result of both NMR and X-ray analysis. However, the computer prediction of tertiary interactions and
base-triples appears to be still beyond the scope of present methodology.
2.4.5 Triple-Stranded RNAs
The first triple-stranded nucleic acid was described in 1957 when poly(rU)poly(rA) was found to form a sta-
ble 2:1 complex in the presence of magnesium chloride. The extra poly(rU) strand is parallel to the poly(rA)
strand and forms Hoogsteen base-triples in the major groove of an A-form Watson–Crick helix. Triplexes of
2poly(rA)poly(rU) can also be formed while poly (rC) can form a triplex with poly(rG) at pH 6 which has
two cytidines per guanine, one of them being protonated to give the CxGC base-triple also seen for triple-
helical DNA (Figure 2.33). Base triples are also a very common feature of tRNA structure (Section 7.1.4).^62
The importance of added cations to overcome the repulsion between the anionic chains of the
Watson–Crick duplex and the poly-pyrimidine third strand is an essential feature of triple-helix formation.
Co^3 (NH 3 ) 6 and spermine are also effective counter-ions as well as the more usual Mg^2 .
Poly(rG) as well as guanosine and GMP can form structures with four equivalent hydrogen-bonded bases in
a plane, with all four strands parallel. It is not clear whether this structure has any relevance to RNA folding.
2.5 Dynamics of Nucleic Acid Structures
Any over-emphasis on the stable structures of nucleic acids runs the risk of playing down the dynamic
activity of nucleic acids that is intrinsic to their function. Pairing and unpairing, breathing and winding are
integral features of the behaviour of these species.^64
Established studies on structural transitions of nucleic acids have for a long time used classical physical
methods, which include light absorption, NMR spectroscopy, ultra-centrifugation, viscometry and X-ray
diffraction (Chapter 11). More recently, these techniques have been augmented by a range of powerful
computational methods (Section 11.7). In each case, the choice of experiment is linked to the time-scale
and amplitude of the molecular motion under investigation.
2.5.1 Helix-Coil Transitions of Duplexes
Double helices have a lower molecular absorptivity for UV light than would be predicted from the sum of
their constituent bases. This hypochromicityis usually measured at 25 6nm while CG base pairs can also
be monitored at 280 nm. It results from coupling of the transition dipoles between neighbouring stacked
bases and is larger in amplitude for AU and AT pairs than for CG pairs. As a result, the UV absorption
of a DNA duplex increasestypically by 20–30% when it is denatured. This transition from a helix to an
unstacked, strand-separated coil has a strong entropic component and so is temperature dependent. The
mid-point of this thermal transition is known as the melting temperature(Tm).
Such dissociation of nucleic acid helices in solution to give single-stranded DNA is a function of base
composition, sequence and chain length as well as of temperature, salt concentration and pH of the solv-
ent. In particular, early observations of the relationship between Tmand base composition for different
DNAs showed that AT pairs are less stable than CG pairs, a fact which is now expressed in a linear cor-
relation between Tmand the gross composition of a DNA polymer by the equation:
Tm X 0.41[% (C G)] (°C)The constant Xis dependent on salt concentration and pH and has a value of 69.3°C for 0.3 M sodium ions
at pH 7 (Figure 2.42).
A second consequence is that the steepness of the transition also depends on base sequence. Thus, melt-
ing curves for homo-polymers have much sharper transitions than those for random-sequence polymers.
This is because AT rich regions melt first to give unpaired regions, which then extend gradually with ris-
ing temperature until, finally, even the pure CG regions have melted (Figures 2.42 and 2.43). In some
64 Chapter 2