Genome copy number predicts extreme evolutionary rate variation in plant mitochondrial DNA.

Nuclear and organellar genomes can evolve at vastly different rates despite occupying the same cell. In most bilaterian animals, mitochondrial DNA (mtDNA) evolves faster than nuclear DNA, whereas this trend is generally reversed in plants. However, in some exceptional angiosperm clades, mtDNA substitution rates have increased up to 5,000-fold compared with closely related lineages. The mechanisms responsible for this acceleration are generally unknown. Because plants rely on homologous recombination to repair mtDNA damage, we hypothesized that mtDNA copy numbers may predict evolutionary rates, as lower copy numbers may provide fewer templates for such repair mechanisms. In support of this hypothesis, we found that copy number explains 47% of the variation in synonymous substitution rates of mtDNA across 60 diverse seed plant species representing ~300 million years of evolution. Copy number was also negatively correlated with mitogenome size, which may be a cause or consequence of mutation rate variation. Both relationships were unique to mtDNA and not observed in plastid DNA. These results suggest that homologous recombinational repair plays a role in driving mtDNA substitution rates in plants and may explain variation in mtDNA evolution more broadly across eukaryotes. Our findings also contribute to broader questions about the relationships between mutation rates, genome size, selection efficiency, and the drift-barrier hypothesis.

Plastid Genome Assembly Using Long-read data.

Although plastid genome (plastome) structure is highly conserved across most seed plants, investigations during the past two decades have revealed several disparately related lineages that experienced substantial rearrangements. Most plastomes contain a large inverted repeat and two single-copy regions, and a few dispersed repeats; however, the plastomes of some taxa harbour long repeat sequences (>300 bp). These long repeats make it challenging to assemble complete plastomes using short-read data, leading to misassemblies and consensus sequences with spurious rearrangements. Single-molecule, long-read sequencing has the potential to overcome these challenges, yet there is no consensus on the most effective method for accurately assembling plastomes using long-read data. We generated a pipeline, plastid Genome Assembly Using Long-read data (ptGAUL), to address the problem of plastome assembly using long-read data from Oxford Nanopore Technologies (ONT) or Pacific Biosciences platforms. We demonstrated the efficacy of the ptGAUL pipeline using 16 published long-read data sets. We showed that ptGAUL quickly produces accurate and unbiased assemblies using only ~50× coverage of plastome data. Additionally, we deployed ptGAUL to assemble four new Juncus (Juncaceae) plastomes using ONT long reads. Our results revealed many long repeats and rearrangements in Juncus plastomes compared with basal lineages of Poales. The ptGAUL pipeline is available on GitHub: https://github.com/Bean061/ptgaul.

Rate accelerations in plastid and mitochondrial genomes of Cyperaceae occur in the same clades.

Cyperaceae, the second largest family in the monocot order Poales, comprises >5500 species and includes the genus Eleocharis with âˆ¼ 250 species. A previous study of complete plastomes of two Eleocharis species documented extensive structural heteroplasmy, gene order changes, high frequency of dispersed repeats along with gene losses and duplications. To better understand the phylogenetic distribution of gene and intron content as well as rates and patterns of sequence evolution within and between mitochondrial and plastid genomes of Eleocharis and Cyperaceae, an additional 29 Eleocharis organelle genomes were sequenced and analyzed. Eleocharis experienced extensive gene loss in both genomes while loss of introns was mitochondria-specific. Eleocharis has higher rates of synonymous (dS) and nonsynonymous (dN) substitutions in the plastid and mitochondrion than most sampled angiosperms, and the pattern was distinct from other eudicot lineages with accelerated rates. Several clades showed higher dS and dN in mitochondrial genes than in plastid genes. Furthermore, nucleotide substitution rates of mitochondrial genes were significantly accelerated on the branch leading to Cyperaceae compared to most angiosperms. Mitochondrial genes of Cyperaceae exhibited dramatic loss of RNA editing sites and a negative correlation between RNA editing and dS values was detected among angiosperms. Mutagenic retroprocessing and dysfunction of DNA replication, repair and recombination genes are the most likely cause of striking rate accelerations and loss of edit sites and introns in Eleocharis and Cyperaceae organelle genomes.

