RNA Processing in C. elegans - Caenorhabditis Elegans: Molecular Genetics and Development - page 191

( Zhang and Blumenthal, 1996 ). In subsequent studies, it was determined that
U2AF 65 recognized these residues in place of a lengthy polypyrimidine tract, while
U2AF 35 interacted specifically with the terminal CAG/R ( Hollins et al., 2005; Zorio
and Blumenthal, 1999a ). Since the remainder of C. elegans splicing machinery has
been conserved with respect to other metazoans, it is thought that, after U2AF
binding, intron splicing proceeds canonically.
Comparison of branch point and 3 0 splice site recognition in different phyla
suggests that they each recognize the intron border by different mechanisms, but
using orthologous proteins. In the yeast Saccharomyces cerevisiae, SF1/BBP tightly
binds the branch point, while there appears to be little consensus to direct U2AF
binding to the polypyrimidine tract, which is often missing altogether ( Rutz and
Seraphin, 1999; Wang et al., 2008 ). The AG at the 3 0 splice site may not be
recognized by any protein, since U2AF 35 is not present in this organism. In mammals
there are relatively weak consensus sequences for both U2AF 65 and BBP/SF1, so
splice-site recognition may be more combinatorial ( Berglund et al., 1998 ). In worms,
there is a tight binding consensus for both U2AF subunits right at the 3 0 splice site,
but not for BBP/SF1. We hypothesize that U2AF recognizes the UUUUCAG/R at the
3 0 splice site, and then the branch-point adenosine is chosen by its proximity to SF1/
BBP, which is perhaps already bound to U2AF, as it is in the fission yeast,
Schizzosaccaromyces pombe ( Huang et al., 2002 ). The fact that some C. elegans
introns contain noncanonical 3 0 splice sites may indicate that the uridine tract in the
UUUUCAG/R is sufficient to direct the splicing machinery to the correct site of
splicing ( Zhang and Blumenthal, 1996 ). This flexibility exists because U2AF 65 and
U2AF 35 work in combination to specify the 3 0 splice site of an intron and direct U2
snRNP binding ( Wu et al., 1999; Zorio and Blumenthal, 1999a ).
Intron length can also influence 3 0 splice site choice in C. elegans ( Zhang and
Blumenthal, 1996 ). In constructs containing 48 nt introns with the engineered 3 0
splice site UUUCAA/AAG, splicing occurred with equal frequency after either the
CAA or the AAG. However, when the intron length was increased to either 171 nt or
283 nt, splicing occurred predominantly after G. Furthermore, analysis of the splice
sites of 139 C. elegans introns suggests that there are actually two classes of introns:
frequent short introns, with an average length of 52 nt; and rarer long introns, with an
average length of 551 nt. Both classes have the same 5 0 splice site consensus
sequence but differ in the amount of variability ( Fields, 1990 ).
5. Paired Splice Sites
The 3 0 ends of some C. elegans introns contain partially duplicated sequences,
paired splice sites. Similar arrangements, which lead to a process called alternative
tandem splicing, have been described in other organisms ( Dou et al., 2006; Hiller
et al., 2004, 2007 ). In mammals, the 3 0 splice site of such an intron contains the
sequence NAGNAG. In this case, U2AF 35 can bind either of these sequences,
demarcating two potential 3 0 splice points. In worms, a similar phenomenon occurs
fairly frequently (Sullivan and Blumenthal, unpublished observations). In these