Biochimie (1992) 74, 867-873 © Soci~t6 francaise de biochimie et biologie mol6culaire / Elsevier, Pads
867
The inhibition of bovine heart hexokinase by 2-deoxy-D-glucose-6phosphate: characterization by alp NMR and metabolic implications W Chen, M Gu6ron* Groupe de Biophysique, Ecole Polytechnique, 91128 Palaiseau, France
(Received 1 June 1992; accepted 30 July 1992)
Summary B The glucose analog, 2-deoxy-D-glucose (2DG), has been used widely for studying the initial steps in the metabolism of glucose by radio-isotope tracer methods and by 31p NMR. In the rat heart perfused with acetate/2DG (both 5 mM) plus insulin, trapping of phosphorus by 2-deoxy-D-glucose-6-phosphate (2DG6P) results in a steady state exhibiting high 2DG6P (55 mM) and low ATP concentrations but near-normal function, as observed in an earlier 31p NMR study. In order to understand how the 2DG6P concentration is stabilized, we studied the inhibition of a mammalian hexokinase by 2DG6P in vitro by a 31p NMR technique. Inhibition, previously unobserved, was found. It is similar to inhibition by G6P in that it is competitive with ATP and not competitive with 2DG, but the inhibition constant (1.4 mM) is much larger. The experimental protocol includes provisions fo~ enzymatic destruction of stray inhibitor:~ such as G6P. The results show that the high 2DG6P and low ATP concentrations found in the steady state of the peffused heart should strongly reduce the rate of phosphorylation of sugars by hexokinase. deoxy.glucose / NMR./positron emission tomography / heart Introduction Glucose derivatives in P E T and N M R metabolic studies
2-Deoxy-D-glucose (2DG), an analog of glucose, has been used widely for studying the initial steps in the metabolism of glucose. The principle of the method is that 2DG behaves similarly to glucose with respect to facilitated transport across the sarcolemma [l] and to subsequent intraceilular phosphorylation by hexokinase (EC 2.7.1.1), resulting in the formation of 2deoxy-D-glucose-6-phosphate (2DG6P). The latter cannot leave the ceil, and is metabolized very slowly. Hence its accumulation provides a measure of glucose metabolism. These properties were first investigated by biochemical methods [2]. Isotopic labeling made possible their application to functional in vivo studies of the brain by [2-t4C]deoxy-glucose autoradiography [3] and [2-1SF]-fluoro-2-deoxy-D-glucose positron emission tomography (PET) [4]. The radio-isotope method is very sensitive but is blind to the nature of the chemical compound which carries the isotope. In particular, it does not distinguish between 2DG and *Correspondence and reprints
2DG6P. Neither does it detect the degradation of 2DG6E whose rate, required for the interpretation of the measurements, must be provided independently. Nuclear magnetic resonance (NMR) provides another means for following the metabolism of 2DG. It suffers from low sensitivity but is attuned to chemical characterization. In particular, 2DG6P can be observed and titrated by 3~p NMR. In a previous study, the rate of degradation of 2DG6P in the perfused rat heart was measured and shown to be insulindependent [5 ]. Due to its low sensitivity, the NMR measurement requires the use of large concentrations of 2DG6E and these have led to the investigation of their effect on heart function [6]. It was discovered that upon perfusion with 2DG and acetate, plus insulin, massive accumulation of 2DG6P occurred, and the rat heart reached a steady state in which function is quasi-normal, despite the very low concentrations of PCr and ATP that result from the sequestration of phosphate groups by 2DG6P. These observations provide strong limitations on the role of the concentrations of PCr or ATP as metabolic regulators. In the steady state, a constant concentration of 2DG6P is maintained by equal rates of 2DG phosphorylation and 2DG6P degradation. The rate of the latter was measured directly. The rate of the fom~er, which must equal that of the latter, was eight times smaller
868 than the initial rate o f 2DG phosphorylation. This could be due to inhibition by 2 D G 6 P o f the import of 2DG. as already invoked in an early study o f the rat diaphragm [2]. It could also be due to inhibition of hexokinase. Kipnis and Cori [2], unaware of 2 D G 6 P degradation, interpreted the constant 2 D G 6 P content as evidence for the arrest o f 2 D G 6 P production; having found no inhibition of rat skeletal muscle hexokinase by 25 m M 2 D G 6 E they ascribed the arrest to blockage o f 2 D G import.
The observation o f 2 D G 6 P degradation in the case o f the perfused rat heart forces us to question this conclusion. In other systems, the observation o f large concentrations o f intracellular 2 D G (non-phosphorylated) also suggests that 2 D G p h o s p h o r y l a t i o n is not limited by import [7]. We have therefore m e a s u r e d the inhibition of a m a m m a l i a n hexokinase b y 2DG6P. Inhibition is observed, m u c h weaker dlan b y G6P, but similar to it in the sense that it is not competitive with respect to glucose and competitive with respect to A T E The observations indicate that inhibition o f hexokinase must be taken into account in the description o f the metabolic steady state o f the rat heart perfused with 2-deoxyglucose.
Materials and m e t h o d s
,.~il
, -.l.d~,~ ,
....
~
. r~
The measurement of enzyme kinetics was performed by NMR. The sensitivity is lower than with classical, UV-monitored, enzymatic methods which could be considered for this study [8], but 31p NMR is well adapted to the inhibitor and metabolite concentrations required. It has the advantage that the concentrations of all the phosphorylated products are traced individually and simultaneously, during the course of the reaction. The pH is also obtained. Lastly, the spectrum would reveal parasitic reactions. The absolute rate of the reaction, adjusted by choosing the enzyme concentration, was chosen in view of the time in which the NMR spectrum could be measured with good sensitivity, which was about 6 min. The experiments were designed to avoid artifacts caused by traces of inhibitor impurities, particularly G6P which is a strong inhibitor of hexokinase (using our definitions (eqn 2), the K~ is 20 I.uM for pig heart type I hexokinase [9]).
18-24
Reagents
Is
o
I-s
i- o
!-15
Fig 1. Phosphorylation of 2DG by hexokinase. The lowest trace (at a vertical scale of 0.5) is the spectrum of data acquired between 6 and 12 min after addition of 0.2 U of hexokinase to the reaction mixture. The other traces correspond to spectra of data taken during successive 6-min intervals as indicated, from which the lowest trace has been subtracted. The loss of ATP and the generation of 2DG6P (ca 7.14 ppm) and ADP can be observed quantitatively, with reference to the known initial ATP concentration, l0 mM. Apparent non-stoichiometry is due to saturation (see Materials and methods). The slight shift of the 2DG6P peak (by. 0.05 ppm) corresponds to an acidification by 0.06 pH units. Experimental conditions: 2DG, 10 mM; ATP, l0 mM; PGI and PFK, 3 U each; reaction volume: 1.5 ml; pH 7.1; T = 37°C. Peak assignments: l, 2DG6P; 2, yATP; 3, ~ADP; 4, ¢tADP; 5, txATP, 6, ~IATE
ATE 2DG, 2DG6P, glucose, G6P, bovine heart hexokinase (HK, type X, 25 U, lyophilized), rabbit muscle phospho-isomerase (PGI, type IV, 1000 U in 0.6 ml) and rabbit muscle phosphofructokinase (PFK, type Ill, 500 U in 0.35 ml) were bought from Sigma (France) and used without further purification. Other reagents were products of Merck, analytic grade. Hexokinase was dissolved in 2.5 ml of a 7:3 mixture of glycerol and water, and stocked at-30°C. Before use, aliquots were diluted with water to a concentration of 0.025 to 0.1 U per 20 pl. PGI and PFK were diluted with water to a concentration of I U/pl and stocked at 4°C.
