To clarify the potential role Rh/RhAG and AQP1 proteins in erythrocyte gas transport, NH3 and CO2 transport was measured in erythrocyte ghost membrane vesicles from rare human variants (Rh(null), CO(null),) and knockout mice (homozygous AQP1-/-, Rh-/- and Rhag-/-) exhibiting well-characterized protein defects. Transport was measured from intracellular pH (pHi) changes in a stopped-flow fluorimeter. NH3 transport was measured in chloride-free conditions with ghosts exposed to 20 mM inwardly directed gradients of gluconate salts of ammonium, hydrazine and methylammonium at 15 degrees C. Alkalinization rates of control samples were 6.5+/-0.3, 4.03+/-0.17, 0.95+/-0.08 s(-1) for each solute, respectively, but were significantly reduced for Rh(null) and CO(null) samples that are deficient in RhAG and AQP1 proteins, respectively. Alkalinization rates of Rh(null) ghosts were about 60%, 83% and 94% lower than that in control ghosts, respectively, for each solute. In CO(null) ghosts, the lack of AQP1 resulted in about 30% reduction of the alkalinization rates as compared to controls, but the transport selectivity of RhAG for the three solutes was preserved. Similar observations were made with ghosts from KO mice Rhag-/- and AQP1-/-. These results confirm the major contribution of RhAG/Rhag in the NH3 conductance of erythrocytes and suggest that the reduction of transport rates in the absence of AQP1 would be better explained by a direct or indirect effect on RhAG/Rhag-mediated transport. When ghosts were preloaded with carbonic anhydrase and exposed to a 25 mM CO2/HCO3- gradient at 6 degrees C, an extremely rapid kinetics of acidification corresponding to CO2 influx was observed. The rate constants were not significantly different between controls and human variants (125+/-6 s(-1)), or between wild-type and KO mice, suggesting no major role of RhAG or AQP1 in CO2 transport, at least in our experimental conditions.