NMR data do not implicate a phosphorane in the T4 DNA ligase reaction

The interesting article by Cherepanov et al. (1) presents ³¹P NMR data as evidence for a pentacoordinate phosphorane intermediate in the T4 DNA ligase reaction. Although some sort of intermediate apparently forms under the low-temperature, solid-state conditions that were used, the data are not consistent with a phosphorane.

The ³¹P resonances assigned to the α, β, and γ phosphoryl groups of bound ATP substrate are −12.38, −22.84, and −7.45 ppm, respectively, and −11.7, −13.04, and −7.7 in the putative phosphorane intermediate. The various coordination states of phosphorus are characterized by chemical shifts that differ from one another by 50 ppm or more (2). Thus, phosphorane formation at the α-phosphorus via nitrogen attack, as proposed, would produce the largest change at the α position, not at the β, as observed. Furthermore, none of the observed chemical shifts resembles those of reported phosphoranes, which range from about −55 to −85 ppm (refs. 2 and 3; see also ref. 4 and references therein). This chemical shift range is characteristic for neutral pentaoxyphosphoranes, whereas that postulated by Cherepanov et al. (1) would be anionic; however, a different protonation state is unlikely to make a phosphorane resemble a phosphate. For example, the protonation of phosphoric acid and its esters results in changes of only 3–4 ppm (5). Similarly, the substitution of a nitrogen for an oxygen ligand causes only a small change in chemical shift, regardless of the coordination number at phosphorus (2). The identity of the intermediate observed in this study is uncertain, but the NMR data are not those expected of a pentacoordinate phosphorane.