Chromosome-level genome assemblies of 2 hemichordates provide new insights into deuterostome origin and chromosome evolution

Abstract

Deuterostomes are a monophyletic group of animals that includes Hemichordata, Echinodermata (together called Ambulacraria), and Chordata. The diversity of deuterostome body plans has made it challenging to reconstruct their ancestral condition and to decipher the genetic changes that drove the diversification of deuterostome lineages. Here, we generate chromosome-level genome assemblies of 2 hemichordate species, Ptychodera flava and Schizocardium californicum, and use comparative genomic approaches to infer the chromosomal architecture of the deuterostome common ancestor and delineate lineage-specific chromosomal modifications. We show that hemichordate chromosomes (1N = 23) exhibit remarkable chromosome-scale macrosynteny when compared to other deuterostomes and can be derived from 24 deuterostome ancestral linkage groups (ALGs). These deuterostome ALGs in turn match previously inferred bilaterian ALGs, consistent with a relatively short transition from the last common bilaterian ancestor to the origin of deuterostomes. Based on this deuterostome ALG complement, we deduced chromosomal rearrangement events that occurred in different lineages. For example, a fusion-with-mixing event produced an Ambulacraria-specific ALG that subsequently split into 2 chromosomes in extant hemichordates, while this homologous ALG further fused with another chromosome in sea urchins. Orthologous genes distributed in these rearranged chromosomes are enriched for functions in various developmental processes. We found that the deeply conserved Hox clusters are located in highly rearranged chromosomes and that maintenance of the clusters are likely due to lower densities of transposable elements within the clusters. We also provide evidence that the deuterostome-specific pharyngeal gene cluster was established via the combination of 3 pre-assembled microsyntenic blocks. We suggest that since chromosomal rearrangement events and formation of new gene clusters may change the regulatory controls of developmental genes, these events may have contributed to the evolution of diverse body plans among deuterostomes.

Fig 1. Highly conserved macrosyntenic structure among deuterostomes was detected based on chromosome-level genome assemblies, including 2 new hemichordate genomes.

(a) A simplified phylogenetic tree of major branches in Planulozoa (Cnidaria+Bilateria). (b) Macrosynteny conservation among deuterostome species including BFL, SCA, PFL, and SPU. Horizontal bars with numbers above represent chromosomes of each species. The conserved synteny blocks between 2 species are connected by curve lines (minimum of 4 gene pairs within a maximum distance of 75 genes between 2 matches). The data underlying this figure can be found in S1 Data. BFL, Branchiostoma floridae; PFL, Ptychodera flava; SCA, Schizocardium californicum; SPU, Strongylocentrotus purpuratus.

Fig 2. Evolutionary history of deuterostome chromosome architectures.

(a, b) A schematic representation of chromosome evolution in deuterostome lineages. The chromosomal architectures of presumed LCAs (bottom in a and left in b) and the chromosomal architectures of living deuterostome species (top in a and right in b). Each box denotes an individual chromosome. Haploid number (1N) and increase (+) or decrease (-) in quantity of chromosomes are indicated. The color code of boxes is taken from the previous study on vertebrate ancestral chromosomes, except for the 9 one-to-one corresponding chromosomes (a, light gray boxes). Chromosomal architecture of the LCA of vertebrates was based on the previous study [20]. In cases where chromosomal fusion events were deduced, types of changes are indicated below color boxes with symbols defined previously [19]; end-end translocation (●), centric insertion (↘), and fusion-with-mixing (⊗). Box sizes do not reflect the actual sizes of chromosomes. (c) Chromosomal positions of the orthologous gene pairs among 5 deuterostome species. Horizontal bars with numbers on top represent chromosomes of each species. In total, 3,668 orthologous gene pairs are illustrated. For ease of comparison, the chromosome sizes are scaled proportionally such that the 5 genome assemblies reach equal sizes. Except for the genes that spread into multiple chromosomes in amphioxus (BFL), gene pairs that are not located on the corresponding chromosomal pairs or cannot be found in all 5 species are not shown. The data underlying this figure can be found in S1 Data. BFL, Branchiostoma floridae; LCA, last common ancestor.

