Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

I. Mt Transformation

A. Introduction

Two types of mt DNA are present in the embryophytes (land plants). The mitochondria of the three clades of bryophytes (liverworts, hornworts and mosses, see chapter “Promiscuous Organellar DNA”) possess circular genomes of around 105–185 kb resembling that of their green algal relatives like Chara and even the early-branching nongreen eukaryote, Reclinomonas americana (Li et al. 2009). They are therefore considered to be of the ancestral type despite the acquisition of several additional features such as moderate size increase, intron gain and RNA editing, which are generally absent in algal genomes (Wang et al. 2009). On the contrary, the mitochondria of seed plants (see Chap. 10 “Horizontal Gene Transfer in Eukaryotes: Fungi-to-Plant and Plant-to-Plant Transfers of organellar DNA”) harbor much larger genomes, from 221 kb for the smallest mt genome sequenced to date (Handa 2003) to more than 2,000 kb in the Cucurbitaceae family (Ward et al. 1981). The size increase does not reflect an increase of the gene content but addition of noncoding sequences such as introns, DNA repeat motifs and insertion of nuclear and chloroplast fragments (Knoop 2004). In addition seed plants possess a complex pool of frequently recombining molecules, the stoichiometry of which is controlled by nuclear genes (Abdelnoor et al. 2003; Arrieta-Montiel et al. 2009). Plant mt transcription is also complex and plant mt RNA transcripts undergo unique post-transcriptional modifications such as editing and trans-splicing (see section “DNA and RNA Delivery into Plant Mitochondria” of this chapter). Investigations of such processes remain difficult due to the impossibility to stably transform plant mitochondria and to manipulate mt gene expression. A plant mt transformation system would therefore be of great value.

Several significant seed plant mtDNA mutations (see chapter “Expression Profiling of Organellar Genes”) were identified. Among them, some are found in subunits of the respiratory-chain complexes and are responsible for visible phenotypes such as the cytoplasmic male sterility in tobacco (Pla et al. 1995; Gutierres et al. 1997; Pineau et al. 2005) or the nonchromosomal stripe phenotype in maize (Karpova and Newton 1999; Kubo and Newton 2008).

In principle, and similarly to the situation in other unicellular organisms like the yeast Saccharomyces cerevisiae (Fox et al. 1988; Johnston et al. 1988) and in the green alga Chlamydomonas (see below for details), such respiratory deficient mutants could serve as recipient strains for mt transformation. A few years ago, cucumber was described as a promising system for mt transformation for several reasons: (1) the mt genome is inherited by the paternal parent and microspores possess a few huge mitochondria that could be transformed before ovule fertilization, (2) mutants with mt deletions, responsible for a mosaic phenotype, exist and could serve as non-reverting recipient strains for mt ­transformation (Havey et al. 2002). However, despite several attempts, no successful transformation of mitochondria in cucumber or any plant system has been reported to date. The reasons for the failure are probably multiple: (1) the lack of a method to transform either microspores or pollen cells or the ovule for plants with maternal transmission of mitochondria; (2) the fact that plant mt genomes are extremely complex and unstable and can rapidly evolve via substoichiometric shifting due to recombination (Small et al. 1987, 1989; Mackenzie 2007); (3) the lack of a selection that is maintained throughout the formation of the adult plant after zygote formation. Auxotrophic markers corresponding to nuclear genes encoding enzymes targeted to mitochondria are attractive for the development of a selection method. In yeast, expression of the synthetic gene ARG8m from the mt genome allows nuclear arg8 mutants to grow without arginine (Steele et al. 1996). The Arg8p protein is normally imported into mitochondria from the cytoplasm, but also functions when synthesized within the organelle in the mt transformants. Thus, arginine prototrophy can become a phenotype dependent on mt gene expression.

