INTRODUCTION

Cotransporters are a class of membrane transport proteins responsible for the accumulation of a diverse range of substrates into cells, including sugars, amino acids, dipeptides, neurotransmitters, drugs, and osmolytes (1). Many have been cloned, sequenced, and functionally expressed in heterologous expression systems such as Xenopus laevis oocytes. Little is known about the regulation of cotransporters, even though they contain potential target sites for phosphorylation by protein kinases (2). In this study we have expressed several mammalian Na⁺/glucose cotransporters in oocytes and used electrophysiological approaches to investigate the role of protein kinases in the regulation of their activity. There is recent evidence to suggest that the intestinal Na⁺/glucose cotransporter is regulated by protein kinase A; forskolin stimulated glucose transport within minutes in both normal mice and mice with cystic fibrosis (3).

Protein kinases may regulate the activity of membrane transport proteins either directly or indirectly. Direct effects occur through phosphorylation of the transporter, and this may change the kinetics of the transporter, such as substrate affinity, maximum velocity, or turnover number. Indirect regulation alters the rate of insertion into, or retrieval from, the plasma membrane. Membrane proteins are cotranslationally inserted into the endoplasmic reticulum and processed through the Golgi apparatus, where vesicles are formed to deliver the proteins to the plasma membrane (Fig. 5). The vesicles fuse with the plasma membrane, resulting in concomitant increases in the number of transporters and in plasma membrane area. Membrane proteins may also be retrieved selectively from the plasma membrane by endocytosis. Protein kinases may therefore alter transport activity by regulating the delivery or retrieval of membrane proteins from the plasma membrane.

Electrophysiological techniques have been developed to measure the kinetics and number of cloned cotransporters expressed in X. laevis oocytes (4, 5, 6). 1) Substrate-induced currents are used to determine substrate affinity (K_m), maximum rate of transport (I_max), and inhibitor constants (K_i). 2) Presteady-state current measurements yield charge movements (Q), which are related to the number of cotransporters (n) in the plasma membrane (n = Q_max/ze, where Q_max is the maximum charge transfer, z is the valence of the cotransporter, and e is elementary charge of an electron). 3) Capacitance measurements (C_m) give the area of the plasma membrane, assuming the specific capacitance of the membrane is 1 µF/cm². In this study we have measured these parameters before and after exposure of single oocytes to membrane-permeable activators of protein kinases.

We demonstrate that protein kinases have a powerful capacity to regulate the functional activity of cotransporters within minutes. The major effect is on protein trafficking between endosomal and plasma membranes. Up- and down-regulation due to protein kinase activation depends critically on the type and isoform of the expressed cotransporter. Activation of PKC¹ leads to reduced sugar transport by rabbit SGLT1, but to enhanced transport by human SGLT1. The information for this diverse effect must be contained in the amino acid sequence of the cotransporter.

EXPERIMENTAL PROCEDURES

Clones from rabbit, human, and rat intestinal Na⁺/glucose transporters, SGLT1 (7, 8, 9), were expressed in X. laevis oocytes. cRNAs of the various transporters were injected into oocytes (7), and oocyte currents were measured with the two-electrode voltage clamp method (6). Only oocytes with resting potentials more negative than 35 mV were used for experiments.

Fig. 1 (A-D) shows a typical pulse protocol, where an oocyte expressing rabbit SGLT1 was clamped to a holding potential (V_h) of 50 mV, and membrane currents were measured after stepping from V_h to the various clamp voltages (V_c) between 50 mV and 150 mV in 20-mV steps. Each pulse was held for 100 ms, and the average current of three sweeps was recorded. Currents were filtered at 500 Hz using a eight-pole low pass Bessel filter and digitized at 100 µs/point. Currents were measured in response to saturating external sugar before and after incubating the oocytes in a standard Na⁺ solution containing effectors of intracellular signaling pathways. Steady-state sugar-induced currents were obtained by taking the difference between steady-state currents at 100 ms in the presence and absence of sugar. Presteady-state or transient currents due to the expression of SGLT1 were isolated from the capacitative and leakage currents by fitting the total current I(t) to

