Multiple losses and transfers to the nucleus of two mitochondrial succinate dehydrogenase genes during angiosperm evolution

Pervasive survival of expressed mitochondrial rps14 pseudogenes in grasses and their relatives for 80 million years following three functional transfers to the nucleus

Background

Many mitochondrial genes, especially ribosomal protein genes, have been frequently transferred as functional entities to the nucleus during plant evolution, often by an RNA-mediated process. A notable case of transfer involves the rps14 gene of three grasses (rice, maize, and wheat), which has been relocated to the intron of the nuclear sdh2 gene and which is expressed and targeted to the mitochondrion via alternative splicing and usage of the sdh2 targeting peptide. Although this transfer occurred at least 50 million years ago, i.e., in a common ancestor of these three grasses, it is striking that expressed, nearly intact pseudogenes of rps14 are retained in the mitochondrial genomes of both rice and wheat. To determine how ancient this transfer is, the extent to which mitochondrial rps14 has been retained and is expressed in grasses, and whether other transfers of rps14 have occurred in grasses and their relatives, we investigated the structure, expression, and phylogeny of mitochondrial and nuclear rps14 genes from 32 additional genera of grasses and from 9 other members of the Poales.

Results

Filter hybridization experiments showed that rps14 sequences are present in the mitochondrial genomes of all examined Poales except for members of the grass subfamily Panicoideae (to which maize belongs). However, PCR amplification and sequencing revealed that the mitochondrial rps14 genes of all examined grasses (Poaceae), Cyperaceae, and Joinvilleaceae are pseudogenes, with all those from the Poaceae sharing two 4-NT frameshift deletions and all those from the Cyperaceae sharing a 5-NT insertion (only one member of the Joinvilleaceae was examined). cDNA analysis showed that all mitochondrial pseudogenes examined (from all three families) are transcribed, that most are RNA edited, and that surprisingly many of the edits are reverse (U→C) edits. Putatively nuclear copies of rps14 were isolated from one to several members of each of these three Poales families. Multiple lines of evidence indicate that the nuclear genes are probably the products of three independent transfers.

Conclusion

The rps14 gene has, most likely, been functionally transferred from the mitochondrion to the nucleus at least three times during the evolution of the Poales. The transfers in Cyperaceae and Poaceae are relatively ancient, occurring in the common ancestor of each family, roughly 80 million years ago, whereas the putative Joinvilleaceae transfer may be the most recent case of functional organelle-to-nucleus transfer yet described in any organism. Remarkably, nearly intact and expressed pseudogenes of rps14 have persisted in the mitochondrial genomes of most lineages of Poaceae and Cyperaceae despite the antiquity of the transfers and of the frameshift and RNA editing mutations that mark the mitochondrial genes as pseudogenes. Such long-term, nearly pervasive survival of expressed, apparent pseudogenes is to our knowledge unparalleled in any genome. Such survival probably reflects a combination of factors, including the short length of rps14, its location immediately downstream of rpl5 in most plants, and low rates of nucleotide substitutions and indels in plant mitochondrial DNAs. Their survival also raises the possibility that these rps14 sequences may not actually be pseudogenes despite their appearance as such. Overall, these findings indicate that intracellular gene transfer may occur even more frequently in angiosperms than already recognized and that pseudogenes in plant mitochondrial genomes can be surprisingly resistant to forces that lead to gene loss and inactivation.

Evolutionary transfers of mitochondrial genes to the nucleus in the Populus lineage and coexpression of nuclear and mitochondrial Sdh4 genes

The transfer of mitochondrial genes to the nucleus is an ongoing evolutionary process in flowering plants. Evolutionarily recent gene transfers provide insights into the evolutionary dynamics of the process and the way in which transferred genes become functional in the nucleus. Genes that are present in the mitochondrion of some angiosperms but have been transferred to the nucleus in the Populus lineage were identified by searches of Populus sequence databases. Sequence analyses and expression experiments were used to characterize the transferred genes. Two succinate dehydrogenase genes and six mitochondrial ribosomal protein genes have been transferred to the nucleus in the Populus lineage and have become expressed. Three transferred genes have gained an N-terminal mitochondrial targeting presequence from other pre-existing genes and two of the transferred genes do not contain an N-terminal targeting presequence. Intact copies of the succinate dehydrogenase gene Sdh4 are present in both the mitochondrion and the nucleus. Both copies of Sdh4 are expressed in multiple organs of two Populus species and RNA editing occurs in the mitochondrial copy. These results provide a genome-wide perspective on mitochondrial genes that were transferred to the nucleus and became expressed, functional genes during the evolutionary history of Populus.

