Biology Reference
In-Depth Information
VIII. SEARCHES FOR MAMMALIAN TRANSCRIPTS EDITED BY ADARS
Inosine occurs at a frequency of approximately one base in every 17,000 nucleo-
tides in mouse brain poly(A)
รพ
RNA (Paul and Bass, 1998). As this frequency of
inosine nucleotides could not be accounted for with known site-specific editing
events, this prompted the search for new ADAR substrates. Before the era of
deep-sequencing, many attempts were made to conclusively describe the list of
ADAR editing events in different model organisms through the detection of A-
G transition changes when comparing cDNA and genomic DNA sequences.
However, this approach proved problematic due to the high frequency of single
nucleotide polymorphisms (SNPs) and sequencing errors.
Levanon
used a bioinformatic approach to search for expressed
sequence tags (ESTs) containing A-Gmismatches (Levanon
et al.
, 2004). As the
substrate for ADAR is dsRNA, they limited their search to predicted double-
stranded regions in ESTs. This approach identified 12,723 putative editing sites in
1637 genes, 26 of which were validated by sequencing genomic DNA, and cDNA
from the same individual. Interestingly, 92% of editing sites identified using this
approach were located within
et al.
elements and 1.3% were in LINE elements.
Using a similar approach, Athanasiadis
Alu
(2004) searched databases
for clusters of multiple A-G changes within short sequences, reasoning that these
were unlikely to be SNPs or sequencing artifacts. They also found A-G changes
clustered at the site of
et al.
repeat element insertions. This approach identified
1445 mRNAs that were edited at 14,500 sites, which would be sufficient to
account for the amount of inosine observed in brain mRNA.
Alu
insertion favors
actively transcribed regions, often occurring in UTR regions and within intronic
regions. The edited
Alu
repeat elements were found to occur in tandem in inverted
repeat orientation such that when transcribed a dsRNA stem-loop structure
would form between the two inverted repeats and be edited by ADAR
(Fig. 3.3). Due to the extensive similarity between
Alu
elements, the stem-loop
structures formed are inherently stable, although bulges and mismatches were
edited more frequently than near-perfect duplexes reflecting the
Alu
substrate
preference of ADARs. Some editing events were found to alter splice sites
resulting in exonization of a partial or whole
in vivo
element, although the novel
alternative splice sites were usually used at a low frequency and rarely constitu-
tively (Athanasiadis
Alu
, 2004). Surprisingly, editing of transcripts containing
other known repeat elements was considerably lower than that of
et al.
Alu
elements,
which is likely to be due to the high level of conservation between
element
subgroups and therefore the ease of creating the required secondary structure.
The prevalence of editing of transcripts containing
Alu
elements was
further confirmed with similar large-scale bioinformatic approaches (Blow
Alu
et al.
,
2004; Kim
et al.
, 2004). It has been postulated that editing of repetitive elements
