We compared acceptor-protein specificities on the formation of ADP-ribose.acceptor adducts by arginine-specific ADP-ribosyltransferase (EC 2.4.2.31) purified from rabbit skeletal muscle sarcoplasmic reticulum (SR) with those of the enzyme purified from chicken peripheral polymorphonuclear cells (heterophils). Major differences are as follows: (1), p33 and beta/gamma-actin, preferential endogenous acceptor proteins for the modification by the heterophil enzyme (Mishima, K., Terashima, M., Obara, S., Yamada, K., Imai, K and Shimoyama, M. (1991) J. Biochem. 110, 388-394 and Terashima, M., Mishima, K., Yamada, K., Tsuchiya, M., Wakutani, T. and Shimoyama, M. (1992) Eur. J. Biochem. 204, 305-311) were not modified by the SR enzyme. (2), The modification of p33 by the heterophil enzyme was enhanced by addition of polyanions such as DNA while the protein did not function as acceptor for modification by the SR enzyme even in the presence of DNA. (3), To ADP-ribosylate endogenous substrate Ca(2+)-transporting ATPase (EC 3.6.1.38) of rabbit skeletal muscle SR, the SR ADP-ribosyltransferase required polycations such as poly(L-lysine), whereas the heterophil enzyme modified the ATPase in the absence of poly(L-lysine). These results suggest that vertebrate arginine-specific ADP-ribosyltransferase prefers its own acceptor protein for the modification. Some other properties of the SR and the heterophil ADP-ribosyltransferases were also compared.