Origin, evolution and genetic effects of nuclear insertions of organelle DNA

In eukaryotes, nuclear genomes are subject to an influx of DNA from mitochondria and plastids. The nuclear insertion of organellar sequences can occur during the illegitimate repair of double-stranded breaks. After integration, nuclear organelle DNA is modified by point mutations, and by deletions. Insertion of organelle DNA into nuclear genes is not rare and can potentially have harmful effects. In humans, some insertions of nuclear mitochondrial DNA are associated with heritable diseases. It remains to be determined whether nuclear organelle DNA can contribute beneficially to gene evolution.

Generation and evolutionary fate of insertions of organelle DNA in the nuclear genomes of flowering plants

Nuclear genomes are exposed to a continuous influx of DNA from mitochondria and plastids. We have characterized the structure of approximately 750 kb of organelle DNA, distributed among 13 loci, in the nuclear genomes of Arabidopsis and rice. These segments are large and migrated to the nucleus quite recently, allowing us to reconstruct their evolution. Two general types of nuclear insertions coexist; one is characterized by long sequence stretches that are colinear with organelle DNA, the other type consists of mosaics of organelle DNA, often derived from both plastids and mitochondria. The levels of sequence divergence of the two types exclude their common descent, implying that at least two independent modes of DNA transfer from organelle to nucleus operate. The post-integration fate of organelle DNA is characterized by a predominance of transition mutations, associated with the gradual amelioration of the integrated sequence to the nucleotide composition of the host chromosome. Deletion of organelle DNA at these loci is essentially balanced by insertions of nonorganelle DNA. Deletions are associated with the removal of DNA between perfect repeats, indicating that they originate by replication slippage.

The complete nucleotide sequence and multipartite organization of the tobacco mitochondrial genome: comparative analysis of mitochondrial genomes in higher plants

Tobacco is a valuable model system for investigating the origin of mitochondrial DNA (mtDNA) in amphidiploid plants and studying the genetic interaction between mitochondria and chloroplasts in the various functions of the plant cell. As a first step, we have determined the complete mtDNA sequence of Nicotiana tabacum. The mtDNA of N. tabacum can be assumed to be a master circle (MC) of 430,597 bp. Sequence comparison of a large number of clones revealed that there are four classes of boundaries derived from homologous recombination, which leads to a multipartite organization with two MCs and six subgenomic circles. The mtDNA of N. tabacum contains 36 protein-coding genes, three ribosomal RNA genes and 21 tRNA genes. Among the first class, we identified the genes rps1 and psirps14, which had previously been thought to be absent in tobacco mtDNA on the basis of Southern analysis. Tobacco mtDNA was compared with those of Arabidopsis thaliana, Beta vulgaris, Oryza sativa and Brassica napus. Since repeated sequences show no homology to each other among the five angiosperms, it can be supposed that these were independently acquired by each species during the evolution of angiosperms. The gene order and the sequences of intergenic spacers in mtDNA also differ widely among the five angiosperms, indicating multiple reorganizations of genome structure during the evolution of higher plants. Among the conserved genes, the same potential conserved nonanucleotide-motif-type promoter could only be postulated for rrn18-rrn5 in four of the dicotyledonous plants, suggesting that a coding sequence does not necessarily move with the promoter upon reorganization of the mitochondrial genome.

