Figure 1. A schematic showing the U2- and U12-type splicing pathways. The dinucleotides at the end of the intron are highly conserved.

Figure 2. Binding of U1 snRNA to 5¢ss and features of 5¢ss in human. Panel A shows the base pairing between the 5¢ end of the U1 snRNA and the consensus 5¢ss sequence. denotes pseudo-uridine, which is a modified uridine nucleotide. The figure shows the conventional numbering of positions, relative to the exon–intron boundary, e.g. –1G refers to nucleotide G at position –1. Panel B shows a pictorial representation of the position weight matrix (PWM) of the human U2-type GT-AG 5¢ss; the bars represent the percentages of nucleotides at each position of the 5¢ss. The table in the figure shows the percentages of nucleotides at each position. Adapted with permission from Roca et al. (2008).

Figure 3. The consensus sequence of 5¢ss in different species. The species are organized according to evolutionary relationship trees that are based on 18S ribosomal RNA conservation (not drawn to scale). Trichomonas vaginalis, Streblomastix strix and G. lamblia introns have not been demonstrated to be spliced by the U2- or U12-type spliceosomes. Bacteria and archaea do not have spliceosomal introns.

Figure 4. Information content in the exonic and intronic portions of GT-AG U2 splice sites. Panel A shows the information content of the 5¢ss GT-AG U2 splice site in several species, whereas Panel B shows the information content of the 3¢ss GT-AG U2 splice site in the same species. Higher conservation is observed within exonic positions in 5¢ss of higher organisms (Homo sapiens, Mus musculus, Arabidopsis thaliana), whereas lower organisms (C. elegans, D. melanogaster) show more conservation in the intronic portions of 5¢ss. The 3¢ss do not show much conservation except right at the exon–intron boundary. C. elegans in 3¢ss shows some conservation in the intron due to pecularities of the U2AF binding to the PPT.

Figure 5. Phylogenetic trees for splice site types from various species, using the information content of the sites. Panel A shows the tree for 5¢ss from various species: A, A. thaliana; C, C. elegans; D, D. melanogaster; M, M. musculus and H, H. sapiens. Panel B shows the corresponding tree for 3¢ss. The 5¢ss predominantly cluster by type, whereas the 3¢ss cluster predominantly by species, suggesting a more malleable architecture for the 3¢ss and a more conserved structure for the 5¢ss.

Figure 6. Associations between pairs of positions within a 5¢ss. Mouse is shown below the diagonal, whereas human is shown above the diagonal. Each square cell represents a pair of nucleotides, e.g. the cell addressed by row –3A and column –2G represents the pair –3A–2G in human 5¢ss, whereas –2G–3A is a pair in the mouse 5¢ss. The maroon colour represents depletion of the pair compared to the expected number, whereas the blue colour represents excess in the pair compared to the expected number. Brighter colours mean stronger depletion or excess over the expected numbers. The patterns seen here are conserved across species, suggesting an evolutionary pressure, probably arising from the spliceosome. Adapted with permission from Roca et al. (2008).