Genistein is a natural inhibitor of hexose and dehydroascorbic acid transport through the glucose transporter, GLUT1.

Genistein is a dietary-derived plant product that inhibits the activity of protein-tyrosine kinases. We show here that it is a potent inhibitor of the mammalian facilitative hexose transporter GLUT1. In human HL-60 cells, which express GLUT1, genistein inhibited the transport of dehydroascorbic acid, deoxyglucose, and methylglucose in a dose-dependent manner. Transport was not affected by daidzein, an inactive genistein analog that does not inhibit protein-tyrosine kinase activity, or by the general protein kinase inhibitor staurosporine. Genistein inhibited the uptake of deoxyglucose and dehydroascorbic acid in Chinese hamster ovary (CHO) cells overexpressing GLUT1 in a similar dose-dependent manner. Genistein also inhibited the uptake of deoxyglucose in human erythrocytes indicating that its effect on glucose transporter function is cell-independent. The inhibitory action of genistein on transport was instantaneous, with no additional effect observed in cells preincubated with it for various periods of time. Genistein did not alter the uptake of leucine by HL-60 cells, indicating that its inhibitory effect was specific for the glucose transporters. The inhibitory effect of genistein was of the competitive type, with a K of approximately 12 μM for inhibition of the transport of both methylglucose and deoxyglucose. Binding studies showed that genistein inhibited glucose-displaceable binding of cytochalasin B to GLUT1 in erythrocyte ghosts in a competitive manner, with a K of 7 μM. These data indicate that genistein inhibits the transport of dehydroascorbic acid and hexoses by directly interacting with the hexose transporter GLUT1 and interfering with its transport activity, rather than as a consequence of its known ability to inhibit protein-tyrosine kinases. These observations indicate that some of the many effects of genistein on cellular physiology may be related to its ability to disrupt the normal cellular flux of substrates through GLUT1, a hexose transporter universally expressed in cells, and is responsible for the basal uptake of glucose.

In mammalian cells, there are two discrete hexose transport systems whose function is to provide the cells with the basic cellular fuel, glucose. These two systems are the family of facilitative hexose transporters (glucose transporters, GLUTs) (1, 2) that transport glucose down a concentration gradient and are widely expressed in all cells and tissues, and the sodiumglucose cotransporters that transport glucose against a concentration gradient and are mainly expressed in small intestine and kidney (3). Six genes encoding different glucose transporter isoforms have been molecularly cloned (4 -9), and more than one isoform is usually expressed in a given cell or tissue. Five isoforms, GLUT1-GLUT5 are expressed on the cell membrane (4 -8) and one, GLUT7, is restricted in expression to the internal membranes of the endoplasmic reticulum (9).
Glucose transporters of the facilitative type are the universal transporters of glucose in mammalian cells (1,2). One isoform, GLUT5, is not a glucose transporter but instead is involved in the transport of fructose in specialized tissues and cells such as spermatozoa (10). GLUT2, the isoform abundantly expressed in liver, has the ability to transport both glucose and fructose but with different affinities (11). Evidence has accumulated indicating that the glucose transporters participate in the cellular accumulation of substrates other than glucose (12)(13)(14)(15)(16)(17). We demonstrated that the glucose transporters GLUT1, GLUT2, and GLUT4 are efficient transporters of the oxidized form of vitamin C, dehydroascorbic acid (15). We found that glucose transporters are also the main pathway mediating the transport of dehydroascorbic acid in normal human neutrophils (15) and HL-60 myeloid leukemia cells (16,17). The data suggest that the facilitative glucose transporters may be the universal transporters of both glucose and dehydroascorbic acid in mammalian cells.