Composition of the reaction mixture The reaction mixture was designed with reference to intracellular physiologic conditions as regards ionic strength, pH and temperature. It included KCI (150 mM), Tris-HC! (200 mM) and EDTA (1 mM). The pH was adjusted to 7.1. The concentration of ATP[Mg] was 2-10 mM and total magnesium was 25 mM in excess of ATE The solution also contained 3 U of PGI and PFK, as discussed below. The water solvent was 30% deuterated.
Experimental procedure The reaction was started by adding 0.025 to 0.2 U of hexokinase (a volume of 20 [al) to 1.5 ml of the reaction mixture,
869 contained in a 10 mm outer diameter (od) NMR sample tube at room temperature. The tube was immediately transferred to the spectrometer which had been set to a sample temperature of 37°C. The data acquired during the first 6 min were not used in the analysis.
.~tp NMR measurements 31p NMR spectra were obtained at 145.6 MHz, using a homebuilt spectrometer [10] equipped with a 10-mm probe. The excitation pulse angle was 60 ° (22 Its) and the repetition time 1.4 s. The block accumulation time was chosen according to the concentration of substrates of hexokinase. It varied from 6 to 24 rain. In most cases, including those displayed in the figures, the free induction decay was multiplied by an exponential-sinus function exp(-~t&)*sin[27t(0.7&+0.07)] with 5 = 10 Hz, a combination chosen for the optimization of the signal-tonoise ratio and of the resolution [ 11 ]. In a few cases, for precise comparison of 2DG6P generation and ATP consumption, we used the integral of the 2DG6P and 13ATP peaks in spectra processed by 10 Hz broadening only. The chemical shifts are referred to phosphocreatine. Quantities were determined by the area of the 13peak of ATP in subtraction spectra (fig 1). The initial spectrum, for which the concentration of ATP is known, provides a reference which requires no correction for NMR saturation. The saturation factors for all peaks were later determined by comparison with NMR signals obtained with a repetition time of 20 s. That for the [3 peak of ATP is 1.9; in retrospect, a repetition time longer than 1.4 s would have been preferable.
Kinetic fotwlalism Hexokinase catalyses the phosphorylation of glucose: ATP + G ADP + G6P + H +
(I)
Glucose-6-phosphate slows down the reaction by competitive inhibition with ATP. The ratio of the reaction rate V to the maximum rate Vma~ is equal to the fraction of enzyme molecules engaged in an active complex with both substrates. If the two substrates bind independently, one finds:
[S,I [S.,l/(gl K2)
V m
Vm.,,
Since 2DG6P is a rather poor inhibitor, one must avoid stray inhibitors of hexokinase, or eliminate them from the reaction medium. Magnesium, ATP (if not complexed with Mg), and ADP are weak inhibitors [ 12]. The first two are controlled by using a slight excess of magnesium over initial ATP_. As for ADP, an inhibiter competitive with ATP, with a Ki of ca 2 mM [ 13|, it is a product of the reaction. We could have destroyed it enzymatically; we chose instead to measure the initial rates, within a range of times such that the ADP concentration had not reached 3 mM. Another problem is the presence of glucose-6-phosphate or of glucose as impurities in the 2DG and 2DG6P reagents. The NMR peaks of the H~ proton of glucose and of glucose-6phosphate, at 5.2 ppm, fall in a spectral region devoid of peaks in the spectrum of the deoxy compounds. We failed to observe these peaks in solutions of 2DG and 2DG6E and this indicated a content of glucose and glucose-6-phosphate smaller than 0.2%. But such quantities could still be troublesome, and the reaction medium was therefore designed to destroy G6E by incorporation of the enzymes PGI and PFK. The former transforms G6P into fructose-6-phosphate (F6P), which PFK transforms quasi-irreversibly into fructose 1,6-diphosphate.
Results Kinetics o f 2DG phosphorylation The results o f a control experiment are shown in figure 1. U p o n addition o f hexokinase to the reaction volume, production o f 2 D G 6 P and A D P begins, together with loss o f ATP. The rate o f the reaction can be measured directly by the variation o f the 2 D G 6 P peak, o f the two A D P peaks and o f the three ATP peaks. Taking into account the saturation factors (see Materials and methods), the variations o f the different phosphorus moieties are consistent with the stoichiometry of equation I.
Reactions o f the ancil l a o ' enzymes P G ! and P F K (2)
(1 + [l]/g,) (1 + [S~i/g~) + [S..,I/K., + [S~I[S,.IItKtK,.)
where S~ and $2 are the two substrates with Michaelis constant K~ and K,., and I is the inhibitor whose dissociaf,~,n constant is K., and which is competitive with $2. It is assumed that simultaneous binding of $2 and I is negligible. We shall see that the model can describe the observations, but it does not pretend to represent the enzymatic mechanism. The equation may be rewritten as:
[Sd IS.,] V m J V = (Kl + [Sd)(K2( l + [ l l / K i ) + [S-,])
Protection against sn'ay inhibitors
(3)
It predicts that the plot of [Sd/V vs lSd is a straight line which cuts the x-axis at -KI, independently of [I1 or [Sz]. The plot of [S2]/V vs IS_,] should also be a straight line, cutting the x-axis at -K2(I + [I]/Ki), and whose slope, K~ + S~, is independent of [I1. This model is known to account for the inhibition of glucose phosphorylation by G6P (I), which is non-competitive with glucose (S:) but competitive with ATP (S,) |9]. The measurements of the present study were also compared to this model.
These e n z y m e s are included so as to avoid inhibition o f the H K reaction by glucose-6-phosphate, which they degrade irreversibly. The degradation o f G 6 P was checked by c o m p a r i n g the rates o f production of 2 D G 6 P either without or with added G 6 F (40 lalVl). The results are shown in table I. Without addition o f 2 D G 6 P or G6P, the rates without (rate_) and with (rate+) PGI and P F K are equal, indicating the absence o f G 6 P in the reagents, including 2DG. The second line shows the inhibition by G6P: according to equation 3, the difference between (1/rate_) and (1/rate+) is proportional to I/K~, where I stands for the inhibitor destroyed by the ancillary e n z y m e combination. The recovery of the rate, up to the control value, upon addition o f PGI and P K F (rate+) shows the effectiveness of this procedure for the elimination o f G6P. The third line shows the inhibition by 2 D G 6 P and conta-
870 Table i. Kinetics of 2DG phosphorylation a. 2DG6P (mM)
G6P (mM)
Control
IlK (units)
Rate_ b
Rate. b
0.2
1.1 c
1.1
0.04
0.2
0.54
1.0
-
G6P 2DG6P
10
-
0.2
0.37
0.50
2DG6P + G6P
10
0.04
0.2
0.29
0.55
aATP and 2DG concentrations: 10 mM; brates are in (lanol of 2DG6P)/(HK unit.min). For rate_, the measurement is made without PGI or PFK; for rate., the solution contains 3 units of each; Cthe relative error on the rates is estimated at + 20%.
bined effect of G6P and 2DG6P. Assuming that G6P is the contaminant in 2DG6P, and that 2 D G 6 P and G6P are competitive, so that their effects on (1/ rate) are additive, then the observed rates correspond to 0.3% contamination (30 pM G 6 P for 10 m M 2DG6P), slightly above the limit o f observability (0.2%) of the experiments designed to search for such impurities, as described in Materials a n d methods.
,J
20
.is
.=.
