Kinetic parameters of transport of 3-O-methylglucose and glucose in adipocytes.

A method is described which for the first time makes it possible to measure the initial velocity of uptake of the nonmetabolizable sugar analogue 3-O-methylglucase in adipocytes. The permeability of the rat adipocyte plasma membrane in the presence of very low concentrations of methylglucose (<<K,) was about 0.7 X 10m6 cm*s-l both at 22°C and at 37°C. This corresponds to a half-time of uptake of about 18 s. Insulin (1 PM) increased the permeability 3to 6-fold at 22°C and 8to la-fold at 37°C. Insulin at 70 pM caused half of the maximal effect. The permeability due to nonmediated diffusion was approximately 2 X lo-’ cm l s-‘. Thus, the permeability to 3-0-methylglucose is almost entirely accounted for by carrier-facilitated diffusion. The following results were obtained at 22°C. The K, for 3-0-methylglucose equilibrium exchange (inSide and outside sugar concentration being equal) was about 3.5 mu both in the absence and in the presence of insulin. V,, was about 0.13 mmol= s-‘01 intracellular water-’ and increased to about 0.8 mmol l s-l l 1-l in the presence of insulin. The K,,, for net uptake (intracellular sugar concentration initially zero) in insulin-stimulated cells was 2.5 to 5 InM suggesting the absence of any marked asymmetry of the transport system. The initial uptake of 3-0-methylglucose was inhibited by D-glucose with an inhibition constant of about 7 mM.

A method is described which for the first time makes it possible to measure the initial velocity of uptake of the nonmetabolizable sugar analogue 3-O-methylglucase in adipocytes.
The permeability of the rat adipocyte plasma membrane in the presence of very low concentrations of methylglucose (<<K,) was about 0.7 X 10m6 cm*s-l both at 22°C and at 37°C. This corresponds to a half-time of uptake of about 18 s. Insulin (1 PM) increased the permeability 3-to 6-fold at 22°C and 8-to la-fold at 37°C. Insulin at 70 pM caused half of the maximal effect. The permeability due to nonmediated diffusion was approximately 2 X lo-' cm l s-'. Thus, the permeability to 3-0-methylglucose is almost entirely accounted for by carrier-facilitated diffusion. The following results were obtained at 22°C. The K, for 3-0-methylglucose equilibrium exchange (inSide and outside sugar concentration being equal) was about 3.5 mu both in the absence and in the presence of insulin.
V,, was about 0.13 mmol= s-'01 intracellular water- ' and increased to about 0.8 mmol l s-l l 1-l in the presence of insulin. The K,,, for net uptake (intracellular sugar concentration initially zero) in insulin-stimulated cells was 2.5 to 5 InM suggesting the absence of any marked asymmetry of the transport system. The initial uptake of 3-0-methylglucose was inhibited by D-glucose with an inhibition constant of about 7 mM. The hexose transport system in adipocytes has attracted much interest because it is markedly stimulated by insulin. Until now, it has not been possible to measure the transport of D-glucose directly due to the rapid metabolism of this sugar. Transport of D-ghCOSe has been measured in adipocyte membrane vesicles, but this system is not stimulated by the addition of insulin to the vesicles (1,2). However, several techniques are now available for the study of nonmetabolizable sugars or sugar analogues in adipocytes. Using modifications of the previously described oil technique (3) it has been possible to measure the rate of uptake of slowly transported sugars such as D-allOSe (4) or L-arabinose (5). However, these sugars are transported slowly, because their Michaelis constants are high (50 InM or more) And this limits their usefulness in the study of the characteristics of the hexose transport system. We have previously described the kinetics of equilibrium exchange of D-3-O-methylglucose which is transported * This work was supported by the Danish Medical Research Council and Nordic Insulin Foundation. 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 USC. Section 1734 solely to indicate this fact.
$ Part of this investigation was carried out while J. G. was a visiting research professor, Diabetes Endocrinology Center. Present address, Department of Physiology, University of Aarhus, 8000 Aarhus C, Denmark.
very rapidly (6). The principle of the method is to follow the efflux of labeled methylglucose from preloaded cells, and the technique is not suited for uptake studies. The half-time of efflux was too short to be measured directly when the permeability of the adipocyte membrane to labeled methylglucose was maximal (i.e. 37"C, total methylglucose concentration <<Km, insulin present), but we were able to extrapolate a value of about 3 s. Czech (7,8) has described a filter technique for measurement of the initial velocity of uptake of 3-O-methylglucose in adipocytes and reported a linear uptake for 20 to 30 s even when the permeability was maximal. The apparent transport rates reported by Czech were at least 20 times lower than those reported by us under comparable conditions (6,7). This suggested to us that initial velocities were underestimated when transport rates were measured with the technique described by Czech. In the present communication, we describe a new technique and we present results which further characterize the hexose transport system in adipocytes. Some preliminary results have been presented previously (9).

