- Email: [email protected]
Heterogeneity of Glucose Transport in Rat Liver Microsomal Vesicles
ARCHIVES OF BIOCHEMISTRY AND BIOPHYSICS
Vol. 359, No. 1, November 1, pp. 133–138, 1998 Article No. BB980888
Heterogeneity of Glucose Transport in Rat Liver Microsomal Vesicles Ga´bor Ba´nhegyi,* Paola Marcolongo,* Ann Burchell,† and Angelo Benedetti*,1 *Istituto di Patologia Generale, Universita` di Siena, 53100 Siena, Italy; and †Department of Obstetrics and Gynaecology, Ninewells Hospital and Medical School, University of Dundee, Dundee, DD1 9SY, Scotland
Received May 15, 1998, and in revised form August 5, 1998
Glucose transport across the membrane of rat liver microsomal vesicles was studied by a rapid filtration method in three different experimental systems: (i) inward transport in the presence of extravesicular glucose, (ii) efflux from passively preloaded vesicles, and (iii) efflux of glucose generated intravesicularly by glucose-6-phosphatase upon addition of glucose 6-phosphate were investigated. The apparent intravesicular glucose space estimated with the rapid filtration method was lower than the total microsomal glucose accessible space both the in the steady-state phase of uptake and at the starting point of efflux: 0.5 versus 2.3 ml/mg protein. The initial rate of influx/efflux was dependent on the extravesicular/intravesicular glucose concentration and was much lower than the rate of influx estimated previously by the light-scattering technique. Both influx and efflux could be inhibited by N-ethylmaleimide and possibly became saturable at high (>100 mM) glucose concentration. Known inhibitors of GLUT transporters (genistein, cytochalasin B, phloretin, and hexoses) did not affect glucose influx. The time course of glucose efflux from vesicles preincubated in the presence of glucose 6-phosphate was similar to that from glucoseloaded vesicles. These data together with that obtained previously (by a light-scattering technique; Marcolongo, P., Fulceri, R., Giunti, R., Burchell, A., and Benedetti, A. (1996) Biochem. Biophys. Res. Commun. 219, 916–922) indicate that microsomal vesicles are heterogeneous regarding their glucose-transporting properties and that glucose transport is bidirectional and its feature meets the requirements of a facilitative transport. © 1998 Academic Press Key Words: glucose; transport; liver microsomes; glucose-6-phosphatase.
1 To whom correspondence and reprint requests should be addressed at Istituto di Patologia Generale, Universita` di Siena, Viale Aldo Moro, 53100 Siena, Italy. Fax: (39) 577 227009. E-mail: [email protected].
0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
The transport of glucose across the membranes of mammalian cells is of vital significance, since glucose has a central role in the homeostasis and intermediary metabolism of the organism. The plasma membrane of all mammalian cells possess member(s) of a family of facilitative glucose transporters (1). The transport of glucose through the endoplasmic reticulum (ER)2 membrane is less clear, despite the fact that this step is supposed to be necessary for the export of glucose produced by the liver. Both oligosaccharide processing of the newly synthesized glycoproteins and the hydrolysis of glucose 6-phosphate result in the release of glucose inside the ER cisternae. In fact, glucose-6-phosphatase, the enzyme catalyzing the last reaction of glucose-producing pathways glycogenolysis and gluconeogenesis, is an integral protein of (hepatic) ER (2) whose active site, according to the recent models of its structure, is situated inside the lumen of the ER (3, 4). The sequence of the enzyme has been reported (5, 6), and functional (7) and genetic (8) evidence for a ER transporter which mediates the translocation of the substrate glucose 6-phosphate from the cytosol to ER lumen has been forwarded. Phenomenological evidence for ER glucose transport system(s) is, however, still not univocal, and the molecular structure of the involved protein(s) is still elusive (9). In the “combined conformational flexibility–substrate transport model” (10 – 13) glucose-6-phosphatase traverses the microsomal membrane forming a water-filled space around the catalytic site in the membrane. Glucose formed in the membrane can enter the lumen of microsomal vesicles, where it accumulates, or can leave toward the cytosol (13). In a recent work (7) we found instead that glucose does derive from the hydrolysis of the intraluminal glucose 6-phosphate. This glucose can, at least in part, accumulate in the microsomal lumen until the hydro2
Abbreviation used: ER, endoplasmic reticulum. 133
134
´ NHEGYI ET AL. BA
lysis of the parent glucose 6-phosphate proceeds, which suggests that the sugar leaves a discrete microsomal pool with a kinetic possibly slower than that of glucose6-phosphate hydrolysis (7). Curiously, it has been also found that glucose is unable to enter the microsomal vesicles from outside (14). An intravesicular glucose space, however, could be detected after prolonged incubation of microsomes in the presence of glucose (15, 16), and it corresponded to 73% of the intravesicular water space (15). Moreover, glucose in high concentration rapidly enters (liver) microsomal vesicles, as it has been revealed by a light-scattering technique in other (17) and our laboratory (15, 18). This “fast” glucose influx cannot be detected by rapid filtration (with radiolabelled glucose) because it was too rapid (t1/2 ranging between 4 and 7.6 s) for this technique as reported/ discussed previously (15). The aim of the present study was to further investigate the influx and efflux of glucose in rat liver microsomal vesicles and the connection of glucose transport with glucose 6-phosphatase activity. Here we describe a glucose influx and efflux pathway, possibly a facilitative transport system for the sugar, which possesses a relatively slow kinetics and can therefore be studied by a conventional filtration technique. This system appears to be present in vesicles other than those provided with the aforementioned rapid transport pathway and is also operative for exporting outside vesicles glucose produced by glucose 6-phosphatase. These data indicate that microsomal vesicles are heterogeneous regarding their glucose transport pathways which suggests the existence of different functional glucose ER pools in the liver. EXPERIMENTAL PROCEDURES Preparation of liver microsomes. Twenty-four-hour-fasted male Sprague–Dawley rats (180 –230 g) were used. Liver microsomes were prepared as reported (19). Microsomal fractions were resuspended in buffer A containing (mM): KCl, 100; NaCl, 20; MgCl2, 1; Mops, 20, pH 7.2. The suspensions were rapidly frozen and maintained under liquid N2 until required. Intactness of microsomal membrane was greater than 90% in preparations employed as ascertained by measuring the latency of mannose-6-phosphatase activity (20). The extent of contamination of the microsomal fraction with plasma membrane was evaluated by measuring 59-nucleotidase activity (21). This activity (units/mg protein, mean 6 SEM, n 5 3) was 67 6 5 in liver homogenates and 90 6 8 in microsomal fractions, and its recovery in the microsomal fractions (as percentage of that of liver homogenate) was 5.9%. The extent of plasma membrane contamination was therefore similar to those previously measured in (purified) liver microsomal fractions (21). To exclude some contribution of plasma membrane-derived vesicles to glucose transport, we preliminary verified that the glucose influx in liver microsomes (see below) was unchanged (e.g., Table I) by digitonin concentration suitable to selectively permeabilise plasma membrane (22). Microsomal protein concentration was determined by the Biuret reaction using BSA as standard. Measurement of glucose influx and efflux. Liver microsomes (1 mg/ml) were incubated in buffer A containing 0.1–100 mM glucose plus [3H]glucose (9 mCi/ml) at 22°C. At the indicated time intervals, samples (0.1 mg protein) were taken to measure 3H associated with
