP as shown in Fig. 2. GKRP binds to the super open apoenzyme form competitively with respect to glucose. Because the k 1 /k Ϫ1 ratio describes the initial binding of glucose to the apoenzyme, a large increase in the k 1 /k Ϫ1 ratio suggests that these mutations (T65I, Y215A, and A456V) adopt a more compact conformation in the apoenzyme, which precludes GK-GKRP interaction. Mutations with small changes in the k 1 /k Ϫ1 ratio (W99R, V455M, and Y214A) show minimal effects on GKRP interaction. Conversely, no such correlations between the slow conformational change step (k 2 /k Ϫ2 ) and K i were observed.
Implications of Structure-Function Correlation in GK-The conformational changes in GK induced by glucose binding mainly involve the small domain through the rearrangement of the ␣13 helix and the connecting region between the large domain and small domain (Fig. 3, A and B). The domain closure forms the allosteric site and causes GKRP to no longer bind to One interesting mutation that has the highest glucose binding affinity is T65I. It increases the onrate k 1 for glucose binding to the apoenzyme by nearly 20-fold and decreases the GKRP binding affinity by 40-fold. In addition, this mutation responds to further activation by GK activators, unlike most of the other activating mutations. 3 T65I is part of a highly flexible loop region that undergoes a substantial rearrangement going from the super open to the closed form (Fig. 3). The crystal structures suggest that the side chains of this loop region can adopt different spatial orientations in the presence of various small molecule GK activators, suggesting a highly dynamic area (14,15). It is reasonable to hypothesize that the T65I mutation causes the protein to adopt a closed conformation that mimics the activated form of GK readily for glucose and activator binding. Further understanding of structure-function relationship in T65I would be facilitated by the protein crystal structures.
The ␣5 and ␣13 helices are structurally important in the allosteric regulation of GK (24). Val-455 and Ala-456 are adjacent residues located in the ␣13 helix. Although the V455M mutation retains interaction with GKRP, A456V nearly abolishes GKRP binding. This is probably because of the difference in the spatial location of Val-455 versus Ala-456 (Fig. 3C). Although Val-455 is buried in a hydrophobic pocket that includes Trp-99, Val-101, Ile-211, Tyr-215, and Leu-451, the side chain of Ala-456 is mainly solvent-exposed and accessible for protein interaction and may therefore play an important role in GKRP interaction. Y214A and Y215A show different effects on GKRP binding where 3 V. V. Heredia and S. Sun, unpublished results. FIGURE 2. Relationship of transient kinetics of glucose binding (k 1 /k ؊1 and k 2 /k ؊2 ) and GKRP inhibition constant (K i ) for GK-activating mutations. A correlation between the association constant of binding to the apoenzyme form (k 1 /k Ϫ1 ) and K i was observed (r 2 ϭ 0.8) (A) but not between the degree of conversion from E*⅐S to E⅐S (k 2 /k Ϫ2 ) and K i (r 2 ϭ 0.1) (B). Because the K i for Y215A is Ͼ5000, a value of 5000 was utilized for this fitting (depicted as an open square). Both plots were double log plots.  A and B show the overall protein conformation of GK, where A depicts the apoform and B depicts the co-crystal form with activator (white) and glucose (yellow) bound. The chemical formula of the small molecule activator was described in the previous publication as N-thiazol-2-yl-2-amino-4-fluoro-5-(1-methylimidazol-2-yl)thiobenzamide (15). Specific regions are highlighted, ␣5-helix in red, ␣13-helix in blue, and the flexible loop region in purple. The residues that were investigated cluster in the allosteric region and are depicted in Corey-Pauling-Koltun. Thr-65 is located within the flexible loop region (purple), Tyr-214 and -215 are located within the ␣5-helix (red), Val-455 and Ala-456 are located within the ␣13-helix (blue), and Trp-99 is located within the allosteric region (green). C and D show a detailed view of the spatial orientation of the residues investigated, where C is the apoform of GK and D is the GK complex with the activator (white) and glucose. In the apoform C, Thr-65 and Ala-456 are solvent-exposed. Trp-99, Tyr-215, and Val-455 are part of a hydrophobic patch. The side chains of Tyr-214 and -215 adopt different geometric orientations in the apoform. The local environment of the activating mutations changes dramatically in the activator-bound form of GK, D, as most of the critical contacts utilize interactions with the activator, Thr-65 shifts to form the flexible loop region that wraps around the compound, Tyr-215 forms a H bond to the activator, Tyr-214 provides van der Waals interactions with the activator, and both Val-455 and Ala-456 shift as part of the ␣13-helix to form the allosteric site.
Y215A disrupts GKRP binding, whereas Y214A has little effect. Tyr-215 and -214 are located in the ␣5 helix that forms the allosteric site and directly interact with GK activators (15). Tyr-215 sits in a hydrophobic patch and forms a hydrogen bond with Gly-72 in the apoenzyme form but loses most of its hydrophobic interactions in the closed form of the enzyme, whereas Tyr-214 is located distal from the hydrophobic patch where Tyr-215 is embedded. The hydrophobic interaction of Y215 may stabilize the super open apoform, which is thought to bind to GKRP. The Y215A mutation may therefore destabilize the apoform and prevent GKRP binding.
The human GK-activating mutations all show decreased fasting blood glucose in vivo and increased glucose binding affinity in vitro. The kinetic basis for the increased glucose binding differs among mutations as a result of their structural perturbation at the allosteric site. The first step of GK activation, formation of the loosely bound E*⅐S complex as described by k 1 /k Ϫ1 , is shown to correlate with the level of GKRP inhibition. This set of activating mutations provides insight regarding the interrelation of kinetic parameters of GK, GKRP binding, and the resulting clinical effects of altering the glucose sensor activity of GK.