Many cellular functions are regulated at the levels of gene expression and of protein function by discrete phosphorylationdephosphorylation events. Tyrosine phosphorylation plays a major role in the transduction of cellular signals induced by the binding of insulin to its cognate receptor (18), and the effect of insulin on GLUT4 is an important point of regulation for glucose transport in mammalian cells. Insulin induces the translocation of GLUT4 from intracellular membrane pools to the cell surface (19,20), and there is also evidence suggesting that insulin may modulate the intrinsic functional activity of the glucose transporters (21). The role of phosphorylation in regulating the activity of the glucose transporters is, however, controversial (22)(23)(24)(25)(26) The isoflavone genistein (4Ј,5,7-trihydroxyisoflavone) is a dietary-derived natural product (27) present in a variety of plant foods that selectively inhibit the activity of protein-tyrosine kinases, as opposed to protein-serine/threonine kinases (28). Genistein affects oncogene-induced tumorigenesis (29), cell proliferation (30 -32), cell differentiation (33,34), angiogenesis (35), and signal transduction mechanisms activated by growth factors (36 -39). The multiple effects of genistein in cellular systems are presumed to reflect its ability to inhibit the activity of protein-tyrosine kinases.
We show here that genistein is a potent inhibitor of the functional activity of the glucose transporters present in human myeloid HL-60 cells, Chinese hamster ovary (CHO) 1 cells overexpressing the glucose transporter GLUT1, and human erythrocytes. The characteristics and the specificity of the inhibition, and the results of studies showing that genistein affects the glucose-displaceable binding of cytochalasin B to GLUT1 in erythrocyte membranes, indicate that the effect of genistein is related to its direct interaction with GLUT1. These results emphasize the ability of GLUT1 to interact with molecules structurally unrelated to glucose and have important implications for our understanding of the effects of genistein on cellular physiology in normal and malignant cells.

EXPERIMENTAL PROCEDURES
Cell Culture-HL-60 cells were cultured in Iscove's-modified Dulbecco's medium supplemented with 10% fetal bovine serum and antibiotics. Cell viability was greater than 95%, as determined by trypan blue exclusion. Chinese hamster ovary cells were cultured in Iscove's-modified Dulbecco's medium supplemented with 10% fetal bovine serum and 0.25 mg/ml G418. CHO cells expressing GLUT1 were a gift from Dr. Michael Czech (Program in Molecular Medicine, University of Massachusetts Medical Center) (40). CHO cells expressing the human placental insulin receptor were constructed in the laboratory of the late Dr. Ora M. Rosen at Sloan Kettering Institute (41). Human erythrocytes were purified from outdated blood samples obtained by the Hematology Service of the Regional Hospital in Valdivia.
Uptake Assays-For uptake assays (17), cells were incubated at room temperature in incubation buffer (17)  and adequate concentrations of the respective unlabeled compounds for the times indicated in the figures. Solutions of ascorbic acid were prepared daily before use and kept at 4°C in incubation buffer containing 0.1 mM dithiothreitol. For dehydroascorbate uptake experiments, ascorbate oxidase (0.1 to 10 units) was added to the incubation mixture to generate dehydroascorbic acid (16). Uptake was stopped by washing the cells in cold (4°C) stopping solution (phosphate-buffered saline without Ca 2ϩ and Mg 2ϩ ). The cells were solubilized in Tris-HCl, pH 8.0, containing 0.2% sodium dodecyl sulfate (SDS) and processed for liquid scintillation counting. When testing the effects of different competitors and inhibitors on uptake, they were added at the beginning of the experiment from concentrated stock solutions freshly prepared, or, alternatively, the cells were preincubated with them.
Cytochalasin B Binding-Pink erythrocyte ghosts were prepared from washed red cells by hypotonic lysis in 5 mM Na 2 HPO 4 buffer, pH 7.4, containing 2 mM EDTA, followed by four successive centrifugation/ wash cycles in the same buffer, and were resealed in 10 mM Na 2 HPO 4 buffer, pH 7.4 (42). Specific binding of cytochalasin B to functional glucose carriers was estimated from the difference between cytochalasin B bound in the presence of 500 mM L-glucose and 500 mM D-glucose. Equilibrium cytochalasin B binding was initiated by the addition of the membrane preparation (suspended in 10 mM Na 2 HPO 4 buffer, pH 7.4) to the binding assay solution. The latter contains 10 mM Na 2 HPO 4 buffer, pH 7.4, cytochalasin E, [4-3 H]cytochalasin B, and D-or Lglucose. The total volume of the reaction mixture was 150 l. The final composition was 0.06 to 0.1 mg/ml erythrocyte membrane protein (equivalent to 1-1.6 ϫ 10 8 cells), 10 M cytochalasin E, 500 mM D-or L-glucose, 0.02-0.04 Ci of [4-3 H]cytochalasin B (11.9 Ci/mmol, DuPont NEN), and cold cytochalasin B for final concentrations of 0.01 to 5 M. The mixture was incubated at room temperature for 10 min before collecting the membranes by centrifugation at 15,000 ϫ g for 10 min. The amount of specifically bound cytochalasin B was estimated by determining the amount of radioactive ligand associated with the membrane pellet and from the difference in the amount of soluble radioactivity before and after centrifugation. Both determinations gave similar results.

