Thermodynamics of Ligand Binding to ABP
A R T I C L E S
Table 2. Temperature Dependence of the Binding of D-Galactose
and 6-Deoxy-D-galactose to ABP
1-deoxy, 2-deoxy-, and 3-deoxy analogues of galactose are each
more favorable than that of galactose by ∼-30 kJ/mol, which
is counterintuitive. While the unfavorable entropic contribution
arising from the loss of a given hydroxyl group may comprise
a number of factors, including the influence of an unpaired polar
side chain, van der Waals interactions, or nonpolar interactions
during complex formation, the conclusion that ligand-protein
hydrogen bonds are enthalpically significantly more favorable
than ligand-solvent hydrogen bonds would appear to be
inescapable, at least in this system. This is in marked contrast
to the conclusions of Connelly et al. in the study of FK506
binding interactions.39 The large negative T∆S° of binding might
in turn arise from significant restriction of protein degrees of
freedom due to the strong nonbonded interactions with the
ligand. Alternatively, as suggested by Lemieux,40 the unfavor-
able T∆S° term might derive from disordered water molecules
adjacent to polyamphiphilic “hydraphobic” surfaces which are
less structured in comparison with bulk water. It should be
possible to distinguish between these two possibilities by
measurement of T∆S° for protein degrees of freedom on a per-
residue basis by use of NMR relaxation experiments.41-43
Resonance assignments in ABP have recently been completed,30
and such measurements are in progress.
Kd
M)
∆
G
°
∆
H
°
T∆S°
error (kJ/mol) error
ligand
(
µ
error
(kJ/mol)
error
(kJ/mol)
308 K
galactose 2.2 0.02 -33.36 0.02 -95.0 0.6 -61.6 0.6
6-deoxy 25.5 1.6 -27.1 0.2 -93
4
-66
4.3
298 K
galactose 0.61 0.02 -35.45 0.1 -87.4 0.84 -52.0 0.84
6-deoxy
6.67 0.09 -29.53 0.03 -81.67 0.5 -52.1 0.45
288 K
galactose 0.52 0.04 -34.63 0.19 -81.8 1.34 -47.2 1.3
6-deoxy 2.43 0.07 -30.95 0.08 -74.43 0.93 -43.5 0.9
temperature dependence of ∆H°. A simple linear model provides
estimates of ∆Cp for the binding of galactose and 6-deoxyga-
lactose of -656 ( 57 and -936 ( 124 J/mol/K, respectively.
36
Since the early work of Kauzmann, large changes in heat
capacity have been identified as a signature for hydrophobic
interactions.37 A significant negative ∆Cp of binding suggests
burial of hydrophobic surface area, based on the good correlation
between ∆Cp and changes in surface area in a number of
systems.38 The larger negative ∆Cp for the more hydrophobic
6-deoxygalactose in comparison with that for galactose is
consistent with this interpretation. By use of solvent-accessible
surface areas for galactose and 6-deoxygalactose computed from
AMSOL 7.0 calculations, together with eq 12 of ref 35, ∆∆C°P
is estimated to be ∼-90 J/mol/K. In addition, the contribution
to ∆∆C°P arising from the sequestration of one additional water
molecule in the complex with 6-deoxygalactose can be estimated
as ∼-75 J/mol/K,12,37 giving a total ∆∆C°P ) -165 J/mol/K.
This result is in satisfactory agreement with the experimental
value of ∆∆C°P ) -280 ( 136 J/mol/K.
Conclusions
In this study we have compared the thermodynamics of
binding of galactose and various deoxy derivatives thereof to
the arabinose binding protein. A combination of isothermal
titration calorimetry experiments together with calculation of
relative ligand solvation free energies, reveals that the contribu-
tion to the solute-solute free energy of binding from the OH1,
OH2, OH3, and OH6 groups appears to be remarkably constant
at ∼-30 kJ/mol, despite the fact that each hydroxyl group
subtends different numbers of hydrogen bonds in the complex.