Unfortunately, there are very few auxotrophic markers that could function within mitochondria of plants. For example, the Chlamydomonas or Arabidopsis ortholog of ARG8 cannot be used as a mitochondrial marker since it actually encodes a plastid protein (Remacle et al. 2009). One notable exception is the BIO2 protein, an iron-cluster enzyme responsible for the last step of biotin synthesis inside the mitochondria. The bio2 mutants defective for the BIO2 protein have been characterized in Arabidopsis and they survive if supplemented with biotin (Patton et al. 1998; Arnal et al. 2006). The bio2 mutants may represent a promising system for mitochondrial transformation, by transforming either protoplasts with subsequent regeneration of plants or the ovule before pollen fertilization, using as selection a medium devoid of biotin.

B. Mt Transformation in the Unicellular Green Alga Chlamydomonas

1. The Mt Genome of Chlamydomonas

The mt genome of C. reinhardtii is a 15.8-kb linear molecule containing at each extremity telomeres corresponding to inverted repeats of about 500 bp, with 40-bp single-stranded extensions (Vahrenholz et al. 1993). Thirteen genes that encode five subunits of the NADH:ubiquinone oxidoreductase or complex I (nd1, nd2, nd4, nd5 and nd6), apocytochrome b of the bc 1 complex or complex III (cob), subunit 1 of cytochrome c oxidase or complex IV (cox1), a reverse transcriptase-like protein (rtl), three tRNAs (trnW, trnQ, trnM) and the ribosomal RNAs (rRNAs) are present in the mt genome (Fig. 1). The rRNA genes are discontinuous and split into mini-sequences encoding four small subunit (S) and eight large subunit (L) rRNA modules, interspersed with one another and with protein and tRNA genes. The small rRNA segments of the two ribosomal subunits are believed to interact by way of extensive intermolecular pairing between one another to form conventional rRNA molecules (Boer and Gray 1988).

Fig. 19.1.
figure 1

Physical map of the 15.8 kb mt genome of C. reinhardtii. The rectangles represent protein-coding genes: cob, gene encoding apocytochrome b of complex III; nd1, 2, 4, 5, and 6, genes encoding the corresponding subunits of complex I; cox1, gene encoding subunit 1 of complex IV, rtl: reverse transcriptase-like protein. L and S represent modules encoding segments of rRNAs of the large and the small ribosomal subunits, respectively. W, Q, and M represent tRNAs for Trp, Gln, and Met, respectively. The inverted telomeric ends are represented by short arrows and the bidirectional origin of transcription between nd5 and cox1 by longer arrows. Positions of the dum1, dum11 and dum22 deletions and of the dum18, dum19 and dum25 mutations are indicated.

Full size image

The mt genome is a multicopy system of around 50–100 copies organized into about 20–30 nucleoids (Nishimura et al. 1998; Hiramatsu et al. 2006). The nucleoids and the mitochondria seem extremely dynamic and undergo changes in their size and shape during the cell cycle (Ehara et al. 1995; Hiramatsu et al. 2006).

2. The Mt Mutants of Chlamydomonas

Several mutations altering the mt cob, cox1 and nd genes have been isolated following random mutagenesis with acriflavine (Remacle et al. 2001b). Phenotypically, the mutants in the cob and cox1 genes have lost the capacity to grow under heterotrophic conditions i.e. in the dark, with acetate as carbon source because they lack the cytochrome pathway of respiration. In contrast, mutants altered in the nd genes, which encode subunits of complex I, are able to grow in the dark, but considerably more slowly than the wild-type strain (Remacle et al. 2001a; Cardol et al. 2002). Most of the mutations located in the cob gene are deletions covering not only the coding sequence but also the left telomere, whereas mutations in the cox1 and nd genes are usually frameshifts caused by deletion or insertion of one thymidine (Remacle et al. 2001b). In addition, the mud2 mutation at codon 129 (Phe TTC → Leu CTC) of the cob gene confers resistance to myxothiazol and mucidine, inhibitors of the cytochrome bc 1 complex (Bennoun et al. 1991).

All the mt point mutants studied so far were found to be homoplasmic, i.e. they contain only mutated mt genomes. However, the mt DNA present in the deletion mutants exists as a mixture of both deleted monomers and dimers arising from head-to-head fusions between deleted monomers (Matagne et al. 1989; Dorthu et al. 1992; Duby and Matagne 1999). The total amount of mt DNA in such mutants is generally lower than in the wild type. The deletion mutants do not revert and are ideally suited as recipients for mt transformation experiments.

