Aims: 

The pathways taken by water through epithelia are not well understood. According to the "molecular water pumping" hypothesis, water transport is not due to osmosis. It occurs when a cotransporter (e.g. the potassium-chloride-cotransporter, KCC1, or the sodium-glucose cotransporter, SGLT1) picks up water molecules from one side of the membrane and a conformational change releases them to the other side. An alternative hypothesis envisions that water passively follows the substrate through water channels. To distinguish between both hypotheses we monitored water permeability Pf of epithelial monolayers. Local osmosis requires high Pf values. In contrast, efficient water pumping against an osmotic gradient calls for rather low Pf values.

Methods: 

We derived the osmotic water flux across a single copy of SGLT1 from spatially resolved fluorescence correlation spectroscopy measurements of (i) SGLT1 concentration in the membrane of stably transfected MDCK cells grown on a permeable support and (ii) the dilution of an aqueous reporter dye in the immediate vicinity of the epithelial mono-layer.

Results: 

SGLT1 and aquaporin-1 have equally high single water channel permeabilities. In contrast, renal KCC1 does not allow osmotic water transport. However, the neighbouring aquaporin-2 ensures a rather high Pf of inner medulla collecting duct (when stimulated with vasopressin). Inhibition of KCC1 reduced Pf of primary cultured cell monolayers by roughly 1/3. The inhibitory effect of furosemid or DIOA on Pf was reversed by the K⁺-H⁺ exchanger nigericin indicating that KCC1 affects water transport solely by K⁺ extrusion.

Conclusion: 

Neither SGLT1 nor KCC1 perform secondary active water transport. Instead, solute and transcellular solvent transport must be osmotically coupled.