Sea cucumber genome provides insights into saponin biosynthesis and aestivation regulation

Abstract

Echinoderms exhibit several fascinating evolutionary innovations that are rarely seen in the animal kingdom, but how these animals attained such features is not well understood. Here we report the sequencing and analysis of the genome and extensive transcriptomes of the sea cucumber Apostichopus japonicus, a species from a special echinoderm group with extraordinary potential for saponin synthesis, aestivation and organ regeneration. The sea cucumber does not possess a reorganized Hox cluster as previously assumed for all echinoderms, and the spatial expression of Hox7 and Hox11/13b potentially guides the embryo-to-larva axial transformation. Contrary to the typical production of lanosterol in animal cholesterol synthesis, the oxidosqualene cyclase of sea cucumber produces parkeol for saponin synthesis and has "plant-like" motifs suggestive of convergent evolution. The transcriptional factors Klf2 and Egr1 are identified as key regulators of aestivation, probably exerting their effects through a clock gene-controlled process. Intestinal hypometabolism during aestivation is driven by the DNA hypermethylation of various metabolic gene pathways, whereas the transcriptional network of intestine regeneration involves diverse signaling pathways, including Wnt, Hippo and FGF. Decoding the sea cucumber genome provides a new avenue for an in-depth understanding of the extraordinary features of sea cucumbers and other echinoderms.

Fig. 1. Genome landscape and polymorphism analysis of the sea cucumber Apostichopus japonicus.

From outer to inner circles: I, marker distribution on 22 chromosomes at the Mb scale; II and III, single nucleotide polymorphism (SNP) density across genome (II) or CDS (III) drawn in 1-Mb sliding windows with a 50-kb step (yellow columns), and polymorphism hotspot regions (p < 1e–4) are colored red; IV and V, gene density and repeat density across the genome, respectively, drawn in 1-Mb non-overlapping windows

Fig. 2. The cluster organization and tempo-spatial expression of Hox genes during the development of A. japonicus.

a Cluster organization of the Hox and ParaHox genes of the sea cucumber A. japonicus and other echinoderms^,,. The sea cucumber has a typical Hox cluster similar to that of the sea star A. planci, invalidating the previous hypothesis that Hox clusters of all echinoderms are reorganized^,. The genes whose identity existed ambiguously are shown with rectangles in dashed lines. b Temporal expression of the sea cucumber Hox and ParaHox cluster genes. Contrary to their ParaHox counterpart, expression of the sea cucumber Hox cluster during development does not exhibit temporal colinearity as typically found in chordate Hox clusters. Compared with other Hox genes, Hox7 and Hox11/13b show prominent expression during gastrulation, likely participating in determination of the larval body plan. c Inferred Hox cluster evolution and spatial expression of Hox7 and Hox11/13b in sea cucumber and sea urchins. Presumably the common ancestor of holothuroids and echinoids contained a typical Hox cluster without Hox6, which is largely preserved in the sea cucumber lineage (except the loss of Hox4) but had undergone a few rearrangements in the sea urchin lineage. The spatial expression of Hox7 and Hox11/13b shows colinearity during gastrulationalong the anterior/posterior (A/P) axis for the sea cucumber A. japonicus and along the oral/aboral (O/A) axis for sea urchins^–. This finding probably suggests the important roles of the two Hox genes in guiding axial transformation from the embryonic to larval stage in echinozoans

Fig. 3. Saponin biosynthesis and convergent evolution of LAS genes in A. japonicus.

a Gene representation of the canonical animal cholesterol synthesis pathway in the sea cucumber genome. The sea cucumber lacks two genes, Cyp51 and Dhcr7, suggesting that it might have lost its de novo cholesterol synthesis ability, consistent with the previous observation of extremely low cholesterol content in sea cucumber. b Evolutionary analysis of the lanosterol synthase (LAS) genes in sea cucumber and other animals. Compared to the inferred ancestral bilaterian LAS sequence (see Methods for details), LAS1 and LAS2 in sea cucumber show the highest sequence divergence and possess more putative plant sites than most other animals. c Overview of plant-like motifs in the sea cucumber LAS sequences and comparison with those of the animal consensus LAS sequence and plant consensus BAS and CAS sequences. The plant-like motifs in the sea cucumber LAS sequences are not present in sea urchin and starfish, suggesting the de novo acquirement of these motifs in the sea cucumber lineage. d Product determination by yeast expression of sea cucumber LAS1 and LAS2. Contrary to the general expectation that animal LAS produces lanosterol, neither sea cucumber LAS1 nor LAS2 produces lanosterol. Sea cucumber LAS1 produces parkeol (previously identified as the triterpene precursor of sea cucumber saponins;), whereas LAS2 produces 9β-lanosta-7, 24 dienol. e Summary of the pathways leading to saponin biosynthesis and steroid biosynthesis. In contrast to the plant kingdom, saponin biosynthesis is rarely found in the animal kingdom^,. The extraordinary ability of saponin synthesis in sea cucumber is enabled by its modification of lanosterol synthase, which possibly occurred through convergent evolution