MATERIALS AND METHODS
Adipocytes were prepared from epididymal and perirenal fat from ad libitum fed male Wistar rats weighing 120 to 150 g and cell diameters were measured as previously described (10). The mean diameter was about 60 pm. The cells were suspended in buffer containing (in mM) Na+:140, K', 4.7; Ca'+, 2.5; Mg",  with the cell pellet was added to a counting vial with 5 ml of scintillation fluid and radioactivity was measured. The method is outlined in Fig. 1 of the miniprint section.'

Evaluation
of Method-It is essential that the stopping solution work efficiently and almost immediately, even in cells with maximal permeability to methylglucose. Table I shows that the intracellular methylglucose space calculated on nondiluted samples as total methylglucose space minus L-glucose space was not significantly different from the intracellular distribution spaces calculated after dilution of the extracellular medium with stopping solution. Fig. 2 shows that the halftime of efflux of methylglucose in insulin-stimulated cells was about 60 min in the presence of 0.3 mu phloretin.
The combined results show that the time required for 0.3 mM phloretin to act is insignificant as compared to the half-time of efflux of labeled methylglucose in buffer, and that the loss of methylglucose from the intracellular pool of phloretin-treated cells is about l%/min.
It should be noted that efflux of 20 mM methylglucose in the presence of phloretin is even slower than that of "tracer" methylglucose alone. A rough estimate of the inhibition constant (KJ of phloretin can be calculated from the curves assuming competitive inhibition. The half-time of efflux of tracer 3-0-methylglucose is about 3 s in the absence of phloretin (see below) and about 700 s in the presence of 30 pM phloretin (Fig. 2). Thus, K, of phloretin is approximately 30 X 3/700 pM or 0.13 pM. It is concluded that the stopping procedure is adequate in the range of methylglucose concentrations used in the present experiments.
The coefficient of variation (n = 20) for methylglucose uptake at equilibrium (as well as for the "uptake" at time zero) was 7.1%. The coefficient of variation for uptake of methylglucose at 2 s in insulin-stimulated cells (intracellular methylglucose space about 30% of that at equilibrium) was 10.5% (n = 20). It appears, therefore, that the timing was sufficiently precise.
Some other aspects of the method are presented in the miniprint section' and the major points are summarized below: 1. The cell preparation is stable, i.e. the same initial uptake was measured in freshly prepared cells and in cells which had been standing for 2 h as 40% suspension at 22°C. It is preferable to use 14C-labeled methylglucose with the label in the methyl group (Fig. 3).
2. Ice cold buffer is not sufficient to stop efflux of labeled methylglucose and the stopping solution becomes increasingly less efficient with increasing albumin concentrations (Fig. 4).

The association of methylglucose
with the cells appears to represent a distribution of methylglucose in aqueous In some cases, the cells were loaded with labeled methylglucose plus 20 mM unlabeled methylglucose for 60 min before efflux was carried out in buffer containing phloretin The points represent means of duplicate values.
phases. The contribution of a binding component to the total distribution space seems small (Table II). 4. Bulk mixing of the isotope with the aqueous phase of the cell suspension appears to be adequate (Table III).
5. The measured kinetic constants apply to the transport system of the plasma membrane and are not influenced to a significant degree by unstirred layers in the extracellular aqueous medium (Fig. 5).
Time of the effect of insulin at this high concentration at 22"C, whereas only about 60 s was required at 37°C. The half-time of uptake of tracer methylglucose was about 3 s in the presence of insulin and this agrees quite well with the values extrapolated from our previous experiments to "infinitely low" methylglucose concentrations (6). An estimate of the permeability of the plasma membrane to the labeled methylglucose can be derived from Fig. 6. The incubation contained 16 ~1 of packed cells with an intracellular distribution volume of 0.3 ~1. It is seen from the curve that in insulin-treated cells about 23% of the equilibrium space is fried in 1 s. In other words, the amount of labeled methylglucose entering the cells per s corresponded to the amount present in 70 nl (7 x 10m5 cm3) of the extracellular medium. The mean surface area per cell was 1.1 x lo4 pm2 (1.1 x low4 cm") and the total surface area was therefore about 18 cm'. Thus, the permeability was about 3.9 X 10e6 cm-s-'.
