Equilibrative sugar uptake in human erythrocytes is characterized by a rapid phase, which equilibrates 66% of the cell water, and by a slow phase, which equilibrates 33% of the cell water. This behavior has been attributed to the preferential transport of β-sugars by erythrocytes (Leitch JM, Carruthers A. Am J Physiol Cell Physiol 292: C974–C986, 2007). The present study tests this hypothesis. The anomer theory requires that the relative compartment sizes of rapid and slow transport phases are determined by the proportions of β- and α-sugar in aqueous solution. This is observed with d-glucose and 3-O-methylglucose but not with 2-deoxy-d-glucose and d-mannose. The anomer hypothesis predicts that the slow transport phase, which represents α-sugar transport, is eliminated when anomerization is accelerated to generate the more rapidly transported β-sugar. Exogenous, intracellular mutarotase accelerates anomerization but has no effect on transport. The anomer hypothesis requires that transport inhibitors inhibit rapid and slow transport phases equally. This is observed with the endofacial site inhibitor cytochalasin B but not with the exofacial site inhibitors maltose or phloretin, which inhibit only the rapid phase. Direct measurement of α- and β-sugar uptake demonstrates that erythrocytes transport α- and β-sugars with equal avidity. These findings refute the hypothesis that erythrocytes preferentially transport β-sugars. We demonstrate that biphasic 3-O-methylglucose equilibrium exchange kinetics refute the simple carrier hypothesis for protein-mediated sugar transport but are compatible with a fixed-site transport mechanism regulated by intracellular ATP and cell shape.
- carrier-mediated transport
- transport kinetics
- transport regulation
the accumulated results of more than 60 years of intensive study reveal important inconsistencies between the experimental behavior of the human erythrocyte glucose transport system and the predictions of models for carrier-mediated solute transport (4, 12, 14, 17, 34, 41, 44, 48, 52). One such inconsistency is multiphasic net and exchange sugar import in which erythrocyte sugar uptake is characterized by the rapid equilibration of 66% of the cell water space followed by the slower equilibration of the remaining cell water (20, 31, 40). The expected result is a single phase of sugar uptake. This and other inconsistencies are not explained by errors in transport determinations or data interpretation (40). Rather, factors extrinsic to the transport process appear to contribute to the operational complexity of erythrocyte sugar transport (14, 52, 53).
Erythrocyte sugar transport is mediated by the glucose transport protein GLUT1 (50) and is directly modulated by cytoplasmic ATP, although the mechanism of GLUT1 regulation is only partially resolved (10). Biphasic sugar transport is present only in ATP-containing red blood cells and is replaced by a single transport process when cytoplasmic ATP is depleted (20, 31, 40). ATP-GLUT1 interactions promote GLUT1 conformational changes (10), and it has been suggested that these changes result in ATP-dependent sugar occlusion or binding within GLUT1 cytoplasmic domains before slow sugar release into cytosol (9, 32). However, demonstrations of multiphasic equilibrium exchange transport at sugar concentrations 1,000-fold greater than red cell GLUT1 concentration are incompatible with this hypothesis (40).
The relative sizes of fast (66% cell volume) and slow (33% cell volume) 3-O-methylglucose (3MG) transport compartments correspond exactly to the equilibrium distribution of β- and α-3MG in aqueous solution (66% β:33% α), suggesting that ATP binding promotes GLUT1 preference for β-sugar (40). Transport specificity for sugar epimers is evident from studies with the glucose epimer galactose (an axial C-4 hydroxyl group), which is transported with 10-fold lower affinity than d-glucose (26). Glucose anomerization is slowed by deuterium oxide (D2O) and acidic pH (35, 47). These agents also modify steady-state sugar transport kinetics (7, 19, 51). GLUT1 specificity for sugar anomers has been studied previously, but with contradictory conclusions. Human erythrocyte β-d-glucose metabolism is faster at physiological sugar concentrations (24, 46), but α-d-glucose utilization is more rapid at subsaturating concentrations (45). Rat erythrocytes have been shown to preferentially utilize α-d-glucose (46) or β-d-glucose (22) as a glycolytic substrate. Faust (23) found that β-d-glucose penetrates human red blood cells nearly three times more rapidly than α-d-glucose. Other studies report no difference in the rate of d-glucose anomer uptake (15), that β-d-glucose is imported 1.5 times faster than α-d-glucose (49), or that α-d-glucose exit is twice as fast as β-d-glucose (39). Competition studies using unlabeled α- or β-d-glucose to inhibit radiolabeled d-glucose zero-trans uptake demonstrate identical affinity for α- and β-d-glucose (15), although we now understand that this assay is not the most discriminating test of anomeric specificity (40). The preference for 1-position fluoroanalogs is impossible to compare directly, because only β-1-deoxy-1-fluoroglucose is transported (43). The apparent inhibition constant [Ki(app)] for β-fluoroanalog isomer inhibition of sorbose import is 15 mM, whereas Ki(app) for the α-isomer is 78 mM (8). α-d-Glucose is 37% more effective than β-d-glucose in promoting GLUT1 conformational changes (3). GLUT1 tryptophan quenching by α- and β-d-glucose indicates that β-d-glucose is favored at higher temperatures, whereas α-d-glucose is preferred at lower temperatures (36).
