All mammalian diacylglycerol kinase (DGK) isoenzymes so far cloned consist of four conserved regions, namely, C1, C2 (tandem EF-hand structures), C3 (tandem cysteine-rich zinc finger sequences) and the C-terminal C4 domains. To determine the catalytic domain we expressed in COS-7 cells various truncation mutants of pig DGK alpha and assessed their enzyme activities. We found that the C4 domain lacking the whole N-terminal region including the zinc fingers possessed DGK activity that was dependent on the concentrations of diacylglycerol and ATP very similarly, as did the wild-type DGK alpha. Furthermore the DGK activity of the wild-type DGK and that expressed by the C4 domain were similarly activated by anionic amphiphiles such as phosphatidylserine, phosphatidylinositol and deoxycholate. It was also shown that a DGK mutant consisting of the zinc fingers and the C4 domain has enzymological properties very similar to those expressed by the C4 domain alone. We also confirmed that the intact DGKs alpha, beta and gamma expressed in COS-7 cells displayed no detectable phorbol ester binding. These results show that the C4 domain of DGK is the catalytic region that is responsible for the enzyme activities sensitive to different activators. We cannot exclude the possibility that the N-terminal portion including the zinc fingers can still interact with diacylglycerol and activators without affecting the enzyme activity measured in vitro. However, it is quite likely that the DGK zinc fingers do not serve as diacylglycerol-binding sites, in contrast with those present in other proteins such as protein kinases C and n-chimaerin. Site-directed mutagenesis of all six putative ATP binding sites (Lys248, Lys383, Lys395, Lys483, Lys492, and Lys554) did not significantly affect the enzyme activity. We therefore suggest that DGK does not contain a typical P-loop of ATP binding sites.