- Email: [email protected]
The regulatory protein of liver glucokinase
Advan. Enzyme Regul., Vol. 32, pp. 133-148, 1992 Printed in Great Britain. All rights reserved
THE
0065-2571/92/$15.00 ~) 1992 Pergamon Press plc.
REGULATORY PROTEIN LIVER GLUCOKINASE
OF
E. VAN SCHAFTINGEN, A. VANDERCAMMEN, M. DETHEUX and D. R. DAVIES* Laboratory of Physiological Chemistry, International Institute of Cellular and Molecular Pathology and Universit6 Catholique de Louvain, Brussels, Belgium
INTRODUCTION
Four different hexokinases (EC 2.7.1.1) catalyze the phosphorylation of glucose in animal tissues. Hexokinase I, II and III display a high affinity for glucose (K m < 0.2 mM) and are inhibited by concentrations of glucose 6-phosphate in the physiological range (reviewed in 1). Their activities can therefore be controlled by the rate of the pathways utilizing glucose 6-phosphate. The fourth hexokinase is usually termed glucokinase (reviewed in 2). This name is somewhat improper since hexokinase IV also catalyzes the phosphorylation of other sugars (mannose, 2-deoxy-glucose), but it remains popular because of the distinctive kinetic properties of this enzyme. Glucokinase displays a sigmoidal saturation curve for glucose (3), and its K m is about 10 mM, a value slightly above the concentration of glucose in the plasma of normal subjects. Furthermore, glucokinase is not inhibited by micromolar concentrations of glucose 6-phosphate and is about half the size of other mammalian hexokinases (55 versus 100 kDa). Hence it is generally accepted that the activity of this enzyme is controlled on a short-term basis by the concentration of glucose. Accordingly, the localization of glucokinase is restricted to the liver parenchymal cells (2) and the pancreatic 13-cells (4, 5), two cell types known to "sense" the glucose concentration. Because glucokinase is a monomeric enzyme with, apparently, only one binding site for glucose, the sigmoidicity of its saturation curve for glucose has intrigued kineticists. This peculiarity has been explained by assuming that the enzyme exists under two conformations with different affinities for glucose (6). After each catalytic cycle, the enzyme is in the high-affinity form for a very short time and, unless it initiates a new catalytic cycle, it rapidly returns to the low-affinity form. This is as if glucokinase can "remember" for some time that it has converted glucose to glucose 6-phosphate. Hence *Present address: Royal Holloway and Bedford New College, Egham Hill, Egham, Surrey, England. 133
134
E. VAN SCHAFTINGEN,et al.
the term "mnemonical" (from the Greek memory) which is used for this type of model (7). Another kinetic peculiarity of glucokinase is that it is inhibited by micromolar concentrations of long chain acyl-CoAs (8), which act under their free form (9). This effect is reversible, and hence is not due to denaturation. However, at present, the physiological meaning of this inhibition is unknown. Furthermore, there is no indication that this inhibition occurs in v i v o (10). The cDNAs and the genes encoding rat liver and islet glucokinase have been cloned (reviewed in 11). The glucokinase gene comprises two different promoters, separated by at least 10 kbp. The furthest upstream of the two promoters is used in the pancreas and the other in the liver and the two resulting mRNAs differ at their 5' end. The sequence of glucokinase is homologous to that of other hexokinases from mammalian tissues and from yeast, suggesting that all glucose phosphorylating enzymes evolved from one single ancestral gene (12). The deduced sequence of the predominant islet enzyme differs by less than 10 aa at the N terminus from the liver enzyme. It is not known if this difference has any effect on the physiological function of glucokinase. SHORT-TERM REGULATION OF GLUCOKINASE ACTIVITY IN ISOLATED HEPATOCYTES In 1978, Bontemps and coworkers (13) showed that the rate of glucose phosphorylation, as measured by the release of tritiated water from [2-3H] glucose, increases when freshly isolated hepatocytes are incubated in a K+-rich medium. This effect is due to an apparent increase in the affinity of glucokinase for glucose. One year later, Clark and coworkers (14) observed that glucose phosphorylation could also be stimulated by addition of 2.5 mM fructose to the cell suspension. Neither of these effects could be explained by the known kinetic properties of glucokinase. When we reinvestigated the effect of fructose, we found that this effect also consisted of an increase in the apparent affinity of glucokinase for its substrate, the K m for glucose decreasing from 15-20 mM to 9-10 mM (Fig. 1). Furthermore, fructose was effective at very low concentrations, a half-maximal effect being reached at ~ 50/xM, well below the concentrations needed to stimulate gluconeogenesis or cause ATP depletion. The effect of fructose was almost maximal 5 min after its addition and it was lost when the cells were washed free of the ketose. This indicated that the increase in glucokinase activity was not due to synthesis of new protein but to a more rapid process, as e.g., an allosteric effect or a covalent modification (10). We also found that sorbitol, at similar concentrations to fructose,
SHORT-TERM R E G U L A T I O N OF GLUCOKINASE
135
10
~ o
8 O.
+ 0.2 mM fructose
6
~E ~_c aE
4
o
E
2 .
0 0
10
20
,
3O
[Glucose] (mM) FIG. 1. Effect of fructose on the rate of detritiation of [2-3H]glucose in isolated hepatocytes. Hepatocytes isolated from an overnight-fasted rat were incubated in the presence of the indicated concentrations of glucose. After 15 min, 0.2 mM fructose was added and [2-3H]glucose, 5 min later. The incubation was arrested 20 rain after the addition of radioactive glucose. From (10) with permission.
Glucose /
/ - - 41~"
" ~/ - - " ~ I / Glucose-6-P t
Fructose
+
,L
Fructose-l-P ~"
D-Glyceraldehyde
-
I
~Fructose-6-P
t
Fructose-1,6-P 2
Dihydroxyacetone-P~l~l~D-Glyceraldehyde-3-P
¢ ¢ ¢
Pyruvate FIG. 2. Metabolism of fructose and D-glyceraldehyde in the liver.
136
E. VAN SCHAFTINGEN, et al.
and D-glyceraldehyde at about 10-fold higher concentrations, stimulated detritiation. In contrast, neither dihydroxyacetone nor glycerol had a positive effect on glucose phosphorylation (10). Thus, the stimulatory effect was mediated by a metabolite common to fructose, sorbitol and D-glyceraldehyde. As can be deduced from Figure 2, the best candidate was fructose 1-phosphate, an intermediate of fructose and sorbitol metabolism, which can also be formed from D-glyceraldehyde and dihydroxyacetone-phosphate by reversal of the reaction catalyzed by aldolase. DISCOVERY
OF THE REGULATORY
PROTEIN
Fructose 1-phosphate was not a known effector of glucokinase. We found however that it stimulated the phosphorylation of 5 mM glucose in liver high-speed supernatants or in a poly(ethyleneglycol) fraction from liver, although it was without effect on glucokinase when this enzyme had been
A
2
A
~ o
2~
/4, cokioose
C
3 ~ ~(..)
ll) m
Q)
B 100
>,
~- 0 ¢'~ U
~ ~.~_g
I
I
5
I0
I
I
l
15 20 25 Froction number
I
I
30
35
/.0
FIG. 3. Separation of glucokinase from a fructose 1-phosphate sensitive factor. A liver fraction prepared by precipitation with poly(ethyleneglycol) was applied to an anion exchange column, which was developed with a KCI gradient. The upper panel shows the activity of glucokinase measured at 50 mM glucose. The lower panel shows measurements of the activity of purified glucokinase in the presence of 25/LI of the fractions with or without 1 mM fructose l-phosphate. From (15) with permission.
SHORT-TERMREGULATIONOF GLUCOKINASE
137
purified by anion-exchange chromatography (15). These results indicated that, to exert its effect, fructose 1-phosphate presumably required a cofactor that was separated from glucokinase during ion-exchange chromatography. We found indeed that the fractions eluted from DEAE-Trisacryl by about 80 mM KCI contained an inhibitor of glucokinase whose effect was cancelled by fructose 1-phosphate (Fig. 3). The inhibitor was a protein, as indicated by its heat- and trypsin-sensitivity and by its molecular mass of ~ 60 kDa. Upon further purification, the regulatory protein lost most of its "activity" when it became free of phosphoglucoisomerase (15). Because glucose 6-phosphate accumulated in the assay that we used, this finding suggested that the regulatory protein required the presence of fructose 6-phosphate. Accordingly, the inhibitory effect of this protein was found to be largely dependent on the presence of fructose 6-phosphate, which, by itself, did not affect the kinetic properties of glucokinase. Figure 4 shows that this phosphate ester had a half-maximal effect at a concentration of about 10 p,M in the absence of fructose 1-phosphate and that the two fructose-phosphates acted competitively with each other. Other experiments showed that the regulatory protein inhibited glucokinase competitively with respect to glucose (15). The regulatory protein was extensively purified by a procedure involving
7.5
{Fru IP] (pM)
o'oII'oo
' 500
[Fru 6 P] (IJM) FIG. 4. Effects of fructose 6-phosphate and of fructose 1-phosphate on the activity of glucokinase in the presence of regulatory protein. Partially purified glucokinase was measured in the presence of 5 mM glucose, 1 mM ATPMg, 3 U/ml regulatory protein and the indicated concentrations of fructose 1-phosphate and of fructose 6-phosphate. From (15) with permission.
