Novel mitochondrial genome rearrangements including duplications and extensive heteroplasmy could underlie temperature adaptations in Antarctic notothenioid fishes

Mitochondrial genomes are known for their compact size and conserved gene order, however, recent studies employing long-read sequencing technologies have revealed the presence of atypical mitogenomes in some species. In this study, we assembled and annotated the mitogenomes of five Antarctic notothenioids, including four icefishes (Champsocephalus gunnari, C. esox, Chaenocephalus aceratus, and Pseudochaenichthys georgianus) and the cold-specialized Trematomus borchgrevinki. Antarctic notothenioids are known to harbor some rearrangements in their mt genomes, however the extensive duplications in icefishes observed in our study have never been reported before. In the icefishes, we observed duplications of the protein coding gene ND6, two transfer RNAs, and the control region with different copy number variants present within the same individuals and with some ND6 duplications appearing to follow the canonical Duplication-Degeneration-Complementation (DDC) model in C. esox and C. gunnari. In addition, using long-read sequencing and k-mer analysis, we were able to detect extensive heteroplasmy in C. aceratus and C. esox. We also observed a large inversion in the mitogenome of T. borchgrevinki, along with the presence of tandem repeats in its control region. This study is the first in using long-read sequencing to assemble and identify structural variants and heteroplasmy in notothenioid mitogenomes and signifies the importance of long-reads in resolving complex mitochondrial architectures. Identification of such wide-ranging structural variants in the mitogenomes of these fishes could provide insight into the genetic basis of the atypical icefish mitochondrial physiology and more generally may provide insights about their potential role in cold adaptation.

Methods
The tissue sampling for C. gunnari, C. esox, and T. borchgrevinki were collected and sequenced using PacBio Sequel II instruments. The raw read data for C. gunnari (SRR20140461, SRR20140462), C. esox (SRR20115482, SRR20115483), and T. borchgrevinki is available under BioProjects PRJNA857989 and PRJNA907802 respectively. For C. aceratus and P. georgianus the available PacBio rawread data was downloaded from PRJNA420419 and PRJEB19273 respectively. The reads aligning to mt reference genomes were extracted using minimap2 and after read correction from Canu 1.8, the reads were assembled using Flye assembler. The resulting genome assemblies were annotated using Mitos2 and tRNAscan-SE 2.0 along with manual confirmation of annotations.