Evidence for the Involvement of Vicinal Sulfhydryl Groups in Insulinactivated Hexose Transport by 3T3-Ll Adipocytes*

Transcription

1 THE JOURNAL OF BIOLOGICAL CHEMISTRY by The American Society of Biological Chernista, Inc. Vol. 260, No. 5, Issue of March 10, pp Printed in ~s.a. Evidence for the Involvement of Vicinal Sulfhydryl Groups in Insulinactivated Hexose Transport by 3T3-Ll Adipocytes* (Received for publication, July 16,1984) Susan C. Frost$ and M. Daniel Lane From the Department of Biological Chemistry, The Johns Hopkins University School of Medicine, Baltimore, Maryland Following the differentiation of 3T3-Ll preadipo- fold by insulin (7). In contrast to isolated adipocyte suspencytes insulin acutely activates the rate of 2-deoxy-[ 1- sions which lose cell viability rapidly, 3T3-Ll adipocytes can l'c]glucoee uptake in the mature 3T3-Ll adipocyte by be studied in stable cell monolayers which maintain cell 15- to 20-fold. Phenylarsine oxide, a trivalent arseni- viability and hormonal responsiveness for extended periods cal that forms stable ring complexes with vicinal di- of time. thiols, prevents insulin-activated hexose uptake in a New insight into the process by which insulin stimulates concentration-dependent manner (Ki = 7 PM) but has glucose transport was provided independently by two groups no inhibitory effect on basal hexose uptake. 2,3-Di- of investigators (9, 10) using different experimental apmercaptopropanol at a level nearly stoichiometric to proaches. It was observed that upon stimulation of adipocytes that of phenylarsine oxide prevents or rapidly reverses by insulin, a membrane-associated intracellular pool of gluthe inhibition of hexose uptake; 2-mercaptoethanol, cose transporters was "translocated" to the plasma membrane, even in high stoichiometric excess over the arsenical, thereby increasing the rate of hexose uptake. The increased does not reverse inhibition of hexose uptake. When V,, for sugar uptake in the presence of insulin was, therefore, phenylarsine oxide is added after adipocytes have been attributed to an increased number of glucose transporters at fully activated by insulin, 2-deo~y-[l-'~C]gl~cose upthe cell surface. While these important findings suggest a take rate decays slowly at a rate corresponding to that basis for the stimulation of glucose transport by insulin, the caused by the withdrawal of insulin = 10 min). Using the same conditions under which phenylarsine mechanism by which insulin binding to its receptor is coupled oxide blocked activation, the K,,, for deoxyglucose upto the apparent translocation of glucose transporters to the take, the rate at which 12"I-insulin became cell-associcell surface is still unknown. ated, and the l2"1-insulin binding isotherm for solubi- Sulfhydryl groups have been implicated in insulin-activated lized insulin receptor were not affected by phenylar- hexose transport by the use of inhibitors (11-15). Thus, N- sine oxide. These results support the transporter trans- ethylmaleimide and dithiobisnitrobenzoate both inhibit inlocation model for insulin-activated hexose transport sulin-activated hexose transport (15). In addition, these and implicate vicinal sulfhydryl groups in a post-in- agents appear to block the return to the basal transport rate sulin binding event essential for the translocation of glucose transporters to the plasma membrane. The acute activation of glucose uptake by animal tissues is historically one of the best-known biochemical actions of insulin (for review, see Ref. 1). This action of insulin is limited primarily to two major cell types, ie. adipocytes and muscle cells, in which the process has been studied extensively. More recently, it was observed that during differentiation of 3T3- L1 preadipocytes into adipocytes in culture, the cells acquired the capacity to undergo insulin-activated sugar uptake (2-7). Accompanying adipose conversion the cells also exhibit a 20- fold increase in the number of functional insulin receptors/ cell (8) and an increase in the number of glucose transporters/ cell.' The use of 3T3-Ll cells to investigate the regulation of the glucose transport system offers several advantages. Fully differentiated 3T3-Ll adipocytes are particularly responsive to insulin, e.g. sugar uptake can be acutely activated * The work was supported by Research Grant AM from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. 3 Supported by Postdoctoral Fellowship AM06899 from the National Institutes of Health. 1 x 106 cytochalasin B binding sites/cell in 3T3-C2 fibroblasts versus 2 X lo6 binding sites/cell in mature 3T3-Ll adipocytes (S. C. Frost, G. E. Lienhard, and M. D. Lane, unpublished results upon removal of insulin. During an investigation of the rapid insulin-induced internalization of cell-surface insulin receptors to a trypsin-resistant cell compartment, we observed that phenylarsine oxide was an effective inhibitor of this process (16). It was also found that this inhibitor has no effect on the recycling of internalized receptor to the cell surface when insulin was withdrawn (16). In the present paper we show that phenylarsine oxide is a potent reversible inhibitor of insulin-activated deoxyglucose uptake by 3T3-Ll adipocytes. Trivalent arsenicals, such as phenylarsine oxide, are known to react specifically with vicinal sulfhydryl groups to form stable ring structures which can be reversed by vicinal dithiol competitors, e.g. 2,3-dimercaptopropanol, but not by monothiols. In this paper it will be shown that insulin-activated deoxyglucose uptake by 3T3-Ll adipocytes is inhibited by phenylarsine oxide and is reversed by 2,3-dimercaptopropanol, but not mercaptoethanol. These and other results implicate vicinal sulfhydryl groups in the coupling of insulin binding (to its receptor) to activation of the glucose transport system. A preliminary account of this work has been presented (32). EXPERIMENTAL PROCEDURES Materials 2-Deoxyglucose, 3-O-methylglucose, Fraction V bovine serum albumin: and ATPluciferin-luciferase were purchased from Sigma. Crude bovine serum albumin (Cohn's Fraction V) was charcoal treated at acid ph to remove insulin-like activity by a modification (17) of the defatting procedure by Chen (18).

