Intracellular zinc movement and its effect on the carbohydrate metabolism of isolated rat hepatocytes.

The effect of zinc ions on carbohydrate metabolism and intracellular Zn2+ was studied in hepatocytes from fed rats. The addition of ZnCl2 to the medium led to an almost 3-fold increase in lactate production and an increase in net glucose production of about 50%. Half-maximal rates occurred at about 40 microM ZnCl2. These effects were not seen with Mn2+, Co2+, or Ni2+ up to 80 microM, whereas Cu2+ at 80 microM and Cd2+ or Pb2+ at 8 microM exhibited similar effects as 80 microM ZnCl2. Changes in intracellular Zn2+ were followed by single cell epifluorescence using zinquin as a specific probe. Intracellular free Zn2+ in isolated hepatocytes was 1.26 +/- 0.27 microM, and the addition of ZnCl2 led to a concentration-dependent increase in epifluorescence. CdCl2 or PbCl2 at 8 microM was as potent as ZnCl2 at 20-80 microM, whereas NiCl2 at 80 microM was without effect. ZnCl2 completely abolished the inhibition of glycolysis by glucagon (cAMP). Glucagon led to a pronounced drop in cytosolic Zn2+. Both glucagon and zinc stimulated glycogenolysis by increasing the phosphorylation of glycogen phosphorylase but acted oppositely on glycolysis. Zinc overcame the inactivation of pyruvate kinase by glucagon without changing the hormone-induced protein phosphorylation. The antagonistic action of zinc and cAMP on glycolysis together with the rapid and marked decrease in free zinc concentration induced by glucagon (cAMP) may indicate an as yet unknown role of zinc as an important mediator of regulation of carbohydrate metabolism.

The effect of zinc ions on carbohydrate metabolism and intracellular Zn 2؉ was studied in hepatocytes from fed rats. The addition of ZnCl 2 to the medium led to an almost 3-fold increase in lactate production and an increase in net glucose production of about 50%. Halfmaximal rates occurred at about 40 M ZnCl 2 . These effects were not seen with Mn 2؉ , Co 2؉ , or Ni 2؉ up to 80 M, whereas Cu 2؉ at 80 M and Cd 2؉ or Pb 2؉ at 8 M exhibited similar effects as 80 M ZnCl 2 . Changes in intracellular Zn 2؉ were followed by single cell epifluorescence using zinquin as a specific probe. Intracellular free Zn 2؉ in isolated hepatocytes was 1.26 ؎ 0.27 M, and the addition of ZnCl 2 led to a concentration-dependent increase in epifluorescence. CdCl 2 or PbCl 2 at 8 M was as potent as ZnCl 2 at 20 -80 M, whereas NiCl 2 at 80 M was without effect. ZnCl 2 completely abolished the inhibition of glycolysis by glucagon (cAMP). Glucagon led to a pronounced drop in cytosolic Zn 2؉ . Both glucagon and zinc stimulated glycogenolysis by increasing the phosphorylation of glycogen phosphorylase but acted oppositely on glycolysis. Zinc overcame the inactivation of pyruvate kinase by glucagon without changing the hormone-induced protein phosphorylation. The antagonistic action of zinc and cAMP on glycolysis together with the rapid and marked decrease in free zinc concentration induced by glucagon (cAMP) may indicate an as yet unknown role of zinc as an important mediator of regulation of carbohydrate metabolism.
Glycolysis in the liver is generally assumed to be regulated at the steps catalyzed by hexokinase/glucokinase, phosphofructokinase-1, and pyruvate kinase. We have shown that in vitro a 11.5-kDa Zn 2ϩ -binding protein (ZnBP), 1 which is identical with parathymosin, interacts in a zinc-specific manner with key enzymes of carbohydrate metabolism, including the liver enzymes aldolase, phosphofructokinase-1, hexokinase, glucose-6-phosphate dehydrogenase, glycerol-3-phosphate dehydrogenase, glyceraldehyde-3-phosphate dehydrogenase, and fructose-1,6-bisphosphatase (1). ZnBP is present in the liver at high concentrations of about 20 M (2), yet its physiological function is unknown, though ZnBP was discovered by its property to inactivate phosphofructokinase-1 in a reversible, zinc-dependent manner (3).
So far only few data are available on the effects of zinc on carbohydrate metabolism. With cytosol from muscle, zinc at half-maximal concentration of about 0.1 mM stimulated lactate production from glucose 6-phosphate (4). Rognstad (5) reported that zinc markedly inhibited glycogen synthesis in hepatocytes from starved rats; moreover, in hepatocytes from fed rats no effect could be observed on the glycogenolysis. In rat adipocytes zinc stimulated glucose transport (6,7), incorporation of glucose into lipids (8), and glucose oxidation by both pathways, glycolysis and hexose monophosphate shunt (9).
Zinc is known to play a pivotal role in the regulation of DNA binding and activation of transcription factors or in the regulation of apoptosis. As for Ca 2ϩ intracellular zinc levels have been suggested to be under homeostatic control. The latter may be essential to maintain the diversity of biochemical functions in the intermediary metabolism, which are dependent on zinc proteins and enzymes (10). However, little is known about the level of Zn 2ϩ in intact cells and to what extent changes in intracellular distribution may influence cellular events. Hitherto the rather high affinity constants of metalloenzymes for zinc have inferred the free zinc concentration in cells to be in the order of 10 Ϫ12 M or even lower (11). Recently, Zalewski et al. (12) have introduced the fluorescent indicator zinquin (ethyl(2methyl-8-p-toluenesulfonamido-6-quinolyloxy) acetate as an intracellular probe for Zn 2ϩ . Using Jurkat lymphoid cells pretreated with pyrithione and zinc, they estimated the free or "labile" intracellular zinc as detected by zinquin to be in the order of 2 nmol/10 7 cells (12), which would correspond to an intracellular concentration of 10 Ϫ5 -10 Ϫ4 M, assuming at least 10 8 cells/g of wet weight. The concentration of free zinc detected by these authors in cultured rat hepatocytes or in liver slices resembled the total zinc concentration in liver, and therefore, zinquin should be used with caution as a probe to measure Zn 2ϩ in living cells (13).
