The water permeability of human red blood cell (RBC) membrane has been monitored by a doping nuclear magnetic resonance (NMR) technique on intact cells and resealed ghosts following exposure to various sulfhydryl-reacting (SH) reagents and proteolytic enzymes. The main conclusions are the following: (i) When appropriate conditions for exposure of erythrocytes or ghosts to mercury-containing SH reagents (concentration, temperature and duration of incubation) were found, the maximal inhibition of water diffusion could be obtained with all mercurials (including HgCl2 and mersalyl that failed to show their inhibitory action on RBC water permeability in some investigations). While previous studies claimed that long incubation times are required for the development of maximal inhibition of water diffusion by mercurials, the present results show that it can be induced in a much shorter time (5-15 min at 37 degrees C) if relatively high concentrations of PCMBS (2-4 mM) are used and no washings of the inhibitor are performed after incubation. Higher than optimal concentrations of mercurials and/or longer incubation times result in lower values of inhibition, sometimes a loss of inhibition, or can even lead to higher values of permeability compared to control RBCs. (ii) The conditions for inhibition by mercurials are drastically changed by preincubation of erythrocytes with noninhibitory SH reagents (such as NEM or IAM) or by exposure to proteolytic enzymes. If the cells are digested with papain, the duration of incubation with PCMBS should be decreased in order for inhibition to occur. This explains the lack of inhibition reported previously, when a relatively long duration of incubation with PCMBS was used subsequent to papain digestion.(ABSTRACT TRUNCATED AT 250 WORDS)