The literature, therefore, presents multiple, contradictory reports of anomeric specificity in human red cell sugar transport. It becomes necessary, therefore, to test the anomer hypothesis with several independent methods. The present study assumes this challenge and demonstrates that GLUT1 lacks anomeric specificity regardless of intracellular ATP content.
MATERIALS AND METHODS
De-identified human blood was obtained from Biological Specialties. Porcine kidney mutarotase was purchased from Calzyme Laboratories. d-[14C]glucose and 2-deoxy-d-[3H]glucose ([3H]2DG) were purchased from Amersham. 3-O-[3H]methylglucose ([3H]3MG), d-[3H]mannose, and all other reagents were purchased from Sigma Chemicals.
Kaline consisted of 150 mM KCl, 5 mM MgCl2, 5 mM EGTA, and 5 mM HEPES, pH 7.4. Lysis buffer contained 10 mM Tris·HCl and 2 mM EDTA, pH 8.0. Sugar-stop solution comprised ice-cold kaline containing 20 μM cytochalasin B (CCB) and 200 μM phloretin.
Red blood cells.
Red blood cells were isolated by washing whole human blood in 4 or more volumes of ice-cold kaline followed by centrifugation at 10,000 g for 15 min at 4°C. Serum and buffy coat were removed by aspiration, and the wash, centrifugation, and aspiration cycle was repeated until the buffy coat was no longer visible. Cells were resuspended in 4 volumes of sugar-free or sugar-containing kaline and incubated for 1 h at 37°C to deplete or load intracellular sugar.
Red cell ghosts.
Ghosts were formed by reversible hypotonic lysis of red blood cells. Red blood cells were suspended in 10 volumes of ice-cold lysis buffer for 10 min. Membranes were harvested by centrifugation at 27,000 g for 20 min and subjected to repeated wash/centrifugation cycles in ice-cold lysis buffer until the membranes appeared light pink (∼3 cycles). Ghosts were washed with 10 volumes of ice-cold kaline and collected by centrifugation at 27,000 g. Harvested membranes were resealed by incubation in 4 volumes of kaline ± 4 mM ATP (37°C) for 1 h and collected by centrifugation at 27,000 g for 15 min at 4°C. Resealed ghosts were stored on ice until used.
Net sugar uptake.
Sugar-depleted cells or ghosts were incubated in 20 volumes of ice-cold kaline containing 100 μM unlabeled sugar and 0.5 μCi/ml labeled sugar. Uptake proceeded for 6 s to 3 h and was arrested by addition of ice-cold stop buffer followed by centrifugation at 14,000 g for 1 min. The supernatant was aspirated, and pelleted ghosts were washed with 20 volumes of stop buffer and recentrifuged, and the supernatant was aspirated. The ghost pellet was extracted with 500 μl of 3% perchloric acid and centrifuged, and samples of the clear supernatant were counted in duplicate. Zero time points were obtained by adding sugar-stop solution to ghosts, followed by sugar uptake medium. Samples were processed immediately. Radioactivity associated with cells at zero time was subtracted from all nonzero time points. Equilibrium time points were collected using an overnight incubation. All time points were normalized to the equilibrium time point. All solutions and tubes used in the assay were preincubated on ice for 30 min before the start of the experiment. Triplicate samples were processed for each time point.