Plastid phylogenomics uncovers multiple species in Medicago truncatula (Fabaceae) germplasm accessions.

Medicago truncatula is a model legume that has been extensively investigated in diverse subdisciplines of plant science. Medicago littoralis can interbreed with M. truncatula and M. italica; these three closely related species form a clade, i.e. TLI clade. Genetic studies have indicated that M. truncatula accessions are heterogeneous but their taxonomic identities have not been verified. To elucidate the phylogenetic position of diverse M. truncatula accessions within the genus, we assembled 54 plastid genomes (plastomes) using publicly available next-generation sequencing data and conducted phylogenetic analyses using maximum likelihood. Five accessions showed high levels of plastid DNA polymorphism. Three of these highly polymorphic accessions contained sequences from both M. truncatula and M. littoralis. Phylogenetic analyses of sequences placed some accessions closer to distantly related species suggesting misidentification of source material. Most accessions were placed within the TLI clade and maximally supported the interrelationships of three subclades. Two Medicago accessions were placed within a M. italica subclade of the TLI clade. Plastomes with a 45-kb (rpl20-ycf1) inversion were placed within the M. littoralis subclade. Our results suggest that the M. truncatula accession genome pool represents more than one species due to possible mistaken identities and gene flow among closely related species.

Effects of Salt Stress on Transcriptional and Physiological Responses in Barley Leaves with Contrasting Salt Tolerance.

Salt stress negatively impacts crop production worldwide. Genetic diversity among barley (Hordeum vulgare) landraces adapted to adverse conditions should provide a valuable reservoir of tolerance genes for breeding programs. To identify molecular and biochemical differences between barley genotypes, transcriptomic and antioxidant enzyme profiles along with several morpho-physiological features were compared between salt-tolerant (Boulifa) and salt-sensitive (Testour) genotypes subjected to salt stress. Decreases in biomass, photosynthetic parameters, and relative water content were low in Boulifa compared to Testour. Boulifa had better antioxidant protection against salt stress than Testour, with greater antioxidant enzymes activities including catalase, superoxide dismutase, and guaiacol peroxidase. Transcriptome assembly for both genotypes revealed greater accumulation of differentially expressed transcripts in Testour compared to Boulifa, emphasizing the elevated transcriptional response in Testour following salt exposure. Various salt-responsive genes, including the antioxidant catalase 3, the osmoprotectant betaine aldehyde dehydrogenase 2, and the transcription factors MYB20 and MYB41, were induced only in Boulifa. By contrast, several genes associated with photosystems I and II, and light receptor chlorophylls A and B, were more repressed in Testour. Co-expression network analysis identified specific gene modules correlating with differences in genotypes and morpho-physiological traits. Overall, salinity-induced differential transcript accumulation underlies the differential morpho-physiological response in both genotypes and could be important for breeding salt tolerance in barley.

Highly Resolved Papilionoid Legume Phylogeny Based on Plastid Phylogenomics.

