The effectiveness of the nonmetabolizable second messenger analogue DL-myo-inositol 1,4,5-trisphosphorothioate (IPS3) described by Cooke, A. M., R. Gigg, and B. V. L. Potter, (1987b. Jour. Chem. Soc. Chem. Commun. 1525-1526.) was examined in triads purified from rabbit skeletal muscle. A Ca2+ electrode uptake-release assay was used to determine the size and sensitivity of the IPS3-releasable pool of Ca2+ in isolated triads. Uptake was initiated by 1 mM MgATP, pCa 5.8, pH 7.5 Release was initiated when the free Ca2+ had lowered to pCa approximately 7. We found that 5-25 microM myo-inositol 1,4,5-trisphosphate (IP3), and separately IPS3, consistently released 5-20% of the Ca2+ pool actively loaded into triads. Single channel recording was used to determine if ryanodine receptor Ca2+ release channels were affected by IPS3 at the same myoplasmic Ca2+ and IPS3 concentrations. Open probability of ryanodine receptor Ca2+ release channels was monitored in triads fused to bilayers over long periods (200 s) in the absence and following addition of 30 microM IPS3 to the same channel. At myoplasmic pCa approximately 7, IPS3 had no effect in the absence of MgATP (Po = 0.0094 +/- 0.001 in control and Po = 0.01 +/- 0.006 after IPS3) and slightly increased activity in the presence of 1 mM MgATP (Po = 0.024 +/- 0.03 in control and Po = 0.05 +/- 0.03 after IPS3). Equally small effects were observed at higher myoplasmic Ca2+. The onset of channel activation by IPS3 or IP3 was slow, on the time scale 20-60 s. We suggest that in isolated triads of rabbit skeletal muscle, IP3-induced release of stored Ca2+ is probably not mediated by the opening of Ca2+ release channels.