138
E. VAN SCHAFFINGEN, et al.
poly(ethyleneglycol) fractionation and chromatographic steps on anionexchangers, hydroxylapatite and Mono S, a cation-exchanger. In the last chromatographic step, the regulatory protein coeluted with a 62 kDa peptide, with which it was identified. Since the molecular mass of the native protein is = 60 kDa, these results indicate a monomeric structure (16). MECHANISM
OF ACTION
OF THE REGULATORY
PROTEIN
Since it is a protein, the regulator may inhibit glucokinase by a number of different mechanisms. It was indeed conceivable, particularly if one took the fructose 6-phosphate effect into consideration, that the regulatory protein inhibits glucokinase indirectly, through the generation of a low-molecular-weight inhibitor of this enzyme. Preincubation of the regulatory protein with fructose 6-phosphate and other components of the glucokinase assay followed by inactivation of the regulatory protein by heat treatment or by incubation with trypsin, did however not result in the formation of an inhibitor of glucokinase (16). The regulatory protein did not appear to cause a stable modification of glucokinase: when glucokinase was preincubated with the inhibitory protein together with fructose 6-phosphate and ATPMg and was then repurified by chromatography on anion-exchanger, it recovered its original affinity for glucose. The possibility that the regulatory protein acts by forming a complex with glucokinase was tested by measuring the apparent molecular mass of glucokinase by centrifugation in isokinetic sucrose gradient, as shown in Figure 5. In the absence of regulatory protein the mass of glucokinase was 55 kDa, in agreement with other work (17), and this value was unchanged when either fructose 6-phosphate or fructose 1-phosphate were included in the gradient buffer. It was also unchanged when glucokinase was centrifuged together with the regulatory protein, but it almost doubled in the simultaneous presence of the regulatory protein and fructose 6-phosphate, although not when fructose 1-phosphate was also included. Conversely, it was found that glucokinase caused an increase in the apparent molecular mass of the regulatory protein in the presence of fructose 6-phosphate. These results indicated that, in the presence of fructose 6-phosphate, the regulatory protein forms an heterodimer with glucokinase. The fact that the formation of this heterodimer and the inhibition of glucokinase by the regulatory protein show the same requirements for fructose 6-phosphate and are both inhibited by fructose 1-phosphate strongly suggests that the formation of the complex is in itself responsible for the inhibition (16). Since neither fructose-phosphate affects the kinetic properties of glucokinase in the absence of regulatory protein, they most likely act by binding to the latter. Accordingly we have recently found that the regulatory protein bound radiolabeled fructose 1-phosphate and sorbitol
SHORT-TERM REGULATION OF GLUCOKINASE i
I
®
,
Regulator
I I
100
-
50
-
Glucokinase
I
+ Regulator
I I
139
@
I
-
100
-
50
E E
v x
>I'-
i,
0
I
I
,
_z
I
I
I
GJucokinase
(9
"6 t~
I-0
I
I
@
I I I I I I
100
0
Glucokinase
I
+ Regulator
I
+
Fru-6-P
I
£1: < .J
~ z ~ I I Molecular mass (kDa) ( C ) I 158 C ~ ¥132
4000
®
m c
-
100
(
Glucokinase
I'
+ Regulator
I
I
v
232 ¥ ~
0
o
)
50
6000
I
I i '
I
I
-
50
E v
t
0
®
x
100
I
+ Fru-6-P O) 0 rr"
+ Fru-l-P
.J
O
2000
~.
[--,.;
13.5
I
I
I
I
I
I
0
10
20
30
0
10
J 20
I
50
0
30
FRACTION NUMBER
FIG. 5. Sedimentation of glucokinase and of the regulatory protein in sucrose gradients: effects of fructose 6-phosphate and fructose 1-phosphate. Glucokinase and the regulatory protein were loaded onto a sucrose gradient either separately (A,B) or together (D-F). The gradient mixtures contained 500/,I,M fructose 6-phosphate (E,F) and fructose 1-phosphate (F). The dashed line indicates the peak of glycerol 3-phosphate dehydrogenase (M, 68,000). (A): glucokinase; (O): regulatory protein. From (16) with permission.
6-phosphate (a potent analog of fructose 6-phosphate, see below) up to about 4 pmol/U of regulatory protein, suggesting a stoichiometry of 1 mol/mol (Vandercammen and Van Schaftingen, unpublished results). The binding of fructose 1-phosphate and of sorbitol 6-phosphate are mutually competitive. This indicates, but does not prove, that positive and negative effectors bind to a single site on the regulatory protein. Alternatively, they could bind to two distinct sites that are not accessible simultaneously, due to steric hindrance or to conformational constraints. MODEL FOR THE ACTION OF THE R E G U L A T O R Y PROTEIN The results described above provide support for the model shown in Figure 6. According to this model, the regulatory protein can exist in
140
E. VAN SCHAFTINGEN, et al. A
I
R'A "q~-----~ R' "qD~-------~ R ~ K3 K
K
E
RI K r
ER ~
K1
ERI
FIG. 6. Model for the action of the regulatoryprotein on glucokinase.The regulatoryprotein exists in two different conformationsdesignated R and R'. R can bind to glucokinase and inhibit this enzyme. I refers to compounds, which like sorbitol 6-phosphate and fructose 6-phosphate, inhibit glucokinasein the presence of regulatory protein. A refers to fructose 1-phosphate and compoundswith a similar effect. From (20) with permission. two different conformations, one (R) which can bind to glucokinase and inhibit this enzyme, and the other (R'), which cannot. Fructose 1-phosphate (symbolized by A, for activator) binds to conformation R', and fructose 6-phosphate (I, for inhibitor), to conformation R. In the absence of inhibitor and activator, the ratio R / R ' is equal to 1/15, accounting for the modest inhibition exerted by the regulatory protein in the absence of effector (15). Fructose 6-phosphate increases this ratio, thus reinforcing the inhibition exerted by the regulatory protein. Fructose 1-phosphate antagonizes the fructose 6-phosphate effect by trapping the regulatory protein in the R' conformation. The model shown in Figure 6 is reminiscent of that proposed for the regulation of the catalytic subunit of cyclic A M P dependent protein kinase by its regulatory subunit (reviewed in 18). In both cases, the free catalytic subunit is active and loses activity upon association to the regulatory subunit, the effect of ligands such as fructose 1-phosphate or cyclic AMP being to prevent the association. One difference between the two systems is that the holoenzyme is an heterotetramer in the case of cyclic AMP dependent protein kinase and an heterodimer in the case of glucokinase. Another difference is that no ligand is required to promote the association in the case of cyclic AMP dependent protein kinase. Finally, the association constant of the glucokinase/regulatory protein complex is lower than that of protein kinase, explaining why the regulatory protein easily dissociates from glucokinase following homogenization of the liver. A L L O S T E R I C NATURE OF THE INHIBITION BY THE R E G U L A T O R Y PROTEIN One characteristic of glucokinase is that it is inhibited competitively with respect to glucose not only by substrate analogs such as glucosamine or
SHORT-TERM REGULATION OF GLUCOKINASE
141
N-acetyl-glucosamine, but also by palmitoyl-CoA and other long chain acyl-CoAs, and by the regulatory protein. We have observed indeed that the regulatory protein also behaved as a fully competitive inhibitor of rat liver glucokinase, i.e., it increased the half-saturating concentration of glucose as a linear function of its concentration without affecting Vma.. It was of interest to know if the various competitive inhibitors bound to different sites. A kinetic analysis indicated that the regulatory protein A
'T,
o.st
1 0
5
,/
0 1 2 3 [ Regulotor ] (U/mr)
O.l.
2.5 1 vo - v i
•
(.umo1-1 . rain. mg protein)
3.0
~
2.8
0 1 [Sorbitot-6-P]
80-
(pM-1)
B
60, 1 ..1_
_
_1_
vo /-,0 (pmot. min-l.mg protein -1) Vi
i
0'5
1
4
lls
(jj M-1 )
[Sorbitol.-6-P] FIG. 7. Two graphical presentations of the inhibition of glucokinase by sorbitoi 6-phosphate at various concentrations of regulatory protein. The inset in A shows the plot of the intercept with the x axis versus the concentration of regulatory protein. From (20) with permission.
