Insights into the Genome of Cyanobacteria and Plastids: tRNA Gene Content and Codon Usage

As described by the endosymbiotic theory, plastids originated from a process involving the engulfment of a cyanobacteria-like cell by a eukaryotic ancestor. Over time, plastid genomes or “plastomes” underwent a drastic reduction, which led to the conservation of a minimal expression apparatus and a set of specific coding sequences. It has been estimated that more than 95% of the plastid proteome would be expressed by the nuclear genome. On contrary, no evidence of import of rRNA and tRNA molecules into plastids have been reported so far. Since a number of plastid genes encode for highly expressed proteins, their sequences could have been evolved to be efficiently translated in the plastids. Assuming that translationally optimal codons are related to the most abundant tRNA isoacceptors and that the abundance of each tRNA can be inferred by its gene copy number, a relationship should exist between codon usage and tRNA gene content. In this PhD thesis, the tRNA gene content and its correlation with the codon usage in protein-coding genes were studied in about 600 plastid genomes and 80 cyanobacteria genomes. In order to analyse the several thousands of GenBank annotations, a set of suitable computer programs called “tRNA tools” was developed and freely distributed. Firstly, despite the reduced number of tRNA genes, most plastid genomes can effectively translate all the codons corresponding to the 20 standard amino acids by using both the wobble and the extended wobble rules (superwobble and “two out of three”). However, a few plastome sequences do not have enough tRNAs to decode all the standard amino acids. A “defective” set of tRNA genes may be ascribed to a loss of plastid functionality or, alternatively, to the existence of tRNA import into plastids. Data on the tRNA gene content were used to evaluate distinctive traits among different groups of eukaryotic organisms, not only among the main kingdoms of life as shown in previous studies. Noteworthy, a set of 23 tRNA genes was recognised as universally shared by most of the plastid genomes, thus extending the results obtained from in vivo experiments made on Nicotiana tabacum to a wide range of organisms. In addition, tRNA(ACG)-Arg was found to be the only tRNA available to decode Arginine codons in Alveolata, Excavata, and Streptophyta, thus confirming the essentiality of the “two out of three” rule. Besides, a number of plastome sequences lack the genes coding for tRNA-Glu which is involved in the pyrrole biosynthesis. Although annotation errors cannot be excluded, the essentiality of tRNA-Glu in plastids could again suggest the occurrence of tRNA import mechanisms. Finally, the hypothesis that synonymous codons related to the most abundant tRNA isoacceptors could have been preferentially conserved during evolution was tested by estimating the correlation between codon usage and tRNA gene copy number. At a global genome-level, the results obtained in this thesis showed that the tRNA gene copy number and the codon usage generally do not correlate when only the standard pairings are taken into account. The results drastically change when the wobbling and superwobbling mechanisms are considered, showing from weak to moderate correlation among all groups. Conversely, plastomes with a reduced tRNA gene set such as Rhizanthella gardneri and Selaginella moellendorfii do not show significant correlation even when considering the superwobble rules. The correlation was further analysed also at singe-gene level, in order to have a deeper insight into the genomes of each group of plastids. As occurred at global genome level, none of the genes analysed in the plastomes of Cyanobacteria, Rhizaria, Glaucophyta, and Streptophyta, showed significant correlation between codon usage and tRNA gene content considering only the standard pairings. In the other groups, the genes that obtained a significant correlation encode subunits of photosystem I, photosystem II, RuBisCO enzyme, ATPase, or Ribosome. Among all the groups of plastids, the psbA gene resulted to have the most optimized codon composition in term of correlation with the tRNA gene content by considering only the standard pairing rules, which are supposed to be the most efficient for translating the codons. Conversely, genes with a lower rate of expression seem to be not optimized in this sense and rely more on the wobble and superwobble rules to be effectively translated.