In hepatocytes the bulk of cellular zinc is thought to be bound as metallothioneins (MT). Accordingly, a rise in total cellular zinc has been observed following increased de novo synthesis of MT in the presence of glucocorticoids or glucagon (14). However, little is known about short term effects on the distribution of zinc ions within the cell.
The present study was designed to look for the effects of zinc ions on carbohydrate metabolism of hepatocytes as an indirect indication of an involvement of ZnBP in the regulation of glycolysis and gluconeogenesis. If the zinc-dependent interaction of ZnBP with key enzymes of the glycolytic and gluconeogenic pathways were of any importance in vivo, a rise in intracellular Zn 2ϩ should result in marked changes of flux rates. Further, it is unknown if and by which mechanism intracellular zinc can be mobilized to increase the cytosolic concentration of free zinc.
Here we report that incubation of hepatocytes from fed rats with ZnCl 2 leads to a 2-3-fold increase in the rate of lactate production and to an enhanced glucose production, due to an enhanced glycogenolysis and an increased flux from glucose-6phosphate through the glycolytic pathway. We show further that zinc counteracts the inhibition of glycolysis by glucagon or cAMP.
In parallel, we measured zinquin-dependent epifluorescence of single hepatocytes. The effect of various metal ions and glucagon or cAMP on the rapid exchangeable cytosolic zinc concentration was investigated. The similarity of increase in glycolysis seen with ZnCl 2 as compared with CdCl 2 or PbCl 2 can be explained by their ability to mobilize intracellular zinc. Moreover, by single cell measurement we can show for the first time that glucagon (or cAMP) rapidly decreases the level of intracellular Zn 2ϩ to less than 20% of that of control.

Methods
Preparation and Incubation of Hepatocytes-Livers from male or female Wistar rats (170 -220 g) fed on Altromin-R stock diet were perfused and used for isolation of hepatocytes by the two-step procedure as described by Berry et al. (15). For incubation, freshly prepared hepatocytes (about 4 g of wet weight) were washed with 250 ml of an albumin-and P i -free incubation medium (4°C, gassed with oxygen) containing 68.44 mM NaCl, 60 mM Hepes, 30 mM Mops, 5.4 mM KCl, 0.70 mM Na 2 SO 4 , 1.22 mM CaCl 2 , and 1.0 mM MgCl 2 adjusted to pH 7.40 with 1 M NaOH at 20°C (HMS medium). After the last washing step, the hepatocytes were diluted to a final concentration of 100 Ϯ 10 mg of wet weight/ml, and aliquots of 2.5-3.5 ml were transferred to open conical 25-ml plastic beakers with a diameter of 3.5 cm at the bottom. The vessels contained maltol to give a final concentration of 0.2 mM in the suspension, and, for zinc incubations, the appropriate amounts of ZnCl 2 . Hormones or dibutyryl-cAMP (Bt 2 -cAMP), when used, were added after the preincubation period. Preincubation time for equilibration was 10 min. The incubations were performed in a waterbath at 37°C with a shaking frequency of 100 ϫ min Ϫ1 , while being gassed with carbogen (95% O 2 , 5% CO 2 ). For the determination of metabolites, aliquots of 0.5 ml were removed after the preincubation time and 20 min later unless otherwise noted.
Cell integrity was measured by trypan blue exclusion and usually resulted in 90 -95% of the cells appearing viable. Because of a possible cell toxicity of ZnCl 2 or other heavy metal ions used, the activity of lactate dehydrogenase in the incubation medium was measured over the length of experiments.
Measurement of Cytosolic Zn 2ϩ -The cells were washed once (50 ϫ g max for 1.5 min) and resuspended in HMS medium plus 15 mM glucose to give a concentration of about 60 mg of wet weight/ml. They were incubated with zinquin ester (5 M) for 30 min at 37°C in a shaking water bath (100 cycles/min) under oxygenation. After this the cells were spun down (100 ϫ g av for 1.5 min), the supernatant was discarded, and the pellet was resuspended in the same volume of HMS medium plus 15 mM glucose and kept on ice until use.
The cells were incubated at 30°C for up to 30 min under conditions as specified in the text. For fluorescence measurements hepatocytes (1-2 mg of wet weight) were withdrawn at time intervals as indicated, suspended in 2 ml of oxygenated HMS medium in a Petri dish (Falcon 3001, with a centered 10-mm glass window), and immediately placed in the light path of the microscope. The fluorescence of single hepatocytes at 380 -390 nm excitation was monitored at 480 -540 nm (see below). Zn 2ϩ was determined from the increase in fluorescence. Usually the epifluorescence of 20 -25 single hepatocytes under each condition was averaged. To correct for autofluorescence, control incubation(s) using hepatocytes "preloaded" in the presence of solvent but otherwise treated identically were always run in parallel. The average in epifluorescence of again 20 -25 single hepatocytes was subtracted from the corresponding values measured in the presence of the dye. The concentration of the solvent(s) in the incubation never exceeded 1% (v/v). The same amount of solvent was added to control incubations.
Fluorescence measurements were performed using a Fura-2 data acquisition system (Luigs & Neumann GmbH, Ratingen, FRG) mounted to an inverted microscope (Zeiss IM 35). The sampling rate was 2 measurements/s. For a more detailed description and evaluation of the equipment see Neher (16).
The concentration of Zn 2ϩ was calculated according to Grynkiewicz If not otherwise stated, the values given are the means Ϯ S.E. of 20 -25 single cells/dish under each condition. The experiments were repeated at least three times with independent cell preparations. Total zinc and calcium were measured by atomic absorption after acid extraction and appropriate dilution of the sample in 0.1 M HCl with 0.3 g of LaCl 3 /100 ml.
