An active homologous recombination pathway in plastids must underlie maintenance of the plastid genome and shape its evolution. Plastid transformation experiments have demonstrated recombination events between short repeats of ~200 bp in length in flowering plant plastids resulting in the deletion of genes (Table 1). Such excision events would be deleterious to WT plastid genomes. Grossly deleted N. tabacum transgenic plastid genomes resulting from recombination between distant psbA 3'UTR sequences do not persist (Svab and Maliga 1993) presumably because of strong selection against dysfunctional plastids with defective plastid genomes. Deleterious recombination events in WT plastid genomes would be avoided if they lacked DNA repeats or contained small repeats that were below the size needed for homologous recombination. One set of inverted repeats would be allowed because they would only flip the relative orientation of the single copy regions (Fig. 7). Any repeats that lie on either side of the large inverted repeat would be converted into direct repeats by large inverted repeat mediated flipping of single copy regions and destabilise plastid genomes (Day and Ellis 1984). Loss of the large inverted repeat in P. sativum is associated with rearrangements in plastid gene order (Palmer and Thompson 1982); presumably because inversions are not restricted by the presence of a large inverted repeat (Day and Ellis 1984).
Whilst most angiosperm plastid genomes contain a large inverted repeat of 2076 kbp they are deficient in repeated DNA sequences over 100 bp in size. The plastid genome of Pelargonium x hortorum contains a large number of repeated sequences, including nine pairs of dispersed repeats of 31-101 bp in size
(Chumley et al. 2006). Over 20% of the C. reinhardtii plastid genome is comprised of repetitive DNA of less than 50 bp in length (Maul et al. 2002). Presumably these repeated sequences of 100 bp or less are too small to act as efficient substrates for homologous recombination. Plastid DNA is uniform within a plant and recombination events that would destabilize and fragment plastid genomes are not normally observed. However, low frequency recombination events between short repeats might give rise to the plastid DNA rearrangements observed during evolution (Hiratsuka et al. 1989; Chumley et al. 2006). Deleted plastid genomes with circular maps (Day and Ellis 1984; Cuzzoni et al. 1995) and circular topologies (Day and Ellis 1985) have been described in albino cereal plants from anther culture demonstrating the instability of the plastid genome when cells are rescued by heterotrophic growth in vitro. A recent report suggests albino Bambusa edulis (bamboo) plants contain deleted plastid genomes (Liu et al. 2007). Aberrant sub-genomic circles have also been described in WT N. tabacum chloroplasts by DNA fibre-based FISH analysis (Lilly et al. 2000). Two factors appear to be required to maintain plastid genomes, first, DNA-RRR surveillance mechanisms that either reduce the frequency of deleterious recombination events or repair deleted plastid genomes, and second, selection for functional plastids with an intact plastid genome.
Was this article helpful?