microsomes. Samples were filtered rapidly through cellulose acetate/ nitrate filter membranes (pore size 0.22 mm; which retained 85 6 2% of the microsomal protein applied) and were washed with 4 ml of Hepes buffer (20 mM, pH 7.2) containing 250 mM sucrose. The time required to execute filtration and washing was 15–20 s. In a series of incubation the pore-forming antibiotic alamethicin (0.05 mg/ml) (23, 24) was added to distinguish the intravesicular and the bound radioactivity. More than 95% of the microsomal protein was retained by filters, indicating that the alamethicin treatment did not affect the vesicular structure of microsomes. The alamethicin-permeabilized microsomes retained amounts of radioactivity #20% of that associated to untreated microsomes. Intravesicular radioactive glucose was bona fide lost during the washing procedure, since the alamethicin nonreleasable portion did not further decrease even after extensive wash of filters (and microsomes). The alamethicin releasable portion of radioactivity was therefore regarded as intravesicular (7). To measure glucose efflux, microsomes (10 mg protein/ml)—preincubated with various concentrations of glucose plus [3H]glucose for 2 h at 22°C—were diluted (10-fold) with the incubation medium including no glucose and aliquots of the diluted incubates (0.1 mg protein) were drawn. In parallel samples, 50 mg/ml of alamethicin was also included in the dilution. For the measurement of glucose efflux derived from glucose 6-phosphate, microsomes (10 mg/ml) were incubated in buffer A in the presence of 0.2 mM glucose 6-phosphate plus D-[14C(U)]glucose 6-phosphate (2–3 mCi/ml) at 22°C. After 5 min incubation microsomes were diluted 10-fold with buffer A and samples were taken as described above. Parallel filters were treated after wash with ZnSO4–Ba(OH)2 to precipitate glucose 6-phosphate and measure D-[14C(U)]glucose derived from the parent D-[14C(U)]glucose 6-phosphate according to Ref. (7). Glucose-6-phosphatase activity was measured on the basis of glucose formation as reported earlier (7). Materials. Alamethicin, pentamidine, cytochalasin B, genistein, digitonin, phloretin, N-ethlymaleimide, glucose kit (Trinder method), DIDS, and D-[1-3H(N)]glucose (0.9 mCi/ml; 15.5 mCi/mmol) were from Sigma. D-[14C(U)]Glucose 6-phosphate (0.1 mCi/ml; 300 mCi/ mmol) was from American Radiolabeled Chemicals Inc. (St. Louis, MO). Cellulose acetate/nitrate filter membranes (pore size 0.22 mm) were from Millipore. All other chemicals were of analytical grade.
RESULTS
Microsomal Uptake of Glucose Rat liver microsomal vesicles incubated in the presence of various concentrations of glucose rapidly took up the radioactive tracer until a steady-state level was reached over a 10-min period of incubation (Fig. 1). The uptake was attributable to intravesicular accumulation, since (i) the majority of radioactivity (.80%) could be released by permeabilizing the vesicles with the pore-forming antibiotics alamehicin and (ii) the shrinkage of vesicles upon the addition of the poorly permeant compound sucrose (250 mM) to the medium reduced the uptake by 70% (Fig. 1). The steady-state level of intravesicular glucose was stable for at least 2 h and its extent was dependent on the extravesicular concentration of glucose (data not shown). In the steady-state phase of glucose uptake the apparent intravesicular glucose space— calculated by dividing the intravesicular glucose content (nmol/mg protein) with the extravesicular glucose concentration (nmol/ml)—was 0.43 6 0.09 (ml/mg protein) at 1 mM
135
GLUCOSE TRANSPORT IN RAT LIVER MICROSOMES
FIG. 1. Glucose uptake by rat liver microsomal vesicles. Liver microsomes were incubated in the presence of 1 mM glucose plus [3H]glucose as tracer and intravesicular glucose contents were measured as detailed under Experimental Procedures. The effect of the addition of 0.1 mM N-ethylmaleimide and 250 mM sucrose on the uptake is also shown. Microsomes were treated with N-ethylmaleimide for 30 min, while sucrose was added immediately prior to start glucose uptake. Sucrose did not affect intactness of microsomes as ascertained by measuring the latency of mannose-6-phosphatase activity (20). Data are means 6 SD of six experiments (control, E) or means of two experiments (sucrose, ❐; and N-ethylmaleimide, ‚).
glucose concentration. The apparent space was independent of the extravesicular glucose concentration; other concentrations than 1 mM gave similar results (0.41– 0.55 ml/mg protein). This space is about 12% of the total intravesicular volume of the (liver) microsomes estimated as 3H2O space (3.6 ml, Ref. 15). Initial rate of glucose influx was estimated by measuring 1-min uptakes. The increase of the concentration of extravesicular glucose from 0.1 to 100 mM resulted in an almost linearly increasing influx rate with a tendency of saturation (Fig. 2). Inconsistent data were obtained at glucose concentrations higher than 100 mM, probably due to the shrinkage of the vesicles caused by hyperosmolarity (not shown). The glucose influx was not affected by digitonin concentration suitable to selectively permeabilize plasma membranederived vesicles (Table I), which makes unlikely their contribution to the glucose uptake by the microsomal fraction used. The effect of potential inhibitors of known glucose transporters was also tested (Table I). Phloretin, cytochalasin B, genistein, pentamidine, fructose, galactose, and mannose were ineffective. Phloretin and genistein resulted instead in an apparent stimulation of glucose uptake (Table I). This could be due to the inhibition of the fast glucose transport (14) which became therefore measurable by the rapid filtration assay. Consistently, we observed by the lightscattering assay that phloretin and genistein increased the t1/2 of (50 mM) glucose influx from 6.5 to 46 and 43 s, respectively (Ba´ nhegyi, Marcolongo, and
FIG. 2. The rate of glucose uptake by and glucose efflux from rat liver microsomal vesicles. Intravesicular glucose content of microsomes—incubated in the presence of or preloaded with 0.1–100 mM glucose plus [3H]glucose as tracer—was measured at “0” min (efflux) and at 1-min incubation time. Glucose did not affect intactness of microsomes as ascertained by measuring the latency of mannose-6phosphatase activity (20). Uptake and efflux rates were calculated as a difference. For the experimental protocol see the Experimental Procedures. Data are means of 2– 8 experiments; SD bars were omitted for clarity.
Benedetti, unpublished observations). Both cytochalasin B and pentamidine have been also previously reported by us (15) to inhibit the microsomal fast glucose
TABLE I
Effect of Inhibitors and Hexoses on the Microsomal Uptake of Glucose
Compound (mM)
Glucose uptake (nmol/min/mg protein)
None Digitonina N-Ethylmaleimide (0.1) N-Ethylmaleimide (0.5) N-Ethylmaleimide (5) Phloretin (0.25) Cytochalasin B (0.25) Genistein (0.1) Pentamidine (0.1) Fructose (50) Galactose (50) Mannose (50) Glucose (50)
0.24 6 0.02 0.23 6 0.02 0.16 6 0.03** 0.12 6 0.03** 0.08 6 0.02** 0.32 6 0.05* 0.24 6 0.03 0.33 6 0.08* 0.22 6 0.04 0.23 6 0.04 0.22 0.22 6 0.03 0.23
Note. Liver microsomes were incubated in the presence of 1 mM glucose plus [3H]glucose as tracer for 1 min and intravesicular glucose content was measured as detailed under Experimental Procedures. Microsomes were treated with the different compounds for 30 min prior to start glucose uptake. Data are means of two or means 6 SD of three to five experiments. Values marked * and ** are significantly different from the control value: P , 0.05 and P , 0.01, respectively. a 20 mg/mg of microsomal protein.