Determination of Metabolites-Aliquots (0.5 ml) of the suspension were removed and mixed with 0.25 ml of 10.5% HClO 4 . After centrifugation, 0.5 ml of the supernatants were neutralized with 0.9 M KOH/50 mM Tris, kept on ice for 30 min, and then centrifuged. The clear supernatants were used for the determination of lactate and glucose according to Bergmeyer (19).
For cross-over studies, at least 6 ml of suspension were used from each incubation. Protein was precipitated by the addition of 0.1 vol of 35% HClO 4 and removed by centrifugation. An aliquot of the supernatant was neutralized with 5 M KOH and centrifuged. The pellet was washed twice with 0.1 M KCl. The combined supernatants were treated with Florisil until they became colorless and were lyophilized. The samples were dissolved in 1-1.5 ml of H 2 O. Determinations of metabolites and adenine nucleotides were performed essentially as described by Bergmeyer (19), but in addition, all assays contained 1 mM diethylenetriaminepentaacetic acid.
[ 32 P]P i Incubation of Hepatocytes and Precipitation of Pyruvate Kinase, Phosphorylase, and Fructose-1,6-bisphosphatase-Hepatocytes were incubated in the presence of HMS medium plus 15 mM glucose and 1 mCi of [ 32 P]P i (carrier-free)/g of wet weight for 30 min and thereafter centrifuged and resuspended in 15 ml of fresh medium without glucose and [ 32 P]P i . The cells were then incubated as given above with and without 0.1 mM ZnCl 2 in the presence and the absence of 0.1 mM Bt 2 -cAMP. After 20 min, 1 ml of the cell suspension from each incubation was diluted with 5 ml of medium (0°C), and cells were sedimented at 50 ϫ g for 1.5 min and resuspended in 1 ml of an extraction/stopping solution containing 45 mM Hepes, 50 mM NaF, 20 mM K 2 HPO 4 , 1.8 mM EDTA, pH 7.5, and 0.2% Triton X-100. After 1 min at room temperature, the suspension was centrifuged in an Eppendorf centrifuge. From each supernatant 0.1-ml aliquots were used for immunoprecipitation of pyruvate kinase/phosphorylase by adding 0.075 ml of a rabbit antipyruvate kinase (L type) antiserum and polyethylene glycol 6000 to a final concentration of 3%. Phosphorylase is coprecipitated as the antiserum cross-reacts with this enzyme. Fructose-1,6-bisphosphatase was precipitated from 0.1-ml supernatants by adding 0.01 ml of rabbit anti-fructose-1,6-bisphosphatase (liver) antiserum and 3% (final concentration) polyethylene glycol 6000. After 30 min, the incubations were centrifuged, and the sediments were washed twice with KRB and analyzed by SDS-polyacrylamide gel electrophoresis. The dried gels were autoradiographed for 4 -6 h on Kodak X-Omat film.
Measurement of Phosphorylase-Aliquots (0.5 ml) were removed from the cell suspension after 10 min of incubation, immediately frozen in liquid nitrogen, and stored at Ϫ60°C until used. Homogenization and determination of phosphorylase a were performed as described by Hue et al. (21). Total phosphorylase was measured as phosphorylase a except that 1.5 mM AMP was added and caffeine was omitted. The total activity at 30°C was 18 units/g of wet weight.
Statistical Analysis-The data are expressed as the means Ϯ S.E. with the number of independent cell preparations (n) in parenthesis. When appropriate, Student's t test for paired data was used for statistical evaluation.

Materials
Collagenase "Worthington" (type CLS II) was purchased from Seromed. Maltol, N,N,NЈ,NЈ-tetrakis (2-pyridylmethyl)ethylenediamine, Hepes, and Mops were from Sigma, and [ 32 P]P i was from Amersham Corp. Zinquin acid and zinquin ester were purchased from Luminis LTD (Adelaide, Australia). Biochemicals and enzymes were from Boehringer; all other chemicals were of analytical grade and came from Merck.

Set-up of Incubation Conditions-Because hepatocytes had
to be incubated in the presence of zinc, anions (bicarbonate, phosphate) maintaining low solubility products with zinc and albumin had to be omitted from the incubation medium. A strongly buffered medium comparable with that reported by Seglen (22) was used but with the omission of phosphate. Under these conditions, hepatocytes from fed rats produced lactate and glucose from endogenous glycogen at a constant rate up to 40 min. Hepatocytes from starved rats with lactate (9 mM) plus pyruvate (1 mM) as gluconeogenic precursors showed glucose production rates of 50 -70 mol of glucose/g of wet weight/h. This gluconeogenic capacity of the cells is comparable with that reported for isolated perfused livers, indicating that in the presence of carbogen no lack in bicarbonate occurred over the time of incubation.
Linearity of metabolic rates was assured by measuring lactate and glucose production in 10-min intervals up to 40 min. Metabolic rates were constant for at least 30 min under all conditions used (minus or plus ZnCl 2 , glucagon, or Bt 2 -cAMP). Therefore, metabolic rates were regularly determined from the 10 and 20 min values; metabolites were determined at 20 min or at an earlier time point.
A preincubation of 10 min proved to be sufficient for hepatocytes to equilibrate in terms of temperature and oxygen. This time was also sufficient to reach a constant zinc level in the medium. This is shown in Fig. 1, where the uptake of zinc is depicted from experiments with 0.05 and 0.10 mM ZnCl 2 added to the medium. With hepatocytes from starved rats, the time course of zinc uptake was comparable with that seen with hepatocytes from fed rats. Interestingly the endogenous zinc concentration was always higher in hepatocytes from rats starved overnight than from fed rats. The mean value Ϯ S.E. was 415 Ϯ 15 nmoles of zinc/g of wet weight (n ϭ 9) in hepatocytes from starved rats, as compared with 264 Ϯ 11 nmoles/g of wet weight (n ϭ 14) in hepatocytes from fed rats. Interestingly, Coyle et al. (23) have found the MT concentration in hepatocytes from fasted rats were double those from fed rats.