2 Phenylarsine oxide and 2,3-dimercaptopropanol were from Aldrich. Insulin for promoting differentiation was from Elanco and for in vitro studies was a gift from Dr. Ronald Chance, Lilly. p-aminophenylarsine oxide was a gift from Dr. Leslie Werbel, Warner Lambert/Parke- Davis. Hexokinase was from Boehringer Mannheim. Sodium ['%I] iodide was purchased from Amersham Corp., and '261-insulinwas synthesized by the chloramine-t (19) or lactoperoxidase (manufacturer's directions; Enzymobead Reagent from Bio-Rad) method. 2- Deoxy-[l-14C]glucose was from Amersham Corp. 3-0-[methyl-"C] Glucose was from New England Nuclear. All other chemicals were of the highest quality available. Methods Cell Culture-3TB-Ll preadipocytes were grown to confluence at 37 "C in 35-mm culture dishes in DMEM? with no added biotin or pantothenate, containing 10% calf serum in incubators equilibrated with 10% COz. Two days postconfluence (day 0) differentiation was induced with methylisobutylxanthine (0.5 mm), dexamethasone (0.25 p ~ ) and, insulin (1 pg/ml) in DMEM containing 10% fetal bovine serum. After 2 days, the methylisobutylxanthine and dexamethasone were removed and insulin was maintained for 2 additional days. On day 4, and thereafter, DMEM (without insulin supplementation) plus 10% fetal bovine serum was replaced every 2 days. Before each experiment, cell monolayers were incubated in serum-free DMEM for 2 h. Cells (2 X lo6 cells/dish) were used for experimentation between days 8 and 12 at which time >95% of the cells expressed the adipocyte phenotype, and insulin-activated 2-deoxy-[l-"C]glucose uptake was maximal.' Deoxyglucose Uptake Rate-Cell monolayers were rapidly washed at 37 "C with 3 X 3.0 ml of Krebs-Ringer phosphate buffer (KRP) at ph 7.4 (NaCl, 128 mm, KCl, 4.7 mm, CaC12, 1.25 mm, MgS04, 1.25 mm, NaPO,, 10.0 mm). For activation experiments, cells were incubated at 37 "C with or without the indicated concentration of phenylarsine oxide for up to 10 min at which time insulin was added. After 10 min, a period of time sufficient to fully activate hexose transport, hexose uptake was initiated by the addition of 0.2 mm 2-deoxy-[l- "C]glucose(0.5 mci/mmol) for the indicated time. Deoxyglucose uptake was linear for min. Uptake was terminated with three rapid 3.0-ml washes of ice-cold phosphate-buffered saline after which cells were dissolved in 0.1% sodium dodecyl sulfate and counted. To examine the effect of phenylarsine oxide on cells already activated by insulin, monolayers were incubated with insulin for 10 min, washed four times with 3.0mlof buffer containing 0.1% bovine serum albumin, and covered with buffer either with or without phenylarsine oxide. Hexose uptake activity was then measured at various times thereafter in 2.5-min uptake assays. The data points in the figures are the average of duplicate determinations. The standard deviation was less than 5%. Day to day variation was less than 25%, the trend always being the same. 3-0-Methylglucose Uptake Rate-Cells were grown and differentiated on 15-mm plastic coverslips (Thermanox, Miles Scientific) set into Lindbro culture dishes (5 X 106 cells/disc). Cells were preincubated in the presence or absence of insulin and/or phenylarsine oxide. Coverslips were placed in a vertical holder designed according to Norton and Munck (20). The holder was lowered manually into a radioactive solution containing 3-0-[methyl-14C]glucose (50 p ~ 4.8, mci/mmol) and the appropriate hormone and/or inhibitor. After the indicated time, the coverslips were passed rapidly through a series of three washes at 25 "C containing phosphate-buffered saline and 0.3 mm phloretin to terminate sugar transport. With each new solution, the holder was quickly raised and lowered five times within the solution resulting in rapid equilibration of the buffer over the surface of the cells (20). The coverslips were transferred to 10-ml glass scintillation vials containing 1.0 ml of 0.1% sodium dodecyl sulfate. After completion, 10 ml of scintillation fluor were added, and vials were shaken until the solution was clear. Fksults were corrected for entrapped intercellular radioactivity (4% of total) using controls run in the presence of 20 FM cytochalasin B. All data are expressed as the average of quadruplicate determinations. The standard deviation was less than 10%. Cellufar ATP Determination-Cell monolayers (2.0 X lo6 cells/35- The abbreviations used are: DMEM, Dulbecco's modified Eagle's medium; HEPES, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid. 