Complete sequence of the maize chloroplast genome: gene content, hotspots of divergence and fine tuning of genetic information by transcript editing

The nucleotide sequence of the chloroplast (cp) DNA from maize (Zea mays) has been completed. The circular double-stranded DNA, which consists of 140,387 base-pairs, contains a pair of inverted repeat regions (IRA and IRB) with 22,748 base-pairs each, which are separated by a small and a large single copy region (SSC and LSC) of 12,536 and 82,355 base-pairs, respectively. The gene content and the relative positions of a total of 104 genes (70 peptide-encoding genes, 30 tRNA genes and four rRNA genes) are identical with the chloroplast DNA of the closely related species rice (Oryza sativa). A detailed analysis of the two graminean plastomes allows the identification of hotspots of divergence which predominate in one region containing a cluster of tRNA genes and in two regions containing degenerated reading frames. One of these length differences is thought to reflect a gene transfer event from the plastome to the nucleus, which is followed by progressive degradation of the respective chloroplast gene resulting in gene fragments. The other divergent plastome region seems to be due to the complete loss of a plastid gene and its functional substitution by a nuclear encoded eukaryotic homologue. The rate of neutral nucleotide substitutions is significantly reduced for protein coding genes located in the inverted repeat regions. This indicates that the existence of inverted repeat regions confers increased genetic stability of the genes positioned in these regions as compared to genes located in the two single copy regions. Editing events cause the primary structures of several transcripts to deviate from the corresponding genomic sequences by C to U transitions. The unambiguous deduction of amino acid sequences from the nucleotide sequences of the corresponding genes is, therefore, not possible. A survey of the 25 editing positions identified in 13 different transcripts of the maize plastome shows that representatives of all protein coding gene classes are subject to editing. A strong bias exists for the second codon position and for certain codon transitions. Based on the number and the codon transition types, and taking into account the frequency of putative editing sites in all peptide encoding genes and unidentified reading frames, a total number of only few more than the experimentally verified 25 editing sites encoded in the maize plastome is estimated. This corresponds to 0.13% of amino acid positions which cannot be derived from the corresponding codons present in the corresponding genes.