Partial purification of the D-glucose transport system in rat adipocyte plasma membranes.

Partially purified rat adipocyte plasma membranes were subjected to selective protein extractions using sodium hydroxide, sodium iodide, and dimethylmaleic anhydride in an effort to identify membrane components associated with glucose transport. Up to 80% of the membrane proteins were extracted in these procedures with the remaining proteins consisting almost entirely of two glycoprotein fractions, 78,000 and 94,000 daltons, as determined by dodecyl sulfate polyacrylamide gel electrophoresis. Sonic disruption of these extracted membrane residues resulted in the formation of vesicular structures as revealed by electron microscopy. These vesicles demonstrated high affinity binding of [3H]cytochalasin B in all three preparations. Vesicles from the dimethylmaleic anhydride-extracted membranes also exhibited a marked stereospecific uptake of o-glucose compared to L-glucose, as measured by a rapid filtration method. This uptake was markedly inhibited by cytochalasin B and this inhibition closely paralleled the high affinity binding of [3H]cytochalasin B to these vesicles. In addition, uptake of o-[3H]glucose was inhibited by phoretin, phlorizin, and dipyridamole, all potent inhibitors of n-glucose transport in the intact adipocyte. Competitive inhibitors of glucose transport such as 3-O-methylglucose and unlabeled D-ghcase itself, also inhibited uptake, while L-glucose and sucrose exhibited almost no effect. These results are consistent with the hypothesis that fat cell hexose transport system activity is associated with membrane components in one or both of the 78,000and 94,000-dalton glycoprotein fractions.

Partially purified rat adipocyte plasma membranes were subjected to selective protein extractions using sodium hydroxide, sodium iodide, and dimethylmaleic anhydride in an effort to identify membrane components associated with glucose transport. Up to 80% of the membrane proteins were extracted in these procedures with the remaining proteins consisting almost entirely of two glycoprotein fractions, 78,000 and 94,000 daltons, as determined by dodecyl sulfate polyacrylamide gel electrophoresis. Sonic disruption of these extracted membrane residues resulted in the formation of vesicular structures as revealed by electron microscopy. These vesicles demonstrated high affinity binding of [3H]cytochalasin B in all three preparations. Vesicles from the dimethylmaleic anhydride-extracted membranes also exhibited a marked stereospecific uptake of o-glucose compared to L-glucose, as measured by a rapid filtration method. This uptake was markedly inhibited by cytochalasin B and this inhibition closely paralleled the high affinity binding of [3H]cytochalasin B to these vesicles. In addition, uptake of o-[3H]glucose was inhibited by phoretin, phlorizin, and dipyridamole, all potent inhibitors of n-glucose transport in the intact adipocyte.
Competitive inhibitors of glucose transport such as 3-O-methylglucose and unlabeled D-ghcase itself, also inhibited uptake, while L-glucose and sucrose exhibited almost no effect. These results are consistent with the hypothesis that fat cell hexose transport system activity is associated with membrane components in one or both of the 78,000-and 94,000-dalton glycoprotein fractions.
Hexose transport has been characterized in a variety of mammalian cell types as operating via a system of facilitated diffusion. Although the human erythrocyte has been the model of major interest in studying this mode of transport, other systems have come under closer scrutiny in recent years (l-4). This has been due, in part, to improved techniques in * This work was supported by grants from the United States Public Health Service (AM 17893) and the Juvenile Diabetes 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 %duertisement" in accordance with 18 U.S.C. Section cell isolation as well as in transport measurement itself (5-7). The rat adipocyte is one such cell type in which hexose transport has been studied extensively, particularly in regard to modulation of this transport by insulin (see Refs. 8 and 9 for reviews). Glucose transport has also been characterized in adipocyte "ghosts" (10,111, as well as various plasma membrane preparations (12,13). Furthermore, it has been shown that elevated rates of stereospecific glucose transport persists in membrane vesicles derived from insulin-treated fat cells (8,14). Fat cell membrane vesicles have also been shown to exhibit high affinity binding sites for cytochalasin B which appear to be associated with the inhibition of glucose transport by this agent (13). Thus, in recent years, these plasma membrane preparations have become excellent models for studying hexose transport as well as hormonal regulation of this system.