'S. C. Frost, I. A. Simpson, S. W. Cushman, and M.D. Lane, manuscript in preparation. Vicinal Sulfhydryls in Insulin-activated Sugar Uptake 2647 mm culture dish) were treated with 1.0 ml of cold 6.0% perchloric acid (w/v) to extract ATP (21). After centrifugation, an aliquot of the supernatant was neutralized with cold 5 M K2COs and held at 0-2 'C for 30 min. After centrifugation, the supernatant was analyzed for ATP content by the luciferase assay method described by Sigma. Sample preparation and ATP assays were performed on the same day. Phenylarsine oxide did not interfere with the luciferase assay for ATP. Cellular Association, Uptake, and Degradation of '26Z-Znsulin-Cells were preincubated in the presence or absence of 20 p~ phenylarsine oxide for 10 min at 37 "C. At zero time, "'I-insulin (k5 p~ unlabeled insulin to assess nonspecific binding) was added in 0.1% bovine serum albumin, and at various times thereafter cells were extracted with trichloroacetic acid (10% w/v). After centrifugation, the pellets and supernatants were counted to determine trichloroacetic acid-soluble (degraded lz1-insulin) and trichloroacetic acid-precipitable (unde- graded 'T-insulin) radioactivity. All results are reported as "'1- insulin that was specifically cell associated or taken up. lz6z-znsulin Binding-To determine cell-surface lzi-insulin binding capacity, cell monolayers were incubated in KRP with or without 20 PM phenylarsine oxide at 37 "C for 10 min. Cell monolayers were then washed to remove excess arsenical, cooled to 4 "C, and incubated overnight (12-15 h) at 4 "C in a modified KRP (4) containing 0.1% bovine serum albumin with various concentrations of lzi-insulin in the absence or presence of 5 p~ unlabeled insulin to determine nonspecific binding. To determine the 12'I-insulin binding capacity of Triton X-100- solubilized insulin receptor, membranes were prepared according to Kohanski and Lane' from cells previously incubated for 10 min at 37 "C with or without phenylarsine oxide. Total cellular membranes were collected by centrifugation at 300,000 X g with 5 intermittent 0.5 M NaCl washes. Membranes were solubilized in 1.0% Triton X- 100 and assayed for '%I-insulin binding capacity at 4 "C (22) also in the presence or absence of phenylarsine oxide (20 p ~ ) Nonspecific. binding was determined in the presence of 1.7 pm unlabeled insulin. 3T3-LZ Adipocyte Hexokinase Actiuity-Cell monolayers, preincubated with or without 20 p~ phenylarsine oxide, were treated with digitonin (0.1 mg/ml KRP) for 10 min at 25 "C to release cytosolic enzymes. After centrifugation the supernatants were assayed for hexokinase activity according to Bergmeyer et al. (23) and analyzed spectrophotometrically at 340 nm. Phenylarsine oxide had no effect on the activity of yeast hexokinase from Boehringer Mannheim or of hexokinase released by digitonin treatment of 3T3-Ll adipocytes. RESULTS During differentiation in cell culture, 3T3-Ll preadipocytes acquire the characteristic insulin-activated sugar uptake of isolated adipocytes (2-7). As illustrated in Fig. la, maximal 2-deo~y-[l-~~C]glucose uptake rate by mature 3T3-Ll adipocytes is achieved within 3-4 min after activation by insulin and then proceeds at a constant rate for at least 10 min. At a saturating insulin level, deoxyglucose uptake was activated by insulin 17-fold (Fig. 1B); activation by insulin ranged from 15- to 20-fold in a large number of other experiments. The K., of insulin for the activation of deoxyglucose uptake was about 7 nm. Cytochalasin B, the classical inhibitor of facilitated glucose transport in animal cells (24,25), inhibited both basal and insulin-stimulated deoxyglucose uptake by 90-95% (Fig. la, Table I). Phenylarsine oxide, which has been found to inhibit the insulin-induced internalization of cell-surface insulin receptors (16), was tested for its effect on insulin-activated hexose transport. When added to 3T3-Ll adipocytes 10 min before insulin, 18 PM phenylarsine oxide blocked insulin-stimulated deoxyglucose uptake but had no effect on the basal uptake rate (Table I). Phenylarsine oxide had a similar inhibitory effect on deoxyglucose uptake activated by Hz02, an insulinomimetic agent which is thought to act at a post-insulin binding site (15, 26, 27). The validity of the deoxyglucose transport assay depends E. Kohanski, R. A., and Lane, M. D. (1985) J. Biol. Chem., in press.

