Modulation by citrate of glycolytic oscillations in skeletal muscle extracts.

Oscillatory behavior of glycolysis in cell-free extracts of skeletal muscle involves repeated bursts of phosphofructokinase activity and associated oscillations in the [ATP]/[ADP] ratio. Addition of citrate, a potent physiological inhibitor of phosphofructokinase, decreased the frequency of the oscillations and delayed the first burst of phosphofructokinase activity in a dose-dependent manner. Citrate decreased the trigger point [ATP]/[ADP] ratio at which bursts of phosphofructokinase activity were initiated but had a much smaller effect on the average [ATP]/[ADP] ratio and did not decrease the peak values of the ratio. When oscillations were prevented by addition of fructose-2,6-P2, the decrease in the [ATP]/[ADP] ratio caused by citrate in the steady state system was similar to the decrease in the trigger point [ATP]/[ADP] ratio in the oscillatory system. The decrease in the average [ATP]/[ADP] ratio was greater in the steady state system than in the oscillating system. These results demonstrate advantages of oscillatory behavior of glycolysis in the regulation of carbohydrate utilization and the maintenance of a high [ATP]/[ADP] ratio.

Oscillatory behavior of the glycolytic pathway occurs under certain conditions when cell-free extracts of rat skeletal muscle are provided with glucose (1-6). The oscillations are generated by repeated bursts of phosphofructokinase activity involving AMP-dependent, autocatalytic activation of the enzyme by its product, fructose-1,6-P2. A burst is initiated once the decreasing [ Citrate is a potent inhibitor of phosphofructokinase that in some cases is thought to mediate the inhibition of glycolysis by alternative fuels, such as fatty acids and ketones (7, 8). Activation of the glycolytic flux in the presence of citrate would therefore require a compensatory decrease in the [ATPJ/[ADP] ratio or an increased energy demand. In the context of oscillatory glycolysis, citrate might be expected to decrease the trigger point [ATP]/[ADP] ratio. What effects citrate might have on the average or peak values of the [ATP]/[ADP] ratio or the oscillation frequency also remain to be determined. These parameters are of particular interest in regard to our proposal that oscillations in glycolysis and the [ATP]/[ADP] ratio may be involved in the signaling mechanism of glucose-stimulated insulin release in the pancreatic @-cell by causing repeated closure of ATP-sensitive potassium channels and resulting oscillations in intracellular free Cap+ (9, 10). The objective of the experiments presented here was to test the effects of citrate under conditions of oscillatory and steady state glycolysis in the muscle extract system, especially with regard to changes in the [ATP]/[ADP] ratio.

EXPERIMENTAL PROCEDURES
The gel-filtered, high speed supernatant of rat hind leg muscle was prepared as described previously ( l l ) , except that the tissue was broken up using a Polytron homogenizer and EDTA was usually omitted from the gel filtration buffer. A larger column (5 X 20 cm) of G-25 Sephadex was used in order to gel filter 50-ml portions of muscle extract a t a time. Aliquots of the gel-filtered extract were quickly frozen in a dry ice/ethanol bath and stored frozen for up to 2 or 3 months; any precipitated protein was removed by brief centrifugation (e.g. 500 X g for 1 min a t 4 "C). Glycolytic oscillations in reaction mixtures were monitored spectrophotometrically on an aliquot by following changes in A260-262 -A2nl.nx:, using a Hewlett-Packard model 8450 spectrophotometer system; such spectral changes are due to corresponding oscillations in the operation of the purine nucleotide cycle in response to the changes in the [ATP]/[ADP] ratio associated with the glycolytic oscillations (3). A light path of 0.5 mm was used by placing a 0.95-cm quartz block in a 1.0-cm cuvette. Samples (1 ml) of the reaction mixture were deproteinized with perchloric acid for enzymatic assay of glycolytic intermediates, ATP, NDP, and AMP, as described previously (2, 3). The centrifugation and neutralization of the samples was done the day of the experiment to minimize ATP hydrolysis. NDP (ADP + GDP) was measured directly because of the nonspecificity of pyruvate kinase, which is normally used for assaying ADP. The assays were performed using a Hewlett-Packard model Biochemicals and enzymes were obtained from Boehringer Mannheim or Sigma. Hexokinase used in the oscillation reaction mixtures was gel-filtered to remove ammonium sulfate as described previously (1). (Endogenous hexokinase is largely bound to mitochondria (13) and hence is lost in the high speed centrifugation of the muscle extract (11.) Male Sprague-Dawley rats were obtained from Charles River Laboratories, Inc.

