gentle right-handed helical twist (12°–20°, the rise is 6.2 Å) between covalently linked residues gives the
C-rich quadruplex a quasi 2D form. The structure has two broad and two narrow grooves. In the latter, the
anti-parallel backbone pairs within intercalated duplexes are in van der Waals contact.
Several crystal structures of C-rich sequences, such as d(CCCC), d(CCCT), d(TCCCCC), d(TCC),
d(CCCAAT) and d(TAACCC), have provided details of the conformation, stabilisation and hydration of
the i-motif.^42 One surprising finding is the absence of effective stacking between the cytosine rings of adja-
cent hemi-protonated CCbase pairs from intercalated duplexes, an obvious difference to the above G-rich
quadruplex. However, a systematic base-on-deoxyribose stacking pattern as well as intracytidine C–H...O
hydrogen bonds may partially compensate for the lack of effective base–base-stacking between layers of
cytosines. The most unusual feature of the i-motif is a systematic, potentially stabilising C–H...O hydro-
gen bonding network between the C2-endopuckered deoxyribose sugars of anti-parallel backbones.
2.3.7.2 The Hoogsteen Duplex. Virtually all nucleic acid duplexes studied in the last 20 years contain
either all GC base pairs or AT base pairs flanked by GC base pairs. Very little 3D-structural work has
been carried out on AT-rich sequences, although the functional relevance of such sequences is well known. For
example, the promoters of many eukaryotic structural genes contain stretches composed exclusively of AT
base pairs. Further, coding sequences in the yeast genome tend to be clustered with AT-rich sequences sep-
arating them, and AT-rich sequences are common in transposable elements (see Section 6.8.3). Crystal
structures of TATA boxes bound to the TBP revealed highly distorted B-form conformations of the DNA.
Fibre diffraction studies of AT-rich sequences provided indications for considerable structural polymorphism.
By contrast, the mostly canonical geometries observed for AT paired regions in the structures of oligonu-
cleotide duplexes may have resulted from the constraints exerted by the GC base pair clamps at both ends.
A new crystal structure determined by Juan Subirana and co-workers of the alternating hexamer
d(ATATAT) raises interesting questions with regard to the existence of double-stranded DNA species that
lack base-pairing of the Watson–Crick type and the possible biological relevance of such alternative DNA
conformations.^43 In the crystal structure, but apparently not in solution, two hexametric fragments adopt
anti-parallel orientation with Hoogsteen pairing between adenine and thymine (Figure 2.37). The
Hoogsteen duplex features an average of 10. 6bp per turn, similar to B-DNA, and all sugars adopt the
C2-endopucker. The diameter of the Hoogsteen duplex is also similar to B-form DNA and the minor
groove widths of the two duplexes differ only marginally. A unique characteristic of the Hoogsteen duplex
is the synconformation of purine nucleosides. This arrangement generates a pattern of hydrogen bond
donors and acceptors in the major and minor grooves that differs between B-DNA and Hoogsteen DNA.
It also confers on the latter – a less electro-negative environment in the minor groove – that may lead to
preferred interactions with relatively hydrophobic groups at that site. Hoogsteen DNA also differs clearly
from triple-stranded arrangements (Section 2.3.6). Thus, in the TxAT triplex, adenine is always in the anti
conformation. Moreover, T and A form the Hoogsteen pair in a parallel orientation while in the case of an
anti-parallel orientation (Figure 2.33a), base-pairing between A and T is of the reverse Hoogsteen type
(Figure 2.33b). Furthermore, in the Ax(AT) triplex, third-strand adenines always base pair with adenine
of the Watson–Crick base pair (Figure 2.33b). Therefore, the anti-parallel Hoogsteen DNA duplex found
for d(ATATAT) is not simply a component of the triple helical motifs.
2.4 STRUCTURES OF RNA SPECIES
As with DNA, studies on RNA structure began with its primary structure. This quest was pursued in par-
allel with that of DNA, but had to deal with the extra complexity of the 2-hydroxyl group in ribonucleo-
sides. Today, we recognise also that RNA has greater structural versatility than DNA in the variety of its
species, in its diversity of conformations and in its chemical reactivity. Different natural RNAs can either
form long, double-stranded structures or adopt a globular shape composed of short duplex domains con-
nected by single-stranded segments. Watson–Crick base-pairing seems to be the norm, though tRNA
structures have provided a rich source of unusual base pairs and base-triplets (Section 7.1.2). In general, it
DNA and RNA Structure 55