Comprising 501 genera and around 14,000 species, Papilionoideae is not only the largest subfamily of Fabaceae (Leguminosae; legumes), but also one of the most extraordinarily diverse clades among angiosperms. Papilionoids are a major source of food and forage, are ecologically successful in all major biomes, and display dramatic variation in both floral architecture and plastid genome (plastome) structure. Plastid DNA-based phylogenetic analyses have greatly improved our understanding of relationships among the major groups of Papilionoideae, yet the backbone of the subfamily phylogeny remains unresolved. In this study, we sequenced and assembled 39 new plastomes that are covering key genera representing the morphological diversity in the subfamily. From 244 total taxa, we produced eight datasets for maximum likelihood (ML) analyses based on entire plastomes and/or concatenated sequences of 77 protein-coding sequences (CDS) and two datasets for multispecies coalescent (MSC) analyses based on individual gene trees. We additionally produced a combined nucleotide dataset comprising CDS plus matK gene sequences only, in which most papilionoid genera were sampled. A ML tree based on the entire plastome maximally supported all of the deep and most recent divergences of papilionoids (223 out of 236 nodes). The Swartzieae, ADA (Angylocalyceae, Dipterygeae, and Amburaneae), Cladrastis, Andira, and Exostyleae clades formed a grade to the remainder of the Papilionoideae, concordant with nine ML and two MSC trees. Phylogenetic relationships among the remaining five papilionoid lineages (Vataireoid, Dermatophyllum, Genistoid s.l., Dalbergioid s.l., and Baphieae + Non-Protein Amino Acid Accumulating or NPAAA clade) remained uncertain, because of insufficient support and/or conflicting relationships among trees. Our study fully resolved most of the deep nodes of Papilionoideae, however, some relationships require further exploration. More genome-scale data and rigorous analyses are needed to disentangle phylogenetic relationships among the five remaining lineages.

Born in the mitochondrion and raised in the nucleus: evolution of a novel tandem repeat family in Medicago polymorpha (Fabaceae).

Plant nuclear genomes harbor sequence elements derived from the organelles (mitochondrion and plastid) through intracellular gene transfer (IGT). Nuclear genomes also show a dramatic range of repeat content, suggesting that any sequence can be readily amplified. These two aspects of plant nuclear genomes are well recognized but have rarely been linked. Through investigation of 31 Medicago taxa we detected exceptionally high post-IGT amplification of mitochondrial (mt) DNA sequences containing rps10 in the nuclear genome of Medicago polymorpha and closely related species. The amplified sequences were characterized as tandem arrays of five distinct repeat motifs (2157, 1064, 987, 971, and 587 bp) that have diverged from the mt genome (mitogenome) in the M. polymorpha nuclear genome. The mt rps10-like arrays were identified in seven loci (six intergenic and one telomeric) of the nuclear chromosome assemblies and were the most abundant tandem repeat family, representing 1.6-3.0% of total genomic DNA, a value approximately three-fold greater than the entire mitogenome in M. polymorpha. Compared to a typical mt gene, the mt rps10-like sequence coverage level was 691.5-7198-fold higher in M. polymorpha and closely related species. In addition to the post-IGT amplification, our analysis identified the canonical telomeric repeat and the species-specific satellite arrays that are likely attributable to an ancestral chromosomal fusion in M. polymorpha. A possible relationship between chromosomal instability and the mt rps10-like tandem repeat family in the M. polymorpha clade is discussed.

Transcriptomic Analysis of Salt-Stress-Responsive Genes in Barley Roots and Leaves.

Barley is characterized by a rich genetic diversity, making it an important model for studies of salinity response with great potential for crop improvement. Moreover, salt stress severely affects barley growth and development, leading to substantial yield loss. Leaf and root transcriptomes of a salt-tolerant Tunisian landrace (Boulifa) exposed to 2, 8, and 24 h salt stress were compared with pre-exposure plants to identify candidate genes and pathways underlying barley's response. Expression of 3585 genes was upregulated and 5586 downregulated in leaves, while expression of 13,200 genes was upregulated and 10,575 downregulated in roots. Regulation of gene expression was severely impacted in roots, highlighting the complexity of salt stress response mechanisms in this tissue. Functional analyses in both tissues indicated that response to salt stress is mainly achieved through sensing and signaling pathways, strong transcriptional reprograming, hormone osmolyte and ion homeostasis stabilization, increased reactive oxygen scavenging, and activation of transport and photosynthesis systems. A number of candidate genes involved in hormone and kinase signaling pathways, as well as several transcription factor families and transporters, were identified. This study provides valuable information on early salt-stress-responsive genes in roots and leaves of barley and identifies several important players in salt tolerance.

In and out: Evolution of viral sequences in the mitochondrial genomes of legumes (Fabaceae).

