Dehydroascorbic acid Transport by GLUT4 in Xenopus Oocytes and Isolated Rat Adipocytes.

Dehydroascorbic acid (DHA), the first stable oxidation product of vitamin C, was transported by GLUT1 and GLUT3 in Xenopus laevis oocytes with transport rates similar to that of 2-deoxyglucose (2-DG), but due to inherent difficulties with GLUT4 expression in oocytes it was uncertain whether GLUT4 transported DHA (J. Biol. Chem. 272:18982-89, 1997). We therefore studied DHA and 2-DG transport in rat adipocytes, which express GLUT4. Without insulin, rat adipocytes transported 2-DG 2-3 fold faster than DHA. Pre-incubation with insulin 0.66 µ M increased transport of each substrate similarly: 7-10-fold for 2-DG and 6-8-fold for DHA. Because intracellular reduction of DHA in adipocytes was complete before and after insulin stimulation, increased transport of DHA was not explained by increased internal reduction of DHA to ascorbate. To determine apparent transport kinetics of GLUT4 for DHA, GLUT4 expression in Xenopus oocytes was re-examined. Pre-incubation of oocytes for > 4 hours with insulin 1 µ M augmented GLUT4 transport of 2-DG and DHA by up to 5-fold. Transport of both substrates was inhibited by cytochalasin B and displayed saturable kinetics. GLUT4 had a higher apparent transport affinity (K M of 0.98 vs. 5.2 mM) and lower maximal transport rate (V MAX of 66 vs. 880 pmole/oocyte/10 min) for DHA compared to 2-DG. The lower transport rate for DHA could not be explained by binding differences at the outer membrane face, as shown by inhibition with ethylidene glucose, or by transporter trans-activation, and therefore was likely due to substrate specific differences in transporter/substrate translocation or release. These novel data indicate that the insulin sensitive transporter GLUT4 transports DHA in both rat adipocytes and Xenopus oocytes. Alterations of this mechanism in diabetes could have clinical implications for ascorbate utilization.


INTRODUCTION
Cellular accumulation of vitamin C (ascorbic acid, ascorbate) is due to transport of both ascorbate and its oxidized metabolite, dehydroascorbic acid (DHA) (1;2). Although ascorbate is the predominant if not the only form in blood, it is possible that DHA is produced in the extracellular milieu in vivo during oxidative stress (3)(4)(5). Experiments in neutrophils demonstrated that the rate of cellular DHA uptake is as much as 30-fold greater than the rate of ascorbate uptake (1;5). Once transported, DHA is immediately reduced intracellularly to ascorbate. DHA uptake followed by intracellular reduction can increase intracellular ascorbate accumulation 5-20-fold within minutes. This process, termed ascorbate recycling, was first demonstrated experimentally in human neutrophils (1;4;6), but may also occur in other cell types if DHA is present extracellularly.
We demonstrated previously that DHA is efficiently transported by glucose transporter isoforms GLUT1 and GLUT3 expressed in Xenopus laevis oocytes, with rates of transport and affinity equal to or greater than that for glucose (7). Isoforms GLUT2, GLUT5, and SGLT1 did not transport DHA, and no glucose transporter isoform transported ascorbic acid (7;8). GLUT4 demonstrated minimal DHA transport activity, and it was not possible to adequately examine DHA transport by this isoform.
GLUT4 is the insulin-sensitive protein responsible for the majority of glucose transport in muscle and adipose tissues. Upon insulin stimulation, intracellular vesicles containing GLUT4 fuse with plasma membranes, greatly increasing glucose transport (9)(10)(11)(12)(13). Ineffective recruitment of GLUT4 to the cell surface results in excessive glucose accumulation in the blood and type II diabetes clinically. In addition to aberrant glucose metabolism, some evidence suggests that diabetics may also have decreased cellular concentrations of ascorbate or increased requirements for it (14)(15)(16)(17). Considering the importance of GLUT4 in glucose uptake and its possible by guest on  http://www.jbc.org/ Downloaded from radiomatic detection confirmed 100% conversion of ascorbate to DHA, which could be completely recovered upon reduction with 2,3 dimercapto-1-propanol.