142
E. VAN SCHAFTINGEN, et al.
and palmitoyl-CoA inhibited synergistically with N-acetyl-glucosamine, but competitively with each other (19). These results suggest that the regulatory protein and palmitoyl-CoA bind to one single site, which is distinct from the catalytic site. Further indication for the existence of an allosteric site came from comparative studies. The purified regulatory protein from rat liver and palmitoyl-CoA have no effect on low K m hexokinases from mammalian tissues, on yeast hexokinase or on glucokinase from Bacillus stearothermophilus (a low K m enzyme, specific for glucose) but they both inhibit liver glucokinase from such distantly related species as Buffo marinus, pig and, of course, rat. KINETIC EFFECTS OFSORBITOL6-PHOSPHATE 1-PHOSPHATE
AND F R U C T O S E
We have also performed a detailed kinetic investigation of the effect of the potent inhibitor, sorbitol 6-phosphate and of fructose 1-phosphate (20). In the presence of regulatory protein, sorbitol 6-phosphate exerts an hyperbolic, partial inhibition, the degree of which increases with the concentration of regulatory protein. Plots of the reciprocal of the difference between the rates in the absence and in the presence of sorbitol 6-phosphate
40'
30'
AO5 (pM) 20
28
10. ~
0
1
2
,
,
50 [Sorbitol-6-P] (,uM}
100
FIG. 8. Effect of the concentration of regulatory protein and of sorbitol 6-phosphate on the apparent affinity for fructose 1-phosphate. The lines are theoretical relationships calculated using the equations derived from the model shown in Fig. 6. From (20) with permission.
143
SHORT-TERM REGULATION OF GLUCOKINASE
versus 1/[sorbitol 6-phosphate] are linear, and allow the calculation of the apparent affinity for sorbitol 6-phosphate. This affinity increases with the concentration of regulatory protein (Fig. 7a). Plots of the reciprocal of the difference between 1/v in the presence and in the absence of sorbitol 6-phosphate versus 1/[sorbitol 6-phosphate] are also linear and cross the x axis at a value independent of the concentration of regulatory protein (Fig. 7b). It can be demonstrated that this value is the inverse of the dissociation constant of the regulatory protein-inhibitor complex (20). Fructose 1-phosphate releases the inhibition exerted by the regulatory protein in an hyperbolic fashion. The concentration of this effector required for a half-maximal effect increases linearly with the concentration of sorbitol 6-phosphate and with the concentration of regulatory protein (Fig. 8). These results are entirely consistent with the model illustrated in Figure 6. SPECIFICITY
OF THE
REGULATORY PROTEIN TO EFFECTORS
WITH
RESPECT
Considering that fructose 6-phosphate and fructose 1-phosphate, two closely related compounds, act in an antagonistic manner, we have investigated the effect of analogs of these compounds to have some idea of the structural features that could be characteristic of the two types of effectors. Table 1 lists the compounds that were found to exert a fructose 6-phosphate-like effect. Most of them are open-chain polyol-phosphates, the most potent being sorbitol 6-phosphate. This suggests that the regulatory protein recognizes fructose 6-phosphate in its open-chain configuration. Accordingly, (tx + 13) methyl fructofuranoside 6-phosphate, a mixture of analogs of the furanose-forms of fructose 6-phosphate, did not inhibit glucokinase in the presence of regulatory protein. Since only a small T A B L E 1. C H A R A C T E R I S T I C S O F T H E E F F E C T O F SORBITOL-6-P A N D O T H E R INHIBITORS A C T I N G T H R O U G H T H E R E G U L A T O R Y P R O T E I N
Compound Sorbitol-6-P Arabitol-5-P 2-Deoxysorbitol-6-P Mannitol-l-P Fructose-6-P Arabinose-5-P
Modification compared to sorbitol-6-P
Extent of inhibition %
10.5 /zM
Ki /~M
none C1 : absent C2 : deoxy C2 : epimer C2 : keto C1 : absent; C2 : keto
70 45 65 63 63 19
1.4 2.5 3.3 5.0 6.7 25.0
5 5 11.8 12.5 20.0 30.0
Glucokinase activity was measured in the presence of 2 U/ml of regulatory protein. The extent of the inhibition and 10.s were calculated graphically from plots of l/(vo - vi) versus 1/I and the K i values, from plots of 1/(llvi - 1/vo) versus IlL From (20) with permission.
144
E. VAN SCHAFTINGEN, et aL
proportion of fructose 6-phosphate is in solution in the non-cyclic, keto form (2.5%; according to 21), this form could be intrinsically the most potent negative effector of the regulatory protein. Several compounds were found to exert a fructose 1-phosphate-like effect. All of them were negatively charged molecules: phosphate esters, Pi and arsenate. The most potent was fructose 1-phosphate, followed by structural analogs, by ribitol 5-phosphate and analogs of this pentitol phosphate, whereas compounds with no clear structural analogy, except for the presence of a highly charged group, were much less active. Several activators are normal constituents of the liver cell: ribose 5-phosphate, fructose-l,6-bisphosphate, 6-phospho-gluconate, 2 and 3-phosphoglycerate, Pi. Taking into consideration their concentration in the liver cell, it appears that Pi is most likely to be the main activator in the absence of fructose. We also found that monovalent anions antagonized the inhibition by the regulatory protein with the following order of potency: I- > Br- > NO 3- > CI- > F- > acetate. Their effect was non-competitive with respect to fructose 6-phosphate. This effect of salts indicates that the interaction between glucokinase and the regulatory protein is at least partially ionic in nature (19). As explained below, changes in the concentration of CIparticipate in the effect of a potassium-rich medium to increase the rate of glucose phosphorylation. PHYSIOLOGICAL IMPLICATIONS
Since, in the absence of fructose, the liver cell contains 10--50 ~M fructose 6-phosphate but no detectable fructose 1-phosphate (22), the regulatory protein can inhibit glucokinase. Thus, the apparent affinity of glucokinase for glucose in intact hepatocytes is lower than that of pure glucokinase. The fact that fructose and other compounds giving rise to fructose 1-phosphate restore the affinity to the expected level is in accordance with this interpretation (10). We have obtained indications that this type of regulation also occurs in vivo (23). To measure the rate of glucose phosphorylation, [2-3H] glucose was injected to anesthetized rats via the portal vein to circumvent the problems due to dilution of the tracer in the whole body. About 90% of the radioactivity remained in the liver for about 30 sec following this injection. The proportion of radioactivity accounted for by tritiated water increased linearly at a rate of about 5%/min in control animals. This value corresponds to a glucokinase flux of ~- 0.6/zmol/min/g, i.e., about 45% of the glucokinase activity that can be measured in a cell-free extract at the concentration of glucose present in the liver. Fructose, when injected via the penile vein, stimulated this rate up to about 2.5-fold in fed and overnight
SHORT-TERM R E G U L A T I O N OF GLUCOKINASE
145
starved rats, a maximal effect being observed at a dose of 50 mg/kg. Fructose was also active when administered by intragastric infusion, a significant effect being observed at as low a dose as 20 mg/kg. The ketose caused increases in the concentration of fructose 1-phosphate, which reached values known to release the inhibition exerted by the regulatory protein. These results indicate that rather low amounts of fructose in the diet (e.g., a few grams for an adult) are likely to stimulate glucose phosphorylation by the liver. The novel mode of regulation of glucokinase provides an explanation for the observation that fructose administration causes a moderate hypoglycemia in anesthetized rats or in perfused livers (23-25), and for the fact that fructose increases the conversion of radiolabeled glucose to glycogen (14, 26). What could be the physiological significance of this "collaboration" between glucose and fructose? Portal blood contains no fructose in the post-absorptive subject. Being present in most food of vegetal origin, together with glucose (with the exception of flour, a relative latecomer in nutritional history), fructose could be a signal that allows an increase in glucose uptake by the liver. MECHANISM
OF THE POTASSIUM
EFFECT
We have mentioned above that the replacement of a Na+-rich medium by a K+-rich medium causes an increase in the apparent affinity of glucokinase for glucose in isolated hepatocytes (13). We recently found that the stimulatory effect of a K÷-rich medium on the rate of glucose phosphorylation was only partially additive with the effect of fructose, suggesting that it was also due to a decrease in the inhibition exerted on glucokinase by its regulatory protein. However, in this case, the effect is not mediated by fructose 1-phosphate. It is not due either to changes in the concentrations of fructose 6-phosphate or Pi, two other effectors of the regulatory protein but it is linked to changes in cellular volume. Replacement of Na + by K + in the medium results indeed in a time- and dose-dependent increase in cell weight and in the intracellular concentration of CI-. These changes as well as the increase in the rate of glucose phosphorylation can be prevented by including in the medium 80 mM trehalose, which acts as an osmotic agent. From these and other results we concluded that the increase in the activity of glucokinase induced by a K+-rich medium is partly due to the increase in the concentration of Cl-, which is known to relieve the inhibition exerted by the regulatory protein on purified glucokinase, and partly to the decrease in the concentration of the regulatory protein by dilution (28). These two effects are presumably antagonized to some extent by a decrease in the intracellular concentration of Pi.