A moderate error is introduced in this calculation since the adipocyte diameters are in fact normally distributed with a coefficient of variation of 10% (10). In separate experiments we, therefore, measured the diameter distribution of 200 cells and calculated the total surface area of the 16 ~1 of packed cells. The maximal permeability to ['4C]methylglucose in the presence of insulin at 22°C varied from 2.1 x lo-" to 4.0 x lo-" cm. s-l. Fig. 8 shows that an increase in temperature from 22°C to 37°C nearly doubled the maximal permeability of insulinstimulated cells whereas the uptake declined slightly in the absence of insulin (cfi Fig. 7). Fig. 9 shows that the steady state insulin dose-response relationship on methylglucose transport at 37°C is closely similar to that obtained on the rate of conversion of glucose to lipids (11). In this experiment, albumin (50 mg/ml) and bacitracin (0.5 mg/ml) were added to the medium in order to reduce protease-mediated degradation of insulin (12). The transport rates (basal or maximally insulin-stimulated) obtained in this medium were not different from those obtained with albumin 10 mg/ml and in the absence of bacitracin (data not shown).
Permeability Due to Nonmediated Diffusion-Before studying the concentration dependence of methylglucose transport, it was important to investigate whether a significant fraction of the uptake of labeled methylglucose was due to nonmediated diffusion. Fig. 10 shows that the rate of uptake of L-['4C]glucose in the absence of insulin was less than 1% of that of ['4C]methylglucose.
Insulin enhanced the uptake suggesting that carrier-mediated transport is involved. Furthermore, 40 mM methylglucose caused a marked inhibition and the half-time of L-glucose uptake was, under these conditions, at least 150 min both in the absence and in the presence of insulin. This means that the component of the permeability to L-glucose, which may be due to nonmediated diffusion, is not more than about 0.5% of the permeability to methylglucase in "basal" cells and about 0.1% of the permeability in insulin-stimulated cells. These figures correspond well to the permeability to methylglucose obtained in the presence of 0.3 IIIM phloretin, cf: Fig. 2. However, methylglucose is more lipophilic than glucose and it may be argued that phloretin in high concentrations changes the general properties of the plasma membrane.
Thus, the minimal permeability for Lglucose or for methylglucose in the presence of phloretin may not necessarily be relevant measures for the nonmediated diffusion permeability under normal incubation conditions. We, therefore, measured uptake of 300 mu methylglucose (about 80 times K,,,, see below) under conditions approaching equilibrium exchange. The results shown in Fig. 11 indicate that the nonmediated diffusion permeability is not more than about 0.5% of the total permeability of insulin-stimulated cells to methylglucose present in a concentration <<K,,, and not more than 2 to 3% for basal cells. These figures must be considered as maximum estimates. The combined data indicated that carrier-mediated diffusion, at least in the presence of insulin, accounts almost entirely for the total transport in the presence of total methylglucose concentrations up to about 5 times K,.
Countertransport-One of the characteristics of carrier-mediated diffusion is that addition of unlabeled sugar can induce a transient transport of the labeled sugar against its concentration gradient. Fig. 12 demonstrates K, was determined graphically as 3.5 mM. Since K,,JV,,,,, for basal cells was 23 1-l X s X liter of intracellular water, V max was calculated as 0.15 mmol x s-' x liter of intracellular water-'. For insulin-stimulated cells V,,, was calculated as 0.78 mmol x SK' x liter-'. The initial velocity is difficult to measure under net uptake conditions (intracellular sugar concentration is initially zero) because the time course of uptake deviates progressively from Uptake of L-glucose then removed to make a 30% packed cell volume and the uptake of 50 pM ['4C]methylglucose was measured after incubation for 2 s. '""Ilabeled insulin was added to parallel incubations and 11% of the radioactivity was soluble in 12% trichloroacetic acid by the time transport was measured.