Plant specific mitoviruses in the 'genus' Mitovirus (Narnaviridae) and their integrated sequences (non-retroviral endogenous RNA viral elements or NERVEs) have been recently identified in various plant lineages. However, the sparse phylogenetic coverage of complete plant mitochondrial genome (mitogenome) sequences and the non-conserved nature of mitochondrial intergenic regions have hindered comparative studies on mitovirus NERVEs in plants. In this study, 10 new mitogenomes were sequenced from legumes (Fabaceae). Based on comparative genomic analysis of 27 total mitogenomes, we identified mitovirus NERVEs and transposable elements across the family. All legume mitogenomes included NERVEs and total NERVE length varied from ca. 2 kb in the papilionoid Trifolium to 35 kb in the mimosoid Acacia. Most of the NERVE integration sites were in highly variable intergenic regions, however, some were positioned in six cis-spliced mitochondrial introns. In the Acacia mitogenome, there were L1-like transposon sequences including an almost full-length copy with target site duplications (TSDs). The integration sites of NERVEs in four introns showed evidence of L1-like retrotransposition events. Phylogenetic analysis revealed that there were multiple instances of precise deletion of NERVEs between TSDs. This study provides clear evidence that a L1-like retrotransposition mechanism has a long history of contributing to the integration of viral RNA into plant mitogenomes while microhomology-mediated deletion can restore the integration site.

The chicken or the egg? Plastome evolution and an independent loss of the inverted repeat in papilionoid legumes.

The plastid genome (plastome), while surprisingly constant in gene order and content across most photosynthetic angiosperms, exhibits variability in several unrelated lineages. During the diversification history of the legume family Fabaceae, plastomes have undergone many rearrangements, including inversions, expansion, contraction and loss of the typical inverted repeat (IR), gene loss and repeat accumulation in both shared and independent events. While legume plastomes have been the subject of study for some time, most work has focused on agricultural species in the IR-lacking clade (IRLC) and the plant model Medicago truncatula. The subfamily Papilionoideae, which contains virtually all of the agricultural legume species, also comprises most of the plastome variation detected thus far in the family. In this study three non-papilioniods were included among 34 newly sequenced legume plastomes, along with 33 publicly available sequences, to assess plastome structural evolution in the subfamily. In an effort to examine plastome variation across the subfamily, approximately 20% of the sampling represents the IRLC with the remainder selected to represent the early-branching papilionoid clades. A number of IR-related and repeat-mediated changes were identified and examined in a phylogenetic context. Recombination between direct repeats associated with ycf2 resulted in intraindividual plastome heteroplasmy. Although loss of the IR has not been reported in legumes outside of the IRLC, one genistoid taxon was found to completely lack the typical plastome IR. The role of the IR and non-IR repeats in the progression of plastome change is discussed.

Plastid Genomes of Flowering Plants: Essential Principles.

The plastid genome (plastome ) has proved a valuable source of data for evaluating evolutionary relationships among angiosperms. Through basic and applied approaches, plastid transformation technology offers the potential to understand and improve plant productivity, providing food, fiber, energy, and medicines to meet the needs of a burgeoning global population. The growing genomic resources available to both phylogenetic and biotechnological investigations is allowing novel insights and expanding the scope of plastome research to encompass new species. In this chapter, we present an overview of some of the seminal and contemporary research that has contributed to our current understanding of plastome evolution and attempt to highlight the relationship between evolutionary mechanisms and the tools of plastid genetic engineering.

Biolistic Plastid Transformation in Lettuce (Lactuca sativa) for Oral Delivery of Biopharmaceuticals.

The interest in producing pharmaceutical proteins in a nontoxic plant host has led to the development of an approach to express such proteins in transplastomic lettuce (Lactuca sativa). A number of therapeutic proteins and vaccine antigen candidates have been stably integrated into the lettuce plastid genome using biolistic DNA delivery. High levels of accumulation and retention of biological activity suggest that lettuce may provide and ideal platform for the production of biopharmaceuticals.