Fig 3. Category clustering analysis based on rearrangement events of the 24 bilaterian ancestral chromosomes.

(a) Three scenarios of phylogenetic relationships among bilaterians can be postulated. In the first scenario (monophyletic deuterostome), 2 deuterostome branches, chordates and ambulacrarians, are grouped together, and their LCA (the LCA of deuterostomes) is denoted with a blue dot. The LCA of bilaterians is indicated with a red dot. In the other 2 scenarios, one of the deuterostome branches is grouped with protostomes, resulting in polyphyletic deuterostomes. In these latter 2 scenarios, the LCA of deuterostomes and bilaterians is the same. (b) Distinct chromosomal rearrangement events of each species, including fusion, split, and spread events, are recorded into the category data based on changes deviated from the 1N = 24 BALGs. For example, there are 3 categories for the bilaterian ALG K, including (1) no rearrangement event (BFL, SPU, POC, SCA, PFL), (2) O2⊗K (RPH, SCO, PYE, PEC), and (3) J1⊗(O2⊗K) (SBE). (c) Conversion of the category data into a binary data matrix. Dark vertical lines distinguish different chromosomes. Red box denotes the chromosomal status of each species as compared to the BALGs. For the 10 species that we examined, the number of the chromosomal rearrangement categories ranges from 2 (BALGs F, G, N, and I) to 8 (BALG B2). A detailed binary code table is provided in S1 Data. (d, e) Bayesian phylogenetic analysis (d) and clustering analysis (e) based on the binary data shows that the 5 deuterostome species (shaded in blue) are grouped together. The data underlying this figure can be found in S1 Data. BALG, bilaterian ancestral linkage group; LCA, last common ancestor.

Fig 4. Counts of TEs around the Hox-bearing genomic regions.

(a) Densities of all TEs (DNA + LTR + LINE + SINE) within the Hox cluster region and non-Hox region of the Hox-bearing chromosome/scaffold of each species. The percent differences in normalized TEs counts between the non-Hox region and the Hox cluster region are illustrated (dashed bars and red values). (b) Distributions of all TEs around the Hox gene cluster of each species. The bin size for each histogram is 10,000 bp. Dotted lines indicate the averaged TE densities of the Hox-bearing chromosomes/scaffolds. Color coding denotes division of Hox genes in “anterior” (dark blue), “group 3” (yellow), “middle” (green), and “posterior” (red) groups. The light blue crosses represent missing Hox genes. Double arrows in light blue indicate inversion events of Hox genes. The data underlying this figure can be found in S1 Data. LINE, long interspersed nuclear elements; LTR, long terminal repeats; SINE, short interspersed nuclear elements; TE, transposable element.

Fig 5. GO enrichment analyses of chromosomes that underwent lineage-specific changes in deuterostomes.

GO enrichment analysis of genes located in the sea star POC9 (a), sea urchin SPU1 (b), hemichordate PFL18 (c), and PFL9 (d). The echinoderm chromosomes (POC12 and SPU3) corresponding to deuterostome DALG R were also analyzed to understand the chordate-specific chromosomal dispersion (e). The enriched GO terms (adjusted p-value <0.1) are clustered and divided into different modules, and selected terms are underlined. The top 3 enriched BP GO terms of each module are shown in S17–S19 Figs. The data underlying this figure can be found in S2, S3, and S4 Data. BP, biological process; GO, gene ontology.

Fig 6. A possible evolutionary history of the pharyngeal gene cluster.

The pharyngeal gene architectures are shown for presumed last common ancestors at key phylogenetic nodes (a) and selected living metazoan species (b) (see S26 Fig for the complete dataset). Genes that are commonly linked together are shown in the same color; homeobox-containing genes, including nkx2.1, nkx2.2 and msx, are in green, mipol1 and foxa genes are in blue, and pax1/9 and slc25a21 are in red. The gray circles indicate genes that are located within the pharyngeal gene cluster. Double slashes are introduced when more than 3 genes are located in between 2 pharyngeal genes. Because pax genes of cnidarians and sponges do not show one-to-one correspondence with those of bilaterians, we surveyed the locations of all potential pax genes and found that none is linked with the other pharyngeal-related genes in cnidarians and sponges.