3. Recombination and Segregation of mt DNA

In Chlamydomonas, homologous mt DNA recombination is only detected after crosses between mt + and mt − strains in mitotic zygotes that do not undergo meiosis. In such zygotes, mt DNA is transmitted by both parents and recombination between the parental genomes is frequent (Remacle et al. 1990; Remacle and Matagne 1993). This demonstrates that enzymes involved in homologous recombination are active in Chlamydomonas mitochondria, an important feature for mt transformation as stable integration of the transforming DNA relies on recombination. Segregation of the mt genomes occurs in the zygotes and their mitotic progeny and after 15–20 divisions, most of the diploid cells are homoplasmic for a mt genome, either recombined or parental.

4. Mt Transformation

a. Deletion Mutants as Recipient for mt Transformation

In the first report of mt transformation in Chlamydomonas, a mutant (dum1) deleted for the left telomere and cob gene (1.5 kb deletion) was successfully transformed to respiratory competence with partially purified mt DNA from C. reinhardtii or Chlamydomonas smithii using a biolistic device (Randolph-Anderson et al. 1993). C. reinhardtii and C. smithii are two interfertile species, which harbor identical mt genomes with the exception of a 1-kb group I intron located in the cob gene that is present in C. smithii but absent in C. reinhardtii (Boynton et al. 1987; Remacle et al. 1990). Later, biolistics was again used successfully to transform the same recipient strain, with purified mt DNA or cloned mt DNA fragments (Yamasaki et al. 2005). In both cases, the wild-type mt sequence of the transforming DNA had replaced the deleted genome in the transformants selected under heterotrophic conditions (dark+acetate). Transformation efficiency was low (0.4–3 transformants/μg DNA). This precluded any genetic manipulation of the mt genome, since isolation of transformants with the desired genotype usually requires the screening of many colonies, as this is also the case in Saccharomyces mt transformation (Bonnefoy et al. 2007).

Subsequently, biolistic transformation was optimized using cloned mt DNA or PCR fragments as transforming molecules (Remacle et al. 2006). Another deletion mutant carrying a 1.2 kb deletion including the left telomere and part of the cob gene (dum11) could be rescued after selection in the dark using a mt DNA fragment covering the deletion and the cob gene as donor DNA (Fig. 1). Homologous recombination occurred between the introduced DNA and the endogenous mt genome and homologous sequences as short as 28 nucleotides could direct recombination (Remacle et al. 2006). Mt transformants were homoplasmic for the 15.8 kb wild-type genome and did not exhibit the dimeric forms of the mt genome that were present in the recipient strain. Moreover, a high transformation efficiency was achieved (100–250 transformants/μg DNA), the best results being obtained with linearized plasmid DNA.

Interestingly, a strain lacking the cob gene and the left telomere could be rescued, although at a very low rate, when the transforming DNA is nearly completely devoid of the left telomere (Remacle et al. 2006). This indicated that the right telomere can be copied to reconstruct the left telomere by recombination. Using the strategy described above, we were able to introduce non-deleterious mutations and also loss-of-function molecular lesions in the mt genome. Myxothiazol-resistant transformants were generated by introducing the nucleotide substitution that is present in the cob gene of the strains displaying myxothiazol resistance (Remacle et al. 2006). Similarly, an in-frame deletion of 23 codons was reconstructed in the nd4 gene with a frequency of one homoplasmic nd4 transformant among 90 transformants analyzed (Remacle et al. 2006). During selection in the dark, recombination events resulted in the co-integration of the cob gene and the deletion in nd4 in some molecules of the mt genome despite the negative effect of the nd4 mutation on complex I assembly and activity as well as on whole cell respiration (Fig. 1) (Remacle et al. 2006). These results open the way to reverse genetics in Chlamydomonas mitochondria and more specifically, to site-directed mutagenesis of mitochondrially encoded subunits of complex I (ND subunits). This is of special interest because the yeast S. cerevisiae, whose mt genome can be manipulated virtually at will, lacks complex I.