Isolation
Isolation of rat adipose cells was performed as previously described (21;22). Briefly, epididymal fat pads were removed from anesthetized Sprague Dawley rats (Charles River Laboratories) and minced in Krebs-Ringer bicarbonate HEPES buffer, 1% BSA without glucose, containing 200nM adenosine. Type I collagenase (Worthington Biochemical) was added for 50-60min and digested tissue was filtered through a 250 micron nylon screen mesh into a 50 ml conical centrifuge tube. Cells were resuspended and washed 5 times with the Krebs-Ringer buffer described. Cell number was quantified using lipid weight determinations.

Transport protocol
Rat adipocytes were diluted in the Krebs-Ringer buffer described above to 3-4 x 10 Oocyte intracellular ascorbate was measured as described (7).

Plasmids and inserts
Rat Glut1 and human Glut3, 4 were obtained as plasmid constructs from G. I. Bell and C.F. Burant (University of Chicago, Chicago, IL). Rat GLUT4 was obtained from M. Birnbaum (University of Pennsylvania, Phila, PA). Plasmid constructs were described previously(7;24-28), and mRNA was prepared in vitro by cutting plasmid vectors with appropriate restriction enzymes followed by in vitro transcription utilizing SP6 or T3 mMessage, mMachine (Ambion). Oocytes were isolated from Xenopus laevis (Xenopus I, Dexter MI) and injected with mRNA using established methods (29) as described (7). Injection volume and mRNA concentration was 30-50 nl at a concentration of 1mg/ml unless specified. After injection oocytes were maintained at 20 o C in OR-2 medium containing 1mM pyruvate (Sigma) for up to 5 days and OR-2 was changed daily. Experiments were performed 3-5 days after mRNA injection. In some experiments oocytes were incubated in the presence of insulin 1 µM for up to 24 hours prior to transport studies.

Oocyte plasma membrane isolation and immunoblot analysis
Oocytes plasma membranes were prepared using established protocols (30 (20). Equal amounts of solubilized plasma membrane complexes were analyzed on SDS/4-20%/polyacrylamide gels.
Proteins were transferred to a PVDF membrane and incubated with polyclonal GLUT4 antisera (Alpha Diagnostic International San Antonio TX). Detection was by colorimetric measurement of alkaline phosphatase activity. Densitometric analyses of colorimetric signals were performed using a flatbed scanner (Hewlett Packard ScanJet 5100C) and Scion Image analysis software (Scion corporation, Fredrick MD).

Statistics and Kinetic Calculations
Data are expressed as the arithmetic mean + the standard deviation (SD) of 10-20 oocytes at each data point, unless otherwise indicated. SD is not displayed when smaller than the symbol size. Transport kinetics were analyzed by best-fit analysis of data points utilizing curve fitting (Jandel Scientific, San Rafael, CA) and Eadie-Hofstee transformation which gave comparable results.

DHA and 2-DG transport in adipocytes
DHA transport by GLUT4 was first studied in isolated rat adipocytes, cells that express GLUT4. In adipocytes without insulin the ratio of GLUT4 to GLUT1 in plasma membrane is approximately 2:1. Upon insulin stimulation this ratio changes to approximately 10:1 (9). Because GLUT4 is disproportionately up-regulated by insulin, adipocytes are a preferred cell system to study GLUT4-mediated transport. If DHA were transported by GLUT4, DHA transport would be insulin stimulated. Adipocytes, pre-incubated with or without insulin, were incubated with either