3 2648 Vicinal Sulfhydryls in Insulin-activated Sugar Uptake W n FIG. 1. Insulin-activated 2-&xyglucose transport by 3T3-Ll adipocytes. A, 2-deoxy-[1-"C]glucose uptake (0.2 mm) was measured over 10 min in the presence (0) or absence (W) of 1 p~ insulin. Some cell monolayers were pretreated for 10 min with cytochalasin (20 PM) and assayed with (0) or without (0) insulin as above. B, cells were preincubated with various concentrations of insulin for 15 min after which 2-deoxy- [ l-l'c]glucose (0.2 mm) uptake was measured for 10 min. Half-maximal transport activation was achieved at 7 nm insulin c INSULIN, M TABLE I Inhibition of insulin- and HzOz-activated hexose uptake by phenylarsine oxide Monolayers of 3T3-Ll adipocytes were preincubated with or without 18 p~ phenylarsine oxide and/or 20 MM cytochalasin for 10 min. Monolayers were then incubated for an additional 15 min with 1 p~ insulin or 8 mm H20~. 2-Deoxy-[l-14C]glucose uptake assays were then conducted for 5 min. Treatment 2-Deoxy-[I-"C]glucose uptake Control Cytochalasin B Phenylarsine oxidetreated Cytochalasin B pnwl/loe cells/min Basal 110f Insulin f 1 Hz f upon adequate ATP and functional hexokinase levels to trap deoxyglucose taken up by cells as deoxyglucose 6-phosphate. As illustrated in Fig. 2, phenylarsine oxide alone or in combination with deoxyglucose led to only a 10 to 20% decrease in cellular ATP level during the time cells were exposed to the arsenical (see Fig. 2, arrow). Since the calculated6 cellular steady-state ATP concentration was 5-7 mm, which greatly exceeds the K,,, of hexokinase for ATP (0.1 mm; Ref. 28), it is unlikely that this change in ATP level would significantly affect the rate of deoxyglucose phosphorylation. In addition, the hexokinase activity in digitonin extracts of cells treated with 20 PM phenylarsine oxide did not differ from the activity of control cells (1.8 versus 2.0 units of hexokinase activity/ lo6 cell, respectively). Thus, the phenylarsine oxide-induced decrease in the transport rate of insulin-stimulated cells is not the result of an inability to phosphorylate deoxyglucose. In contrast to phenylarsine oxide which inhibited insulinactivated, but not basal hexose uptake, cytochalasin B inhibited both basal and insulin-stimulated deoxyglucose uptake in the presence or absence of phenylarsine oxide (Table I). These results (and those to be described below), indicate that This calculation uses a value for the intracellular water volume of 2.0 pl/106 cells determined from 3-O-[rnethyl-14C]glucose uptake (at equilibrium) experiments presented in this paper daoxyglucosr I 1 I I FIG. 2. Effect of phenylarsine oxide and 2-deoxyglucose on the concentration of cellular ATP. 3T3-Ll adipocyte monolayers were preincubated with or without 18 PM phenylarsine oxide for 10 min. To some cells, 0.2 mm 2-deoxyglucose was added 2.5 min prior to acidification to mimic transport assay conditions. At the indicated times, the incubations were terminated with 1.0 ml of 6% perchloric acid (w/v). After centrifugation, an aliquot was neutralized with cold KzC03 (5 M). The clear supernatant was then analyzed for ATP content as described in the text. The arrou indicates the time that cells were exposed to the arsenical in most of the experiments presented in this study. bas0, phenylarsine oxide. the action of phenylarsine oxide is focused only on the activation of hexose transport and not on basal hexose transport per se. The dependence of the inhibition of deoxyglucose uptake on phenylarsine oxide concentration was investigated both at 10 nm insulin, near the K., for sugar uptake, and at 1 PM insulin, a saturating level (Fig:3A). The Ki for phenylarsine oxide at both insulin concentrations was found to be 7 PM (Fig. 3, A and B). These results show that once phenylarsine oxide has interacted with the system, even a high level of insulin, added subsequently, is not capable of reversing the inhibition (Fig. 3B). It will be demonstrated below, however, that inhibition by phenylarsine oxide can be reversed by vicinal dithiol reagents. The effectiveness of two other trivalent arsenicals, p-aminophenylarsine oxide and arsenite, were compared with that of phenylarsine oxide for inhibition. While all three reagents inhibited insulin-activated deoxyglucose uptake, they differed considerably with respect to their Ki values which were found

4 Vicinal Sulfhydryls in Insulin-activated Sugar Uptake 2649 w 1 I z Y ap 100 IP - " A / b 1pM Insulin 10 nm Insul in 00 I I #ARSINE OXIDE, pm FIG. 3. Effect of phenylarsine oxide on 2-deoxyglucose uptake. A, 3T3-Ll adipocytes were preincubated for 10 min with phenylarsine oxide at the indicated concentrations followed by a 15- min incubation with or without insulin. Uptake of 2-deoxy-[l-"C] glucose (0.2 mm) was measured for 5 min. Cell monolayers were then washed and dissolved in 0.1% SDS and counted for radioactivity. Inset, o"-o, no insulin. Ins, insulin. E, results from A are presented as per cent of inhibition of the insulin-stimulated transport rate. to be 7, 30, and 1,500 p~ for phenylarsine oxide, p-aminophenylarsine oxide, and arsenite, respectively. These results indicate that the more hydrophobic arsenicals are more effec- tive inhibitors. This suggests that the more hydrophobic compartments of the cells (particularly membranes) may concentrate and stabilize the inhibitor at its site of action. Alternatively, the aromatic ring may interact with a hydrophobic site on the inhibited protein. It is of interest that in the absence of insulin, low levels of phenylarsine oxide actually