Class Ia ribonucleotide reductases (RNRs) are allosterically regulated by ATP and dATP to maintain the appropriate deoxyribonucleotide levels inside the cell for DNA biosynthesis and repair. RNR activity requires precise positioning of the β2 and α2 subunits for the transfer of a catalytically essential radical species. Excess dATP inhibits RNR through the creation of an α-β interface that restricts the ability of β2 to obtain a position that is capable of radical transfer. ATP breaks the α-β interface, freeing β2 and restoring enzyme activity. Here, we investigate the molecular basis for allosteric activity regulation in the well-studied Escherichia coli class Ia RNR through the determination of six crystal structures and accompanying biochemical and mutagenesis studies. We find that when dATP is bound to the N-terminal regulatory cone domain in α, a helix unwinds, creating a binding surface for β. When ATP displaces dATP, the helix rewinds, dismantling the α-β interface. This reversal of enzyme inhibition requires that two ATP molecules are bound in the cone domain: one in the canonical nucleotide-binding site (site 1) and one in a site (site 2) that is blocked by phenylalanine-87 and tryptophan-28 unless ATP is bound in site 1. When ATP binds to site 1, histidine-59 rearranges, prompting the movement of phenylalanine-87 and trytophan-28, and creating site 2. dATP hydrogen bonds to histidine-59, preventing its movement. The importance of site 2 in the restoration of RNR activity by ATP is confirmed by mutagenesis. These findings have implications for the design of bacterial RNR inhibitors.