Nuclear SDH2-1 and SDH2-2 genes, encoding the iron-sulfur subunit of mitochondrial complex II in Arabidopsis, have distinct cell-specific expression patterns and promoter activities

Three different nuclear genes encode the essential iron-sulfur subunit of mitochondrial complex II (succinate dehydrogenase) in Arabidopsis (Arabidopsis thaliana), raising interesting questions about their origin and function. To find clues about their role, we have undertaken a detailed analysis of their expression. Two genes (SDH2-1 and SDH2-2) that likely arose via a relatively recent duplication event are expressed in all organs from adult plants, whereas transcripts from the third gene (SDH2-3) were not detected. The tissue- and cell-specific expression of SDH2-1 and SDH2-2 was investigated by in situ hybridization. In flowers, both genes are regulated in a similar way. Enhanced expression was observed in floral meristems and sex organ primordia at early stages of development. As flowers develop, SDH2-1 and SDH2-2 transcripts accumulate in anthers, particularly in the tapetum, pollen mother cells, and microspores, in agreement with an essential role of mitochondria during anther development. Interestingly, in contrast to the situation in flowers, only SDH2-2 appears to be expressed at a significant level in root tips. Strong labeling was observed in all cell layers of the root meristematic zone, and a cell-specific pattern of expression was found with increasing distance from the root tip, as cells attain their differentiated state. Analysis of transgenic Arabidopsis plants carrying SDH2-1 and SDH2-2 promoters fused to the beta-glucuronidase reporter gene indicate that both promoters have similar activities in flowers, driving enhanced expression in anthers and/or pollen, and that only the SDH2-2 promoter is active in root tips. These beta-glucuronidase staining patterns parallel those obtained by in situ hybridization, suggesting transcriptional regulation of these genes. Progressive deletions of the promoters identified regions important for SDH2-1 expression in anthers and/or pollen and for SDH2-2 expression in anthers and/or pollen and root tips. Interestingly, regions driving enhanced expression in anthers are differently located in the two promoters.

Mitochondria of Protists

Over the past several decades, our knowledge of the origin and evolution of mitochondria has been greatly advanced by determination of complete mitochondrial genome sequences. Among the most informative mitochondrial genomes have been those of protists (primarily unicellular eukaryotes), some of which harbor the most gene-rich and most eubacteria-like mitochondrial DNAs (mtDNAs) known. Comparison of mtDNA sequence data has provided insights into the radically diverse trends in mitochondrial genome evolution exhibited by different phylogenetically coherent groupings of eukaryotes, and has allowed us to pinpoint specific protist relatives of the multicellular eukaryotic lineages (animals, plants, and fungi). This comparative genomics approach has also revealed unique and fascinating aspects of mitochondrial gene expression, highlighting the mitochondrion as an evolutionary playground par excellence.

Sequence and comparative analysis of the maize NB mitochondrial genome

The NB mitochondrial genome found in most fertile varieties of commercial maize (Zea mays subsp. mays) was sequenced. The 569,630-bp genome maps as a circle containing 58 identified genes encoding 33 known proteins, 3 ribosomal RNAs, and 21 tRNAs that recognize 14 amino acids. Among the 22 group II introns identified, 7 are trans-spliced. There are 121 open reading frames (ORFs) of at least 300 bp, only 3 of which exist in the mitochondrial genome of rice (Oryza sativa). In total, the identified mitochondrial genes, pseudogenes, ORFs, and cis-spliced introns extend over 127,555 bp (22.39%) of the genome. Integrated plastid DNA accounts for an additional 25,281 bp (4.44%) of the mitochondrial DNA, and phylogenetic analyses raise the possibility that copy correction with DNA from the plastid is an ongoing process. Although the genome contains six pairs of large repeats that cover 17.35% of the genome, small repeats (20-500 bp) account for only 5.59%, and transposable element sequences are extremely rare. MultiPip alignments show that maize mitochondrial DNA has little sequence similarity with other plant mitochondrial genomes, including that of rice, outside of the known functional genes. After eliminating genes, introns, ORFs, and plastid-derived DNA, nearly three-fourths of the maize NB mitochondrial genome is still of unknown origin and function.