Apart from following the exclusion of trypan blue, the leakage of lactate dehydrogenase was followed over the time of experiments as an additional indicator of cell membrane integrity. In general, the activity of lactate dehydrogenase found in the medium was about 5% of the total activity, and after 30 min of incubation this amount increased to about 10%. Intracellular calcium was measured as an additional indicator of cell viability. During the incubation period of 20 min, the total cellular calcium remained constant in the control incubations and did not change in the presence of 0.1 mM ZnCl 2 . The total calcium concentration was 2.5 mol/g of wet weight, which is about twice that found in the intact liver (24).
Effect of Zinc on Lactate and Glucose Formation of Hepatocytes from Fed Rats-Hepatocytes from fed rats were incubated without exogenous substrate in the absence or the presence of different ZnCl 2 concentrations up to 0.1 mM. All incubations contained 0.2 mM of the membrane permeable zinc-chelator maltol. Lactate and glucose concentrations were determined, and the rates were calculated as mol/g of wet weight/h. The results are shown in Fig. 2. The addition of more than 0.05 mM ZnCl 2 to the medium resulted in a doubling of the rate of lactate formation and an increase in the rates of glucose production by 50%. Although the relative increase in each experiment depended on the control rates, distinct effects of zinc were seen in the mean values as given in Fig. 2. In general, incubations with 0.1 mM ZnCl 2 resulted in rates of 30 -40 mol of lactate/g/h. Maltol (0.2 mM) itself had no effect on the metabolic rates in the absence of exogenous zinc. Maltol has been used to facilitate uptake of zinc into erythrocytes, releasing liganded zinc to hemoglobin (25). Although with hepatocytes a better uptake of zinc was not observed in the presence of maltol, the effects of zinc were more reproducible, indicating that zinc offered to the cell in the liganded form is better available for biological effect than the unliganded form.
Under all conditions, the addition of the membrane-impermeable chelating agent diethylenetriaminepentaacetic acid to the medium at concentrations equimolar to those of ZnCl 2 abolished the effects of zinc (data not shown, but see Fig. 3). Moreover, no difference in the oxygen consumption of hepatocytes could be detected in the presence of maltol or maltol plus Because we had to use a medium without phosphate or bicarbonate, the zinc effect was compared in parallel incubations using HMS medium and KRB buffer with 1.2 mM CaCl 2 . Because zinc is completely insoluble in KRB buffer alone, maltol had to be added at much higher concentrations (0.42 mM) to maintain a Zn 2ϩ concentration of 0.1 mM. This titration was verified by atomic absorption spectroscopy. Under these conditions we could show that the zinc effects in both media were comparable. Glycolysis was doubled in both media, the glycogenolysis (lactate plus glucose formation) in terms of glucose equivalents increased by 44% Ϯ 13 (3) in KRB buffer and 55% Ϯ 16 (3) in HMS medium. In both media the complete inhibition of glycolysis by 0.05 mM Bt 2 -cAMP (see below) was abolished by 0.1 mM zinc.
Effect of Other Metal Ions on Lactate and Glucose Formation-To evaluate the specificity of zinc, the effect of iron, nickel, cobalt, manganese, copper, cadmium, and lead ions was examined. Calcium and magnesium were always present in the incubation medium at 1.2 and 1.0 mM, respectively. The incubation conditions were as given in the methods section, except that ZnCl 2 was replaced by the other metal ions. Up to a concentration of 80 M, no effect was observed with Fe 2ϩ , Ni 2ϩ , and Co 2ϩ . With MnCl 2 a small but inhibitory effect (ϳ20%) on the rates of lactate production was detectable. With CuCl 2 , no effect on the lactate formation occurred up to a concentration of 40 M, but at 80 M lactate formation increased to values that were comparable with those found in the presence of 80 M ZnCl 2 .
CdCl 2 and PbCl 2 were cell toxic at concentrations above 40 M. At 8 M, however, both cations were as effective stimulating the rates of lactate and glucose production as ZnCl 2 at 80 M.
Hormone Effects in the Presence of Zinc-The effects of glucagon and vasopressin on lactate and glucose formation from glycogen were examined in the presence and the absence of ZnCl 2 (Table I). With glucagon (10 Ϫ7 M) alone, lactate formation was nearly abolished or was even converted into lactate consumption. However, the stimulatory effect of zinc on lactate production was not overcome by glucagon (Table I).
Glucose formation was increased by glucagon due to enhanced glycogenolysis. ZnCl 2 had no additional effect at maximally effective concentrations of glucagon, whereas at submaximal rates of glucose formation, it had an additive effect (not shown).
Whereas stimulation of glycogenolysis by glucagon is caused by increased levels of cAMP, vasopressin acts by increasing the cytosolic concentration of free calcium via inositol 1,4,5trisphosphate. Vasopressin (0.5-5 ϫ 10 Ϫ8 M) itself had no effect on the net lactate production, although the cells were hormonesensitive as indicated by the enhanced glucose formation (ϳ143%). The addition of ZnCl 2 increased the rate of lactate formation in the presence of vasopressin (see Table I) to the rate found with ZnCl 2 alone. It is therefore unlikely that the effect of zinc on glycolysis is due to a mobilization of calcium. This conclusion is strengthened by our observation that in the presence of ZnCl 2 the O 2 consumption of hepatocytes was not different from control incubations (results not shown).
