Evidence for Electron Transfer Reactions Involved in the Cu’+-dependent Thiol Activation of Fat Cell Glucose Utilization*

Abstract Previous studies indicated that the ability of thiols to mimic insulin action on fat cell glucose oxidation and lipolysis was dependent on Cu2+. Evidence now presented indicates that H2O2 formed by the reaction of thiols, Cu2+, and O2 mediates the effects of these agents on isolated fat cells. Oxidation of thiols to the respective disulfides in the presence of Cu2+ accompanied the stimulatory effects on glucose metabolism. Diphenyl-1, 10-phenanthroline, which binds tightly to Cu+ but not Cu2+, blocks copper-catalyzed thiol oxidation and the stimulatory effect on fat cell glucose utilization, indicating that reduction of Cu2+ is an obligatory step for thiol action. The thiol-Cu2+ effect on fat cells is readily reversible since stimulated rates of glucose oxidation returned to control levels following addition of the chelator. Catalase inhibited the stimulatory effect of thiols but not insulin on fat cell glucose oxidation, whereas H2O2 mimicked the action of thiols and insulin on this process. As little as 10 µm H2O2 stimulated glucose utilization, whereas 1 mm H2O2 was maximally effective. Higher concentrations were less effective. The stimulatory effect of various mercaptoethanol and Cu2+ concentrations on labeled CO2 production from d-[1-14C]glucose by fat cells paralleled the peroxide formed under these conditions. Other oxidants such as MnO4= and diamide were also effective in enhancing fat cell glucose oxidation. The stimulatory effect of all three oxidants, as well as insulin, was inhibited by phlorizin and cytochalasin B, known glucose transport inhibitors. Fat cells enhanced the decomposition of added H2O2 or that generated by cysteine, Cu2+, and O2. These data are consistent with the concept that an electron transfer reaction between a fat cell component or components and oxidants results in stimulated glucose transport or metabolism or both.

Among the large number of agents which have been reported to mimic the ability of insulin to stimulate glucose metabolism and inhibit lipolysis in isolated fat cells are the thiols (1, 2). We have previously reported that the effects of thiols on fat cell function were dependent on Cu2+, a contaminant of albumin preparations used in fat cell incubation media (2). The data presented in this report indicate that HzOz, which is produced nonenzymatically by the combination of thiols, Cut+, and 02 under the conditions used, mediates the biological actions of these agents. Further, other oxidants such as Mn04= and diamide (diazenedicarboxylic acid bis-N, N-dimethylamide) also stimulated glucose utilization in isolated fat cells.

METHODS
White fat cells were obtained by enzymatic digestion of the parametrial adipose tissue from 130 to 150-g female rats (Charles River CD strain) fed laboratory chow ad Zibitum (3). For each experiment the parametrial adipose tissue from two or more rats was pooled and then cut into small pieces with scissors and blotted, and the desired amount added to l-ounce polyethylene bottles.
Depending on the requirements of the particular experiments 1 to 3 g of tissue were incubated in each bottle with 3 to 8 ml of 37, albumin in phosphate buffer and 1 mg of crude collagenase per ml (Ctostridium histolyticum, Worthington) at 37" for 1 hour. The phosphate buffer contained 128 mM NaCl, 1.4 mM CaC&, 1.4 mhz MgS04, 5.2 mM KCl, 10 rnM NazHP04, pH 7.4, and 3 7. bovine serum albumin, Fraction V (Armour).
At the end of the digestion period cells were filtered through cheesecloth and washed twice with 6 ml of 1% albumin buffer at 37". The fat cells were resuspended in albumin buffer and incubated in plastic culture tubes (16 mm x 100 mm) at 37" in a shaking water bath. The final incubation mixture volume in each tube was 1.2 ml and contained 0.1 or 0.2 mM D-[l-'4c]glUcase. The reaction was stopped by addition of 0.2 ml of 0.5 M HzS04.
Glucose conversion to COZ was determined as described by Fain et al. (4). The values for each experiment are the averages of duplicate tubes and are based on changes during the incubation period over those of initial controls incubated for 5 min. The data presented in this report for each esperimental design are the averages of two or more experiments performed on different days.
Cuzf at a concentration of 2 X 1O-5 M in 1 ml of Krebs-Ringer phosphate buffer containing diphenyl-1 , IO-phenanthroline (1 mM), was reduced by 1 mM cysteine, and the time course of reduction of Cu2+ was measured at 479 nm. In other tubes 1 mM cysteine was added to 1 ml of 3% albumin in Krebs-Ringer buffer and the reduction of copper which contaminated the albumin and was accessible to cysteine was measured similarly.
