Transition state structure for ADP-ribosylation of eukaryotic elongation factor 2 catalyzed by diphtheria toxin

Bacterial protein toxins are the most powerful human poisons known, exhibiting an LD(50) of 0.1-1 ng kg(-)(1). A major subset of such toxins is the NAD(+)-dependent ADP-ribosylating exotoxins, which include pertussis, cholera, and diphtheria toxin. Diphtheria toxin catalyzes the ADP ribosylation of the diphthamide residue of eukaryotic elongation factor 2 (eEF-2). The transition state of ADP ribosylation catalyzed by diphtheria toxin has been characterized by measuring a family of kinetic isotope effects using (3)H-, (14)C-, and (15)N-labeled NAD(+) with purified yeast eEF-2. Isotope trapping experiments yield a commitment to catalysis of 0.24 at saturating eEF-2 concentrations, resulting in suppression of the intrinsic isotope effects. Following correction for the commitment factor, intrinsic primary kinetic isotope effects of 1.055 +/- 0.003 and 1.022 +/- 0.004 were observed for 1(N)'-(14)C]- and 1(N)-(15)N]NAD(+), respectively; the double primary isotope effect was 1.066 +/- 0.004 for 1(N)'-(14)C, 1(N)-(15)N]NAD(+). Secondary kinetic isotope effects of 1.194 +/- 0.002, 1.101 +/- 0.003, 1.013 +/- 0.005, and 0.988 +/- 0.002 were determined for 1(N)'-(3)H]-, 2(N)'-(3)H]-, 4(N)'-(3)H]-, and 5(N)'-(3)H]NAD(+), respectively. The transition state structure was modeled using density functional theory (B1LYP/6-31+G) as implemented in Gaussian 98, and theoretical kinetic isotope effects were subsequently calculated using Isoeff 98. Constraints were varied in a systematic manner until the calculated kinetic isotope effects matched the intrinsic isotope effects. The transition state model most consistent with the intrinsic isotope effects is characterized by the substantial loss in bond order of the nicotinamide leaving group (bond order = 0.18, 1.99 A) and weak participation of the attacking imidazole nucleophile (bond order = 0.03, 2.58 A). The transition state structure imparts strong oxacarbenium ion character to the ribose ring even though significant bond order remains to the nicotinamide leaving group. The transition state model presented here is asymmetric and consistent with a dissociative S(N)1 type mechanism in which attack of the diphthamide nucleophile lags behind departure of the nicotinamide.