activated basal deoxyglucose uptake (Fig. 3A, inset). While the basis of this activation is not presently understood, this result taken together with those above strongly suggests that phenylarsine oxide interacts with a component closely associated with the insulin-activated transport system. Since maximal inhibition of insulin-activated sugar uptake was achieved at 20 p~ phenylarsine oxide, all subsequent studies were carried out at this inhibitor concentration. The inhibitory effect of phenylarsine oxide on insulinstimulated hexose transport was further substantiated in experiments utilizing the sugar analog methylglucose which is transported but, unlike 2-deoxyglucose, not further metabolized. Fig. 4 shows that insulin activated the 3-O-[rnethyl-"C] glucose uptake rate by 15-fold (106 nl of extracellular fluid/ IO6 cells uersus 7 n1/106 cells for insulin-stimulated and basal cells, respectively, during the initial phase of uptake). The half-time for achieving complete equilibration of 3-0-methylglucose was 12 and 125 s for insulin-stimulated and basal cells, respectively. The distribution of methylglucose between extracellular and intracellular pools allows the calculation of intracellular water volume at equilibrium ( pl/106 FIG. 4. Effect of phenylarsine oxide on 3-0-methylglucose uptake. Cells grown on coverslips were preincubated for 5 min with or without phenylarsine oxide (20 p ~ ) Cells. were exposed to 1 p~ insulin for an additional 10 min. 3-0-[m~thyl-"C]Glucose uptake was determined as described in the text for the indicated time. Uptake was terminated with rapid washes in phloretin-phosphate-buffered saline, and cell-associated radioactivity determined. All data are corrected for intercellular space using cytochalasin B-inhibited transport which represents <3% of the equilibrium influx value. a, insulin; 0, insulin and phenylarsine oxide; A, control. W t 1 1 FIG. 5. Effect of phenylarsine oxide on deoxyglucose uptake by 3T3-Ll adipocytes previously activated by insulin. Cell monolayers were incubated for 15 min with or without 1 p~ insulin. At zero time insulin was withdrawn or maintained where indicated, and 20 p~ phenylarsine oxide was added. At given intervals, 2-deoxy- [l-14c]glucose(0.2 mm) uptake was measured for a 2.5-min period which is included in the times indicated. cells). Insulin does not alter this parameter. Preincubation for 5 min with phenylarsine oxide inhibits insulin-stimulated methylglucose transport comparable to its effect on insulinstimulated deoxyglucose uptake, i.e. the rate of sugar uptake is reduced to the basal level. The action of phenylarsine oxide is essentially immediate since its addition concomitant with insulin caused maximal inhibition of insulin-activated deoxyglucose uptake (results not shown). In contrast, however, when phenylarsine oxide was added after hexose uptake had been activated by insulin, the hexose uptake rate declined slowly toward the basal rate (tli2 = 10 min) despite the continued presence of insulin (Fig. 5). This slow rate of inhibition by phenylarsine oxide in the presence of insulin occurred at the same rate, i.e. exhibited the same tljz, as the rate of return to the basal rate of sugar

5 2650 Vicinal Sulfhydryls in Insulin-activated Uptake Sugar uptake, when insulin was withdrawn from activated cells (Fig. for control and insulin-stimulated cells, respectively. Simi- 5). These results are consistent with the widely accepted model larly, cells activated by insulin and then treated with phenylarsine oxide for 60 min as in Fig. 5 exhibited a reduced V,,, for insulin-activated glucose transport in which translocation for sugar uptake (2.4 nmol/min/106 cells), but no change in of the glucose transporter to the cell surface from an intra- K, for deoxyglucose (1.4 mm). This result is consistent with cellular compartment occurs upon activation (9, 10). Our the view expressed above that phenylarsine oxide blocks the results suggest that phenylarsine oxide blocks the movement insulin-induced translocation of the glucose transporter to of transporters from one compartment to another, specifically the plasma membrane but does not affect the activity of the from the intracellular location to the plasma membrane. How- transporter per se at the cell surface. ever, in the "activated" state when most of the glucose trans- To determine whether phenylarsine oxide alters the binding porters are already at the plasma membrane (as monitored by and/or uptake of insulin by intact cells under the exact transport rate; Fig. 5, zero time), phenylarsine oxide does not conditions where the inhibitor totally blocks insulin-activated block the re-entry of the transporter to its internal "inactive" sugar uptake (Table I), the rate at which '251-insulin becomes site. This is in contrast to the effects of other thiol-reactive cell associated was measured at 37 "C. To distinguish between reagents which block not only activation but also reversion cell-associated intact '251-insulin and '251-insulin degradation to the basal state (15). Once the transporter is at the intraproducts, both trichloroacetic acid-precipitable and trichlocellular site, return to the cell surface is blocked by phenylarroacetic acid-soluble cellular lz5i activity were measured. As sine oxide even in the presence of insulin. These results shown in Fig. 7A, prior treatment of cells with phenylarsine suggest that the glucose transporter undergoes recycling. oxide had little, if any, effect on the rate at which '251-insulin, It is worth noting that the basal rate of sugar uptake is