Protein Engineering vol.4 no 3 pp.229 —231, 1991
Linear free energy relationships in enzyme binding interactions studied by protein engineering
Alan R.Fersht and Tim N.C.Weils1 MRC Unit lor Protein Function and Design, Cambridge IRC for Protein Engineering, University Chemical Laboratory, Cambridge CB2 1EW, UK and 'Glaxo Institute for Molecular Biology S A., Route des Acacias 46, 1211 Geneva 24, Switzerland
Introduction We have analysed by protein engineering methods the effects of binding interactions at the active site of tyrosyl-tRNA synthetase on the kinetics and equilibria for the formation of tyrosyl adenylate (Fersht et al., 1986, 1987; Fersht, 1987a). Changes in the rate constants for formation of tyrosyl adenylate, e.g. equation (1), were measured, as were the changes in equilibrium E.Tyr.ATP
s*
E.Tyr-AMP.PPi
(1)
constant for its formation on deletion of moieties of the side chains that interact with the substrates. Interestingly, certain of the changes may be described by an LFER: AG$ = A + /3AG,eq
(2)
where, for example, AG$, is the activation energy for the formation of enzyme-bound tyrosyl adenylate and pyrophosphate from enzyme-bound tyrosine and ATP (equation 1) and AGeq is the equilibrium free energy change of the reaction. The value of (3 for equation (1) for mutation of those residues studied that interact with the side chain of tyrosine and the ribose ring of ATP is 0.83 ± 0.05. This was calculated from a plot of log k3 against log £_3 assuming a linear functional relationship between the two and transforming the resulting slope parameters © Oxford University Press
Consider the binding of a substrate S to an enzyme that proceeds to products via a transition state S$. The structure of the active site of the enzyme cannot be complementary to the structures of all three of S, SJ and final products, but can be perfectly complementary only to one of the three components and/or to parts of all three. Since the hypotheses of Haldane (1930) and Pauling (1946), it is currently accepted that, in most cases, enzyme transition-state complementarity is optimal for catalysis. This was directly confirmed by protein engineering experiments on the tyrosyl-tRNA synthetase that showed that there are groups on the enzyme that contribute no stabilization energy whatsoever to ground states but bind strongly to the transition state of the reactants. There are in addition, however, groups that bind non-reacting parts of the substrate more strongly in the enzyme-product complex than in the enzyme-substrate and enzyme-transition state complexes (Figure 1, right). Enzyme-product (or enzyme - intermediate) complementarity had earlier been proposed to be important for reactions that form very unstable products to displace equilibria in favour of products (Jencks, 1969) and to sequester reactive intermediates (Fersht, 1974). Tyrosyl-tRNA synthetase, for example, binds tyrosyl 229
Downloaded from http://peds.oxfordjournals.org/ at Memorial University of Newfoundland on July 18, 2014
Experiments on mutants of tyrosyl-tRNA synthetase have shown that there can be linear free energy relationships (LFERs) between changes in activation free energies and changes in binding energies when groups are deleted that bind to non-reacting parts of the substrate (Fersht et al., 1986, 1987). It has now been proposed (Straub and Karplus, 1990) that such LFERs can occur for the mutation of hydrogen bonding groups only for the limiting examples of Br0nsted /3 of 0, 1 or oo, and that fractional values of /3 are not permissible. The reasoning behind this is that the energy of a hydrogen bond is not linear with distance and the (false) premise that an LFER requires that there is a linear relationship between bond energy and distance. We show from a simple model how LFERs can arise for binding interactions and how they can give fractional values of /3, in accord with experimental evidence. An LFER occurs between binding and catalysis when a set of interactions exists in which each member contributes to the binding energy of the transition state the same fraction of the binding energy it contributes to the products (both relative to the ground state).
(Fersht, 1986b; Fersht et at., 1987; Wells and Fersht, 1989). The probability of/3 being 1.0 is
Linear free energy relationships in enzyme binding interactions studied by protein engineering
Alan R.Fersht and Tim N.C.Weils1 MRC Unit lor Protein Function and Design, Cambridge IRC for Protein Engineering, University Chemical Laboratory, Cambridge CB2 1EW, UK and 'Glaxo Institute for Molecular Biology S A., Route des Acacias 46, 1211 Geneva 24, Switzerland
Introduction We have analysed by protein engineering methods the effects of binding interactions at the active site of tyrosyl-tRNA synthetase on the kinetics and equilibria for the formation of tyrosyl adenylate (Fersht et al., 1986, 1987; Fersht, 1987a). Changes in the rate constants for formation of tyrosyl adenylate, e.g. equation (1), were measured, as were the changes in equilibrium E.Tyr.ATP
s*
E.Tyr-AMP.PPi
(1)
constant for its formation on deletion of moieties of the side chains that interact with the substrates. Interestingly, certain of the changes may be described by an LFER: AG$ = A + /3AG,eq
(2)
where, for example, AG$, is the activation energy for the formation of enzyme-bound tyrosyl adenylate and pyrophosphate from enzyme-bound tyrosine and ATP (equation 1) and AGeq is the equilibrium free energy change of the reaction. The value of (3 for equation (1) for mutation of those residues studied that interact with the side chain of tyrosine and the ribose ring of ATP is 0.83 ± 0.05. This was calculated from a plot of log k3 against log £_3 assuming a linear functional relationship between the two and transforming the resulting slope parameters © Oxford University Press
Consider the binding of a substrate S to an enzyme that proceeds to products via a transition state S$. The structure of the active site of the enzyme cannot be complementary to the structures of all three of S, SJ and final products, but can be perfectly complementary only to one of the three components and/or to parts of all three. Since the hypotheses of Haldane (1930) and Pauling (1946), it is currently accepted that, in most cases, enzyme transition-state complementarity is optimal for catalysis. This was directly confirmed by protein engineering experiments on the tyrosyl-tRNA synthetase that showed that there are groups on the enzyme that contribute no stabilization energy whatsoever to ground states but bind strongly to the transition state of the reactants. There are in addition, however, groups that bind non-reacting parts of the substrate more strongly in the enzyme-product complex than in the enzyme-substrate and enzyme-transition state complexes (Figure 1, right). Enzyme-product (or enzyme - intermediate) complementarity had earlier been proposed to be important for reactions that form very unstable products to displace equilibria in favour of products (Jencks, 1969) and to sequester reactive intermediates (Fersht, 1974). Tyrosyl-tRNA synthetase, for example, binds tyrosyl 229
Downloaded from http://peds.oxfordjournals.org/ at Memorial University of Newfoundland on July 18, 2014
Experiments on mutants of tyrosyl-tRNA synthetase have shown that there can be linear free energy relationships (LFERs) between changes in activation free energies and changes in binding energies when groups are deleted that bind to non-reacting parts of the substrate (Fersht et al., 1986, 1987). It has now been proposed (Straub and Karplus, 1990) that such LFERs can occur for the mutation of hydrogen bonding groups only for the limiting examples of Br0nsted /3 of 0, 1 or oo, and that fractional values of /3 are not permissible. The reasoning behind this is that the energy of a hydrogen bond is not linear with distance and the (false) premise that an LFER requires that there is a linear relationship between bond energy and distance. We show from a simple model how LFERs can arise for binding interactions and how they can give fractional values of /3, in accord with experimental evidence. An LFER occurs between binding and catalysis when a set of interactions exists in which each member contributes to the binding energy of the transition state the same fraction of the binding energy it contributes to the products (both relative to the ground state).
(Fersht, 1986b; Fersht et at., 1987; Wells and Fersht, 1989). The probability of/3 being 1.0 is
