The flash-induced proton-binding behavior of reaction centers from Rhodobacter sphaeroides was examined, over a wide range of pH, as a function of the one-electron redox states of the acceptor quinones ( Q A Q − A and Q B Q − B ) and the primary donor ( P + P ). Below about pH 9, the P +Q − states (P +Q − A and P +Q − B), generated in the absence of exogenous electron donor to P +, fail to take up protons stoichiometrically, as established in the previous paper (Maróti, P. and Wraight, C.A. (1988) Biochim. Biophys. Acta 934, 314–328). When P + is rereduced by a donor, to yield the state PQ −, proton binding is enhanced in this lower pH range. In the case of Q B-reconstituted reaction centers, the net proton binding stoichiometry ( H + P + ) for PQ − B is about 0.85, between pH 7 and pH 9, approaching the stoichiometric value expected for a pH-dependent redox midpoint potential of Q B Q − B . The shortfall from H + P + = 1.0 can be accounted for by the involvement of four protonatable groups with different p K values depending on the redox state of Q B. The p K shifts vary from 0.3 to 1.5 pH units, with the lower p K groups exhibiting the smaller p K shifts. The enhancement of proton binding, associated with the rereduction of P +, is interpreted as a response of the same groups to the redox state of P + P , with the lower p K groups exhibiting the largest p K shifts - up to 1.0 pH unit. A similar general behavior is seen for reaction centers lacking Q B, or in the presence of terbutryn, a competitive inhibitor of Q B-binding. Quantitatively, the rereduction of P + does not restore such high levels of H + binding for PQ − A or PQ − A + terbutryn ( H + P + ⩽ 0.5 at all pH values), but the behavior can be similarly accounted for by four protonatable groups that are somewhat less responsive to the redox states of Q A and P. The p K values for the three different acceptor configurations (Q B, Q A and Q A + terbutryn ) are similar but not identical, and depend on the redox states of the primary donor and the acceptor quinones, and on the occupancy of the Q B-binding site. The p K values are discussed in terms of possible structural determinants of the quinone binding sites. The data define protonation networks for the reaction center states PQ, P +Q − and PQ −, and allow one to deduce the properties of a fourth state: P +Q. The derived p K values predict the occurrence of H + release from the state P +Q, at low pH, and this was confirmed by using ferricyanide to reoxidize Q − following a flash to generate P +Q. The proposed protonation scheme allows the calculation of the pH dependence of the one-electron transfer equilibrium between P +Q − AQ B and P +Q AQ − B. This agrees well with the measured value, derived from the kinetics of charge recombination. However, the p K changes, derived from the enhanced proton binding that accompanies rereduction of P +, give rise to a substantial discrepancy between the calculated and measured values for the PQ − AQ B ↔ PQ A Q − B equilibrium. The pH dependences of the redox midpoint potentials ( E m) of Q A Q − A , Q B Q − B and P + P are also calculated. Good agreement between calculated and measured values is obtained for P + P , and between the calculated value for Q B Q − B and that expected from studies on chromatophores. However, the calculated pH dependence of E m ( Q A Q − A ) is at variance with that measured in isolated reaction centers or chromatophores. These discrepancies are discussed but not resolved.