not reduced by even a 60-min exposure to phenylarsine oxide (Fig. 5). At present we cannot provide an explanation for why, when phenylarsine oxide is added to insulin-stimulated cells, the sugar uptake rate does not return completely to the basal unstimulated rate (Fig. 5). The possibility was tested that the inhibitory effect of phenylarsine oxide on deoxyglucose transport results, at least in part, from an effect on the K,,, for the sugar substrate. Previous work with several systems has shown that insulin increases the V,,, for deoxyglucose transport without significantly affecting the apparent K,,, (for review, see Ref. 1). As shown by others with 3T3-Ll adipocytes (5,6) and confirmed here (Fig. 6), insulin increased the Vmax for deoxyglucose uptake from 0.82 to 12.3 nmol/106 cells/min without significantly affecting the K,,, for deoxyglucose, i.e. 1.3 and 1.5 mm i.e. trichloroacetic acid-precipitable lz5i activity, became specifically cell associated or achieved a steady-state level (>5 min after 1251-insulin addition). Although the inhibitor caused n W k 5 V 0>- 15- a AFTER '251-INSULIN ADDITION I I 2-DEOXYGLUCOSE, rnm I.L / In."lm+!'[OEOXYGLUCOSE] FIG. 6. Effect of phenylarsine oxide on K, for deoxyglucose uptake. A, cell monolayers, treated as in Fig. 5 for the 60-min time interval, were assayed for 2-deo~y-[l-'~C]glucose uptake rate at the indicated concentration of hexose. B, a Lineweaver-Burk plot of the data from A ADDED INSULIN, nm FIG. 7. Effect of phenylarsine oxide on the rate of association of lasi-insulin with 3T3-Ll adipocytes at 37 "C and on 1251-insulin binding to detergent-solubilized receptor at 4 "C. A, cell monoiayers were preincubated with (0, m) or without (0,O) 20 WM phenylarsine oxide for 10 min in a buffer containing 0.1% bovine serum albumin before addition of 5 nm '9-insulin. At the indicated times, the buffer was removed and cells were washed rapidly with cold phosphate-buffered saline (3 X 3.0 ml). Cell monolayers were treated with cold 20% trichloroacetic acid and soluble (0, m) and precipitable (0,O) radioactivity determined. TCA, trichloroacetic acid &As0 = phenylarsine oxide. B, 3T3-Ll adipocytes were scraped from cell monolayers, homogenized in hypotonic HEPES (10 mm, ph 7.6), membranes collected by centrifugation (180,000 X g), and washed 5 times with high salt buffer (0.5 M NaC1). Membranes were solubilized in 1% Triton X-100 buffer (HEPES, 50 mm, ph 8.45) and centrifuged at 180,000 X g. Solubilized extract was batch adsorbed to wheat germ agglutinin Sepharose and receptor eluted with 0.3 M N- acetylglucosamine in 0.1% Triton X-100. '251-Insulin binding capacity was assayed overnight at 4 "C in the presence or absence of 20 p~ phenylarsine oxide. bas0, phenylarsine oxide.

6 a small initial lag in the degradation of insulin, the final rate at which cell-associated insulin underwent degradation to trichloroacetic acid-soluble form was unchanged (Fig. 7A). Similarly, when cell monolayers were treated with phenylarsine oxide as above, '251-insulin binding by receptor extracted from total cellular membranes with Triton X-100 was not affected (results not shown). Furthermore, Triton X-100- solubilized receptor from control cell monolayers exhibited no alteration of insulin binding when incubated at 4 "C in the presence of20 pm phenylarsine oxide (Fig. 7B). Taken together these results indicate that phenylarsine oxide treatment of cells, under conditions where insulin-activated deoxyglucose uptake is completely blocked, has little effect on the insulin-binding ability of the receptor. Unexpectedly, however, when intact cell monolayers were incubated with 20 p~ phenylarsine oxide at 4 "C for 14 h the apparent affinity of the cell-surface receptor for insulin was reduced by about 10-fold. It should be noted that these conditions differ considerably from those used to demonstrate inhibition of insulin-activated sugar uptake by phenylarsine oxide (Fig. 3). While the basis for these contrasting results is not known, it appears that prolonged incubation with phenylarsine oxide at 4 "C either disrupts the integrity/accessibility of plasma membrane insulin receptors or stabilizes an inhibitory interaction between the receptor and a coupling component which is the target of phenylarsine oxide. Since the inhibition of insulin binding by long-term exposure to phenylarsine oxide can be largely reversed by subsequent treatment of the cell monolayers with 2,3-dimercaptopropanol (results not shown), the latter interpretation appears more reasonable. Reversibility of the inhibition of hexose uptake by 2,3-dimercaptopropanol (see below and Fig. 8) demonstrates that disruption of cellular integrity does not occur under conditions of brief exposure to phenylarsine oxide at 37 "C which block insulin-activated hexose uptake, but not insulin binding and degradation (Fig. 7A). Trivalent arsenicals, including phenylarsine oxide, form stable ring complexes with vicinal dithiols, such as the lipoyl prosthetic group of pyruvate and a-ketoglutarate dehydrogenase (for review, see Ref. 29); hence, trivalent arsenicals, are potent inhibitors of these enzymes (30). The formation of stable dithiol-trivalent arsenical ring complexes can be reversed by competing with vicinal dithiol compounds, such as 2,3-dimercaptopropanol, but not by monothiol compounds. Thus, reversibility of the inhibition of enzymatic activity by trivalent arsenicals with dimercaptopropanol (but not by monothiol) is diagnostic for the participation of vicinal dithiol groups in a metabolic process (31). To test the possibility