146
E. VAN SCHAFTINGEN, et al. THE REGULATORY
PROTEIN
IN P A N C R E A T I C
ISLETS
Glucokinase is also present in pancreatic 13-cells where it plays an important role in the control of the glycolytic flux and, consequently, in the control of insulin secretion (27). In collaboration with W. Malaisse's group, we have obtained the following indications that the regulatory protein is present also in this cell type: (1) glucokinase from islet tissue is inhibited by the regulatory protein from rat liver; (2) in islet extracts, the rate of glucose phosphorylation is stimulated by fructose 1-phosphate; (3) in intact islets, the rate of detritiation of [2-3H]glucose is stimulated by D-glyceraldehyde, which causes a significant increase in the intracellular concentration of fructose 1-phosphate (29). Fructose is however unable to stimulate the phosphorylation of glucose in islets, because the fructokinase activity is extremely low (30). In our present state of knowledge, the exact role of the regulatory protein in islets is unclear. It remains nonetheless that 13-cells have a potential means of controlling the affinity for glucose of the rate-limiting enzyme of glycolysis. SUMMARY
Fructose, sorbitol and D-glyceraldehyde stimulate the rate of glucose phosphorylation in isolated hepatocytes. This effect is mediated by fructose 1-phosphate, which releases the inhibition exerted by a regulatory protein on liver glucokinase. In the presence of fructose 6-phosphate, the regulatory protein binds to, and inhibits, liver glucokinase. Fructose 1-phosphate antagonizes this inhibition by causing dissociation of the glucokinase-regulatory protein complex. Both phosphate esters act by binding to the regulatory protein, and by presumably causing changes in its conformation. The regulatory protein behaves as a fully competitive inhibitor. It inhibits liver glucokinase from various species, and rat islet glucokinase, but has no effect on hexokinases from mammalian tissues or from yeast, or on glucokinase from microorganisms. Kinetic studies indicate that the regulatory protein binds to glucokinase at a site distinct from the catalytic site. Several phosphate esters, mainly polyol-phosphates, were found to mimick the effect of fructose 6-phosphate. The most potent is sorbitol 6-phosphate, suggesting that fructose 6-phosphate is recognized by the regulatory protein in its open-chain configuration. Other phosphate esters and Pi have a fructose 1-phosphate-like effect. The stimulatory effect of fructose on glucose phosphorylation is observed not only in isolated hepatocytes but also in the livers of anesthetized rats. This suggests that fructose could be a nutritional signal causing an increase in the hepatic glucose uptake.
SHORT-TERM REGULATION OF GLUCOKINASE
147
ACKNOWLEDGEMENTS S u p p o r t e d by the B e l g i a n F o n d s de la R e c h e r c h e Scientifique M6dicale, by the B e l g i a n P r i m e M i n i s t e r ' s Office Science Policy P r o g r a m m i n g a n d by N I H g r a n t D K 9235. E . V . S . is M a i t r e de R e c h e r c h e s of the F N R S , A . V . , a s p i r a n t U C L a n d M . D . , a fellow of the I R S I A .
REFERENCES 1. S.P. COLOWICK, The hexokinases, pp. 1-48 in The Enzymes, vol. 9 (P. BOYER, ed.) Academic Press, New York (1973). 2. S. WEINHOUSE, Regulation of glucokinase in the liver, Curr. Top. Cell. ReguL 11, 1-50 (1976). 3. A, C. STORER and A. CORNISH-BOWDEN, Kinetics of rat liver glucokinase. Cooperative interactions with glucose at physiologically significant concentrations, Biochem. J. 159, 7-14 (1976). 4. M. D. MEGLASSON and F. M. MATSCHINSKY, Pancreatic islet metabolism and regulation of insulin secretion, Diabetes Metab. Rev. 2, 163-214 (1986). 5. S. LENZEN and U. PANTEN, Signal recognition by pancreatic 13-cells, Biochem. Pharmacol. 37, 371-378 (1988). 6. A.C. STORER and A. CORNISH-BOWDEN, Kinetic evidence for a 'mnemonical' mechanism for rat liver glucokinase, Biochem. J. 165, 61--69 (1977). 7. J. RICARD, J.-C. MEUNIER and J. BUC, Regulatory behavior of monomeric enzymes. 1. The mnemonical enzyme concept, Eur. J. Biochem. 49, 195-208 (1974). 8. C. M. DAWSON and C. N. HALES, The inhibition of rat liver glucokinase by palmitoyI-CoA, Biochim. Biophys. Acta 176, 657-659 (1969). 9. P. S. TIPPE'IT and K. E. NEET, Specific inhibition of glucokinase by long chain acyl coenzymes A below the critical micelle concentration, J. Biol. Chem. 257, 12839-12845 (1982). 10. E. VAN SCHAFFINGEN and A. VANDERCAMMEN, Stimulation of glucose phosphorylation by fructose in isolated hepatocytes, Eur. J. Biochem. 179, 173-177 (1989). 11. M.A. MAGNUSON, Glucokinase gene structure. Functional implications of molecular genetic studies, Diabetes 39, 523-527 (1990). 12. R. J. MIDDLETON, Hexokinases and glucokinases. Biochem. Soc. Trans. 18, 180-183 (1990). 13. F. BONTEMPS, L. HUE and H. G. HERS, Phosphorylation of glucose in isolated rat hepatocytes. Sigmoidal kinetics explained by the activity of glucokinase alone, Biochem. J. 174, 603-611 (1978). 14. D.G. CLARK, O. H. FILSELL and D. L. TOPPING, Effects of fructose concentration on carbohydrate metabolism, heat production and substrate cycling in isolated rat hepatocytes, Biochem. J. 184, 501-507 (1979). 15. E. VAN SCHAFFINGEN, 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, 179-184 (1989). 16. A. VANDERCAMMEN and E. VAN SCHAFTINGEN, The mechanism by which rat liver glucokinase is inhibited by the regulatory protein, Eur. J. Biochem. 191, 483-489 (1990). 17. T.L. ANDREONE, R. L. PRINTZ, S. J. PILKIS, M. A. MAGNUSON and D. K. GRANNER, The amino acid sequence of rat liver glucokinase deduced from cloned cDNA, J. Biol. Chem. 264, 363-369 (1989). 18. S. J. BEEBE and J. D. CORBIN, Cyclic nucleotide-dependent protein kinases, pp. 43-111 in The Enzymes, vol. 17 (P. BOYER and E. G. KREBS, eds.) Academic Press, Orlando (1986).
148
E. VAN SCHAFI'INGEN, et al.
19. A. VANDERCAMMEN and E. VAN SCHAFTINGEN, Competitive inhibition of liver glucokinase by its regulatory protein, Eur. J. Biochem. 200, 545-551 (1991). 20. M. DETHEUX, A. VANDERCAMMEN and E. VAN SCHAFTINGEN, Effectors of the regulatory protein acting on liver glucokinase: a kinetic investigation, Eur. J. Biochem. 200, 553-561 (1991). 21. C. A. SWENSON and R. BARKER, Proportion of keto and aldehydo forms in solutions of sugars and sugar phosphates, Biochemistry 10, 3151-3154 (1971). 22. D . R . DAVIES, M. DETHEUX and E. VAN SCHAFTINGEN, Fructose 1-phosphate and the regulation of glucokinase activity in isolated hepatocytes, Eur. J. Biochem. 192, 283-289 (1990). 23. E. VAN SCHAFTINGEN and D. R. DAVIES, Fructose administration stimulates glucose phosphorylation in the livers of anaesthetized rats, F A S E B J. 5, 326-330 (1991). 24. D. L. TOPPING and P. A. MAYES, The concentrations of fructose, glucose and lactate in the splanchnic blood vessels of rats absorbing fructose, Nutr. Metab. 13, 331-338 (1971). 25. D. L. TOPPING and P. A. MAYES, Comparative effects of fructose and glucose on the lipid and carbohydrate metabolism of perfused rat liver, Brit. J. Nutr. 36, 113-126, (1976). 26. P.O. SEGLEN, Autoregulation of glycolysis, respiration, gluconeogenesis and glycogen synthesis in isolated parenchymal rat liver cells under aerobic and anaerobic conditions, Biochim. Biophys. Acta 338, 317-336 (1974). 27. F.M. MATSCHINSKY, Glucokinase as glucose sensor and metabolic signal generator in pancreatic 13-cells and hepatocytes, Diabetes 39, 647~552 (1990). 28. E. VAN SCHAFFINGEN and A. VANDERCAMMEN, Mechanism of the stimulatory effect of a potassium-rich medium on the phosphorylation of glucose in isolated hepatocytes, Eur. J. Biochem., in press (1992). 29. W.J. MALAISSE, F. MALAISSE-LAGAE, D. R. DAVIES, A. VANDERCAMMEN and E. VAN SCHAFFINGEN, Regulation of glucokinase by a fructose 1-phosphatesensitive protein in pancreatic islets, Eur. J. Biochern. 190, 539-545 (1990). 30. W. J. MALAISSE, F. MALAISSE-LAGAE, D. R. DAVIES and E. VAN SCHAFTINGEN, Presence of fructokinase in pancreatic islets, FEBS Letters 255, 175-178 (1989).