RESULTS
Addition of increasing concentrations of citrate in the physiological range progressively decreased the frequency (increased the period) of the glycolytic oscillations in muscle extracts (Fig. 1). Furthermore, the start of the oscillations was progressively delayed.
The effect of citrate on the oscillating levels of glycolytic intermediates and adenine nucleotides is shown in detail in Fig. 2. The timing of the repeated bursts of phosphofructokinase activity is indicated by the sudden drops in glucose-6-P and fructose-6-P and rises in fructose-1,6-P2 and triose-]?. The time of the first burst of phosphofructokinase activity was delayed in the presence of citrate; there was a greater accumulation of AMP and fructose-6-P and loss of ATP before the initiation of the activation of the enzyme occurred. In the subsequent oscillations, where there was less relative difference in fructose-6-P levels with and without citrate, accumulation of 50-100% more AMP was apparently required to initiate the burst of phosphofructokinase activity in the presence of citrate as in its absence. This correlates with about a 30% lower trigger point [ATP]/[ADP] ratio ( Fig. 3 and Table I (Fig. 4).

DISCUSSION
The results presented here show that citrate modulates glycolytic oscillations by decreasing their frequency and delaying their onset. These effects are due to inhibition of phosphofructokinase, such that a lower trigger value of the [ATP]/[ADP] ratio must be reached before a burst of phosphofructokinase can be initiated. However, citrate inhibition has little effect on the maximum or average values of the [ATP]/[ADP] ratio in the oscillating system, compared with the pronounced drop in the ratio caused by citrate in steady state glycolysis. These experimental results verify predictions made previously on the basis of mathematical modeling (4). The citrate concentrations used here are similar to the physiological levels of 0.2-0.3 mM calculated for rat leg muscle ( 5 ) .
It has been well established that glycolysis in heart is inhibited by increased levels of fatty acids or ketones, by nutritional states such as diabetes or starvation in which the concentrations of these substances are high, and by alternate acetyl-coA precursors such as pyruvate or acetate (7, 8). These effects are at least in part mediated by an increased tissue level of citrate (14-18). Glucose uptake and oxidation are also inhibited under these conditions. Whether citratemediated inhibition of glycolysis occurs in skeletal muscle is less clear, in view of the negative results of some studies with incubated muscles and the perfused rat hindquarter (19-21). I t was suggested that a rise in citrate might occur only in Conditions were similar to those described for Figs. 1 and 3, except that 0.5 pM fructose-2,6-P2 was added, the NAD concentration was 10 p~, buffer pH was 6.8, the muscle protein concentration was 1.3 mg/ml, and aspartate was omitted. found an increase in citrate in red muscle from starved rats after exercise, but not with starvation or exercise alone or diabetes. Increased citrate has also been observed in skeletal muscle perfused with acetate or acetoacetate (26) and following a fat diet (27). In human leg muscle, too, the combination of a fat diet and short term exercise can lead to a rise in citrate (28). The effects of exercise in raising citrate may result in part from increased operation of the purine nucleotide cycle to provide oxaloacetate (via fumarate) (29), together with increased acetyl-coA production from carbohydrate and fatty acid oxidation.
It was previously shown that oscillatory behavior of glycolysis can maintain a higher average [ATP]/[ADP] ratio than steady state or steady flux behavior (5, 6) and has a greater thermodynamic efficiency (30, 31). The experiments presented here demonstrate a related advantage of such oscillatory behavior in the context of use of competing fuels. Since in the oscillatory behavior a burst of phosphofructokinase activity only occurs once the trigger point [ flux behavior. Thus, glycolytic oscillations provide a better mechanism for regulation by alternative fuels without compromising the capacity of glycolysis to produce a high energy state on demand.
Glycolytic oscillations have also been studied in heart extracts (32), yeast cells and extracts (33)(34)(35)(36), and ascites cells (37). Frenkel (38) reported that addition of citrate blocked the oscillations in heart extracts; however, the 4 mM concentration used was much higher than the physiologically relevant concentrations used here. Citrate was reported to have no effect on oscillations in yeast extracts (33). Citrate inhibition of yeast phosphofructokinase may be relatively weak (50% inhibition by 5 mM citrate) and was reported to be lost on storage or partial purification of crude preparations of the enzyme (39).
Recently there has been increasing interest in oscillatory biochemical systems because of their distinct advantages for signal transduction, especially with such frequency modulation as shown here by citrate (40)(41)(42). We have proposed that stimulus secretion coupling in the pancreatic islet involves glycolytic oscillations and closure of ATP-sensitive potassium channels by the peaks in the [ATP]/[ADP] ratio (9,lO). Islet phosphofructokinase is inhibited by citrate (43), and the citrate concentration rises following glucose stimulation (44). The present results with the muscle extract model system indicate that citrate does not reduce the peaks of the [ATP]/ [ADP] ratio during oscillatory glycolysis (Fig. 3) and, therefore, should not interfere with closure of the ATP-sensitive potassium channels in the islet coupling mechanism. However, under some conditions the rise in citrate might alter the frequency of the linked oscillations in glycolysis, the [ATP]/ [ADP] ratio, intracellular free Cap+, and insulin release. Citrate might also serve to counter the actions of fructose-2,6-P p , glucose-1,6-Pz, and phosphate, which have the opposite modulatory effect (1,5,6).