136
´ NHEGYI ET AL. BA
vesicles loaded in the presence of 1 mM glucose by 63 6 13% (mean 6 SD, n 5 4). Efflux of Glucose Deriving from Glucose 6-Phosphate Hydrolysis
FIG. 3. Glucose efflux from glucose preloaded rat liver microsomal vesicles. Microsomes were loaded with 10 mM glucose plus [3H]glucose as tracer for 2 h and glucose efflux was provoked by a 10-fold dilution in a glucose-free buffer. Intravesicular glucose contents were measured as detailed under Experimental Procedures. Data are means 6 SD of six experiments.
transport, but the t1/2 of glucose transport was still lower (about 12 s) than the time required for accomplishing the rapid filtration assay (about 15 s). This presumably resulted in the lack of measurability of the fast glucose transport/uptake (as discussed in detail in Ref. 15). N-Ethylmaleimide was inhibitory in a concentration-dependent way, although a complete inhibition was not attained (Table I). The inhibitory effect of 0.1 mM N-ethylmaleimide on the time course of glucose influx is shown in Fig. 1. We have previously observed (15) that 0.1 mM N-ethylmaleimide did not inhibit the fast glucose transport.
It was previously observed by using a rapid filtration technique that a portion of glucose derived from the hydrolysis of glucose 6-phosphate by glucose-6-phosphatase enzyme is still retained into microsomal vesicles until the hydrolysis of the sugar phosphoester proceeds (7, 13). The remaining portion of intralumenally formed glucose should more rapidly leave the vesicles since it was not recovered in the microsomal fraction by the ultrafiltration technique and appeared soon in the extravesicular fluid (7). Further experiments, therefore, were drawn to characterize the efflux of this glucose once formed from glucose-6-phosphate hydrolysis. Rat liver microsomal vesicles (10 mg/ml) were incubated in the presence of 0.2 mM glucose 6-phosphate for 5 min. Due to the high hydrolytic activity (7.4 6 1.8 nmol/min/mg protein) at that time more than 90% of added glucose 6-phosphate had been converted to glucose. At this time microsomes were diluted 10-fold to decrease the rate of further hydrolysis (its rate was less than 0.04 nmol/min/mg protein during the efflux measurement) and the efflux of the intravesicularly accumulated glucose was measured. The initial intravesicular glucose contents in these experiments were similar to those in which microsomes were loaded by incubating them in the presence of 1 mM glucose; moreover, the shape of the two efflux curves was also similar (Fig. 4).
Glucose Efflux from Glucose-Loaded Microsomes Rat liver microsomal vesicles (10 mg protein/ml) were incubated in the presence of various concentrations (0.1–100 mM) of glucose and [3H]glucose for 2 h, then the medium was diluted 10-fold with glucose-free buffer and the intravesicular glucose content was measured with the filtration method. The dilution started a glucose efflux lasting for more than 60 min, when a new equilibrium was reached (Fig. 3). The initial rate of the resulting efflux calculated as a difference of “0” min (in accordance with the fact it was about 5–10 s) and 1-min measurements was linearly dependent on the extravesicular glucose concentrations during the preincubation and, consequently, on the intravesicular glucose content of microsomes (Fig. 2). The inclusion of 100 mM glucose in the dilution buffer did not cause trans stimulation of the initial rate of glucose efflux (not shown). N-Ethylmaleimide was also inhibitory in case of glucose efflux; the addition of 1 mM N-ethylmaleimide to the dilution buffer decreased the efflux from
FIG. 4. Glucose efflux deriving from glucose 6-phosphate hydrolysis in rat liver microsomal vesicles. Liver microsomes were incubated for 5 min in the presence of 0.2 mM glucose 6-phosphate plus [14C]glucose 6-phosphate and then glucose efflux was provoked by a 10-fold dilution in glucose and glucose 6-phosphate-free buffer (E). For comparison glucose efflux from microsomes passively loaded with 1 mM glucose is also shown (❐). Intravesicular glucose contents were measured selectively as detailed under Experimental Procedures. Data are means 6 SD of four experiments.
GLUCOSE TRANSPORT IN RAT LIVER MICROSOMES
137
DISCUSSION
The results of the present study demonstrate the heterogeneity of microsomal permeability to glucose. The nature of the microsomal glucose transport and the extent of intravesicular glucose pool have been disputed. Measurement of the glucose isotope space in microsomal pellets and the light-scattering behavior of glucose clearly indicated that the majority of microsomal vesicles are permeable for glucose (15–18). On the other hand, filtration-based methods resulted in the underestimation or even in the apparent absence of microsomal glucose permeability (11–14). The rapid filtration technique applied in the present study is obviously useless in the investigation of rapid glucose transport processes, since despite the low-resolution time of filtering, the time required by the washing procedure leads to the partial loss of intravesicular compounds. However, the disadvantage of this method can be advantageous for the selective investigation of a slow glucose transport process in rat liver microsomes. In contrast to previous studies we have found that microsomal vesicles take up glucose. The rate of uptake was dependent on the extracellular glucose concentration, started to be saturated at very high glucose concentration (Fig. 2), and reached a steady-state level in 10 min of incubation (Fig. 1). The steady-state intravesicular glucose space of the vesicles was about 0.5 ml/mg protein at all the extravesicular glucose concentrations tested. This value is far from the total intravesicular glucose space reported earlier (2.34 ml/mg protein) (15); this difference is attributable to the loss of the majority of intravesicular glucose from a fast releasable pool. The low intravesicular glucose content together with the overestimation of extravesicular water space by the erroneous use of mannitol as nonpermeant compound could lead to the assumption that microsomal membrane is not permeable to glucose in the outside to inside direction (14). Important questions are whether the slow glucose transport process is an independent entity, or whether it is only the tail of the fast glucose transport observed by light-scattering, or whether it is caused by the uneven distribution of the same transporter in microsomal vesicles. The latter two possibilities can be excluded because of the following points. (i) The inhibitors of the slow and fast transport processes are different: N-ethylmaleimide, for example, inhibited only the slow while pentamidine, cytochalasin B, genistein, and phloretin only the fast transport (see 15, and unpublished results mentioned under Results). (ii) t1/2 of the fast transport (influx) is a few seconds (15), while that of the slow transport (influx or efflux) is a few minutes. (iii) The presence of the fast transport in all vesicles would lead to a quicker equilibration during the influx measurement. However, both transport pro-
FIG. 5. Schematic representation of glucose spaces in rat liver microsomal vesicles. The figures indicate the relative size of various intravesicular spaces in ml/mg protein. Data are also taken from Refs. (7) and (15).
