The complexity of alternative splicing of hagoromo mRNAs is increased in an explosively speciated lineage in East African cichlids

Abstract

The adaptive radiation of cichlid fishes in the lakes of East Africa is a prime example of speciation. The choice of cichlid mates on the basis of a variety of coloration represents a potential basis for speciation that led to adaptive radiation. Here, we characterize the cichlid homolog of the zebrafish hagoromo (hag) gene that was recently cloned and characterized from a pigmentation mutant. Although only one hag mRNA was reported in zebrafish, cichlids express nine different hag mRNAs resulting from alternative splicing. The hag mRNAs are expressed between the myotome and the epidermis where pigment cells are located, suggesting the cichlid hag gene is involved in pigmentation. The hag mRNA splicing pattern does not fluctuate among individuals from each of two species, suggesting that alternative splice site choice is fixed within species. Furthermore, cichlids in lineages that underwent explosive speciation expressed a greater variety of hag mRNAs than those in lineages that did not undergo such a degree of speciation, suggesting that species in the explosively speciated lineage acquired a complex regulatory mechanism of alternative splicing over a very short evolutionary period. Here, we provide an example in which alternative splicing may play a role in mate choice, leading to cichlid speciation through diversification of gene function by production of multiple mRNAs from a single gene.

Fig. 1.

(a) Structure of the cichlid hag. Positions of exons are indicated by numbers. Sequences at exon–intron boundaries are given for exons 1–5. Vertical lines delineate exon–intron junctions; conserved GT sequences for splicing are underlined, and termination codons are italicized. The checkered box indicates the conserved 3′UTR in the mRNAs terminating within intron 5 (i5terminal). (b) RT-PCR amplification of hag mRNAs. mRNAs terminating at exon 9 (e9terminal) and i5terminal mRNAs were amplified from M. auratus cDNA by RT-PCR (lanes 1 and 2). Positions of expected mobilities of amplified products of mRNAs are indicated to the right of lanes 1 and 2, respectively. Asterisks indicate intron 2- and 3-containing cDNA in lanes 1 and 2. (c) Nested PCR amplification of hag mRNAs. Intron-containing and exon-skipped mRNAs were amplified by nested PCR by using the RT-PCR products of e9terminal and i5terminal mRNAs as templates (lanes 1–8). The fragment was intron 2- and 3-containing e9terminal cDNA in lane 1. The longer fragments were intron 2- and 3-containing i5terminal cDNA in lane 5. “M” indicates a 200-bp ladder of size markers. (d) Schematic representation of hag mRNA structures. Exons are indicated by “e,” and introns are indicated by “i.” Half-arrows indicate positions of primers specific to each mRNA. Gray boxes represent introns.

Fig. 3.

Examination of intron and exon content from 14 cichlid species by using RT-PCR and nested PCR. The following mRNAs from various species were amplified by RT-PCR and nested PCR: (a) e9terminal-fl,(b) e9terminal-i2ret, (c) e9terminal-i3ret, (d) e9terminal-e3less, (e) i5terminal-fl,(f) the coding region of i5terminal, (g) i5terminal-i1ret, (h) i5terminal-i2ret, (i) i5terminal-i3ret, and (j) i5terminal-e3less. Filled arrows indicate expected mobilities of amplified products of each mRNA. (a) The longer and shorter fragments relative to e9terminal-fl were intron-containing and exon 3-less e9terminal cDNAs, respectively. (b) The longer fragments (open arrow), relative to e9terminal-i2ret, were intron 2- and 3-containing e9terminal cDNAs. (e) The longer and shorter fragments, relative to i5terminal-fl, were intron-containing and exon 3-less i5terminal cDNAs. (f) The longer and shorter fragments, relative to i5terminal coding region, were intron-containing and exon 3-less i5terminal cDNAs. (g) The longer fragments (open arrow), relative to i5terminal-i1ret, were intron 1- and 3-containing i5terminal cDNAs. (h) The longer fragments (open arrow), relative to i5terminal-i2ret, were intron 2- and 3-containing i5terminal cDNAs. Each of five individual D. compressiceps and D. strigatus specimens were used for lanes 15–19 and lanes 20–24, respectively. Refer to Fig. 4 for full names of species.

Fig. 4.

Phylogeny of the cichlid species examined in this study, showing the numbers of hag mRNA types found in each species and the speciation rates in each lineage. The phylogenetic tree is based on molecular data from several sources (, –47). The net speciation intervals (SI) for Lakes Malawi and Victoria haplochromine, Lamprologini, and river species (tilapiine) were calculated in previous works (31, 32). The SI for Tropheus was calculated on the basis of the method described in refs. and by using the age of radiation for Tropheus (48). N.D., no data, because the age of radiation for each lineage was not deduced.

Fig. 2.

Expression of hag. (a) A juvenile specimen of Haplochromis sp. from Lake Victoria at day 60 was used for in situ hybridization. (b) In situ hybridization by using full-length antisense RNA terminating at exon 9 (e9terminal-fl) as a probe (head and venter removed from specimen). The signal (blue) was detected in almost all areas of the skin. (c) Cross section through the solid line with arrow in b. The expression of hag (blue) was detected beneath the epidermis (e). The myotome is indicated by “m.” (d) Higher magnification of the bracketed region in c, showing hag expression between the m and the e between the arrowheads.