E
10
r..--i
24-36 i
, ~
-2
,~,
,
0
I
,
~
I
2
I
,
,
4
I
,
,
,
6
II
8
,
,
,
I
,
,
10
[2DG] m M . . . .
io
?5
?lo
1-15
. . . .
I
. . . .
I
~ , , I , , , , I , , , , I
0
5
I
-20
-10
0
10
. . . .
I ' ' ' ' 1 ' ' ' '
. . . .
I = = , , I , , , ,
20
1
30
[2DG6P] mM Fig 2. Phosphorylation of 2DG by hexokinase in the presence of 2DG6P. Same conditions as in figure 1, except that the initial reaction mixture contains 30 mM 2DG6P. The difference spectrum is taken over a 12-min interval and scaled to 0.5, so that its peak iN~ensities are directly comparable to those of figure !. The rate of the reaction is much reduced. The slight pH change is conspicuous in the spectral shift of the 2DG6P peak in the difference spectrum. The inset shows the 2DG6P difference peak obtained after compensation of the spectral shift. minants (rate_). The partial suppression of the inhibition by PGI and PFK (rate.) is ascribed to the destruction of contaminants. The last line shows the com-
Fig 3. A. Plot of [2DG]/V vs [2DG], for different concentrations of 2DG6P: A, 0 mM; ×, 6 raM; o, 10 mM, ~, 20 mM; ~ , 30 raM. The conditions of the experiments are those of figure I. Straight lines converging to a point on the x-axis correspond to an inhibition by 2DG6P which is not competitive with 2DG, and -K2u3 is then the common intercept (eqn 3). For each concentration of 2DG6P, a straight line is traced according to the best visual fit. The relative errors for 30 mM 2DG6P are large because of the low rate (strong inhibition)" these data are ignored in the evaluation of K2~. B. The relative slopes of the plots in A are plotted vs inhibitor concentration. The straight line extrapolates to an intercept of-Ki(l + [ATP]/KAw). The point at 30 mM is not used for the tracing of the straight line. For symbols see A.
871 Side effects of the ancillary enzymes were searched for in a reaction mixture containing ATP and 2DG6P (10 mM each), HK (0.2 U), and prepared as described except that 2DG was omitted. After 30 min, the 3~p NMR spectrum displayed no peaks of ADP or other products, and no degradation of 2DG6P was observed. At that time 2DG was added. The initial rate of phosphorylation was the same as in an experiment in which the initial delay was omitted. Hence no product affecting the kinetics of HK had been generated during the delay.
2o ° ~
lO
Inhibition of 2DG phosphorylation by 2DG6P Figure 2 shows a typical measurement of the HK kinetics in the presence of the inhibitor 2DG6P. Comparison with figure 1 shows the reduction of the rate of 2DG phosphorylation. A small acidification (0.06 pH units) which occurs during the reaction produces a slight shift in frequency of the 2DG6P peak [14], which the subtraction procedure makes conspicuous. Whereas a full characterization of the enzyme involved rate measurements versus A T E 2 D G and 2 D G 6 P concentrations, we studied the H K reaction only at constant A T P (I0 raM) or at constant 2 D G (I0 raM). The rates in the various conditions are shown in figures 3 A and 4A. The variation of (substrate concentration/rate) vs substrate concentration indicates that the 2DG6P inhibitor is not competitive with 2DG (converging straight lines), and is competitive with ATP (parallel straight lines). Figures 3B and 4B show that the measurements are compatible with the linear dependence on the inhibitor concentration which is expected from equation 3. The intercepts with the xaxis provide the constants of the model: K2ua from figure 3A; /('^TP from the line measured without inhibitor in figure 4A; Ki from figure 4B. The inhibition constant for G6P is also obtained, assuming that G6P and 2DG6P react in the same fashion with the enzyme. The results are collected in table II. Data for the glucose-G6P system [9] are shown for comparison. A second value of Ki is provided by the intercept in figure 3B, which is predicted at Ki(I+[ATP]/KA~). The two values of Ki are 1.7 and 1.2 mM respectively. The difference is in the range of experimental errors and does not warrant a reconsideration of the model of equation 3. Discussion
Inhibition by 2DG6P Our results demonstrate that 2DG6P inhibits the phosphorylation of 2DG by bovine heart hexokinase. The
o -2
o
2
4
6
8
lO
[ATP] mM ...............................
10 8
!.
6
2 i
-5
i
0
10
20
i
,
i
30
[2DG6P] mM Fig 4. A. Plot of [ATP]/V vs [ATP], for different concentrations of 2DG6P. According to equation 3, parallel straight lines correspond to inhibition by 2DG6P in competition with ATE The common slope is then proport;onal to 1 + K2uJ[2DG]. The intercept in the absence of inhibitor determines KATp.B. Plot of the y-axis intercept in A vs inhibitor concentration. The straight line extrapolates to an x-axis intercept of-Ki, which is thus determined. Same conditions as in figure 1. For symbols see legend to figure 3.
result is genuine, and not an artifact due to contamination by the potent inhibitor G6P which we have eliminated by the procedures described above. The inhibition is competitiye with respect to ATP and not competitive with respect to 2DG. The inhibition constant (Ki = 1.45"(1 _+ 0.3) mM is about 250 times larger than that de:rived for G6P (first line of table II). In combination with the competition by ATE this may explain that inhibition of 2DG phosphorylation by 2DG6P wa,; not observed in earlier studies [2, 15].
872 The inhibition measured here probably extends to that of glucose phosphorylation, since HK phosphorylates these two sugars in similar fashion and at similar rates [161. The similarity of 2DG6P and G6P regarding competition with substrates suggests that the inhibition mechanism may be the same in both cases, namely allosteric inhibition by binding to a second site of the pseudo-dimeric mammalian enzyme [15]. This assumption is supported by our unpublished observation that 2DG6P does not inhibit 2DG phosphorylation by the monomeric yeast hexokinase, again like G6P [ 171.
Implications for the metabolism of 2-deoxy-glucose and glucose Knowing the rate parameters of phosphorylation of 2DG by HK makes it possible to investigate whether inhibition of the hexokinase reaction by 2DG6P contributes to the establishment of the low-ATE lowPCr, steady-state of the perfused rat heart mentioned in the Introduction. The hearts were studied under four different conditions [5]: with and without insulin, with glucose or with acetate perfusion. When glucose is the carbon source, its phosphorylation (and concomitantly that of 2DG) is regulated by the work output, and the regulation involves inhibition of hexokinase by G6P. (As a result, the phosphorylation rate is hardly affected by insulin [5, 18].) There is no such regulation when acetate is the carbon source, therefore the rate of 2DG phosphorylation is larger. The effect of insulin shows that in its absence, 2DG import is a limiting factor for phosphorylation [19]. But in its presence, hexokinase may become limiting, the more so when the concentration of the inhibitor 2DG6P is large and that of ATP is low, conditions which are those of the steady state. Can this explain the reduction by a factor of eight in the rate of phosphorylation in the steady state, with respect to the initial rate, where the concentration of 2DG6P is negligible?
