The photoproduct of carbon monoxide myoglobin generated at 4 K and lower has a resonance Raman spectrum characteristic of a high-spin heme but in which the high-frequency core size-sensitive lines are at lower frequency than those in the deoxy preparation. Such differences are not detected in the photoproduct generated at higher temperatures (50 K) or in that generated at room temperature with 10-nsec pulses. The data indicate that at the low temperature (4 K), the heme in the photoproduct is not fully relaxed, and from the data we conclude that the photoproduct has an expanded porphyrin core. We infer that the core size exceeds that in deoxymyoglobin because the rigid protein prevents the highspin iron atom from moving to its full out-of-plane displacement at the very low temperatures.To gain an understanding of the basic mechanisms of ligand association in heme proteins, the elementary steps in the ligand binding and release process must be uncovered. An approach to the elucidation of these steps is to photodissociate a bound ligand and follow the relaxation of the deoxy heme that has been generated. At room temperature the evolution of this photoproduct occurs over a wide temporal scale ranging from the subpicosecond absorption of the photon (1) to structural relaxation of the protein residues, which may occur on a time scale of 10s of microseconds (2,3). The specific relaxation pathway of the photoproduct is very complex and depends on the protein as well as the ligand.An alternate approach in the identification of the intermediate structures in the photoproduct relaxation is through the use of cryogenic temperatures, where metastable states may be isolated (4). By this method changes in the properties of photodissociated heme proteins were detected many years ago with optical absorption techniques (4-6), and studies of the metastable species have been since extended with many other techniques. The relationship between the structure, as inferred from the resonance Raman spectra, of the lowtemperature photoproduct and the transient photoproduct generated at 300 K was discussed recently, and in carbon monoxide hemoglobin, HbCO, it was observed that the differences in the Raman spectra between the stabilized photoproduct and the deoxy preparation at 80 K were the same as those obtained transiently at room temperature with 10-nsec pulses (7). These findings led us to investigate the low-temperature (1.6-100 K) photodissociation of carbon monoxide myoglobin, MbCO, with resonance Raman scattering to determine the properties of the unrelaxed photoproduct and its relationship to the room-temperature transient photoproduct. The experiments were carried out by freezing CO-bound Mb to low-temperature and then directing a laser beam on the sample. The laser photodissociates the CO, resulting in a deoxy heme trapped in the frozen protein matrix. At the lowest temperatures we find significant differences in the resonance Raman spectra ofthe photoproduct and the equilibrium deoxy species. We interpret these results...