In the red cell a variety of methods have been employed in attempts to identify the membrane components associated with hexose transport activity. Methods using affinity and differential labeling of membranes have begun to prove successful in localizing transport-associated components to specific membrane protein bands on polyacrylamide gels (E-18). Other approaches have utilized various extraction procedures in attempts to purify transport-associated (18)(19)(20)(21) and high affinity cytochalasin B-binding proteins (18,21,22). Very recently, successful reconstitution of hexose transport activity has been achieved subsequent to incorporation of partially purified erythrocyte membrane proteins into phospholipid vesicles (20,(23)(24)(25). We report here that extraction of partially purified fat cell plasma membranes with 2,3-dimethylmaleic anhydride yields a membrane preparation which consists almost entirely of two glycoprotein bands on SDSI-polyacrylamide gel electrophoresis.
These extracted membranes retain stereospecific n-glucose transport activity that is sensitive to a variety of known inhibitors of glucose transport. Furthermore, the extracted membranes contain high affinity cytochalasin B binding sites which appear to be involved in the inhibition of glucose transport by cytochalasin B. with a crude fat cell plasma membrane preparation as well as a purified plasma membrane preparation (26). We found that the crude membrane preparation was often more sensitive to selective extraction by dimethylmaleic anhydride than the more purified preparation obtained by high speed centrifugation on sucrose density gradients.
The crude preparations of plasma membranes were prepared by homogenization of cells in icecold Buffer A in a loose fitting glass homogenizing tube with seven up and down strokes using a Teflon pestle. The homogenate was centrifuged at 8500 x g for 10 min and the supernatant as well as a small amount of fluffy white material collected from the surface of the brown (mitochondrial) pellet were then centrifuged at 40,000 x g for 30 min and the resulting pellet resuspended in ice-cold 1 mM EDTA, 5 rnM Tris (pH 7.5).
Membrane Extractions ~ Extraction of plasma membranes with NaOH and dimethylmaleic anhydride was similar to the method of Steck and Yu (27). For NaOH extraction 1 volume of membrane suspension (4 mg/ml) was suspended in 10 volumes of ice cold water previously brought to pH 12 with 1 N NaOH. The extraction proceeded for 5 min after which the solution was neutralized to pH 7.0 with 0.1 N HCl and centrifuged at 1.2 x 10' g-min on a Sorvall RC-2 centrifuge.
For dimethylmaleic anhydride extractions, 1 volume of membranes was added to 15 volumes of water and 2 mg/ml of solid dimethylmaleic anhydride was added with constant stirring while maintaining the pH at 8.0 with continuous aliquots of a 2 N NaOH solution.