We compute the rates by equation 3, using the appropriate concentrations of the different products. The intracellular 2DG concentration is assumed to be equal to that of the perfusate (5 mM), and therefore remains constant. With $2 standing for ATP and I for 2DG6E and with subscripts • for the initial state and • , for the steady state, we have: (l + [I**]/Ki)K2/[S2** ] + 1 V:~/V**
.-
([I**I/Ki)(K2/[S2**I) (4)
K2][S2.I + 1
where the last expression is obtained by keeping only the largest terms. The initial ATP concentration is 18 mM. The steady-state value is affected by a large relative error because of its low value, and of the possibility of a contribution of mitochondrial ATP to the observed signal. Based on figure 1 of [5], we take a value of 4.5 mM, but a value twice as small could be chosen on the basis of similar measurements [6]. The concentration of 2DG6E initially zero, reaches 55 mM in the steady state. We note that the last expression in equation 4 involves a ratio of concentrations and is therefore independent of the cytosolic volume (1.4 ml/(g protein) [6, 20]) used to compute them. With the values of table II, the ratio of rates is 7.5. It would be 15 if one took the smaller value mentioned for the steady-state ATP concentration. These values are in the range of the in vivo ratio of 8. Hence, inhibition of hexokinase by 2DG6P may be responsible for the slow phosphorylation in the steady state. At the very least, a strong iohibition of hexokinase is expected in the steady state of the acetate/2DG/insulin peffusion experiment. On the other hand, inhibition of hexokinase cannot be responsible for the steady state which is also observed in the absence of insulin. In this case, the steady-state concentration of 2DG6P is only 20 mM whereas that of ATP is close to the initial one. The ratio of initial to steady-state rates is again about eight, whereas the first expression in equation 4 predicts a ratio of 1.6. Hence other mechanisms must be
Table il. Mich,~elis and inhibition constants of the hexokinase reaction a.
2DG --->2DG6P G ---->G6~
~o~ ~
K~
~T~
g,,.~o~6~,
/t',,~,
0.6
-
0.8
1.45c
0.006d
-
0.02
0,4
-
0.020
apH 7.1, T = 37°C. bAll constants are in mM; the relative error, gauged from the spectral signal-to-noise ratio and from the dispersion in the plots, is approximately + 30%. CAverage of the values, 1.7 and 1.2, computed from figures 4B and 3B, respectively. dThe oata in lines 2 and 3 of table I show that llKi is the same for 40 l.tM of G6P and for 10 mM of 2DG6E Ilence K,c~p, = K,2~6p]250. eThese data, from [9], are for pig heart type I hexokinase; pH 7.6; T = 30°C; no added salt.
873
at work. One element is that 2DG6P may interfere with sugar import, as suggested by reports of anomalies in the intracellular concentration of glucose (nonphosphorylated) in the presence of sugars such as 2DG which are phosphorylated to products which accumulate in the cell [21, 22]. In summary, we have demonstrated the inhibition of a mammalian hexokinase by 2-deoxy-o-glucose-6phosphate, and determined the inhibition coefficient. Inhibition by 2DG6P is competitive v s ATP and not v s 2DG. Inhibition of hexokinase by 2DG6P must be taken into account in the interpretation of the 2DG metabolism, and it provides at least a partial explanatiot~ for the steady state of the rat heart perfused with acetate/2DG/insulin. Acknowledgment
8
9 10 11 12 13
14
WC was supported by Association fran~aise contre ies Myopathies.
15
References
16
1 Takala TES, Hassinen IE (1981) Effect of mechanical work load on the transmural distribution of glucose uptake in the isolated perfused rat heart, studied by regional deoxyglucose trapping. Circ Res 49, 62--69 2 Kipnis DM, Cori CF (1958) Studies of tissue permeability V. The penetration and phosphorylation of 2-deoxy-glucose ill the rat diaphragm. J Biol Chem 234, 171-177 3 Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, Pettigrew KD, Sakurada O, Shirohara M (1977) The [t4Cldeoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure and normal values in the conscious anesthetized albino rat. J Neurochem 28, 897-916 4. Phelps ME, Hoffman El, Selin C, Huang SC, Robinson G, MacDonald N, Schelbert H, Kuhl DE (1978) Investigation of [IXFl2-fluoro-2-deoxyglucose for the measure of myocardial glucose metabolism. J Nucl Med 19, 1311-1319 5 Hoerter JA, Dorrnont D, Girault M, Gu6ron M, Syrota A (1991) 31p NMR evidence for insulin-dependent 2-deoxyglucose metabolism in the peffused rat heart. J Mol Cell Cardio123, 1101-1 i 15 6 Hoerter JA, Lauer C, Vassort G, Gu~ron M (1988) Sustained function of normoxic hearts depleted in ATP and phosphocreatine: a 31P-NMR study. Am J Physio1255, C ! 92---C201 7 Jacobs AEM, Oosterhof A, Veerkamp JH (1990) 2-DeoxyD-glucose uptake in cultured human muscle cells. Biochim Biophys Acta 1051, 230-236
17
18
19
20
21
22
Wilson JE, Chung V (1989) Rat brain hexokinase: further studies on the specificity of the hexose and hexose 6phosphate binding sites. Arch Biochem Biophys 269, 5 i 7-525 Easlerby JS, O'Brien MJ (1973) Purification and properdes of pig heart hexokinase. Eur J Biochem 38, 210211 Caron F, Gu6ron M, Nguyen Ngoc Quoc Thuy, Herzog RF (1980) A high-resolution multipurpose Fir" NMR spectrometer. Revue Phys Appl 15, 1267-1274 Gu6ron, M ¢1978) Line-narrowing and line-broadening using trigonometric functicms. J Magn Res 36, 515-520 Bachelard HS (1971) Allosteric activation of brain hexokinase by magnesium ions and by magnesium ion-adenosine triphosphate complex. Biochem J 125, 249-54 Rees BB, Ropson IJ, Hand SC (1989) Kinetic properties of hexokinase under near-physiologic condition relation to metabolic arrest in artemia embryos during anoxia. J Biol Chem 264, 15410-15417 Bailey IA, Williams SR, Radda GK, Gadian DG (1981) Activity of phosphorylase in total ischaemia in the rat heart. A phosphorus-31 nuclear magnetic resonance study. Biochem J 196, 171-178 Crane RK, Sols A (1954) The non-competitive inhibition of brain hexokinse by glucose-6-phosphate and related compounds. J Biol Chem 210, 597-606 Sois A, Crane RK (1954) Substrate specificity of brain hexokinase. J Biol Chem 210, 581-595 White TK, Wilson JE (1990) Binding of nucleoside triphosphates, inorganic phosphate, and other polyanionic ligands to the N-terminal region of rat brain hexokinase: relationship to regulation of hexokinase activity by antagonistic interaction between glucose 6-phosphate and inorganic phosphate. Arch Biochem Biophys 277, 2634 Bailey IA, Radda GK, Seymour AML, Williams SR, (1982) The effecis of insulin on myocardial metabolism and acidosis in normoxia and ischemia. Biochim Biophys Acta 720, 17-27 Morgan HE, Henderson MJ, Regen DM, Park CR (1961) Regulation of glucose uptake in muscle. I The effects of insulin and anoxia on glucose transport and phosphorylation in the isolated peffused heart of normal rats. J Bio/ Chem 236, 253-261 Geisbuhler T, Altschuld RA, Trewyn RW, Ansel AZ, Lamka K, Brierley GP (1984) Adenine nucleotide metabolims and compartimentalisation in isolated rat heart cells. Circ Res 54, 536-547 Paris S, Pouissegur J, Ailhaud G (1980) Sugar transport in chick embryo cardiac cells in culture. Analysis by countertransport, relationship to phosphorylation and effect of glucose starvation. Biochim Biophys Acta 602, 644--652 Thompson KA, Kleinzeller A (1989) 2-Deoxy-a-glucose accumulation in adipocytes: apparent transport discrimination between 2-deoxy-o-glucose and 3-O-methyl-oglucose. Biochim Biophys Acta 101 l, 58-60
867
The inhibition of bovine heart hexokinase by 2-deoxy-D-glucose-6phosphate: characterization by alp NMR and metabolic implications W Chen, M Gu6ron* Groupe de Biophysique, Ecole Polytechnique, 91128 Palaiseau, France