Genetic Correction of Mitochondrial Diseases: Using the Natural Migration of Mitochondrial Genes to the Nucleus in Chlorophyte Algae as a Model System

Mitochondrial diseases display great diversity in clinical symptoms and biochemical characteristics. Although mtDNA mutations have been identified in many patients, there are currently no effective treatments. A number of human diseases result from mutations in mtDNA-encoded proteins, a group of proteins that are hydrophobic and have multiple membrane-spanning regions. One method that has great potential for overcoming the pathogenic consequences of these mutations is to place a wild-type copy of the affected gene in the nucleus, and target the expressed protein to the mitochondrion to function in place of the defective protein. Several respiratory chain subunit genes, which are typically mtDNA encoded, are nucleus encoded in the chlorophyte algae Chlamydomonas reinhardtii and Polytomella sp. Analysis of these genes has revealed adaptations that facilitated their expression from the nucleus. The nucleus-encoded proteins exhibited diminished physical constraints for import as compared to their mtDNA-encoded homologues. The hydrophobicity of the nucleus-encoded proteins is diminished in those regions that are not involved in subunit-subunit interactions or that contain amino acids critical for enzymatic reactions of the proteins. In addition, these proteins have unusually large mitochondrial targeting sequences. Information derived from these studies should be applicable toward the development of genetic therapies for human diseases resulting from mutations in mtDNA-encoded polypeptides.

Transfer of RPS14 and RPL5 from the mitochondrion to the nucleus in grasses

Gene transfer from the mitochondrion to the nucleus, a process of outstanding importance to the evolution of the eukaryotic cell, is an on-going phenomenon in higher plants. After transfer, the mitochondrial gene has to be adapted to the nuclear context by acquiring a new promoter and targeting information to direct the protein back to the organelle. To better understand the strategies developed by higher plants to transfer organellar genes during evolution, we investigated the fate of the mitochondrial RPL5-RPS14 locus in grasses. While maize mitochondrial genome does not contain RPS14 and RPL5 genes, wheat mitochondrial DNA contains an intact RPL5 gene and a nonfunctional RPS14 pseudogene. RPL5 and PsiRPS14 are co-transcribed and their transcripts are edited. In wheat, the functional RPS14 gene is located in the nucleus, within the intron of the respiratory complex II iron-sulfur subunit gene (SDH2). Its organization and expression mechanisms are similar to those previously described in maize and rice, allowing us to conclude that RPS14 transfer and nuclear activation occurred before divergence of these grasses. Unexpectedly, we found evidence for a more recent RPL5 transfer to the nucleus in wheat. This nuclear wheat RPL5 acquired its targeting information by duplication of an existing targeting presequence for another mitochondrial protein, ribosomal protein L4. Thus, mitochondrial and nuclear functional RPL5 genes appear to be maintained in wheat, supporting the hypothesis that in an intermediate stage of the transfer process, both nuclear and mitochondrial functional genes coexist. Finally, we show that RPL5 has been independently transferred to the nucleus in the maize lineage and has acquired regulatory elements for its expression and a mitochondrial targeting peptide from an unknown source.

Abundance of plastid DNA insertions in nuclear genomes of rice and Arabidopsis

Pairwise comparison of whole plastid and draft nuclear genomic sequences of Arabidopsis thaliana and Oryza sativa L. ssp. indica shows that rice nuclear genomic sequences contain homologs of plastid DNA covering about 94 kb (83%) of plastid genome and including one or more full-length intact (without mutations resulting in premature stop codons) homologues of 26 known protein-coding (KPC) plastid genes. By contrast, only about 20 kb (16%) of chloroplast DNA, including a single intact plastid-derived KPC gene, is presented in the nucleus of A. thaliana. Sixteen rice plastid genes have at least one nuclear copy without any mutation or with only synonymous substitutions. Nuclear copies for other ten plastid genes contain both synonymous and non-synonymous substitutions. Multiple ESTs for 25 out of 26 KPC genes were also found, as well as putative promoters for some of them. The study of substitutions pattern shows that some of nuclear homologues of plastid genes may be functional and/or are under the pressure of the positive natural selection. The similar comparative analysis performed on rice chromosome 1 revealed 27 contigs containing plastid-derived sequences, totalling about 84 kb and covering two thirds of chloroplast DNA, with the intact nuclear copies of 26 different KPC genes. One of these contigs, AP003280, includes almost 57 kb (45%) of chloroplast genome with the intact copies of 22 KPC genes. At the same time, we observed that relative locations of homologues in plastid DNA and the nuclear genome are significantly different.