After a 2 month selection in the dark, heteroplasmy could still be detected in some transformants, suggesting that the segregation process of the mt molecules was extremely slow. To circumvent the problem of heteroplasmy, we recently used a deletion mutant with a deletion extending up to nd4 (the dum22 mutant) (Remacle et al. 2001b) to force the insertion of mutations into nd4 by recombination. The dum22 mutant is absolutely dependent on glycolysis and the chloroplast for ATP formation since it lacks the three key protein complexes of the respiratory chain (complex I, complex III and IV). This mutant has proven to be a good recipient strain for isolation of homoplasmic nd4 transformants, although transformation efficiency is very low (1–2 transformants/μg of DNA). For example, using the dum22 mutant as recipient strain, we have been recently able to isolate two homoplasmic transformants affected in nd4. One transformant bears a point mutation (Leu TTG → Pro CCA) corresponding to a human disease (Chronic Progressive External Ophthalmoplegia) at codon 158 of nd4 (V. Larosa, unpublished). The other one has been created to study mt codon usage and bears a set of 11 codons, which differ from the usual mt codon usage (T. Salinas and C. Remacle, unpublished). It is worth mentioning that a similar transformant remained heteroplasmic when using the dum11 mutant only affected in cob as a recipient strain (T. Salinas, unpublished).

b. Frameshift Mutants as Recipients for mt Transformation

Point mutants can also be used for mt transformation. A double frameshift mutant in both the cox1 and nd1 genes (dum19 dum25) (Fig. 1) could be rescued for heterotrophic growth and the 23 codon in-frame deletion in nd4 cited above could be reconstructed (Remacle et al. 2006). Another frameshift mutant (dum18) in the cox1 gene was also employed as a recipient strain for transformation (Colin et al. 1995). This mutant seems best suited for site-directed mutagenesis of the nd genes since the cox1 gene lies between the nd4 and nd5 genes on one end and the nd2 and nd6 genes on the other end of the genome (Fig. 1). Unfortunately, the high frequency of reversion of this frameshift mutation precluded the use of this strain as a recipient for transformation experiments.

5. Toward a Selection Independent of the Restoration of Heterotrophic Growth

An obvious limitation to mt transformation in Chlamydomonas is the selection process that requires a 2-month incubation period in the dark before any molecular characterization can be performed. The development of a more rapid selection using phototrophic growth for the generation of mt transformants is therefore a high priority. We first tested a co-transformation strategy using a primary selection in the light, by bombarding the recipient strain with a plasmid carrying a nuclear marker and another construct containing the mt DNA of interest. Nuclear transformants were first selected in the light and then tested for their respiratory competence to detect mt transformation events (Remacle et al. 2006). This type of selection, commonly used for yeast transformation (Bonnefoy et al. 2007), was not successful when applied to Chlamydomonas, presumably because simultaneous transformation of the nuclear and mitochondrial genomes is a extremely rare event (Remacle et al. 2006). A novel method of selection for mt transformation based on the use of the maize URF13 protein is currently being tested. URF13 is a chimeric mt inner membrane protein that arose from the recombination of mt molecules (Hanson 1991). The presence of URF13 in the mitochondria causes cytoplasmic male sterility and susceptibility to methomyl, an insecticide commonly used for crops (Hanson 1991; Levings and Siedow 1992; Rhoads et al. 1995). Interestingly, the expression of a mitochondrially targeted URF13 in the fungi S. cerevisiae and Pichia pastoris also confers methomyl sensitivity (Glab et al. 1990; Huang et al. 1990; Soderholm et al. 2001). A Chlamydomonas recipient strain carrying a codon-optimized version of URF13 in place of the nd4 gene will be created by biolistic transformation using dark selection (nd4::URF13). We expect the nd4::URF13 strain to be methomyl sensitive (and deficient for complex I) based on the fact that expression of URF13 in the mitochondria of Saccharomyces, Pichia and tobacco confers this trait (Glab et al. 1990; Huang et al. 1990; von Allmen et al. 1991; Chaumont et al. 1995; Soderholm et al. 2001). As a proof of concept, the methomyl sensitive nd4::URF13 recipient strain will be transformed with a construct containing the wild-type nd4 gene and selection of methomyl resistant transformants will be attempted in the light. Such transformants are expected to arise from homologous recombination between the nd4::URF13 region in the mt DNA and the transforming DNA carrying the wild-type nd4 gene. If such a selection is successfully established, mutations in the nd4 gene resulting in complex I deficiency in humans will be reconstructed in the Chlamy­domonas mt genome of the nd4::URF13 strain using the methomyl resistance selection. The same methodology can be applied to manipulate other mt nd genes.