Oxidation of mercaptoethanol was monitored using the sulfhydryl blocking reagent 6,6'-dithiodinicotinic acid by following the absorbance at 344 nm (6). Mercaptoethanol was incubated in 1.2 ml of buffer under the conditions studied (Figs. 4 and 6), and 50.~1 aliquots were withdrawn at the appropriate times and added to 10 ~1 of 10 mM 6,6'-dithiodinicotinic acid. After dilution with 1 ml of Hz0 the absorbance was measured at 344 nm. Essentially no difference was observed in the rate of mercaptoethanol oxidation in the presence or absence of fat cells.
HzOz was assayed as follows.
One-milliliter aliquots of buffer with appropriate additions of thiols and Cu2+ (Fig. 6) were added to the assay mixture.
This mixture was allowed to stand at room temperature for 30 min before adding 1 drop of 4 ;"; HCl.
After an additional 30 min the absorbance was measured at 420 nm. For the experiments described in Fig. 7 aliquots of fat cell in cubation medium were filtered on glass fiber filters (Whatman GF/C) to free the aliquots from cells before assaying.
n-[l-14C]glucose was obtained from New England Nuclear Corp. Diphenyl-1 , lophenanthroline was obtained from the G. Frederick Smith Chemical Co., Columbus, Ohio, and 30% HzOz from Allied Chemical.
Diamide was obtained from Calbiochem and 6,6'-dithionicotinic acid from Newcell Biochemicals, Berkeley, Calif. RESULTS We have previously demonstrated that albumin preparations which were freed of heavy metal contaminants by reaction with o-phenanthroline and passage through Sephadcx G-50 were un able to support the thiol effect on fat cell glucose oxidation; addition of divalent copper to the treated albumin restored the potentiation of cysteine action (2). Fig. 1 shows the results of experiments designed to test the effect of various concentrations of Cu2+ on mercaptoethanol-enhanced glucose oxidation in the presence of a very low albumin concentration (0.1 yO). This amount of albumin was able to potentiate the thiol effect only slightly.
Under these conditions 30 PM Cu2+ maximally potentiated the effects of 0.2 mM and 1 mM mercaptoethanol on fat cell glucose utilization.
In contrast to our earlier observations using higher albumin levels, 0.1 mM Cu2+ itself markedly stimulated glucose oxidation by fat cells incubated in 0.1% albumin.
Mercaptoethanol at both concentrations inhibited this effect of Cu2f by about 50%.
In order to determine whether the permissive effect of Cu2+ on t.hiol action involves reduction of the metal to the monovalent form we tested the effect of diphenyl-1 ,lO-phenanthroline. Cysteine (1 mM) was added to 1 ml of Krebs-Ringer buffer containing 2 X 10m5 M Cu2+ and 1 rnM diphenyll,lO-phenanthroline and the time course of absorbance at 479 nm was measured (A-A).
The time course of reduction of Cu2+ contaminating albumin preparations by cysteine was monitored by addition of 1 mM thiol to 1 ml of 3c/; albumin buffer and 1 rnM diphenyl-1, lo-phenanthroline No increase in absorbance of the chelator in the absence of added cysteine was found b---m). monitored by diphenyl-1 , 10.phenanthroline absorbance at 479 nm, rapid reduction of Cu2+ by thiols was found to occur (Fig.  3). The amount of Cu%ontaminating albumin which is accessible to reduction by cysteine was approximately 6.6 nm per mg of albumin. Fig. 4 (top panel) presents the time course of labeled COZ production by fat cells incubated in the presence of 1 mM mercaptoethanol.
Diphenyl-1 , 10.phenanthroline added at 15 min reversed thiol-activated glucose oxidation back to control levels. The chelator had essentially no effect on control levels of fat cell glucose oxidation.
Oxidation of mercaptoethanol accompanied the stimulation of glucose utilization over the 60.min incubation period and this oxidation was also blocked by diphenyl-1 , 10.phenanthroline.
Under the conditions of this series of experiments about 65% of the initial amount of mercaptoethanol was oxidized (Fig. 4). The presence of fat cells had little or no effect on the rate of mercaptoethanol oxidation in 2% albumin buffer.