cesses are saturable at high (.100 mM) glucose concentration. Thus, we conclude that rat liver microsomal vesicles are heterogeneous concerning their glucose transport properties. Similar conclusion was drawn by Meissner and Allen (16), who found measuring radioisotope efflux from preloaded vesicles that about 70% of rat liver microsomes was highly permeable to different small molecules including glucose, while the remaining 30% was less permeable. In our experiments about 20% (apparent vesicular space of 0.5 ml/mg protein) of the glucose-permeable vesicles (apparent vesicular space of 2.34 ml/mg protein (15)) do not contain the component(s) responsible for the fast glucose transport. Interestingly this pool coincides with a part of the glucose-6-phosphatase system (Fig. 5). The coincidence results in the partial intraluminal retention of glucose deriving from glucose 6-phosphate hydrolysis (11–13). Due to the heterogeneity of transport, glucose produced intraluminally by glucose 6-phosphate hydrolysis is partially released fast from one part of vesicles and is partially accumulated in another part of vesicles. This phenomenon was observed previously (11–13) and interpreted as a bidirectional efflux from an intramembraneous catalytic site toward both the cytosol and the lumen of the ER. The initial rate of glucose influx and efflux (Fig. 2) and the efflux of glucose deriving from glucose 6-phosphate hydrolysis (Fig. 4) are similar, suggesting that they are mediated by the same (bidirectional) transport process. According to our previous paper (7) in the presence of 0.2 mM glucose 6-phosphate at the steady-state phase of uptake rat liver microsomal vesicles contain 1.4 nmol/mg intravesicular glucose. Since this value was obtained by using the same filtration method, it should be present in the slowly releasable intravesicular glucose pool. Taking into account that this pool represents
138
´ NHEGYI ET AL. BA
only the one-seventh of microsomes, it contains approx 10 nmol glucose/mg protein. The rate of efflux at this intravesicular glucose concentration— calculated on the basis of Fig. 2—is roughly 2.5 nmol/min/mg protein. During steady-state conditions this should be equal to the glucose production of this pool. It means that the slowly releasable glucose pool is responsible for approximately the one-third of microsomal total glucose production (7.4 nmol/min/mg protein), in accordance with that this pool represents one-third of the microsomal glucose 6-phosphate-accessible space. The characteristic features of the slow component of glucose transport (Table I) are different from those of mammalian plasma membrane GLUT transporters (1). In addition, the predominant liver plasma membrane GLUT 2 transporter has been shown to be absent from intracellular membranes using antiGLUT 2 C-terminal antibodies (25, 26, and Burchell and Benedetti, unpublished observation). In any case, the high KM of the microsomal glucose transport fits well into the decreasing order of the KM values (ER membrane, .100 mM; GLUT 2 in the plasma membrane of intact hepatocytes, 66 mM (27); GLUT 1,3,4 in the plasma membrane of nonhepatic cells, 1–5 mM (1); and hexokinase, 0.01– 0.1 mM (28)) which can serve efficiently the unidirectional glucose transport from the liver to the peripheral organs during gluconeogenic conditions. The existence of vesicles having different glucose transport properties in isolated microsomal fraction may be explained by assuming that in vivo the ER contains a limited number of highly permeable structures. Fragmentation during homogenization can yield vesicles containing this structure and others do not. Nevertheless, the hydrolytic component of the glucose6-phosphatase system has been reported to distribute evenly in microsomal vesicles. The existence of a slowly releasable glucose pool in intact hepatic ER would reduce the futile cycling between glucose and glucose 6-phosphate and it could mean that a significant amount of glucose produced by the liver has the potential to leave eventually by the exocytotic pathway. The in vivo role of these glucose pools, particularly in the functioning of the glucose-6-phosphatase system, remains to be established. ACKNOWLEDGMENTS The financial support of Telethon, Italy (Grant E.638 to A. Benedetti) is gratefully acknowledged. This work was also supported by grants from EC (Biotechnology Project PL97 0286), the Medical Research Council and Scottish Home and Health Department to A. Burchell who was a Lister Institute Research Fellow, and the Ciba Fellowship Trust to A. Burchell and A. Benedetti. G. Ba´nhegyi was a recipient of a Research Fellowship in Siena from the European Science Foundation Program of Fellowships in Toxicology and the “Eo¨tvo¨s” Hungarian State Fellowship.
REFERENCES 1. Gould, G. W., and Holman, G. D. (1993) Biochem. J. 295, 329 – 341. 2. Nordlie, R. C., and Sukalski, K. A. (1985) in The Enzymes of Biological Membranes (Martonosi, A. N., Ed.), Vol. 2, pp. 349 – 398, Plenum, New York. 3. Hemrika, W., and Wever, R. (1997) FEBS Lett. 409, 317–319. 4. Pan, C.-J., Lei, K.-J., Annabi, B., Hemrika, W., and Chou, J. Y. (1998) J. Biol. Chem. 273, 6144 – 6148. 5. Lei, K.-J., Shelly, L. L., Pan, C.-J., Sidbury, J. B., and Chou, J. Y. (1993) Science 262, 580 –583. 6. Shelly, L. L., Lei, K.-J., Pan, C.-J., Sakata, S. F., Ruppert, S., Schutz, G., and Chou, J. Y. (1993) J. Biol. Chem. 268, 21482– 21485. 7. Ba´nhegyi, G., Marcolongo, P., Fulceri, R., Hinds, C., Burchell, A., and Benedetti, A. (1997) J. Biol. Chem. 272, 13584 –13590. 8. Gerin, I., Veiga-da-Cunha, M., Achouri, Y., Collet, J.-F., and Van Schaftingen, E. (1998) FEBS Lett. 419, 235–238. 9. Burchell, A. (1998) Biochem. J. 331, 973. 10. Schulze, H.-U., Nolte, B., and Kannler, R. (1986) J. Biol. Chem. 261, 16571–16578. 11. Berteloot, A., Vidal, H., and van de Werve, G. (1991) J. Biol. Chem. 266, 5497–5507. 12. St-Denis, J.-F., Berteloot, A., Vidal, H., Annabi, B., and van de Werve, G. (1995) J. Biol. Chem. 270, 21092–21097. 13. Berteloot, A., St-Denis, J-F., and van de Werve, G. (1995) J. Biol. Chem. 270, 21098 –21102. 14. Romanelli, A., St-Denis, J-F., Vidal, H., Tchu, S., van de Werve, G. (1994) Biochem. Biophys. Res. Commun. 200, 1491–1497. 15. Marcolongo, P., Fulceri, R., Guinti, R., Burchell, A., and Benedetti, A. (1996) Biochem. Biophys. Res. Commun. 219, 916 – 922. 16. Meissner, G., and Allen, R. (1981) J. Biol. Chem. 256, 6413– 6422. 17. Meissner, G. (1988) Methods Enzymol. 157, 417– 437. 18. Fulceri, R., Gamberucci, A., Scott, H. M., Giunti, R., Burchell, A., and Benedetti, A. (1995) Biochem. J. 307, 391–397. 19. Henne, V., and So¨ling, H. D. (1986) FEBS Lett. 202, 267–273. 20. Burchell, A., Hume, R., and Burchell, B. (1988) Clin. Chim. Acta 173, 183–192. 21. Benedetti, A., Fulceri, R., Romani, A., and Comporti, M. (1988) J. Biol. Chem. 263, 3466 –3473. 22. Inamitsu, T., and Ohtsuki, I. (1984) Eur. J. Biochem. 145, 115– 121. 23. Ojcius, D. M., and Young, J. D.-E. (1991) Trends Biochem. Sci. 16, 225–229. 24. Fulceri, R., Ba´nhegyi, G., Gamberucci, A., Giunti, R., Mandl, J., and Benedetti, A. (1994) Arch. Biochem. Biophys. 309, 43– 46. 25. Thorens, B., Sarker, H. K., Kaback, H. R., and Lodish, H. F. (1988) Cell 55, 281–290. 26. Thorens, B., Cheng, Z.-Q., Brown, D., and Lodish, H. F. (1990) Am. J. Physiol. 259, 6279 – 6285. 27. Eliott, K. R. F., and Craik, J. D. (1983) Biochem. Soc. Trans. 10, 12–13. 28. Chen, Y. T., and Burchell, A. (1995) in The Metabolic and Molecular Basis of Inherited Disease (Scriver, C. R., Beaudet, A. L., Sly, S. W., and Valle, D., Eds.), pp. 935–965, McGraw–Hill, New York.
