The study, published in BMC Plant Biology, assembled the complete mitochondrial genomes of four species in the genus: Rehmannia glutinosa, the source of the famous herbal medicine dihuang, along with R. henryi, R. piasezkii, and R. chingii. All four belong to the broomrape family, Orobanchaceae, a group better known for its parasitic members but which also contains these therapeutic and horticulturally valuable plants. Despite decades of research into the bioactive compounds and metabolic pathways that make Rehmannia medicinally important, its mitochondrial genomes had remained essentially unexplored territory until now.

What the team found is a striking demonstration of structural plasticity. The mitochondrial genome of R. glutinosa spans 547,002 base pairs and exists as a typical single circular molecule, as does that of R. piasezkii, which is considerably larger at 742,657 base pairs. But the other two species broke the mold entirely. R. chingii, at 662,985 base pairs, is organized as one linear molecule accompanied by two circular structures, while R. henryi, at 663,966 base pairs, is split across four separate circular molecules. In other words, four species from the same genus carry mitochondrial genomes that differ not just in size but in their fundamental molecular architecture.

This kind of multipartite organization is not merely a curiosity. Plant mitochondrial genomes are known to exist as dynamic populations of molecules whose shapes depend on recombination between repeated sequences scattered throughout the DNA. When repeat-mediated recombination occurs at certain sites, a single master circle can be resolved into multiple smaller circles or linear fragments, and different configurations can coexist within the same cell. The assembly approach used in this study relied on long-read PacBio HiFi sequencing, which allowed the researchers to resolve the complex assembly graphs and confirm which alternative connections between contigs were genuinely supported by the read data, a critical step when the genome refuses to conform to a single circular model.

Beneath the architectural drama, the researchers documented a detailed landscape of repetitive DNA. Simple sequence repeats, short tandem motifs that are useful as molecular markers, showed a broadly conserved composition across the four species. However, the numbers of dispersed repeats, longer duplicated sequences that can mediate recombination, and tandem repeats varied quantitatively among the species. This variation in repeat content is exactly what would be expected to drive the differences in genome structure, since each pair of sufficiently large repeats represents a potential recombination point that can reshape the molecule into alternative conformations. The findings thus connect the fine-scale repeat inventory directly to the large-scale structural diversity observed among the four genomes.

Perhaps the most surprising result is how little the functional content of these genomes varies despite their structural chaos. The team predicted 394 C-to-U RNA editing sites across 31 protein-coding genes in each of the four species, a profile that was highly conserved throughout the genus. RNA editing of this kind is a hallmark of plant mitochondria, where specific cytidines in the transcribed RNA are converted to uridines, often restoring codons that would otherwise be nonfunctional. The consistency of this editing landscape suggests that while the physical packaging of the genome is fluid, the information it encodes and the post-transcriptional processing it requires are under strong stabilizing selection.

The study also traced the ongoing conversation between two organelles. Mitochondrial genomes in plants frequently absorb DNA from the chloroplast, and these mitochondrial plastid DNA sequences, or MTPTs, were abundant in all four Rehmannia species, with 21 to 28 such transfers identified in each genome. Among them was a conserved region of approximately 4.5 kilobases containing complete sequences of ndhB, a photosynthesis-related gene, along with two plastid ribosomal protein genes, rps7 and rps12. These fragments are essentially molecular fossils of past DNA traffic from chloroplast to mitochondrion, and their conservation across the genus points to a shared history of inter-organelle gene transfer predating the divergence of these species.

To place these genomes in an evolutionary context, the researchers conducted phylogenetic analysis that strongly supported the monophyly of Rehmannia, confirming that the genus forms a single coherent lineage within Orobanchaceae. They then measured nucleotide diversity and calculated the ratio of nonsynonymous to synonymous substitution rates, known as Ka/Ks, across the shared protein-coding genes. The overall picture was one of conservation: most mitochondrial genes are evolving slowly and under purifying selection, as expected for genes encoding essential components of the respiratory chain. Yet a handful of genes stood out, with sdh3, atp6, and matR showing relatively higher sequence variation among the species, hinting at localized hotspots of divergence within an otherwise conservative genome.

The contrast between these two layers of evolution, structural and sequence-based, is what makes the study compelling. On one hand, the mitochondrial genomes of Rehmannia have diversified dramatically in size and organization over what is presumably a modest evolutionary timescale, with one species carrying its genome in four circles and another in a hybrid of linear and circular molecules. On the other hand, the gene content, RNA editing patterns, and rates of sequence evolution remain remarkably stable. This coexistence of pronounced structural plasticity with conserved genomic features appears to be a defining characteristic of plant mitochondrial genomes, and the Rehmannia data provide a particularly clean example because the four species are close relatives, minimizing confounding effects from deep evolutionary divergence.

For the broader research community, the work delivers a set of complete, well-annotated mitochondrial genomes for a medicinally significant genus, along with a framework for understanding how repeat-mediated recombination, chloroplast DNA influx, and selective constraint interact to shape mitochondrial genome evolution in flowering plants. It also underscores a practical lesson for genome assembly: assuming a single circular mitochondrial chromosome can lead researchers astray, and long-read sequencing with careful validation of alternative assembly paths is increasingly essential. As more species across Orobanchaceae and beyond receive the same treatment, comparisons like this one will help reveal whether the structural exuberance seen in Rehmannia is the rule or the exception in plant mitochondrial evolution, and whether the genes that vary most, such as sdh3 and atp6, hold clues to adaptation in these remarkable medicinal plants.

Subject of Research: Comparative assembly and evolutionary analysis of mitochondrial genomes in four Rehmannia species

Article Title: Assembly and comparative analysis of the four complete mitochondrial genomes in Rehmannia (Orobanchaceae): structural diversity and evolutionary insights

Article References: Jiang, X., Yan, C., Qiu, A., Hou, R., Wu, Q., Huang, S., Wang, J., Ren, M., Wang, S., Li, J., & Liu, Z.-L. (2026). Assembly and comparative analysis of the four complete mitochondrial genomes in Rehmannia (Orobanchaceae): structural diversity and evolutionary insights. BMC Plant Biology. https://doi.org/10.1186/s12870-026-10054-2

Image Credits: AI Generated

DOI: 10.1186/s12870-026-10054-2

Keywords: Rehmannia, mitochondrial genome, Orobanchaceae, genome assembly, RNA editing, MTPTs, chloroplast gene transfer, repeat-mediated recombination, plant evolution, PacBio HiFi, Ka/Ks selection, medicinal plants