Sugar equilibrium exchange uptake.
Ghosts were resealed in the presence of 1 or 2.5 mM unlabeled sugar(s) and centrifuged, and the supernatant was aspirated. Ghosts were then incubated in 20 volumes of ice-cold kaline containing the same concentration of unlabeled sugar(s) plus 0.5 μCi/ml radiolabeled sugar. Exchange proceeded for 6 s to 10 h and was stopped by the addition of ice-cold sugar stop. Ghosts were processed as described above with net sugar uptake.
Data were collected in D2O at 24°C using a 400-MHz Oxford NMR.
Sugar anomerization kinetics.
Anomerization was measured by optical rotation using a Rudolph Autopol II polarimeter. Freshly dissolved sugar (10 or 100 mM) in temperature-equilibrated kaline ± 4 mM ATP was placed in a 10-ml, 10-cm path length thermojacketed cell, and optical rotation was measured over time. For mutarotase experiments, an appropriate amount of 1,000 U/ml mutarotase in kaline was placed into the thermojacketed cell and allowed to equilibrate before addition of freshly dissolved 10 mM d-glucose in kaline.
HPLC analysis of sugar anomers.
Sugar anomers were chromatographically separated using a Bio-Rad HPX-87C column as previously described (6). The column was equilibrated on ice with distilled H2O as the mobile phase. Unlabeled sugars were detected using a Waters model 410 differential refractometer, and radiolabeled sugars were detected using a Packard 505TR flow scintillation analyzer.
Red blood cell and ghost electron microscopy.
Scanning electron microscopy of red blood cells and ghosts was carried out as previously described (40).
Data analysis and simulations.
Curve fitting was performed by nonlinear regression using the software package Kaleidagraph (version 4.04; Synergy Software, Reading, PA). Time course simulations and parameter fitting were made by fourth-order Runge-Kutta numerical integration and the method of least squares using the software program Berkeley Madonna (version 8.3.22).
Sugar transport in the presence of intracellular ATP is assumed to follow simple Michaelis-Menten kinetics where net sugar uptake sugar by a fixed site transporter (12) or an alternating carrier (63) is described by (1) where v is unidirectional uptake of α-sugar, v is unidirectional uptake of β-sugar, v is unidirectional exit of α-sugar, and v is unidirectional exit of β-sugar. In expanded form, this is given as (2) where S1 and S2 are intra- and extracellular sugar, respectively, superscipts α and β indicate α- and β-anomers, Kα and Kβ are carrier intrinsic affinity constants for α- and β-anomers, respectively, R is 1/Vmax for zero-trans exit of α-sugar, R is 1/Vmax for zero-trans exit of β-sugar, R is 1/Vmax for zero-trans entry of α-sugar, R is 1/Vmax for zero-trans entry of β-sugar, R is 1/Vmax for exchange transport of α-sugar, R is 1/Vmax for exchange transport of β-sugar, and R= R= 1/Vmax for heteroexchange transport of α- and β-sugar. Because transport is passive, the following must hold true: (3)
In the absence of ATP, differential transport of anomers is lost and Eq. 2 collapses to its normal form: (4) where K and R terms are as described above. Under these conditions, the transporter displays equal affinity and capacity for α- and β-anomers of any given sugar, and each anomer should be transported in proportion to its concentration in solution.
When S1 = S2 (equilibrium exchange) and unidirectional sugar fluxes are measured using radiolabeled sugar (Q), transport in the presence of intracellular ATP is described by Eqs. 5 and 6: (5) (6) where Q1 and Q2 are intra- and extracellular radiolabeled sugar, respectively. When intracellular ATP is absent, differential transport of anomers is lost and Eq. 6 reduces to (7) Under these conditions, the transporter displays equal affinity and capacity for α- and β-anomers of any given sugar, and each anomer should be transported in proportion to its concentration in solution.