Vol. 359, No. 1, November 1, pp. 133–138, 1998 Article No. BB980888
Heterogeneity of Glucose Transport in Rat Liver Microsomal Vesicles Ga´bor Ba´nhegyi,* Paola Marcolongo,* Ann Burchell,† and Angelo Benedetti*,1 *Istituto di Patologia Generale, Universita` di Siena, 53100 Siena, Italy; and †Department of Obstetrics and Gynaecology, Ninewells Hospital and Medical School, University of Dundee, Dundee, DD1 9SY, Scotland
Received May 15, 1998, and in revised form August 5, 1998
Glucose transport across the membrane of rat liver microsomal vesicles was studied by a rapid filtration method in three different experimental systems: (i) inward transport in the presence of extravesicular glucose, (ii) efflux from passively preloaded vesicles, and (iii) efflux of glucose generated intravesicularly by glucose-6-phosphatase upon addition of glucose 6-phosphate were investigated. The apparent intravesicular glucose space estimated with the rapid filtration method was lower than the total microsomal glucose accessible space both the in the steady-state phase of uptake and at the starting point of efflux: 0.5 versus 2.3 ml/mg protein. The initial rate of influx/efflux was dependent on the extravesicular/intravesicular glucose concentration and was much lower than the rate of influx estimated previously by the light-scattering technique. Both influx and efflux could be inhibited by N-ethylmaleimide and possibly became saturable at high (>100 mM) glucose concentration. Known inhibitors of GLUT transporters (genistein, cytochalasin B, phloretin, and hexoses) did not affect glucose influx. The time course of glucose efflux from vesicles preincubated in the presence of glucose 6-phosphate was similar to that from glucoseloaded vesicles. These data together with that obtained previously (by a light-scattering technique; Marcolongo, P., Fulceri, R., Giunti, R., Burchell, A., and Benedetti, A. (1996) Biochem. Biophys. Res. Commun. 219, 916–922) indicate that microsomal vesicles are heterogeneous regarding their glucose-transporting properties and that glucose transport is bidirectional and its feature meets the requirements of a facilitative transport. © 1998 Academic Press Key Words: glucose; transport; liver microsomes; glucose-6-phosphatase.
1 To whom correspondence and reprint requests should be addressed at Istituto di Patologia Generale, Universita` di Siena, Viale Aldo Moro, 53100 Siena, Italy. Fax: (39) 577 227009. E-mail: [email protected].
0003-9861/98 $25.00 Copyright © 1998 by Academic Press All rights of reproduction in any form reserved.
The transport of glucose across the membranes of mammalian cells is of vital significance, since glucose has a central role in the homeostasis and intermediary metabolism of the organism. The plasma membrane of all mammalian cells possess member(s) of a family of facilitative glucose transporters (1). The transport of glucose through the endoplasmic reticulum (ER)2 membrane is less clear, despite the fact that this step is supposed to be necessary for the export of glucose produced by the liver. Both oligosaccharide processing of the newly synthesized glycoproteins and the hydrolysis of glucose 6-phosphate result in the release of glucose inside the ER cisternae. In fact, glucose-6-phosphatase, the enzyme catalyzing the last reaction of glucose-producing pathways glycogenolysis and gluconeogenesis, is an integral protein of (hepatic) ER (2) whose active site, according to the recent models of its structure, is situated inside the lumen of the ER (3, 4). The sequence of the enzyme has been reported (5, 6), and functional (7) and genetic (8) evidence for a ER transporter which mediates the translocation of the substrate glucose 6-phosphate from the cytosol to ER lumen has been forwarded. Phenomenological evidence for ER glucose transport system(s) is, however, still not univocal, and the molecular structure of the involved protein(s) is still elusive (9). In the “combined conformational flexibility–substrate transport model” (10 – 13) glucose-6-phosphatase traverses the microsomal membrane forming a water-filled space around the catalytic site in the membrane. Glucose formed in the membrane can enter the lumen of microsomal vesicles, where it accumulates, or can leave toward the cytosol (13). In a recent work (7) we found instead that glucose does derive from the hydrolysis of the intraluminal glucose 6-phosphate. This glucose can, at least in part, accumulate in the microsomal lumen until the hydro2
Abbreviation used: ER, endoplasmic reticulum. 133
134
´ NHEGYI ET AL. BA
lysis of the parent glucose 6-phosphate proceeds, which suggests that the sugar leaves a discrete microsomal pool with a kinetic possibly slower than that of glucose6-phosphate hydrolysis (7). Curiously, it has been also found that glucose is unable to enter the microsomal vesicles from outside (14). An intravesicular glucose space, however, could be detected after prolonged incubation of microsomes in the presence of glucose (15, 16), and it corresponded to 73% of the intravesicular water space (15). Moreover, glucose in high concentration rapidly enters (liver) microsomal vesicles, as it has been revealed by a light-scattering technique in other (17) and our laboratory (15, 18). This “fast” glucose influx cannot be detected by rapid filtration (with radiolabelled glucose) because it was too rapid (t1/2 ranging between 4 and 7.6 s) for this technique as reported/ discussed previously (15). The aim of the present study was to further investigate the influx and efflux of glucose in rat liver microsomal vesicles and the connection of glucose transport with glucose 6-phosphatase activity. Here we describe a glucose influx and efflux pathway, possibly a facilitative transport system for the sugar, which possesses a relatively slow kinetics and can therefore be studied by a conventional filtration technique. This system appears to be present in vesicles other than those provided with the aforementioned rapid transport pathway and is also operative for exporting outside vesicles glucose produced by glucose 6-phosphatase. These data indicate that microsomal vesicles are heterogeneous regarding their glucose transport pathways which suggests the existence of different functional glucose ER pools in the liver. EXPERIMENTAL PROCEDURES Preparation of liver microsomes. Twenty-four-hour-fasted male Sprague–Dawley rats (180 –230 g) were used. Liver microsomes were prepared as reported (19). Microsomal fractions were resuspended in buffer A containing (mM): KCl, 100; NaCl, 20; MgCl2, 1; Mops, 20, pH 7.2. The suspensions were rapidly frozen and maintained under liquid N2 until required. Intactness of microsomal membrane was greater than 90% in preparations employed as ascertained by measuring the latency of mannose-6-phosphatase activity (20). The extent of contamination of the microsomal fraction with plasma membrane was evaluated by measuring 59-nucleotidase activity (21). This activity (units/mg protein, mean 6 SEM, n 5 3) was 67 6 5 in liver homogenates and 90 6 8 in microsomal fractions, and its recovery in the microsomal fractions (as percentage of that of liver homogenate) was 5.9%. The extent of plasma membrane contamination was therefore similar to those previously measured in (purified) liver microsomal fractions (21). To exclude some contribution of plasma membrane-derived vesicles to glucose transport, we preliminary verified that the glucose influx in liver microsomes (see below) was unchanged (e.g., Table I) by digitonin concentration suitable to selectively permeabilise plasma membrane (22). Microsomal protein concentration was determined by the Biuret reaction using BSA as standard. Measurement of glucose influx and efflux. Liver microsomes (1 mg/ml) were incubated in buffer A containing 0.1–100 mM glucose plus [3H]glucose (9 mCi/ml) at 22°C. At the indicated time intervals, samples (0.1 mg protein) were taken to measure 3H associated with
