The case of biparental mitochondrial inheritance
The pea (Pisum sativum L.) is an important crop and traditional genetic object. It is frequently considered an obligate selfer (Blixt 1971), nevertheless cross-pollination was shown to occur in cultivated peas grown outdoors (reviewed in Blixt 1971; Loenning 1984; our unpublished data) and heterozygocity was shown to occur in natural populations of wild peas (Smýkal et al. 2018). Hybridisation was supposed to play an important role in microevolution of wild representatives of P. sativum in the past (Smýkal et al. 2017; Trneny et al. 2018; Bogdanova et al. 2021; Rispail et al. 2023; Shatskaya et al. 2023). Maternal organellar inheritance has been assumed for peas like for most angiosperms (Birky 2001; Camus et al. 2022). However, in pea, pollen tube produced by male gametophyte was shown to contain variable amount of plastid DNA (Corriveau et al. 1988; 1989), and few proplastids were registered there cytologically (Polans et al. 1990). Microscopic studies showed that pollen tubes also contained mitochondria, but specific staining did not reveal mitochondrial DNA (Corriveau and Coleman 1991). Earlier we discovered cases of biparental inheritance of plastids in crosses between cultivated and certain wild peas (Pisum sativum subsp. elatius s.l.) (Bogdanova & Kosterin 2006; Bogdanova et al. 2012; 2014; Kosterin & Bogdanova 2015) and between some wild peas (Kosterin & Bogdanova 2021). Some of these crosses were associated with nuclear-plastid conflict which could be alleviated by non-canonically inherited paternal plastids (Bogdanova 2007; Bogdanova et al. 2014; Kosterin & Bogdanova 2015 and unpublished), the situation favouring paternal plastid transmission (Chiu and Sears 1993). Similar paternal shift in the composition of mitochondrial genomes was observed in populations of interspecies hybrids between rye and wheat (Siniavskaya et al. 2004) or barley and wheat (Aksyonova et al. 2005). Nevertheless, in pea, we up to now failed to reveal paternal or biparental inheritance of mitochondria (Bogdanova 2007).
In the present work we performed crosses between the pea lines WL_1238 and JI_1794 and studied F2 populations from reciprocal crosses. Based on our previous experience, we did not expect to meet cases of biparental inheritance of mitochondria. However, the here presented finding demonstrates that paternal inheritance of mitochondria is possible in peas but is a rarer phenomenon than that of plastids, 0.8 vs. 1.2%, respectively, in the F2 of the cross WL_1238 x JI_1794. Mitochondria are supposed to be inherited independently of plastids (Nagata et al. 1999; Nagata, 2010). In this study, both observed plants with paternal mtDNA also contained paternal plastid DNA (Fig. 1). In one plant from this study, as well as in all earlier observed cases, paternal plastid inheritance was not accompanied by non-canonical mtDNA inheritance.
Heteroplasmy for mitochondrial DNA in wild pea accessions
Heteroplasmy, that is, presence of mitochondrial DNA of different kinds in the same cells, may occur via different pathways. In natural populations of cross-pollinating plants it appears to be rather common and persists via paternal leakage and transfer by hetroplasmic mothers to offspring (Pearl et al. 2009; Levsen et al. 2016; Mandel et al. 2020). Hybridisation was implied to play an important role in wild pea microevolution (Smýkal et al. 2017; Trneny et al. 2018; Bogdanova et al. 2021; Rispail et al. 2023; Shatskaya et al. 2023) and here we demonstrate possibility of paternal leakage of mitochondria that would result in heteroplasmy. Alternatively, heteroplasmy could have occurred due to relatively recent mutations which are still retained despite rapid vegetative sorting of organelles (Birky 2001; Broz et al. 2022).
In P. fulvum accession VIR_6071, nine of 10 plants had the three analyzed mtDNA markers apparently homoplasmic and typical of all accessions of this species. Yet, one of these plants was heteroplasmic for all the three markers. This implies that heteroplasmy affected the entire mitochondrial genome rather than separate loci. It is remarkable that the same plant was heteroplasmic for the tested plastid DNA marker. Thus, both foreign mitochondria and plastids were acquired by a progenitor of the VIR_6071 sample which we have at our disposal. This excludes the version of de novo mutations, and points to the genetic exchange of organelles between different pea lineages. The observed homoplasmy of other plants could be true or apparent. In the first case we probably observe the ongoing process of organelle sorting. In the second case the minor mtDNA fraction escapes being detected. We may suppose that some additional mtDNA fraction persists not only in accession VIR_6071 but probably also in other accessoins in which we observed admixtures of ‘alien’ mtDNA reads. Occasionally such fraction may be detected, due to either its increase for some reason to substantial amount or just because of its uneven distribution along a plant and random involvement into analysis.
