It has recently been shown that a sodium background current, ib,Na, exists in cardiac muscle cells whose effect is to depolarize the membrane so that the resting potential, Vm, is positive to the potassium equilibrium potential, EK. In ventricular cells, where ib,Na is smallest, Vm is about 10 mV positive to EK (EK = -87 mV at 37¡C). Yet, replacement of Na+ ions by large impermeant cations does not cause the expected hyperpolarization. We have studied this problem in guinea-pig myocytes using a single microelectrode recording technique in combination with a rapid external solution switch. Cells depolarized [less than or equal to] 0·5 mV from potentials between -80 and -73 mV and hyperpolarized up to 5 mV from potentials between -73 and -64 mV when 70 mM choline chloride or N-methyl-D-glucamine chloride were used to replace 70 mM Na+ in the bathing solution. Replacement by 70 mM lithium chloride, however, only caused hyperpolarization in very depolarized cells when the voltage change was much smaller. The changes were complete almost as soon as the solution change, i.e. within 250 ms, indicating that the actions are attributable to the external solution change rather than to secondary changes in intracellular concentrations. Patch clamp recording was used to investigate the mechanism involved. These experiments showed that the presence or absence of the inward rectifier current iK1 determines in which direction Na+ removal acts. In the absence of iK1 the changes are attributable to removal of ib,Na, whereas in the presence of iK1 the changes resemble the i(V) relation for iK1, implying that Na+ regulates iK1 in a way that can mask the changes in ib,Na. These results explain why removal of Na+ does not lead to hyperpolarization in ventricular cells as would be expected if changes in ib,Na were solely responsible. Computer reconstruction shows that the effects may be attributed to actions of sodium removal on the conductance and gating of iK1.

Under physiological conditions, calcium-dependent ([Ca2+]i-dependent) Cl- currents (ICl(Ca)) have been suggested to participate in the repolarizing processes. In this paper, the possible contribution of ICl(Ca) to transient inward currents and, hence to arrhythmias, has been studied in myocytes from the working myocardium and from the conductive system. Single atrial, ventricular and Purkinje cells, isolated enzymatically from rabbit heart, have been studied under whole-cell voltage-clamp and were internally perfused with the fluorescent Ca2+ indicator, fura-2 (100 µM). Ca2+ release from the sarcoplasmic reticulum was either induced by external application of caffeine or occurred spontaneously in Ca2+-overloaded cells. Membrane currents accompanying Ca2+ transients showed linear current-voltage characteristics between -60 and +80 mV as evidenced from fast voltage ramps. When intra- and extracellular Cl- concentrations were kept symmetrical in the absence of the Na+-Ca2+ exchange mechanism, transient currents had a reversal potential close to 0 mV. Reduction of external Cl- concentration shifted this reversal potential towards the new Cl- equilibrium potential. Neither the time course of the transient currents nor the shift in their reversal potentials was significantly affected by the presence of Na+. Approximately 90 % of this current was blocked by the application of the Cl- channel blocker, anthracene-9-carboxylic acid (0·5 mM) at +80 mV. It is concluded, that [Ca2+]i-activated transient membrane currents in atrial, ventricular and Purkinje cells of rabbit heart are mainly due to the activation of a [Ca2+]i-dependent Cl- current.

Short chain fatty acids (SCFAs) and CO2 have been shown to stimulate net Mg2+ efflux from the isolated reticulorumen in vivo. To investigate the underlying mechanisms of Mg2+ transport we performed Ussing chamber and microelectrode experiments and measured 28Mg2+ fluxes across sheep rumen epithelium in vitro. In the presence of SCFAs mucosal-to-serosal Mg2+ flux (JMgm-s) amounted to 82·3 ± 7·8 nmol cm-2 h-1 and serosal-to-mucosal Mg2+ flux (JMgs-m) to 3·2 ± 0·7 nmol cm-2 h-1. Replacing SCFAs with gluconate caused a 50 % reduction of JMgm-s, whereas JMgs-m was not affected. Among the SCFAs, n-butyrate was more effective in stimulating JMgm-s than acetate, propionate or iso-butyrate. Eliminating HCO3--CO2 from SCFA-containing solutions did not affect Mg2+ fluxes, whereas the same replacement in SCFA-free solutions led to a further reduction in JMgm-s. JMgm-s decreased after the addition of ethoxyzolamide to SCFA-free, bicarbonate buffered solutions. Decreasing mucosal pH from 6·4 to 5·4 increased JMgm-s in SCFA-free, bicarbonate buffered solutions. SCFAs had no effect on the apical membrane potential of rumen epithelial cells. The experiments show that both SCFAs and CO2 stimulate Mg2+ transport through an increase in JMgm-s, most probably via stimulation of a Mg2+-2H+ exchange mechanism. SCFAs may have additional metabolic effects on Mg2+ transport.