Effect of Genistein in HL-60
Cells-Genistein inhibited the uptake of deoxyglucose and dehydroascorbic acid by HL-60 cells in a dose-dependent manner, with 50% inhibition observed at 12-15 M genistein (Fig. 1A). 100 M genistein completely blocked the uptake of dehydroascorbic acid and deoxyglucose. The HL-60 cells express the glucose transporter GLUT1, and the functional characteristics of this transporter are similar to those described for GLUT1 present in other cell types, including its ability to transport dehydroascorbic acid and D-but not L-hexoses, and its sensitivity to inhibition by cytochalasin B but not cytochalasin E (16,17).
The above assays were carried out using cells preincubated with genistein for 30 min, a typical protocol used in studies analyzing the effect of genistein on the activity of cellular protein-tyrosine kinases. As uptake under these conditions is a complex function of transport and intracellular trapping of the transported substrate, genistein could be inhibiting uptake at either of these steps. Total inhibition of deoxyglucose (Fig. 1B) and dehydroascorbic acid (data not shown) transport was observed in cells preincubated with 100 M genistein from 0 to 60 min, indicating that the full inhibitory effect of genistein on uptake develops instantaneously and no preincubation is required.
Genistein affects multiple cellular activities as a result of its ability to inhibit the activity of protein-tyrosine kinases (29 -39). The observed inhibition of deoxyglucose and dehydroascorbate uptake by genistein therefore could be linked to the inhibition of phosphorylation events involved in the cellular uptake of these compounds. While protein phosphorylation can modulate the activity of GLUT4 (23)(24)(25)(26)43), there is no evidence indicating that GLUT1 is phosphorylated on tyrosine residues. Therefore, we hypothesized that the inhibitory effects of genistein on transport reflect a direct action of genistein on the glucose transporters rather than inhibition of cellular proteintyrosine kinases. Support for this hypothesis was obtained from experiments analyzing the effect of genistein under experimental conditions 1 The abbreviation used is: CHO, Chinese hamster ovary. tailored to measure transport as distinct to accumulation (17). We determined the dose dependence of genistein's effect on the transport of deoxyglucose, dehydroascorbic acid, methylglucose, and leucine in HL-60 cells using a 30-s uptake assay. Methylglucose is not trapped or metabolized, and, therefore, its kinetics of transport can define the time frame at which transport is still occurring under initial velocity conditions. We measured the transport of leucine to control for the specificity of the effect of genistein on the activity of the glucose transporters, as opposed to possible nonspecific effects on membrane transport. Leucine is transported into cells by transport systems functionally unrelated to the glucose transporters (44). The initial rate of transport of dehydroascorbic acid, deoxyglucose, methylglucose, and leucine by the HL-60 cells was linear for the first 60 s of uptake (data not shown), validating the assay for the determination of the kinetic constants of transport. When genistein was added at the beginning of the uptake assays, it inhibited the transport of dehydroascorbic acid, deoxyglucose, and methylglucose, with 50% inhibition observed at 10 -15 M genistein (Fig. 2, A-C). Complete inhibition of transport was observed at 100 M genistein, but some transport was observed at concentrations of genistein greater than 100 M. No effect of genistein on the transport of leucine by the HL-60 cells was observed in these studies, with less than 20% inhibition at 100 M genistein (Fig. 2D), indicating the speci-ficity of the effect of genistein on the activity of the glucose transporters and militating against the possibility that it could be caused by a general biological effect on the cell membrane. The specificity of the effect of genistein was further confirmed in experiments which showed that daidzein (4Ј,7-dihydroxyisoflavone), an analog of genistein that lacks a hydroxyl group at position 4 (Fig. 2, E and F) did not affect the transport of dehydroascorbic acid, deoxyglucose, methylglucose, or leucine in the HL-60 cells (Fig. 2, A-D). Similarly, staurosporine, a general inhibitor of protein phosphorylation, did not affect the transport of dehydroascorbic acid, deoxyglucose, methylglucose, or leucine (data not shown) at concentrations (1-10 M) known to completely block protein phosphorylation in HL-60 cells (45). Genistein competitively inhibited the transport of both deoxyglucose and methylglucose in HL-60 cells with a K i of approximately 15 M (Fig. 3, A-D). These observations indicate a direct interaction of genistein with the glucose transporter expressed by HL-60 cells.
