The specific binding of [125I]T3 and [125I]rT3 to nuclear extracts of rat and pork liver was studied. The Ka values of the binding of T3 and rT3 to nuclear extracts were similar (0.65 +/- 0.12 vs. 0.68 +/- 0.10 x 10(9) M-1; P = NS), but the maximal binding capacity for T3 was smaller than that for rT3 (333 +/- 36 vs. 1209 +/- 338 fmol/mg DNA; P less than 0.05). The amount of nonradioactive T4 required to cause a 50% displacement of [125I]T3 from the nuclear binding sites was, on a molar basis, 27 times greater than that of T3 in the case of pork liver and 110 times greater than that of T3 in the case of rat liver. Similarly, rT3-binding sites were highly specific. Relative to rT3, 16- and a 100-fold greater molar excesses of T4 were required to cause a 50% displacement of [125I[rT3 from the nuclear binding sites in pork and rat liver, respectively. Similarly, 700--2150 times more rT3 than T3 was required to obtain a 50% displacement of [125I]T3; in contrast, 50--250 times more T3 than rT3 was needed to displace 50% specifically bound [125I]rT3. Reduced glutathione and other sulfhydryl (SH)-reducing agents increased, whereas oxidized glutathione and SH-oxidizing or -binding agents decreased the binding of [125I]T3 and [125I]rT3 to nuclear extracts, with one exception. Dithiothreitol, a potent SH-reducing agent, reduced the binding of rT3 whereas it increased the binding of T3 to the nuclear binding sites. A 48-h fast was associated with a significant reduction in the maximal binding capacity of nuclear extracts for T3 (control vs. fasting rats, 589 +/- 92 vs. 339 +/- 50 fmol T3/mg DNA; P less than 0.05) and rT3 (1620 +/- 162 vs. 693 +/- 102 fmol rT3/mg DNA; P less than 0.005) without changes in affinity. The data suggest that 1) there exist specific high affinity, low capacity nuclear rT3-binding sites in rat and pork liver which are distinct from the nuclear T3 receptors, and 2) rT3- and T3-binding sites are reduced in parallel during starvation.