II. DNA and RNA Delivery into Plant Mitochondria

A.Cytosolic tRNA Import into Plant Mitochondria

Mitochondria perform protein biosynthesis. Therefore, they require a complete set of transfer RNAs (tRNAs). The availability of several complete mt genomes of land plants and algae allowed the identification of mt trn genes encoding tRNAs (O’Brien et al. 2009). In algae (with the exception of C. reinhardtii and Polytomella capuana) and in bryophytes, the number of trn genes seems sufficient or nearly sufficient for mt translation (O’Brien et al. 2009). In contrast, in seed plants the number of trn genes is clearly insufficient as trn genes for tRNAs corresponding to 5–7 amino acids are absent from the mt genomes. Experimental studies in a number of these organisms showed that this lack is compensated by the import of the corresponding cytosolic tRNAs (Glover et al. 2001; Vinogradova et al. 2009). Transport of tRNAs from the cytosol to mitochondria is not restricted to plants but is a widespread process that also occurs in organisms that would not need to import tRNAs to sustain mt translation (Salinas et al. 2008; Alfonzo and Soll 2009; Lithgow and Schneider 2010). An important feature in plants is that the number and the identity of imported tRNAs vary from one species to another and are not always consistent with the assigned phylogenetic position. The significance of this observation is currently unclear. Interestingly, with a few notable exceptions, the mt population of nuclear-encoded tRNAs was found to be primarily complementary to those encoded in the mt genome. Therefore, tRNA import in plant mitochondria represents a highly specific process as only a subset of cytosolic tRNAs appears to be routed to the mitochondria. In vivo studies in tobacco demonstrated the involvement of aminoacyl-tRNA synthetases in tRNA import (Dietrich et al. 1996). These studies also revealed the presence of different import determinants within tRNAs depending on the tRNA studied, showing the complexity and selectivity of the import process (Delage et al. 2003b; Salinas et al. 2005). The development of an in vitro tRNA import system (Delage et al. 2003a) together with biochemical approaches in potato mitochondria allowed the identification of some components of the translocation machinery of tRNAs through the mt membranes. These investigations implicated the Voltage Dependent Anion Channel (VDAC), a known player in metabolite transport, as the major component of the tRNA transport system through the outer mitochondrial membrane. Moreover, TOM20 and TOM40, two major components of the TOM (Translocase of the Outer mitochondrial Membrane) complex, are likely to be important for tRNA binding at the surface of mitochondria (Salinas et al. 2006).

B. In Vitro Import of DNA

Isolated plant mitochondria are able to take up double-stranded DNA without sequence specificity. This was first documented using a 2.3 kb linear DNA plasmid originally described in maize mitochondria (Leon et al. 1989). Additional experiments indicate the ability of mitochondria to take up DNA molecules of more than 10 kb. This uptake appeared to be an active transmembrane potential-dependent mechanism and DNA was found to be transcribed in organello after import in the mt matrix (Koulintchenko et al. 2003). Mitochondria from mammalian and yeast cells also display the ability to take up DNA, which subsequently is transcribed in organello (Koulintchenko et al. 2006; Weber-Lotfi et al. 2009). The authors concluded that the process of DNA uptake may involve VDAC and the adenine nucleotide translocator, which are core components of the mt permeability transition pore complex in animal cells (Zamzami and Kroemer 2001). At the same time the authors ruled out mt membrane permeabilization as a possible mechanism for the DNA uptake in the organelle. In a yeast system, it was shown that DNA import is inhibited by VDAC effectors. Yeast strains deleted for the VDAC-1 or VDAC-2 gene are severely reduced in mt DNA import (Weber-Lotfi et al. 2009), supporting the role of this component in DNA import into mitochondria. Understanding the mechanism that mitochondria use to import DNA in vitro may be useful for the genetic engineering of plant or animal mitochondria in vivo. This has not yet been possible, even with vectors designed for mt gene expression. If VDACs are able to import DNA into mitochondria in living cells, this would greatly increase the chances to establish mt transformation (Weber-Lotfi et al. 2009).

