142 7 Splicing and Alternative Splicing Impact on Gene Design
It is important to note that the same splicing factors can stimulate as well as
inhibit the inclusion of cassette exons depending on their respective binding site
position. This has been shown for SR proteins (see previous text) and hnRNPs.
For example, hnRNP H has been shown to promote exon inclusion when bound
to intronic positions, but induce exon skipping when bound to exonic sequences
[147, 148]. The hnRNP-like splicing factor Nova1 is exclusively expressed in CNS
neurons and recognizes YCAY clusters. A genome-wide map revealed that the
position of its binding site relative to the regulated exon dictates if Nova1 pro-
motes exon inclusion or skipping [118].
The accumulated knowledge on the impact of cis-regulatory motifs, exon fea-
tures (e.g., length, splice site strength), and RNA structure was successfully com-
bined to build a “splicing code” that accurately predicts tissue-specific expression
of alternatively spliced cassette exons [149].7.5.4 Transcription-Coupled Alternative Splicing
Splicing is not only controlled by a plethora of different splicing factors, but it is
also coupled to transcription, already shown in early studies [150]. Global
sequencing analyses of multiple tissues and cell types in different organisms indi-
cate that co-transcriptional splicing is widespread ([25, 151–157], reviewed in
[158]). In budding yeast, fly, and human cell lines and tissues, the vast majority of
introns are co-transcriptionally spliced [25, 151, 154, 156, 157]. Due to their
experimental and analytical differences, it is sometimes hard to compare the
studies. While some findings show that intron length negatively correlates with
co-transcriptional splicing frequency in mouse, human, and fly [25, 155, 156],
another study, focusing on highly expressed genes with long introns, came to the
exact opposite conclusion [151]. However, numerous studies agree that constitu-
tive splicing occurs to a greater degree in a co-transcriptional manner than alter-
native splicing [25, 151, 155, 156]. One study in mouse macrophages found that
full-length yet incompletely spliced transcripts accumulated in the chromatin
fraction [152]. The relatively low frequency of co-transcriptional splicing in this
and in another mouse study [155] is in contrast to the high numbers found in
yeast, fly, and human cells. To provide clear evidence, analysis of directly compa-
rable human and mouse cell types should be addressed.
The alternative splicing decision can be influenced by several elements,
including promoters [159, 160], transcription factors [161, 162], and coactivators
[163–165], as well as transcription enhancers [166], chromatin remodelers [167],
and factors affecting chromatin structure [168–172]. Two models are currently
discussed that are not mutually exclusive: the recruitment model and the kinetic
model (reviewed in [173]).
The recruitment model involves the recruitment of splicing factors to tran-
scription sites by the transcription machinery. The carboxy-terminal domain
(CTD) of RNA polymerase II (Pol II) has a key role in functionally coupling tran-
scription to capping and 3′ processing. Additionally, several alternative splicing
factors associate with the CTD, implicating this domain in alternative splicing.
One example is the splicing factor SRSF3, which interacts with the CTD and
inhibits inclusion of cassette exon 33 in the fibronectin mRNA [174].