Variable presence of the inverted repeat and plastome stability in Erodium

Abstract

Background and aims: Several unrelated lineages such as plastids, viruses and plasmids, have converged on quadripartite genomes of similar size with large and small single copy regions and a large inverted repeat (IR). Except for Erodium (Geraniaceae), saguaro cactus and some legumes, the plastomes of all photosynthetic angiosperms display this structure. The functional significance of the IR is not understood and Erodium provides a system to examine the role of the IR in the long-term stability of these genomes. We compared the degree of genomic rearrangement in plastomes of Erodium that differ in the presence and absence of the IR.

Methods: We sequenced 17 new Erodium plastomes. Using 454, Illumina, PacBio and Sanger sequences, 16 genomes were assembled and categorized along with one incomplete and two previously published Erodium plastomes. We conducted phylogenetic analyses among these species using a dataset of 19 protein-coding genes and determined if significantly higher evolutionary rates had caused the long branch seen previously in phylogenetic reconstructions within the genus. Bioinformatic comparisons were also performed to evaluate plastome evolution across the genus.

Key results: Erodium plastomes fell into four types (Type 1-4) that differ in their substitution rates, short dispersed repeat content and degree of genomic rearrangement, gene and intron content and GC content. Type 4 plastomes had significantly higher rates of synonymous substitutions (dS) for all genes and for 14 of the 19 genes non-synonymous substitutions (dN) were significantly accelerated. We evaluated the evidence for a single IR loss in Erodium and in doing so discovered that Type 4 plastomes contain a novel IR.

Conclusions: The presence or absence of the IR does not affect plastome stability in Erodium. Rather, the overall repeat content shows a negative correlation with genome stability, a pattern in agreement with other angiosperm groups and recent findings on genome stability in bacterial endosymbionts.

Keywords: Erodium; Geraniaceae; chloroplast; genome evolution; inversions; inverted repeat; plastome; repeated sequences.

Fig. 1.

Phylogram depicting relationships among selected Erodium species. The maximum likelihood tree (score 70^914·4940 lnL) was generated from a concatenated alignment of sequences of 19 protein-coding genes (26 985 bp) for 18 Erodium species and the outgroup California macrophylla and represents the constraint topology for rates analyses. The two major clades (sensu Fiz et al., 2006) within the genus are labelled along with the long branch clade (LBC) and the four types of plastome characterized. The branch leading to LBC has been interrupted for concision. The number of inversions relative to California is given in parentheses after each species. Numerals at the nodes indicate divergence time estimates (Fiz et al., 2008). The two hatched and two solid symbols (+/–) indicating IR status represent paired events that are alternatives of each other as discussed in the text. Gene and intron losses are indicated on the relevant branches. The scale bar indicates the number of substitutions per site.

Fig. 2.

Genomic rearrangement correlates with repetitive DNA content in Erodium plastomes. Repetitive DNA was identified by BLAST search of each genome against itself using blastn under default parameters and an e-value of 1e⁻¹⁰. One copy of the IR was removed from genomes with an IR (C. macrophylla and three Erodium species). The number of genomic rearrangements was determined by counting rearranged co-linear blocks of genes in MAUVE alignments of each genome against the unrearranged Type 2 genome of E. carvifolium. Coloured points represent Type 1 (pink), Type 2 (black), Type 3 (blue) and Type 4 (green) plastomes.

Fig. 3.

Erodium species lacking the IR have all lost IRa. Annotated nucleotide alignment of the region formerly flanking the copy of the IR lost on the branch leading to Erodium. Pseudogenes are indicated by the (ψ) symbol. Upper histogram indicates nucleotide identities across the alignment. Nucleotide identities were calculated within and between plastome types and are reported to the left (within) and below (between) the alignment. For within-type comparisons (left) square brackets indicate which species were included in each comparison with the length of sequences compared (nt) and per cent identity. For Clade II (Type 2), species comparisons within subclades are also given (left). The bracket below indicates the extent of the region compared for identity values, and the length of these regions for each type is reported (nt) to the left of each group. For between-type comparisons, two Type 2 species (asterisks) were selected representing the two major clades within Clade II. Identity values were based on alignments of two species from each type, for a total of four species included in each between-clade comparison.

Fig. 4.

Schematic linear representation of Type 4 Erodium plastomes. The inferred ancestral plastome with gene annotations was adapted from Weng et al. (2014). Type 4 plastome maps were adapted from maps generated using OrganellarGenomeDRAW (Lohse et al., 2013). Pseudogenes are indicated by the (ψ) symbol. Thick, black lines indicate IR extent.