Interconversion of α- and β-d-glucose by mutarotation is described by (E8a) (E8b) where Glc is the total d-glucose concentration, α and β are the fractions of sugar that exit as α- and β-anomers, and kα and kβ are first-order rate constants for conversion of α-Glc to β-Glc and β-Glc to α-Glc, respectively. kβ/kα = 2 for d-glucose and 3MG, kβ/kα = 1.22 for [3H]3MG, and kobs = kα + kβ. Mutarotase increases kα and kβ but does not alter the relationships kβ/kα = 2 and kobs = kα + kβ.
Sugar transport via a fixed site transporter displaying geminate exchange between endo- (e1) and exofacial (e2) sugar binding sites and bulk solution sugar or intersite cavity sugar is illustrated in Fig. 6. With G defined as radiolabeled sugar and S as unlabeled sugar, sugar flows (JG or JS) between interstitium (o), e2, intersite cavity (c), e1, and cytoplasm (i) are described by: (E9a) (E9b) (E9c) (E9d) (E9e) (E9f) (E9g) (E9h) where G2 and G1 refer to the concentrations of G bound at sites 1 and 2, respectively, and Gc is the concentration of intersite cavity G. All true second-order rate constants (ko2, kc2, kc1, and ki1) are converted to pseudo first-order constants by multiplying by the GLUT1 concentration. We have simplified the solution by assuming kex = kexo2 = kex2c = kex1c = kex1i and ko2 = kc2 = kc1 = ki1. Ko = k2o/ko2 and Ki = k1i/ki1. For a passive transport system, ko2k2ckc1k1i = ki1k1ckc2k2o.
Compartment sizes do not correlate with equilibrium anomer distributions.
Analysis of equilibrium anomer distributions of aqueous sugars by proton NMR and by HPLC yielded similar results (Table 1). Aqueous d-glucose and 3MG are characterized by equilibrium anomer distributions of ∼66% β-sugar and 33% α-sugar. The 2-position appears to determine anomer distributions, because 2DG (missing the 2-position hydroxyl) is almost equally populated as β- and α-sugar (55 and 45%, respectively), whereas d-mannose (the 2-position epimer of d-glucose) equilibrates as 33% β- and 66% α-sugar. Exchange 3MG, d-glucose, 2DG, or d-mannose transport was examined in the presence and absence of intracellular ATP. Intracellular ATP induced biphasic transport (Fig. 1) of all sugars. Transport compartment sizes, however, did not always correlate with equilibrium anomer distributions of sugars (Table 1). The fast transport compartment was 67% of total transport space for all substrates tested. This is higher than expected for 2DG and d-mannose but is consistent with the equilibrium β-sugar contents of d-glucose and 3MG. Experiment-to-experiment variations in transport compartment sizes made it necessary to simultaneously measure exchange transport of different sugars within the same population of ghosts by dual isotope separation. Simultaneous exchange uptake measurements of 2.5 mM d-glucose and 3MG (Fig. 1A) and of d-glucose and 2DG (Fig. 1B) in ghosts resealed with ATP demonstrated that the transport compartment sizes of d-glucose, 3MG, and 2DG were not significantly different. Thus compartment sizes are not consistently related to relative equilibrium proportions of α- and β-sugars in solution.
Anomerization rate is without effect on the slow phase of transport.
The anomer transport hypothesis predicts that biphasic sugar transport kinetics are observed only when the rate of sugar anomerization is significantly slower than the slowest phase of sugar transport (40). The rate of 3MG anomerization (as measured by circular dichroism, CD) is 10-fold slower than the slow phase of transport (40). d-Glucose, 2DG, and d-mannose do not have measurable CD signals. Anomerization of these sugars was measured by polarimetry. Anomerization is first-order in sugar concentration. The temperature dependence of d-glucose, 3MG, 2DG, and d-mannose anomerization is summarized in Table 2.
d-Glucose and 3MG anomerization at ice temperature were much slower than the slow rate of transport. However, the same was not true for 2DG and d-mannose. Both 2-position glucose analogs displayed accelerated anomerization kinetics relative to d-glucose, but the temperature dependence of anomerization was unchanged. For these sugars, anomerization rates may exceed the slow rate of transport. Nevertheless, these sugars were transported even more slowly than their more slowly anomerizing parent sugar, d-glucose. Thus there is no correlation between the rate of anomerization of the individual hexoses and their slow rate of uptake.