Clade-Specific Plastid Inheritance Patterns Including Frequent Biparental Inheritance in Passiflora Interspecific Crosses.

Plastid inheritance in angiosperms is presumed to be largely maternal, with the potential to inherit plastids biparentally estimated for about 20% of species. In Passiflora, maternal, paternal and biparental inheritance has been reported; however, these studies were limited in the number of crosses and progeny examined. To improve the understanding of plastid transmission in Passiflora, the progeny of 45 interspecific crosses were analyzed in the three subgenera: Passiflora, Decaloba and Astrophea. Plastid types were assessed following restriction digestion of PCR amplified plastid DNA in hybrid embryos, cotyledons and leaves at different developmental stages. Clade-specific patterns of inheritance were detected such that hybrid progeny from subgenera Passiflora and Astrophea predominantly inherited paternal plastids with occasional incidences of maternal inheritance, whereas subgenus Decaloba showed predominantly maternal and biparental inheritance. Biparental plastid inheritance was also detected in some hybrids from subgenus Passiflora. Heteroplasmy due to biparental inheritance was restricted to hybrid cotyledons and first leaves with a single parental plastid type detectable in mature plants. This indicates that in Passiflora, plastid retention at later stages of plant development may not reflect the plastid inheritance patterns in embryos. Passiflora exhibits diverse patterns of plastid inheritance, providing an excellent system to investigate underlying mechanisms in angiosperms.

Extensive variation in nucleotide substitution rate and gene/intron loss in mitochondrial genomes of Pelargonium.

Geraniaceae organelle genomes have been shown to exhibit several highly unusual features compared to most other photosynthetic angiosperms. This includes massively rearranged plastomes with considerable size variation, extensive gene and intron loss, accelerated rates of nucleotide substitutions in both mitogenomes and plastomes, and biparental inheritance and cytonuclear incompatibility of the plastome. Most previous studies have focused on plastome evolution with mitogenome comparisons limited to only a few taxa or genes. In this study, mitogenomes and transcriptomes were examined for 27 species of Geraniales, including 13 species of Pelargonium. Extensive gene and intron losses were detected across the Geraniales with Pelargonium representing the most gene depauperate lineage in the family. Plotting these events on the Geraniaceae phylogenetic tree showed that gene losses occurred multiple times, whereas intron losses more closely reflected the relationships among taxa. In addition, P. australe acquired an intron by horizontal transfer. Comparisons of nucleotide substitution rates in Pelargonium showed that synonymous changes in nuclear genes were much lower than in mitochondrial genes. This is in contrast to the previously published studies that indicated that nuclear genes have 16 fold higher rates than mitochondrial genes across angiosperms. Elevated synonymous substitutions occurred for each mitochondrial gene in Pelargonium with the highest values 783 and 324 times higher than outgroups and other Geraniaceae, respectively. Pelargonium is one of four unrelated genera of angiosperms (Ajuga, Plantago and Silene) that have experienced highly accelerated nucleotide substitutions in mitogenomes. It is distinct from most angiosperms in also having elevated substitution rates in plastid genes but the cause of rate accelerations in Pelargonium plastomes and mitogenomes may be different.

Nucleotide substitution rates of diatom plastid encoded protein genes are positively correlated with genome architecture.

Diatoms are the largest group of heterokont algae with more than 100,000 species. As one of the single-celled photosynthetic organisms that inhabit marine, aquatic and terrestrial ecosystems, diatoms contribute ~ 45% of global primary production. Despite their ubiquity and environmental significance, very few diatom plastid genomes (plastomes) have been sequenced and studied. This study explored patterns of nucleotide substitution rates of diatom plastids across the entire suite of plastome protein-coding genes for 40 taxa representing the major clades. The highest substitution rate was lineage-specific within the araphid 2 taxon Astrosyne radiata and radial 2 taxon Proboscia sp. Rate heterogeneity was also evident in different functional classes and individual genes. Similar to land plants, proteins genes involved in photosynthetic metabolism have lower synonymous and nonsynonymous substitutions rates than those involved in transcription and translation. Significant positive correlations were identified between substitution rates and measures of genomic rearrangements, including indels and inversions, which is a similar result to what was found in legume plants. This work advances the understanding of the molecular evolution of diatom plastomes and provides a foundation for future studies.