C. Electroporation of Isolated Mitochondria with DNA and RNA

Electroporation changes the conductance and molecular permeability of cell membranes and has been used for delivery of nucleic acids into a broad spectrum of cells (see literature in Rao et al. 2009). Delivery of small RNAs into plant mitochondria has been reported upon electroporation of protoplasts with tRNAs, which subsequently were imported into mitochondria (Wintz and Dietrich 1996). The first reports of successful electroporation of isolated organelles were published some 20 years ago. One group demonstrated the introduction of RNA into isolated chloroplasts and used this system to study the control of mRNA stability in chloroplasts by 3′ inverted repeats (Adams and Stern 1990), while another group used mitochondria from maize and rice (Mulligan et al. 1989). These attempts were not continued with plant organelles at first, but some work was done using mitochondria isolated from mice liver cells (Collombet et al. 1997). Electroporation of maize, Sorghum and wheat mitochondria was successfully established a few years later (Farré and Araya 2001; Staudinger and Kempken 2003). Recently, isolated mitochondria from Arabidopsis and cauliflower were also used for electroporation (Bolle and Kempken 2006; Kempken et al. 2009) and the details of the methodology have been published (Farré et al. 2007; Kempken et al. 2007).

The data from the mt electroporation systems differ with respect to efficiency. While in animal and wheat mitochondria, electroporation at 1.4 kV/cm (Collombet et al. 1997) and 1.3 kV/cm (Farré and Araya 2001), respectively, gave optimal results, in maize, 1.8–2.0 kV/cm appeared to be most efficient (Staudinger and Kempken 2003). In contrast, Mulligan et al. (1989) reported 8.8 kV/cm to be opti­mal. It is possible that these older data may be due to differences in the type of apparatus used, while the deviation of the newer data may be due to differences in the amount of mitochondria and plasmid DNA used.

There appears to be a size limit for the DNA to be introduced. While plasmid DNA of up to 11 kb could be introduced into mitochondria via electroporation, larger DNA molecules of about 30 kb were not (Staudinger and Kempken 2003). Moreover, while both linear and covalently closed circular (ccc) plasmid DNA can be imported into the mt matrix, the ccc DNA is mostly converted to open circular and linear DNA during the process (Collombet et al. 1997; Staudinger and Kempken 2003). Furthermore, there is no evidence for integration of the introduced DNA into the mt chromosome (Staudinger and Kempken 2003). The use of a mt plasmid from Chenopodium album for electroporation (Backert et al. 1997) gave some indication that replication of mt plasmids may be possible in isolated mitochondria (Kempken, unpublished data, see Fig. 2).

Fig. 19.2.
figure 2

Introduction of mp1 plasmid DNA (Backert et al. 1997) in maize mitochondria via electroporation. DNA was introduced in mitochondria as described in (Staudinger and Kempken 2003), and in organello incubation was carried out in the presence of α-32P-dCTP for up to 16 h. DNA was isolated after incubation and subjected to agarose gel electrophoresis. Over time, an increase in the mp1 signal is observed in the autoradiogram.

Full size image

In addition to DNA, RNA can also be transformed into mitochondria using electroporation. This was demonstrated recently for maize and cauliflower mitochondria (Hinrichsen et al. 2009). The data presented demonstrate that plant mt RNA processing appears to be independent of both transcription and respiratory regulation. Moreover, introduction of RNA offers interesting experimental strategies to investigate RNA processing events.