DHA and 2DG transport in Xenopus laevis oocytes
Xenopus oocytes present an ideal means to specifically examine DHA transport by GLUT 4. In oocytes adequate membrane expression of GLUT4 can be problematic due to its poor insertion in the plasma membrane (19;20). For this reason and because previously reported DHA transport was very low in GLUT4-expressing oocytes (7), it was necessary to increase transport rates to study DHA transport kinetics. Oocytes and mammalian cells might have similar insulinresponsive intracellular pathways for translocation of transmembrane proteins (31). Therefore, we tested whether insulin enhanced DHA and 2-DG transport in GLUT4-expressing oocytes.
These oocytes were incubated from 1-24 hours with insulin1 µM, washed, and incubated with either radiolabelled DHA or 2-DG ( Fig. 3A, 3B). Both DHA and 2-DG transport increased up to 5-fold after 4 hours of insulin pre-incubation. Transport of both substrates was completely inhibited by 10 µM cytochalasin B (data not shown). Control water-injected oocytes showed very little transport and were unresponsive to insulin pre-incubation. Pre-incubation with 1µM insulin produced the maximum effect on both DHA and 2-DG transport (Fig. 3C, 3D). The effect of insulin was specific for GLUT4, because oocytes expressing either GLUT1 or SGLT1 showed no increased transport in response to insulin (Fig. 3E).
We tested whether the effect of insulin on increasing DHA and 2-DG transport was due to increased GLUT4 translocation to the oocyte plasma membrane. Oocytes previously injected with or without GLUT4 were incubated for 4 hours with insulin, and GLUT4 in isolated oocyte plasma membranes was quantitated by Western blot (Fig. 4). Insulin increased GLUT4 in the membrane approximately 5 fold, based on densitometric analyses, in agreement with the transport findings.
Based on these results, DHA transport experiments in GLUT4-expressing oocytes were performed after a 4-hour pre-incubation with insulin 1 µM except as indicated. mRNA injection amount and post-injection incubation time influenced DHA and 2-DG transport. Transport of both substrates increased linearly with injected amounts of mRNA from 0.3-40 ng (data not shown), and 30 ng of mRNA was used for injections. Transport increased linearly with substrate incubation time from days 2 to 5 after mRNA injection (data not shown). Experiments were performed using oocytes 3-5 days post-injection, because oocyte fragility increased after day 5.

Transport kinetics and transport properties of DHA and 2DG in GLUT4-expressing oocytes
To determine kinetics injected oocytes were pre-incubated with insulin, washed, and Because of the difference in maximal transport rates between DHA and 2-DG in GLUT4expressing oocytes, uptake rates of DHA and 2-DG were compared to those in oocytes expressing GLUT1 or GLUT3. At comparable extracellular concentrations, 2-DG uptake was 6-12-fold greater than that of DHA in GLUT4-expressing oocytes ( Fig. 6A and 6B). In contrast, oocytes expressing GLUT1 or GLUT3 transported both substrates equally (Fig. 6B). Maximal transport rates of DHA for GLUT1-and GLUT3-expressing oocytes were approximately 100 pmole/oocyte/min compared to a maximal rate of 6.6 pmole/oocytes/min for GLUT4-expressing oocytes. The apparent transport affinity of GLUT4 for DHA was similar to that of GLUT1 and GLUT3 (7) and higher than that for 2-DG (lower KM ), but the rate of GLUT4 mediated transport was at least 6-fold lower for DHA compared with 2-DG. The maximal transport rate differences Uptake rates of both substrates were similar regardless of 3-OMG loading (Fig. 7).  (Fig. 8). These results suggest that DHA and 2DG binding to the external face of GLUT4 are comparable.
In summary, GLUT4-expressing oocytes transported DHA with an apparent affinity higher than that of 2-DG, but with a 6-12 fold lower maximal transport rate. This could not be attributed to intracellular reduction of DHA, substrate trans-activation, or decreased DHA binding at the extracellular surface. Therefore, the lower transport rate is likely due to differences between DHA and 2-DG in either transmembrane translocation or intracellular release.

