We all know that glucose occupies a central position in cellular carbon and energy metabolism. However, glucose has another little-known function; it also works as a signalling molecule regulating food intake, body weight, blood glucose levels and gastrointestinal reflexes. Which poses the question, are eukaryotic cells able to sense external glucose concentrations directly?

Apart from a few gustatory chemoreceptors, the only proteins that have so far been suggested as possible glucose sensors are two glucose transporter homologues found in yeast, both of which transduce the external glucose signal to control gene expression in the cell. However, little else was known about these sensor proteins until scientists from Los Angeles and Würzburg recently discovered a novel human glucose sensor hiding in a large family of Na⁺/glucose cotransporters called SLC5.

Ernest Wright and Hermann Koepsell had been interested for many years in these types of glucose transporters and knew that the SLC5 cotransporters usually facilitate glucose uptake by coupled transport of sodium into the cell. Therefore, it was not surprising that they would examine a previously undescribed member of the SLC5 family, discovered by the human genome project in 1999, called human sodium coupled glucose transporter 3 (hSGLT3). After cloning the protein, the team began searching for hSGLT3 expression in human tissue, finding it in the plasma membranes of skeletal muscles, at the neuromuscular junction and also in the cholinergic neurons of the small intestine. But what was the protein's physiological role in cholinergic neurons and skeletal muscles?

To answer this question, the team examined the functional properties of the apparent glucose transporter. Using the Xenopus laevis oocyte expression system, they first injected hSGLT3 cRNA into the oocytes and then tested for evidence that the RNA was translated to produce protein that was inserted into the plasma membrane. Next, they tested whether the modified oocytes could transport radiolabelled sugars across the cell's membrane with the human glucose transporter. However, they were unable to detect any sugar transport. hSGLT3 seemed to be incapable of transporting sugar.

At first, Wright and his colleagues must have assumed that hSGLT3 was nonfunctional, either due to misfolding or some other artefact. But when they studied the electrical properties of hSGLT3 in Xenopus oocytes they found depolarization of the membrane; the cotransporter was able to transport sodium in response to glucose. hSGLT3 was functional, and the membrane depolarization turned out to be reversible and specific for d-glucose. And when the team looked at the Na⁺-dependent membrane potential, it was directly proportional to the glucose concentration. All of these properties make hSGLT3 an ideal glucose sensor. Rather than modulating gene expression as the yeast sensors do, hSGLT3 is a completely novel type of glucose sensor that resembles the behaviour of chemoreceptors and signals glucose levels through membrane depolarisation.

Thus, hSGLT3 may influence both gastrointestinal cholinergic nerve activity and skeletal muscle in response to varying extracellular glucose concentrations by modulating membrane potential. This new and unexpected insight into glucose sensing may also serve as a warning not to rely exclusively on sequence homologies when assigning functions to unknown proteins. Although hSGLT3 looked just like any other transporter, in reality it turned out to have a completely unexpected function.