Under the rug: Abandoning persistent misconceptions that obfuscate organelle evolution.

The advent and advance of next generation sequencing over the past two decades made it possible to accumulate large quantities of sequence reads that could be used to assemble complete or nearly complete organelle genomes (plastome or mitogenome). The result has been an explosive increase in the availability of organelle genome sequences with over 4000 different species of green plants currently available on GenBank. During the same time period, plant molecular biologists greatly enhanced the understanding of the structure, repair, replication, recombination, transcription and translation, and inheritance of organelle DNA. Unfortunately many plant evolutionary biologists are unaware of or have overlooked this knowledge, resulting in misrepresentation of several phenomena that are critical for phylogenetic and evolutionary studies using organelle genomes. We believe that confronting these misconceptions about organelle genome organization, composition, and inheritance will improve our understanding of the evolutionary processes that underly organelle evolution. Here we discuss four misconceptions that can limit evolutionary biology studies and lead to inaccurate phylogenies and incorrect structure of the organellar DNA used to infer organelle evolution.

Rampant Nuclear Transfer and Substitutions of Plastid Genes in Passiflora.

Gene losses in plastid genomes (plastomes) are often accompanied by functional transfer to the nucleus or substitution of an alternative nuclear-encoded gene. Despite the highly conserved gene content in plastomes of photosynthetic land plants, recent gene loss events have been documented in several disparate angiosperm clades. Among these lineages, Passiflora lacks several essential ribosomal genes, rps7, rps16, rpl20, rpl22, and rpl32, the two largest plastid genes, ycf1 and ycf2, and has a highly divergent rpoA. Comparative transcriptome analyses were performed to determine the fate of the missing genes in Passiflora. Putative functional transfers of rps7, rpl22, and rpl32 to nucleus were detected, with the nuclear transfer of rps7, representing a novel event in angiosperms. Plastid-encoded rps7 was transferred into the intron of a nuclear-encoded plastid-targeted thioredoxin m-type gene, acquiring its plastid transit peptide (TP). Plastid rpl20 likely experienced a novel substitution by a duplicated, nuclear-encoded mitochondrial-targeted rpl20 that has a similar gene structure. Additionally, among rosids, evidence for a third independent transfer of rpl22 in Passiflora was detected that gained a TP from a nuclear gene containing an organelle RNA recognition motif. Nuclear transcripts representing rpoA, ycf1, and ycf2 were not detected. Further analyses suggest that the divergent rpoA remains functional and that the gene is under positive or purifying selection in different clades. Comparative analyses indicate that alternative translocon and motor protein complexes may have substituted for the loss of ycf1 and ycf2 in Passiflora.

Unprecedented Intraindividual Structural Heteroplasmy in Eleocharis (Cyperaceae, Poales) Plastomes.

Plastid genomes (plastomes) of land plants have a conserved quadripartite structure in a gene-dense unit genome consisting of a large inverted repeat that separates two single copy regions. Recently, alternative plastome structures were suggested in Geraniaceae and in some conifers and Medicago the coexistence of inversion isomers has been noted. In this study, plastome sequences of two Cyperaceae, Eleocharis dulcis (water chestnut) and Eleocharis cellulosa (gulf coast spikerush), were completed. Unlike the conserved plastomes in basal groups of Poales, these Eleocharis plastomes have remarkably divergent features, including large plastome sizes, high rates of sequence rearrangements, low GC content and gene density, gene duplications and losses, and increased repetitive DNA sequences. A novel finding among these features was the unprecedented level of heteroplasmy with the presence of multiple plastome structural types within a single individual. Illumina paired-end assemblies combined with PacBio single-molecule real-time sequencing, long-range polymerase chain reaction, and Sanger sequencing data identified at least four different plastome structural types in both Eleocharis species. PacBio long read data suggested that one of the four E. dulcis plastome types predominates.