We have previously presented evidence consistent with the concept that some substance formed by the reaction of thiols and Cu2+ mediates the stimulatory effect on fat cell glucose oxidation (2). The data presented in Table I indicate that this mediator appears to be HzOz formed by the interaction of CL?, thiol, and OZ. The effect of cysteine and mercaptoethanol on fat cell glucose utilization was blocked by catalase, but there was no effect of catalase on insulin action nor basal glucose oxidation.
Further, 1 mM H202, like insulin and thiols, enhanced labeled CO* production by fat cells and this effect was also abolished by catalase (Table I). When H202 was added to fat cells at zery) time and glucose oxidation monitored for 15 min, a biphasic effect of the agent was observed (Fig. 5) at around 1 mM, whereas 4 mM HzOz was significantly less effective (Fig. 5).
Further evidence which suggests that Hz02 mediates the effect of thiols is presented in Fig. 6. Fat cell glucose oxidation in the presence of 0.2 mM and 1 mM mercaptoethanol was enhanced to a greater extent with 30 pM Cu2+ than with 3 pM.
Both the amount of thiol oxidized and the amount of HzOz formed during this 15.min incubation period paralleled the magnitude of the response under conditions identical to those used to monitor CO:! production except without cells, since in their presence Hz02 was difficult to assay due to its rapid disappearance (Fig.  7). The rate of mercaptoethanol oxidation is the same in the presence and absence of cells.' Fig. 7 presents the results of an experimental design where HzOz generated by 1 mM cysteine, 30 FM Cu*+, and O2 was assayed in the presence and absence of fat cells over a 15-min incubation period.
In the absence of fat cells the maximum net amount of H202 produced was about 110 nmoles and occurred at about 3 min. At this point, the net amount of Hz02 produced in the presence of fat cells was only about 35 nmoles and this dropped to about 10 nmoles by 5 min. This enhanced rate of HZ02 decomposition by fat cells also occurred when the Hz02 was added directly rather than generated  and Cu2+. The incubation was carried out at 37" for 15 min, and labeled CO2 production and mercaptoethanol oxidized were monitored.
HZ02 formed under the appropriate conditions was determined in duplicate tubes treated as above, but in the absence of fat cells. The amount of I-I202 formed by the reaction of cysteine with Cu2+ and Oa in 15 miu exceeds that produced by mercaptoethauol under the same conditions (Fig.  6). We have also found that the rate of H202 formation by cysteine oxidation is much more rapid than that of mercaptoethanol.
1 M. P. Czech, J. C. Lawrence, and W. S. Lynn, unpublished data. The effect of other oxidants on fat cell glucose metabolism is shown in Table II. bInOl= and diamide markedly stimulated the conversion of n-[1-i*C!]glucose to labeled COz by isolated fat cells. The stimulatory effects of these agents as well as those of Hz02 and insulin were inhibited by 2 mM phlorizin and cytochalasin Bi, which are glucose transport inhibitors (Table II). We also tested whether the effect of Cu2+ on fat cell glucose oxidation was mediated by peroxide.
This stimulatory effect of Cu2+ was much more marked at 0.1 rnaf than at 0.3 mM (Fig. 8). Neither diphenyl-1 , 10.phenanthroline nor catalase significantly inhibited the effect of Cu2+ on glucose utilization, whereas 2 mM EDTA abolished its effect. Thus, these data indicate that large amounts of CL@ do not produce Hz02; rather it seems that divalent copper at these high concentrations acts directly on fat cells to stimulate glucose utilization. DISCUSSION One approach to the study of insulin action has been the study of chemically simple or specific agents which either mimic or inhibit the effects of insulin on fat cell metabolism.
The general expericncc has been that a large number of agents with very different chemical specificities can mimic some of these insulin effects. For example, thiols (1, 2), mild proteolysis (7,8), polyamines (9), hypertonicity (lo), sulfhydryl reagents (ll), prostaglandins (12), vitamin KS (13, 14), lectins (15), and others have been reported to stimulate glucose oxidation and inhibit lipolysis in fat 1~11s. However, none of these have yet been shown to mimic all of the effects of insulin, and it is unlikely that any act exactly as does insulin.
A major disappointment in this approach has been that despite the existence of so many active agents little or ii0 progress has been made at identifying the specific cell components involved in the action of any of these agents.