microsomes. Samples were filtered rapidly through cellulose acetate/ nitrate filter membranes (pore size 0.22 mm; which retained 85 6 2% of the microsomal protein applied) and were washed with 4 ml of Hepes buffer (20 mM, pH 7.2) containing 250 mM sucrose. The time required to execute filtration and washing was 15–20 s. In a series of incubation the pore-forming antibiotic alamethicin (0.05 mg/ml) (23, 24) was added to distinguish the intravesicular and the bound radioactivity. More than 95% of the microsomal protein was retained by filters, indicating that the alamethicin treatment did not affect the vesicular structure of microsomes. The alamethicin-permeabilized microsomes retained amounts of radioactivity #20% of that associated to untreated microsomes. Intravesicular radioactive glucose was bona fide lost during the washing procedure, since the alamethicin nonreleasable portion did not further decrease even after extensive wash of filters (and microsomes). The alamethicin releasable portion of radioactivity was therefore regarded as intravesicular (7). To measure glucose efflux, microsomes (10 mg protein/ml)—preincubated with various concentrations of glucose plus [3H]glucose for 2 h at 22°C—were diluted (10-fold) with the incubation medium including no glucose and aliquots of the diluted incubates (0.1 mg protein) were drawn. In parallel samples, 50 mg/ml of alamethicin was also included in the dilution. For the measurement of glucose efflux derived from glucose 6-phosphate, microsomes (10 mg/ml) were incubated in buffer A in the presence of 0.2 mM glucose 6-phosphate plus D-[14C(U)]glucose 6-phosphate (2–3 mCi/ml) at 22°C. After 5 min incubation microsomes were diluted 10-fold with buffer A and samples were taken as described above. Parallel filters were treated after wash with ZnSO4–Ba(OH)2 to precipitate glucose 6-phosphate and measure D-[14C(U)]glucose derived from the parent D-[14C(U)]glucose 6-phosphate according to Ref. (7). Glucose-6-phosphatase activity was measured on the basis of glucose formation as reported earlier (7). Materials. Alamethicin, pentamidine, cytochalasin B, genistein, digitonin, phloretin, N-ethlymaleimide, glucose kit (Trinder method), DIDS, and D-[1-3H(N)]glucose (0.9 mCi/ml; 15.5 mCi/mmol) were from Sigma. D-[14C(U)]Glucose 6-phosphate (0.1 mCi/ml; 300 mCi/ mmol) was from American Radiolabeled Chemicals Inc. (St. Louis, MO). Cellulose acetate/nitrate filter membranes (pore size 0.22 mm) were from Millipore. All other chemicals were of analytical grade.
RESULTS
Microsomal Uptake of Glucose Rat liver microsomal vesicles incubated in the presence of various concentrations of glucose rapidly took up the radioactive tracer until a steady-state level was reached over a 10-min period of incubation (Fig. 1). The uptake was attributable to intravesicular accumulation, since (i) the majority of radioactivity (.80%) could be released by permeabilizing the vesicles with the pore-forming antibiotics alamehicin and (ii) the shrinkage of vesicles upon the addition of the poorly permeant compound sucrose (250 mM) to the medium reduced the uptake by 70% (Fig. 1). The steady-state level of intravesicular glucose was stable for at least 2 h and its extent was dependent on the extravesicular concentration of glucose (data not shown). In the steady-state phase of glucose uptake the apparent intravesicular glucose space— calculated by dividing the intravesicular glucose content (nmol/mg protein) with the extravesicular glucose concentration (nmol/ml)—was 0.43 6 0.09 (ml/mg protein) at 1 mM
135
GLUCOSE TRANSPORT IN RAT LIVER MICROSOMES
FIG. 1. Glucose uptake by rat liver microsomal vesicles. Liver microsomes were incubated in the presence of 1 mM glucose plus [3H]glucose as tracer and intravesicular glucose contents were measured as detailed under Experimental Procedures. The effect of the addition of 0.1 mM N-ethylmaleimide and 250 mM sucrose on the uptake is also shown. Microsomes were treated with N-ethylmaleimide for 30 min, while sucrose was added immediately prior to start glucose uptake. Sucrose did not affect intactness of microsomes as ascertained by measuring the latency of mannose-6-phosphatase activity (20). Data are means 6 SD of six experiments (control, E) or means of two experiments (sucrose, ❐; and N-ethylmaleimide, ‚).
glucose concentration. The apparent space was independent of the extravesicular glucose concentration; other concentrations than 1 mM gave similar results (0.41– 0.55 ml/mg protein). This space is about 12% of the total intravesicular volume of the (liver) microsomes estimated as 3H2O space (3.6 ml, Ref. 15). Initial rate of glucose influx was estimated by measuring 1-min uptakes. The increase of the concentration of extravesicular glucose from 0.1 to 100 mM resulted in an almost linearly increasing influx rate with a tendency of saturation (Fig. 2). Inconsistent data were obtained at glucose concentrations higher than 100 mM, probably due to the shrinkage of the vesicles caused by hyperosmolarity (not shown). The glucose influx was not affected by digitonin concentration suitable to selectively permeabilize plasma membranederived vesicles (Table I), which makes unlikely their contribution to the glucose uptake by the microsomal fraction used. The effect of potential inhibitors of known glucose transporters was also tested (Table I). Phloretin, cytochalasin B, genistein, pentamidine, fructose, galactose, and mannose were ineffective. Phloretin and genistein resulted instead in an apparent stimulation of glucose uptake (Table I). This could be due to the inhibition of the fast glucose transport (14) which became therefore measurable by the rapid filtration assay. Consistently, we observed by the lightscattering assay that phloretin and genistein increased the t1/2 of (50 mM) glucose influx from 6.5 to 46 and 43 s, respectively (Ba´ nhegyi, Marcolongo, and
FIG. 2. The rate of glucose uptake by and glucose efflux from rat liver microsomal vesicles. Intravesicular glucose content of microsomes—incubated in the presence of or preloaded with 0.1–100 mM glucose plus [3H]glucose as tracer—was measured at “0” min (efflux) and at 1-min incubation time. Glucose did not affect intactness of microsomes as ascertained by measuring the latency of mannose-6phosphatase activity (20). Uptake and efflux rates were calculated as a difference. For the experimental protocol see the Experimental Procedures. Data are means of 2– 8 experiments; SD bars were omitted for clarity.
Benedetti, unpublished observations). Both cytochalasin B and pentamidine have been also previously reported by us (15) to inhibit the microsomal fast glucose
TABLE I
Effect of Inhibitors and Hexoses on the Microsomal Uptake of Glucose
Compound (mM)
Glucose uptake (nmol/min/mg protein)
None Digitonina N-Ethylmaleimide (0.1) N-Ethylmaleimide (0.5) N-Ethylmaleimide (5) Phloretin (0.25) Cytochalasin B (0.25) Genistein (0.1) Pentamidine (0.1) Fructose (50) Galactose (50) Mannose (50) Glucose (50)
0.24 6 0.02 0.23 6 0.02 0.16 6 0.03** 0.12 6 0.03** 0.08 6 0.02** 0.32 6 0.05* 0.24 6 0.03 0.33 6 0.08* 0.22 6 0.04 0.23 6 0.04 0.22 0.22 6 0.03 0.23
Note. Liver microsomes were incubated in the presence of 1 mM glucose plus [3H]glucose as tracer for 1 min and intravesicular glucose content was measured as detailed under Experimental Procedures. Microsomes were treated with the different compounds for 30 min prior to start glucose uptake. Data are means of two or means 6 SD of three to five experiments. Values marked * and ** are significantly different from the control value: P , 0.05 and P , 0.01, respectively. a 20 mg/mg of microsomal protein.
