A recent C-13 NMR experiment (Smith et al. Nature Struct. Biol. 1996. 3, 946-950) on the Asp 25-Asp25 ' dyad in pepstatin A/HIV-1 protease measured two separate resonance lines. which were interpreted as being a singly protonated dyad. We address this issue by performing ab initio molecular dynamics calculations on models for this site accompanied by calculations of C-13 NMR chemical shifts and isotopic shifts. We find that already on the picosecond time-scale the model proposed by Smith et al. is not stable and evolves toward a different monoprotonated form whose NMR pattern differs from the experimental one. We suggest, instead a different protonation state in which both aspartic groups are protonated. Despite the symmetric protonation state. the calculated C-13 NMR properties are in good agreement with the experiment. We rationalize this result using a simple valence bond model, which explains the chemical inequality of the two C sites. The model calculations, together with our calculations on the complex, allow also the rationalization of C-13 NMR properties on other HIV-1 PR/inhibitor complexes. Both putative binding of the substrate to the free enzyme, which has the dyad singly protonated (Piana. S.; Carloni, P. Proteins: Struct., Funct., Genet. 2000. 39, 26-36), and pepstatin A binding to the diprotonated form are consistent with the inverse solvent isotope effect on the onset of inhibition of pepsin by pepstatin and the kinetic iso-mechanism proposed for aspartic proteases.