Some of the most intriguing examples of adaptation in eukaryotes arise within the plant kingdom, many in response to a plant's immotility and consequent inability to escape environmental stresses (Hawkesford and Buchner, 2001). These unique attributes occur in various forms to produce wonders of plant architecture, specialized physiology and reproductive strategies. At a cellular level, several unique features of plant metabolism and organellar genome maintenance are evident in plants (Mackenzie and Mcintosh, 1999). Some of these cellular attributes are thought to be the outcome of the endosymbiotic process that has led to the present-day plastid and consequent mito-chondrial-plastid co-evolution (Allen, 1993; Adams et al., 2002; Elo et al, 2003).
As in the case of most animal systems, organellar genomes generally show strict maternal inheritance in plants. However, there are exceptions to this pattern. In some cases, paternal inheritance is observed, though varying degrees of biparental inheritance are also seen (Reboud and Zeyl, 1994; Zhang et al., 2003). In nearly all such exceptions, the plastid has been more likely to show variation from strict maternal inheritance than the mitochondrion. Why the relaxation of strict maternal inheritance pat terns might be tolerated more in plant systems than in animal, and the mechanisms underlying these selective organelle transmission patterns, are not yet well understood. Mechanisms for selective exclusion of paternal organelles vary. Whereas some systems appear to target paternally derived organellar DNA for selective destruction or suppressed replication (Nagata et al, 1999; Sodmergen et al, 2002; Moriyama and Kawano, 2003), some animal systems are postulated to exclude or destroy the paternal organelles themselves (Sutovsky, 2003).
When considering organelle inheritance and segregation processes, one must keep in mind the distinct dynamics of organelle behaviour. In contrast to nuclear genetic information, which undergoes replication at a precise point within a tightly regulated cell cycle, segregating to daughter cells in equal and unchanging copy number each cellular generation, organellar genomes obey very different rules. Organellar DNA replication does not maintain tight synchrony with the cell cycle (Birky, 1983), and the numbers of genomes per organelle and organelles per cell vary dramatically depending on tissue type. In general, mitochondrial biogenesis is highest in meristematic and reproductive tissues, where numbers of genomes per mitochondrion and mitochondria per cell generally range in the hundreds, while
© CAB International 2005. Plant Diversity and Evolution: Genotypic and Phenotypic Variation in Higher Plants (ed. R.J. Henry)
mitochondrial numbers decrease markedly in vegetative tissues (Lamppa and Bendich, 1984; Fujie et al, 1993; Robertson et al, 1995). The 'segregation' of organelles, existing as multi-genomic populations within cells, occurs by a process of cytoplasmic sorting throughout development. Although cytoplasmic sorting is generally considered a stochastic process, nuclear gene influence on cytoplasmic segregation is evident.
This review will describe the unusual nature of plant mitochondrial genomes, contrasting their features and behaviour with what is known of mammalian and fungal systems. One anticipates that the considerable divergence observed in plants derives, at least in part, from the unusual plant cellular context of mitochondrial-chloroplast co-evolution. With availability of complete plant genome sequence information, considerable evidence has accumulated recently in support of this assumption. However, it is also important to note that the vast majority of genes essential for mitochondrial processes are nuclear encoded; the mitochondrial genome, though essential, encodes less than 5% of the information required for its varied functions. Therefore, it is impossible to consider plant mitochondrial genome evolution and adaptation without addressing the critical role of the nucleus in these ongoing processes.
Nuclear Regulation of Mitochondrial DNA and RNA Metabolism
The availability of complete genome sequences for Arabidopsis and rice has allowed the identification of several candidate genes predicted to function in organellar DNA and RNA metabolism functions. Two striking features of nuclear genes that appear to direct organellar genome maintenance are evident. The first surprising property of these genes is their organization within the nuclear genome. Nuclear genes encoding organellar DNA and RNA metabolism loci appear to be largely clustered in a few regions of the plant genome (Elo et al., 2003). Moreover, this genomic arrangement may not be limited to plants alone (Lefai et al., 2000).