(Eq. 1)

where I₁ is the capacitative current with the time constant ₁, t is time, I₂ is the SGLT1 transient current with the time constant ₂, and I_ss is the steady-state current (6). Charge movement Q due to SGLT1 was calculated by integrating the presteady-state currents at each V_c. Charge-voltage relationships were obtained by fitting the charge movement Q at various clamped voltages with the Boltzmann equation (6):

(Eq. 2)

in which Q_max = Q_dep  Q_hyp, Q_dep and Q_hyp being Q at depolarizing and hyperpolarizing limits; T, absolute temperature; F, Faraday's constant; R, the gas constant; V_0.5, the potential for 50% charge transfer; and z, the apparent valence of the movable charge.

To determine the oocyte membrane capacitance (C_m), the oocyte was held at 50 mV and a saturating concentration of substrate (5 mM MDG, -methyl-D-glucopyranoside) was added to the bath solution to eliminate the presteady-state currents. Depolarizing and hyperpolarizing voltage steps of 20 and 40 mV were applied for 100 ms, and the integral of the current transients (Q) due to charging of the membrane capacitance was calculated for each voltage step (V_c  V_h). A plot of Q versus (V_c  V_h) yielded a linear relationship with the slope C_m (6). Assuming a specific capacitance of 1 µF/cm², the capacitance of control oocytes of 300 nF (Fig. 2C) is equivalent to a membrane area of 30 × 10⁶ µm², which is consistent with morphological measurements (10).

Oocytes were bathed in a standard Na⁺ solution containing (in mM): 100 NaCl, 2 KCl, 1 CaCl₂, 1 MgCl₂, and 10 HEPES (pH 7.4). The composition of the bathing solution was varied by replacing NaCl with choline chloride, adding MDG and/or phlorizin, a specific inhibitor of the Na⁺/glucose cotransporter.

To test the effects of activators and inhibitors of intracellular signaling components on cotransporters, presteady-state and steady-state currents were measured before and after incubation of the oocytes with activators or inhibitors for 5-120 min: 0.1 mM 8-bromoadenosine 3,5-cyclic monophosphate (8-Br-cAMP), 1 µM sn-1,2-dioctanoylglycerol (DOG), 1 µM calyculin A, and 0.1 mM 8-bromoguanosine 3,5-cyclic monophosphate (8-Br-cGMP). Reagents were purchased from either Calbiochem or Sigma.

Summarized data are presented as arithmetic mean values ± S.E. with n referring to the number of observations. A paired t test was used to test for statistically significant differences. p  0.05 was set as significance level.

The putative PKA and PKC phosphorylation sites for the rabbit, human, and rat SGLT1 transporters were analyzed using GeneWorks^® 2.4 software (IntelliGenetics, Mountain View, CA). There are four putative intracellular PKC sites in all three isoforms, and one putative intracellular PKA site in rabbit and human, but not rat isoforms.

RESULTS

Injection of rabbit SGLT1 cRNA into Xenopus oocytes resulted in high levels of expression of the Na⁺/glucose cotransporter in the plasma membrane. The Na⁺-dependent MDG uptake and the MDG-dependent Na⁺ currents both increased 1000-fold (4, 5). The maximal transport rate (I_max), the number of transporters in the plasma membrane (Q_max), and the plasma membrane area (C_m) increased linearly with time over 14 days (Fig. 2, A-C). Non-injected or water-injected oocytes showed no transporter-related currents or charge movements, and their capacitance did not increase (Fig. 2C). After 6 days the current induced by 5 mM MDG was 1007 ± 103 nA (Fig. 2A, n = 5), the apparent K_m for MDG was 0.25 mM, and the apparent inhibitor constant (K_i) for phlorizin was 10 µM (see also Ref. 5). The number of plasma membrane SGLT1 transporters was 1 × 10¹¹/oocyte, calculated from Q_max measurements (71 ± 13 nC, n = 5, Fig. 2B) and assuming z = 3.5 (10). This is equivalent to an insertion rate of 2.5 × 10⁵ transporter/s.