A disturbed calcium homeostasis characterizes diabetic pregnancy. This study documents changes in bone mineral composition in diabetic pregnant rats and examines the effect of insulin replacement. Control pregnant (CP), diabetic pregnant (DP) and insulin-treated DP (DPi) rats were assessed for femoral calcium and magnesium content, bone mineral density (BMD) and the ratio of hypertrophic to maturing and proliferative cells in the femoral growth plate. DP rats showed a significantly (P < 0·01) lower body weight, femoral weight and length than CP rats. Femoral calcium and magnesium content was also significantly (P < 0·05) lower in DP rats, as was ash weight. When calcium and magnesium were normalized for ash weight no signficant differences were apparent. A significantly (P < 0·05) lower total BMD at the distal femur was seen in DP rats. This comprised a significantly (P < 0·01) lower trabecular BMD with no significant change in cortical BMD. A significantly (P < 0·05) higher ratio of hypertrophic to maturing and proliferative cells of the femoral growth plate was evident in DP animals. DPi rats showed normal blood glucose concentrations and femoral growth plate histology. DPi rats also showed normal femoral weight and length but only partially restored femoral ash weight and mineral content. Insulin failed to normalize total or trabecular BMD. Diabetes mellitus clearly has a marked effect on bone growth and mineral content in pregnancy which may be relevant to overall calcium homeostasis. The lower bone growth, bone calcium content and trabecular BMD may be unfortunate consequences of the marked hypercalciuria reported elsewhere in diabetes and may serve to maintain normocalcaemia in the disease.

Experiments were performed on twenty-two neurologically normal subjects in order to investigate the effects of skin brushing on H reflex excitability in triceps surae. H reflex amplitude was observed to decrease during brushing of a 2 cm × 10 cm area of skin overlying triceps surae, returning to control levels when brushing ceased. Alterations in the duration of the brushing period did not affect the magnitude of the H reflex inhibition, although increased frequency of brush strokes per minute slightly increased the amount of inhibition observed. Brushing of smaller skin areas overlying the muscle only sometimes resulted in H reflex inhibition. Brushing over other skin areas of the lower limb produced only slight effects on triceps surae H reflex amplitude. Removal of cutaneous input from the skin overlying triceps surae abolished any effects of brushing in this area on H reflex amplitude. These results demonstrate an inhibitory effect of brushing on H reflex excitability in normal subjects. If such results are confirmed in subjects with neurological deficit, they could have implications for the use of brushing in clinical practice.