that vicinal thiol groups are the site of inhibition of insulin-activated hexose uptake by phenylarsine oxide, attempts were made to reverse the inhibition with 2,3-dimercaptopropanol or 2-mercaptoethanol. When cells were inhibited by exposure to phenylarsine oxide for 10 min and then were incubated with insulin and dimercaptopropanol added together, inhibition was immediately and completely reversed (Fig. 8A). In contrast, 2-mercaptoethanol was ineffective in reversing this inhibition. It should be noted that neither dimercaptopropanol nor mercaptoethanol in the absence of the arsenical affected the rate of insulin-activated deoxyglucose uptake under these conditions (results not shown). Of interest is the fact that addition of dimercaptopropanol, 10 min after insulin, to cells previously treated with the arsenical somewhat decreased the rate and extent of reactivation (Fig. &I). Finally, it was determined that a concentration of 16 JLM dimercaptopropanol was required for either half-maximal pre- Vicinal Sulfhydryls in Insulin-activated Sugar Uptake 2651 FIG. 8. Prevention and reversal of inhibition by phenylarsine oxide of insulin-stimulated deoxyglucose uptake. A, cell monolayers were preincubated with (0, A, V, 0) or without (0) phenylarsine oxide for 10 min. Insulin was added to all cells in the absence (0) or presence (U) of 200 pm 2,3-dimercaptopropanol or 400 p~ 2-mercaptoethanol (A) for the indicated time. To some cells, 200 p~ 2,3-dimercaptopropanol (0) was added 10 min after insulin as indicated by the arrow. 2-Deoxy-[l-''C]glu~ose uptake rates were measured for 2.5 min at each time point. B, prevention of inhibition (X-X). Cell monolayers were preincubated for 10 min with 20 p~ phenylarsine oxide and 2,3-dimercaptopropanol. Insulin (1 p ~ was ) added for an additional 15 min. 2-Deo~y-[l-'~C]glucose uptake rates were measured for a 10-min period. Reversal of inhibition (0-0). The same.time frame wasfollowed. 2,3-Dimercaptopropanol was added with insulin, after the 10-min incubation with 20 p~ phenylarsine oxide. vention or half-maximal reversal of inhibition of deoxyglucose uptake by 20 JLM phenylarsine oxide (Fig. 88). Thus, dimercaptopropanol is extremely effective in reversing inhibition with intact cells acting at a concentration nearly stoichiometric with that of phenylarsine oxide. In contrast, 2-mercaptoethanol at a molar stoichiometry of 20:l relative to phenylar- sine oxide had no detectable effect on the prevention or reversal of inhibition of insulin-activated deoxyglucose uptake (results not shown). The reversal of the inhibition by dimercaptopropanol shows that phenylarsine oxide does not disrupt the integrity of the plasma membrane or irreversibly damage the insulin-activated deoxyglucose transport system under these conditions. 2 (12 DISCUSSION The results presented in this paper show that phenylarsine oxide, a trivalent arsenical which forms stable complexes with vicinal dithiols, inhibits insulin-activated deoxyglucose uptake by 3T3-Ll adipocytes. The characteristics of this inhibition indicate that the inhibitor uncouples the activation process, rather than directly affecting either hexose transport per se or the binding of insulin to its receptor. Several lines of evidence indicate that phenylarsine oxide does not directly inhibit hexose transport. As shown in Table I and in Figs. 1 and 2, the arsenical has no inhibitory effect on the basal rate of deoxyglucose uptake. To validate this conclusion it was necessary to establish that the basal hexose uptake measured represents bonafide facilitated uptake via the glucose transporter. This was demonstrated (Table I and unpublished results) by the fact that basal deoxyglucose uptake (and insulin-activated uptake as well) is inhibited >90% by cytochalasin B, an established inhibitor of facilitated glucose transport in animal cells (24, 25). Furthermore, the kinetics of inhibition revealed that the hexose transport system is subject to inhibition only when the transporter is in the basal state. Thus, whereas phenylarsine oxide inhibits deoxyglucose uptake instantly (tl/z < 0.2 min) when added with or prior to insulin, inhibition is delayed considerably (tip = 10 min) when the inhibitor is added after hexose uptake has been activated by insulin (Fig. 5). This delay of inhibition corresponds to the slow rate loss (tllz = 10 min, Fig. 5) of

7 2652 Vicinal Sulfhydryls in Insulin-activated Sugar Uptake transport activity when insulin is withdrawn from cells previously activated by insulin. Based on the translocation mechanism for insulin-activated glucose transport (9, 10) these results indicate that inhibition by phenylarsine oxide can only occur when the transport system is in the basal state, i.e. when most of the transporters are in the intracellular compartment. The hypothesis that the coupling of the interaction of insulin with its receptor to the hexose transport system is the focus of phenylarsine oxide action is consistent with the observations that the inhibitor had little or no effect on the apparent affinity of the transporter for hexose (Fig. 6) nor did it affect binding of insulin to its receptor (Fig. 7). Despite the fact that the inhibitor does not appear to affect the binding of insulin to its receptor or of hexose to the transporter, the possibility is not ruled out that the arsenical interacts with the receptor or the transporter per se at a site(s) involved (either directly or indirectly) in coupling the receptor to the transport system. Alternatively, the signaling by the receptor to the transport