The four subunits of mitochondrial respiratory complex II are encoded by multiple nuclear genes and targeted to mitochondria in Arabidopsis thaliana

Mitochondrial respiratory complex II contains four subunits: a flavoprotein (SDHI), an iron-sulphur subunit (SDH2) and two membrane anchor subunits (SDH3 and SDH4). We have found that in Arabidopsis thaliana SDH I and SDH3 are encoded by two, and SDH4 by one nuclear genes, respectively. All these encoded polypeptides are found to be imported into isolated plant mitochondria. While both SDHI proteins are highly conserved when compared to their counterparts in other organisms, SDH3 and SDH4 share little similarity with non-plant homologues. Expression of SDH1-1, SDH3 and SDH4 genes was detected in all tissues analysed, with the highest steady-state mRNA levels found in flowers and inflorescences. In contrast, the second SDH1 gene (SDH1-2) is expressed at a low level.

Intracellular gene transfer: reduced hydrophobicity facilitates gene transfer for subunit 2 of cytochrome c oxidase

Subunit 2 of cytochrome c oxidase (Cox2) in legumes offers a rare opportunity to investigate factors necessary for successful gene transfer of a hydrophobic protein that is usually mitochondrial-encoded. We found that changes in local hydrophobicity were necessary to allow import of this nuclear-encoded protein into mitochondria. All legume species containing both a mitochondrial and nuclear encoded Cox2 displayed a similar pattern, with a large decrease in hydrophobicity evident in the first transmembrane region of the nuclear encoded protein compared with the organelle-encoded protein. Mitochondrial-encoded Cox2 could not be imported into mitochondria under the direction of the mitochondrial targeting sequence that readily supports the import of nuclear encoded Cox2. Removal of the first transmembrane region promotes import ability of the mitochondrial-encoded Cox2. Changing just two amino acids in the first transmembrane region of mitochondrial-encoded Cox2 to the corresponding amino acids in the nuclear encoded Cox2 also promotes import ability, whereas changing the same two amino acids in the nuclear encoded Cox2 to what they are in the mitochondrial-encoded copy prevents import. Therefore, changes in amino acids in the mature protein were necessary and sufficient for gene transfer to allow import under the direction of an appropriate signal to achieve the functional topology of Cox2.

Punctuated evolution of mitochondrial gene content: high and variable rates of mitochondrial gene loss and transfer to the nucleus during angiosperm evolution

To study the tempo and pattern of mitochondrial gene loss in plants, DNAs from 280 genera of flowering plants were surveyed for the presence or absence of 40 mitochondrial protein genes by Southern blot hybridization. All 14 ribosomal protein genes and both sdh genes have been lost from the mitochondrial genome many times (6 to 42) during angiosperm evolution, whereas only two losses were detected among the other 24 genes. The gene losses have a very patchy phylogenetic distribution, with periods of stasis followed by bursts of loss in certain lineages. Most of the oldest groups of angiosperms are still mired in a prolonged stasis in mitochondrial gene content, containing nearly the same set of genes as their algal ancestors more than a billion years ago. In sharp contrast, other plants have rapidly lost many or all of their 16 mitochondrial ribosomal protein and sdh genes, thereby converging on a reduced gene content more like that of an animal or fungus than a typical plant. In these and many lineages with more modest numbers of losses, the rate of ribosomal protein and sdh gene loss exceeds, sometimes greatly, the rate of mitochondrial synonymous substitutions. Most of these mitochondrial gene losses are probably the consequence of gene transfer to the nucleus; thus, rates of functional gene transfer also may vary dramatically in angiosperms.