After acid ceased to evolve, the suspension was centrifuged as above and the pellet resuspended in 5 mM Tris buffer, 1 mM EDTA, pH 6.8. This suspension was either frozen overnight and the next day centrifuged and the pellet resuspended in Krebs-Ringer phosphate buffer, pH 7.4, for uptake and binding studies, or centrifuged and resuspended immediately after preparation. NaI extraction was based on the method of Kahlenberg (19). One volume of membrane suspension was suspended in 7 volumes of 1 M NaI in water, adjusted to pH 7.5, and incubated for 30 min after which the suspension was centrifuged as above.  (13) to exhibit stereospecific uptake of Dglucose after brief sonication. We subjected such fat cell mem-Purification of the Adipocyte Hexose Transport System branes to various protein extraction procedures which have proved selective in red cells (19,21,27) and tested the sedimented membrane residues for n-glucose uptake after extraction. Fig. 1 shows the remaining membrane polypeptides following extraction of a crude plasma membrane preparation under alkaline conditions using NaOH, hyperosmolar NaI, or dimethylmaleic anhydride. Although all of these substances effected selective elution of membrane proteins, dimethylmaleic anhydride was by far the most effective, followed by NaOH and NaI, respectively. In all three cases the membrane proteins in the 94,000-and 78,000-dalton region of dodecyl sulfate gels remained with the pellet while most other Coomassie blue-staining proteins were selectively eluted to varying degrees. After extraction, the membrane residues were washed, resuspended in Krebs-Ringer phosphate buffer, and dispersed in this buffer using a Brinkmann Polytron apparatus before [3H]cytochalasin B binding and n-glucose uptake were monitored. Although all three preparations exhibited high affinity cytochalasin B binding capacity, only the sodium iodide and dimethylmaleic anhydride-extracted membrane residues exhibited stereospecific n-glucose uptake (data not shown). Since the dimethylmaleic anhydride extraction was the most selective in eluting membrane proteins, this preparation was used for all subsequent investigations. Fig. 2 represents densitometric scans of SDS-polyacrylamide gels stained with Coomassie blue or Schiff reagent following electrophoresis of the control membranes and those extracted with dimethylmaleic anhydride. Crude unextracted plasma membranes were resolved into approximately 14 bands on 5% acrylamide gels and resembled the pattern previously observed by several laboratories (34)(35)(36) in that two major glycoprotein species (Bands I and II) were observed at 94,000 and 78,000 daltons. The prominence of the 56,000-dalton band (Band III) in this crude plasma membrane preparation rela-tive to purer preparations (26) reflects additional protein contamination in this region from other subcellular fractions (see Ref. 34). In some experiments up to 80% of the total membrane protein was released from the plasma membrane preparation in the presence of dimethylmaleic anhydride. The dimethylmaleic anhydride-extracted pellet consisted chiefly of glycoprotein Bands I and.11 and a very small amount of the material in the Band III region (Fig. 2B). The values indicated above each region represent the per cent contribution these bands make to the total protein content of the membrane based on Coomassie blue-staining intensity. In the experiment illustrated in Fig. 2, Band I (94,000 daltons) constituted approximately 10% of the plasma membrane preparation, while it composed 69% of the dimethylmaleic anhydride pellet preparation. Similarly, Band II (78,000 daltons) was enriched from 4.4% in control membranes to 25% of the extracted pellet. On the other hand, the major Coomassie blue-staining protein in the control plasma membrane preparation, Band III (56,000 daltons), constituted 37% of the original preparation, but only 5.6% of the extracted pellet. Band III was the only other constituent of this pellet which routinely accounted for greater than 1% of the total staining polypeptides. However, several dimethylmaleic anhydride-extracted preparations have had virtually no detectable protein material in this region of dodecyl sulfate gels (not shown).
Results of analyses of the composition of extracted membranes in three experiments are presented in Table I. While the average total protein recovered in the extracted pellets in the experiments presented was about 25% of the starting material, the relative phospholipid content increased by 35%. Both neutral sugar and sialic acid content markedly increased on a per mg of protein basis in the dimethylmaleic anhydrideextracted membrane pellet. There was significant solubilization of all three membrane components by dimethylmaleic FIG. 1. Selective release of membrane polypeptides by various extraction procedures. Following extraction of plasma membrane suspensions by various agents (see below) the suspensions were centrifuged and each pellet was sampled for electrophoresis. A, unextracted rat adipocyte plasma membrane preparation; B, NaOH: 30-min incubation of 1 ml of membrane suspension (4 mg/ml) in 7 ml of ice cold water adjusted to pH 12 with NaOH; C, Nal: 1 ml of membrane suspension was mixed with 10 ml of 1 M NaI and incubated on ice for 30 min; D, 2,3-dimethylmaleic anhydride (DMMA): 1 ml of membrane suspension was suspended in 7 ml of H,O (22") and solid 2,3-dimethylmaleic anhydride (1.25 mg/mg of protein) was added while maintaining the pH of the mixture at 8.0 with 2 N NaOH until the evolution of acid ceased.  anhydride, however. Interestingly, extraction of membranes with this agent resulted in the solubilization of the low molecular weight PAS-staining glycoprotein designated Band III (Fig. 2), which is suspected of being an endoplasmic reticulum contaminant (34). In these extractions, over 60% of the membrane phospholipid was eluted, but the resulting pellet was still enriched in phospholipids, due to the even greater extraction of proteins from the membrane (70 to 80%). Theoretically, the extraction of all bands other than I and II should result in a recovery of about 13% of the total protein in the membrane pellet based on the gel scan in Fig. 2A. However, since membrane glycoproteins often stain poorly with Coomassie blue, their relative contribution to the total membrane protein may be underestimated.
