Biomedical Engineering Reference
In-Depth Information
RNA used in the structure determination was optimised for NMR rather than genuine
in vivo targets such as
oskar
,
prospero
and
bicoid
. The dsRDB domain was thought
to be a non-sequence specific interaction as the nucleotides are not accessible; how-
ever, two recent studies suggest this may not be the case. The first study is the NMR
structure of the 44 nt fs(1)K10 mRNA localisation signal and is also the first such
RNA signal to have its structure determined (Bullock et al.
2010
) . K10 localises to
the anterior in
Drosophila
oocytes helping set up the dorsoventral axis. The struc-
ture of the stem-loop was consistent with the secondary structure predictions and
unexpectedly revealed that the stem forms A¢ RNA. Typically double stranded RNA
is A form; however, the runs of stacked purines (As and Gs) in the stem lead to a
widened major groove allowing access to the bases and hence specific recognition
by RNA binding proteins (Fig.
11.2
). Another study has shown that adenosine
deaminase (ADAR2) is able to specifically recognise double stranded RNA in a
stem-loop (Ste fl et al.
2010
). Although ADAR2 is not involved in RNA localisation,
it shows that RNA binding proteins are able to specifically recognise double stranded
A-form RNA.
Finally, zipcode binding protein 1 (Zbp1) has been shown to bind and be required
for the localisation of b-actin mRNA (Ross et al.
1997
). Zbp1 contains two RNA
recognition motif (RRM) and four KH domains and specifically recognises a 28 nt
region of the b -actin 3 ¢ UTR. The structure of the human ortholog of ZBP1, IMP1
KH34, reveals that the RNA loops around the third and forth KH domains (Chao
et al.
2010
). The looping is required as the RNA binding faces of the KH domains
are on opposite sides of the protein, and is thought to aid in the assembly of addi-
tional proteins involved in post-transcriptional regulation.
11.3.5
RNA Tertiary Structure Prediction
The experimental determination of RNA structure by X-ray crystallography and
NMR spectroscopy is both costly and time consuming, resulting in an ever-widen-
ing gap between the sequence and structure databases. This has led to the devel-
opment of computational methods to bridge this gap. RNA folds in a hierarchical
manner, first into secondary structures then building up into tertiary structures
(Brion and Westhof
1997
). As with the RNA secondary structure methods, the
non-helical regions provide the greatest challenges. In the last 5 years there have
been several methods published for the prediction of RNA tertiary structures, with
the most promising methods outlined below. MC-Sym is part of a pipeline and
generates RNA tertiary structures from the secondary structures predicted by
MC-Fold (Parisien and Major
2008
). Databases of experimentally derived struc-
tures are used to generate small fragments, which are built up to generate first a set
of secondary structures, then a full tertiary structure. A recent comparison between
an NMR structure and an MC-Sym prediction of a 4 × 4 internal loop highlights
the promise of the computational prediction of tertiary structure (Lerman et al.
2011
). However a close match to the NMR structure was only selected with the
Search WWH ::
Custom Search