(Received 1 June 1992; accepted 30 July 1992)
Summary B The glucose analog, 2-deoxy-D-glucose (2DG), has been used widely for studying the initial steps in the metabolism of glucose by radio-isotope tracer methods and by 31p NMR. In the rat heart perfused with acetate/2DG (both 5 mM) plus insulin, trapping of phosphorus by 2-deoxy-D-glucose-6-phosphate (2DG6P) results in a steady state exhibiting high 2DG6P (55 mM) and low ATP concentrations but near-normal function, as observed in an earlier 31p NMR study. In order to understand how the 2DG6P concentration is stabilized, we studied the inhibition of a mammalian hexokinase by 2DG6P in vitro by a 31p NMR technique. Inhibition, previously unobserved, was found. It is similar to inhibition by G6P in that it is competitive with ATP and not competitive with 2DG, but the inhibition constant (1.4 mM) is much larger. The experimental protocol includes provisions fo~ enzymatic destruction of stray inhibitor:~ such as G6P. The results show that the high 2DG6P and low ATP concentrations found in the steady state of the peffused heart should strongly reduce the rate of phosphorylation of sugars by hexokinase. deoxy.glucose / NMR./positron emission tomography / heart Introduction Glucose derivatives in P E T and N M R metabolic studies
2-Deoxy-D-glucose (2DG), an analog of glucose, has been used widely for studying the initial steps in the metabolism of glucose. The principle of the method is that 2DG behaves similarly to glucose with respect to facilitated transport across the sarcolemma [l] and to subsequent intraceilular phosphorylation by hexokinase (EC 2.7.1.1), resulting in the formation of 2deoxy-D-glucose-6-phosphate (2DG6P). The latter cannot leave the ceil, and is metabolized very slowly. Hence its accumulation provides a measure of glucose metabolism. These properties were first investigated by biochemical methods [2]. Isotopic labeling made possible their application to functional in vivo studies of the brain by [2-t4C]deoxy-glucose autoradiography [3] and [2-1SF]-fluoro-2-deoxy-D-glucose positron emission tomography (PET) [4]. The radio-isotope method is very sensitive but is blind to the nature of the chemical compound which carries the isotope. In particular, it does not distinguish between 2DG and *Correspondence and reprints
2DG6P. Neither does it detect the degradation of 2DG6E whose rate, required for the interpretation of the measurements, must be provided independently. Nuclear magnetic resonance (NMR) provides another means for following the metabolism of 2DG. It suffers from low sensitivity but is attuned to chemical characterization. In particular, 2DG6P can be observed and titrated by 3~p NMR. In a previous study, the rate of degradation of 2DG6P in the perfused rat heart was measured and shown to be insulindependent [5 ]. Due to its low sensitivity, the NMR measurement requires the use of large concentrations of 2DG6E and these have led to the investigation of their effect on heart function [6]. It was discovered that upon perfusion with 2DG and acetate, plus insulin, massive accumulation of 2DG6P occurred, and the rat heart reached a steady state in which function is quasi-normal, despite the very low concentrations of PCr and ATP that result from the sequestration of phosphate groups by 2DG6P. These observations provide strong limitations on the role of the concentrations of PCr or ATP as metabolic regulators. In the steady state, a constant concentration of 2DG6P is maintained by equal rates of 2DG phosphorylation and 2DG6P degradation. The rate of the latter was measured directly. The rate of the fom~er, which must equal that of the latter, was eight times smaller
868 than the initial rate o f 2DG phosphorylation. This could be due to inhibition by 2 D G 6 P o f the import of 2DG. as already invoked in an early study o f the rat diaphragm [2]. It could also be due to inhibition of hexokinase. Kipnis and Cori [2], unaware of 2 D G 6 P degradation, interpreted the constant 2 D G 6 P content as evidence for the arrest o f 2 D G 6 P production; having found no inhibition of rat skeletal muscle hexokinase by 25 m M 2 D G 6 E they ascribed the arrest to blockage o f 2 D G import.
The observation o f 2 D G 6 P degradation in the case o f the perfused rat heart forces us to question this conclusion. In other systems, the observation o f large concentrations o f intracellular 2 D G (non-phosphorylated) also suggests that 2 D G p h o s p h o r y l a t i o n is not limited by import [7]. We have therefore m e a s u r e d the inhibition of a m a m m a l i a n hexokinase b y 2DG6P. Inhibition is observed, m u c h weaker dlan b y G6P, but similar to it in the sense that it is not competitive with respect to glucose and competitive with respect to A T E The observations indicate that inhibition o f hexokinase must be taken into account in the description o f the metabolic steady state o f the rat heart perfused with 2-deoxyglucose.
Materials and m e t h o d s
,.~il
, -.l.d~,~ ,
....
~
. r~
The measurement of enzyme kinetics was performed by NMR. The sensitivity is lower than with classical, UV-monitored, enzymatic methods which could be considered for this study [8], but 31p NMR is well adapted to the inhibitor and metabolite concentrations required. It has the advantage that the concentrations of all the phosphorylated products are traced individually and simultaneously, during the course of the reaction. The pH is also obtained. Lastly, the spectrum would reveal parasitic reactions. The absolute rate of the reaction, adjusted by choosing the enzyme concentration, was chosen in view of the time in which the NMR spectrum could be measured with good sensitivity, which was about 6 min. The experiments were designed to avoid artifacts caused by traces of inhibitor impurities, particularly G6P which is a strong inhibitor of hexokinase (using our definitions (eqn 2), the K~ is 20 I.uM for pig heart type I hexokinase [9]).
18-24
Reagents
Is
o
I-s
i- o
!-15
Fig 1. Phosphorylation of 2DG by hexokinase. The lowest trace (at a vertical scale of 0.5) is the spectrum of data acquired between 6 and 12 min after addition of 0.2 U of hexokinase to the reaction mixture. The other traces correspond to spectra of data taken during successive 6-min intervals as indicated, from which the lowest trace has been subtracted. The loss of ATP and the generation of 2DG6P (ca 7.14 ppm) and ADP can be observed quantitatively, with reference to the known initial ATP concentration, l0 mM. Apparent non-stoichiometry is due to saturation (see Materials and methods). The slight shift of the 2DG6P peak (by. 0.05 ppm) corresponds to an acidification by 0.06 pH units. Experimental conditions: 2DG, 10 mM; ATP, l0 mM; PGI and PFK, 3 U each; reaction volume: 1.5 ml; pH 7.1; T = 37°C. Peak assignments: l, 2DG6P; 2, yATP; 3, ~ADP; 4, ¢tADP; 5, txATP, 6, ~IATE
ATE 2DG, 2DG6P, glucose, G6P, bovine heart hexokinase (HK, type X, 25 U, lyophilized), rabbit muscle phospho-isomerase (PGI, type IV, 1000 U in 0.6 ml) and rabbit muscle phosphofructokinase (PFK, type Ill, 500 U in 0.35 ml) were bought from Sigma (France) and used without further purification. Other reagents were products of Merck, analytic grade. Hexokinase was dissolved in 2.5 ml of a 7:3 mixture of glycerol and water, and stocked at-30°C. Before use, aliquots were diluted with water to a concentration of 0.025 to 0.1 U per 20 pl. PGI and PFK were diluted with water to a concentration of I U/pl and stocked at 4°C.