DISCUSSION
This study presents new information showing that GLUT4 transported dehydroascorbic acid, the oxidized form of ascorbate. Transport was initially examined using isolated rat adipocytes because of their abundance of GLUT4. Rat adipocytes express GLUT1 as well as GLUT4, and in response to insulin both are translocated to the cell surface. Upon insulin stimulation, GLUT4 concentration in plasma membrane is approximately 10 fold greater than GLUT1, and increased glucose transport after insulin stimulation can be attributed predominantly to GLUT4(9). We report here that DHA and 2-DG were transported by rat adipocytes without insulin, insulin pre-incubation increased uptake of both DHA and 2-DG 6-8 fold, and the rate of DHA transport was approximately 3 fold less than that of 2-DG independent of insulin. These novel data support the conclusion that DHA transport in adipocytes was mediated by GLUT4.
To validate that GLUT4 could transport DHA, GLUT4 was expressed in Xenopus oocytes. Advantages of oocytes are that specific GLUT transporters can be tested individually, endogenous glucose and DHA transport is nearly undetectable in control oocytes, and the human GLUT4 construct could be tested. These experiments showed conclusively for the first time that GLUT4 transported DHA, that the affinity of GLUT4 for DHA was similar to that of GLUT 1 and GLUT3, and that the affinity of GLUT4 for DHA was higher than for 2-DG. We showed that extracellular binding of DHA and 2-DG were the same by performing inhibition experiments with ethylidene glucose, a glucose analog not translocated by glucose transporters.
Because the intracellular side of the transport protein may also be available to bind substrate, we examined whether the differences in transport might be attributed to a trans-activation effect.
Once inside oocytes, DHA is rapidly reduced to ascorbate, which does not interact with glucose transport proteins (7). Therefore, only 2-DG could possibly interact with intracellular binding sites. Using oocytes pre-loaded with 20 mM 3-OMG, we demonstrated that intracellular sugar had no effect GLUT4 transport of either 2-DG or DHA. These data are consistent with previous reports that GLUT4, unlike GLUT1, does not demonstrate mechanisms of transactivation (32;33). Taken together, the findings suggest that the different rate of uptake of DHA compared to 2-DG is likely due differences in membrane translocation and/or intracellular substrate release.
To clarify this issue, additional biochemical tools are needed that are not currently available, such as a non-reducible DHA analog or a high affinity ligand that inhibits DHA but not 2-DG transport.
DHA transport by GLUT4 may be a novel aspect of ascorbate recycling, in which DHA formed by oxidation of extracellular ascorbate is rapidly transported into cells and immediately reduced to ascorbate intracellularly. Ascorbate recycling allows cells to accumulate ascorbate rapidly (4), and may be a new mechanism for ascorbate accumulation in GLUT 4 expressing cells such as adipocytes, skeletal muscle cells, or cardiac myocytes. Because GLUT 4 transports DHA as shown in this paper, GLUT1 and GLUT3 transport DHA (7), and nearly all cells possess at least one of these transporters, we predict that most cells will transport DHA if available.
The contribution of GLUT 4 -mediated transport of DHA to ascorbate accumulation in vivo is unknown. Plasma concentrations of DHA in healthy people cannot be distinguished from zero (3). DHA was detected in plasma of diabetic patients, but this may be due to assay artifact(2;42;43). DHA plasma concentrations, however, may not indicate substrate availability to tissues. Cellular transport may be a consequence of local formation of DHA extracellularly, which would not be reflected by plasma measurements (2). Local oxidation of ascorbate to DHA is followed by its rapid uptake into neutrophils(4). It is unknown whether similar local oxidation occurs in the extracellular milieu of tissues that express GLUT 4. Transport of ascorbate itself may be responsible for ascorbate accumulation under conditions without oxidation. With oxidative stress, DHA uptake and subsequent intracellular reduction might contribute to ascorbate accumulation(1;2).
Type II diabetes is characterized by insulin insensitivity, decreased GLUT4 translocation to the cell membrane, decreased glucose transport, and accumulation of glucose in plasma.
Although unresolved, a number of investigators have suggested that ascorbate requirements of diabetics may be increased (14)(15)(16)(17). The data here provide several potential mechanisms by which this may occur. DHA uptake followed by internal reduction could be one mechanism accounting for ascorbate accumulation in GLUT4 expressing tissues. If in diabetics DHA uptake is diminished due to decreased GLUT4 at the cell membrane, ascorbate intracellular concentrations might decline. It is also possible that DHA clearance by GLUT1, GLUT3, or GLUT4 may decline due to high extracellular glucose concentrations, and DHA residence time outside cells may increase. Because DHA is labile, irreversible extracellular degradation of DHA may occur and ascorbate reducing equivalents would be lost. Another possibility is that insulininduced increases in GLUT1 at the plasma membrane may partially mediate DHA uptake. In oocytes, the rate of GLUT1-mediated DHA transport is approximately 20 fold faster than that