Solvent dynamics and mechanism of proton transfer in human carbonic anhydrase II

In this paper we describe the first all-atom aqueous-phase MD
simulations of human carbonic anhydrase II in three protonation states
relevant to the rate-limiting intramolecular proton-transfer step. In
particular, we have examined the zinc-water form of the enzyme (CHOH),
the zinc-hydroxide form of the enzyme with a doubly protonated His-64
(COHH, the putative intramolecular proton-transfer proton-accepting
residue), and the native zinc-hydroxide form (COH) of the enzyme (i.e.,
with an unprotonated His-64). From these MD simulations (up to similar
to 1 ns in length) we have studied in detail the dynamics of these three
systems. overall the dynamics of the three systems do not vary
significantly (e.g., the active site region is rigid, the number of
long-lived hydrogen bonds is constant, etc.) with the exception of COHH.
In this case the residues that Line the entrance to the active site
cavity (near the location of His-64) undergo significantly higher
fluctuations than in the CHOH and COH cases. It is postulated that this
facilitates solvent and buffer exchange around His-64, thereby
facilitating the intermolecular proton-transfer step. We also find that
the motion of His-64 is limited in all three cases to occupying the
``in'' orientation (similar to 7 Angstrom from the zinc ion, while the
so-called ``out'' conformer is further away), which suggests that
fluctuations of this residue between the in and out conformers have a
limited influence on the intramolecular proton transfer. However, due to
the limited time scales of our simulations, this needs to be examined in
more detail. importantly, though, we find that His-64 acts as a
``gate-keeper'' between the inner active site region (characterized by
localized water molecules) and the outer (bulk) region, which is
characterized by relatively freely diffusing water molecules. This
function of His-64 has not been realized previously. In the inner active
site we have identified relatively long-lived water bridges between the
zinc-bound water or hydroxide and the imidazole or imidazolium side
chain of His-64. The lengths of these bridges vary between two and six
water molecules, and the preferred bridge depends on the protonation of
the active site. We estimate that the probability of water bridge
formation is low (at most similar to 1.5 kcal/mol) and that water bridge
formation is not the rate-limiting step in the proton-transfer process
(transfer from zinc-bound water to an active site water is
rate-limiting).