In order to exclude the possibility of an interaction of zinc with the glucagon receptor, experiments were performed with Bt 2 -cAMP. As shown in Table I (part C) Effect of Zinc on Gluconeogenesis-The effect of zinc on gluconeogenesis was examined on hepatocytes from rats starved overnight. Regardless of the gluconeogenic precursor (lactate plus pyruvate or dihydroxyacetone) and the rate of gluconeo-   genesis under control conditions, the addition of 0.08 -0.1 mM ZnCl 2 to the medium resulted in a small (16 -18%) but significant inhibition of the rate of glucose formation (Table II). With dihydroxyacetone as the glucogenic precursor, gluconeogenesis was inhibited by the addition of zinc, but the rate of lactate formation was increased about 100-fold (from 0.18 to 18.9 mol/g of wet weight/h), demonstrating directly that the flux of triosephosphates had changed in part from the gluconeogenic to the glycolytic direction.

Measurement of Free Zinc in Hepatocytes-
The free zinc concentration in eight independent preparations of rat hepatocytes was 1.26 Ϯ 0.27 M (range 0.61-2.7 M). This value increased during prolonged incubation in the presence of zinc, especially in cells showing blebbing and signs of beginning cell death.

Effect of Various Divalent Cations on Zinquin
Fluorescence-The addition of ZnCl 2 led to a concentration-dependent increase in epifluorescence, which was abolished by the presence of diethylenetriaminepentaacetic acid. CdCl 2 or PbCl 2 at 8 M were as potent as ZnCl 2 at 20 M in raising Zn-dependent epifluorescence. NiCl 2 at 80 M was without effect (Fig. 3). This rank of cations resembles their "zinc-like" effects on carbohydrate metabolism and resembles their ability to mobilize zinc from Zn-MT (26). Apart from Zn 2ϩ , only Cd 2ϩ shows a weak reaction with zinquin (12). The quantum yield in emitted light would be less than 30% of that observed in the presence of equimolar Zn 2ϩ . The observed effect clearly exceeds this proportion. All other cations tested do not react with the fluorophore.
Short Term Effects of Hormones or cAMP on Cytosolic Zn 2ϩ -Glucagon (10 Ϫ7 M) induced a rapid decrease in free zinc concentration (Fig. 4A) that could be mimicked by cAMP (0.1 mM) (Fig. 4B). The effect was distinct within 10 min and clearly present at glucagon concentrations as low as 10 Ϫ9 M (Fig. 4A, inset) but was partly restored at 20 min. In accordance, Zn 2ϩ levels measured in the presence of the slowly degradable analogue 8-Br-cAMP remained more stable (Fig. 4B).
To assess whether the rapid decrease in cytosolic Zn 2ϩ was due to an increased binding of the cation by MT, the potential of Cd 2ϩ (8 M) to discharge zinc from MT was investigated. The ability of Cd 2ϩ to raise cellular Zn 2ϩ was similar in the absence or the presence of cAMP, indicating that a portion of zinc not associated with MT is reduced (not shown).
Cross-over Studies-In order to locate the points of control at which zinc acts on carbohydrate metabolism, intracellular metabolite concentrations were determined in hepatocytes incubated with and without 0.1 mM ZnCl 2 , in the presence or the absence of glucagon. The data are given in Table III. In the absence of glucose and in the presence of zinc, the glucose production was increased and the glycolytic flux was doubled, indicating enhanced glycogenolysis. This is supported by the almost 2-fold increase in concentration of hexose-6-phosphates (glucose-6-phosphate plus fructose-6-phosphate) in the presence of ZnCl 2 . Under this condition the level of Fru-1,6-P 2 was also increased.
A distinct cross-over point at pyruvate kinase indicates that the conversion of phosphoenolpyruvate to pyruvate was increased in the presence of zinc. Furthermore, the data show that changes in levels of metabolites caused by glucagon are counteracted by the addition of ZnCl 2 .
Effect of Zinc on the Concentration of Fructose-2,6-bisphosphate-Fru-2,6-P 2 as a positive effector for phosphofructokinase-1 and a negative effector for fructose-1,6-bisphosphatase plays a major role in the control of glycolysis and gluconeogenesis in rat liver (see Ref. 27 for a review). In two experiments with hepatocytes from fed rats, we measured a concentration of FIG. 4. Effect of glucagon or 8-bromo-cAMP on Zn 2ϩ level in isolated rat hepatocytes. Hepatocytes were preloaded with zinquin as given under "Experimental Procedures." Cells (25-30 mg of wet weight/ml) were thermoequilibrated at 37°C in a shaking water bath. After 5 min a sample of 50 l was withdrawn for immediate measurement of epifluorescence, the agonists were added, and the incubation was continued for 20 min. At 10 and 20 min, identical aliquots were removed for epifluorescence measurements. Other conditions were as given in Fig. 3. The epifluorescence at 390 nm excitation was averaged for 20 -30 cells/dish, and the average in autofluorescence of again 20 -30 cells/dish was subtracted. The calculation of free Zn 2ϩ was done as given under "Experimental Procedures." Representative experiments are depicted. A, glucagon, 10 Ϫ7 M; B, 8-bromo-cAMP, 0.1 mM. Open symbols, control; filled symbols, agonist. Inset, dose dependence of glucagon effects on free Zn 2ϩ concentration. The conditions were as given above, except that 34 mg of wet weight/ml were used. The samples were taken after 10 min of incubation. 10.6 and 12 nmol of fru-2,6-P 2 /g of wet weight under control conditions. In the presence of ZnCl 2 (0.03 and 0.1 mM), the values were almost identical (11.2 and 12.1 nmol/g of wet weight, respectively), although the net lactate formation in these experiments increased from 21 to 38 mol/g/h. In the presence of Bt 2 -cAMP (0.1 mM), the concentration of fru-2,6-P 2 decreased by about 50% (to 5.5 and 6.7 nmol/g of wet weight, respectively), but in the presence of Bt 2 -cAMP (0.1 mM) plus ZnCl 2 (0.1 mM), the concentration of fru-2,6-P 2 was restored almost to control values (9.5 and 9.3 nmol/g of wet weight).