This series of papers is presented as a tribute to Wilfred F. Widdas, Emeritus Professor, Bedford and Royal Holloway College, University of London. The papers are the outcome of a meeting held by the Physiological Society on January 5th 1997 at Sheffield as a celebration of Widdas's scientific work. The work referred to of course is the series of papers by Widdas and co-workers describing mediated transfer of sugars across cell membranes. The ideas therein relating to carrier-mediated transport of sugars across membranes have become one of the most influential scientific paradigms in modern biology. The history of science shows that a successful paradigm has its supporters, who work either to extend its influence to ever wider spheres, or to understand the nature of its constraints. The natural tendency is for this support group to diminish the importance of anomalous results; this consequently breeds an opposition group whose aim is to show that these anomalies require major adjustments to the paradigm. This arrangement leads to the formation of scientific factions and eventually either to the overthrow of the paradigm or to a major reconstruction and thus scientific advances. A special feature of sugar transport is the relative scientific longevity of the workers in the field, of whom Widdas is a prime example; a similar tendency was found in quite a few of the participants at the meeting. Another is that they do not fall easily into the facile categories of supporters or adversaries of the carrier paradigm. Although Wilfred Widdas's early work led both to the formation and refinement of the carrier molecule, he always has had a more open view about the nature of the transporter than many of its more ardent supporters. His indefatigable scientific curiosity has led him to investigate some nagging anomalies in the theory of glucose transport and consequently somewhat away from the mainstream. This rare and courageous stance stems, perhaps, from the Non-Conformist tradition from which he comes and certainly demonstrates a monumental dedication to science. All the participants are grateful to the Physiological Society and Glaxo-Wellcome for their financial support and the editors of Experimental Physiology for their help in testimonial publication.

This series of papers is presented as a tribute to Wilfred F. Widdas, Emeritus Professor, Bedford and Royal Holloway College, University of London. The papers are the outcome of a meeting held by the Physiological Society on January 5th 1997 at Sheffield as a celebration of Widdas's scientific work.

The work referred to of course is the series of papers by Widdas and co-workers describing mediated transfer of sugars across cell membranes. The ideas therein relating to carrier-mediated transport of sugars across membranes have become one of the most influential scientific paradigms in modern biology.

The history of science shows that a successful paradigm has its supporters, who work either to extend its influence to ever wider spheres, or to understand the nature of its constraints. The natural tendency is for this support group to diminish the importance of anomalous results; this consequently breeds an opposition group whose aim is to show that these anomalies require major adjustments to the paradigm. This arrangement leads to the formation of scientific factions and eventually either to the overthrow of the paradigm or to a major reconstruction and thus scientific advances.

A special feature of sugar transport is the relative scientific longevity of the workers in the field, of whom Widdas is a prime example; a similar tendency was found in quite a few of the participants at the meeting. Another is that they do not fall easily into the facile categories of supporters or adversaries of the carrier paradigm. Although Wilfred Widdas's early work led both to the formation and refinement of the carrier molecule, he always has had a more open view about the nature of the transporter than many of its more ardent supporters. His indefatigable scientific curiosity has led him to investigate some nagging anomalies in the theory of glucose transport and consequently somewhat away from the mainstream.

This rare and courageous stance stems, perhaps, from the Non-Conformist tradition from which he comes and certainly demonstrates a monumental dedication to science.All the participants are grateful to the Physiological Society and Glaxo-Wellcome for their financial support and the editors of Experimental Physiology for their help in testimonial publication.

The shortcomings of diffusion as a mechanism of glucose transfer across cell membranes date back to the last century. The early work, mostly qualitative, was greatly extended by LeFevre in America and by Wilbrandt in Switzerland, who reported their quantitative studies at the Bangor Symposium of The Society of Experimental Biology in 1953 (LeFevre, 1954; Wilbrandt, 1954). Widdas (1952) published simple carrier kinetics for glucose transfer which predicted the phenomenon of uphill transfer by counterflow. However, in that 1952 paper it was carefully pointed out that a good fit to the kinetics did not define the details of the mechanism. Nevertheless, the reverse is true: namely, any viable scheme of working must be consistent with the established kinetic properties. For instance the kinetics of uphill transfer by counterflow (confirmed four years later by Park, Post, Kalman, Wright, Johnson & Morgan, 1956; Rosenherg & Wilbrandt, 1957) clearly exclude a simple water channel through the membrane.The concept of a mobile membrane carrier on which the early kinetics were based was variously modified to meet newer developments, which included asymmetric affinities on the outside and inside of the red cell membrane. The mobile carrier has now been superseded by the concept of membrane protein transporters. Nevertheless, some of the nomenculture, such as carrier transport and carrier catalysis (Krämer, 1994), has persisted although it no longer conveys all the subtleties of the transfer mechanism.

