The molecular evolution of spiggin nesting glue in sticklebacks

Abstract

Gene duplication and subsequent divergence can lead to the evolution of new functions and lineage-specific traits. In sticklebacks, the successive duplication of a mucin gene (MUC19) into a tandemly arrayed, multigene family has enabled the production of copious amounts of 'spiggin', a secreted adhesive protein essential for nest construction. Here, we examine divergence between spiggin genes among three-spined sticklebacks (Gasterosteus aculeatus) from ancestral marine and derived freshwater populations, and propose underpinning gene duplication mechanisms. Sanger sequencing revealed substantial diversity among spiggin transcripts, including alternatively spliced variants and interchromosomal spiggin chimeric genes. Comparative analysis of the sequenced transcripts and all other spiggin genes in the public domain support the presence of three main spiggin lineages (spiggin A, spiggin B and spiggin C) with further subdivisions within spiggin B (B1, B2) and spiggin C (C1, C2). Spiggin A had diverged least from the ancestral MUC19, while the spiggin C duplicates had diversified most substantially. In silico translations of the spiggin gene open reading frames predicted that spiggins A and B are secreted as long mucin-like polymers, while spiggins C1 and C2 are secreted as short monomers, with putative antimicrobial properties. We propose that diversification of duplicated spiggin genes has facilitated local adaptation of spiggin to a range of aquatic habitats.

Keywords: Gasterosteus aculeatus; adaptation; gene duplication; gene family; nest building; retrotransposon.

Figure 1

Spiggin gene phylogenies and alignments. (A) Results from sliding‐window phylogenetic analysis of five Gasterosteus aculeatus draft genome spiggin transcript predictions annotated as A, B1, B2, C1 and C2. In profile 1, these transcripts were rooted with a MUC19 out‐group (O). Profiles 2 and 3 were rooted using spiggin A and show the impact of adding nonchimeric spiggin C1 (SpgC1, AB910016 from this study) and chimeric spiggin B3 (G. aculeatus draft genome) sequences, respectively. Posterior probabilities of the two most frequent topologies along the alignment are indicated by solid and dashed lines. Positions along the alignment in base pairs (bp) are set below a generalized in silico spiggin translation with the protein domains: von Willebrand factor type D domain, C8 domain and trypsin inhibitor‐like cysteine‐rich domain. A simplified view of the tandemly arrayed spiggin genes on chromosome IV of the stickleback draft genome is shown below the phylogenies. (B) Nucleotide alignment of a representative selection of spiggin transcripts sequenced in this study (West Sweden and Edinburgh, UK), along with all published spiggin sequences and the six spiggin transcript predictions from the draft genome assembly of G. aculeatus (Alaska) used in the sliding‐window phylogenies (A). Light grey indicates consensus between sequences, and black indicates nucleotide differences. WS, West Sweden; ED, Edinburgh; alt spl, alternatively spliced; chim, intrachromosomal chimeric; inter, interchromosomal chimeric. All spiggin transcripts have been further annotated with the spiggin A, B and C nomenclature.

Figure 2

Nucleotide alignments of Gasterosteus aculeatus and Pungitius pungitius interchromosomal spiggin chimerics with the 2006 draft (Broad/gasAcu1) assembly of G. aculeatus and parental genes. Alignments of the G. aculeatus (A) and P. pungitius (B) interchromosomal spiggin chimerics with spiggin B (AB910011) and the recruited region of the draft assembly. The spiggin region of the P. pungitius interchromosomal chimeric has not been annotated with a specific spiggin gene as the spiggin multigene family in P. pungitius has not been fully characterized. Putatively assigned introns are indicated in grey. Black indicates nucleotide differences between sequences within the alignment. Black triangles represent start codons, grey triangles represent stop codons, and the white triangles indicate polyadenylation signals (AAUAAA). Chr, chromosome.

Figure 3

Genomic and cDNA alignments of spiggin chimerics. (A) Nucleotide alignments of cDNA and genomic DNA of the spiggin B/ChrIX interchromosomal spiggin chimeric with the recruited regions of chromosomes IV and IX and spiggin B (AB910011). Black and grey triangles indicate start and stop codons, respectively. (B) Nucleotide alignments of intrachromosomal spiggin chimerics with parental genes, spiggin B, spiggin C1 and spiggin C2. Black indicates nucleotide differences between each sequence and the consensus sequence. Putatively assigned introns and alignments with parental genes are annotated below each chimeric gene. B, Spiggin B; IX, chromosome IX.

Figure 4

Nucleotide alignments of Gasterosteus aculeatus interchromosomal spiggin chimerics identified from previously published spiggin genes. Black indicates nucleotide differences between each sequence and the consensus sequence. Black triangles represent start codons, grey triangles represent stop codons, and the white triangles indicate polyadenylation signals (AAUAAA). Chr, chromosome; Inter‐con, inter‐contig (on the draft stickleback genome assembly).

Figure 5

The spiggin multigene family as annotated on the 2006 draft (Broad/gasAcu1) assembly of Gasterosteus aculeatus. RepeatMasker has been used to highlight Long Interspersed Nuclear Element‐1s (LINE‐1s) for Chr:groupIV 20672561–21551524 (A) and Chr:groupIV 21002172–21221912 (B). The spiggin transcript nomenclature as revised in this study is shown above each spiggin gene (B). Figure is adapted from Ensembl Genome Browser data. CR1, Chicken repeat 1; REX1, retrotransposable elements first described in X iphophorus fish genome.

Figure 6

Alignments of in silico translations of spiggin transcript putative open reading frames. Below each translation are annotated predicted O‐linked glycosylation sites (light blue), N‐linked glycosylation sites (purple) and dimerization/multimerization motifs (red arrows). Coloured boxes in the protein alignments represent the following domains: vWD, von Willebrand factor type D domain (blue); C, C8 domain (red); T, trypsin inhibitor‐like cysteine‐rich domain (yellow). Light grey indicates amino acid differences between each of the translations.

Figure 7

Proposed model of spiggin gene amplification by L1 retrotransposon‐mediated unequal crossing over, gene conversion and retrotransposition. Boxes 1 and 2 show how insertion of L1 retrotransposons on either side of the ancestral single‐copy MUC19 may have been responsible for the initial gene duplication through unequal crossing over. Box 3 indicates how further gene duplication and divergence could have led to the three major spiggin gene lineages, A, B and C. Box 4 indicates how retrotransposition has led to further gene duplication.