Effect of Zinc on Glycogenolysis-Because all experiments described so far in this paper clearly showed that zinc had a strong glycogenolytic effect, the activation state of phosphorylase was measured in hepatocytes under different incubation conditions. As a positive control hepatocytes incubated with Bt 2 -cAMP were also checked for their ratio of phosphorylase a/b. The results are given in Table IV. Zinc as well as Bt 2 -cAMP increases the relative amount of phosphorylase a from about 35% to 60% of total phosphorylase activity. The active portion of phosphorylase in the control hepatocytes (35%) is somewhat higher than the 20% normally reported for the intact liver (20).
The observed increase in glycogenolysis and glycolysis by zinc could be due to a stimulation of phosphorylase b and of phosphofructokinase-1 by a rise in AMP. In our experiments the ATP concentrations were between 1.5 and 2.1 mol/g of wet weight, and the AMP concentrations were about 34 nmol/g of wet weight. Both parameters did not change considerably under conditions where the rate of lactate production increased 3-fold due to the addition of ZnCl 2 . The ATP/ADP ratios were between 3.02 and 3.5 and also not affected by the presence of zinc.
Phosphorylation State of Pyruvate Kinase, Glycogen Phosphorylase, and Fructose-1,6-bisphosphatase-Because zinc appears to act partly by increasing the activities of glycogen phosphorylase and pyruvate kinase, the phosphorylation states of these proteins and of fructose-1,6-bisphosphatase were examined. An enhanced protein dephosphorylation would explain the effect of zinc on pyruvate kinase, although an activation of a protein phosphatase per se would contradict the simultaneously observed enhanced glycogenolysis, because, in contrast to the active dephospho form of pyruvate kinase, glycogen phosphorylase is less active in its dephosphorylated form. Hepatocytes were preincubated with [ 32 P]P i for 30 min. Thereafter, these cells were incubated in the presence and the absence of 0.1 mM ZnCl 2 with and without 0.1 mM Bt 2 -cAMP. Pyruvate kinase, fructose-1,6-bisphosphatase, and phosphorylase were precipitated by antisera as described under "Experimental Procedures." In the presence of Bt 2 -cAMP, the phosphorylation of pyruvate kinase increased, but zinc did not affect the phosphorylation of this enzyme (Fig. 5). In contrast, incubation with ZnCl 2 resulted in a clearly higher proportion of phosphorylated phosphorylase, an effect that was comparable with that of cAMP. Zinc and cAMP had no additive effects. These results agree with the increased enzyme activity of phosphorylase (Table IV). Fructose-1,6-bisphosphatase was already phosphorylated under control conditions, and a further increase in phosphorylation with either cAMP or with zinc could not be observed (Fig. 5).

on metabolite concentration of hepatocytes in the presence and the absence of glucagon
Hepatocytes from fed rats were incubated as described under "Experimental Procedures" without and with 0.1 mM ZnCl 2 in the presence or the absence of 10 Ϫ7 M glucagon. After 20 min, the incubations were stopped by the addition of 0.1 volume of 35% HClO 4 . The values are the means Ϯ S.E. of three independent cell preparations incubated under all four conditions. Further details are given under "Experimental Procedures." The abbreviations are: G6P, glucose-6-phosphate; F6P, fructose-6-phosphate; FBP, fructose-1,6-bisphosphate; DHAP, dihydroxyacetone phosphate; GAP, glyceraldehyde-3-phosphate; 3PG, 3-phosphoglycerate; 2PG, 2-phosphoglycerate; PEP, phosphoenol pyruvate.   Hepatocytes from fed rats were preincubated with [ 32 P]P i (1 mCi/g of wet weight) for 30 min. The medium was changed, and the cells were thereafter incubated without (lanes 1) and with 0.1 mM ZnCl 2 (lanes 2), with 0.1 mM Bt 2 -cAMP (lanes 3), and with Bt 2 -cAMP plus 0.1 mM ZnCl 2 (lanes 4), as given under "Experimental Procedures." After 20 min of incubation, cells were sedimented and extracted with 0.2% triton X-100. From aliquots (0.1 ml) of the supernatants pyruvate kinase, glycogen phosphorylase, and fructose-1,6-bisphosphatase were immunoprecipitated, separated by SDSpolyacrylamide gel electrophoresis, and autoradiographed as described under "Experimental Procedures." F, fructose-1,6-bisphosphatase; GP, glycogen phosphorylase; PK, pyruvate kinase. DISCUSSION We have shown here that incubation of hepatocytes from fed rats with zinc in the medium strongly enhances the formation of lactate and glycogenolysis. The effect depends on the concentration of ZnCl 2 added to the medium and is significant at concentrations of 33 M and higher. A comparable effect of zinc was also seen with KRB buffer, provided it had been supplemented with enough maltol. The failure of Rognstad (5) to observe this effect was most probably due to the insolubility of zinc in a medium containing phosphate and bicarbonate. Using his medium we were unable to detect zinc in the solution by atomic absorption spectroscopy, even with concentrations as high as 0.5 mM ZnCl 2 . Therefore, genuine zinc effects cannot be expected under these conditions without an appropriate membrane-permeable chelator (i.e. maltol).
The change in intracellular free zinc is much lower than the changes of total zinc found in the medium after the equilibration period (Fig. 1). In liver cytosol various proteins (e.g. MT) and low molecular weight compounds will act as buffers by binding large amounts of the added zinc, depending on their stability constants, which in general are higher than that of the maltol-Zn complex, which is 5.62 (26).
