The Lingula genome provides insights into brachiopod evolution and the origin of phosphate biomineralization

Abstract

The evolutionary origins of lingulid brachiopods and their calcium phosphate shells have been obscure. Here we decode the 425-Mb genome of Lingula anatina to gain insights into brachiopod evolution. Comprehensive phylogenomic analyses place Lingula close to molluscs, but distant from annelids. The Lingula gene number has increased to ∼34,000 by extensive expansion of gene families. Although Lingula and vertebrates have superficially similar hard tissue components, our genomic, transcriptomic and proteomic analyses show that Lingula lacks genes involved in bone formation, indicating an independent origin of their phosphate biominerals. Several genes involved in Lingula shell formation are shared by molluscs. However, Lingula has independently undergone domain combinations to produce shell matrix collagens with EGF domains and carries lineage-specific shell matrix proteins. Gene family expansion, domain shuffling and co-option of genes appear to be the genomic background of Lingula's unique biomineralization. This Lingula genome provides resources for further studies of lophotrochozoan evolution.

Figure 1. Deuterostomic development of the brachiopod, L. anatina, and its close relationship to molluscs.

(a) Adult (shell length ∼4–5 cm). (b–i) Embryogenesis: egg (b), embryos at 4-cell (c), 16-cell (d), 32-cell (e) and 128-cell stages (f), blastula (g), late gastrula (h) and 2-pair cirri larva (i). Scale bar, 50 μm. bp, blastopore; cr, cirri; ct, chaeta; gt, gut; ml, mantle lobe; mo, mouth; pd, pedicle; sh, shell; st, stone. (j) Phylogenetic position of Lingula among lophotrochozoans (orange box; molluscs are blue; annelids are green). The tree was constructed using the maximum likelihood method with 150 one-to-one orthologues (46,845 amino-acid positions) with LG+Γ4 model. Circles at all nodes indicate 100% bootstrap support.

Figure 2. Evolution of the Lingula genome is revealed by comparative genomics of lophotrochozoan gene families.

(a) Venn diagram of shared and unique gene families in four metazoans. Gene families were identified by clustering of orthologous groups using OrthoMCL. The number in parentheses shows unique gene families compared among 22 selected metazoan genomes. (b) Gene family history analyses with CAFE. Divergence times were estimated with PhyloBayes using calibration based on published fossil data. Gene families expanded or gained (red), contracted or lost (green). (c) Frequency of pair-wise genetic divergence calculated with synonymous substitution rate (Ks) among all possible paralogous pairs in the Lingula, Lottia and Capitella genomes.

Figure 3. Expansion and expression of Lingula chitin synthase genes indicate roles in shell formation and digestion.

(a) Phylogenetic analysis of chitin synthase (CHS) genes using the neighbour-joining method with the JTT model (90 genes, 358 amino acids and 1,000 bootstrap replicates). Three-letter code: adi, coral (Acropora digitifera); aqu, sponge (Amphimedon queenslandica); bfl, amphioxus (B. floridae); cel, nematode (Caenorhabditis elegans); cgi, Pacific oyster (Cr. gigas); cte, polychaete (Ca. teleta); dme, fly (Drosophila melanogaster); dpu, water flea (Daphnia pulex); hro, leech (H. robusta); lgi, sea snail (L. gigantea); nve, sea anemone (Nematostella vectensis); pfu, pearl oyster (Pinctada fucata); sce, baker's yeast (Saccharomyces cerevisiae); tca, beetle (Tribolium castaneum); uma, corn smut fungus (Ustilago maydis). Numbers are Lingula gene IDs. (b) CHS genes detected with BLASTP among 17 selected metazoan genomes. It is noteworthy that CHS genes with myosin head domains are only present among lophotrochozoans (grey area). (c) The expression of Lingula CHS genes in embryonic stages and adult tissues (separated by a vertical dashed line). FPKM, fragments per kilobase of transcript per million mapped reads.

Figure 4. Comparative transcriptomics and genomics reveal different origins of biomineralization-related genes.

(a) Spearman's correlation coefficient (ρ) and hierarchical clustering analyses of transcriptome data from adult tissues of the brachiopod, Lingula, and Pacific oyster, Crassostrea, in which 6,315 orthologous gene pairs were identified. An adult Lingula is shown with the dorsal shell removed and the anus opening to the right. (b) Genes involved in formation of vertebrate bone, mollusc shell and Lingula shell are compared in biomineralization-capable metazoans. Hierarchical clustering was performed in vertebrate bone formation-associated genes. Numbers of genes analysed are indicated in the parentheses. Shark, C. milii; pearl oyster, Pinctada fucata; sea snail, L. gigantea. BMP, bone morphogenetic protein; FGF, fibroblast growth factor; SCPPs, secreted calcium-binding phosphoproteins; SPARCs, secreted proteins acidic and rich in cysteine.

Figure 5. BMP signalling may be involved in larval shell formation.

(a–h) Confocal images of Lingula larvae from one-pair cirri to three-pair cirri stages. (a–d) Filamentous actin (F-actin) staining shows the cellular structure of larvae. Cytoplasmic membranes and nuclei are labelled with CellMask (grey) and DAPI (blue), respectively. Inset in c shows the cell boundary at higher magnification. (e–h) Activation of BMP signalling is monitored by nuclear signals of phosphorylated Smad1/5/9 (pSmad, red). Inset in e shows nuclear pSmad signals at higher magnification. It is noteworthy that signals are localized at the margin of mantle lobes (arrows). Orientation of embryos is indicated at the bottom-right corner of each panel. cr, cirrus (cirri); mf, muscle fiber; ml, mantle lobe; mo, mouth; sh, embryonic shell; tn, tentacle. Scale bars, 50 μm.

Figure 6. Fibrillar collagens in Lingula and vertebrates have different origins.

(a) Phylogenetic analysis of the collagen triple helix region using the maximum likelihood method with the LG model (159 genes, 542 amino acids, 100 bootstrap replicates). Expression of gene models is supported by the shell proteome (square) and transcriptome (circle). Numbers indicate Lingula gene IDs. (b) Domain structure of selected collagens. Expression of proteins is shown in grey boxes. C4, type IV collagen C4 domain; COLFI, fibrillar collagens C-terminal domain; EGF, epidermal growth factor-like domain; EGF_CA, calcium-binding EGF domain; Fxa_inhibition, coagulation Factor Xa inhibitory site; VWC, von Willebrand factor type C domain. (c) Genomic organization of tandem-duplicated collagen genes expressed in mantle and shell. Arrows indicate the direction of transcription. Grey boxes denote exons.

Figure 7. Genes related to biomineralization expressed during Lingula shell formation.

A schematic illustration of genes involved in Lingula biomineralization identified in the present study. Genes are coloured by their known functions in shell or bone formation in molluscs and vertebrates, respectively. Dashed outlines indicate gene families expanded specifically in Lingula. BMPR, bone morphogenetic protein receptor; ECM, extracellular matrix; GAG, glucosaminoglycan; SEVP1, Sushi von Willebrand factor type A, EGF and pentraxin domain-containing protein 1; WVA, von Willebrand factor type A domain containing protein. Proteins with ion-binding domains are labelled with Ca²⁺, Fe²⁺ or Cu²⁺. P and S in white circles indicate phosphate and sulfate groups.