The endosymbiotic processes believed to have given rise to present day organelles are generally thought to have involved the transfer of large amounts of genetic information from mitochondrial and plastid progenitor genomes to the nucleus. Very early on, these transfers might have involved large genomic segments that encompassed many genes simultaneously. More recent gene transfers, following the advent of RNA editing processes, have probably occurred as singular gene events that involve an RNA intermediate in the process (Nugent and Palmer, 1991; Covello and Gray, 1992). With the massive nuclear genomic rearrangements that have occurred in plants subsequent to the endosymbiotic events (Blanc et al., 2000), it is difficult to envisage how genetic linkage has been maintained for large numbers of transferred genes without selection. One possibility is that maintenance of related genes in a linkage might facilitate coordinate gene regulation during key points in development (Boutanaev et al., 2002). For example, at the point immediately following pollination, a maternally derived cytoplasm must immediately establish compatibility with the newly introduced paternal nuclear contribution. Possibly epigenetic regulation would be crucial for re-establishing necessary interge-nomic compatibility.
A second intriguing observation regarding the nuclear genes that participate in organellar genome maintenance is the large number predicted to encode proteins functional in both mitochondria and plastids (Hedtke et al., 1997; Beardslee et al, 2002; Elo et al., 2003). This assumption is supported by individual genes that encode dual-targeted proteins, as well as genes that have undergone duplication, with duplicate members targeting distinct organelle types. The prevalence of proteins that apparently function in both mitochondria and plastids suggests that a substantial component of the DNA and RNA metabolism apparatus is overlapping in the two organelle types (Small et al., 1998; Peeters and Small, 2001). At one time this would have seemed an incongruous idea, given the numerous differences that exist in mitochondrial and plastid functions, genome structure and gene organization. These distinctions are likely to be remnant features of their progenitors, as the number of reports of shared, probably acquired, features increases.
Plant mitochondrial genomes are distinguished by their extreme variation in size, ranging from about 200 to 2400 kbp, in comparison to the 16-kbp genome of human mitochondria and intermediate but less variably sized genomes of fungi. With the recent availability of complete mitochondrial genome sequences for at least four plant species, we now find that the dramatic differences in genome size are not accounted for by vast discrepancies in coding capacity. In fact, plant mitochondrial genomes encode somewhere between 55 and 70 genes (Oda et al., 1992; Unseld et al, 1997; Notsu et al, 2002; Handa, 2003), less than twice the number of genes found in the human mitochondrion. Considerable sequence redundancy, integration of non-mitochondrial DNA, and ectopic recombination have contributed to the observed variation.
The mitochondrial genome of plants consists of a heterogeneous population of both circular and linear DNA molecules, many existing in highly branched configurations (Backert et al, 1996; Bendich, 1996; Oldenburg and Bendich, 1996; Backert and Borner, 2000). To date, an origin of replication has not been defined in plants, and evidence suggests that replication may occur, at least in part, by a rolling circle mechanism. In fact, it has been suggested that replication may initiate by a strand invasion process perhaps resembling that of T4 phage (Backert and Borner, 2000).
In contrast to mammalian mitochondria, plant mitochondrial genomes are unusually dynamic in their structure, in part a consequence of prolific intra- and intergenomic recombination activity. Within most plant mitochondrial genomes are dispersed several repeated sequences, present in both inverted and direct orientations. High-
frequency DNA exchange at these sites produces a complex assemblage of large inversions and subgenomic DNA molecules, each containing a portion of the genome. Whether this recombination activity is continuous throughout plant development, or restricted to a particular cell type, is not clear. The technical difficulties that have complicated these studies arise from the entwined nature of replicative and recombi-national processes.
In addition to high-frequency homologous recombination, plant mitochondrial genomes commonly undergo low-frequency recombinations at non-homologous, often intragenic, sites. This ectopic recombination activity gives rise to chimeric gene configurations, often expressive, within a wide array of plant species. In many cases, these unusual gene chimeras are discovered by their causative association with cytoplasmic male sterility (Schnable and Wise, 1998). However, not all ectopic recombinations necessarily produce a detectable phenotype (Marienfeld et al, 1997).
