From evaluation of Cpone can calculate the possible contribution to binding free energy that arises from the
hydrophobic transfer of the drug from bulk solvent into the DNA minor groove ( Ghyd) as 110.4 kJ mol^1.
These data, as well as crystallographic studies, indicate that binding of dye to A3T3 induces negligible
conformational changes in either the DNA or the drug. In effect it is a rigid body interaction. Therefore the
value of Gconfcan be set to zero. The value of Grtcan be set at62.8 kJ mol^1 by use of arguments
discussed by Spolar and Record.^50 Thus the calculated free energy change for the Hoechst 33258–A3T3
interaction is 55.0 kJ mol^1 , very close to the experimental value of 48.9 kJ mol^1 for Gobs. Therefore,
by difference it is possible to estimate Gmol (the contribution to free energy from weak noncovalent inter-
actions between the drug and DNA), which is6.1 kJ mol^1.
This analysis suggests that for this particular ligand–DNA complex, noncovalent molecular interactions
such as hydrogen bonds make a net unfavourable contribution to binding free energy. Although this result is
apparently at odds with NMR and crystallographic studies, the two positions can be reconciled when one
considers the hydration state of the minor groove. Hydrogen-bond formation between the ligand and DNA
is at the expense of hydrogen-bond breakage between the DNA and site-specific waters located in the minor
groove. The net result is an iso-energetic exchange reaction involving Hoechst 33258 and water, such that
a near zero free energy change results. Instead, the major driving force for Hoechst 33258 binding to A3T3
is the hydrophobic transfer of the drug from bulk solvent into the duplex minor groove binding site. This
is reflected in the observed negative Cpand the positive binding enthalpy. The large energetic cost resulting
from losses in rotational and translational freedom that occurs upon complex formation is more than compen-
satedfor by the favourable contributions from hydrophobic transfer and polyelectrolyte effects. Thus affinity
might be best modulated by changes in drug hydrophobicity. However, hydrogen bond formation and
other noncovalent interactions are key modulators of specificity in drug–DNA interactions and they are also
important in fine-tuning the free energy of binding in response to the interactions available in a given DNA
binding site.
9.8 Intercalation VersusMinor Groove Binding
There are some interesting compounds that possess structural features that could, in principle, involve
binding by either intercalation or groove binding. In reality, one is favoured over the other. What are the
factors that dictate choice of binding mode? Specific examples are the ligand SN 6999 and the related
compound chloroquine (Figure 9.18).
Both molecules contain a fused bicyclic quinoline ring system, yet biophysical and structural studies
have shown that while chloroquine can intercalate, SN 6999 binds in the minor groove. SN 6999 has a
curved shape that fits in the minor groove and it forms hydrogen bonds in a similar manner to other groove
binding drugs. Modelling studies have shown that SN 6999 could form favourable contacts within a base
pair stack as well as possible electrostatic interactions with the phosphate backbone in an intercalation
complex. However, if such an intercalation complex formed instead of the groove-bound complex, it
would be at the expense of favourable free energy from hydrogen bonding, non-bonded contacts with the
groove walls, and solvent release from the minor groove. Hence for SN 6999, binding in the minor groove
is much more favourable energetically than intercalation, even though the latter is possible. Chloroquine
however does not have the correct hydrogen bond donor/acceptor groups or any optimal structure for appro-
priate minor groove interactions. Thus it binds to DNA by intercalation but only with low affinity.
SN 6999 is an example of a minor groove binder having a benzene-fused heterocycle. Conversely, there
are examples of unfused aromatic cations that bind by intercalation rather than groove binding. Netropsin,
distamycin, and Hoechst 33258 are all examples of ligands that have torsional freedom and this property
allows these compounds to adopt an optimal twist for minor groove binding. Strekowski and Wilson have
designed and synthesised a new group of diphenylpyridine compounds with unfused aromatic rings pos-
sessing torsional freedom. So one might expect them to be good minor groove-binding ligands (Figure 9.18).
Molecular modelling and X-ray crystallography indicate that the molecule has a twist of between 20° and
25° between the phenyl and pyrimidine planes. The existence of twisted rings along with cationic end
370 Chapter 9