D-type cyclin-dependent kinase activities have not so far been detected in mammalian cells. Lysis of rodent fibroblasts, mouse macrophages, or myeloid cells with Tween 20 followed by precipitation with antibodies to cyclins Dl, D2, and D3 or to their major catalytic partner, cyclin-dependent kinase 4 (cdk4), yielded kinase activities in immune complexes which readily phosphorylated the retinoblastoma protein (pRb) but not histone H1 or casein. Virtually all cyclin Dl-dependent kinase activity in proliferating macrophages and fibroblasts could be attributed to cdk4. When quiescent cells were stimulated by growth factors to enter the cell cycle, cyclin Dl-dependent kinase activity was first detected in mid G₁, reached a maximum near the G₁/S transition, and remained elevated in proliferating cells. The rate of appearance of kinase activity during G₁ phase lagged significantly behind cyclin induction and correlated with the more delayed accumulation of cdk4 and formation of cyclin Dl-cdk4 complexes. Thus, cyclin D1-associated kinase activity was not detected during the G₀-to-G₁ transition, which occurs within the first few hours following growth factor stimulation. Rodent fibroblasts engineered to constitutively overexpress either cyclin Dl alone or cyclin D3 together with cdk4 exhibited greatly elevated cyclin D-dependent kinase activity, which remained absent in quiescent cells but rose to supraphysi-ologic levels as cells progressed through G₁. Therefore, despite continued enforced overproduction of cyclins and cdk4, the assembly of cyclin D-cdk4 complexes and the appearance of their kinase activities remained dependent upon serum stimulation, indicating that upstream regulators must govern formation of the active enzymes.