system could involve one or more intermediary component(s) with which phenylarsine oxide reacts. The possibility was considered that interaction of phenylarsine oxide might block insulin-activated autophosphorylation and thereby interrupt signal relay to the glucose transport system. It was found that neither basal nor insulin-activated autophosphorylation of partially purified Triton X-100 solubilized receptor was affected by prior incubation with 20 ~ L M phenylarsine oxide.' Hence, it appears that neither insulin binding nor autophosphorylation by the solubilized receptor is affected by the inhibitor. The mechanism by which trivalent arsenicals inhibit certain enzymatic processes, e.g. catalysis by pyruvate and a-ketoglutarate dehydrogenases, involves the reaction of vicinal sulfhydryl groups on the enzyme with the arsenic atom to form a stable ring complex, the 5- to 6- membered ring structures being the most stable (29). Reversal of this inhibition with a stoichiometric amount of a vicinal dithiol compound, such as 2,3-dimercaptopropanol which forms a stable 5-membered ring with trivalent arsenical, is considered diagnostic for the involvement of vicinal thiol groups in the inhibited process (31). In the case of the a- ketoacid dehydrogenases, the reduced dithiol form of the lipoyl prosthetic group, which is bound covalently to the enzyme, possesses a vicinal thiol group separated by three methylene groups and thus gives rise to a 6-membered ring upon reaction with the arsenical. The similar characteristics of the inhibition (and its reversal) of insulin-activated hexose transport by phenylarsine oxide to those of the a-ketoacid dehydrogenases described above suggest a similar mechanism. It is of interest that the insulin-induced internalization of the insulin receptor is blocked by phenylarsine oxide, while its translocation back to the plasma membrane when insulin is withdrawn is unaffected by the inhibitor (16). In contrast, the situation appears to be reversed with the hexose transporter, that is insulin-induced translocation of the transporter to the cell surface is inhibited while its internalization upon removal of insulin is not affected. The common feature of these two insulin-dependent processes may be internalization of the receptor. Thus, insulin-induced internalization of its receptor may be required to trigger translocation of the trans- porter to the cell surface, perhaps even through a direct receptor-transporter interaction. Alternatively, the signal that R. A. Kohanski, S. C. Frost, and M. D. Lane, unpublished results. triggers receptor internalization may also trigger transport externalization without a direct interaction between receptor and transporter. Whatever the exact mechanism for the activation of the glucose transporter translocation process, it is evident that phenylarsine oxide reacts with a component of the coupling system involved in this sequence of events. Further work will be necessary to identify this component and to define the event. Acknowledgments-We would like to thank Dr. Ron Kohanski for providing partially purified insulin receptor and for assistance in performing the solubilized insulin-binding assay, our colleagues in the laboratory for helpful suggestions during the course of this work, and Norma Mitchell for typing the manuscript. REFERENCES 1. Gliemann, J., and Rees, W. D. (1983) Curr. Top. Membr. Tramp. 18, Karlsson, F. A., Grunfeld, C., Kahn, C. R., and Roth, J. (1979) Endocrinology 104, Rosen, 0. M., Smith, C. J., Fung, C. and Rubin, C. S. (1978) J. Biol. Chem. 253, Reed, B. C., and Lane, M. D. (1980) Adu. Enzyme Regul. 18, Grunfeld, C., Baird, K., Van Obberghen, E., and Kahn, C.R. (1981) Endocrinology 109, Resh, M. D. (1982) J. Biol. Chem. 257, Frost, S. C., Kohanski, R. A., and Lane, M. D. (1983) Fed. Proc. 42, Reed, B. C., Ronnett, G. V., Clements, P. R., and Lane, M. D. (1981) J. Biol. Chem. 256, Cushman, S. W., and Wardzala, L. J. (1980) J. Biol. Chem. 256, Suzuki, K., and Kono, T. (1980) Proc. Natl. Acad. Sci. U. S. A. 77, Cart&, J. R., and Martin, D. B. (1969) Biochim. Biophys. Acta 177, Cadenas, E., Kahi, H., Park, C. R., and Rasmussen, H. (1961) J. Biol. Chem. 236, Fong, C. T. O., Silver, L., Popenoe, E. A., and Debons, A. F. (1962) Biochim. Biophys. Acta 56, Czech, M. P., Lawrence, J. C., Jr., and Lynn, W. S. (1974) Proc. Natl. Acad. Sci. U. S. A. 71, Czech, M. P. (1976) J. BWZ. Chem. 251, Knutson, V. P., Ronnett, G. V., and Lane, M. D. (1983) J. Biol. Chem. 258, Hanson, R. W., and Ballard, F. J. (1968) J. Lipid Res. 9, Chen, R. F. (1967) J. Biol. Chem. 242, Roth, J. (1975) Methods Enzymol. 37, Norton, J. M., and Munck, A. (1980) J. Immunol. 125, Williamson, J. R., and Corkey, B. E. (1969) Methods Enzymol. 13, Kohanski, R.A., and Lane, M. D. (1983) J. Biol.Chem. 258, Bergmeyer, H. U., Bernt, E., Schmidt, F., and Stork, H. (1974) in Methods of Enzymatic Analysis (Bergmeyer, H. U., ed) pp , Academic Press, New York 24. Junp, C. Y.. and RamDal. A. L. (1977).. J. Biol. Chem ; ' 25. Lin. S.. and Saudich. J. A. (1974) J. Biol. Chem May, J. M., and de Haen, C. (1979) J. Biol. Chem. 254, Muchmore, D. B., Little, S. A., and de Haen, C. (1982) Biochemistry 21, Walsh, C. (1979) Enzymatic Reaction Mechanism, pp , W. H. Freeman and Co., San Francisco. 29. Webb, J. L. (1966) Enzyme and Metabolic Inhibitors, Vol. 111, pp , Academic Press, New York 30. Peters, R. A., and Wakelin, R. N. (1946a) Biochem. J. 40, Peters, R. A., Sinclair, H. M., and Thompson, R. H. S. (1946b) Biochem. J. 40, Frost, S. C., and Lane, M. D. (1984) Fed. Proc. 43, 1895