## Saturday, August 11, 2012

### Transition states of native and drug-resistant HIV-1 protease are the same

What is the rate determining step?
The general features of the HIV-1 protease mechanism appear to be 1) nucleophilic attack on the amide carbonyl group by a water molecule to form a gem-diol intermediate, 2) protonation of the amide nitrogen, and 3) cleavage of the amide C-N bond.

As pointed out in the introduction, computational studies such as Okimoto et al. and. Piana et al. both  seem to indicate that Step 2 is the rate limiting step.

However, a closer look at these studies suggest that the issue is not well settled.  Okimoto et al.'s (small gas phase) model suggests that the gem-diol intermediate is more stable than the enzyme-substrate complex and computed the barrier of the second step relative to this intermediate.  Piana et al.'s (QM/MM) model suggests that the gem-diol intermediate is less stable than the enzyme-substrate complex and computed the barrier of the second step relative to this enzyme substrate.

If one switches the reference state in both studies Okimoto et al.'s model would predict the first two steps to have essentially the same barrier, while Piana et al.'s model would predict the first step to be rate determining.  Furthermore, the primary $^{15}$N isotope effect computed by Piana et al. (0.97$\pm$0.2) is larger than the experimental value obtained by Rodrigez et al. (0.995$\pm$0.002) (large isotope effects mean significant deviations from 1).

Kinetic isotope effects confirm the rate limiting step and the transition state structure
Schramm and co-workers have used labelled substrates to measure primary $^{14}$C and $^{15}$N and secondary $^3$H and $^{18}$O isotope effects.  Furthermore, they used small gas phase models to compute the corresponding isotope effects for all five stationary points at the ONIOM (M06-2X/6-31+G**:am1) level of theory.

Overall the measured isotope effects best match the computed values for the transition state for Step 2, which is therefore likely to be the rate determining step.  This isotope effect was computed relative to the enzyme-substrate complex; it would be interesting to recompute this value relative to the gem-diol intermediate, which is sufficiently stable to be observed experimentally at low-pH conditions (Das et al.)

Transition states and drug design
Very similar isotope effects was also measured for a mutant (I84V) which has displayed resistance to all nine FDA approved-inhibitors, indicating that the transition state structure in this mutant is quite similar to that of the wild-type.  Thus, transition state-mimics would likely inhibit both forms of the enzyme and may lead to new inhibitors that are less prone to resistance.

Acknowledgement: Thanks to Luca De Vico for alerting me to this paper