where S,, is the concentration of sugar in the extracellular aqueous phase and Si the concentration in the intracellular aqueous phase in that small time interval. This flux equation has been integrated (13,14) and in this form the equation expresses the intracellularly accumulated amount of sugar as a function of time. Fig. 14 shows an experiment in which equilibrium exchange of methylglucose in insulin-stimulated cells was measured both with a very low methylglucose concentration (50 pM) and with 20 mM. From these two curves, K,,, (equilibrium exchange) was calculated as 3.7 mM and the curve for net uptake of 20 mM methylglucose was calculated assuming a symmetrical transport system. It is seen that the calculated curve for net uptake agrees well with the experimental points. identical with K,,, for net uptake of methylglucose provided that the initial velocity of uptake is measured. In practice, the nonexponential nature of the net uptake curve (cf Fig. 14) will cause a small error in the estimate of the initial velocity in the presence of high concentrations of sugar. Fig. 15 shows that K, for net uptake of methylglucose is of the same magnitude as K, for equilibrium exchange. K, ranged in four experiments from 2.5 InM to 5 InM. This result, combined with the data shown in Fig. 13, suggests that the transport system in insulin-stimulated cells is symmetric. half of that of methylglucose. The prefix apparent is used because it cannot be excluded that a small amount of labeled glucose is converted to CO2 and escapes. However, this seems rather unlikely since the uptake of glucose, by the time samples were taken (3 s), corresponded to less than 30% equilibration with the intracellular water space and since chilling prevented metabolism after uptake was stopped (cf text to Fig. 16). In any case, the inhibition constant of glucose was also about twice that of methylglucose, measured on the initial velocity of uptake of [ '?]glucose. DISCUSSION The Maximal Permeability to Methylglucose-The halftime for equilibration of a nonmetabolizable sugar or sugar analogue in a particular cell system must be the same for equilibrium exchange and net influx when the substrate concentration is much lower than K, for influx or exchange. The finding of similar half-times with the efflux method (6) and the present technique, therefore, supports the validity of the methods. Czech (7) reported average intracellular distribution spaces for methylglucose of 0.5 to 1.5 pl in cells from small rats, i.e. with mean diameters of about 60 pm. This is slightly smaller than the values reported by us and should, if anything, tend to decrease the half-time for equilibration. However, Czech (7) reported a linear uptake of tracer ["Hlmethylglucose for at least 30 s at 37°C and in the presence of insulin, i.e. under conditions where we find a half-time of uptake of about 2 s (Fig. 9). These data indicate that the maximal permeabilities reported by us are at least 20-fold higher than those reported by Czech (7). It should be noted that Chandramouli et al. (2) found that 100 pM methylglucose was 80% equilibrated in insulin-stimulated cells after incubation for 5 s at 24°C followed by centrifugation of the cells through oil for 15 s. This supports our results even though the incubation time was not well defined because no stopping solution was applied (2).
The transport rates in terms of pmol x s-' are difficult to compare directly since different authors have used different concentrations of methylglucose. However, it is reasonable to assume that the initial rate of uptake is proportional to the sugar concentration in a range up to 1 mM. Table IV shows the calculated initial rates of uptake of 100 pM methylglucose in insulin-stimulated cells. It appears that the rates of uptake reported in the present paper are much higher than previously reported rates. There is about a loo-fold difference between our results and those reported by Czech (7,15). It seems unlikely that this difference can be explained by differences in rats or collagenase preparations which, according to the authors' experience, may cause variations in the maximal permeability of not more than a factor of 3. Table IV also shows the rate of conversion of glucose to metabolic products. This figure must necessarily be smaller than the initial rate of glucose uptake (when measured on the same cells) which, again, is smaller than the initial rate of methylglucose uptake (Figs. 15 and 16). It appears that the rate of conversion of glucose to lipids plus COZ as reported by Czech (1.5) is 2 to 20 times higher than the initial rate of uptake of methylglucose. The data of Ref. 16 show that the apparent initial velocity of methylglucose uptake measured by the oil flotation method (3) after incubation for 20 s at 21°C is much lower than the rate of glucose conversion to products at 37°C. Czech (15) has reported that cells which responded at least 20-fold to insulin with respect to conversion of 0.2 mM [l-'4C]glucose to CO2 (or lipids, CL Fig. 2 of Ref. 15) responded with a 2-to 3-fold increase of the initial velocity of uptake of tracer methylglucase at 37°C (7,8). In our hands, the effect of insulin (fold increase) on methylglucose transport is considerably higher (Ref. 6,Figs. 7 and 8). Taken together, these data suggest that the initial velocity of tracer methylglucose in insulin-stimulated cells is markedly underestimated when measured as   Table I 22 (lipids) described by Czech (7,8). In our view, the main problems are as follows. 1) Addition of 10 ~1 of isotope solution to 100 ~1 of cell suspension does not give instantaneous mixing of the aqueous phases (Table III). 2) It is too late to take the first time point after 15 to 30 s. 3) Ice cold buffer does not stop efflux adequately (Fig. 4). 4) The use of ["Hlmethylglucose may have been a problem in some cases (Fig. 3). On the other hand, it should also be emphasized that it is clearly possible to demonstrate effects of insulin and other factors which increase the permeability of the adipocyte membrane to hexose with the technique described by Czech (8), and that it has provided the basis for important observations in relation to insulin's mechanism of action.