136
´ NHEGYI ET AL. BA
vesicles loaded in the presence of 1 mM glucose by 63 6 13% (mean 6 SD, n 5 4). Efflux of Glucose Deriving from Glucose 6-Phosphate Hydrolysis
FIG. 3. Glucose efflux from glucose preloaded rat liver microsomal vesicles. Microsomes were loaded with 10 mM glucose plus [3H]glucose as tracer for 2 h and glucose efflux was provoked by a 10-fold dilution in a glucose-free buffer. Intravesicular glucose contents were measured as detailed under Experimental Procedures. Data are means 6 SD of six experiments.
transport, but the t1/2 of glucose transport was still lower (about 12 s) than the time required for accomplishing the rapid filtration assay (about 15 s). This presumably resulted in the lack of measurability of the fast glucose transport/uptake (as discussed in detail in Ref. 15). N-Ethylmaleimide was inhibitory in a concentration-dependent way, although a complete inhibition was not attained (Table I). The inhibitory effect of 0.1 mM N-ethylmaleimide on the time course of glucose influx is shown in Fig. 1. We have previously observed (15) that 0.1 mM N-ethylmaleimide did not inhibit the fast glucose transport.
It was previously observed by using a rapid filtration technique that a portion of glucose derived from the hydrolysis of glucose 6-phosphate by glucose-6-phosphatase enzyme is still retained into microsomal vesicles until the hydrolysis of the sugar phosphoester proceeds (7, 13). The remaining portion of intralumenally formed glucose should more rapidly leave the vesicles since it was not recovered in the microsomal fraction by the ultrafiltration technique and appeared soon in the extravesicular fluid (7). Further experiments, therefore, were drawn to characterize the efflux of this glucose once formed from glucose-6-phosphate hydrolysis. Rat liver microsomal vesicles (10 mg/ml) were incubated in the presence of 0.2 mM glucose 6-phosphate for 5 min. Due to the high hydrolytic activity (7.4 6 1.8 nmol/min/mg protein) at that time more than 90% of added glucose 6-phosphate had been converted to glucose. At this time microsomes were diluted 10-fold to decrease the rate of further hydrolysis (its rate was less than 0.04 nmol/min/mg protein during the efflux measurement) and the efflux of the intravesicularly accumulated glucose was measured. The initial intravesicular glucose contents in these experiments were similar to those in which microsomes were loaded by incubating them in the presence of 1 mM glucose; moreover, the shape of the two efflux curves was also similar (Fig. 4).
Glucose Efflux from Glucose-Loaded Microsomes Rat liver microsomal vesicles (10 mg protein/ml) were incubated in the presence of various concentrations (0.1–100 mM) of glucose and [3H]glucose for 2 h, then the medium was diluted 10-fold with glucose-free buffer and the intravesicular glucose content was measured with the filtration method. The dilution started a glucose efflux lasting for more than 60 min, when a new equilibrium was reached (Fig. 3). The initial rate of the resulting efflux calculated as a difference of “0” min (in accordance with the fact it was about 5–10 s) and 1-min measurements was linearly dependent on the extravesicular glucose concentrations during the preincubation and, consequently, on the intravesicular glucose content of microsomes (Fig. 2). The inclusion of 100 mM glucose in the dilution buffer did not cause trans stimulation of the initial rate of glucose efflux (not shown). N-Ethylmaleimide was also inhibitory in case of glucose efflux; the addition of 1 mM N-ethylmaleimide to the dilution buffer decreased the efflux from
FIG. 4. Glucose efflux deriving from glucose 6-phosphate hydrolysis in rat liver microsomal vesicles. Liver microsomes were incubated for 5 min in the presence of 0.2 mM glucose 6-phosphate plus [14C]glucose 6-phosphate and then glucose efflux was provoked by a 10-fold dilution in glucose and glucose 6-phosphate-free buffer (E). For comparison glucose efflux from microsomes passively loaded with 1 mM glucose is also shown (❐). Intravesicular glucose contents were measured selectively as detailed under Experimental Procedures. Data are means 6 SD of four experiments.
GLUCOSE TRANSPORT IN RAT LIVER MICROSOMES
137
DISCUSSION
The results of the present study demonstrate the heterogeneity of microsomal permeability to glucose. The nature of the microsomal glucose transport and the extent of intravesicular glucose pool have been disputed. Measurement of the glucose isotope space in microsomal pellets and the light-scattering behavior of glucose clearly indicated that the majority of microsomal vesicles are permeable for glucose (15–18). On the other hand, filtration-based methods resulted in the underestimation or even in the apparent absence of microsomal glucose permeability (11–14). The rapid filtration technique applied in the present study is obviously useless in the investigation of rapid glucose transport processes, since despite the low-resolution time of filtering, the time required by the washing procedure leads to the partial loss of intravesicular compounds. However, the disadvantage of this method can be advantageous for the selective investigation of a slow glucose transport process in rat liver microsomes. In contrast to previous studies we have found that microsomal vesicles take up glucose. The rate of uptake was dependent on the extracellular glucose concentration, started to be saturated at very high glucose concentration (Fig. 2), and reached a steady-state level in 10 min of incubation (Fig. 1). The steady-state intravesicular glucose space of the vesicles was about 0.5 ml/mg protein at all the extravesicular glucose concentrations tested. This value is far from the total intravesicular glucose space reported earlier (2.34 ml/mg protein) (15); this difference is attributable to the loss of the majority of intravesicular glucose from a fast releasable pool. The low intravesicular glucose content together with the overestimation of extravesicular water space by the erroneous use of mannitol as nonpermeant compound could lead to the assumption that microsomal membrane is not permeable to glucose in the outside to inside direction (14). Important questions are whether the slow glucose transport process is an independent entity, or whether it is only the tail of the fast glucose transport observed by light-scattering, or whether it is caused by the uneven distribution of the same transporter in microsomal vesicles. The latter two possibilities can be excluded because of the following points. (i) The inhibitors of the slow and fast transport processes are different: N-ethylmaleimide, for example, inhibited only the slow while pentamidine, cytochalasin B, genistein, and phloretin only the fast transport (see 15, and unpublished results mentioned under Results). (ii) t1/2 of the fast transport (influx) is a few seconds (15), while that of the slow transport (influx or efflux) is a few minutes. (iii) The presence of the fast transport in all vesicles would lead to a quicker equilibration during the influx measurement. However, both transport pro-
FIG. 5. Schematic representation of glucose spaces in rat liver microsomal vesicles. The figures indicate the relative size of various intravesicular spaces in ml/mg protein. Data are also taken from Refs. (7) and (15).