Genes for two mitochondrial ribosomal proteins in flowering plants are derived from their chloroplast or cytosolic counterparts

Often during flowering plant evolution, ribosomal protein genes have been lost from the mitochondrion and transferred to the nucleus. Here, we show that substitution by a duplicated, divergent gene originally encoding the chloroplast or cytosolic ribosomal protein counterpart accounts for two missing mitochondrial genes in diverse angiosperms. The rps13 gene is missing from the mitochondrial genome of many rosids, and a transferred copy of this gene is not evident in the nucleus of Arabidopsis, soybean, or cotton. Instead, these rosids contain a divergent nuclear copy of an rps13 gene of chloroplast origin. The product of this gene from all three rosids was shown to be imported into isolated mitochondria but not into chloroplasts. The rps8 gene is missing from the mitochondrion and nucleus of all angiosperms examined. A divergent copy of the gene encoding its cytosolic counterpart (rps15A) was identified in the nucleus of four angiosperms and one gymnosperm. The product of this gene from Arabidopsis and tomato was imported successfully into mitochondria. We infer that rps13 was lost from the mitochondrial genome and substituted by a duplicated nuclear gene of chloroplast origin early in rosid evolution, whereas rps8 loss and substitution by a gene of nuclear/cytosolic origin occurred much earlier, in a common ancestor of angiosperms and gymnosperms.

Gene transfer from mitochondrion to nucleus: novel mechanisms for gene activation from Cox2

The evolutionarily recent transfer of the gene for cytochrome c oxidase subunit 2 (cox2) from the mitochondrion to the nucleus in legumes is shown to have involved novel gene-activation steps. The acquired mitochondrial targeting presequence is bordered by two introns. Characterization of the import of soybean Cox2 indicates that the presequence is cleaved in a three-step process which is independent of assembly. The final processing step takes place only in the mitochondria of legume species, and not in several non-legume plants. The unusually long presequence of 136 amino acids consists of three regions: the first 20 amino acids are required for mitochondrial targeting and can be replaced by another presequence; the central portion of the presequence is required for efficient import of the Cox2 protein into mitochondria; and the last 12 amino acids, derived from the mitochondrially encoded protein, are required for correct maturation of the imported protein. The acquisition of a unique presequence, and the capacity for legume mitochondria to remove this presequence post-import, are considered to be essential adaptations for targeting of Cox2 to the mitochondrion and therefore activation of the transferred gene in the nucleus.

Mitochondrial gene transfer in pieces: fission of the ribosomal protein gene rpl2 and partial or complete gene transfer to the nucleus

Mitochondrial genes are usually conserved in size in angiosperms. A notable exception is the rpl2 gene, which is considerably shorter in the eudicot Arabidopsis than in the monocot rice. Here, we show that a severely truncated mitochondrial rpl2 gene (termed 5' rpl2) was created by the formation of a premature stop codon early in eudicot evolution. This 5' rpl2 gene was subsequently lost many times from the mitochondrial DNAs of 179 core eudicots surveyed by Southern hybridization. The sequence corresponding to the 3' end of rice rpl2 (termed 3' rpl2) has been lost much more pervasively among the mitochondrial DNAs of core eudicots than has 5' rpl2. Furthermore, where still present in these mitochondrial genomes, 3' rpl2 always appears to be a pseudogene, and there is no evidence that 3' rpl2 was ever a functional mitochondrial gene. An intact and expressed 3' rpl2 gene was discovered in the nucleus of five diverse eudicots (tomato, cotton, Arabidopsis, soybean, and Medicago). In the first three of these species, 5' rpl2 is still present in the mitochondrion, unlike the two legumes, where both parts of rpl2 are present in the nucleus as separate genes. The full-length rpl2 gene has been transferred intact to the nucleus in maize. We propose that the 3' end of rpl2 was functionally transferred to the nucleus early in eudicot evolution, and that this event then permitted the nonsense mutation that gave rise to the mitochondrial 5' rpl2 gene. Once 5' rpl2 was established as a stand-alone mitochondrial gene, it was then lost, and was probably transferred to the nucleus many times. This complex history of gene fission and gene transfer has created four distinct types of rpl2 structures or compartmentalizations in angiosperms: (1) intact rpl2 gene in the mitochondrion, (2) intact gene in the nucleus, (3) split gene, 5' in the mitochondrion and 3' in the nucleus, and (4) split gene, both parts in the nucleus.