We have compared the electron-transfer kinetics in reaction centers (RCs) cooled in the dark with those cooled under illumination (i.e., in the charge-separated state). Large differences between the two cases were observed. We interpreted these findings in terms of light-induced structural changes. The kinetics of charge recombination D+QA-----DQA in RCs containing one quinone were modeled in terms of a distribution of donor-acceptor electron-transfer distances. For RCs cooled under illumination the distribution broadened and shifted to larger distances compared to the distribution for RCs cooled in the dark. The model accounts for the nonexponential decay observed at low temperatures [McElroy, J. D., Mauzerall, D. C., & Feher, G. (1974) Biochim. Biophys. Acta 333, 261-277; Morrison, L.E., & Loach, P.A. (1978) Photochem. Photobiol. 27, 751-757]. A possible physiological role of the structural changes is an enhanced charge stabilization. For RCs with two quinones, the recombination kinetics D+QAQB-----DQAQB were found to be strongly temperature dependent. This was interpreted in terms of temperature-dependent transitions between structural states [Agmon, N., & Hopfield, J.J. (1983) J. Chem. Phys. 78, 6947-6959]. This interpretation requires that these transitions occur at cryogenic temperatures on a time scale t greater than or approximately 10(3) s. The electron transfer from QA- to QB was found to not take place in RCs cooled in the dark (tau ABdark greater than 10(-1) s). In RCs cooled under illumination, we found tau ABlight less than 10(-3) s. We suggest the possibility that the drastic decrease in tau AB observed in RCs cooled under illumination is due to the trapping of a proton near QB-.