Exogenous porcine kidney mutarotase accelerates α- and β-d-glucose interconversion by as much as 1,000-fold without altering the equilibrium distribution of anomers. Mutarotase is without effect on 3MG anomerization (Table 2). d-Glucose anomerization in the presence of 50 U/ml mutarotase at 20°C is too fast to measure using manual mixing methodologies.
Assuming the anomer specificity hypothesis is correct, our simulations show that resealing a significant mutarotase activity within red cell ghosts will promote monophasic transport even in the presence of ATP (Fig. 2). Extracellular mutarotase should have minimal effect on transport rates and compartment sizes (the fast phase would be a little faster and larger), and transport would remain biphasic. Experimental analysis of d-glucose transport in ghosts lacking and containing 500 U/ml mutarotase ± 4 mM ATP is summarized in Fig. 2A. Predicted transport kinetics are also illustrated. Transport was biphasic in the presence of mutarotase. However, differences would be difficult to observe if mutarotase were active only in the external medium (compare data with curves b and c in Fig. 2A). Our measurements show that ATP did not inhibit mutarotase (see Table 2) and that intracellular mutarotase was indeed active. Mutarotase loaded ghosts contained greater quantities of intracellular β-d-glucose after 5-min exposure to α-d-glucose than did ghosts lacking intracellular mutarotase. This effect was inhibited by the sugar transport inhibitors CCB and phloretin (Fig. 2B, curve c), confirming that glucose must enter the cell to encounter intracellular, catalytically active mutarotase.
Direct measurements of anomer transport.
If the anomer specificity hypothesis is correct, red blood cells should import a disproportional amount of β-sugar and the interstitium should lose more β-sugar during the early phase of sugar uptake. We used HPLC to separate anomers and refractive index change to measure the anomeric composition of cytosolic and interstitial sugar. The method was verified by monitoring 3MG anomerization in vitro. The initial proportion of β-3MG:α-3MG in freshly prepared “α-3MG” solutions was 20:80, and this ratio increased with time at a rate consistent with the rates of anomerization reported in Table 2.
The amount of sugar within ghosts at early times during transport from media containing 2.5 mM total initial 3MG was too small to measure by refractive index, but changes in extracellular anomer composition were measurable. Simulations suggest that the maximal difference between starting and remaining anomer content occurs at 5 min of uptake at 4°C. Our experiments show that the relative anomer content of extracellular medium was unaffected by sugar import into cells lacking or containing ATP (Table 3).
These experiments were repeated using radiolabeled 3MG that, following extraction of cytosol, was chromatographically separated into α- and β-anomers and detected using an in-line radiometric scintillation counter. Radiometric detection of cytosolic 3MG allows direct measurement of [3H]3MG anomers transported into ghosts, greatly increases signal to noise, and permits measurement of unidirectional equilibrium exchange fluxes. We routinely observed that [3H]3MG and unlabeled 3MG had different equilibrium anomer distributions (45:55 α:β vs. 33:66 α:β, respectively; see Table 3). To ensure that the extraction procedure for measuring intracellular sugar did not alter anomer proportions, we subjected medium containing freshly dissolved, unlabeled 3MG to mock cytosol extraction (lysis in ice-cold 3% perchloric acid and neutralization with sodium bicarbonate) without measurable effect on anomer ratios (20:80 β:α). Even at the earliest time points of transport, when the β- to α-anomer ratio of imported radiolabeled sugar is predicted to be greatest, there was no difference in anomeric proportions of intra- and extracellular [3H]3MG (Fig. 3 and Table 4). Thus GLUT1 is able to transport both anomers equally well.
Exchange transport is biphasic in the presence of transport inhibitors.
If biphasic sugar exchange transport reflects fast and slow β- and α-d-glucose transport, transport inhibitors should inhibit both fast and slow phases of transport. CCB (10 μM) inhibited both fast and slow transport phases (3- to 5-fold; Fig. 4A and Table 5). 3MG exchange exit was also biphasic, and both phases were inhibited by 10 μM CCB (exit data not shown). However, phloretin (100 μM) inhibited only the fast phase (5-fold; Fig. 4A).