The Kinetic Constants-The two major requirements for evaluation of kinetic constants are the ability to measure the initial velocity in the relevant concentration range and the absence of a large component of nonmediated diffusion. These requirements seem to be fulfiied in the present experiments. Czech reported (Fig. 3 of Ref. 7) that nearly half of the initial uptake of 20 mM methylglucose was not inhibited by 40 pM cytochalasin B which, as shown by Vinten (17), is a competitive inhibitor with an inhibition constant of 250 nM. The present results (Fig. 2) show that the half-time of efflux of 20 mM methylglucose in the presence of 0.3 mu phloretin is at least 60 to 120 min whereas the half-time for 20 mM methylglucose efflux (equilibrium exchange) in nonstimulated cells is not more than 4 min (6). We would therefore estimate nonmediated diffusion to be maximally 7% under these conditions. Czech (7) found that a component of the transport system has a high affinity to methylglucose with a K, of about 1.5 mM. Czech (7) and Olefsky (18), who used 10-s incubations followed by centrifugation through oil without stopping solution, have also reported that the initial rate of uptake was proportional to the methylglucose concentration in the range 5 to 20 mM after subtraction of the values obtained in the presence of cytochalasin B. This high K, component was large and accounted for about half of the total mediated transport of 10 mM methylglucose.
It should be noted that the cited studies were carried out as net uptake experiments (intracellular methylglucose concentration initially zero). The true initial velocity is difficult to measure due to the nonexponential nature of the uptake curves (Fig. 14), and the bias will be different at different concentrations of methylglucose. This might be the reason for the apparent heterogeneity of the transport system (7,18). It may be concluded from our studies that, if a heterogeneity exists, the present method is not precise enough to reveal it. This argument also applies to a possible asymmetry of the transport system (Fig. 14).
The finding that the inhibition constant of glucose (about 7 mM at 22'C) on the initial uptake of tracer methylglucose is about twice as high as that of methylglucose (about 3.5 mM) is in agreement with the results of Loten et al. (4) who determined the inhibition constant of glucose on the initial uptake of allose in adipocytes as about 13 mM at 37°C and that of methylglucose as about 4 mM. The differences in the values may be due to differences in temperature.
There is an apparent discrepancy between these results and Olefsky's (18) finding that the inhibition constant of glucose on tracer deoxyglucose uptake is 1.0 to 2.3 mu and that of methylglucose is 4.5 to 9.0 mM. However, deoxyglucose is phosphorylated by hexokinase and trapped in the cell as deoxyglucose phosphate. Olefsky (18) incubated the cells for 3 min and by that time essentially all intracellular labeled deoxyglucose was in the phosphorylated form. n-Glucose is also phosphorylated and the apparent K, for glucose metabolism, particularly in the presence of insulin, is much lower than the apparent K, for glucose transport reported in the present paper (4,6). Glucose transport cannot be considered rate-limiting for glucose metabolism under all conditions and it might, therefore, inhibit the phosphorylation of deoxyglucose at lower concentrations than those required to inhibit deoxyglucose transport. We have actually found that the inhibition constant of glucose on deoxyglucose transport (incubation 3 s) is considerably higher than the inhibition constant on deoxyglucose phosphorylation (incubation 3 min).' The data suggest that the correct K, for transport of n-glucose at physiological temperature is of the order of 10 mM.
Is the Insulin-stimulated Transport System Different from the Nonstimulated?-The finding that insulin increases V,,, without changing K,,, significantly can be interpreted in two ways. Insulin might either increase the number of carriers or it might increase the turnover on already functioning carriers (or both mechanisms might be involved). Vinten (17) found that V maX of stimulated and nonstimulated methylglucose transport showed the same temperature dependence (18 to 37"(Z), and that insulin did not affect K, significantly in this range of temperatures. This is compatible with the hypothesis that insulin increases the number of functioning carriers. This mechanism was also suggested by Olefsky (18) on the basis of studies on the temperature dependence of 2-deoxyglucose uptake. However, the present studies show that the uptake of tracer methylglucose (i.e. concentration <<K,,J in stimulated cells exhibits a temperature dependence different from that of nonstimulated cells (22 to 37°C Fig. 8). This phenomenon was also noted by Czech (7). The data suggest that the effect of insulin is not exclusively brought about by an increase in the number of carriers available to the hexose. In further support of this hypothesis, we have noted that the stimulated and nonstimulated transport system show different dependencies on changes in PH.' These phenomena are at present under further investigation.