Associated with the expression of SGLT1, the plasma membrane area of the cell, determined from C_m measurements, increased in 6 days from 3 × 10⁷ µm² to 6 × 10⁷ µm² (304 ± 10 nF (n = 21) to 594 ± 101 nF (n = 5), Fig. 1C) or by 58 µm²/s. This increase in area is equivalent to the net insertion of 1,300 membrane vesicles/s into the plasma membrane, given an average vesicle diameter of 120 nm, and a specific capacitance of the membrane of 1 µF/cm². We estimate that the rate of exocytosis of SGLT1 transport vesicles must be about an order of magnitude higher, 10,000/s, since freeze-fracture electron microscopic analysis² reveals that each transport vesicle contains only about 20 SGLT1 cotransporters, not the 200 (estimated from the ratio of the SGLT1 insertion rate (2.5 × 10⁵/s) to the net vesicle insertion rate (1,300/s).

What evidence is there that changes in oocyte C_m are in fact due to changes in membrane area? The clearest evidence comes from a recent study on the expression of brain sodium channels in oocytes (11), where a -Subunit increased functional expression of the sodium channel, and this was accompanied by up to a 4-fold increase in C_m. Electron microscopic examination of the oocytes indicated that this was due to an increase in the density and size of the microvilli. It was suggested that the -subunit increased functional expression by fusion of intracellular transport vesicles containing sodium channels with the plasma membrane.

We used 8-Br-cAMP, an activator of PKA (12); DOG, an activator of the Ca²⁺/diacylglycerol-dependent protein kinase, PKC (13); and 8-Br-cGMP, an activator of the cGMP-dependent protein kinase (14). Furthermore, we tested a blocker of the dephosphorylation of serine and threonine residues targeted by PKA and PKC, i.e. the specific inhibitor of phosphatases 1 and 2A, calyculin A (15). All were membrane-permeable and showed no effect in control oocytes (n = 6), except that calyculin A (1 µM) led to a small reduction in plasma membrane area (12 ± 8%, n = 6). We did not use phorbol esters to activate PKC because we found that phorbol 12-myristate 13-acetate (50 nM) increased the background current, decreased oocyte resistance, and depolarized the membrane potential in control, non-injected oocytes. The effects of 8-Br-cAMP and DOG on rabbit SGLT1-expressing oocytes were reversible after incubation for 60 min in buffer free of activators.

Sugar-dependent currents of oocytes injected with rabbit SGLT1 cRNA were measured, before and after these oocytes were incubated in 0.1 mM 8-Br-cAMP, 1 µM DOG, or 0.1 mM 8-Br-cGMP for 5-120 min and the effects on Na⁺/glucose cotransport were recorded. Fig. 3 shows that 8-Br-cAMP increased the maximum rate of transport (I_max) by 27 ± 9% (n = 6), whereas DOG reduced the rate by 62 ± 5% (n = 6). Full responses were obtained within 30 min, and significant changes in current were observed after 5 min with DOG and after 20 min with 8-Br-cAMP. 8-Br-cGMP did not affect sugar transport (3 ± 9%, n = 8). Neither 8-Br-cAMP nor DOG altered the affinity of rabbit SGLT1 for sugar or phlorizin (data not shown). The effects of 8-Br-cAMP or DOG were independent of the level of expression for I_max ranging from 244 to 3172 nA, in direct contrast to the effects described for the GABA-transporter expressed in Xenopus oocytes (16).

