A. Warshel et al., LINEAR FREE-ENERGY RELATIONSHIPS IN ENZYMES - THEORETICAL-ANALYSIS OFTHE REACTION OF TYROSYL-TRANSFER-RNA SYNTHETASE, Journal of the American Chemical Society, 116(19), 1994, pp. 8437-8442
Recent studies of genetically modified enzymes have indicated that cha
nges in activation free energies, Delta Delta g(double dagger), and ch
anges in reaction free energies, Delta Delta G(0), are correlated by t
he relationship Delta Delta g(double dagger) = beta Delta Delta G(0).
The present work explores the basis for such linear free energy relati
onships (LFERs) in enzymatic reactions, focusing on the effects of mut
ations in tyrosyl-tRNA-synthetase (TTS). It is demonstrated that the o
ptimal way to analyze LFERs is by describing the reaction in terms of
pure valence bond (VB) resonance structures rather than in terms of pa
rtially formed bonds. The use of the pure VB representation allows one
to evaluate the relevant LFER using Marcus-type concepts and to compa
re the predicted beta to the observed one. Using a two-resonance-struc
ture VB model for TTS produces beta similar or equal to 0.5, which dis
agrees with the observed values of beta similar or equal to 0.83 and b
eta >> 1 for two classes of mutations. Noting, however, that the phosp
horyl transfer process in TTS has been described before as going throu
gh a high-energy intermediate, we describe this reaction in terms of t
hree VB resonance structures. This accounts for the observed values of
beta and supports the validity of LFER in TTS. It is pointed out that
LFERs are valid in proteins even when the changes in Delta g(double d
agger) involve very anharmonic interactions like hydrogen bonds, since
such relationships reflect the correlation between Delta Delta g(doub
le dagger) and Delta Delta G(0) rather than the correlation between De
lta Delta g(double dagger) and the effect of specific residues. Howeve
r, obtaining LFER in proteins requires that the active site environmen
t responds linearly to the change of charges during the reaction, and
such a linear response is far from obvious. Fortunately, the simulatio
n study presented in this work as well as previous simulations has dem
onstrated that active sites of proteins obey the linear response appro
ximation. Such a behavior of highly anharmonic systems is due to the a
vailability of many compensating polar interactions. This finding prov
ides a theoretical basis for the experimental observation of LFER in T
TS.