Isoelectric focusing (IEF) combined with Western blotting detected the modification of carbonic anhydrase III (CA III) by S-thiolation (glutathione-protein mixed disulfide formation) when primary cultures of rat hepatocytes were subjected to oxidative stress. Carbonic anhydrase III constitutes 5-8% of the soluble protein in male rate liver, and is the major hepatic protein modified by S-thiolation in males. As little as 10 ng of CA III could be clearly identified on Western blots of IEF gels. With this method, two S-thiolated forms of CA III, one partially- and one fully S-thiolated forms, could be detected during treatment with t-butyl hydroperoxide (t-BuOOH) or menadione. S-thiolation of CA III by t-BuOOH reached a maximum at 3 min and decreased to control levels by 15 min, whereas menadione gradually increased CA III S-thiolation. These results support the concepts that CA III has two reactive sulfhydryl groups and that S-thiolation can occur in different ways. Oxidation of GSH to GSSG was coincident with CA III S-thiolation. A similar CA III S-thiolation pattern was observed in female hepatocytes under oxidative stress, although CA III levels in females were 15-fold less than in males. CA III S-thiolation was also detected in rat muscle and heart and the blotting pattern of S-thiolated proteins was similar to the pattern in liver;Se-deficient hepatocytes, in which GSH peroxidase activity was inhibited and thus GSSG formation was blocked, were used to determine whether S-thiolation occurred largely by mixed-disulfide exchange, a process requiring GSSG. In Se-deficient hepatocytes, CA III S-thiolation did occur after the treatment of t-BuOOH or menadione but to a much lesser extent than in Se-adequate cells. When intracellular GSH was depleted to 5% of the control level by buthionine sulfoximine (BSO), no detectable change of CA III S-thiolation occurred in the treatment with t-BuOOH or menadione;In an in vivo study, no difference in CA III S-thiolation could be detected between Se-adequate and Se-deficient male rates 8 h after the administration of 10 mg/kg of paraquat with or without prior BSO treatment. Intracellular GSSG level was not increased by paraquat or decreased by Se-deficiency. Lipid peroxidation by paraquat was potentiated in Se-deficient rats. Similarly, GSH depletion by BSO also sensitized Se-deficient rats to paraquat-induced lipid peroxidation. The total amount of CA III protein and the amount of nondethiolatable CA III was significantly less in Se-deficient rats, by 35%. The proportion of nondethiolatable CA III was 15% less in the Se-deficient rats compared with Se-adequate rats (23% vs 27% of CA III was modified form, respectively). The significance of this protein modification to metabolic regulation during oxidative stress deserves further study.