De novo draft assembly of the Botrylloides leachii genome provides further insight into tunicate evolution

Abstract

Tunicates are marine invertebrates that compose the closest phylogenetic group to the vertebrates. These chordates present a particularly diverse range of regenerative abilities and life-history strategies. Consequently, tunicates provide an extraordinary perspective into the emergence and diversity of these traits. Here we describe the genome sequencing, annotation and analysis of the Stolidobranchian Botrylloides leachii. We have produced a high-quality 159 Mb assembly, 82% of the predicted 194 Mb genome. Analysing genome size, gene number, repetitive elements, orthologs clustering and gene ontology terms show that B. leachii has a genomic architecture similar to that of most solitary tunicates, while other recently sequenced colonial ascidians have undergone genome expansion. In addition, ortholog clustering has identified groups of candidate genes for the study of colonialism and whole-body regeneration. By analysing the structure and composition of conserved gene linkages, we observed examples of cluster breaks and gene dispersions, suggesting that several lineage-specific genome rearrangements occurred during tunicate evolution. We also found lineage-specific gene gain and loss within conserved cell-signalling pathways. Such examples of genetic changes within conserved cell-signalling pathways commonly associated with regeneration and development that may underlie some of the diverse regenerative abilities observed in tunicates. Overall, these results provide a novel resource for the study of tunicates and of colonial ascidians.

Figure 1

B. leachii phylogenetic position and life cycle. (A) Schematic showing phylogeny of tunicates with respect to the chordate clade (consensus based on^,,). (B) Life cycle of B. leachii. The colony expands and grows by asexual reproduction (right loop). During favourable conditions such as warmer water temperatures, members of the colonies start sexual reproduction (left loop). The embryo develops viviparously within the colony in brood pouches until hatching. Motile larvae attach to nearby substrates and begin metamorphosis into oozooids. Abbreviations: zooid (z), system (y), tunic (c), vascular system (v), terminal ampullae (a), buccal siphon (p_b), atrial siphon (p_a), fertilized oocyte (o), notochord (n), larval tadpole (l), oozooid (z_o), bud (b), budlet (b_t), regressing zooid (r).

Figure 2

Comparison of tunicate genomes. (A) Clustering of orthologous protein sequences. Indicated are the number of cluster groups, each of which contains at least two proteins. (B) TreeMap representation of the overrepresented GO Biological Processes terms within the ortholog groups shared between B. leachii and B. schlosseri genomes but not with C. robusta, O. dioica and M. oculata. Each rectangle size is proportional to the GOrilla minimum hypergeometric p-value of each GO term. (C) Distribution of the three classes of GO terms for each species. The colour-codes (left) are common for the entire row.

Figure 3

Hox genes are dispersed and reduced in number within tunicate genomes. Schematic depicting Hox gene linkages retained in five tunicate genomes in comparison to the ancestral Hox complex, which included thirteen genes. Orthologous genes are indicated by common colours. Chromosome (chr) or scaffold number (S) is shown, along with gene ID when available for newly annotated genomes. For B. floridae and H. sapiens, the length of each Hox gene cluster is given in brackets, and for B. leachii, the total scaffold length is shown. If a gene ID is not available (for unannotated genes), the co-ordinates of the BLAST hit (for the homeobox protein domain) is either given in File S5 or shown in the figure under the putative gene. Transcript IDs for B. leachii Hox genes identified in our transcriptome data are also provided in File S5.

Figure 4

Ancestral gene linkages remain between a few pharyngeal cluster genes in tunicate genomes. Schematic depicting the organization of the pharyngeal cluster genes among the studied chordate genomes. Double-parallel lines indicate >1 Mb distance between genes. Chromosome (chr) or scaffold (S) number is shown, along with gene ID when available for newly annotated genomes. Orthologous genes are indicated by common colours. Transcript IDs for B. leachii genes identified in our transcriptome data are provided in File S5.

Figure 5

NK homeobox cluster genes are fragmented within tunicate genomes. Schematic depicting the organization of the NK homeobox cluster genes among the studied chordate genomes. Double-parallel lines indicate >1 Mb distance between genes. Chromosome (chr) or scaffold (S) number is shown, along with gene ID when available for newly annotated genomes. Orthologous genes are indicated by common colours. Transcript IDs for B. leachii genes identified in our transcriptome data are provided in File S5.

Figure 6

Duplication of components of the Wnt signalling pathway in tunicate genomes. Schematic showing the organization of (A) the Wnt genes within each indicated genome and (B) of the downstream effectors. Note that no Wnt5 ortholog is present in the O. dioica genome. Genome browser images for the Wnt5 genes are shown in Fig. S4.

Figure 7

Tunicate Delta proteins. Bayesian phylogenetic tree depicting the relationship between tunicate and vertebrate DSL proteins, using Drosophila Delta to root the tree. Tunicate proteins are shown in bold and shaded areas correspond to Delta and Jagged groupings. Branch support values (probabilities) are indicated.

Figure 8

Evolution of the RA pathway in tunicates. (A) Overview of the RA synthesis and degradation pathway. In bold are the major proteins that contribute to RA signalling during animal development. Indicated below these are changes to the number of copies present in examined genomes. (B) Maximum likelihood phylogenetic tree depicting the relationship between invertebrate and vertebrate CYP26 proteins using CYP4 and CYP51 proteins as an out-group. Tunicate proteins are shown in bold. No Cyp26 gene has been identified in the O. dioica genome. Values for the approximate likelihood-ratio test (aLRT) are indicated.