We have good evidence that the effects of zinc on glycogenolysis and glycolysis are specific for zinc ions. Ca 2ϩ , Mg 2ϩ , Mn 2ϩ , Fe 2ϩ , Ni 2ϩ , and Co 2ϩ were ineffective. CuCl 2 , however, at comparable concentration and CdCl 2 or PbCl 2 at 10-fold lower concentration showed the same effects as ZnCl 2 . Those cations showing the zinc-like effect on metabolism did indeed raise intracellular free zinc (Fig. 3). It seems feasible to conclude that these metal ions mobilize Zn 2ϩ from Zn-MTs. This conclusion is supported by in vitro displacement studies (26) that show that Co 2ϩ and Ni 2ϩ could not displace zinc from Zn-thionein, whereas Cd 2ϩ , Pb 2ϩ , and Cu 2ϩ could. The rank of affinities in this study was: Cd Ͼ Pb Ͼ Cu Ͼ Zn Ͼ Ni Ͼ Co. The metabolic effects observed here as well as the data on intracellular Zn 2ϩ mobilization fit exactly into this order of affinities. Furthermore, using the isolated perfused liver, Kingsley and Frazier (28) have shown that exposure to cadmium increases both zinc secretion into the perfusate and biliary excretion of zinc.
Here we have determined the concentration of free Zn 2ϩ by monitoring single cell epifluorescence using zinquin as an intracellular probe. The advantage of this technique as compared with bulk measurements of zinquin fluorescence in cell suspension (12,13) is that specific hepatocytes of uniform size and round shape and lacking signs of cell damage can be examined for zinquin-dependent fluorescence. Recently, Zalewski et al. (12) reported that apoptotic HL-60 cells revealed bright zinquin fluorescence, indicating an increased zinc concentration, membrane blebbing, and often a decrease in cellular volume. Conversely, zinquin seems to escape from necrotic hepatocytes, these cells showing therefore no or a markedly reduced fluorescence. Coyle et al. (13), using zinquin to follow changes in labile zinc in hepatocytes in primary culture or in liver slices treated with dexamethasone and interleukine-6 to induce metallothionein synthesis, have advised not to compare samples with different concentrations of MT. All effects reported here are short term effects (ϳ30 min), where only negligible changes, if any, in MT levels should occur. In freshly isolated rat hepatocytes, induction of MT by Zn 2ϩ was detectable after 2 h. Glucagon was less effective increasing MT synthesis only by 28 -35% after 5 h (23).
The concentration of free Zn 2ϩ reported here (about 10 Ϫ6 M) is considerably higher than would be predicted from measurements with enzymes, such as alkaline phosphatase and alcohol dehydrogenase, both being metalloproteins with extremely high affinities for Zn 2ϩ (11). Although we cannot exclude that some of the zinc we are detecting could have been derived from Zn-MT, the observed changes in Zn 2ϩ -dependent fluorescence mediated by glucagon or Cd 2ϩ or Pb 2ϩ do not support a major interference.
Both zinc and glucagon stimulated glycogenolysis due to an enhanced phosphorylation of glycogen phosphorylase. The mechanism by which intracellular Zn 2ϩ activates phosphorylase kinase has to be further investigated.
With respect to the fate of glucose-6-phosphate, zinc and glucagon (or cAMP) acted in completely different ways. Glucagon shifted glucose-6-phosphate mainly in the direction of glucose formation and inhibited the glycolytic pathway, whereas zinc, in addition to the enhancement of glycogenolysis, also markedly stimulated glycolysis. In combination, Zn 2ϩ acted antagonistically to the hormone and was able to overcome the inhibition of glycolysis by glucagon or by Bt 2 -cAMP.
The rapid decrease in intracellular Zn 2ϩ concentration in the presence of glucagon (Fig. 4) is a new and unexpected observation and offers an explanation for the antagonistic effects of glucagon and zinc on glycolysis. The rapid decrease in intracellular free zinc was also seen with 8-bromo-cAMP (Fig. 4), indicating that the action of glucagon on intracellular Zn 2ϩ is mediated via cAMP and most probably catalyzed by protein kinase A. This could lead to a sequestration of the cation.
In consequence, in order to locate the points at which zinc acts on carbohydrate metabolism, we focused on points where phosphorylation of enzymes is known to be of regulatory importance. We have analyzed the activity and grade of phosphorylation of glycogen phosphorylase and pyruvate kinase and determined the concentration of fru-2,6-P 2 , an indicator of the activity of the bifunctional enzyme fructose-2,6-bisphosphatase/phosphofructokinase-2. The proportion of phosphorylase a increased in the presence of Zn 2ϩ as well as in the presence of glucagon from 35% to about 60% of total phosphorylase activity (Table IV). This could be verified in intact hepatocytes by [ 32 P]P i incorporation into phosphorylase (Fig. 5). Moreover, the effect of zinc and Bt 2 -cAMP on the phosphorylation of glycogen phosphorylase was not additive, suggesting that zinc acts rather via an increased activity of phosphorylase kinase and not by an inhibition of the protein phosphatases 1 or 2a. That the effects of zinc and of Bt 2 -cAMP on phosphorylase activation are comparable and not additive can also be calculated from the metabolic rates given in Table I by converting the rates of glucose and lactate production into glycosyl units (not shown). However, the mechanism by which the increase in cytosolic Zn 2ϩ activates the phosphorylase kinase remains to be elucidated. A binding of Zn 2ϩ by calmodulin has been reported (29), but the measured affinity (K D about 100 M) seems to be too low to be of physiological importance.
The mechanism by which zinc activates glycolysis remains obscure. Based on the metabolite measurements reported here, zinc exerts its effect most likely at the steps catalyzed by pyruvate kinase and by phosphofructokinase-1. The substrate concentration (fructose-6-phosphate) of the rate-limiting enzyme phosphofructokinase-1 was doubled, and at the step of pyruvate kinase the inverse changes in substrate and product concentration (cross-over) indicate another control point at this step.