Electroporation has been used for transformation of isolated organelles only. However, it may be possible to employ electroporation for genetic engineering of organelles in intact cells as well. Recently, a mathematical cell model was established, which indicates the possibility to use electric fields for intracellular manipulations (Esser et al. 2010). It was indeed demonstrated that high-intensity electric pulses induce mitochondria-dependent apoptosis in mice cells (Li et al. 2008). Therefore electroporation may be useful in the future for the in vivo transformation of mitochondria. This certainly will have applications not only in plant science but also in the design of therapies for human mt disorders (Cwerman-Thibault et al. 2010).

D. In organello Analysis

1. DNA Replication

Plant and fungal mt in organello systems have been used to analyze replication (Bedinger and Walbot 1986; Kempken et al. 1989) and transcription of mt genomes (Carlson et al. 1986). One specific type of genetic elements, the so-called linear plasmids, has been of particular interest in these studies. Linear plasmids are several thousand bases in size, characterized by long terminal inverted repeats and the presence of 5′-end bound terminal proteins. They are believed to be remnants of bacteriophages with similar structures (Meinhardt et al. 1990; Kempken et al. 1992). Studies on plant mitochondria undergoing in organello replication while using radiolabeled nucleotides support the existence of DNA repair and DNA replication activities within the organelle. Linear plasmids showed much higher incorporation of radiolabeled nucleotides than the high molecular weight DNA, which might be due to the activity of DNA polymerases encoded by linear plasmids. Similar results were obtained in a fungal in organello system employing mitochondria from Ascobolus immersus. End fragments from the plasmid were found to be more heavily labeled than internal fragments, indicating a potential role of the terminal proteins in DNA replication. It is assumed that terminal proteins function as primers for the replication of linear plasmids, as is the case for certain viruses with linear plasmids (Bedinger and Walbot 1986; Kempken et al. 1989). However, the fungal system provided stronger evidence for true DNA replication versus DNA repair, as the use of strand specific probes gave evidence for protein-primed DNA replication (Kempken et al. 1989).

2. Transcription and RNA Processing

Molecular analysis of plant mt transcription and RNA processing was long hampered due to the inability to transform plant mitochondria. Alternatively, in vitro systems have ­successfully been employed (e.g. Takenaka et al. 2004; Kühn et al. 2005) and in organello systems provide another substitute for in planta experiments (e.g. Farré and Araya 2001; Staudinger and Kempken 2003).

Two processes have mainly been studied using in organello systems: RNA editing and splicing. RNA editing is a post-transcriptional process that alters the information content of RNA (see chapter “Organellar Proteomics: Close Insights into the Spatial Breakdown and Functional Dynamics of Plant Primary Metabolism”). Several functional types are known (Knoop 2010), one of which occurs mostly by conversion of specific cytidine residues to uridine residues in the RNAs of higher plant mitochondria (Covello and Gray 1989; Gualberto et al. 1989; Hiesel et al. 1989). In addition, in several plant clades such as hornworts or ferns, many U-to-C changes have also been observed to occur in mitochondria (for reviews see Shikanai 2006; Takenaka et al. 2008). RNA editing is also observed in mosses, some liverworts and in gymnosperms (Knoop 2010). The number of RNA editing sites differs, but there are up to maybe more than 500 in higher plant mitochondria (Takenaka et al. 2008; Bruhs and Kempken 2010; Knoop 2010), and even more than 1,500 in lycophyte mitochondria (Grewe et al. 2011).

In attempts to elucidate higher plant mt RNA editing and the mechanism by which the C-to-U transition is achieved, in organello systems based on mitochondria from pea seedlings, potato tubers (Yu and Schuster 1995) or maize seedlings (Rajasekhar and Mulligan 1993) were developed. Taken together, the studies have excluded nucleotide excision and base exchange as possible editing reactions, while deamination or transamination reactions both are possible mechanisms. However, so far no conclusive evidence for either reaction has been obtained (Takenaka et al. 2008).