cesses are saturable at high (.100 mM) glucose concentration. Thus, we conclude that rat liver microsomal vesicles are heterogeneous concerning their glucose transport properties. Similar conclusion was drawn by Meissner and Allen (16), who found measuring radioisotope efflux from preloaded vesicles that about 70% of rat liver microsomes was highly permeable to different small molecules including glucose, while the remaining 30% was less permeable. In our experiments about 20% (apparent vesicular space of 0.5 ml/mg protein) of the glucose-permeable vesicles (apparent vesicular space of 2.34 ml/mg protein (15)) do not contain the component(s) responsible for the fast glucose transport. Interestingly this pool coincides with a part of the glucose-6-phosphatase system (Fig. 5). The coincidence results in the partial intraluminal retention of glucose deriving from glucose 6-phosphate hydrolysis (11–13). Due to the heterogeneity of transport, glucose produced intraluminally by glucose 6-phosphate hydrolysis is partially released fast from one part of vesicles and is partially accumulated in another part of vesicles. This phenomenon was observed previously (11–13) and interpreted as a bidirectional efflux from an intramembraneous catalytic site toward both the cytosol and the lumen of the ER. The initial rate of glucose influx and efflux (Fig. 2) and the efflux of glucose deriving from glucose 6-phosphate hydrolysis (Fig. 4) are similar, suggesting that they are mediated by the same (bidirectional) transport process. According to our previous paper (7) in the presence of 0.2 mM glucose 6-phosphate at the steady-state phase of uptake rat liver microsomal vesicles contain 1.4 nmol/mg intravesicular glucose. Since this value was obtained by using the same filtration method, it should be present in the slowly releasable intravesicular glucose pool. Taking into account that this pool represents
138
´ NHEGYI ET AL. BA
only the one-seventh of microsomes, it contains approx 10 nmol glucose/mg protein. The rate of efflux at this intravesicular glucose concentration— calculated on the basis of Fig. 2—is roughly 2.5 nmol/min/mg protein. During steady-state conditions this should be equal to the glucose production of this pool. It means that the slowly releasable glucose pool is responsible for approximately the one-third of microsomal total glucose production (7.4 nmol/min/mg protein), in accordance with that this pool represents one-third of the microsomal glucose 6-phosphate-accessible space. The characteristic features of the slow component of glucose transport (Table I) are different from those of mammalian plasma membrane GLUT transporters (1). In addition, the predominant liver plasma membrane GLUT 2 transporter has been shown to be absent from intracellular membranes using antiGLUT 2 C-terminal antibodies (25, 26, and Burchell and Benedetti, unpublished observation). In any case, the high KM of the microsomal glucose transport fits well into the decreasing order of the KM values (ER membrane, .100 mM; GLUT 2 in the plasma membrane of intact hepatocytes, 66 mM (27); GLUT 1,3,4 in the plasma membrane of nonhepatic cells, 1–5 mM (1); and hexokinase, 0.01– 0.1 mM (28)) which can serve efficiently the unidirectional glucose transport from the liver to the peripheral organs during gluconeogenic conditions. The existence of vesicles having different glucose transport properties in isolated microsomal fraction may be explained by assuming that in vivo the ER contains a limited number of highly permeable structures. Fragmentation during homogenization can yield vesicles containing this structure and others do not. Nevertheless, the hydrolytic component of the glucose6-phosphatase system has been reported to distribute evenly in microsomal vesicles. The existence of a slowly releasable glucose pool in intact hepatic ER would reduce the futile cycling between glucose and glucose 6-phosphate and it could mean that a significant amount of glucose produced by the liver has the potential to leave eventually by the exocytotic pathway. The in vivo role of these glucose pools, particularly in the functioning of the glucose-6-phosphatase system, remains to be established. ACKNOWLEDGMENTS The financial support of Telethon, Italy (Grant E.638 to A. Benedetti) is gratefully acknowledged. This work was also supported by grants from EC (Biotechnology Project PL97 0286), the Medical Research Council and Scottish Home and Health Department to A. Burchell who was a Lister Institute Research Fellow, and the Ciba Fellowship Trust to A. Burchell and A. Benedetti. G. Ba´nhegyi was a recipient of a Research Fellowship in Siena from the European Science Foundation Program of Fellowships in Toxicology and the “Eo¨tvo¨s” Hungarian State Fellowship.
REFERENCES 1. Gould, G. W., and Holman, G. D. (1993) Biochem. J. 295, 329 – 341. 2. Nordlie, R. C., and Sukalski, K. A. (1985) in The Enzymes of Biological Membranes (Martonosi, A. N., Ed.), Vol. 2, pp. 349 – 398, Plenum, New York. 3. Hemrika, W., and Wever, R. (1997) FEBS Lett. 409, 317–319. 4. Pan, C.-J., Lei, K.-J., Annabi, B., Hemrika, W., and Chou, J. Y. (1998) J. Biol. Chem. 273, 6144 – 6148. 5. Lei, K.-J., Shelly, L. L., Pan, C.-J., Sidbury, J. B., and Chou, J. Y. (1993) Science 262, 580 –583. 6. Shelly, L. L., Lei, K.-J., Pan, C.-J., Sakata, S. F., Ruppert, S., Schutz, G., and Chou, J. Y. (1993) J. Biol. Chem. 268, 21482– 21485. 7. Ba´nhegyi, G., Marcolongo, P., Fulceri, R., Hinds, C., Burchell, A., and Benedetti, A. (1997) J. Biol. Chem. 272, 13584 –13590. 8. Gerin, I., Veiga-da-Cunha, M., Achouri, Y., Collet, J.-F., and Van Schaftingen, E. (1998) FEBS Lett. 419, 235–238. 9. Burchell, A. (1998) Biochem. J. 331, 973. 10. Schulze, H.-U., Nolte, B., and Kannler, R. (1986) J. Biol. Chem. 261, 16571–16578. 11. Berteloot, A., Vidal, H., and van de Werve, G. (1991) J. Biol. Chem. 266, 5497–5507. 12. St-Denis, J.-F., Berteloot, A., Vidal, H., Annabi, B., and van de Werve, G. (1995) J. Biol. Chem. 270, 21092–21097. 13. Berteloot, A., St-Denis, J-F., and van de Werve, G. (1995) J. Biol. Chem. 270, 21098 –21102. 14. Romanelli, A., St-Denis, J-F., Vidal, H., Tchu, S., van de Werve, G. (1994) Biochem. Biophys. Res. Commun. 200, 1491–1497. 15. Marcolongo, P., Fulceri, R., Guinti, R., Burchell, A., and Benedetti, A. (1996) Biochem. Biophys. Res. Commun. 219, 916 – 922. 16. Meissner, G., and Allen, R. (1981) J. Biol. Chem. 256, 6413– 6422. 17. Meissner, G. (1988) Methods Enzymol. 157, 417– 437. 18. Fulceri, R., Gamberucci, A., Scott, H. M., Giunti, R., Burchell, A., and Benedetti, A. (1995) Biochem. J. 307, 391–397. 19. Henne, V., and So¨ling, H. D. (1986) FEBS Lett. 202, 267–273. 20. Burchell, A., Hume, R., and Burchell, B. (1988) Clin. Chim. Acta 173, 183–192. 21. Benedetti, A., Fulceri, R., Romani, A., and Comporti, M. (1988) J. Biol. Chem. 263, 3466 –3473. 22. Inamitsu, T., and Ohtsuki, I. (1984) Eur. J. Biochem. 145, 115– 121. 23. Ojcius, D. M., and Young, J. D.-E. (1991) Trends Biochem. Sci. 16, 225–229. 24. Fulceri, R., Ba´nhegyi, G., Gamberucci, A., Giunti, R., Mandl, J., and Benedetti, A. (1994) Arch. Biochem. Biophys. 309, 43– 46. 25. Thorens, B., Sarker, H. K., Kaback, H. R., and Lodish, H. F. (1988) Cell 55, 281–290. 26. Thorens, B., Cheng, Z.-Q., Brown, D., and Lodish, H. F. (1990) Am. J. Physiol. 259, 6279 – 6285. 27. Eliott, K. R. F., and Craik, J. D. (1983) Biochem. Soc. Trans. 10, 12–13. 28. Chen, Y. T., and Burchell, A. (1995) in The Metabolic and Molecular Basis of Inherited Disease (Scriver, C. R., Beaudet, A. L., Sly, S. W., and Valle, D., Eds.), pp. 935–965, McGraw–Hill, New York.