Fig. 4.

Structure and alternative splicing of stickleback MSX2A. a Transcript diagram of MSX2A, showing exons (yellow segments) and intron (connecting line). Both a full-length transcript (above) and a shortened transcript based on an alternative 5 ^′ splice site can be observed by RT-PCR. Primer binding sites used for the allele-specific expression assay (P1 and P2) and for RT-PCR (P3 and P4) are shown. The region encoding the homeodomain (DNA-binding domain) is highlighted in purple. The intron spans approximately 580 bp (800 bp if the alternative splice site is used) and is not shown to scale. In the shortened transcript, a stop codon occurs 30 bp into the second exon, and the remainder of the exon is untranslated (shown in gray). b Selected parts of a sequence alignment between the freshwater allele and two versions of the marine allele, highlighting sites where nucleotide changes lead to amino acid changes. Nucleotide coordinates are listed above; amino acid positions are indicated below. The only consistent amino acid difference between the freshwater allele and both marine alleles is the E15G polymorphism. The nucleotide change at that site also creates a five-nucleotide poly-G tract (GGGGG) that is specific to the freshwater allele. c The freshwater poly-G tract is a motif preferred by hnRNP F/H family proteins, the binding of which can promote the usage of an adjacent 5 ^′ splice site [45]. Self-interaction between hnRNP proteins bound at two different sites can also work to define the intervening sequence as intronic [46]. The presence of poly-G tracts near both the normal 5 ^′ splice site and the alternative site in the freshwater allele (underlined) suggests that an hnRNP interaction could occur and promote usage of the alternative site. hnRNP heterogeneous nuclear ribonucleoprotein, RT-PCR reverse-transcription polymerase chain reaction