The results of Fig. 3 show that activation of PKA increases, and activation of PKC decreases, the maximum rate of Na⁺/glucose cotransport in oocytes expressing rabbit SGLT1. These effects could have been accomplished by changing either the turnover number of the transporter, or the number of transporters in the plasma membrane. Fig. 4 summarizes a series of experiments where we simultaneously recorded changes in transport, numbers of transporters, and membrane area. In oocytes expressing rabbit SGLT1 (Fig. 4A) 8-Br-cAMP increased transport 28 ± 4% (n = 10), and DOG decreased transport 51 ± 5% (n = 13). Fig. 4B shows that 8-Br-cAMP increased the number of cotransporters in the membrane by 19 ± 7% (n = 10), and DOG decreased the number of cotransporters by 41 ± 4% (n = 13). Fig. 4C shows that 8-Br-cAMP increased the area of the plasma membrane by 13 ± 5% (n = 10), and DOG decreased the area by 30 ± 6% (n = 13). Plots of the 8-Br-cAMP- and DOG-induced changes in maximum transport rate (I_max) against the changes in number of cotransporters (Q_max) and the changes in plasma membrane area (C_m) were both linear (I_max/Q_max = 0.9 ± 0.1; I_max/C_m = 1.0 ± 0.1). This indicates that PKA and PKC regulate sugar transport by increasing and decreasing the trafficking of rabbit SGLT1 to the plasma membrane. Calyculin A (1 µM) had no significant effect on the transport rate, the number of cotransporters in the plasma membrane, or the membrane surface area (not shown), consistent with the opposing effects of PKA and PKC.

1

1

Similar results were obtained with 8-Br-cAMP on oocytes expressing the human and rat isoforms of SGLT1. With human SGLT1 8-Br-cAMP (Fig. 4, D-F) increased the rate of sugar transport 39 ± 15% (n = 9), the number of transporters in the plasma membrane 22 ± 10% (n = 9), and the membrane area by 15 ± 7% (n = 9). In the case of rat SGLT1 (data not shown), 8-Br-cAMP also significantly increased I_max 12 ± 5% (n = 9) and Q_max 16 ± 8% (n = 9), even though there are no consensus phosphorylation sites for PKA. There was no change in plasma membrane area with rat SGLT1 (1 ± 4%, n = 9), suggesting that in this case there was no significant change in the net rate of vesicle fusion with the plasma membrane. This may simply reflect that the rate of exocytosis is an order of magnitude greater than the net increase in membrane area and that the PKA may effect both exo- and endocytosis (Fig. 4).

Strikingly different results were obtained with DOG in oocytes expressing the three SGLT1 isoforms. While DOG caused a decrease in all three rabbit SGLT1 parameters (I_max, Q_max, and C_m; Fig. 4, A-C), DOG increased the rate of sugar transport by human SGLT1 by 41 ± 10% (n = 11), and there were proportionate increases in the number of transporters (33 ± 11%) and the area of the plasma membrane (30 ± 7%) (Fig. 4, D-F). The results obtained with DOG on oocytes expressing rat SGLT1 were very similar to those for rabbit SGLT1: DOG decreased transport 30 ± 5%, decreased the number of transporters 29 ± 7%, and decreased the plasma membrane area 13 ± 3% (n = 8).

In four experiments with oocytes expressing human SGLT1, calyculin A increased I_max 13-92% and Q_max 3-55%. These increases were similar to those produced by 8-Br-cAMP and DOG (Fig. 4) and suggest that regulation occurs by phosphorylation of serine and/or threonine residues in targeted proteins, the cotransporters or chaperons.

DISCUSSION

To answer the question about the role of protein kinases in the regulation of sodium/glucose cotransport by SGLT1, we have employed: 1) the powerful oocyte expression system to express cloned isoforms of SGLT1, 2) a unique combination of electrophysiological techniques to measure the kinetics, number of plasma membrane transporters, and the trafficking of these transporters between the cell and the plasma membrane, and 3) standard protocols to activate PKA and PKC in tissues and cells, including oocytes (12, 13).

