RNA Processing in C. elegans - Caenorhabditis Elegans: Molecular Genetics and Development - page 189

2. Noncanonical Splice Sites
As is the case in other metazoans, there are also a few examples of introns with
GC-AG splice sites in the C. elegans genome ( Farrer et al., 2002; Sheth et al.,
2006 ). They are processed in the same manner as the GT-AG introns, indicating
that the splicing machinery can also recognize introns with these borders.
Presumably the surrounding nucleotides are able to direct the U1 snRNP into
position in spite of the mismatched cytosine. Indeed, it has been observed that
GC-AG introns typically have 5 0 splice sites that, in all other regards, more
closely complement the splice site recognition sequence on the U1 snRNP.
Although many of these introns appear to be constitutively spliced, some have
been shown to be alternatively spliced. In these cases, the weaker 5 0 splice site
consensus created by the mismatched cytosine is thought to influence splice site
selection ( Farrer et al., 2002 ).
Other instances of splicing from noncanonical splice sites have been reported in C.
elegans, and additional examples can be found by inspection of sequences annotated
as splice sites in the genome. For example, it has been reported that introns with
mutations in the 3 0 splice site - such as AA, AT, GG, or TG - are nonetheless spliced
at these mutant sites, albeit less efficiently than when the wild-type sequence is
present ( Aroian et al., 1993 ). Also, there are numerous instances of use of non-AG 3 0
splice sites in wild-type C. elegans genes (unpublished observations; WormBase,
WS210 ). Furthermore, when genes containing a Tc1 transposon insertion are tran-
scribed, the transposon RNA is often spliced out from 5 0 splice sites as varied as TT
or AT, and from 3 0 splice sites like GG, TG, AC, or GC ( Rushforth et al., 1993;
Rushforth and Anderson, 1996 ). The mechanism by which these reactions occur is
not understood. However, it is clear from all of these findings that the cellular
machinery responsible for splice site recognition in C. elegans is significantly less
stringent than it is in other studied systems.
3. Intron Properties
These examples of noncanonical splicing indicate that, in spite of the extensive
conservation of the intron-splicing mechanism among higher eukaryotes, some intron
sequences and splicing factors have acquired specialized features in C. elegans.In
general, introns in this organism have an A+U content of about 70%, while the
average A+U content found in the exons flanking them is about 54%. Conversely,
there is little elevation in the A+U content of yeast and mammalian introns
( Blumenthal and Steward, 1997; Csank et al.,1990 ). Studies in C. elegans and in
plants have shown that steep transitions in A+U content assist in defining intron
borders, and that artificial introns with high C+G content are not efficiently spliced
( Conrad et al., 1995; Goodall and Filipowicz, 1989 ).
Introns in C. elegans tend to be shorter than those in other metazoans. While some
introns within the C. elegans genome are well over 1 kb long, the median intron
length is only 65 nt, and the most common intron size is 47 nt. By comparison, the