Net sugar import by human erythrocytes consists of ATP-modulated rapid and slow phases while sugar export consists of a single slow phase. We have proposed that this behaviour results from obligate substrate tunnelling from transporter to bulk cytosol through a complex containing high-affinity, low-capacity sugar binding sites (Cloherty, Sultzman, Zottola & Carruthers, 1995). Glyceraldehyde-3-phosphate dehydrogenase (GAPDH) is known to compartmentalize ATP delivery to erythrocyte membrane ATPases and interact directly with the erythrocyte glucose transporter in vitro. The present study examines the possibility that GAPDH is an obligate component of the hypothesized sugar-binding complex. GAPDH remains associated with the erythrocyte membrane following cell lysis and remains associated with the cytoskeleton under conditions where more than 99 % of the membrane glucose transport protein (GLUT1) is released by detergent (Triton X-100). GAPDH is released from erythrocyte membranes upon exposure to Mg.ATP or to NADH. ATP displacement of membrane-bound GAPDH is half-maximal at 200 µM ATP and appears to involve ATP interaction with multiple, co-operative sites. GAPDH interaction with purified tetrameric GLUT1 is saturable, co-operative and also inhibited by ATP. ATP inhibition of GAPDH binding to purified tetrameric GLUT1 is less effective than ATP inhibition of GAPDH binding to intact erythrocyte membranes. Removal of cellular GAPDH by exposing erythrocyte membranes to NADH prior to membrane resealing neither affects ATP modulation of sugar transport nor reduces biphasic net sugar uptake to a single phase. We conclude that ATP-sensitive GAPDH interaction with the cytoplasmic surface of erythrocyte membranes and GLUT1 is responsible neither for ATP modulation of sugar transport nor for multiphasic net sugar import by human red cells.

The inward transport of two purines, adenosine and hypoxanthine, at 37¡C by horse erythrocytes was compared. No mediated transport of adenosine was detected in horse erythrocytes, nor was saturable, high-affinity binding of the potent facilitated-diffusion inhibitor nitrobenzylthioinosine demonstrable in horse erythrocyte membranes. In contrast, erythrocytes from most horses possessed a saturable sodium-dependent hypoxanthine transporter (apparent Km, 100 ± 28 µM; Vmax, 0·20 ± 0·08 mmol (l cells)-1 h-1; means ± S.E.M., n = 5). Guanine inhibited hypoxanthine influx (apparent Ki, 24 ± 6 µM), but adenine and xanthine had no effect. Unlike human erythrocytes, no sodium-independent hypoxanthine transporter was detected in horse erythrocytes. There are, however, a small number of animals ([similar]15 %) whose erythrocytes fail to transport hypoxanthine. This variation appears to be under genetic control, but the precise nature of the control is unknown.

The properties are discussed of system y+L, a broad scope amino acid transporter which was first identified in human erythrocytes. System y+L exhibits two distinctive properties: (a) it can bind and translocate cationic and neutral amino acids, and (b) its specificity varies depending on the ionic composition of the medium. In Na+ medium, the half-saturation constant for L-lysine influx was 9·5 ± 0·67 µM and the inhibition constant (Ki) for L-leucine was 10·7 ± 0·72 µM. L-Leucine is the neutral amino acid that binds more powerfully, whereas smaller analogues, such as L-alanine and L-serine interact less strongly (the corresponding inhibition constants were Ki,Ala, 0·62 ± 0·11 mM; Ki,Ser, 0·49 ± 0·08 mM). In the presence of K+, the carrier functions as a cationic amino acid specific carrier, but Li+ is able to substitute for Na+ facilitating neutral amino acid binding. The effect of the inorganic cations is restricted to the recognition of neutral amino acids; translocation occurs at similar rates in the presence of Na+, K+ and Li+. The only structural feature that appears to impair translocation is bulkiness and substrates with half-saturation constants differing by more than 100-fold translocate at the same rate. This suggests that translocation is largely independent of the forces of interaction between the substrate and the carrier site. System y+L activity has been observed in Xenopuslaevis oocytes injected with the cRNA for the heavy chain of the 4F2 human surface antigen. 4F2hc is an integral membrane protein with a single putative membrane-spanning domain and it remains to be clarified whether it is part of the transporter or an activator protein.