Effect of Genistein in CHO Cells-We next analyzed the effect of genistein on the uptake of dehydroascorbic acid and deoxyglucose in CHO cells transfected with a plasmid carrying the cDNA for GLUT1. Immunoblotting experiments using anti-GLUT1 antibodies revealed that the transfected cells expressed an increased number of glucose transporters as compared to control cells stably transfected with a plasmid carrying the cDNA for the human placental insulin receptor (data not shown). CHO cells overexpressing GLUT1 had an increased capacity to take up deoxyglucose and dehydroascorbic acid as compared to the control cells, and uptake was linear for at least 10 min (Fig. 4, A and B). Genistein caused a dose-dependent inhibition of uptake and, at 100 M, completely inhibited uptake of both dehydroascorbic acid and deoxyglucose in the GLUT1-expressing as well as the control cells (Fig. 4, C  and D). In both cell lines, 50% inhibition of uptake was observed at approximately 12 M genistein. These results show that the effect of genistein on the activity of GLUT1 is independent of the cell context in which the transporters are expressed. The simplest interpretation of the above results is that genistein interferes with the transport of dehydroascorbic acid, deoxyglucose, and methylglucose by directly interacting with GLUT1.
Effect of Genistein in Human Erythrocytes-Genistein inhibited the uptake of deoxyglucose in human erythrocytes in a dose-dependent manner, with 50% inhibition observed at approximately 15 M genistein (Fig. 5A). Kinetic analyses indicated that genistein competitively inhibited the uptake of deoxyglucose with a K i of about 12 M (Fig. 5, B and C). The data, therefore, are consistent with the direct interaction of genistein with the erythrocyte glucose transporter. We further tested this hypothesis by analyzing the effect of genistein on the binding of radiolabeled cytochalasin B to the hexose transporters present in purified human erythrocyte ghosts. Affinity labeling experiments show that cytochalasin B binds covalently to the glucose transporter in a D-glucose-displaceable manner (46,47). We reasoned that if genistein and cytochalasin B compete for the glucose binding or transport sites in the transporter, they may also compete with each other. The pink ghosts used in these experiments bound 0.49 -0.55 nmol of cytochalasin B per mg of protein in a D-glucose-sensitive fashion, with a dissociation constant of 0.25 M (data not shown). From these data, we calculated that there are 1.8 -2.1 ϫ 10 5 cytochalasin B binding sites related to the glucose transporter per cell, based on estimating the amount of membrane protein per cell at 0.61 pg. These sites are competed by high concentrations of D-but not L-glucose. The calculated number of sites per cell agrees with previous estimates of 2-3 ϫ 10 5 sites/cell derived from cytochalasin B binding studies (48) and chemical labeling of the GLUT1 protein (49). Increasing concentrations of genistein efficiently competed for the glucose-sensitive cytochalasin B binding sites present in the erythrocyte membranes (Fig. 6A). Approximately 15 M genistein inhibited the binding of 0.1 M cytochalasin B by 50%, while total inhibition of binding was observed at 100 M genistein. A small increase in binding was observed at concentrations of genistein greater than 100 M, a result similar to the increase in transport observed in HL-60 cells treated with similar concentrations of genistein ( Fig. 6A and Fig. 2, B and C). Genistein was a linear competitive inhibitor of the binding of cytochalasin B to the erythrocyte ghosts (Fig. 6B). A secondary plot of the apparent extent of cytochalasin B binding versus genistein concentration was linear and gave an inhibition constant for genistein of approximately 8.0 M (Fig. 6C). The above observations are consistent with the concept that genistein interacts directly with the glucose transporters. DISCUSSION We show here that genistein is a potent inhibitor of the cellular uptake of dehydroascorbic acid, deoxyglucose, and methylglucose, all substrates that enter cells through hexose transporters of the facilitative type. We used as experimental systems HL-60 cells and human erythrocytes, cells that express the facilitative hexose transporter GLUT1. Additionally, we used stably transfected CHO cells expressing GLUT1. It is possible that genistein could be inhibiting the activity of protein-tyrosine kinases whose phosphorylating activity is fundamental in maintaining the activity of GLUT1. There is, however, no available evidence indicating that GLUT1 is phosphorylated on tyrosine residues. Our data are compatible with the concept that genistein inhibits the cellular uptake of hexoses and dehydroascorbic acid by a mechanism unrelated to its ability to inhibit the activity of protein-tyrosine kinases. Rather, genistein interacts directly with GLUT1 and competitively inhibits the transport of hexoses and dehydroascorbic acid across the cell membrane.