Composition of the reaction mixture The reaction mixture was designed with reference to intracellular physiologic conditions as regards ionic strength, pH and temperature. It included KCI (150 mM), Tris-HC! (200 mM) and EDTA (1 mM). The pH was adjusted to 7.1. The concentration of ATP[Mg] was 2-10 mM and total magnesium was 25 mM in excess of ATE The solution also contained 3 U of PGI and PFK, as discussed below. The water solvent was 30% deuterated.
Experimental procedure The reaction was started by adding 0.025 to 0.2 U of hexokinase (a volume of 20 [al) to 1.5 ml of the reaction mixture,
869 contained in a 10 mm outer diameter (od) NMR sample tube at room temperature. The tube was immediately transferred to the spectrometer which had been set to a sample temperature of 37°C. The data acquired during the first 6 min were not used in the analysis.
.~tp NMR measurements 31p NMR spectra were obtained at 145.6 MHz, using a homebuilt spectrometer [10] equipped with a 10-mm probe. The excitation pulse angle was 60 ° (22 Its) and the repetition time 1.4 s. The block accumulation time was chosen according to the concentration of substrates of hexokinase. It varied from 6 to 24 rain. In most cases, including those displayed in the figures, the free induction decay was multiplied by an exponential-sinus function exp(-~t&)*sin[27t(0.7&+0.07)] with 5 = 10 Hz, a combination chosen for the optimization of the signal-tonoise ratio and of the resolution [ 11 ]. In a few cases, for precise comparison of 2DG6P generation and ATP consumption, we used the integral of the 2DG6P and 13ATP peaks in spectra processed by 10 Hz broadening only. The chemical shifts are referred to phosphocreatine. Quantities were determined by the area of the 13peak of ATP in subtraction spectra (fig 1). The initial spectrum, for which the concentration of ATP is known, provides a reference which requires no correction for NMR saturation. The saturation factors for all peaks were later determined by comparison with NMR signals obtained with a repetition time of 20 s. That for the [3 peak of ATP is 1.9; in retrospect, a repetition time longer than 1.4 s would have been preferable.
Kinetic fotwlalism Hexokinase catalyses the phosphorylation of glucose: ATP + G ADP + G6P + H +
(I)
Glucose-6-phosphate slows down the reaction by competitive inhibition with ATP. The ratio of the reaction rate V to the maximum rate Vma~ is equal to the fraction of enzyme molecules engaged in an active complex with both substrates. If the two substrates bind independently, one finds:
[S,I [S.,l/(gl K2)
V m
Vm.,,
Since 2DG6P is a rather poor inhibitor, one must avoid stray inhibitors of hexokinase, or eliminate them from the reaction medium. Magnesium, ATP (if not complexed with Mg), and ADP are weak inhibitors [ 12]. The first two are controlled by using a slight excess of magnesium over initial ATP_. As for ADP, an inhibiter competitive with ATP, with a Ki of ca 2 mM [ 13|, it is a product of the reaction. We could have destroyed it enzymatically; we chose instead to measure the initial rates, within a range of times such that the ADP concentration had not reached 3 mM. Another problem is the presence of glucose-6-phosphate or of glucose as impurities in the 2DG and 2DG6P reagents. The NMR peaks of the H~ proton of glucose and of glucose-6phosphate, at 5.2 ppm, fall in a spectral region devoid of peaks in the spectrum of the deoxy compounds. We failed to observe these peaks in solutions of 2DG and 2DG6E and this indicated a content of glucose and glucose-6-phosphate smaller than 0.2%. But such quantities could still be troublesome, and the reaction medium was therefore designed to destroy G6E by incorporation of the enzymes PGI and PFK. The former transforms G6P into fructose-6-phosphate (F6P), which PFK transforms quasi-irreversibly into fructose 1,6-diphosphate.
Results Kinetics o f 2DG phosphorylation The results o f a control experiment are shown in figure 1. U p o n addition o f hexokinase to the reaction volume, production o f 2 D G 6 P and A D P begins, together with loss o f ATP. The rate o f the reaction can be measured directly by the variation o f the 2 D G 6 P peak, o f the two A D P peaks and o f the three ATP peaks. Taking into account the saturation factors (see Materials and methods), the variations o f the different phosphorus moieties are consistent with the stoichiometry of equation I.
Reactions o f the ancil l a o ' enzymes P G ! and P F K (2)
(1 + [l]/g,) (1 + [S~i/g~) + [S..,I/K., + [S~I[S,.IItKtK,.)
where S~ and $2 are the two substrates with Michaelis constant K~ and K,., and I is the inhibitor whose dissociaf,~,n constant is K., and which is competitive with $2. It is assumed that simultaneous binding of $2 and I is negligible. We shall see that the model can describe the observations, but it does not pretend to represent the enzymatic mechanism. The equation may be rewritten as:
[Sd IS.,] V m J V = (Kl + [Sd)(K2( l + [ l l / K i ) + [S-,])
Protection against sn'ay inhibitors
(3)
It predicts that the plot of [Sd/V vs lSd is a straight line which cuts the x-axis at -KI, independently of [I1 or [Sz]. The plot of [S2]/V vs IS_,] should also be a straight line, cutting the x-axis at -K2(I + [I]/Ki), and whose slope, K~ + S~, is independent of [I1. This model is known to account for the inhibition of glucose phosphorylation by G6P (I), which is non-competitive with glucose (S:) but competitive with ATP (S,) |9]. The measurements of the present study were also compared to this model.
These e n z y m e s are included so as to avoid inhibition o f the H K reaction by glucose-6-phosphate, which they degrade irreversibly. The degradation o f G 6 P was checked by c o m p a r i n g the rates o f production of 2 D G 6 P either without or with added G 6 F (40 lalVl). The results are shown in table I. Without addition o f 2 D G 6 P or G6P, the rates without (rate_) and with (rate+) PGI and P F K are equal, indicating the absence o f G 6 P in the reagents, including 2DG. The second line shows the inhibition by G6P: according to equation 3, the difference between (1/rate_) and (1/rate+) is proportional to I/K~, where I stands for the inhibitor destroyed by the ancillary e n z y m e combination. The recovery of the rate, up to the control value, upon addition o f PGI and P K F (rate+) shows the effectiveness of this procedure for the elimination o f G6P. The third line shows the inhibition by 2 D G 6 P and conta-
870 Table i. Kinetics of 2DG phosphorylation a. 2DG6P (mM)
G6P (mM)
Control
IlK (units)
Rate_ b
Rate. b
0.2
1.1 c
1.1
0.04
0.2
0.54
1.0
-
G6P 2DG6P
10
-
0.2
0.37
0.50
2DG6P + G6P
10
0.04
0.2
0.29
0.55
aATP and 2DG concentrations: 10 mM; brates are in (lanol of 2DG6P)/(HK unit.min). For rate_, the measurement is made without PGI or PFK; for rate., the solution contains 3 units of each; Cthe relative error on the rates is estimated at + 20%.
bined effect of G6P and 2DG6P. Assuming that G6P is the contaminant in 2DG6P, and that 2 D G 6 P and G6P are competitive, so that their effects on (1/ rate) are additive, then the observed rates correspond to 0.3% contamination (30 pM G 6 P for 10 m M 2DG6P), slightly above the limit o f observability (0.2%) of the experiments designed to search for such impurities, as described in Materials a n d methods.
,J
20
.is
.=.
E
10
r..--i
24-36 i
, ~
-2
,~,
,
0
I
,
~
I
2
I
,
,
4
I
,
,
,
6
II
8
,
,
,
I
,
,
10
[2DG] m M . . . .
io
?5
?lo
1-15
. . . .