CCB is an endofacial site erythrocyte glucose transport inhibitor, whereas phloretin is an exofacial site inhibitor (38). We therefore tested the effects of a second exofacial site inhibitor, maltose (13). Extracellular maltose causes significant cell lysis at concentrations (100 mM) required for effective exchange transport inhibition. We therefore repeated transport measurements using 100 mM sucrose as an osmotic control and incorporating tracer [14C]sucrose (a nontransported sugar) into ghosts to adjust for lysis-induced loss of radioisotope during postquenching washing steps. As with phloretin, we observed significant (7-fold) inhibition of the rapid phase of sugar exchange by 100 mM maltose, but the loss of cellular integrity following long incubation in hyperosmotic medium prevented accurate determination of slow-phase kinetics in both sucrose- and maltose-treated cells.
Incubation of human red blood cells with 2,4-dinitrophenol (DNP) caused echinocytosis (Fig. 5A) (59–61). DNP (10 mM for 10 min) abolished biphasic 3MG exchange transport (Fig. 5B) in freshly isolated red blood cells, resulting in exchange kinetics that resemble those of ATP-depleted ghosts. HPLC analysis of cytosolic nucleotides confirms that DNP treatment does not alter the AMP, ADP, or ATP contents of the red blood cell. This result suggests that altered cell shape modifies erythrocyte sugar transport.
Both CCB and cytochalasin D (CCD) modify the cytoskeleton by capping the barbed ends of actin filaments, reducing the steady-state length of actin filaments and inducing the formation of actin nuclei by interacting with monomeric actin with high affinity [Kd(app) ≈ 3 μM (27)]. Unlike CCB, CCD is an extremely low-affinity inhibitor of erythrocyte glucose transport [Ki(app) ≥ 100 μM (Robichaud TK, Henderson PJ, Carruthers A, unpublished observations; Ref. 55)]. CCD was without significant effect on the rapid phase of equilibrium exchange 3MG uptake but inhibited the slow phase with Ki(app) = 4 μM (Fig. 4B and Table 5). Compartment sizes were unaffected by CCD.
The present study evaluates the hypothesis that biphasic, equilibrative sugar transport in human red blood cells results from ATP-dependent differential transport of α- and β-sugars by GLUT1. The results of experiments testing four specific predictions of the hypothesis refute this idea. We conclude that the red cell glucose transport protein GLUT1 transports α- and β-sugars with equal avidity and that biphasic sugar transport results from nucleotide-dependent regulation of the intrinsic catalytic properties of GLUT1.
The differential α- and β-sugar transport hypothesis makes four strict predictions. 1) The relative sizes of cellular transport compartments (fast and slow equilibrative volumes) is related to the relative proportions of α- and β-sugars in solution. 2) Acceleration of the rate of α- and β-sugar interconversion (anomerization) to rates that exceed the slow phase of transport will accelerate or eliminate the slow phase of transport. 3) β-3MG and β-d-glucose are transported more rapidly than α-3MG and α-d-glucose. 4) Glucose transport inhibitors must inhibit both fast transport of β-sugar and slow transport of α-sugar.
Our experiments show the following results. 1) The relative sizes of cellular transport compartments (fast and slow equilibrative volumes) are independent of the relative proportions of α- and β-sugars in solution. 2) Acceleration of the rate of α- and β-sugar interconversion (anomerization) is without effect on the slow phase of transport. 3) α- and β-3MG are transported with equal avidity by red blood cells. 4) Exofacial inhibitors of glucose transport (maltose, phloretin) inhibit the fast phase of sugar transport, an endofacial inhibitor and cytoskeleton disruptor (CCB) inhibits both fast and slow phases of transport, and a cytoskeleton disruptor (CCD) inhibits only the slow phase of transport.
Trivial explanations of biphasic transport kinetics (e.g., 3MG transport into functionally distinct populations of cells; see Ref. 40) are excluded by the simultaneous demonstrations of monophasic uridine transport and biphasic 3MG transport in ATP-loaded red cell ghosts and by the observation that extracellular uridine and 3MG equilibrate with identical total cellular volumes (40). Only two explanations remain for ATP-dependent biphasic sugar uptake: 1) biphasic kinetics reflect the intrinsic properties of the transporter when complexed with ATP (11) or 2) ATP-replete cells contain two cytoplasmic compartments separated by a diffusion barrier, and only the larger compartment is bounded by cell membrane containing functional GLUT1 (40). The smaller compartment (33% of the cell water) must also be bound by cell membrane, but in this case the membrane is either devoid of GLUT1 or contains inhibited GLUT1. The juxtaposition of the small compartment and plasma membrane is necessary to explain ENT1-mediated monophasic and rapid equilibration of erythrocyte total cell water with uridine in the presence of intracellular ATP (40).