Therefore, it is not feasible to make an exact quantitative comparison between the data in Table I and Fig. 2. While it is possible that other proteins are present in the extracted membrane pellets, the data in Figs. 1 and 2 indicate that if this is the case either these proteins must be present in very small amounts or do not stain with Coomassie blue stain.
Vesicles were prepared from these extracted pellets with a Brinkmann Polytron ( Activity also markedly decreased after several hours at room temperature. In addition to cytochalasin B, other known inhibitors of Dglucose transport in fat cells were tested for the ability to inhibit n-glucose transport by these dimethylmaleic anhydride-extracted membranes (Table  II). In these experiments phlorizin, phloretin, and dipyridamole were potent inhibitors of transport while unlabeled sugars such as 90 mM 3-O-methyl glucose and n-glucose also inhibited, but to a lesser degree. Interestingly, L-glucose itself inhibited n-glucose uptake about 20%. This may reflect an osmotic shrinking of the vesicles rather than inhibition of influx since 90 mM sucrose also exhibited a similar effect (data not shown).
Alternatively, this small inhibition may reflect a low affinity of L-glucose for the n-glucose transport system as has been observed in intact fat cells (9) binding sites through purification, or a combination of both. The existence of several classes of high affinity cytochalasin B binding sites is certainly not the only way to interpret these data. It is also possible that the nonlinearity of binding may be due to negative cooperativity.
However, a dissociation constant of 5 x 10m7 M agrees well with previous findings for the intact plasma membrane (13). Furthermore, there is good agreement between this dissociation constant and the value of 7 x 10m7 M as the concentration of cytochalasin B required to half-maximally inhibit glucose transport activity (Fig. 6). Thus, these results support the notion that the way in which cytochalasin B exerts its potent inhibitory effects on hexose transport is closely associated with its high affinity binding to the membrane.

DISCUSSION
Rat adipocyte plasma membrane preparations have previously been shown (12, 13) to exhibit preferential uptake of Dversus L-glucose. In the present study we have demonstrated that this facilitated transport system is retained after extraction of plasma membranes with dimethylmaleic anhydride which results in elution of up to 80% of the total membrane protein in these preparations. Steck and Yu (2'7) have shown that dimethylmaleic anhydride reacts covalently with membrane proteins and electrostatically dissociates the polypeptides from the plasma membrane. In the red cell these workers found that membrane glycoproteins exhibiting strong hydrophobic associations with the membrane were the most resistant to extraction by this agent. In addition, Lin and Spudich (21) observed that high affinity cytochalasin B binding was also resistant to extraction of red cell membranes with dimethylmaleic anhydride, while Kahlenberg (19) demonstrated an increase in the specific activity of n-glucose uptake after such an extraction. In the present work we observed that subsequent to dimethylmaleic anhydride extraction of the rat adipocyte plasma membrane proteins, only the two membrane glycoprotein fractions, 94,000 and 78,000 daltons, were retained in the membrane. Although variable traces of some of the other bands remained, these generally constituted less than 6% of the Coomassie blue-staining proteins. The fact that only two  (19)(20)(21)23). This band has already been shown to contain the anion transport system (38). More re-