The substantially unfavorable enthalpy of binding (∼30 kJ/mol)
of 1-deoxygalactose, 2-deoxygalactose, and 3-deoxygalactose
in comparison with that of galactose, cannot be readily ac-
counted for by differences in ligand solvation, suggesting that
ligand-protein hydrogen bonds are enthalpically significantly
more favorable than ligand-solvent hydrogen bonds. The
significant affinity of ABP for 6-deoxygalactose, which is only
∼1 order of magnitude weaker than galactose, derives in part
from the incorporation of an additional water molecule that
substitutes for OH6. In contrast, the substantially higher affinity
for 2-deoxygalactose in comparison with that for either 1-deoxy-
galactose or 3-deoxygalactose derives not from differences in
solvation of the binding site but from differences in the solvation
free energies of the free ligands. We have recently noted similar
thermodynamics governing the specificity of binding of pyrazine
derivatives to the major urinary protein,44 suggesting that the
modulation of binding specificity by ligand solvation thermo-
dynamics may be a general phenomenon.
Overall Thermodynamics of Binding. Binding to ABP of
all ligands described in the present study is enthalpy driven.
With the exception of charged ligands such as heparin and
heparin sulfate, both ∆H° and particularly T∆S° are significantly
larger than typical values reported for the vast majority of
carbohydrate-protein interactions,13 including oligosaccharides.
The reason for these anomalously large values cannot be
delineated with certainty from global thermodynamics measure-
ments derived from ITC. It is noteworthy, however, that the
∆H° value for binding of galactose compared with that for
1-deoxy-, 2-deoxy-, or 3-deoxygalactose is favorable by ∼30
kJ/mol. Naively, one might interpret this additional enthalpic
contribution to binding in the case of galactose as arising from
the additional hydrogen bond(s) that form due to the additional
hydroxyl group in the complex. However, prior to binding, the
ligand is hydrogen bonded to solvent water, the enthalpic
contribution of which is contained within ∆∆Gïsolv. Intuitively,
one would anticipate that the enthalpic component (∆∆Hï
)
solv
for solvation of deoxy analogues of galactose compared with
that for galactose is endothermic, i.e., solvation of deoxy
analogues is enthalpically less favorable than that of galactose.
If one assumes momentarily that the enthalpic contribution from
ligand-protein hydrogen bonds is effectively zero, on the basis
that hydrogen bonds to solvent exist prior to the association,
one must conclude that the enthalpies of solvation of the
(39) Connelly, P.; Aldape, R. A.; Bruzzese, F. J.; Chambers, S. P.; Fitzgibbon,
M. J.; Fleming, M. A.; Itoh, S.; Livingston, D. J.; Navia, M. A.; Thomson,
J. A.; Wilson, K. P. Proc. Natl. Acad. Sci. 1994, 91, 1964-1968.
(40) Lemieux, R. U. Acc. Chem. Res. 1996, 29, 373-380.
(41) Akke, M.; Bruschweiler, R.; Palmer, A. G. J. Am. Chem. Soc. 1993, 115,
9832-9833.
(42) Yang, D. W.; Kay, L. E. J. Mol. Biol. 1996, 263, 369-382.
(43) Yang, D.; Mok, Y.-K.; Forman-Kay, J. D.; Farrow, N.; Kay, L. E. J. Mol.
Biol. 1997, 272, 790-804.
(44) Bingham, R.; Bodenhausen, G.; Findlay, J. H. B. C.; Hsieh, S.-Y.; Kalverda,
A. P.; Kjellberg, A.; Perazzolo, C.; Phillips, S. E. V.; Seshadri, K.; Turnbull,
W. B.; Homans, S. W. J. Am. Chem. Soc. 2004, 126, 1675-1681.
(36) Kauzmann, W. AdV. Protein Chem. 1959, 14, 1-63.
(37) Cooper, A. Biophys. Chem. 2000, 85, 25-39.
(38) Spolar, R. S.; Livingstone, J. R.; Record, M. T. J. Biochemistry 1992, 31,
3947-3955.
9
J. AM. CHEM. SOC. VOL. 126, NO. 38, 2004 11875