CODON USAGE AND BASE COMPOSITION IN RICKETTSIA-PROWAZEKII

Codon usage and base composition in sequences from the A + T-rich geno me of Rickettsia prowazekii, a member of the alpha Proteobacteria, hav e been investigated. Synonymous codon usage patterns are roughly simil ar among genes, even though the data set includes genes expected to be expressed at very different levels, indicating that translational sel ection has been ineffective in this species. However, multivariate sta tistical analysis differentiates genes according to their G + C conten ts at the first two codon positions. To study this variation, we have compared the amino acid composition patterns of 21 R. prowazekii prote ins with that of a homologous set of proteins from Escherichia coil. T he analysis shows that individual genes have been affected by biased m utation rates to very different extents: genes encoding proteins highl y conserved among other species being the least affected. Overall, pro tein coding and intergenic spacer regions have G + C content values of 32.5% and 21.4%, respectively. Extrapolation from these values sugges ts that P. prowazekii has around 800 genes and that 60-70% of the geno me may be coding.