THE
0065-2571/92/$15.00 ~) 1992 Pergamon Press plc.
REGULATORY PROTEIN LIVER GLUCOKINASE
OF
E. VAN SCHAFTINGEN, A. VANDERCAMMEN, M. DETHEUX and D. R. DAVIES* Laboratory of Physiological Chemistry, International Institute of Cellular and Molecular Pathology and Universit6 Catholique de Louvain, Brussels, Belgium
INTRODUCTION
Four different hexokinases (EC 2.7.1.1) catalyze the phosphorylation of glucose in animal tissues. Hexokinase I, II and III display a high affinity for glucose (K m < 0.2 mM) and are inhibited by concentrations of glucose 6-phosphate in the physiological range (reviewed in 1). Their activities can therefore be controlled by the rate of the pathways utilizing glucose 6-phosphate. The fourth hexokinase is usually termed glucokinase (reviewed in 2). This name is somewhat improper since hexokinase IV also catalyzes the phosphorylation of other sugars (mannose, 2-deoxy-glucose), but it remains popular because of the distinctive kinetic properties of this enzyme. Glucokinase displays a sigmoidal saturation curve for glucose (3), and its K m is about 10 mM, a value slightly above the concentration of glucose in the plasma of normal subjects. Furthermore, glucokinase is not inhibited by micromolar concentrations of glucose 6-phosphate and is about half the size of other mammalian hexokinases (55 versus 100 kDa). Hence it is generally accepted that the activity of this enzyme is controlled on a short-term basis by the concentration of glucose. Accordingly, the localization of glucokinase is restricted to the liver parenchymal cells (2) and the pancreatic 13-cells (4, 5), two cell types known to "sense" the glucose concentration. Because glucokinase is a monomeric enzyme with, apparently, only one binding site for glucose, the sigmoidicity of its saturation curve for glucose has intrigued kineticists. This peculiarity has been explained by assuming that the enzyme exists under two conformations with different affinities for glucose (6). After each catalytic cycle, the enzyme is in the high-affinity form for a very short time and, unless it initiates a new catalytic cycle, it rapidly returns to the low-affinity form. This is as if glucokinase can "remember" for some time that it has converted glucose to glucose 6-phosphate. Hence *Present address: Royal Holloway and Bedford New College, Egham Hill, Egham, Surrey, England. 133
134
E. VAN SCHAFTINGEN,et al.
the term "mnemonical" (from the Greek memory) which is used for this type of model (7). Another kinetic peculiarity of glucokinase is that it is inhibited by micromolar concentrations of long chain acyl-CoAs (8), which act under their free form (9). This effect is reversible, and hence is not due to denaturation. However, at present, the physiological meaning of this inhibition is unknown. Furthermore, there is no indication that this inhibition occurs in v i v o (10). The cDNAs and the genes encoding rat liver and islet glucokinase have been cloned (reviewed in 11). The glucokinase gene comprises two different promoters, separated by at least 10 kbp. The furthest upstream of the two promoters is used in the pancreas and the other in the liver and the two resulting mRNAs differ at their 5' end. The sequence of glucokinase is homologous to that of other hexokinases from mammalian tissues and from yeast, suggesting that all glucose phosphorylating enzymes evolved from one single ancestral gene (12). The deduced sequence of the predominant islet enzyme differs by less than 10 aa at the N terminus from the liver enzyme. It is not known if this difference has any effect on the physiological function of glucokinase. SHORT-TERM REGULATION OF GLUCOKINASE ACTIVITY IN ISOLATED HEPATOCYTES In 1978, Bontemps and coworkers (13) showed that the rate of glucose phosphorylation, as measured by the release of tritiated water from [2-3H] glucose, increases when freshly isolated hepatocytes are incubated in a K+-rich medium. This effect is due to an apparent increase in the affinity of glucokinase for glucose. One year later, Clark and coworkers (14) observed that glucose phosphorylation could also be stimulated by addition of 2.5 mM fructose to the cell suspension. Neither of these effects could be explained by the known kinetic properties of glucokinase. When we reinvestigated the effect of fructose, we found that this effect also consisted of an increase in the apparent affinity of glucokinase for its substrate, the K m for glucose decreasing from 15-20 mM to 9-10 mM (Fig. 1). Furthermore, fructose was effective at very low concentrations, a half-maximal effect being reached at ~ 50/xM, well below the concentrations needed to stimulate gluconeogenesis or cause ATP depletion. The effect of fructose was almost maximal 5 min after its addition and it was lost when the cells were washed free of the ketose. This indicated that the increase in glucokinase activity was not due to synthesis of new protein but to a more rapid process, as e.g., an allosteric effect or a covalent modification (10). We also found that sorbitol, at similar concentrations to fructose,
SHORT-TERM R E G U L A T I O N OF GLUCOKINASE
135
10
~ o
8 O.
+ 0.2 mM fructose
6
~E ~_c aE
4
o
E
2 .
0 0
10
20
,
3O
[Glucose] (mM) FIG. 1. Effect of fructose on the rate of detritiation of [2-3H]glucose in isolated hepatocytes. Hepatocytes isolated from an overnight-fasted rat were incubated in the presence of the indicated concentrations of glucose. After 15 min, 0.2 mM fructose was added and [2-3H]glucose, 5 min later. The incubation was arrested 20 rain after the addition of radioactive glucose. From (10) with permission.
Glucose /
/ - - 41~"
" ~/ - - " ~ I / Glucose-6-P t
Fructose
+
,L
Fructose-l-P ~"
D-Glyceraldehyde
-
I
~Fructose-6-P
t
Fructose-1,6-P 2
Dihydroxyacetone-P~l~l~D-Glyceraldehyde-3-P
¢ ¢ ¢
Pyruvate FIG. 2. Metabolism of fructose and D-glyceraldehyde in the liver.
136
E. VAN SCHAFTINGEN, et al.
and D-glyceraldehyde at about 10-fold higher concentrations, stimulated detritiation. In contrast, neither dihydroxyacetone nor glycerol had a positive effect on glucose phosphorylation (10). Thus, the stimulatory effect was mediated by a metabolite common to fructose, sorbitol and D-glyceraldehyde. As can be deduced from Figure 2, the best candidate was fructose 1-phosphate, an intermediate of fructose and sorbitol metabolism, which can also be formed from D-glyceraldehyde and dihydroxyacetone-phosphate by reversal of the reaction catalyzed by aldolase. DISCOVERY
OF THE REGULATORY
PROTEIN
Fructose 1-phosphate was not a known effector of glucokinase. We found however that it stimulated the phosphorylation of 5 mM glucose in liver high-speed supernatants or in a poly(ethyleneglycol) fraction from liver, although it was without effect on glucokinase when this enzyme had been
A
2
A
~ o
2~
/4, cokioose
C
3 ~ ~(..)
ll) m
Q)
B 100
>,
~- 0 ¢'~ U
~ ~.~_g
I
I
5
I0
I
I
l
15 20 25 Froction number
I
I
30
35
/.0
FIG. 3. Separation of glucokinase from a fructose 1-phosphate sensitive factor. A liver fraction prepared by precipitation with poly(ethyleneglycol) was applied to an anion exchange column, which was developed with a KCI gradient. The upper panel shows the activity of glucokinase measured at 50 mM glucose. The lower panel shows measurements of the activity of purified glucokinase in the presence of 25/LI of the fractions with or without 1 mM fructose l-phosphate. From (15) with permission.
SHORT-TERMREGULATIONOF GLUCOKINASE
137
purified by anion-exchange chromatography (15). These results indicated that, to exert its effect, fructose 1-phosphate presumably required a cofactor that was separated from glucokinase during ion-exchange chromatography. We found indeed that the fractions eluted from DEAE-Trisacryl by about 80 mM KCI contained an inhibitor of glucokinase whose effect was cancelled by fructose 1-phosphate (Fig. 3). The inhibitor was a protein, as indicated by its heat- and trypsin-sensitivity and by its molecular mass of ~ 60 kDa. Upon further purification, the regulatory protein lost most of its "activity" when it became free of phosphoglucoisomerase (15). Because glucose 6-phosphate accumulated in the assay that we used, this finding suggested that the regulatory protein required the presence of fructose 6-phosphate. Accordingly, the inhibitory effect of this protein was found to be largely dependent on the presence of fructose 6-phosphate, which, by itself, did not affect the kinetic properties of glucokinase. Figure 4 shows that this phosphate ester had a half-maximal effect at a concentration of about 10 p,M in the absence of fructose 1-phosphate and that the two fructose-phosphates acted competitively with each other. Other experiments showed that the regulatory protein inhibited glucokinase competitively with respect to glucose (15). The regulatory protein was extensively purified by a procedure involving
7.5
{Fru IP] (pM)
o'oII'oo
' 500
[Fru 6 P] (IJM) FIG. 4. Effects of fructose 6-phosphate and of fructose 1-phosphate on the activity of glucokinase in the presence of regulatory protein. Partially purified glucokinase was measured in the presence of 5 mM glucose, 1 mM ATPMg, 3 U/ml regulatory protein and the indicated concentrations of fructose 1-phosphate and of fructose 6-phosphate. From (15) with permission.
