Effect of Natural Mutations on the Function of Human AQP5

In the epidermis, AQP5 is expressed in the keratinocytes of the granular layer. So far, eight AQP5 mutations—W35S, A38E, I45S, N123D, N123Y, I177F, Y178C, and R188C—have been identified in humans, and these natural mutations are linked to autosomal dominant diffuse non‐epidermolytic palmoplantar keratoderma (NEPPK). However, the role of AQP5 mutations in NEPPK is still not clear, and the functional characterization of these natural human mutations has only just begun. Here, we express wild‐type (WT) AQP5 or one of the eight mutants in Xenopus oocytes, and then analyze CO 2 permeability, using microelectrodes to assess the maximal rise in surface pH (DpH S ) as we expose an oocyte to CO 2 /HCO 3 − ; osmotic H 2 O permeability ( P f ), as we transfer the same oocyte to a hypotonic medium and monitor swelling; and surface‐membrane expression of each AQP construct, performing surface biotinylation and then using an AQP5 antibody to detect the construct by western blotting in a batch of at least 10 oocytes. Our controls were day‐matched oocytes injected with H 2 O rather than cRNA encoding one of the nine AQP5 constructs. The channel‐specific DpH S * (or P f *) is the difference between DpH S (or P f ) in an AQP5‐expressing oocyte vs a control. We observed: (1) The gross DpH S * is similar, regardless of whether the oocyte expresses WT AQP5 or any one of the eight mutants. (2) The gross P f * is similar among WT AQP5 and seven of the mutants. Only R188C was different, having only 1/3 the P f * of the WT. (3) Oocytes expressing the eight mutants have surface‐membrane expression levels 2‐4 times higher than oocytes expressing WT AQP5. When we normalize gross DpH S * or gross P f * to surface‐membrane expression, we find that, compared to WT AQP5, all mutants have (1) significantly lower DpH S * per molecule, and (2) significantly lower P f * per molecule. When we divide DpH S * by P f * to obtain a measure of CO 2 /H 2 O selectivity, we find that DpH S */ P f * = ~0.2 for WT AQP5, and obtain similar values for the seven of mutants. However, R188C has a DpH S */ P f * ratio that is 3‐fold higher than WT. Considering that R188 follows the second NPA motif and contributes to the ar/R selectivity filter, we generated several other R188 mutations, including R188S, R188Y, R188D, R188E, R188H, and R188K. The mutants containing an –OH group (R188S and R188Y) have a surface‐membrane expression similar to WT, have 1/3 the P f * of WT, and have no CO 2 conductance (DpH S * = 0). Among the charge‐containing mutants, R188E has the lowest P f *, just slightly greater than that of H 2 O‐injected oocytes. However, R188E still retains almost 1/3 of the normal DpH S *. Thus, the R188E mutant has the highest selectivity for CO 2 over H 2 O, with DpH S */ P f * = ~2, or nearly 12‐fold greater than WT AQP5. The above data show that it is possible to manipulate CO 2 and H 2 O permeability independently, consistent with the idea that, in the AQP5 tetramer, H 2 O moves through the monomeric pore, whereas CO 2 moves predominantly through the central pore. A corollary is that a mutation to the monomeric pore can also affect the central pore.