The electrophysiological methods provide sensitive assays of both the kinetics and number of cotransporters in the plasma membrane. The major advantages are that these parameters can be determined in a single cell before and after activation of protein kinases. In addition, the capacitance measurements provide a simple and direct method to determine the surface area of the cell, and changes in area due to endocytosis and exocytosis. The validity of capacitance as a measure of surface area of the oocyte was confirmed by careful morphometric measurements in oocytes (10), and electron microscopic studies have also established up to 4-fold increases in capacitance are due to increases in oocyte surface area (11, 17).

8-Br-cAMP has long been used to activate PKA in oocytes expressing cloned membrane proteins (e.g. Refs. 18 and 19) and phorbol esters have been used to activate PKC (16, 17, 19). 8-Br-cAMP and DOG are membrane-permeable reagents that act on the last step in the activation of PKA and PKC in cells. They rapidly, and reversibly, modulated the functional expression of SGLT1 in oocytes. Additional evidence that PKA and PKC are activated by these agonists in our experiments is that calyculin A, a specific blocker of the dephosphorylation of serine and threonine residues targeted by PKA and PKC, produced similar results to 8-Br-cAMP and DOG in oocytes expressing human SGLT1.

This study demonstrates that SGLT1 cotransporters are regulated by activation of protein kinases in oocytes. The maximum transport rate of rabbit SGLT1 was stimulated 30% by activation of PKA and inhibited by up to 50% by activation of PKC. Changes in transport rate were observed within minutes after adding the membrane permeable activators to the bath. Given the rapid effects, and the fact that these experiments were carried out on cRNA-injected oocytes, it is reasonable to conclude that protein kinases are acting at a post-translational step in the functional expression of the transporters in the plasma membrane. Even with the use of a single heterologous expression system, oocytes, and the same protein kinase activators, different responses were obtained with different isoforms of SGLT1. Activation of PKA increased the transport rate of rabbit, human, and rat SGLT1 s, but decreased the rate of transport by the mouse brain taurine (TAUT) cotransporter.³ Activation of PKC reduced the transport rates of rabbit and rat SGLT1, and mouse TAUT,³ but increased the transport rate of human SGLT1 (see Fig. 4D). Therefore, the effect of a protein kinase depends critically on the transporter and the isoform being expressed.

Similar conclusions can be drawn from the literature, where it has been reported that activation of PKC increased the rate of GABA uptake into oocytes expressing the cloned rat brain Na⁺/Cl/GABA cotransporter (16), and decreased the rate of phosphate transport in oocytes expressing the cloned rat renal Na⁺/phosphate cotransporter (19).

Protein kinases regulate Na⁺/glucose cotransport by altering the maximum rate of transport. We did not observed any effect on the apparent affinity for glucose or the inhibitor constant for the non-transported, competitive inhibitor phlorizin (data not shown), or the turnover number of the transporter (I_max/Q_max, Fig. 4). Similar results were obtained for Na⁺/Cl/taurine transport,³ and the effect of PKC on GABA uptake by GAT1 expressed in oocytes and glycine uptake by GLYT1b expressed in HEK 293 cells (16, 21).

Our studies with SGLT1 (Fig. 4) demonstrate that changes in the maximum rate of transport are due to changes in the number of cotransporters in the plasma membrane, and the concurrent changes in membrane area suggest that the protein kinases regulate the number of cotransporters by modulating the rate of insertion into, or retrieval from, the plasma membrane (Fig. 5). Regulated exo- or endocytosis also account for the effects of 8-Br-cAMP, DOG, and calyculin A on the retinal Na⁺/Cl/taurine cotransporter.³ Furthermore, fractionation studies also suggest that PKC regulates GABA transport in oocytes by controlling the distribution of GAT1 cotransporters between cytoplasmic compartments and the plasma membrane (16). Vesicle fusion with or vesicle retrieval from the plasma membrane is known to regulate the activity of other transporters, ion channels and pumps in cells and tissues, e.g. facilitated glucose transport in adipocytes by GLUT4, water transport in the renal collecting duct via CHIP28, Cl transport in exocrine glands, and acid secretion in the stomach, are regulated by insertion and retrieval of the transport protein into or from the plasma membrane (22).

