Draft genome of Dugesia japonica provides insights into conserved regulatory elements of the brain restriction gene nou-darake in planarians

Abstract

Background: Planarians are non-parasitic Platyhelminthes (flatworms) famous for their regeneration ability and for having a well-organized brain. Dugesia japonica is a typical planarian species that is widely distributed in the East Asia. Extensive cellular and molecular experimental methods have been developed to identify the functions of thousands of genes in this species, making this planarian a good experimental model for regeneration biology and neurobiology. However, no genome-level information is available for D. japonica, and few gene regulatory networks have been identified thus far.

Results: To obtain whole-genome information on this species and to study its gene regulatory networks, we extracted genomic DNA from 200 planarians derived from a laboratory-bred asexual clonal strain, and sequenced 476 Gb of data by second-generation sequencing. Kmer frequency graphing and fosmid sequence analysis indicated a complex genome that would be difficult to assemble using second-generation sequencing short reads. To address this challenge, we developed a new assembly strategy and improved the de novo genome assembly, producing a 1.56 Gb genome sequence (DjGenome ver1.0, including 202,925 scaffolds and N50 length 27,741 bp) that covers 99.4% of all 19,543 genes in the assembled transcriptome, although the genome is fragmented as 80% of the genome consists of repeated sequences (genomic frequency ≥ 2). By genome comparison between two planarian genera, we identified conserved non-coding elements (CNEs), which are indicative of gene regulatory elements. Transgenic experiments using Xenopus laevis indicated that one of the CNEs in the Djndk gene may be a regulatory element, suggesting that the regulation of the ndk gene and the brain formation mechanism may be conserved between vertebrates and invertebrates.

Conclusion: This draft genome and CNE analysis will contribute to resolving gene regulatory networks in planarians. The genome database is available at: http://www.planarian.jp.

Keywords: Conserved non-coding elements; Dugesia japonica; Genome; Nou-darake; Planarian.

Fig. 1

Kmer (k = 17) frequency analysis. a Kmer species frequency graph. The horizontal axis shows the depth of kmer species, and the vertical axis shows the percentage of each kmer species value (blue curve). b. Kmer individuals frequency graph. The horizontal axis shows depth of kmer individuals, the left vertical axis shows the percentage of each kmer individual’s value (blue curve), and the right vertical axis shows the accumulative frequency of the kmer individuals (red curve)

Fig. 2

One fosmid insert sequence of the gene Djth (DJF-016O13). This figure shows the alignment of genomic DNA and RNA sequencing reads to one fosmid insert sequence of the gene Djth (DJF-016O13) and the repeated elements annotation of this fosmid sequence by Repbase. Green arrowheads show the10th and 11th exon of the Djth gene on the fosmid insert sequence. Grey spots show matches between the sequencing reads and the reference fosmid sequence, and black spots show mismatches. Red bars represent repetitive sequences

Fig. 3

D. japonica genome assembly workflow

Fig. 4

Alignment between the Djth fosmid insert sequence and its corresponding genome scaffold. In the illustration of the Djth fosmid structure, green arrowheads show the 10th and 11th exons of the Djth gene in the fosmid (see Fig. 2). In the illustration of the Djth fosmid repeated sequences, red bars represent repetitive sequences. In the illustration of the alignment between Djth fosmid and scaffold, pink blocks show matched sequences between the fosmid and the corresponding scaffold, while white blocks show mismatches, which were mainly caused by gaps in the scaffold

Fig. 5

Category of gene ontology annotation

Fig. 6

Conserved non-coding elements between D. japonica and S. mediterranea

Fig. 7

Transgenic experiments suggested that Djndk CNE3 might be a regulatory element. a The arrowhead shows CNE3 inserted actGFP vector, CNE (140)-actGFP driven GFP to express in the anterior region at the end of gastrulation in transgenic Xenopus embryos (n = 132). b Putative transcription factor-binding motifs are boxed in different colors; those subjected to mutation analysis are indicated by asterisks. The detailed point mutation design of three transcription factor-binding motifs are shown by red colored words. c Mutation analysis of CNE3 (the 140 bp element). actGFP is an empty reporter construct that contains the β-actin basal promoter; wt (140) is the construct of CNE (140)-actGFP used in (a); mt1, m2 and mt3 are point mutations (Msx (M), Tcf/Lef (T), and Jun/Fos (J)) generated from wt (140), and the detailed mutation design is shown in (b). The bar chart shows the percentage of the embryos that showed GFP expression in the neural plate among total developed embryos injected with the vector constructs. Actual numbers of GFP-positive cases and total numbers of scored embryos are indicated in parentheses. The chi-square test showed that the percentage of positive cases in the wt (140) and the Jun/Fos mutant constructs are significantly different (P < 0.0001), whereas the differences observed in other cases were not significant (P > 0.05)