Most cancer deaths result from the cancer either being intrinsically resistant to chemotherapeutic drugs, or becoming resistant after being initially sensitive (Leyland-Jones, Dalton, Fisher & Sikic, 1993). Often, in cells grown in cell culture, drug resistance correlates with the presence of one or more of the so-called P-glycoproteins or multidrug resistance (MDR) proteins, products of the mdr family of genes (Endicott & Ling, 1989). The mdr genes (with two forms in humans (MDR1 and MDR2), and three (mdr1, mdr2 and mdr3) in mice and hamsters) have been cloned, sequenced (Gros, Croop & Housman, 1986) and shown to be members of the 'ABC' (ATP binding cassette) or 'traffic ATPase' superfamily (Doige & Ames, 1993). All members of this superfamily (which includes the CFTR chloride ion channel, the product of the STE6 gene in yeast that transports the [alpha]-mating factor polypeptide, bacterial protein translocator HlyB that exports haemolysin, and many bacterial permeases) have a membrane transport function and share homologous sequences that often bind ATP (Kane, Pastan, Gottesman, 1990; Gros & Buschman, 1993) and can hydrolyse it. Human MDR1 and mouse mdr1 and mdr3 (Tang-Wai, Kajiji, Dicapua, de Graaf, Roninson & Gros, 1995) confer multidrug resistance, while human MDR2 and mouse mdr2 transport phosphatidylcholine and function as lipid translocases (Smit et al. 1993). The genes have tandem repeated structures, each coding for polypeptide sequences which putatively straddle the membrane six times, and each containing an ATP-binding region. The MDR genes code for glycoproteins of molecular weight 150000-180000. In drug-resistant cells, the genes are generally present as multiple repeats on episomes. Some clinical studies show that expression of the MDR1 gene is enhanced in drug-resistant tumours in patients and that this expression is a good predictor of resistance (Chan et al. 1991).Expression of the mdr genes is associated with a greatly increased resistance to the killing action of many cytotoxic drugs (Biedler & Riehm, 1970; Johnson, Chitnis, Embrey & Gregory, 1978). It brings about a marked (up to fiftyfold) reduction in the cellular steady-state level of its numerous 'substrates' (Dano, 1973; Inaba, Kobayashi, Sakurai & Johnson, 1979; Fojo, Akiyama, Gottesman & Pastan, 1985) which are fairly hydrophobic compounds but possess also some hydrophilicity. Many of these substrates are weak bases. Some bear permanent positive charges (Lampidis et al. 1989), but zwitterions seem not to be substrates. The MDR protein transports such drugs actively out of the cells in which the mdr gene is expressed (Lankelma, Spoelstra, Dekker & Broxterman, 1990; Dordal, Winter & Atkinson, 1992), that is, it is a drug pump. The protein can be present in as many as 8 × 105 (Sehested, Simpson, Skovsgaard & Jesnsen, 1989; Demant, Sehested & Jensen, 1990) or even 3 × 106 (Shapiro & Ling, 1994) copies per cell. Resistance is reversed by numerous compounds, termed 'reversers', 'modulators', or 'chemo-sensitizers' (e.g. verapamil; Tsuruo, Iida, Yamashiro, Tsukagoshi & Sakurai, 1982) and also by other MDR substrates (Horio, Lovelace, Pastan & Gottesman, 1991). P-glycoprotein has been isolated from cell membranes and demonstrated to be an ATPase, with a turnover number of some 25 s-1 at 37¡C (Ambudkar, Lelong, Zhang, Cardarelli, Gottesman & Pastan, 1992; Urbatsch, Al-Shawi & Senior, 1994; Sharom, Yu, Chu & Doige, 1995). P-glycoprotein is often present in tissues and at tissue surfaces that transport water and nutrients (kidney, gut, the blood-brain barrier; Thiebaut, Tsuruo, Hamada, Gottesman, Pastan & Willingham, 1987; Lum & Gosland, 1995), but is also present in the pancreas.What might be the physiological role of the P-glycoproteins? It was early suggested that they may be concerned in the protection of the organism from toxic, perhaps mutagenic, xenobiotic substances (Gottesman & Pastan, 1988). Their appearance in cytotoxin-resistant tumours would be an example of the ectopic expression of genes that is so often encountered in tumours. P-glycoprotein thus acts as a double-edged sword, protecting its wielder in most situations, at the expense of aiding in his or her destruction when tumours have already developed (Gottesman & Pastan, 1988). Studies using animals in which one or other mdr gene has been knocked out have thrown light on the matter. Thus, Schinkel et al. (1994), in an elegant study using mice with a knock-out mutation in the mdr1A gene, have shown that the P-glycoprotein is a major component of the blood-brain barrier, ensuring that lipid-soluble toxins (ivermectin, for example) are pumped out of the brain tissues across that barrier, protecting the organism against brain damage. P-glycoprotein acts also to keep intracellular drug concentrations low in a number of other body tissues. Blocking P-glycoprotein using knock-out mice (Schinkel et al. 1994) or by the addition of reversers of P-glycoprotein (Lyubimov, Lan, Pashinsky, Ayesh & Stein, 1995) leads to a significant rise in the tissue content of vinblastine. The tissues in which P-glycoprotein brings about a reduced concentration of cytotoxin are, in many cases, just those that are intrinsically resistant to chemotherapy by cytotoxins (Lum & Gosland, 1995).There is no fully accepted model for the mechanism of action of this important molecule. There is much evidence that P-glycoprotein behaves quite differently from conventional membrane pumps. The 'vacuum cleaner' model (Raviv, Pollard, Bruggemann, Pastan & Gottesman, 1990), in which P-glycoprotein is a machine that cleans its substrates out of the plasma membrane, and delivers them to the exterior aqueous phase, provides an intriguing general framework for discussion of the pump's action and is the basic standpoint of this paper.