138
E. VAN SCHAFFINGEN, et al.
poly(ethyleneglycol) fractionation and chromatographic steps on anionexchangers, hydroxylapatite and Mono S, a cation-exchanger. In the last chromatographic step, the regulatory protein coeluted with a 62 kDa peptide, with which it was identified. Since the molecular mass of the native protein is = 60 kDa, these results indicate a monomeric structure (16). MECHANISM
OF ACTION
OF THE REGULATORY
PROTEIN
Since it is a protein, the regulator may inhibit glucokinase by a number of different mechanisms. It was indeed conceivable, particularly if one took the fructose 6-phosphate effect into consideration, that the regulatory protein inhibits glucokinase indirectly, through the generation of a low-molecular-weight inhibitor of this enzyme. Preincubation of the regulatory protein with fructose 6-phosphate and other components of the glucokinase assay followed by inactivation of the regulatory protein by heat treatment or by incubation with trypsin, did however not result in the formation of an inhibitor of glucokinase (16). The regulatory protein did not appear to cause a stable modification of glucokinase: when glucokinase was preincubated with the inhibitory protein together with fructose 6-phosphate and ATPMg and was then repurified by chromatography on anion-exchanger, it recovered its original affinity for glucose. The possibility that the regulatory protein acts by forming a complex with glucokinase was tested by measuring the apparent molecular mass of glucokinase by centrifugation in isokinetic sucrose gradient, as shown in Figure 5. In the absence of regulatory protein the mass of glucokinase was 55 kDa, in agreement with other work (17), and this value was unchanged when either fructose 6-phosphate or fructose 1-phosphate were included in the gradient buffer. It was also unchanged when glucokinase was centrifuged together with the regulatory protein, but it almost doubled in the simultaneous presence of the regulatory protein and fructose 6-phosphate, although not when fructose 1-phosphate was also included. Conversely, it was found that glucokinase caused an increase in the apparent molecular mass of the regulatory protein in the presence of fructose 6-phosphate. These results indicated that, in the presence of fructose 6-phosphate, the regulatory protein forms an heterodimer with glucokinase. The fact that the formation of this heterodimer and the inhibition of glucokinase by the regulatory protein show the same requirements for fructose 6-phosphate and are both inhibited by fructose 1-phosphate strongly suggests that the formation of the complex is in itself responsible for the inhibition (16). Since neither fructose-phosphate affects the kinetic properties of glucokinase in the absence of regulatory protein, they most likely act by binding to the latter. Accordingly we have recently found that the regulatory protein bound radiolabeled fructose 1-phosphate and sorbitol
SHORT-TERM REGULATION OF GLUCOKINASE i
I
®
,
Regulator
I I
100
-
50
-
Glucokinase
I
+ Regulator
I I
139
@
I
-
100
-
50
E E
v x
>I'-
i,
0
I
I
,
_z
I
I
I
GJucokinase
(9
"6 t~
I-0
I
I
@
I I I I I I
100
0
Glucokinase
I
+ Regulator
I
+
Fru-6-P
I
£1: < .J
~ z ~ I I Molecular mass (kDa) ( C ) I 158 C ~ ¥132
4000
®
m c
-
100
(
Glucokinase
I'
+ Regulator
I
I
v
232 ¥ ~
0
o
)
50
6000
I
I i '
I
I
-
50
E v
t
0
®
x
100
I
+ Fru-6-P O) 0 rr"
+ Fru-l-P
.J
O
2000
~.
[--,.;
13.5
I
I
I
I
I
I
0
10
20
30
0
10
J 20
I
50
0
30
FRACTION NUMBER
FIG. 5. Sedimentation of glucokinase and of the regulatory protein in sucrose gradients: effects of fructose 6-phosphate and fructose 1-phosphate. Glucokinase and the regulatory protein were loaded onto a sucrose gradient either separately (A,B) or together (D-F). The gradient mixtures contained 500/,I,M fructose 6-phosphate (E,F) and fructose 1-phosphate (F). The dashed line indicates the peak of glycerol 3-phosphate dehydrogenase (M, 68,000). (A): glucokinase; (O): regulatory protein. From (16) with permission.
6-phosphate (a potent analog of fructose 6-phosphate, see below) up to about 4 pmol/U of regulatory protein, suggesting a stoichiometry of 1 mol/mol (Vandercammen and Van Schaftingen, unpublished results). The binding of fructose 1-phosphate and of sorbitol 6-phosphate are mutually competitive. This indicates, but does not prove, that positive and negative effectors bind to a single site on the regulatory protein. Alternatively, they could bind to two distinct sites that are not accessible simultaneously, due to steric hindrance or to conformational constraints. MODEL FOR THE ACTION OF THE R E G U L A T O R Y PROTEIN The results described above provide support for the model shown in Figure 6. According to this model, the regulatory protein can exist in
140
E. VAN SCHAFTINGEN, et al. A
I
R'A "q~-----~ R' "qD~-------~ R ~ K3 K
K
E
RI K r
ER ~
K1
ERI
FIG. 6. Model for the action of the regulatoryprotein on glucokinase.The regulatoryprotein exists in two different conformationsdesignated R and R'. R can bind to glucokinase and inhibit this enzyme. I refers to compounds, which like sorbitol 6-phosphate and fructose 6-phosphate, inhibit glucokinasein the presence of regulatory protein. A refers to fructose 1-phosphate and compoundswith a similar effect. From (20) with permission. two different conformations, one (R) which can bind to glucokinase and inhibit this enzyme, and the other (R'), which cannot. Fructose 1-phosphate (symbolized by A, for activator) binds to conformation R', and fructose 6-phosphate (I, for inhibitor), to conformation R. In the absence of inhibitor and activator, the ratio R / R ' is equal to 1/15, accounting for the modest inhibition exerted by the regulatory protein in the absence of effector (15). Fructose 6-phosphate increases this ratio, thus reinforcing the inhibition exerted by the regulatory protein. Fructose 1-phosphate antagonizes the fructose 6-phosphate effect by trapping the regulatory protein in the R' conformation. The model shown in Figure 6 is reminiscent of that proposed for the regulation of the catalytic subunit of cyclic A M P dependent protein kinase by its regulatory subunit (reviewed in 18). In both cases, the free catalytic subunit is active and loses activity upon association to the regulatory subunit, the effect of ligands such as fructose 1-phosphate or cyclic AMP being to prevent the association. One difference between the two systems is that the holoenzyme is an heterotetramer in the case of cyclic AMP dependent protein kinase and an heterodimer in the case of glucokinase. Another difference is that no ligand is required to promote the association in the case of cyclic AMP dependent protein kinase. Finally, the association constant of the glucokinase/regulatory protein complex is lower than that of protein kinase, explaining why the regulatory protein easily dissociates from glucokinase following homogenization of the liver. A L L O S T E R I C NATURE OF THE INHIBITION BY THE R E G U L A T O R Y PROTEIN One characteristic of glucokinase is that it is inhibited competitively with respect to glucose not only by substrate analogs such as glucosamine or
SHORT-TERM REGULATION OF GLUCOKINASE
141
N-acetyl-glucosamine, but also by palmitoyl-CoA and other long chain acyl-CoAs, and by the regulatory protein. We have observed indeed that the regulatory protein also behaved as a fully competitive inhibitor of rat liver glucokinase, i.e., it increased the half-saturating concentration of glucose as a linear function of its concentration without affecting Vma.. It was of interest to know if the various competitive inhibitors bound to different sites. A kinetic analysis indicated that the regulatory protein A
'T,
o.st
1 0
5
,/
0 1 2 3 [ Regulotor ] (U/mr)
O.l.
2.5 1 vo - v i
•
(.umo1-1 . rain. mg protein)
3.0
~
2.8
0 1 [Sorbitot-6-P]
80-
(pM-1)
B
60, 1 ..1_
_
_1_
vo /-,0 (pmot. min-l.mg protein -1) Vi
i
0'5
1
4
lls
(jj M-1 )
[Sorbitol.-6-P] FIG. 7. Two graphical presentations of the inhibition of glucokinase by sorbitoi 6-phosphate at various concentrations of regulatory protein. The inset in A shows the plot of the intercept with the x axis versus the concentration of regulatory protein. From (20) with permission.