In organello assays in combination with uptake of DNA via electroporation have been successfully employed to analyze RNA processing (Farré et al. 2001; Staudinger and Kempken 2003). Using biotinylated Uridine Tri-Phosphate (UTP), de novo transcription and RNA processing were observed in organello. The in organello systems were shown to faithfully reflect the in planta situation. Transcripts that are fully edited in planta are fully edited in organello, while those partially edited in planta are also partially edited in the in organello system (Staudinger and Kempken 2003, 2004). Using a wheat in organello system, a core sequence including 16 nucleotides upstream and six nucleotides downstream of the edited nucleotide was defined for recognition of two editing sites (Farré et al. 2001; Choury et al. 2004). However, for efficient editing, an upstream sequence of 40 nucleotides appeared to be required as shown by in vitro experiments (Takenaka et al. 2004). While it is now established that pentatricopeptide proteins (see chapter “Organellar Proteomics: Close Insights into the Spatial Breakdown and Functional Dynamics of Plant Primary Metabolism”) participate in the recognition of mt RNA editing sites (e.g. Verbitskiy et al. 2009; Zehrmann et al. 2009), experiments using cauliflower and maize in organello systems also suggest an influence of a RNA secondary or tertiary structure (Bolle and Kempken 2006).

The way splicing and editing function in cross-species analyses is not fully predictable. While the cox2 RNA is spliced and edited in mono- and di-cotyledonous mitochondria regardless of the mono- or dicot origin of the gene (Staudinger and Kempken 2003), the rps10 RNA from potato (dicot) is neither spliced nor edited in wheat (monocot) mitochondria (Choury et al. 2005). The atp6 mRNA from sorghum (monocot) is not even edited in maize (monocot) mitochondria despite very high sequence similarity. However, a chimeric atp6 transcript composed of maize and sorghum DNA sequences gave rise to partial RNA editing, which may be due to presence of a RNA editing recognition sequence in the atp6 transcript (Staudinger and Kempken 2003). An in organello system has also demonstrated the inability of plant mitochondria to recognize chloroplast editing sites (Bolle et al. 2007).

RNA editing can be essential for subsequent RNA splicing (Farré and Araya 2002; Castandet et al. 2010). However, both RNA splicing and editing are independent of the transcription machinery, as demonstrated using the cox2 mRNA electroporated into mitochondria (Hinrichsen et al. 2009).

3. Translation

In organello translation of proteins employing 35S labeled amino acids has long been established (Grohmann 1995). It has been employed for analysis of cytoplasmic male sterility (e.g. Horn et al. 1991; Moneger et al. 1994) and for the analysis of translation of partially edited transcripts. Amino acid sequencing of immunoprecipitated NAD9 protein indicated that only proteins from fully edited nad9 mRNAs accumulate in organello (Grohmann et al. 1994).

While transcription and RNA processing can occur from DNA electroporated into isolated mitochondria and subsequent in organello incubation, there is no experimental evidence supporting that introduced RNA can be translated. In animal mitochondria, RNAs from a mt reporter gene (modified luciferase) could be stably maintained in the presence of functioning mitochondrial protein synthesis. However, the RNAs were not translated (McGregor et al. 2001). Likewise there is no proof for translation of RNA transcribed from introduced DNA into higher plant mitochondria (Kempken, unpublished data). As the introduced mRNA is faithfully processed (Hinrichsen et al. 2009), the reason for why translation cannot occur remains unclear.

III. Conclusion

Mt transformation of the unicellular alga C. reinhardtii can be achieved with surprisingly high efficiency and is of great value for the analysis of mt gene functions. Unfortunately comparable attempts to transform vascular plant mitochondria have been unsuccessful so far. While plant mt transformation has not yet been achieved, mt in organello systems and the ability of mitochondria to uptake DNA provide an attractive alternative to study transcription and RNA processing mechanisms, such as RNA splicing and RNA editing. However, although in organello translation has long been established, the translation of RNA either directly introduced into isolated mitochondria, or transcribed from DNA electroporated into mitochondria, has not been reported yet. It is conceivable that a close coupling between transcription and translation as well as the proximity of the mt membrane could be necessary to synthesize polypeptides within mitochondria, but other parameters could be important, too. In addition, there could be a connection between the failure to translate imported nucleic acids within mitochondria and the inability to transform plant mitochondria. Therefore, understanding what might be the reason(s) for this lack of translation represents a major challenge that will provide clues to successful transformation of plant mitochondria.