sense, with an associated scalar field, such as found for the electric forces in which two charged particles
attract or repel each other. Instead, the hydrophobic effect is a consequence of the behaviour of bulk water
at a non-polar surface.^6 In its optimal orientation, a water molecule can form four hydrogen bonds in a
tetrahedral geometry: two through its lone pairs of electrons (the acceptors) and two through its protons
(the donors). In bulk solvent at physiological temperatures and pressures, the water molecules are in
continual exchange of hydrogen-bonding partners. When the water molecules encounter a non-polar
molecule that cannot accept or donate, these hydrogen bonding interactions are broken. If two such
non-polar molecules are in proximity, they will associate so as to minimise the disruption to the dynamic
water interactions. In general, macromolecules such as proteins and nucleic acids tend to associate in such
a way as to expose a minimum number of non-polar features on the external surface to the aqueous solv-
ent. The non-polar parts of nucleic acid include the 5-methyl group of thymine, the heterocyclic carbon
atoms within the purine and pyrimidine rings, and the ribose carbon atoms. These atoms are often
contacted by the aliphatic portions of amino acid side chains in protein–nucleic acid complexes (e.g.
Figure 10.4d).
The association of hydrophobic molecules in aqueous solvent is favoured entropically. This might
seem surprising given that a hydrogen-bonding pattern is being affected and that bond breaking is associ-
ated mostly with enthalpicchange. However, it is thought that water molecules become immobilized on
non-polar surfaces because of restricted interactions with partner water molecules. Thus, by burying the
exposed non-polar surface, water molecules are liberated and their translational and rotational entropy
increases, with a concomitant lowering of the overall free energy. In this way, shape complementarity can
help to bury non-polar surfaces optimally and to contribute to the affinity of a macromolecular interaction.
Shape complementarity also favours optimal van der Waals interactions (Section 10.2.4).
10.2.4 How Dispersions Attract: van der Waals Interactions and Base Stacking
Van der Waals interactionsarise from the dispersiveforces originating from transient dipoles inherent in the
electronic structure of atoms and molecules, and even molecules without a net permanent dipole can form
these interactions. Dispersive forces are proportional to the inverse sixth power of the distance between the
dipoles. Although the interactions are very weak and sensitive to structural fluctuations, their cumulative con-
tribution becomes significant over an extensive intermolecular interface. In a specific protein–DNA complex,
van der Waals contacts account for roughly two-thirds of the interfacial surface area, whilst the remaining one-
third involves direct or water-mediated hydrogen-bonding interactions.^7 Thus, the contacting protein and
nucleic acid surfaces in specific complexes are not only characterized by the stereochemical match of donors
and acceptors, but also by the shape complementarity that produces a snug fit so as to benefit from the r^6 law.
Dispersive effects also contribute to the stability of base-stacking interactionsand thus are an import-
ant aspect of the sequence-dependent conformational effects mentioned earlier (Section 2.3.1). As a con-
sequence of stacking and base-pair hydrogen bonding, the packing density of atoms within duplex DNA
is greater than for equivalent atoms in related small molecules packing in crystalline lattices.^8 As a conse-
quence of the hydrophobic effect, the bases tend to propeller-twistin the horizontal plane so as to opti-
mize the buried surface area between neighbouring base steps. However, this feature is not sufficient in
itself to account fully for the observed sequence-dependent conformational variations in duplex DNA.
Aside from the general effect of hydrophobicity on the general shape of the molecule, electrostatic effects
are important components of DNA conformation, and they tune the detailed geometry of the base stack-
ing. Each of the aromatic atoms of the bases can be treated as sandwiches of charges, from which the con-
formational propensities of the heterocyclic bases can be calculated based on the optimized electrostatic
interactions (Section 2.3.1). When combined with a simulated ‘backbone link’, these calculations can
account for many of the experimentally observed conformations of different base steps.^5
Dipolar effectsalso contribute to DNA conformation and protein–DNA interactions, but the importance
of this effect is difficult to quantify since it depends greatly on the organization of fixed charges and on the
dielectric constant. For example, a permanent dipole is associated with the G bases because of the
390 Chapter 10