I
. . . .
I
~ , , I , , , , I , , , , I
0
5
I
-20
-10
0
10
. . . .
I ' ' ' ' 1 ' ' ' '
. . . .
I = = , , I , , , ,
20
1
30
[2DG6P] mM Fig 2. Phosphorylation of 2DG by hexokinase in the presence of 2DG6P. Same conditions as in figure 1, except that the initial reaction mixture contains 30 mM 2DG6P. The difference spectrum is taken over a 12-min interval and scaled to 0.5, so that its peak iN~ensities are directly comparable to those of figure !. The rate of the reaction is much reduced. The slight pH change is conspicuous in the spectral shift of the 2DG6P peak in the difference spectrum. The inset shows the 2DG6P difference peak obtained after compensation of the spectral shift. minants (rate_). The partial suppression of the inhibition by PGI and PFK (rate.) is ascribed to the destruction of contaminants. The last line shows the com-
Fig 3. A. Plot of [2DG]/V vs [2DG], for different concentrations of 2DG6P: A, 0 mM; ×, 6 raM; o, 10 mM, ~, 20 mM; ~ , 30 raM. The conditions of the experiments are those of figure I. Straight lines converging to a point on the x-axis correspond to an inhibition by 2DG6P which is not competitive with 2DG, and -K2u3 is then the common intercept (eqn 3). For each concentration of 2DG6P, a straight line is traced according to the best visual fit. The relative errors for 30 mM 2DG6P are large because of the low rate (strong inhibition)" these data are ignored in the evaluation of K2~. B. The relative slopes of the plots in A are plotted vs inhibitor concentration. The straight line extrapolates to an intercept of-Ki(l + [ATP]/KAw). The point at 30 mM is not used for the tracing of the straight line. For symbols see A.
871 Side effects of the ancillary enzymes were searched for in a reaction mixture containing ATP and 2DG6P (10 mM each), HK (0.2 U), and prepared as described except that 2DG was omitted. After 30 min, the 3~p NMR spectrum displayed no peaks of ADP or other products, and no degradation of 2DG6P was observed. At that time 2DG was added. The initial rate of phosphorylation was the same as in an experiment in which the initial delay was omitted. Hence no product affecting the kinetics of HK had been generated during the delay.
2o ° ~
lO
Inhibition of 2DG phosphorylation by 2DG6P Figure 2 shows a typical measurement of the HK kinetics in the presence of the inhibitor 2DG6P. Comparison with figure 1 shows the reduction of the rate of 2DG phosphorylation. A small acidification (0.06 pH units) which occurs during the reaction produces a slight shift in frequency of the 2DG6P peak [14], which the subtraction procedure makes conspicuous. Whereas a full characterization of the enzyme involved rate measurements versus A T E 2 D G and 2 D G 6 P concentrations, we studied the H K reaction only at constant A T P (I0 raM) or at constant 2 D G (I0 raM). The rates in the various conditions are shown in figures 3 A and 4A. The variation of (substrate concentration/rate) vs substrate concentration indicates that the 2DG6P inhibitor is not competitive with 2DG (converging straight lines), and is competitive with ATP (parallel straight lines). Figures 3B and 4B show that the measurements are compatible with the linear dependence on the inhibitor concentration which is expected from equation 3. The intercepts with the xaxis provide the constants of the model: K2ua from figure 3A; /('^TP from the line measured without inhibitor in figure 4A; Ki from figure 4B. The inhibition constant for G6P is also obtained, assuming that G6P and 2DG6P react in the same fashion with the enzyme. The results are collected in table II. Data for the glucose-G6P system [9] are shown for comparison. A second value of Ki is provided by the intercept in figure 3B, which is predicted at Ki(I+[ATP]/KA~). The two values of Ki are 1.7 and 1.2 mM respectively. The difference is in the range of experimental errors and does not warrant a reconsideration of the model of equation 3. Discussion
Inhibition by 2DG6P Our results demonstrate that 2DG6P inhibits the phosphorylation of 2DG by bovine heart hexokinase. The
o -2
o
2
4
6
8
lO
[ATP] mM ...............................
10 8
!.
6
2 i
-5
i
0
10
20
i
,
i
30
[2DG6P] mM Fig 4. A. Plot of [ATP]/V vs [ATP], for different concentrations of 2DG6P. According to equation 3, parallel straight lines correspond to inhibition by 2DG6P in competition with ATE The common slope is then proport;onal to 1 + K2uJ[2DG]. The intercept in the absence of inhibitor determines KATp.B. Plot of the y-axis intercept in A vs inhibitor concentration. The straight line extrapolates to an x-axis intercept of-Ki, which is thus determined. Same conditions as in figure 1. For symbols see legend to figure 3.
result is genuine, and not an artifact due to contamination by the potent inhibitor G6P which we have eliminated by the procedures described above. The inhibition is competitiye with respect to ATP and not competitive with respect to 2DG. The inhibition constant (Ki = 1.45"(1 _+ 0.3) mM is about 250 times larger than that de:rived for G6P (first line of table II). In combination with the competition by ATE this may explain that inhibition of 2DG phosphorylation by 2DG6P wa,; not observed in earlier studies [2, 15].
872 The inhibition measured here probably extends to that of glucose phosphorylation, since HK phosphorylates these two sugars in similar fashion and at similar rates [161. The similarity of 2DG6P and G6P regarding competition with substrates suggests that the inhibition mechanism may be the same in both cases, namely allosteric inhibition by binding to a second site of the pseudo-dimeric mammalian enzyme [15]. This assumption is supported by our unpublished observation that 2DG6P does not inhibit 2DG phosphorylation by the monomeric yeast hexokinase, again like G6P [ 171.
Implications for the metabolism of 2-deoxy-glucose and glucose Knowing the rate parameters of phosphorylation of 2DG by HK makes it possible to investigate whether inhibition of the hexokinase reaction by 2DG6P contributes to the establishment of the low-ATE lowPCr, steady-state of the perfused rat heart mentioned in the Introduction. The hearts were studied under four different conditions [5]: with and without insulin, with glucose or with acetate perfusion. When glucose is the carbon source, its phosphorylation (and concomitantly that of 2DG) is regulated by the work output, and the regulation involves inhibition of hexokinase by G6P. (As a result, the phosphorylation rate is hardly affected by insulin [5, 18].) There is no such regulation when acetate is the carbon source, therefore the rate of 2DG phosphorylation is larger. The effect of insulin shows that in its absence, 2DG import is a limiting factor for phosphorylation [19]. But in its presence, hexokinase may become limiting, the more so when the concentration of the inhibitor 2DG6P is large and that of ATP is low, conditions which are those of the steady state. Can this explain the reduction by a factor of eight in the rate of phosphorylation in the steady state, with respect to the initial rate, where the concentration of 2DG6P is negligible?
