Cytochrome aco purified from an alkalophilic bacterium grown at pH 10 contains hemes a, b, and c as prosthetic groups, and their redox behavior was examined by using stopped-flow and rapid-scan techniques. Under anaerobic conditions the reduction of both heme a and c moieties with dithionite proceeded exponentially but with different rates, usually the former being reduced about 4 times faster than the latter. The reduction of protoheme was much slower, and a time-difference spectrum for this species was of a high spin type with absorption peaks at 433, 557, and 609 nm. Only the protoheme combined with CO, fulfilling the criteria for cytochrome o. Potentiometric titrations determined a midpoint potential of c heme to be 95 mV at pH 7.0 and 25 degrees C and suggested the presence of two forms of a heme with midpoint potentials of 250 and 323 mV. Cytochrome aco utilizes ascorbate plus N,N,N',N'-tetramethyl-p-phenylenediamine (TMPD) to reduce oxygen relatively rapidly without added cytochrome c (Qureshi, M. H., Yumoto, I., Fujiwara, T., Fukumori, Y., Yamanaka, T. (1990) J. Biochem. 107, 480-485). During the steady state, however, heme a stayed almost fully reduced in contrast to a partial reduction of heme c. Even after exhaustion of the dissolved oxygen the extent of reduction of heme c was 60-70% that attained by the dithionite reduction. When ascorbate plus TMPD-reduced cytochrome aco was exposed to oxygen the reduced heme c was oxidized rapidly whereas the oxidation of reduced a heme was negligibly slow. The full reduction of heme a during the steady state and its extremely slow oxidation rendered participation of heme a in the oxidase reaction less likely. A novel peak appearing transiently around 567 nm during the reaction was tentatively ascribed to an intermediate form of protoheme, or o heme, which was thus supposed to react directly with molecular oxygen. These results suggest strongly that the main electron transfer pathway would be c----o----oxygen. A possible role of a in regulating the electron flow through the main pathway and its functional relationship to a heme in the aa3-type cytochrome oxidase were discussed.