Sarcoplasmic reticulum (SR) and plasma membranes from canine left ventricle were used to evaluate the presence of the enzyme CDPdiglyceride-inositol transferase in these membranes. (K+,-Ca2+)-ATPase activity, a marker for SR, was 79.2 +/- 5.0 (SE) and 11.2 +/- 2.0 mumol.mg-1.h-1 in SR and plasma membrane preparations, respectively, and (Na+,K+)-ATPase activity, a marker for plasma membranes, was 5.6 +/- 1.2 and 99.2 +/- 8.0 mumol.mg-1.h-1, respectively. Contamination of SR and plasma membrane preparations by mitochondria was estimated to be 2% and 8%, respectively, and by Golgi membranes, 0.9% and 1.8%, respectively. Transferase activity, measured at pH 6.8, was 1.32 +/- 0.04 (SE) and 0.28 +/- 0.04 nmol of [3H]phosphatidylinositol ([3H]PtdIns).mg-1.min-1 in three SR and plasma membrane preparations, respectively. The transferase activity detected in the plasma membrane preparation could be accounted for largely, but not entirely, by contaminating SR membranes. The pH optimum for the SR transferase activity was between 8.0 and 9.0; little or no activity was detectable at pH 6.3 and 5.5, the lowest pH tested. Ca2+ inhibited the enzyme, half-maximal inhibition occurring at about 10 microM Ca2+; removal of the Ca2+ by addition of ethylene glycol bis(beta-aminoethyl ether)-N,N,N',N'-tetraacetic acid restored activity. No loss of [3H]PtdIns could be detected when membranes were incubated in the presence or absence of Ca2+. The Ca2+ inhibition of the transferase was noncompetitive with respect to CDP-dipalmitin while that with respect to myo-inositol was slightly noncompetitive at low [Ca2+] and became uncompetitive at higher [Ca2+].(ABSTRACT TRUNCATED AT 250 WORDS)