Next generation sequencing-based analysis of repetitive DNA in the model dioecious [corrected] plant Silene latifolia

Abstract

Background: Silene latifolia is a dioecious [corrected] plant with well distinguished X and Y chromosomes that is used as a model to study sex determination and sex chromosome evolution in plants. However, efficient utilization of this species has been hampered by the lack of large-scale sequencing resources and detailed analysis of its genome composition, especially with respect to repetitive DNA, which makes up the majority of the genome.

Methodology/principal findings: We performed low-pass 454 sequencing followed by similarity-based clustering of 454 reads in order to identify and characterize sequences of all major groups of S. latifolia repeats. Illumina sequencing data from male and female genomes were also generated and employed to quantify the genomic proportions of individual repeat families. The majority of identified repeats belonged to LTR-retrotransposons, constituting about 50% of genomic DNA, with Ty3/gypsy elements being more frequent than Ty1/copia. While there were differences between the male and female genome in the abundance of several repeat families, their overall repeat composition was highly similar. Specific localization patterns on sex chromosomes were found for several satellite repeats using in situ hybridization with probes based on k-mer frequency analysis of Illumina sequencing data.

Conclusions/significance: This study provides comprehensive information about the sequence composition and abundance of repeats representing over 60% of the S. latifolia genome. The results revealed generally low divergence in repeat composition between the sex chromosomes, which is consistent with their relatively recent origin. In addition, the study generated various data resources that are available for future exploration of the S. latifolia genome.

Figure 1. Size distribution and repeat composition of clusters generated by similarity-based partitioning of S. latifolia 454 reads.

Bars on the histogram represent individual clusters, bar sizes correspond to number of reads in the clusters and colors to the type of repetitive sequences. The cumulative proportion of clusters in the genome is shown along the X-axis.

Figure 2. Repeat identification in the clone of S. latifolia genomic DNA using similarity search against repeat-specific 454 read databases.

Sequence of the chromosome Y-derived BAC clone MS2–9d12F (109.5 kb, GenBank accession AB257588) containing genomic regions surrounding molecular marker MS2 was searched for similarities to a set of databases compiled from the 454 reads assigned to individual repeat families. The search was performed using the PROFREP server employing blastn program with an e-value cutoff of 1e⁻¹⁵. Numbers of similarity hits along the sequence were converted into copy numbers by their normalization to the genome sequencing coverage of 454 sequencing (0.056x). The plot colors correspond to different families of repetitive DNA and the position of the MS2 sequence is marked by a yellow triangle. Bars above the plot show the positions of putative retrotransposon coding domains detected by Ishii et al. .

Figure 3. Comparison of genomic proportions of individual repeat families in male and female genomes.

Symbols on the plot representing individual clusters are color-coded according to the repeat type (A) or proportion of GC in their sequences (B). The position along the X-axis corresponds to the average proportion of the repeat in male and female genomes, while its position along the Y-axis is determined by its relative abundance in male (M) and female (F) genomes. This is expressed as the repeat proportion in the male divided by the sum of its proportions in the male and female genomes, resulting in the value of 0.5 for sequences with the same proportions in male and female genomes (marked by a solid line on the graph). The dashed lines mark two- and three-fold enrichment of a sequence in the male (corresponding to the values of 0.67 and 0.75, respectively) and female (0.33 and 0.25, respectively) genome. The positions of the sequence quantification controls are marked by triangles, showing their expected (open symbols) and observed (black symbols) values.

Figure 4. Consensus sequences of S. latifolia satellite repeats reconstructed from the most frequent k-mers.

K-mer frequencies for each satellite were calculated from Illumina sequencing data and used to reconstruct their most conserved fragments. These fragments were merged to create full-length monomers as depicted in Fig. S4. The resulting consensus sequences are displayed as sequence logos where the height of the letters reflects the frequencies of the corresponding k-mers, and major sequence variants are displayed along with the prevailing nucleotides . The sequence regions used to design the FISH probes are marked with orange lines and the palindrome sequences common to the CL68 and CL156 repeats are marked with black arrows. The dashed underlines in the CL12 sequence show the positions of two regions that are duplicated in some of the X43.1 monomers.

Figure 5. Localization of satellite repeats on metaphase chromosomes of S. latifolia.

FISH experiments were performed simultaneously with two probes labeled by different fluorochromes (red or green, as indicated) in order to investigate the co-localization of sequence variants of the STAR repeats (A–B), 15Ssp and X43.1 (C) and TRAYC-like repeats (D–F). (A) Co-localization of the CL16/STAR-C repeat consensus (16HL2) and its sequence variant (16HL3). (B) The probes for a different region of the CL16/STAR-C consensus (16HL1) and for the chromosome Y-enriched subfamily CL398/STAR-Y (398H2). (C) Consensus probes for CL87/15Ssp (87H1) and X43.1 (12H5) satellites. Co-localization of the CL68/TRAYC-like consensus (68H1) with X43.1 (12H5) (D), CL156 (156H1) (E) and with the CL68 subfamily (68H2) (F). The inset in (F) is an example of the Y chromosome hybridized to 68H2 and 398H2, showing the localization of their interstitial signals on different chromosome arms. The chromosomes were counterstained with DAPI (blue). Sex chromosomes are indicated with X and Y. The positions of probes within repeat monomers are shown on Fig. 4 and their sequences are provided in Table 2.

Figure 6. Graph representation and annotation of clusters containing rDNA genes and the satellite X43.1.

(A) Graph layout displaying individual 454 reads as nodes and their similarities as connecting edges . The circular shape of the graph reflects the tandem organization of multiple copies of the rDNA repeat units (reconstructed in B) in the genome. The graph nodes representing sequence reads from various parts of the rDNA unit are distinguished by colors. Sequences representing X43.1 repeats form due to their high copy number and tandem arrangement globular structure on the graph located at the 5' end of the intergenic spacer (IGS) and surrounded by another tandem repeat (TR1). The graph was created using merged sets of reads from clusters CL12 (containing rDNA genes, X43.1 repeats and part of the IGS) and CL59 (remaining part of the IGS).