Fig. 4. Key regulators and transcriptional network of aestivation in A. japonicus.

a Identification of differentially expressed genes (DEGs) and differentially expressed transcriptional factors (DE-TFs) in four organs (body wall, muscle, respiratory tree and intestine) during different states of aestivation. Venn diagrams and histograms show the shared gene numbers between organs and the absolute gene numbers in each organ. Among the four organs, body wall shows the most DEGs and DE-TFs, representing the most responsive organ during sea cucumber aestivation. b Expression profiles of nine TFs showing differential expression during aestivation in all four organs. Compared with other TFs, Klf2 and Egr1 are the most significant TFs, especially in body wall, likely playing important roles in the regulation of aestivation. Aestivation states: non-aestivation (Non_aes); early aestivation (Early_aes); deep aestivation (aes); and arousal from aestivation (Aro). c Expression heatmap of Klf2, Egr1 and clock-related genes during different aestivation states (according to quantitative PCR results), and the inferred clock gene-controlled regulation model. Klf2 and Egr1 may trigger the upregulation of Cry1 (either directly or indirectly through Clock and Bmal1) during sea cucumber aestivation, which propels the animal into an extended sleep phase, and decreased Cry1 expression makes the animal awaken from aestivation. Aestivation states are the same as depicted in (b) except Pre_aro representing initial arousal from aestivation. d Co-expression TF network of the aestivation-responsive model AM7. Klf2 and Egr1 are recognized as hub transcription factors in the network. The TFs showing differential expression in all organs are labeled in red, whereas for the remaining DE-TFs showing differential expression in at least one organ are labeled in yellow. e KEGG enrichment analysis of the AM7 module. The AM7 module governs diverse gene pathways, including those participating in cell proliferation and differentiation, seasonal rhythmicity and immune responses, suggesting the complex mechanism of molecular regulation during sea cucumber aestivation. The circle size and filled portion represent the gene numbers (from the AM7 module) and percentage of differentially expressed genes (DGEs) in a given pathway, respectively. The statistical significance is colored according to Q values

Fig. 5. Epigenetic regulation of intestine hypometabolism and participation of the expanded Fgfr family in intestine regeneration of A. japonicus.

a Identification of differentially methylated sites during different aestivation states. The intestine of sea cucumber shows prominent hypermethylation during aestivation. b Expression profiles of significantly hypermethylated genes (HMGs), showing the overall transcriptional suppression of these HMGs. c KEGG enrichment analysis of HMGs. HMGs are involved in numerous metabolic pathways, suggesting that intestine hypometabolism is caused by transcriptional suppression of metabolic pathways mediated through DNA hypermethylation. d The phylogeny of the Fgfr gene family in sea cucumber and other animals. The Fgfr gene family shows significant expansion in the sea cucumber genome (38 in contrast to 4–13 in other echinoderms or chordates). The expanded gene members in sea cucumber mostly form a separate clade (indicated by the red cluster). Numbers above the branches are support percentages for 1000 bootstrap replicates. The accession numbers or IDs of corresponding genes displayed in the tree are provided in Supplementary Table S37. e Expression heat maps of the Fgfr genes in sea cucumber intestine during different stages of aestivation and regeneration. Expression of the Fgfr genes is mostly suppressed during aestivation (corresponding to intestine atrophy), whereas it is activated during the regeneration process. f Expression of the FGF signaling pathway during intestine regeneration in sea cucumber. The potential roles of the FGF signaling pathway in intestine regeneration are supported by observation of the activation of various downstream cascades during the intestine regeneration process in sea cucumber. Fold-change (regeneration stages vs. the control stage) is color-coded