Insight into the rates of SGLT1 protein insertion into the plasma membrane of oocytes can be obtained from our electrophysiological studies. Activation of PKA by 8-Br-cAMP results in the insertion of approximately 2.5 × 10¹⁰ proteins into the plasma membrane within 30 min (Fig. 4), or 1 × 10⁷/s (Q_max increased by 14 nC, and n = Q_max/ze = 14 nC/3.5 × 1.6 × 10¹⁹ C = 2.5 × 10¹⁰ proteins). This compares with a constitutive insertion rate of 2.5 × 10⁵/s (71 nC/6 days). The 100-fold increase in transporter insertion brought about by 8-Br-cAMP was accompanied by an increase in oocyte capacitance of 130 nF/30 min (from 445 ± 20 nF to 575 ± 70 nF), which compares to the base-line level increase in membrane capacitance of 1 nF/30 min after the injection of cRNA of rabbit SGLT1 (Fig. 2C). An increase of 130 nF is equivalent to a fusion rate of 110,000 vesicles/s, assuming that the increase in membrane capacitance was due to the insertion of membrane vesicles with an average diameter of 120 nm and that the specific capacitance of the vesicle membrane is 1 µF/cm², i.e. 8-Br-cAMP increases the net rate of vesicle fusion 100-fold above the constitutive rate. The rates of vesicle fusion may seem high relative to that in other cells such as neurons, chromaffin cells, and pituitary cells (20 to 10⁴ vesicles/s) (23), but it should be noted that the area of a 1-mm diameter oocyte is about 250,000 times larger than these cells.

The ratio of the rates of SGLT1 protein insertion to the rate of vesicle insertion was similar in untreated oocytes, 8-Br-cAMP-activated, and DOG-activated oocytes (Q_max/C_m). This suggests a density of 100-200 SGLT1 proteins/intracellular transport vesicle, whereas freeze-fracture electron micrographs suggest a density of 20/vesicle.² This discrepancy is due to the fact that oocyte capacitance measurements measures the net change in surface area, and it indicates that the rates of exocytosis are about an order of magnitude higher than the net rate. The fact that the ratios of protein to vesicle insertion and retrieval are similar in resting and activated oocytes further indicate that the protein kinases only alter the rate, not the mechanism of SGLT1 insertion or retrieval. This implies that there is a pool of around 10⁸ transport vesicles at day 6 in the oocyte ready for regulated insertion. Electron microscopic examination of thin sections of Xenopus oocytes does in fact document the presence of numerous coated and uncoated membrane vesicles in close proximity to the plasma membrane (see Ref. 10).

Since PKC decreases transport rate for rabbit SGLT1, but increases transport in the case of human SGLT1 (see Fig. 4, A and D), we conclude that it is the transport protein itself that determines the direction of its regulated trafficking. It is unlikely that regulation occurs by direct phosphorylation of the cotransporters, because: 1) transport by rat SGLT1 was activated by 8-Br-cAMP, even though this transporter does not contain phosphorylation sites for PKA; and 2) human PEPT1 (24) shows no response to PKC,⁴ even though it has several putative PKC phosphorylation sites. Additional supporting evidence is that: 1) PKC activation still occurs after deletion of all PKC consensus phosphorylation sites in the Na⁺/Cl/GABA (GAT1), Na⁺/phosphate (NaP_i-2), and Na⁺/Cl/glycine (GLYC1b) cotransporters (16, 19, 21); and 2) PKA activation still occurs after deletion of the facilitative glucose transporter GLUT4 is down-regulated by PKA even after deletion of the PKA consensus phosphorylation site (25). Proteins may be phosphorylated at non-consensus sites, and so direct experiments are required to determine the phosphorylation state of SGLT1. To date it has not been feasible to isolate SGLT1 protein from oocytes to measure phosphorylation, largely due to the lack of suitable antibodies for immunoprecipitation and the complexity of oocytes.