Comparative Mitogenome Analysis of the Genus Trifolium Reveals Independent Gene Fission of ccmFn and Intracellular Gene Transfers in Fabaceae.

The genus Trifolium is the largest of the tribe Trifolieae in the subfamily Papilionoideae (Fabaceae). The paucity of mitochondrial genome (mitogenome) sequences has hindered comparative analyses among the three genomic compartments of the plant cell (nucleus, mitochondrion and plastid). We assembled four mitogenomes from the two subgenera (Chronosemium and Trifolium) of the genus. The four Trifolium mitogenomes were compact (294,911-348,724 bp in length) and contained limited repetitive (6.6-8.6%) DNA. Comparison of organelle repeat content highlighted the distinct evolutionary trajectory of plastid genomes in a subset of Trifolium species. Intracellular gene transfer (IGT) was analyzed among the three genomic compartments revealing functional transfer of mitochondrial rps1 to nuclear genome along with other IGT events. Phylogenetic analysis based on mitochondrial and nuclear rps1 sequences revealed that the functional transfer in Trifolieae was independent from the event that occurred in robinioid clade that includes genus Lotus. A novel, independent fission event of ccmFn in Trifolium was identified, caused by a 59 bp deletion. Fissions of this gene reported previously in land plants were reassessed and compared with Trifolium.

Fluctuations in Fabaceae mitochondrial genome size and content are both ancient and recent.

BACKGROUND: Organelle genome studies of Fabaceae, an economically and ecologically important plant family, have been biased towards the plastid genome (plastome). Thus far, less than 15 mitochondrial genome (mitogenome) sequences of Fabaceae have been published, all but four of which belong to the subfamily Papilionoideae, limiting the understanding of size variation and content across the family. To address this, four mitogenomes were sequenced and assembled from three different subfamilies (Cercidoideae, Detarioideae and Caesalpinioideae). RESULTS: Phylogenetic analysis based on shared mitochondrial protein coding regions produced a fully resolved and well-supported phylogeny that was completely congruent with the plastome tree. Comparative analyses suggest that two kinds of mitogenome expansions have occurred in Fabaceae. Size expansion of four genera (Tamarindus, Libidibia, Haematoxylum, and Leucaena) in two subfamilies (Detarioideae and Caesalpinioideae) occurred in relatively deep nodes, and was mainly caused by intercellular gene transfer and/or interspecific horizontal gene transfer (HGT). The second, more recent expansion occurred in the Papilionoideae as a result of duplication of native mitochondrial sequences. Family-wide gene content analysis revealed 11 gene losses, four (rps2, 7, 11 and 13) of which occurred in the ancestor of Fabaceae. Losses of the remaining seven genes (cox2, rpl2, rpl10, rps1, rps19, sdh3, sdh4) were restricted to specific lineages or occurred independently in different clades. Introns of three genes (cox2, ccmFc and rps10) showed extensive lineage-specific length variation due to large sequence insertions and deletions. Shared DNA analysis among Fabaceae mitogenomes demonstrated a substantial decay of intergenic spacers and provided further insight into HGT between the mimosoid clade of Caesalpinioideae and the holoparasitic Lophophytum (Balanophoraceae). CONCLUSION: This study represents the most exhaustive analysis of Fabaceae mitogenomes so far, and extends the understanding the dynamic variation in size and gene/intron content. The four newly sequenced mitogenomes reported here expands the phylogenetic coverage to four subfamilies. The family has experienced multiple mitogenome size fluctuations in both ancient and recent times. The causes of these size variations are distinct in different lineages. Fabaceae mitogenomes experienced extensive size fluctuation by recruitment of exogenous DNA and duplication of native mitochondrial DNA.