The classical explanation for metabolic regulation by allosteric activators does hold for the antagonistic effect of zinc on the inhibition of glycolysis by glucagon. The most potent activator of phosphofructokinase-1, namely fru-2,6-P 2 , was reduced by glucagon from 12 to about 6 nmoles/g of wet weight, whereas in the presence of glucagon plus zinc the concentration of this effector was restored to 9 nmoles/g of wet weight. This would lead to a higher flux through phosphofructokinase-1 in the presence of glucagon plus zinc as compared with glucagon alone. The effect of zinc on the fru-2,6-P 2 level in the presence of Bt 2 -cAMP and the failure to get an effect in the absence of Bt 2 -cAMP may well be explained by a counteraction of Zn 2ϩ exclusively on the phosphorylated and therefore less active form of phosphofructokinase-2. We have shown that under this condition protein kinase A is as active as in the absence of ZnCl 2 (see phosphorylation of pyruvate kinase, Fig. 5), and phosphofructokinase-2 should be phosphorylated as well, unless there exists an as yet unknown specific and Zn 2ϩ -dependent protein phosphatase.
It is also possible that the higher concentration of fructose-6-phosphate, which is almost doubled in the presence of ZnCl 2 , has more effect on the phosphorylated than on the unphosphorylated form of phosphofructokinase-2. The mechanism of Zn 2ϩ action on this enzyme seems to be different from that of vanadate, which also raises glycolytic flux in hepatocytes (30) but exerts also glycogenolytic effects by activating glycogen phosphorylase and inactivating glycogen synthase (31). The "insulin-like" effect of vanadate on glycolysis has been related to the increase in fru-2,6-P 2 . Rider et al. (32) have found vanadate to be an inhibitor of chicken liver fructose-2,6-bisphosphatase, which offers a likely explanation for the increased flux through phosphofructokinase-1 in the presence of vanadate. However, this explanation is not valid for the pronounced glycolytic effect of zinc on its own, because under this condition the concentration of fru-2-6-P 2 did not change.
The inhibition of glycolysis by glucagon at the step of pyruvate kinase has been explained by the increased protein phosphorylation. The concentration of zinc, however, which overcame the glucagon-induced inhibition of glycolysis, did not at all affect the glucagon-induced phosphorylation of the enzyme. This makes it also unlikely that a protein phosphatase is involved in the antagonistic effect of zinc. In the presence of zinc alone, pyruvate kinase remains dephosphorylated (Fig. 5) and should be in its active form. Therefore, the elevated fru-1,6-P 2 concentration measured under this condition (15-23 M) should not activate pyruvate kinase more than under control conditions, because the fru-1,6-P 2 concentration necessary for half-maximal stimulation is reported to be much lower. K 0.5 values of 0.06 M and 0.13 M fru-1,6-P 2 are given for the unphosphorylated and the phosphorylated form of the purified rat liver pyruvate kinase, respectively (33). Even the highest K 0.5 values we could find in the literature, resulting from a study using crude extracts from hepatocytes (4.2 and 9.8 M fru-1,6-P 2 for control and for hepatocytes incubated with glucagon, respectively (34)), would not readily explain the activation of glycolysis in the presence of zinc alone.
The stimulation of lactate formation in hepatocytes by zinc could also result from a strong inhibition of fructose-1,6bisphosphatase, a regulatory enzyme in gluconeogenesis. This enzyme is strongly inhibited by zinc in vitro at concentrations below 1 M (35), but it has also been reported that zinc at 0.5-2 M counteracts the inhibition of fructose-1,6-bisphosphatase by fru-2,6-P 2 , shifting the K i from 3 M to values higher than 50 M (36). In our experiments an inhibition of glucose formation from pyruvate, alanine, or dihydroxyacetone in hepatocytes from starved rats was indeed observed but only at concentrations of 0.1 mM ZnCl 2 or higher, and this inhibition never exceeded 15-20% (see Table II). At lower concentrations of zinc, an inhibition of gluconeogenesis was not observed, although at these concentrations a significant stimulation of glycolysis occurred with hepatocytes from fed rats. In hepatocytes from fed rats, the fructose-1,6-bisphosphatase was already in the phosphorylated state. Although this enzyme is a substrate for pro-tein kinase A in vitro, the incubation with Bt 2 -cAMP did not lead to a further incorporation of phosphate into the enzyme (Fig. 5). This observation resembles results from our earlier studies analyzing the phosphorylation of phosphofructokinase-1 in hepatocytes from fed and starved rats. In those experiments the degree of phosphorylation was clearly determined by the metabolic state rather than by the activation of the protein kinase A (37).
In summary we show here for the first time that effects on carbohydrate metabolism induced by Zn 2ϩ , Cd 2ϩ , or Pb 2ϩ in the medium are accompanied by an intracellular increase in free zinc and that glucagon or cAMP drastically reduce the free zinc concentration. Our observation that zinc counteracts the effect of cAMP on glycolysis as catalyzed by protein kinase A is in line with the concept that Zn 2ϩ plays a regulatory role in signal transduction. For instance, zinc has been shown to induce the translocation of protein kinase C to membranes that will convert the enzyme into a more active form due to binding of zinc to the cysteine-rich domains required for the formation of the phorbol ester binding site (38,39). In Chinese hamster ovary cells overexpressing protein kinase C ␣ , phorbol ester activation led to an increased phosphorylation of the insulin receptor. This serine/threonine phosphorylation by excessive protein kinase C seems to inhibit insulin-stimulated responses (40).
We have shown earlier that the 11.5-kDa ZnBP binds in vitro Zn 2ϩ -dependently to several enzymes of the carbohydrate metabolism, including phosphofructokinase-1 and fructose-1,6bisphosphatase. In liver these two regulatory enzymes are both covalently modified by phosphorylation but without measurable effects on the kinetic of these enzymes. It is tempting to speculate that zinc ions could transduce certain effects of glucagon and/or insulin, permitting a cross-talk between these, in many respects, antagonistic hormones. However, there is evidence that insulin-like effects of Zn 2ϩ on adipocytes occur by a mechanism unrelated to insulin (9). But in contrast to adipocytes, hepatocytes have the capacity for both glycolysis and gluconeogenesis and have a high zinc content and specific types of phosphorylatable enzymes, so that the regulation mechanisms by insulin in hepatocytes may differ from those in adipocytes.