The data from time course experiments were consistent with the hypothesis that the effect of genistein was not mediated through inhibition of protein-tyrosine kinases. No genistein preincubation step was necessary to observe its effect on the uptake of methylglucose, deoxyglucose, and dehydroascorbic acid. The effect of genistein was instantaneous and maximum when genistein was added to the uptake assay simultaneously with the test substrate at time 0. On the other hand, the available evidence indicates that the effect of genistein on protein tyrosine phosphorylation is a time-dependent phenomenon that in some cases requires long incubation times to fully develop (28,34,36,37,39). Thus, the instantaneous effect of genistein on uptake argues against the involvement of inhibition of protein tyrosine phosphorylation in the mechanism of action of genistein. Further support for the lack of involvement of protein-tyrosine kinases in the effect of genistein on transport was provided by data from the uptake experiments in human erythrocytes. The similar dose-effect curves for the effect of genistein on uptake in the HL-60 cells, transfected CHO cells, and human erythrocytes indicated that the effect of genistein was cell-independent and likely occurred through the same mechanism in the three cell types analyzed. Although HL-60 and CHO cells are sensitive to the effect of tyrosine kinase inhibitors, human erythrocytes do not express tyrosine kinase activities inhibited by genistein.
Our data indicate that genistein affects the facilitated transport of hexoses and dehydroascorbic acid as distinct from the trapping/accumulation of the transported substrates. The similar dose-response curves for inhibition of uptake in both short uptake studies measuring transport and in extended uptake assays measuring accumulation of the transported substrate suggest that genistein is inhibiting uptake at the common transport step previous to the intracellular trapping of the transported substrates. Trapping of deoxyglucose depends on its intracellular phosphorylation to deoxyglucose 6-phosphate which is not further metabolized and cannot leave the cell because is not transported by GLUT1 (1,2). On the other hand, trapping of dehydroascorbic acid depends on its reduction to ascorbic acid that is not a substrate for GLUT1 and therefore accumulates intracellularly (15,16). These are two very different biochemical processes that depend on different enzyme activities unlikely to be similarly affected by genistein. The effect of genistein on uptake of the nonmetabolizable methylglucose is confirmative. Thus, our data indicate a direct action of genistein on transport as opposed to accumulation.
The data showing that genistein inhibited the transport of methylglucose and deoxyglucose in HL-60 cells and deoxyglucose in human erythrocytes in a competitive manner strongly support the concept that genistein exerts its effect on transport by directly interacting with GLUT1. This notion is consistent with the binding data indicating that genistein blocked the glucose-sensitive binding of cytochalasin B to GLUT1 present in human erythrocytes. Although the precise nature of the interaction between genistein and GLUT1 cannot be deduced from these experiments, the data indicate that genistein interacts with sites in the transporter involved in the binding or transport of dehydroascorbic acid and hexoses and that this interaction is also responsible for interfering with the binding of cytochalasin B.
