Evolution of sex-dependent mtDNA transmission in freshwater mussels (Bivalvia: Unionida)

Abstract

Doubly uniparental inheritance (DUI) describes a mode of mtDNA transmission widespread in gonochoric freshwater mussels (Bivalvia: Palaeoheterodonta: Unionida). In this system, both female- and male-transmitted mtDNAs, named F and M respectively, coexist in the same species. In unionids, DUI is strictly correlated to gonochorism and to the presence of the atypical open reading frames (ORFans) F-orf and M-orf, respectively inside F and M mtDNAs, which are hypothesized to participate in sex determination. However, DUI is not found in all three Unionida superfamilies (confirmed in Hyrioidea and Unionoidea but not in Etherioidea), raising the question of its origin in these bivalves. To reconstruct the co-evolution of DUI and of ORFans, we sequenced the mtDNAs of four unionids (two gonochoric with DUI, one gonochoric and one hermaphroditic without DUI) and of the related gonochoric species Neotrigonia margaritacea (Palaeoheterodonta: Trigoniida). Our analyses suggest that rearranged mtDNAs appeared early during unionid radiation, and that a duplicated and diverged atp8 gene evolved into the M-orf associated with the paternal transmission route in Hyrioidea and Unionoidea, but not in Etherioidea. We propose that novel mtDNA-encoded genes can deeply influence bivalve sex determining systems and the evolution of the mitogenomes in which they occur.

Figure 1

Maps of the seven mt genomes sequenced in this study. The corresponding GenBank accession numbers are given inside each genome map. Genomes are not in scale among each other, see main text for their length. Outer ring comprises all standard and putative coding sequences, identified with the following colour code: yellow, genes encoding electron transport chain and ATP-synthase subunits; dark blue, tRNA genes (see Supplementary Fig. S1 in Supplementary Information 1 for their names); pale green, rRNA genes; red, F-orf; bright blue, M-orf; bright green, additional ORFs cited in the text. Sequences are located on the outer or inner side of this circle according to their coding direction, respectively forward (clockwise) and reverse (anti-clockwise). In A. trapesialis mtDNA, the position of the two sequencing gaps is indicated on this ring. Middle ring comprises the repetitive regions, indicated in grey (not found in N. margaritacea and H. menziesii F): in A. trapesialis are specified the names R1–3 for the large tandem repeats found between nad4L and atp6; in H. menziesii M, two regions of high similarity found in different locations are marked with an asterisk (*), and a large palindrome sequence with “p”. Alignments of all repetitive regions can be found in Supplementary Information 2. Innermost ring displays the position of the unassigned regions as black segments.

Figure 2

Phylogeny of Palaeoheterodonta. Trees were constructed using (a) nucleotide sequence of 12 mtDNA-encoded genes and (b) the respective inferred protein sequences of species in Table 1. Only ML trees are shown: topology of BI trees was largely congruent; the few small differences are only described in the main text and below. Support values at a node are shown only if (1) they were not 100% bootstrap support on the ML tree and 1.0 for the posterior probability values on the BI tree, or (2) when a node was only present in the ML tree. Support values, when shown, are presented next to the node as ‘ML bootstrap value/BI posterior probability’. Species names in the middle column are coded with the following colours, according to taxonomy and/or mtDNA type: black, non-Palaeoheterodonta outgroups; brown, Trigoniida; red, Etherioidea; pink, F mtDNA of DUI species; blue, M mtDNA of a DUI species; violet, H mtDNA of a secondarily hermaphroditic unionid. Branches inside Palaeoheterodonta are coloured according to taxa: brown, Trigoniida; red, Etherioidea; bright green, Hyriidae (Hyrioidea); dark green, Margaritiferidae (Unionoidea); aqua blue, Unionidae (Unionoidea). (a): A dissimilarity in the ‘Unionoidea F’ clade between BI and ML trees resides in the different position of the branch comprising five species of the subfamily Gonideinae (i.e., H. cumingii, H. schlegelii, I. japanensis, S. carinatus, and S. oleivora; Table 1) relative to all other Unionidae F/H mtDNAs: in the ML tree here shown, this clade is sister to the cluster containing Ambleminae and Lampsilinae (see Table 1 for those species), while in the BI (not shown) it is sister to the clade containing the other ten F and H Unionidae mt genomes. Because of the positions in both ML and BI trees of L. tortuosa, A. woodiana and U. pictorum, subfamilies Gonideinae and Unioninae (Table 1) are not supported as being monophyletic. (b): Both ML and BI protein-based trees differ from the nucleotide-based one in the relative position of L. tortuosa, which branches differently inside the same cluster (compare the two trees in figure). Again, Gonideinae and Unioninae are not monophyletic (Table 1).

Figure 3

Structural characterization of proteins encoded by lineage-specific genes and ORFs. Visual summary of the characterization made with Quick2D on MCOX2, F-ORF, M-ORF, and the putative proteins encoded by the ORFs CmonM_UR_24_11 and Atra_UR_21_9 from Hyridella menziesii, Cumberlandia monodonta, and Anodontites trapesialis mtDNAs. Complete Quick2D outputs are displayed in Supplementary Information 5. In each panel, on the X axis are the amino acid positions of the protein, while on the Y is the support for each feature (described below) shown in terms of how many methods indicated a certain characteristic at a given position. Colour code of features: full green areas, transmembrane regions; striped red areas, regions forming helices; striped blue areas, regions forming beta-sheets; purple lines, signal peptide signature; orange lines, disordered regions. The two grey squares on the background of panels (a) and (b) represent, from left to right, the boundaries of the conserved transmembrane and periplasmic domains of COX2.

Figure 4

Model for the evolution of mt genomes and DUI in Palaeoheterodonta. The backbone tree represents the phylogeny obtainable by removing the M mtDNAs from the trees in Fig. 2. DUI presence/absence is specified for all major clades. At the tip of each final branch, the names of the species for which we obtained the mtDNA sequences in this study are indicated together with the respective family (no species are enlisted for Unionidae); superfamily affiliations are specified with bars on the right side of the figure. Circles beside a final branch represent a schematic mt genome structure with the following colour code: black, mtDNA of a non-DUI species; red, F mtDNA in a DUI species/family; blue, M mtDNA in a DUI species/family. Purple circles in parentheses represent the typical H mtDNA of secondarily hermaphroditic species in a given family. For N. margaritacea mt genome, the interrogation point highlights the uncertain presence of DUI in this species, while the red colour the fact that the genome we obtained, if DUI is present, would be its F. Bars on the mtDNAs represent atp8-derived ORFs (green), M-orfs (blue), F-orfs (red), or H-orfs (purple), while blue triangles represent the elongated cox2 genes in M mtDNAs. The position of these latter features on the schematic mtDNAs reflects their actual location in the mt genomes (compare with Fig. 1). Bars inside the Unionida branches of the tree indicate evolutionary events and/or character states of lifestyle-related features as enlisted by ref. (grey bars) or evolutionary events of mt genomes (black bars), described inside the figure.