Reduced glucose exchange between compartments could result from decreased intracellular diffusion (i.e., increased tortuosity or molecular crowding and increased cytoplasmic viscosity). However, the coefficient for sugar self-diffusion must be reduced by at least 1,000-fold to explain limited intracellular diffusion over distances of 3.5 μm or less (62). This seems untenable given estimates of intracellular tortuosity in neurons, brain, and red blood cells of 1–3 (7, 25, 29) and estimates of intracellular viscosity in red blood cells and Xenopus oocytes of 6–15 cP (16, 28). These increases in tortuosity and viscosity relative to aqueous solution could reduce intracellular glucose diffusion by as much as 20-fold but not by the necessary 1,000-fold.
Diminished glucose exchange between compartments is also consistent with an intracellular permeability barrier separating large and small intracellular compartments. We estimate the upper limit for permeation across this hypothetical barrier is 1 × 10−8 cm/s.1 However, ultrastructural studies fail to observe intracellular compartments in erythrocytes (64). Disruption of red cell membrane detergent-resistant microdomains or “lipid rafts” by methyl-β-cyclodextrin-induced cholesterol depletion is without effect on red cell sugar transport (58), suggesting either that red cell GLUT1 is excluded from membrane rafts or that GLUT1 function is unaffected by nonuniform GLUT1 distribution in the cell membrane. In summary, these considerations indicate that the membrane and cytoplasm compartmentalization hypothesis for biphasic, erythrocyte sugar transport is untenable.
If biphasic transport kinetics result from ATP modulation of the intrinsic catalytic properties of GLUT1, what is the underlying mechanism? Two hypotheses are considered. The alternating carrier model posits that the transporter isomerizes between two states: one presenting an endofacial sugar binding site (e1) to the cytoplasm, and a second presenting an exofacial sugar binding site (e2) to the interstitium (37, 42). Sugar import occurs when extracellular sugar complexes with e2 to form eS2, which then undergoes the eS2 → eS1 conformational change, and complexed sugar dissociates into cytoplasm. After reorientation of e1 → e2, the carrier is ready for another round of sugar import. The fixed-site transporter model hypothesizes that the transporter presents e1 and e2 sites simultaneously and that sugars exchange between these sites via a translocation pore (5, 52). Ligand binding studies support the latter hypothesis (13, 33), whereas analysis of available and homology-modeled crystal structures of major facilitator superfamily (MFS) transporters suggests either multiple, simultaneous substrate binding sites (21, 56) or a single site at any instance (1).
The alternating carrier hypothesis cannot account for ATP-dependent biphasic exchange sugar transport without invoking factors extrinsic to the carrier such as 1) the compartment hypothesis (refuted above) or 2) a time- and ATP-dependent inactivation of the carrier. The latter hypothesis is refuted by demonstrations of biphasic equilibrium exchange transport in freshly made ATP- and 3MG-containing ghosts and in ATP- and 3MG-containing ghosts allowed to rest for 10 h before transport measurement (40). The alternating carrier model also fails in other ways. For example, the model is incompatible with a large body of transport data and may also contravene basic thermodynamic principles (54). Naftalin (54) proposed a modified fixed-site transporter in which the external and internal sugar binding sites of the fixed-site transporter are connected via an internal cavity and that net sugar import involves sugar binding to the exofacial site and dissociation of bound sugar into the internal cavity, whence sugar may reassociate with the exofacial site or associate with the endofacial binding site. Dissociation from the endofacial site into cytoplasm completes the transport cycle (Fig. 6). Accelerated exchange (a hallmark of “carrier-mediated” transport; Ref. 63) is explained when the rate of movement of sugar from endo- or exofacial binding sites into the intersite cavity is accelerated when both binding sites are occupied by sugar or by the presence of intercavity sugar (54). This model builds on the principle of geminate recombination (57) and is consistent with docking studies showing that homology-modeled GLUT1 presents endo- and exofacial sugar binding sites connected by an intersite cavity (21, 56). The model is also compatible with rapid-quench glucose transport studies demonstrating that a transported sugar molecule is biochemically trapped within the GLUT1 translocation pathway when endo- and exofacial binding sites are blocked by inhibitors (9) and that GLUT1-ATP interaction increases GLUT1 sugar occlusion (32). The model is also consistent with the observation that a common membrane protein architecture can mediate both anion channel function and proton-Cl antiport (2) and thus provides a conceptual framework for coevolution of channel and transporter mechanisms from a common ancestor.
