Ribonucleotide reductase (RDPR) from Escherichia coli is composed of two subunits, R1 and R2, and catalyzes the conversion of nucleotides to deoxynucleotides. The mechanism of inactivation of RDPR by 2'-azido-2'-deoxynucleoside 5'-diphosphate (N3UDP) has been examined using a variety of isotopically labeled derivatives: (1'-, 2'-, 3'-, or 4'-[2H])-N3UDPs and 2'-[15N3, 13C]-N3UDP. Electron paramagnetic resonance (EPR) and electron spin echo envelope modulation (ESEEM) spectroscopy studies using these compounds indicate that the 2' carbon-nitrogen bond to the azide moiety is cleaved prior to or upon formation of the nitrogen-centered radical derived from the azide moiety of N3UDP. EPR studies reveal no hyperfine interactions of the nitrogen-centered radical with the 1', 2', 3', or 4' hydrogens of N3UDP. ESEEM studies however, reveal that the 1' and 4' deuterons are 3.3 +/- 0.2 and 2.6 +/- 0.3 A, respectively, from the nitrogen-centered radical. Further support for carbon-nitrogen bond cleavage is derived from studies of the interaction of oxidized R1, C225SR1, and C462SR1 with R2 and N3UDP. In all three cases, in contrast to the results with the wild type R1, azide is detected. Nitrogen-centered radical is not observed with either oxidized R1 or C225SR1 but is observed with C462SR1. These results suggest that C225 is required for the conversion of azide into N2 and a nitrogen-centered radical. The dynamics of the inactivation of RDPR by N3UDP have also been examined. Use of [3'-2H]N3UDP reveals an isotope effect of approximately 2 on the loss of the tyrosyl radical and the rate of inactivation of RDPR. In both cases the kinetics are complex, suggesting multiple modes of inactivation. In addition, several modes of inactivation are required to explain the observation that loss of the tyrosyl radical is slower than the rate of inactivation. Studies using [5'-3H]N3UDP reveal that the rapid inactivation is the result of the formation of a tight noncovalent complex between modified nucleotide, nitrogen-centered radical and RDPR. Destruction of the nitrogen-centered radical is a slow process which appears to be accompanied by decomposition of the modified nucleotide into PPi, uracil, and 2-methylene-3(2H)-furanone. The latter covalently modifies R1 and ultimately leads to loss of approximately 50% of the activity of R1.

The initial step in the mechanism-based inactivation of ribonucleotide reductases by 2‘-chloro-2‘-deoxynucleotides is abstraction of H3‘ by a proximal free radical on the enzyme. The C3‘ radical is postulated to undergo spontaneous loss of chloride, and the resulting cationic radical loses a proton to give a 3‘-keto intermediate. Successive β-eliminations produce a Michael acceptor which causes inactivation. This hypothesis would predict rapid loss of mesylate or tosylate anions from C2‘, but sluggish loss of azide or thiomethoxide. In contrast, loss of azido and methylthio radicals from C2‘ should occur readily whereas homolysis to give (methyl or tolylsulfonyl)oxy and fluoro radicals should be energetically prohibitive. Protected 3‘-O-(phenoxythiocarbonyl)-2‘-substituted nucleosides were treated with tributylstannane/AIBN or triphenylsilane/dibenzoyl peroxide in refluxing toluene. The 2‘-O-(mesyl and tosyl) and 2‘-fluoro compounds underwent direct radical-mediated hydrogenolysis of the thionocarbonate group to give 3‘-deoxy-2‘-substituted products, whereas 2‘-(azido, bromo, chloro, iodo, and methylthio)-3‘-thionocarbonates gave 2‘,3‘-didehydro-2‘,3‘-dideoxy derivatives via loss of 2‘-substituents from an incipient C3‘ radical. These results are in harmony with loss of radicals, but not anions, from C2‘. The well-known radical-mediated hydrogenolytic cleavage of halogen and methylthio (slow) groups from C2‘ of the 3‘-hydroxy (unprotected) precursors and reduction of 2‘-azides to amines occurred with tributylstannane/AIBN. Triphenylsilane/dibenzoyl peroxide gave parallel (but slower) hydrogenolysis with the 2‘-(iodo, bromo, and methylthio) compounds, but cleavage of the 2‘-chloro group was very slow and no reduction of 2‘-azides to amines was detected. Rather, the latter system effected slow hydrogenolytic removal of the 2‘-azido group. Thus, chemoselective differentiation of certain functional groups is possible with triphenylsilane and tributylstannane. Reduction of azides to amines with tributylstannane is known, but hydrogenolytic deazidation (slow) with triphenylsilane in the absence of amine formation appears to be novel.

No abstract

No abstract