The C-cluster of carbon monoxide dehydrogenase (CODH) appears to be the active site for the oxidation of CO to CO2. We have studied with EPR and Mossbauer spectroscopy the enzymes from Rhodospirillum rubrum (CODHRr; similar to 8 Fe atoms and 1 Ni atom per alpha) and Clostridium themzoaceticum (CODHCt; similar to 12 Fe atoms and 2 Ni atoms per alpha beta). The study of CODHRr offers two advantages. First, the enzyme lacks the A-cluster responsible for the synthase activity of CODHCt. Second, a Ni-deficient protein (Ni-deficient CODHRr) containing all Fe components of the holoenzyme can be isolated. The holoenzymes of both species can be prepared in a state for which the C-cluster exhibits the so-called g(av) = 1.82 EPR signal (C-red1); the spectra of Ni-deficient CODHRr do not exhibit this signal. Our results are as follows: The Mossbauer data show that all iron atoms of Ni-deficient CODHRr belong to two [Fe4S4](1+/2+) clusters. The so-called B-cluster, which functions in electron transfer, is diamagnetic in the [Fe4S4](2+) state, B-ox, and exhibits an S = 1/2 (g = 1.94) EPR signal in the [Fe4S4](+) state, B-red. The spectroscopic properties of the B-cluster are the same in Ni-deficient, holo-CODHRr and CODHCt. The precursor to the C-cluster of Ni-deficient CODHRr, labeled C*, is diamagnetic in the [Fe4S4](2+) state, but has an S = 3/2 spin in the [Fe4S4](+) form. Upon incorporation of Ni, the properties of the C*-cluster change substantially. At E(m)(,) = -110 mV, the C-cluster undergoes a 1-electron reduction from the oxidized state, C-ox, to the reduced state, C-red1, which exhibits the g(av) = 1.82 EPR signal. A study of a sample poised at -300 mV shows that this signal originates from an S = 1/2 [Fe4S4](+) cluster. In this state, the cluster has a distinct subsite, ferrous component II (FCII), having Delta E(Q) = 2.82 mm/s and delta = 0.82 mm/s; these parameters suggest a pentacoordinate site somewhat similar to subsite Fe-a of the Fe4S4 cluster of active aconitase. The same values for Delta E(Q) and delta were observed for CODHCt. Upon addition of CN-, a potent inhibitor of CO oxidation, the Delta E(Q) of FCII of CODHCt changes from 2.82 to 2.53 mm/s, suggesting that CN- binds to the FCII iron. The Mossbauer studies of CODHRr showed that only similar to 60% of the C-clusters were capable of attaining the C-red1 state; the remainder were C-ox (or C-ox(*)). For the Mossbauer sample, the EPR spin concentration of the g(av) = 1.82 signal was similar to 65% of that determined for the g = 1.94 signal of B(red)of the fully reduced sample, a result consistent with the similar to 60% obtained from Mossbauer spectroscopy. When CODHRr was reduced with CO or dithionite, a fraction of the C-clusters developed a signal similar to the g(av) = 1.86 signal (C-red2) of CODHCt. The Mossbauer and EPR spectra of dithionite-reduced CODHRr show that a large fraction of the C-centers are in a state for which the [Fe4S4](+) cluster has S = 3/2. While the assumption of an [Fe4S4](+) cluster with an aconitase-type subsite electronically isolated from the Ni site can explain the g values of the g(av) = 1.82 signal and the absence of Ni-61 hyperfine interactions, published resonance Raman and EPR data suggest that the Ni site may be electronically linked to the Fe-S moiety of the C-cluster. We present a model that considers a weak exchange interaction (effective coupling constant j) between an S = 1 Ni-II site (zero-field splitting, D) and the S = 1/2 ground state of the [Fe4S4](+) cluster. This model suggests \j\ < 2 cm(-1), accounts for the g values of C-red1, and provides an explanation for the unusual g values (g(av) approximate to 2.16) reported by S. W. Ragsdale and co-workers for the adducts of CODHCt with thiocyanate and cyanate. The coupling model is consistent with Ni-61 EPR studies of CODH.
