Ejaculated ram sperm were demembranated with Triton X-100, separated from the detergent-soluble matrix, and reactivated [San Agustin and Witman (1993): Cell Motil. Cytoskeleton 24:264–273]. The percent motility of models prepared from freshly washed sperm was comparable to that of the washed sample before demembranation, regardless of whether cAMP was included in the reactivation medium. However, demembranated models derived from aging or metabolically inhibited sperm exhibited a lower percent reactivation and required cAMP to attain the level of motility of freshly washed sperm. Cyclic AMP was ∼100 times more effective than cGMP. The requirement for cAMP could be bypassed by addition of porcine heart cAMP-dependent protein kinase (PKA) catalytic subunit to the reactivation medium, demonstrating that cAMP was acting via PKA. The cAMP stimulation of reactivation was not affected by inclusion of the PKA inhibitor PKI(5–24) in the reactivation medium, but was decreased when the models were preincubated with PKI(5–24) prior to reactivation. The cytosol-free models retained >90% of the sperm PKA activity; therefore, the PKA appears to be anchored to internal sperm structures. This PKA could not be extracted by cAMP or Triton X-100 alone, but only by cAMP and Triton X-100 in combination. We conclude that cAMP-dependent protein phosphorylation is critical for sperm motility, but that the essential protein phosphate sites turn over slowly under our reactivation conditions, so that the cAMP requirement is apparent only in models prepared from sperm having a low internal ATP or cAMP content. Interestingly, reactivation was rapidly blocked by the peptide arg-lys-arg-ala-arg-lys-glu, which has been reported to be a selective inhibitor of cGMP-dependent protein kinase. © 1994 Wiley-Liss, Inc.