We have simulated biphasic 3MG equilibrium exchange transport in ATP-containing cells and monophasic exchange transport in ATP-free cells using such a fixed-site transport mechanism (Fig. 6). These simulations demonstrate that Kd for sugar binding to e2 is at least 10-fold lower than Kd for sugar binding to e1 and are thus consistent with the body of erythrocyte sugar transport data indicating asymmetry in transport [Km(app) for sugar uptake < Km(app) for sugar exit (54)]. Biphasic exchange kinetics are critically dependent on 1) asymmetry in binding affinity (as Kd for sugar binding to e2 approaches Kd for sugar binding to e1, the slow component of transport is accelerated and biphasic transport is lost) and 2) the rate of exchange between intersite cavity sugar and e1 or e2. This rate constant (kex) determines the rate constant for the fast component of transport, and as exchange rates fall, the fast component collapses into the slow transport component. Biphasic transport is unaffected by intersite cavity volume over the range 7.5 × 10−25 liter (the cavity volume predicted in Ref. 56) through 1.5 × 10−22 liter (the approximate volume of a single GLUT1 molecule; Ref. 56).
Our simulations (see Table 6) indicate that ATP modulates fixed-site transport by 1) increasing transport asymmetry (ATP reduces Kd for sugar binding to e2 by 2.7-fold but reduces Kd for sugar binding to e1 by only 1.7-fold) and 2) increasing the rate of exchange between bound sugar and intersite cavity sugar or bulk solvent sugars by 4.5-fold. Our simulations also suggest how endo- and exofacial inhibitors influence transport. The exofacial inhibitors maltose and phloretin inhibit the rapid phase of transport by reducing exchange (kex) between intersite cavity sugar and sugar binding sites by 10-fold. Interestingly, maltose and phloretin also act to reduce Ko for 3MG binding to e2 sites by fivefold, in part by increasing ko2. CCB inhibits rapid and slow transport by reducing kex and Ko by 27- and 10-fold, respectively. These effects on Ko are consistent with previous demonstrations indicating that single occupancy of tetrameric GLUT1 by exofacial maltose (30) or by endofacial CCB (18) allosterically increases the affinity of e2 for 3MG by as much as fivefold. CCD is without effect on the rapid phase of transport but decelerates the slow phase by reducing Ko. CCD has only very low affinity for GLUT1 (55), suggesting that its actions are mediated either by direct but low-affinity effect on GLUT1 to reduce Ko or indirectly by alterations in erythrocyte cytoskeleton. The apparent lack of affinity of CCD for GLUT1 may reflect the lack of effect of CCD on kex and the fact that previous studies did not extend transport measurements out to 60 min or longer, the time at which the slow phase of transport is observable and the effects of CCD on transport become measurable (see Fig. 4B). DNP mimics the ability of ATP depletion to modulate transport, suggesting that its effects are manifest via nonspecific membrane perturbations of GLUT1 function or that DNP displaces ATP from the GLUT1-nucleotide complex. Future studies will need to address these various possibilities.
This work was supported by National Institute of Diabetes and Digestive and Kidney Diseases Grants DK36081 and DK44888 and American Diabetes Association Grant 1-06-IN-04 (Gail Patrick Innovation Award supported by a generous gift from the Estate of Gail Patrick).
↵1 We assume the small compartment is a sphere, and we calculate the permeability coefficient P as P = k/A, where k = rate constant for filling the small compartment and A = compartment surface area.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2009 the American Physiological Society