The effects of all-zinc metallothionein (Zn-metallothionein) and predominantly cadmium metallothionein (Cd/Zn-metallothionein) on free radical lipid peroxidation have been investigated, using erythrocyte ghosts as the test system. When treated with xanthine and xanthine oxidase, Zn-metallothionein and Cd/Zn-metallothionein underwent thiolate group oxidation and metal ion release that was catalase-inhibitable, but superoxide dismutase-non-inhibitable. Similar treatment in the presence of ghosts and added Fe(III) resulted in metallothionein oxidation that was significantly inhibited by superoxide dismutase. Ghosts incubated with xanthine/xanthine oxidase/Fe(III) underwent H2O2- and O2--dependent lipid peroxidation, as measured by thiobarbituric acid reactivity. Neither type of metallothionein had any effect on xanthine oxidase activity, but both strongly inhibited lipid peroxidation when added to the membranes concurrently with xanthine/xanthine oxidase/iron. This inhibition was far greater and more sustained than that caused by dithiothreitol at a concentration equivalent to that of metallothionein thiolate. Significant protection was also afforded when ghosts plus Cd/Zn-metallothionein or Zn-metallothionein were preincubated with H2O2 and Fe(III), and then subjected to vigorous peroxidation by the addition of xanthine and xanthine oxidase. These results could be mimicked by using Cd(II) or Zn(II) alone. Previous studies suggested that Zn(II) inhibits xanthine/xanthine oxidase/iron-driven lipid peroxidation in ghosts by interfering with iron binding and redox cycling. Therefore, the primary determinant of metallothionein protection appears to be metal release and subsequent uptake by the membranes. These results have important implications concerning the antioxidant role of metallothionein, a protein known to be induced by various prooxidant conditions.