A Multidisciplinary Approach to RNA Localisation - Biophysical Approaches to Translational Control of Gene Expression - page 219

endogenous RNA binding interactions. Purification of the RNA:protein complexes
is then followed by sequencing of the RNAs. A digestion step also allows the exact
RNA fragment binding the protein to be identified (Konig et al. 2010 ) . One limita-
tion, however, is that the cross-linking only works for single stranded RNA targets,
while many RNA binding targets contain structured elements. There are also biases
for uracil tracts in the RNA (Ule et al. 2005b ), certain amino acids in the protein
(Williams and Konigsberg 1991 ) and requirement of a favourable topology of the
interaction to make the covalent bond (Maquat and Kiledjian 2008 ) . Photoactivatable
ribonucleoside-enhanced CLIP (PAR-CLIP) method uses modified photoactivat-
able nucleotides with increased cross-linking efficiency to overcome some of these
limitations (Hafner et al. 2010 ). A recent analysis of the CLIP and PAR-CLIP meth-
ods found both to be able to resolve binding sites on a single nucleotide level, but
also found biases introduced by the digestion steps (Kishore et al. 2011 ) .

11.3.3 Computational Characterisation of RNA:

Protein Binding Preferences

There are several computational methods for the prediction of the amino acids in
RNA binding proteins that directly interact with the RNA: RNABindR (Terribilini
et al. 2006 ), NAPS (Carson et al. 2010 ) and PiRaNhA (Murakami et al. 2010 ) . These
methods are trained on databases of experimentally determined structures of
RNA:protein complexes and are based on amino acid properties such as hydropho-
bicity, accessibility and substitution scores. However none of these methods directly
take into account the RNA itself, so the RNA sequence and secondary structure infor-
mation are largely ignored. The inclusion of the secondary structure prediction of the
RNA targets has been shown to greatly enhance the prediction of the binding sites of
RNA binding proteins (Li et al. 2010 ), as the nucleotides in the target RNA must be
accessible to bind to the individual RBDs. The inclusion of binding affinity data
further improves the determination of binding preferences for RNA binding proteins.
However this is at the level of single RBDs and their individual RNA target motifs.

11.3.4 Structure Determination of RNA Localisation Transport

Particle Components

The first structure to be determined for a transport particle component was that of
Staufen (Ramos et al. 2000 ; Bycroft et al. 1995 ) . Staufen contains fi ve dsRBDs and
was also the first protein component to be identified being required for RNA locali-
sation (St Johnston 1995 ). The NMR structure of the third dsRDB domain from
Drosophila Staufen revealed the critical residues for RNA binding and the mutation
of these residues abolished localisation with the full length protein. However the