142
E. VAN SCHAFTINGEN, et al.
and palmitoyl-CoA inhibited synergistically with N-acetyl-glucosamine, but competitively with each other (19). These results suggest that the regulatory protein and palmitoyl-CoA bind to one single site, which is distinct from the catalytic site. Further indication for the existence of an allosteric site came from comparative studies. The purified regulatory protein from rat liver and palmitoyl-CoA have no effect on low K m hexokinases from mammalian tissues, on yeast hexokinase or on glucokinase from Bacillus stearothermophilus (a low K m enzyme, specific for glucose) but they both inhibit liver glucokinase from such distantly related species as Buffo marinus, pig and, of course, rat. KINETIC EFFECTS OFSORBITOL6-PHOSPHATE 1-PHOSPHATE
AND F R U C T O S E
We have also performed a detailed kinetic investigation of the effect of the potent inhibitor, sorbitol 6-phosphate and of fructose 1-phosphate (20). In the presence of regulatory protein, sorbitol 6-phosphate exerts an hyperbolic, partial inhibition, the degree of which increases with the concentration of regulatory protein. Plots of the reciprocal of the difference between the rates in the absence and in the presence of sorbitol 6-phosphate
40'
30'
AO5 (pM) 20
28
10. ~
0
1
2
,
,
50 [Sorbitol-6-P] (,uM}
100
FIG. 8. Effect of the concentration of regulatory protein and of sorbitol 6-phosphate on the apparent affinity for fructose 1-phosphate. The lines are theoretical relationships calculated using the equations derived from the model shown in Fig. 6. From (20) with permission.
143
SHORT-TERM REGULATION OF GLUCOKINASE
versus 1/[sorbitol 6-phosphate] are linear, and allow the calculation of the apparent affinity for sorbitol 6-phosphate. This affinity increases with the concentration of regulatory protein (Fig. 7a). Plots of the reciprocal of the difference between 1/v in the presence and in the absence of sorbitol 6-phosphate versus 1/[sorbitol 6-phosphate] are also linear and cross the x axis at a value independent of the concentration of regulatory protein (Fig. 7b). It can be demonstrated that this value is the inverse of the dissociation constant of the regulatory protein-inhibitor complex (20). Fructose 1-phosphate releases the inhibition exerted by the regulatory protein in an hyperbolic fashion. The concentration of this effector required for a half-maximal effect increases linearly with the concentration of sorbitol 6-phosphate and with the concentration of regulatory protein (Fig. 8). These results are entirely consistent with the model illustrated in Figure 6. SPECIFICITY
OF THE
REGULATORY PROTEIN TO EFFECTORS
WITH
RESPECT
Considering that fructose 6-phosphate and fructose 1-phosphate, two closely related compounds, act in an antagonistic manner, we have investigated the effect of analogs of these compounds to have some idea of the structural features that could be characteristic of the two types of effectors. Table 1 lists the compounds that were found to exert a fructose 6-phosphate-like effect. Most of them are open-chain polyol-phosphates, the most potent being sorbitol 6-phosphate. This suggests that the regulatory protein recognizes fructose 6-phosphate in its open-chain configuration. Accordingly, (tx + 13) methyl fructofuranoside 6-phosphate, a mixture of analogs of the furanose-forms of fructose 6-phosphate, did not inhibit glucokinase in the presence of regulatory protein. Since only a small T A B L E 1. C H A R A C T E R I S T I C S O F T H E E F F E C T O F SORBITOL-6-P A N D O T H E R INHIBITORS A C T I N G T H R O U G H T H E R E G U L A T O R Y P R O T E I N
Compound Sorbitol-6-P Arabitol-5-P 2-Deoxysorbitol-6-P Mannitol-l-P Fructose-6-P Arabinose-5-P
Modification compared to sorbitol-6-P
Extent of inhibition %
10.5 /zM
Ki /~M
none C1 : absent C2 : deoxy C2 : epimer C2 : keto C1 : absent; C2 : keto
70 45 65 63 63 19
1.4 2.5 3.3 5.0 6.7 25.0
5 5 11.8 12.5 20.0 30.0
Glucokinase activity was measured in the presence of 2 U/ml of regulatory protein. The extent of the inhibition and 10.s were calculated graphically from plots of l/(vo - vi) versus 1/I and the K i values, from plots of 1/(llvi - 1/vo) versus IlL From (20) with permission.
144
E. VAN SCHAFTINGEN, et aL
proportion of fructose 6-phosphate is in solution in the non-cyclic, keto form (2.5%; according to 21), this form could be intrinsically the most potent negative effector of the regulatory protein. Several compounds were found to exert a fructose 1-phosphate-like effect. All of them were negatively charged molecules: phosphate esters, Pi and arsenate. The most potent was fructose 1-phosphate, followed by structural analogs, by ribitol 5-phosphate and analogs of this pentitol phosphate, whereas compounds with no clear structural analogy, except for the presence of a highly charged group, were much less active. Several activators are normal constituents of the liver cell: ribose 5-phosphate, fructose-l,6-bisphosphate, 6-phospho-gluconate, 2 and 3-phosphoglycerate, Pi. Taking into consideration their concentration in the liver cell, it appears that Pi is most likely to be the main activator in the absence of fructose. We also found that monovalent anions antagonized the inhibition by the regulatory protein with the following order of potency: I- > Br- > NO 3- > CI- > F- > acetate. Their effect was non-competitive with respect to fructose 6-phosphate. This effect of salts indicates that the interaction between glucokinase and the regulatory protein is at least partially ionic in nature (19). As explained below, changes in the concentration of CIparticipate in the effect of a potassium-rich medium to increase the rate of glucose phosphorylation. PHYSIOLOGICAL IMPLICATIONS
Since, in the absence of fructose, the liver cell contains 10--50 ~M fructose 6-phosphate but no detectable fructose 1-phosphate (22), the regulatory protein can inhibit glucokinase. Thus, the apparent affinity of glucokinase for glucose in intact hepatocytes is lower than that of pure glucokinase. The fact that fructose and other compounds giving rise to fructose 1-phosphate restore the affinity to the expected level is in accordance with this interpretation (10). We have obtained indications that this type of regulation also occurs in vivo (23). To measure the rate of glucose phosphorylation, [2-3H] glucose was injected to anesthetized rats via the portal vein to circumvent the problems due to dilution of the tracer in the whole body. About 90% of the radioactivity remained in the liver for about 30 sec following this injection. The proportion of radioactivity accounted for by tritiated water increased linearly at a rate of about 5%/min in control animals. This value corresponds to a glucokinase flux of ~- 0.6/zmol/min/g, i.e., about 45% of the glucokinase activity that can be measured in a cell-free extract at the concentration of glucose present in the liver. Fructose, when injected via the penile vein, stimulated this rate up to about 2.5-fold in fed and overnight
SHORT-TERM R E G U L A T I O N OF GLUCOKINASE
145
starved rats, a maximal effect being observed at a dose of 50 mg/kg. Fructose was also active when administered by intragastric infusion, a significant effect being observed at as low a dose as 20 mg/kg. The ketose caused increases in the concentration of fructose 1-phosphate, which reached values known to release the inhibition exerted by the regulatory protein. These results indicate that rather low amounts of fructose in the diet (e.g., a few grams for an adult) are likely to stimulate glucose phosphorylation by the liver. The novel mode of regulation of glucokinase provides an explanation for the observation that fructose administration causes a moderate hypoglycemia in anesthetized rats or in perfused livers (23-25), and for the fact that fructose increases the conversion of radiolabeled glucose to glycogen (14, 26). What could be the physiological significance of this "collaboration" between glucose and fructose? Portal blood contains no fructose in the post-absorptive subject. Being present in most food of vegetal origin, together with glucose (with the exception of flour, a relative latecomer in nutritional history), fructose could be a signal that allows an increase in glucose uptake by the liver. MECHANISM
OF THE POTASSIUM
EFFECT
We have mentioned above that the replacement of a Na+-rich medium by a K+-rich medium causes an increase in the apparent affinity of glucokinase for glucose in isolated hepatocytes (13). We recently found that the stimulatory effect of a K÷-rich medium on the rate of glucose phosphorylation was only partially additive with the effect of fructose, suggesting that it was also due to a decrease in the inhibition exerted on glucokinase by its regulatory protein. However, in this case, the effect is not mediated by fructose 1-phosphate. It is not due either to changes in the concentrations of fructose 6-phosphate or Pi, two other effectors of the regulatory protein but it is linked to changes in cellular volume. Replacement of Na + by K + in the medium results indeed in a time- and dose-dependent increase in cell weight and in the intracellular concentration of CI-. These changes as well as the increase in the rate of glucose phosphorylation can be prevented by including in the medium 80 mM trehalose, which acts as an osmotic agent. From these and other results we concluded that the increase in the activity of glucokinase induced by a K+-rich medium is partly due to the increase in the concentration of Cl-, which is known to relieve the inhibition exerted by the regulatory protein on purified glucokinase, and partly to the decrease in the concentration of the regulatory protein by dilution (28). These two effects are presumably antagonized to some extent by a decrease in the intracellular concentration of Pi.