Cytoplasmic male sterility (CMS) is a condition in which the plant is unable to produce or shed viable pollen as a consequence of mitochondrial mutation. In nearly all cases investigated, the associated mitochon-drial mutations are dominant, stemming from expression of unique sequence chimeras that form aberrant open reading frames. To date, all cases of CMS appear to be associated with ectopic recombination within the genome, implying an adaptive advantage to this activity. No two CMS mutations have been identical, although striking similarities have been observed in some cases (Tang et al, 1996). Interestingly, the frequency of non-homologous recombination, or the relative copy number of the derived recombinant forms, appears to be controlled by nuclear genes.
Nuclear Regulation of Mitochondrial Genome Structure
Whereas evolutionary pressures appear to have selected for a highly conservative, stable and compact mitochondrial genome configuration within animals, adaptive selection within the plant kingdom has produced a system that appears to benefit from a high degree of variability (Marienfeld et al, 1999). A fascinating example of the dynamic nature of the plant mitochondrial genome is the widespread phenomenon termed stoichiometric shifting (Small et al, 1987).
Certain subgenomic mitochondrial DNA configurations change dramatically in relative copy number during the development of the plant. When present substoichiomet-rically, these components of the mitochon-drial genome have been estimated at one copy per every 100-200 cells of the plant (Arrieta-Montiel et al, 2001), representing a heteroplasmic (heterogeneous cytoplasmic) condition. However, stoichiometric shifting can result in preferential amplification of these molecules to levels equimolar with the principal mitochondrial genome. The selective amplification or suppression of particular portions of the mitochondrial genome is influenced by nuclear genotype.
Stoichiometric shifting has been reported in a wide array of plant species (Mackenzie and Mcintosh, 1999), and shifting events are apparently induced under conditions of cell culture or cybridization (Kanazawa et al, 1994; Gutierres et al, 1997; Bellaoui et al., 1998), alloplasmy (Kaul, 1988), spontaneous CMS reversion to pollen fertility (Mackenzie et al, 1988; Smith and Chowdhury, 1991), and the introduction or mutation of specific nuclear genes.
in the CMS system of common bean, the mitochondrial sequence associated with pollen sterility, designated pvs-orf239, resides on a mitochondrial molecule that is shifted to substoichiometric levels in response to the introduction of the dominant nuclear gene Fr (Mackenzie and Chase, 1990). This is an appealing system for study because introduction of Fr, by standard crossing, results in Mendelian segregation for a particular, reproducible mitochondrial rearrangement. The pvs-orf239 sequence, when present in high copy number, is expressed and the plant is male sterile. When Fr is introduced, and the mitochondrial pvs-orf239 sequence is reduced to substoichiometric levels, the plant is male fertile. Interestingly, the condition of male fertility is not reversed by the segregation of Fr, suggesting that the product of Fr acts unidirectionally or in a limited context, and the reversal of Fr action might require additional nuclear components.
In Arabidopsis, mutation at the nuclear locus CHM results in the copy number amplification of a mitochondrial chimeric DNA configuration and the appearance of green-white variegation (Martinez-Zapater et al., 1992; Sakamoto et al., 1996). Several mutant alleles of CHM are available in Arabidopsis, presenting an opportunity to clone the gene.
Recently, the product of the CHM locus was shown to resemble the MUTS protein of Escherichia coli, and the gene has now been designated MSH1 (MutS Homologue 1; Abdelnoor et al., 2003). MutS is a component of the DNA mismatch repair apparatus, and several of its homologues have been identified to function within the nuclear genome of higher eukaryotes. One other nuclear gene encoding a mitochondrial MutS homologue was reported several years ago in yeast (Reenan and Kolodner, 1992). In that case, the mitochondrial protein was suggested to function in mismatch repair (Chi and Kolodner, 1994).