We compute the rates by equation 3, using the appropriate concentrations of the different products. The intracellular 2DG concentration is assumed to be equal to that of the perfusate (5 mM), and therefore remains constant. With $2 standing for ATP and I for 2DG6E and with subscripts • for the initial state and • , for the steady state, we have: (l + [I**]/Ki)K2/[S2** ] + 1 V:~/V**
.-
([I**I/Ki)(K2/[S2**I) (4)
K2][S2.I + 1
where the last expression is obtained by keeping only the largest terms. The initial ATP concentration is 18 mM. The steady-state value is affected by a large relative error because of its low value, and of the possibility of a contribution of mitochondrial ATP to the observed signal. Based on figure 1 of [5], we take a value of 4.5 mM, but a value twice as small could be chosen on the basis of similar measurements [6]. The concentration of 2DG6E initially zero, reaches 55 mM in the steady state. We note that the last expression in equation 4 involves a ratio of concentrations and is therefore independent of the cytosolic volume (1.4 ml/(g protein) [6, 20]) used to compute them. With the values of table II, the ratio of rates is 7.5. It would be 15 if one took the smaller value mentioned for the steady-state ATP concentration. These values are in the range of the in vivo ratio of 8. Hence, inhibition of hexokinase by 2DG6P may be responsible for the slow phosphorylation in the steady state. At the very least, a strong iohibition of hexokinase is expected in the steady state of the acetate/2DG/insulin peffusion experiment. On the other hand, inhibition of hexokinase cannot be responsible for the steady state which is also observed in the absence of insulin. In this case, the steady-state concentration of 2DG6P is only 20 mM whereas that of ATP is close to the initial one. The ratio of initial to steady-state rates is again about eight, whereas the first expression in equation 4 predicts a ratio of 1.6. Hence other mechanisms must be
Table il. Mich,~elis and inhibition constants of the hexokinase reaction a.
2DG --->2DG6P G ---->G6~
~o~ ~
K~
~T~
g,,.~o~6~,
/t',,~,
0.6
-
0.8
1.45c
0.006d
-
0.02
0,4
-
0.020
apH 7.1, T = 37°C. bAll constants are in mM; the relative error, gauged from the spectral signal-to-noise ratio and from the dispersion in the plots, is approximately + 30%. CAverage of the values, 1.7 and 1.2, computed from figures 4B and 3B, respectively. dThe oata in lines 2 and 3 of table I show that llKi is the same for 40 l.tM of G6P and for 10 mM of 2DG6E Ilence K,c~p, = K,2~6p]250. eThese data, from [9], are for pig heart type I hexokinase; pH 7.6; T = 30°C; no added salt.
873
at work. One element is that 2DG6P may interfere with sugar import, as suggested by reports of anomalies in the intracellular concentration of glucose (nonphosphorylated) in the presence of sugars such as 2DG which are phosphorylated to products which accumulate in the cell [21, 22]. In summary, we have demonstrated the inhibition of a mammalian hexokinase by 2-deoxy-o-glucose-6phosphate, and determined the inhibition coefficient. Inhibition by 2DG6P is competitive v s ATP and not v s 2DG. Inhibition of hexokinase by 2DG6P must be taken into account in the interpretation of the 2DG metabolism, and it provides at least a partial explanatiot~ for the steady state of the rat heart perfused with acetate/2DG/insulin. Acknowledgment
8
9 10 11 12 13
14
WC was supported by Association fran~aise contre ies Myopathies.
15
References
16
1 Takala TES, Hassinen IE (1981) Effect of mechanical work load on the transmural distribution of glucose uptake in the isolated perfused rat heart, studied by regional deoxyglucose trapping. Circ Res 49, 62--69 2 Kipnis DM, Cori CF (1958) Studies of tissue permeability V. The penetration and phosphorylation of 2-deoxy-glucose ill the rat diaphragm. J Biol Chem 234, 171-177 3 Sokoloff L, Reivich M, Kennedy C, Des Rosiers MH, Patlak CS, Pettigrew KD, Sakurada O, Shirohara M (1977) The [t4Cldeoxyglucose method for the measurement of local cerebral glucose utilization: theory, procedure and normal values in the conscious anesthetized albino rat. J Neurochem 28, 897-916 4. Phelps ME, Hoffman El, Selin C, Huang SC, Robinson G, MacDonald N, Schelbert H, Kuhl DE (1978) Investigation of [IXFl2-fluoro-2-deoxyglucose for the measure of myocardial glucose metabolism. J Nucl Med 19, 1311-1319 5 Hoerter JA, Dorrnont D, Girault M, Gu6ron M, Syrota A (1991) 31p NMR evidence for insulin-dependent 2-deoxyglucose metabolism in the peffused rat heart. J Mol Cell Cardio123, 1101-1 i 15 6 Hoerter JA, Lauer C, Vassort G, Gu~ron M (1988) Sustained function of normoxic hearts depleted in ATP and phosphocreatine: a 31P-NMR study. Am J Physio1255, C ! 92---C201 7 Jacobs AEM, Oosterhof A, Veerkamp JH (1990) 2-DeoxyD-glucose uptake in cultured human muscle cells. Biochim Biophys Acta 1051, 230-236
17
18
19
20
21
22
Wilson JE, Chung V (1989) Rat brain hexokinase: further studies on the specificity of the hexose and hexose 6phosphate binding sites. Arch Biochem Biophys 269, 5 i 7-525 Easlerby JS, O'Brien MJ (1973) Purification and properdes of pig heart hexokinase. Eur J Biochem 38, 210211 Caron F, Gu6ron M, Nguyen Ngoc Quoc Thuy, Herzog RF (1980) A high-resolution multipurpose Fir" NMR spectrometer. Revue Phys Appl 15, 1267-1274 Gu6ron, M ¢1978) Line-narrowing and line-broadening using trigonometric functicms. J Magn Res 36, 515-520 Bachelard HS (1971) Allosteric activation of brain hexokinase by magnesium ions and by magnesium ion-adenosine triphosphate complex. Biochem J 125, 249-54 Rees BB, Ropson IJ, Hand SC (1989) Kinetic properties of hexokinase under near-physiologic condition relation to metabolic arrest in artemia embryos during anoxia. J Biol Chem 264, 15410-15417 Bailey IA, Williams SR, Radda GK, Gadian DG (1981) Activity of phosphorylase in total ischaemia in the rat heart. A phosphorus-31 nuclear magnetic resonance study. Biochem J 196, 171-178 Crane RK, Sols A (1954) The non-competitive inhibition of brain hexokinse by glucose-6-phosphate and related compounds. J Biol Chem 210, 597-606 Sois A, Crane RK (1954) Substrate specificity of brain hexokinase. J Biol Chem 210, 581-595 White TK, Wilson JE (1990) Binding of nucleoside triphosphates, inorganic phosphate, and other polyanionic ligands to the N-terminal region of rat brain hexokinase: relationship to regulation of hexokinase activity by antagonistic interaction between glucose 6-phosphate and inorganic phosphate. Arch Biochem Biophys 277, 2634 Bailey IA, Radda GK, Seymour AML, Williams SR, (1982) The effecis of insulin on myocardial metabolism and acidosis in normoxia and ischemia. Biochim Biophys Acta 720, 17-27 Morgan HE, Henderson MJ, Regen DM, Park CR (1961) Regulation of glucose uptake in muscle. I The effects of insulin and anoxia on glucose transport and phosphorylation in the isolated peffused heart of normal rats. J Bio/ Chem 236, 253-261 Geisbuhler T, Altschuld RA, Trewyn RW, Ansel AZ, Lamka K, Brierley GP (1984) Adenine nucleotide metabolims and compartimentalisation in isolated rat heart cells. Circ Res 54, 536-547 Paris S, Pouissegur J, Ailhaud G (1980) Sugar transport in chick embryo cardiac cells in culture. Analysis by countertransport, relationship to phosphorylation and effect of glucose starvation. Biochim Biophys Acta 602, 644--652 Thompson KA, Kleinzeller A (1989) 2-Deoxy-a-glucose accumulation in adipocytes: apparent transport discrimination between 2-deoxy-o-glucose and 3-O-methyl-oglucose. Biochim Biophys Acta 101 l, 58-60
