An "open"-state crystal structure of lysine 5,6-amino-mutase suggests that transition to a hypothetical "closed"-state is required to bring the cofactors adenosylcobalamin (AdoCbl) and pyridoxal-5'-phosphate (PLP) and the substrate into proximity for the radical-mediated 1,2-amino group migration. This process is achieved by transaldimination of the PLP-Lys144 beta internal aldimine with the PLP substrate external aldimine. A closed state crystal structure is not available. UV-vis and electron paramagnetic resonance studies show that homologues of substrate D-lysine, 2,5-DAPn, 2,4-DAB, and 2,3-DAPr bind to PLP as an external aldimine and elicit the AdoCbl Co-C bond homolysis and the accumulations of cob(II)alamin and analogue based radicals, demonstrating the existence of a closed state. H-2- and P-31-electron nuclear double resonance studies, supported by computations, show that the position for hydrogen atom abstraction from 2,5-DAPn and 2,4-DAB by the 5'-deoxyadenosyl radical occurs at the carbon adjacent to the amine, resulting in overstabilized radicals by spin delocalization through the triune into the pyridine ring of PLP. These radicals block the active site, inhibit the enzyme, and poise the enzyme into two distinct conformations: for evennumbered analogues, the cob(II)alamin remains proximal to and spin-coupled with the analogue-based radical in the closed state while odd-numbered analogues could trigger the transition to the open state of the enzyme. We provide here direct spectroscopic evidence that strongly support the existence of a closed state and its analogue-dependent transition to the open state, which is one step that was proposed to complete the catalytic turnover of the substrate lysine.