Since mitochondrial heteroplasmy is sorted out quite rapidly (Broz et al. 2022), the presumed hybridisation should have taken place not so many generations ago and we would see its signs in the plant phenotype, determined by the nuclear genome, which would retain some features of hybrids between P. fulvum and P. sativum, since the admixed mtDNA had the allele characteristic for the latter. Nevertheless, accession VIR_6071 demonstrated the pure Pisum fulvum phenotype. During recent decades a notion appeared that genetic exchange of organelles may occur via so-called horizontal gene transfer (Bergthorsson et al. 2003; Filip and Skuza 2021) associated with plastid (Stegemann et al. 2012) or mitochondria (Gurdon et al. 2016) capture from foreign species or even kingdom (Filip and Skuza 2021). We can imagine that such an event affected VIR6071 accession of P. fulvum, that would explain phenotypic uniformity of the sample we analyze.
In the case of PIS_2845 all of the tested plants manifested heteroplasmy for the mitochondrial DNA marker with no apparent signs for plastid heteroplasmy. This might imply that co-existing variants of mitochondrial genomes are maintained by natural selection, either due to "selfish" behaviour (Havird et al. 2019), or contribution to plant's fitness (Allen 2005; Flood et al. 2020), possibly via ensuring citonuclear compatibility (Sambatti et al. 2008). At the same time, foreign plastids could have been sorted out. Alternatively, we face a nuclear sequence transferred from mitochondria.
A recombinant mitochondrial genome
In plants, mitochondrial DNA recombination facilitated by abundant repeated regions is rather common process which involves DNA repair mechanism (Davila et al. 2011; Gualberto and Newton 2017; Chevigny et al. 2020; Wu et al. 2020a; Camus et al. 2022), and is manifested in existence of a series of structural isoforms (Palmer and Hebron 1988; Davila et al. 2011; Sloan et al. 2012; Gualberto et al. 2014; Smith et al. 2015; Wu and Sloan 2019; Kozik et al. 2019; Wu et al. 2020a; Xia et al. 2020; Camus et al. 2022). These processes occur routinely within or between mtDNA molecules of the same organism due to recurrent fusion/fission events (Logan 2003, 2006; Arimura et al. 2004; Sheahan et al. 2005). However, paternal leakage can result in physical proximity of DNA from mitochondria of different origin. In this situation, recombination between foreign mitogenomes is expected to occur and create novel genotypes (McCauley and Ellis 2008; Levsen et al. 2016), provided that nuclear genotype does not prevent mtDNA recombination (Wu et al. 2020). The finding in accession W6_2107 of a ~ 13 kb insertion of mtDNA originating from another, more ancient evolutionary lineage M1 in terms of Bogdanova et al. (2021) can be interpreted only in terms of a recombination event between mitochondrial genomes representing lineages M1 and M2, which had occurred in an ancestor of accession W6_2107 after its divergence from other accessions studied. This event should have been preceded by hybridisation between carriers of M1 and M2 mitochondria and biparental mitochondrial inheritance. It is impossible to judge which of the genomes, M1 or M2, came via paternal inheritance. The accession W6_2107 earlier has been supposed to be of hybrid origin, since it harbours unusual combination of nuclear, plastid and mitochondrial markers (Shatskaya et al. 2023).
In the above cited paper we reconstructed another case of recombination between mitochondrial genomes of the same lineages M1 and M2 during pea evolution, namely an acquisition by the ancestor of the present-day M2 of ca 10 kb of the mitogenome mitogenome from a representative of lineage M1. Curiously, this region appeared to be deleted in a related (sub)species Pisum abyssinicum A. Br., the mitogemones of which belong to the lineage M2 (Shatskaya et al. 2023). So, recombination between diverged mitochondrial genomes took place in evolution of the genus Pisum L. more than once.
The three findings reported in this work, occasional non-canonical inheritance of mitochondria, naturally occurring heteroplasmy and recombination of diverged mitochondrial genomes in pea elucidate microevolutionary processes which shaped diversity in this important crop and the oldest genetic object.