Genuine linear plastid DNA molecules can be distinguished from linear products of broken DNA circles by studying their ends. Breakage products would be expected to possess ends that map to randomly selected regions of the plastid genome, and these ends would be expected to be indistinguishable from doublestrand breaks with flush or short single-stranded 5' or 3' DNA extensions. Analysis of plastid DNA deletion mutants in albino cereal plants regenerated from pollen provided the first evidence for the presence of linear plastid DNA molecules with special ends (Day and Ellis 1985; Ellis and Day 1986). Small linear sub-genomic molecules have also been found in albino somatic cells from cereals (Kawata et al. 1997; Zubko and Day 2002) and can represent the predominant plastid DNA species in albino cereal plants. Their abundance facilitates the analysis of their ends, which have been examined in detail.
Small linear plastid DNA molecules are inverted repeat palindromes with hairpin ends, which map to a number of sites in the large single copy region of plastid DNA near the trnE(UUC) gene (Ellis and Day 1986; Harada et al. 1992; Kawata et al. 1997). The centres of a subset of these linear palindromes are located between trnG(GCC) and trnfM(CAU) of the 135 to 140 kbp cereal plastid genome (Ogihara et al. 2000) and retain only 5.2 kbp of plastid DNA (Zubko and Day 2002). Small linear DNA molecules all contain the plastid trnE(UUC) gene, which is probably essential for heme synthesis (Howe and Smith 1991; Zubko and Day 2002). Linear hairpin DNA molecules are also found in eubacteria including the spirochete genus Borrelia (Casjens 1999) and prophage N15 of Escherichia coli (Rybchin and Svarchevsky 1999). Models for the origin of these linear DNA molecules include: strand switching during DNA replication (Ellis and Day 1986), possibly promoted by short inverted repeats (Fig. 4a); repair of double strand DNA breaks by intra-strand annealing at inverted repeats (Fig. 4b; Qin and Cohen 2000); and an E. coli linear prophage N15-like mechanism involving two cleavages, sealing DNA ends to form hairpins and resolution of replicated DNA into a linear palindrome (Fig. 4c; Rybchin and Svarchevsky 1999). Hairpin ends provide a mechanism to overcome the end-replication-problem and stabilise the ends of linear DNA molecules by protecting them from nucleases (Cavalier-Smith 1974).
Studies on albino cereal plants demonstrate that plastids contain the enzymes required to maintain and replicate linear DNA molecules. If hairpin molecules play a role in maintenance of intact plastid DNA they should also be found in the green chloroplasts of WT plants. Revealingly, hairpin molecules are found in WT H. vulgare chloroplasts. The hairpin ends do not appear to be localised but map to various sites within the plastid genome (Collin and Ellis 1991), which is consistent with their derivation from a population of linear DNA molecules with heterogeneous ends. Dispersed ends that are not defined in location are also found in the heterogeneous populations of linear DNA molecules in plant (Backert and Borner 2000; Oldenburg and Bendich 2001) and Saccharomyces cerevisiae (bakers' yeast) mitochondria. Mitochondrial DNA in S. cerevisiae is comprised of a polydisperse population of linear DNA molecules ranging in size between the 75 kb monomer and 150 kbp dimer (Williamson 2002).
Was this article helpful?