146
E. VAN SCHAFTINGEN, et al. THE REGULATORY
PROTEIN
IN P A N C R E A T I C
ISLETS
Glucokinase is also present in pancreatic 13-cells where it plays an important role in the control of the glycolytic flux and, consequently, in the control of insulin secretion (27). In collaboration with W. Malaisse's group, we have obtained the following indications that the regulatory protein is present also in this cell type: (1) glucokinase from islet tissue is inhibited by the regulatory protein from rat liver; (2) in islet extracts, the rate of glucose phosphorylation is stimulated by fructose 1-phosphate; (3) in intact islets, the rate of detritiation of [2-3H]glucose is stimulated by D-glyceraldehyde, which causes a significant increase in the intracellular concentration of fructose 1-phosphate (29). Fructose is however unable to stimulate the phosphorylation of glucose in islets, because the fructokinase activity is extremely low (30). In our present state of knowledge, the exact role of the regulatory protein in islets is unclear. It remains nonetheless that 13-cells have a potential means of controlling the affinity for glucose of the rate-limiting enzyme of glycolysis. SUMMARY
Fructose, sorbitol and D-glyceraldehyde stimulate the rate of glucose phosphorylation in isolated hepatocytes. This effect is mediated by fructose 1-phosphate, which releases the inhibition exerted by a regulatory protein on liver glucokinase. In the presence of fructose 6-phosphate, the regulatory protein binds to, and inhibits, liver glucokinase. Fructose 1-phosphate antagonizes this inhibition by causing dissociation of the glucokinase-regulatory protein complex. Both phosphate esters act by binding to the regulatory protein, and by presumably causing changes in its conformation. The regulatory protein behaves as a fully competitive inhibitor. It inhibits liver glucokinase from various species, and rat islet glucokinase, but has no effect on hexokinases from mammalian tissues or from yeast, or on glucokinase from microorganisms. Kinetic studies indicate that the regulatory protein binds to glucokinase at a site distinct from the catalytic site. Several phosphate esters, mainly polyol-phosphates, were found to mimick the effect of fructose 6-phosphate. The most potent is sorbitol 6-phosphate, suggesting that fructose 6-phosphate is recognized by the regulatory protein in its open-chain configuration. Other phosphate esters and Pi have a fructose 1-phosphate-like effect. The stimulatory effect of fructose on glucose phosphorylation is observed not only in isolated hepatocytes but also in the livers of anesthetized rats. This suggests that fructose could be a nutritional signal causing an increase in the hepatic glucose uptake.
SHORT-TERM REGULATION OF GLUCOKINASE
147
ACKNOWLEDGEMENTS S u p p o r t e d by the B e l g i a n F o n d s de la R e c h e r c h e Scientifique M6dicale, by the B e l g i a n P r i m e M i n i s t e r ' s Office Science Policy P r o g r a m m i n g a n d by N I H g r a n t D K 9235. E . V . S . is M a i t r e de R e c h e r c h e s of the F N R S , A . V . , a s p i r a n t U C L a n d M . D . , a fellow of the I R S I A .
REFERENCES 1. S.P. COLOWICK, The hexokinases, pp. 1-48 in The Enzymes, vol. 9 (P. BOYER, ed.) Academic Press, New York (1973). 2. S. WEINHOUSE, Regulation of glucokinase in the liver, Curr. Top. Cell. ReguL 11, 1-50 (1976). 3. A, C. STORER and A. CORNISH-BOWDEN, Kinetics of rat liver glucokinase. Cooperative interactions with glucose at physiologically significant concentrations, Biochem. J. 159, 7-14 (1976). 4. M. D. MEGLASSON and F. M. MATSCHINSKY, Pancreatic islet metabolism and regulation of insulin secretion, Diabetes Metab. Rev. 2, 163-214 (1986). 5. S. LENZEN and U. PANTEN, Signal recognition by pancreatic 13-cells, Biochem. Pharmacol. 37, 371-378 (1988). 6. A.C. STORER and A. CORNISH-BOWDEN, Kinetic evidence for a 'mnemonical' mechanism for rat liver glucokinase, Biochem. J. 165, 61--69 (1977). 7. J. RICARD, J.-C. MEUNIER and J. BUC, Regulatory behavior of monomeric enzymes. 1. The mnemonical enzyme concept, Eur. J. Biochem. 49, 195-208 (1974). 8. C. M. DAWSON and C. N. HALES, The inhibition of rat liver glucokinase by palmitoyI-CoA, Biochim. Biophys. Acta 176, 657-659 (1969). 9. P. S. TIPPE'IT and K. E. NEET, Specific inhibition of glucokinase by long chain acyl coenzymes A below the critical micelle concentration, J. Biol. Chem. 257, 12839-12845 (1982). 10. E. VAN SCHAFFINGEN and A. VANDERCAMMEN, Stimulation of glucose phosphorylation by fructose in isolated hepatocytes, Eur. J. Biochem. 179, 173-177 (1989). 11. M.A. MAGNUSON, Glucokinase gene structure. Functional implications of molecular genetic studies, Diabetes 39, 523-527 (1990). 12. R. J. MIDDLETON, Hexokinases and glucokinases. Biochem. Soc. Trans. 18, 180-183 (1990). 13. F. BONTEMPS, L. HUE and H. G. HERS, Phosphorylation of glucose in isolated rat hepatocytes. Sigmoidal kinetics explained by the activity of glucokinase alone, Biochem. J. 174, 603-611 (1978). 14. D.G. CLARK, O. H. FILSELL and D. L. TOPPING, Effects of fructose concentration on carbohydrate metabolism, heat production and substrate cycling in isolated rat hepatocytes, Biochem. J. 184, 501-507 (1979). 15. E. VAN SCHAFFINGEN, 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, 179-184 (1989). 16. A. VANDERCAMMEN and E. VAN SCHAFTINGEN, The mechanism by which rat liver glucokinase is inhibited by the regulatory protein, Eur. J. Biochem. 191, 483-489 (1990). 17. T.L. ANDREONE, R. L. PRINTZ, S. J. PILKIS, M. A. MAGNUSON and D. K. GRANNER, The amino acid sequence of rat liver glucokinase deduced from cloned cDNA, J. Biol. Chem. 264, 363-369 (1989). 18. S. J. BEEBE and J. D. CORBIN, Cyclic nucleotide-dependent protein kinases, pp. 43-111 in The Enzymes, vol. 17 (P. BOYER and E. G. KREBS, eds.) Academic Press, Orlando (1986).
148
E. VAN SCHAFI'INGEN, et al.
19. A. VANDERCAMMEN and E. VAN SCHAFTINGEN, Competitive inhibition of liver glucokinase by its regulatory protein, Eur. J. Biochem. 200, 545-551 (1991). 20. M. DETHEUX, A. VANDERCAMMEN and E. VAN SCHAFTINGEN, Effectors of the regulatory protein acting on liver glucokinase: a kinetic investigation, Eur. J. Biochem. 200, 553-561 (1991). 21. C. A. SWENSON and R. BARKER, Proportion of keto and aldehydo forms in solutions of sugars and sugar phosphates, Biochemistry 10, 3151-3154 (1971). 22. D . R . DAVIES, M. DETHEUX and E. VAN SCHAFTINGEN, Fructose 1-phosphate and the regulation of glucokinase activity in isolated hepatocytes, Eur. J. Biochem. 192, 283-289 (1990). 23. E. VAN SCHAFTINGEN and D. R. DAVIES, Fructose administration stimulates glucose phosphorylation in the livers of anaesthetized rats, F A S E B J. 5, 326-330 (1991). 24. D. L. TOPPING and P. A. MAYES, The concentrations of fructose, glucose and lactate in the splanchnic blood vessels of rats absorbing fructose, Nutr. Metab. 13, 331-338 (1971). 25. D. L. TOPPING and P. A. MAYES, Comparative effects of fructose and glucose on the lipid and carbohydrate metabolism of perfused rat liver, Brit. J. Nutr. 36, 113-126, (1976). 26. P.O. SEGLEN, Autoregulation of glycolysis, respiration, gluconeogenesis and glycogen synthesis in isolated parenchymal rat liver cells under aerobic and anaerobic conditions, Biochim. Biophys. Acta 338, 317-336 (1974). 27. F.M. MATSCHINSKY, Glucokinase as glucose sensor and metabolic signal generator in pancreatic 13-cells and hepatocytes, Diabetes 39, 647~552 (1990). 28. E. VAN SCHAFFINGEN and A. VANDERCAMMEN, Mechanism of the stimulatory effect of a potassium-rich medium on the phosphorylation of glucose in isolated hepatocytes, Eur. J. Biochem., in press (1992). 29. W.J. MALAISSE, F. MALAISSE-LAGAE, D. R. DAVIES, A. VANDERCAMMEN and E. VAN SCHAFFINGEN, Regulation of glucokinase by a fructose 1-phosphatesensitive protein in pancreatic islets, Eur. J. Biochern. 190, 539-545 (1990). 30. W. J. MALAISSE, F. MALAISSE-LAGAE, D. R. DAVIES and E. VAN SCHAFTINGEN, Presence of fructokinase in pancreatic islets, FEBS Letters 255, 175-178 (1989).