The effect of anaesthetics on the relation between local cerebral blood flow (rCBF) and extracellular glucose was studied in rat striatum. Cerebral blood flow was measured using the hydrogen clearance method and extracellular glucose using an implanted glucose oxidase-based biosensor. Rats were given an intraperitoneal (I.P.) injection of either sodium pentobarbitone (60 mg kg-1) or chloral hydrate (350 mg kg-1). The effect of the I.P. injection, as demonstrated by an I.P. saline injection, was a brief increase in rCBF accompanied by a decrease in glucose. Sodium pentobarbitone produced a decrease in both rCBF and glucose, whereas chloral hydrate caused a decrease in glucose but an increase in rCBF. These findings show a dissociation between rCBF and extracellular glucose and suggest that glucose in the extracellular compartment is not derived directly from the blood vascular compartment.

It has been suggested that the glucose transport system of human erythrocytes contains an arginine shield to prevent the leak of potassium through the transporter. To investigate this suggestion we treated human erythrocytes with the specific arginine reagent 1,2-cyclohexanedione. Under conditions which produce a covalent reaction between arginine and the reagent, a steady leak of potassium occurs. If glucose, maltose or the inhibitor phloretin are present during the reaction the extent of the leak is reduced. These findings support the view that arginines have a role in preventing potassium loss through the glucose transporter.

