ESR Distance Measurements Support a Model for Long Range Radical Initiation in E. Coli Ribonucleotide Reductase

Escherichia coli ribonucleotide reductase (RNR) catalyzes the conversion of nucleoside diphosphates (NDPs) to deoxynucleoside diphosphates (dNDPs). The active protein is composed of two homodimeric subunits (R1 and R2) thought to form a 1:1 complex (see figure). R1 binds the NDP substrates, houses the essential cysteines required for catalysis and the binding sites for the allosteric effectors that govern substrate specificity and turnover rates. R2 harbors the essential di-iron tyrosyl radical cofactor on residue 122 (Y.). A major unresolved issue is the mechanism of radical initiation: how the tyrosyl radical (Y.) in R2 generates a transient thiyl radical (C439.) in R1 required for nucleotide reduction. The current model for the radical initiation process involves a specific pathway composed of aromatic amino acids and traverses a distance of 3.5 nm, derived from a docking model of the R1 and R2 structures, but a substantial part of the electron transfer pathway is not apparent from the available structural information.& A measurement of the distance between Y122 and C439 was imperative to establish the radical initiation model. Using DQC-ESR and DEER we performed the first measurements of the distance between Y. in R2 and a nitrogen centered radical (N.) directly attached to C225 in R1. The 6 pulse 17.4 GHz DQC results at two temperatures are illustrated below. The oscillating signal at 25K shows the presence of fixed pairs at a range of distances. The corresponding distance distribution that was obtained by Tikhonov regularization is also shown. There are pairs at different distances that are possible in this formally 4-spin system. These and our DEER results show the presence of three types of pairs with average distances of 3.3 nm. (Y.-Y.), 3.9 and 4.84 nm (two possible distances for N.-Y., with the former just a minor component). At 80K only the signal from the N.-N. pairs survives the increased effects of the fast-relaxing di-iron cluster. We could thus estimate the amount of N.-N. pairs to be less than 10% of the N.-Y. pairs detected at 25K, with most of the 85 K signal coming from intermolecular dipolar interactions.

From considerations of the docking model and the absence of a significant N.-N. interaction, we can assign the 4.8 nm distance as between Y. and N., which is consistent with structure b in the figure. This distance is consistent with the docking model, and it rules out a large conformational change between R1 and R2 on active complex formation.