Several of the nbovc agents have been tested on trypsinized cells which 110 longer respond to insulin and have been found to be still active (16-19); therefore, if their actions do share any common pathways with those of insulin, they probably occur at steps beyond the hormone-receptor interaction.
The greatest utility of studies with these agents may be to provide clues to the chemical events which occur subsequent to insulin binding and which lead to the enhancement of glucose transport.
Our previous finding that the effects of thiols were dependent on divslent copper (2) prompted us to further characterize the basis for t,his dependence.
The present findings (Fig. I) demonstrate the potentiation of t,hiol action by <'us+ at a very low albumin concentration (0.1%).
Attempts to demonstrate this effect in the complete absence of albumin have been variable,' probably due to the markedly diminished viability of fat cells under these conditions (20). Furthermore, the data presented show that albumin preparations used in the cell incubation buffer contain sufficient amounts of Cu*+ to catalyze production of Hz02 from thiols (Fig. 2).
Under 110 circumstances have we found an effect of thiols on fat cell glucose utilization without the concomitant oxidation of a portion of the sulfhydryl groups to the disulfide (Figs. 4 and 6). Divalent copper is well known for its catalysis of this oxidation r&action and its subsequent reduction of molecular osygen (21). We therefore propose that the Cu*+-dependent thiol activation of fat cell glucose oxidation is mediated by peroside formed by the following reactions.
The effectiveness of the osidants Mn04=, CL+, and diamide as well as previously reported effects of phenazine methosulfate (23) and vitamin KS (13,14) suggest that these agents as well as HzOz act by catalyzing the oxidation of a fat cell component or components.
It is known that diamide, Mn04=, and H20z can rapidly catalyze the oxidation of sulfhydryls.
It is not clear whether the actions of these oxidants reside at the level of glucose transport or intracellular glucose metabolism or both.
Lavis and Williams found that the effect of thiols was additive to that of submasimal, but not maximal, doses of insulin (1). These workers also showed that the thiols were ineffective at stimulating glucose oxidation of fat cell homogenates and suggested that the primary effect may be at the level of glucose transport (1). On the other hand, Jacob and Jandl (30) concluded from their careful studies that H202 stimulated red cell I)-[l-14C]glucose conversion to labeled CO* secondary to lowering intracellular reduced nicotinamide adenine dinucleotide phosphate levels. This effect was found to be mediated through oxidation of intracellular glutathione which is then reduced at the expense of reducing equivalents from the dinucleotide phosphat,e in a reaction catalyzed by glutathione reductase. The increased level of oxidized nicotinamide adenine dinucleotide phosphate is then thought to stimulate reactions in the pathway of glucose metabolism which form the reduced state of the dinucleotide phosphate.
That t,his series of events also occurs in fat cells upon addition of H,02 is likely, since diamide, which is a rather specific agent for oxidizing glutathione, is also a potent stimulator of glucose utilization (Table II). Whether osidants also directly or indirectly modulate the n-glucose transport system in the fat cell plasma membrane is the subject of furbher studies.
We have recently developed a method to monitor 2-deoxy-I)-glucose transport in fat cell ghosts 2 H+ + 2 Cu+ + 02 + 2 Cr12+ + IIzOz (2) which may be useful in answering this question (31) This scheme predicts that the absence of molecular oxygen should abolish the ability of thiols and Cu2+ to mimic insulin action.
We have thus tested the effects of 1 m&f cysteine and 200 punits per ml of insulin on fat cells incubated anaerobically for 20 min in 20:) albumin buffer and found that the fat cell response to the thiol was inhibited by about 95"/" whereas that to insulin was unaffcctcd.' That the resulting H202 formed from these electron transfers is the species which actually interacts with the fat cells is strongly supported by the present data. Catalasc which catalyzes the rapid decomposition of H202, inhibited the effect of thiols but not insulin on fat cell glucose metabolism, whereas added H202 mimicked the effect of t,hc thiols (Table I). Further, the amount of peroxide formed by various concentrations of mcrcaptoethanol and Cu2+ was proportional to the magnitude of the fat cell response (Fig. 6).
Of interest was the finding that the enhanced fat cell glucose oxidation rates which occurred in response to thiol-Cu*+ returned to control levels upon inhibition of further peroxide formation by diphcnyl-1 , 10phcnanthroline (Fig. 4). This ready reversibility of thiol action argues against the involvement of irreversible damage to cells due to, for example, lipid peroxidation.
In support of this conclusion, Lavis et al. (22) showed that as-