Mismatch repair components appear to serve two important functions within the eukaryotic genome: to bind and repair nucleotide mismatches and to suppress non-homologous recombination activity (Modrich and Lahue, 1996; Harfe and Jinks-Robertson, 2000). Enhanced ectopic recombination appears to be the effect within the plant mitochondrial genome in response to MSH1 mutation (Abdelnoor et al., 2003).
Interestingly, allelic variation for MutS in microbial populations appears to provide certain adaptive advantage under severe selection conditions by enhancing mutation frequency (LeClerc et al., 1996; LeClerc and Cebula, 1997; Bjedov et al., 2003; Chopra et al., 2003). This 'mutator' phenomenon, arising from MutS variation, has been reported in a number of organisms including humans, where such variation contributes to cancer incidence (Li, 1999). A phenomenon resembling mitochondrial stoichiometric shifting has also been described in Drosophila, although the nuclear effector is not yet identified (Le Goff et al, 2002). It now appears likely that evolutionary advantage might have been realized by permitting a degree of genetic variation within the mitochondrial MSH1 gene of higher plants. A possible adaptive role of MSH1 variation in plant populations will be discussed in a later section.
Nuclear Regulation of Mitochondrial Transcript Processing and Fertility Restoration
A striking example of a derived plant feature shared by both mitochondria and plas-tids is their dependence on RNA processing, editing and stabilizing functions for organellar gene expression (Hoffmann et al., 2001; Binder and Brennicke, 2003). RNA processing, which involves the cleavage of RNA at precise sites as part of RNA maturation, and RNA editing, which generally involves specific C to U conversions within a transcript, are both found to occur in a wide array of plastid and mitochondrial RNAs. The expansion of these processes has apparently been accompanied, or driven by, a concomitant expansion in the number of nuclear genes associated with these functions.
The pentatricopeptide repeat (PPR) family of proteins in Arabidopsis numbers over 500 members, with over two-thirds encoding proteins predicted to target mitochondria or plastids (Small and Peeters, 2000). The PPR proteins share almost no detectable sequence homology. Rather, they are linked by their unusual structural similarities. Although highly divergent at their amino termini, each PPR protein contains a series of 35-amino-acid repeat structures, present in variable numbers. These repeats are predicted to confer a helical structure to the protein that is postulated to interact directly with RNA or proteins. It has been suggested that this family of nuclear proteins may provide the RNA recognition specificity necessary for RNA processing activities.
Interestingly, molecular studies of fertility restoration mechanisms in several CMS systems reveal a role of most restorer genes in RNA processing (Schnable and Wise, 1998). Over the past few years, five nuclear genes that restore pollen fertility to CMS mutants have been cloned. Of these, four have been shown to encode PPR proteins. These include the restorers of fertility in petunia (Bentolila et al., 2002), Kosena radish (Koizuka et al., 2003), Ogura radish (Brown et al., 2003; Desloire et al., 2003) and rice (Kazama and Toriyama, 2003). It would not be surprising to find PPR proteins implicated in the fertility restoration of several other CMS plant species in the near future. The RNA specificity that is postulated by PPR protein:RNA binding, combined with the intragenic recombination origins of most CMS mutations, appears an ideal system for nuclear control of CMS-associated aberrant gene expression.
Evolutionary Implications of Mitochondrial Genome Dynamics in Higher Plants
The mitochondrial mutations that confer cytoplasmic male sterility have been of great interest for their value to the hybrid seed industry (Fig. 5.1). In a broad range of plant species, the phenomenon of heterosis, or hybrid vigour, is well documented (Tsaftaris and Kafka, 1998; Rieseberg et al., 2000). The genomic condition produced by hybridization, probably associated with higher levels of gene heterozygosity and perhaps epige-netic processes, provides markedly enhanced reproductive capacity and plant vigour. Although of obvious agricultural benefit, the heterotic state is also clearly advantageous in natural populations (Rieseberg et al, 2000).
Most domesticated crop species are categorized as predominantly outcrossing or self-pollinating. However, in natural populations exist the more heterogeneous and dynamic conditions of gynodioecy. Gynodioecious populations are composed of both female and hermaphroditic individuals, permitting population expansion within geographically isolated environments
Was this article helpful?