Differences in regulated trafficking by activation of the protein kinases must be due to differences in the protein sequence; rabbit and human SGLT1 are 85% identical in their amino acid sequence (2). Inspection of the aligned human, rabbit, and rat SGLT1 sequences (26) reveals three non-conserved regions, at the amino terminus, the extracellular loop between helices 6 and 7, and the intracellular loop between helices 13 and 14. The cytoplasmic loop, residues 550-636, is the most likely domain involved in the sequence specific regulation. In all isoforms this domain does contain consensus sites for PKC phosphorylation, but none for PKA phosphorylation.

A similar distinction was reported for the regulation of the GLUT1 and GLUT4 isoforms by insulin (22), where a NH₂-terminal cytoplasmic motif, -FQQI, mediates the intracellular sequestration of GLUT4 (25). In the polarized kidney cell line, Madin-Darby canine kidney cells, protein kinases regulate the delivery of influenza virus hemagglutinin from the Golgi to the plasma membrane, and mutation of a single amino acid residue in hemagglutinin dramatically alters its sorting (27, 28). Preliminary studies with mutant human SGLT1 proteins have identified single amino acid residues that appear to determine the direction of regulation by a protein kinase.⁴ These considerations suggest that single residues or motifs in the cotransporter sequences determine how protein kinases regulate SGLT1 trafficking between the cytoplasm and the plasma membrane.

In both neuronal and non-neuronal cells, the molecular machinery involved in vesicle transport between the trans-Golgi and the plasma membrane has been identified (29, 30). The vesicles that bud off from the trans-Golgi bear unique address markers (synaptobrevins or cellulobrevins) that target membranes identified by receptors (e.g. syntaxin in neuronal presynaptic membranes). Preliminary studies⁴ suggest that cellulobrevins are involved in the both the constitutive and protein kinase regulated insertion of SGLT1 cotransporters into the oocyte plasma membrane. Botulinum toxin F, a Zn²⁺-dependent protease that specifically cleaves synaptobrevins and cellulobrevins, partially inhibits the constitutive expression of rabbit SGLT1 in oocytes, and blocks the response to both 8-Br-cAMP and DOG.

We cannot preclude a direct effect of protein kinases on the SGLT1, and this may account for some of the variation observed in PKA- and PKC-related changes in maximum transport rate, number of transporters, and the area of the plasma membrane (Fig. 4). Nevertheless, we have not detected changes in the kinetics of the transporter (sugar affinity, the phlorizin inhibitor constant, and the turnover number, I_max/Q_max) other than the maximal transport rate. We would expect changes in these parameters if the transporter was directly phosphorylated. In the case of at least one cotransporter, the glutamate cotransporter (GLT-1), transport is increased by PKC, and this occurs by direct phosphorylation of the protein (29). Mutation of the serine residue that forms part of the consensus PKC phosphorylation site (Ser-113) eliminated the response to phorbol esters (31).

Our studies clearly demonstrate that human, rabbit, and rat Na⁺/glucose cotransporters expressed in Xenopus oocytes are regulated by protein kinases, and clearly suggest that this occurs through regulated exo- and endocytosis. In the rat small intestine, the glucagon induced increase in glucose transport across the brush border membrane is thought to occur through cAMP (20), and in mice with cystic fibrosis, forskolin enhanced glucose transport 2-4-fold within minutes (3). It is possible that these rapid responses to activators of intracellular cAMP are also due to the redistribution of SGLT1 proteins between the intracellular compartment and the brush border membrane. The different responses of the rabbit and human SGLT1s in oocytes to activation of PKC (Fig. 4) also suggest that there may be species differences in the regulation of sugar transport by kinases in the intestine. However, while the physiological significance of such species differences in the regulation of SGLT1 is puzzling, there are marked differences in the kinetics and substrate specificity of the rat, rabbit, and human isoforms (26).