Mind the gap! The mitochondrial control region and its power as a phylogenetic marker in echinoids

Abstract

Background: In Metazoa, mitochondrial markers are the most commonly used targets for inferring species-level molecular phylogenies due to their extremely low rate of recombination, maternal inheritance, ease of use and fast substitution rate in comparison to nuclear DNA. The mitochondrial control region (CR) is the main non-coding area of the mitochondrial genome and contains the mitochondrial origin of replication and transcription. While sequences of the cytochrome oxidase subunit 1 (COI) and 16S rRNA genes are the prime mitochondrial markers in phylogenetic studies, the highly variable CR is typically ignored and not targeted in such analyses. However, the higher substitution rate of the CR can be harnessed to infer the phylogeny of closely related species, and the use of a non-coding region alleviates biases resulting from both directional and purifying selection. Additionally, complete mitochondrial genome assemblies utilizing next generation sequencing (NGS) data often show exceptionally low coverage at specific regions, including the CR. This can only be resolved by targeted sequencing of this region.

Results: Here we provide novel sequence data for the echinoid mitochondrial control region in over 40 species across the echinoid phylogenetic tree. We demonstrate the advantages of directly targeting the CR and adjacent tRNAs to facilitate complementing low coverage NGS data from complete mitochondrial genome assemblies. Finally, we test the performance of this region as a phylogenetic marker both in the lab and in phylogenetic analyses, and demonstrate its superior performance over the other available mitochondrial markers in echinoids.

Conclusions: Our target region of the mitochondrial CR (1) facilitates the first thorough investigation of this region across a wide range of echinoid taxa, (2) provides a tool for complementing missing data in NGS experiments, and (3) identifies the CR as a powerful, novel marker for phylogenetic inference in echinoids due to its high variability, lack of selection, and high compatibility across the entire class, outperforming conventional mitochondrial markers.

Keywords: Control region; Echinoidea; Mitochondrial markers; Molecular phylogeny; NGS; Sea urchins.

Fig. 1

Representation of echinoid complete mitochondrial genomes assembled from NGS data, showing gene annotation and coverage. The annotated genomes are represented by four echinoid species: Hemicentrotus pulcherrimus, Strongylocentrotus fragilis, Mesocentrotus franciscanus, and Strongylocentrotus intermedius, corresponding to GenBank accession numbers: KC898202, KC898198, KC898199, and KC898200, respectively. Annotations are given at the outer margin of the external circle. Concentric circles represent the corresponding coverage for each of the represented species mitogenomes. Data was obtained from Kober and Bernardi [86, 87]. Enlarged segment illustrates the position of the various primers used in the current study. Coverage was calculated in BRIG [88], after read mapping with Bowtie2 [89] (using the predefined alignment threshold “very-sensitive”). Annotations are based on those for H. pulcherrimus (GenBank accession no. NC_023771) and radial plots generated using BRIG

Fig. 2

Pairwise tree comparisons for phylogenetic trees based on commonly used mitochondrial markers. Trees include the two most commonly used phylogenetic mitochondrial markers: a fragment of the cytochrome c oxidase subunit 1 (a) gene and a fragment of the 16S ribosomal RNA (c) as well as the novel tRNAs and control region (e). To facilitate independent comparisons, the genetically inferred trees were restricted to the 35 publicly available complete echinoid mitochondrial genomes. Genera represented by more than one species were collapsed and are depicted by single branches. Supporting values (> 0.85 posterior probabilities and > 75% ML bootstrap values) are shown next to nodes. Topological comparisons between the genetically inferred trees and current classification (b, d, f) (see text for details) were visualised using Phylo.io [62]. Colour scale for the comparison metric (a variant of the Jaccard Index as implemented in Phylo.io) ranges from 0 (subtrees completely different) to 1 (subtree structure of the respective node is identical)

Fig. 3

Substitution saturation plot of the CRA marker based on the CRA-All dataset. The number of transitions (s) and transversions (v) is plotted against F84 genetic distance. A linear correlation is sustained for both transitions and transversions as expected in the absence of saturation

Fig. 4

Phylogenetic tree reconstruction of the echinoid control region and adjacent areas (CRA). The BI tree presented is based on 86 unique haplotypes retrieved from a total of 110 sequences, 405 bp long (see Table 1 for details on the sequences used for this tree). Supporting values (> 0.5 posterior probabilities and > 50% ML bootstrap values) are shown above the nodes

Fig. 5

Coverage (orange curve) and GC content (black curve; 200 bp sliding window, 10 bp step width) through the mitogenome of Hemicentrotus pulcherrimus (GenBank accession no. KC898202) illustrating moderate (R² = 0.335), but highly significant correlation (t-test, p < 10^− 100) between the two graphs. Note extreme drop of coverage towards the end of the CR (highlighted in grey), which coincides with a slight decrease in GC-content, but shows a much stronger negative excursion than other GC-poor areas in the mitogenome of this species (e.g. at nucleotide positions 4.4, 8.5, or 12.6 kb)