The intermediate, designated X, formed during the self-assembly reaction of the tyrosyl radical/µoxo-bridged diferric cofactor in the R2 subunit of Escherichia coli ribonucleotide reductase (RNR) is directly involved in the oxidation of Y122 to the catalytically essential ‚Y122. Earlier rapid freeze-quench (RFQ) Q-band ENDOR studies led to the formulation of X as a spin-coupled Fe III /Fe IV center, with an S ) 1 / 2 ground state, and showed that X contains a single terminal aqua ligand (water molecule or 2-fold disordered hydroxyl) bound to Fe III but does not contain an hydroxyl bridge. That ENDOR data, coupled with RFQ-EXAFS data, plus the strong spin coupling between the iron ions constrain the structure of X to a di-or tribridged species whose inorganic core (defined as iron and exogenous ligands) contains the [(H x O)Fe III OFe IV ] fragment. To determine whether the core contains a second oxo bridge and to establish the fate of the atoms derived from O 2 , we have now performed CW and pulsed Q-band 17 O ENDOR experiments on samples of X prepared in both H 2 17 O and 17 O 2 , using a uniformly 15 N-labeled protein, [U-15 N]-R2. These measurements, along with kinetic studies on the formation of X in both wild-type and Y122F R2, as monitored by both ENDOR and S-band EPR spectroscopies, reveal that X contains two oxygen atoms. Both are initially derived from O 2 , with one present as a µ-oxo bridge and one as the terminal aqua ligand; with time the latter of these atoms exchanges with solvent. These and our previous studies indicate that X does not contain a di-µ-oxo-or µ-oxo,hydroxobridged core structure. A structure for X is proposed that contains a single oxo bridge, one terminal aqua ligand bound to the Fe III , and one or two additional mono-oxo bridges provided by the carboxylate oxygens of E115 and/or E238. In addition, the time course of the formation of X in the presence of 17 O 2 provides important insights into the dynamics of cluster assembly.

We report CW and pulsed Q-band 1,2H ENDOR measurements of intermediate X formed during the assembly of the diferric tyrosyl radical cofactor of the R2 subunit in ribonucleotide reductase. These studies, performed with H2O and D2O buffers, were designed to determine whether the exchangeable proton signals are associated with an hydroxo bridge, a terminal water, or both. In doing so, we identify the types of protonated oxygen (OH x ) species coordinated to the iron ions of X and their disposition relative to the ferric and ferryl iron ions. The exchangeable proton signals displayed by intermediate X belong to two protons associated with a terminal water bound to Fe,III and not to an hydroxo bridge; within the precision of the modeling, this picture of a terminal water is indistinguishable from that of a 2-fold disordered terminal hydroxyl. The fact that X displays strong spin-coupling between iron ions requires that there be one or more oxo/hydroxo bridges. These findings then establish that X contains the [(H x O)FeIIIOFeIV] fragment.

The g tensor of the tyrosyl radical present in the active R2 protein of ribonucleotide reductase is anisotropic, and the g 1 component is influenced by hydrogen bonding to the oxygen of the tyrosyl ring. We have studied the tyrosyl radical in the R2 protein of Escherichia coli, mouse, and herpes simplex virus type 1 (HSV1) with high-frequency EPR and pulsed ENDOR after reconstitution in D2O. From the high-frequency EPR measurements the g tensor of the radical in HSV1 RNR R2 was found to be identical to that in mouse R2, indicating the presence of a hydrogen bond to the phenolic oxygen in both cases, and in contrast to that in E. coli R2. The pulsed ENDOR spectra confirmed the absence of an exchangeable proton near the tyrosyl radical in E. coli R2. For mouse and HSV1 R2 a clear ENDOR signal of exchanged deuterium was found with a hyperfine splitting of −0.53 MHz (mouse) and −0.56 MHz (HSV1). This was interpreted as a proton at a distance of 1.89 Å (mouse) and 1.86 Å (HSV1) from the phenolic oxygen with an orientation, derived from simulations, in the plane of the tyrosyl ring. The most likely origin of this proton is the water ligand at Fe1. This is in contrast with photosystem II where the hydrogen bonding to the radical YD • was formed by a nearby histidine. The presence of the hydrogen bond to the tyrosyl radical may be related to the faster spin−lattice relaxation for the mouse and HSV1 radical compared to that for the E. coli radical, as measured before by Galli et al. [J. Am. Chem. Soc. 1995, 117, 740−746]. It seems therefore likely that the distance between the tyrosyl radical and the iron−oxygen cluster in mouse and HSV1 R2 proteins is shorter compared to that in E. coli R2. Since the tyrosyl radicals in the HSV1 and mouse R2 proteins are much more accessible to the solvent, the hydrogen bond may play a useful role in stabilizing the tyrosyl radical.