structures from NMR data has been an important issue until recently.12,13Approaches are now available to
enable the measurement of residual dipolar couplings, which occur viaa through-space process but over
a longer range than nOes. To measure such couplings, it is necessary to partially align the nucleic acid
molecule (not just the nuclear spin) with Bo, in a manner akin to a crystal lattice. To achieve molecular
alignment, it is generally necessary for an agent to be introduced that has a preferred orientation, such as
phospholipid bicellesor filamentous phage Pf1, to induce partial alignment of nucleic acid sequences.
Such approaches are now aiding the conformational analysis of nucleic acids. NMR structures are now
being generated of structural motifs, such as stem-loops, G-tetraplexes, I-motifs, pseudoknots and
triplexes (Chapter 2),^14 as well as complexes of nucleic acids with drugs (Chapter 9) and proteins (Chapter 10).
Although the upper molecular mass limit for NMR structure determination for nucleic acids is around
25kDa, a 100kDa RNA has been studied by NMR.^15 Here only part of the RNA sequence was^13 C and^15 N
enriched, thus simplifying the assignment procedure and allowing the RNA fragment to be studied within
a larger molecule.
11.3 X-ray Crystallography
X-ray crystallography is the key method for determination of the 3D structure of biological macromolecules
at atomic resolution, and therefore has had a major impact on our understanding of the structure and function
of the nucleic acids.
In three decades of crystallographic structure determination, hundreds of structures of DNA, RNA and
protein–nucleic acid complexes have accumulated, as witnessed by the 2334 structures deposited in the
Nucleic Acid Database (NDB) as of March 16, 2004 (http://ndbserver.rutgers.edu),^16 and the 24,785 structures
in the Research Collaboratory for Structural Biology/Protein Data Bank (RCSB/PDB) as of March 23, 2004
(http://www.rcsb.org/pdb).^17 Dramatic advances have been made in virtually all areas of X-ray crystallography,
including crystallization (sparse-matrix crystallization screens) and crystal handling (cryo-protection), data
collection and resolution (synchrotron radiation and CCD detectors), phasing (multi-wavelength anomalous
dispersion, MAD), electron density map interpretation and model building (computer graphics and automatic
chain tracing) and structure refinement (more computer power, simulated annealing and full-matrix least-
squares refinement) (Figure 11.11a–e).18–22
The first step is the production of well-diffracting crystals. While oligonucleotides are relatively easy to
crystallize, crystals with good diffraction qualities are less common. In order to obtain crystals with the
necessary diffraction characteristics, it has proven desirable to consider the specific properties of a DNA and
RNA molecule, to determine the physicochemical parameters that favour 3D assembly, and to analyse the
inter-molecular contacts of the packing. In contrast to proteins, the packing environment of nucleic acids can
stabilize or induce structural transitions. In general, it is necessary to vary the particular construct (lengths,
sequence, blunt ends vs.dangling ends, etc.) and to test a large number of crystallization conditions.
Single-crystal diffraction data contain all the information needed to reconstruct the 3D structure of the unit
cell and molecules in that structure except for the phase information, which can still present a formidable hur-
dle to solving some structures (Figure 11.11). The basic approaches to determining phase information are
multiple isomorphous replacement(MIR), multi- and single-wavelength anomalous dispersion(MAD
and SAD, respectively), and molecular replacement(MR). The first two techniques require the incorporation
of ‘heavy’ atoms or heavy atom–containing derivatives into specific positions within the molecular lattice.
Such incorporation can be achieved by either soaking the crystals in a solution containing the heavy atom, by
crystallizing the compound in the presence of the heavy atom, or by incorporation of a heavy atom into the
compound by chemical synthesis or modification, for example 5-bromo-2 -deoxyuridine. For MIR to be suc-
cessful, it is an absolute requirement that crystals containing heavy atoms are of the same form as the original,
so that their X-ray diffraction patterns can be used together with the diffraction pattern from the non-derivatised
crystal to identify the positions of heavy atoms. The consequent phase and amplitude information can be
used to calculate the phases of all other reflections. In many oligonucleotide structure determinations, the
phase problem can be circumvented by use of MR, where a related structure or a model structure is used to
438 Chapter 11