The multiple effects of genistein on cell physiology are believed to represent its ability to inhibit protein tyrosine phosphorylation. Genistein has been shown, however, to have cellular targets other than the protein-tyrosine kinases such as protein-histidine kinase (50) and DNA topoisomerase II (51). Here, we provide evidence indicating that genistein is a potent inhibitor of the facilitative glucose transporter GLUT1. Our findings have implications for the interpretation of experiments revealing a major effect of genistein on a general cellular function. The literature contains reports of genistein's effects in different cellular systems where the concentrations used are well within the range that causes strong inhibition of the functional activity of the glucose transporters. These include inhibition of endothelial cell proliferation and in vitro angiogenesis (35) and cell cycle and apoptotic events induced in HL-60 and Molt 4 cells (30). Furthermore, genistein inhibits the growth of the human prostatic carcinoma cell lines LNCaP and PC-3 without inhibiting the tyrosine kinase activity of the epidermal growth factor receptor in these cells (52). Genistein also caused apoptosis in thymocytes without altering protein tyrosine phosphorylation (53). Given the presence of glucose transporters in all cells and tissues (1,2), and the importance of these proteins for the provision of cellular nutrients essential for normal cell function, it seems reasonable to consider that some of the effects of genistein on cell proliferation and differentiation may be related to its capacity to inhibit the activity of the glucose transporters. GLUT1 is expressed in all cells and tissues, is especially abundant in erythrocytes and brain, and is responsible for the basal cellular uptake of glucose (1,2).
The finding that genistein inhibits the functional activity of GLUT1 may be particularly relevant to the physiology of cancer cells. It has been known for a long time that one of the primary characteristics of cancer cells is an increased metabolism of glucose (54). Since cancer cells do not accumulate intracellular stores of glucose in the form of glycogen or fat as does liver and adipose tissue, glucose must be obtained continuously from external sources and transported intracellularly. The increased ability of cancer cells to transport glucose is used clinically to locate tumors in patients and to assess their metabolism and response to therapy in a noninvasive manner by positron emission tomography scanning using [ 18 F]fluorodeoxyglucose, a molecule transported by facilitative hexose transporters (55,56). The mechanism whereby cancer cells increase their ability to take up glucose involves the selective overexpression of GLUT1 (57-60), a transporter we show here is inhibited by genistein. In this regard, there is now a growing interest in the possible use of genistein or genistein-containing soy matrices in chemoprevention trials for breast and prostate cancer (61).
The molecular details of the interaction of GLUT1 with genistein, a compound functionally defined by its specific interaction with protein-tyrosine kinases, are unknown. It has been suggested that the glucose transporters possess a nucleotide binding site and experimental evidence has been presented showing that they can be photolabeled with azido-ATP (62) and azidoadenosine (63). Furthermore, there is controversy regarding the ability of ATP and ADP to modulate the functional activity of the glucose transporters (42, 64 -66). Data are available indicating that genistein, while unable to inhibit the insulin receptor kinase activity, markedly decreased basal and insulin-stimulated glucose uptake in rat adipocytes (67). Although the authors did not address the mechanism of action of genistein in their experimental system, they observed that treatment with genistein decreased the labeling of GLUT4 present in the plasma membrane to the same extent that it inhibited glucose uptake without inhibiting the insulin-induced recruitment of GLUT4 to the plasma membrane. These data are consistent with the evidence presented here indicating a direct interaction of genistein with GLUT1 and suggest that genistein has the capacity to interact with other members of the family of facilitative glucose transporters. Our data show that the interaction of genistein with GLUT1 is quite specific, as exemplified by the fact that daidzein, an isoflavone that differs from genistein in lacking one hydroxyl group, did not affect the transport of dehydroascorbic acid, deoxyglucose, or methylglucose in HL-60 cells. Although daidzein does not inhibit the activity of the epidermal growth factor receptor, a receptor whose activity is inhibited by genistein, it inhibits thromboxane A2-stimulated protein tyrosine phosphorylation (28,68). Overall, these observations suggest the existence of a variety of compounds able to interact with the glucose transporters in a highly specific manner that are potentially capable of modulating their functional activity. We show here that genistein directly inhibits the transport function of GLUT1.