1. Extracellular ATP (EC50=146+/-57 microM) and various ATP analogues activated cyclic AMP production in undifferentiated HL-60 cells. 2. The order of agonist potency was: ATPgammaS (adenosine 5'-O-[3-thiotriphosphate]) > or = BzATP (2'&3'O-(4-benzoylbenzoyl)-adenosine-5'-triphosphate) > or = dATP > ATP. The following agonists (in order of effectiveness at 1 mM) were all less effective than ATP at concentrations up to 1 mM: beta,gamma methylene ATP > or = 2-methylthioATP > ADP > or = Ap4A (P1, P4-di(adenosine-5') tetraphosphate) > or = Adenosine > UTP. The poor response to UTP indicates that P2Y2 receptors are not responsible for ATP-dependent activation of adenylyl cyclase. 3. Several thiophosphorylated analogs of ATP were more potent activators of cyclic AMP production than ATP. Of these, ATPgammaS (EC50=30.4+/-6.9 microM) was a full agonist. However, adenosine 5'-O-[1-thiotriphosphate] (ATPalphaS; EC50=45+/-15 microM) and adenosine 5'-O-[2-thiodiphosphate] (ADPbetaS; EC50=33.3+/-5.0 microM) were partial agonists. 4. ADPbetaS (IC50=146+/-32 microM) and adenosine 5'-O-thiomonophosphate (AMPS; IC50=343+/-142 microM) inhibited cyclic AMP production by a submaximal concentration of ATP (100 microM). Consistent with its partial agonist activity, ADPbetaS was estimated to maximally suppress ATP-induced cyclic AMP production by about 65%. AMPS has not been previously reported to inhibit P2 receptors. 5. The broad spectrum P2 receptor antagonist, suramin (500 microM), abolished ATP-stimulated cyclic AMP production by HL-60 cells but the adenosine receptor antagonists xanthine amine congener (XAC; 20 microM) and 8-sulpho-phenyltheophylline (8-SPT; 100 microM) were without effect. 6. Extracellular ATP also activated protein kinase A (PK-A) consistent with previous findings that PK-A activation is involved in ATP-induced differentiation of HL-60 cells (Jiang et al., 1997). 7. Taken together, the data indicate the presence of a novel cyclic AMP-linked P2 receptor on undifferentiated HL-60 cells.