The process by which free energy from ATP or an ion gradient is coupled to work - either osmotic work, by a membrane pump, or mechanical work, by a molecular motor - as well as the development of force by a molecular motor, can be explained by mechanisms dependent on substrate binding energy. Coupling involves a reaction sequence that combines the driving and driven reactions and that is controlled at switch points where the mobility of the coupling protein and its specificity in binding and catalysis can be abruptly altered; the altered state is an intermediate (or transition state) in the coupled reaction, the unaltered state an intermediate in the uncoupled reaction (slippage). The balance between these states, which is determined by the increase in substrate binding energy in the transformation, decides the relative importance of the coupled and uncoupled paths. On this basis a general expression for switching may be derived: the tightness of coupling is limited by a ratio of substrate dissociation constants before and after a controlling change in state: Ratecoupled/Rateuncoupled [less than or equal to] Kinitial state/Kfinal state.When binding energy is used in this way to distort the conformation of a protein, mechanical work is done; the work is internal but becomes external if the protein is connected to an external load. External work, force F multiplied by distance d, is then limited by a ratio of substrate dissociation constants before and after the conformational change: W = Fd < RT ln(Kinitial state/Kfinal state).The ratios of dissociation constants, estimated from published values of the force exerted by molecular motors, are > 1 × 104 for myosin and > 2 × 104 for kinesin. From the tightness of coupling of the anion exchange carrier of red cells the ratio of constants for coupling is [greater than or equal to] 4 × 104. These increments can be accounted for by conversion of a surface complex to an enclosed complex, as in chelate formation. The work done by an ATP-driven motor is limited by the free energy of ATP hydrolysis, but the force developed is independent of the ATP reaction, and comes from binding energy with either the driving or driven substrate.

The inhibition constant of L-sorbose efflux (Ki(sorbose)) from human erythrocytes for inhibition by D-glucose increases from 5·15 ± 0·89 to 12·24 ± 1·9 mM on cooling from 50¡C to 30¡C; the Ki(sorbose) of D-mannose increases similarly on cooling. The activation energy Ea(sorbose) of net L-sorbose exit from human erythrocytes is 62·9 ± 3·1 kJ mol-1; but in the co-presence of 5 mM D-glucose Ea(sorbose) is reduced to 41·7 ± 1·6 kJ mol-1 (P < 0·005). These data are consistent with the view that when D-glucose binds to the hexose transporter it leads to an activated transporter state which remains transiently activated after glucose dissociates; if L-sorbose binds to this excited state it is more mobile than otherwise and consequently the apparent Ki(sorbose) of D-glucose is raised. Cooling prolongs the decay time of the activated state; hence the Ki(sorbose) of D-glucose rises as temperature is reduced.

A number of mitochondrial carrier systems function both in homologous and in heterologous exchange mode, which, in the case of the phosphate carrier is a homologous Pi--Pi- and a heterologous Pi--OH- exchange. In addition, we showed that mitochondrial carriers, e.g. the aspartate/glutamate and the phosphate carrier, can undergo a functional shift from coupled antiport to uncoupled uniport after modification of cysteine residues. In this transport mode a mixture of carrier- and channel-type properties is observed. To address this question on the molecular level, the phosphate carrier from yeast (S. cerevisiae) mitochondria was expressed in E. coli, solubilized, purified and functionally reconstituted. From three cysteine residues present in the yeast phosphate carrier at positions 28, 134 and 300, only one single cysteine residue (C28) was found responsible for the functional switch from antiport to uniport. Upon replacement by a serine residue, this interconversion was blocked. After incorporation of the carrier into giant liposomes, electrophysiological methods (patch clamp) were applied. Under these conditions, a fourth transport mode of the phosphate carrier was observed, namely an action as anion-selective channel, which could be reversibly blocked by phosphate.

The cellular uptake of glucose catalysed by the facilitated glucose transporter (GLUT) family is further regulated by metabolites and hormones, most importantly by insulin. All of the six isoforms known in this family possess a large cytoplasmic domain of divergent amino acid sequence. A body of evidence indicates that this domain is important for GLUT regulation. Exactly how this domain participates in the regulation, however, is not known. A likely possibility is that a specific cellular protein interacts with GLUT at this domain, and thus modulates the function. This putative, glucose transporter binding protein (GTBP) may be an enzyme, or a non-enzymic adaptor or docking protein. Indeed, we have identified several cellular proteins that bind to the cytoplasmic domain of GLUT proteins; these include glyceraldehyde-3-phosphate dehydrogenase, glucokinase, GTBP70, GTBP85, GTBP28 and L-3-hydroxyacyl-CoA dehydrogenase. Some of these GLUT-GTPB interactions are functionally coupled. Whether any of these interactions actually participates in the insulin-induced GLUT regulation is yet to be determined.