The molecular determinants of a Ca(2+) spark, those events that determine the sudden opening and closing of a small number of ryanodine receptor (RyR) channels limiting Ca(2+) release to a few milliseconds, are unknown. As a first step we investigated which of two RyR isoforms present in mammalian embryonic skeletal muscle, RyR type 1(RyR-1) or RyR type 3 (RyR-3) has the ability to generate Ca(2+) sparks. Their separate contributions were investigated in intercostal muscle cells of RyR-1 null and RyR-3 null mouse embryos. A comparison of Ca(2+) spark parameters of RyR-1 null versus RyR-3 null cells measured at rest with fluo-3 showed that neither the peak fluorescence intensity (DeltaF/F(o) = 1.25 +/- 0.7 vs. 1.55 +/- 0.6), spatial width at half-max intensity (FWHM = 2.7 +/- 1.2 vs. 2.6 +/- 0.6 microm), nor the duration at half-max intensity (FTHM = 45 +/- 49 vs. 43 +/- 25 ms) was significantly different. Sensitivity to caffeine (0.1 mM) was remarkably different, with sparks in RyR-1 null myotubes becoming brighter and longer in duration, whereas those in RyR-3 null cells remained unchanged. Controls performed in double RyR-1/RyR-3 null cells obtained by mice breeding showed that sparks were not observed in the absence of both isoforms in >150 cells imaged. In conclusion, 1) RyR-1 and RyR-3 appear to be the only intracellular Ca(2+) channels that participate in Ca(2+) spark activity in embryonic skeletal muscle; 2) except in their responsiveness to caffeine, both isoforms have the ability to produce Ca(2+) sparks with nearly identical properties, so it is rather unlikely that a single RyR isoform, when others are also present, would be responsible for Ca(2+) sparks; and 3) because RyR-1 null cells are excitation-contraction (EC) uncoupled and RyR-3 null cells exhibit a normal phenotype, Ca(2+) sparks result from the inherent activity of small clusters of RyRs regardless of the participation of these RyRs in EC coupling.