Codon usage in Entamoeba histolytica, E. dispar and E. invadens

We analyzed the frequencies of genetic codon usage of 68 non-redundant protein coding genes from the human-pathogenic E. histolytica (28 117 codons), 6 from the non-pathogenic E. dispar (1744 codons), and 4 from the reptilian E. invadens (933 codons). The A + U contents of the protein coding sequences from E. histolytica, E. dispar, and E. invadens were 67%, 66%, and 58%, respectively. The nucleotide frequency in the third position was strongly biased toward A + U in E. histolytica and E. dispar (85% and 82%, respectively); the degree of the A + U bias was higher in the third position than that in the first or second position. In contrast, the nucleotide frequency in the third position was less biased in E. invadens (60% A + U) than in E. histolytica and E. dispar. Codon usage was biased in accordance with the A + U preference in the third position in E. histolytica and E. dispar. However, no apparent difference in the codon usage was found between E. histolytica and E. dispar. The codon usage in E. invadens was found less biased; the nucleotide biases observed in the third position of the synonymous codons for several amino acids including leucine, tyrosine, cysteine, and histidine of the E. histolytica and E. dispar genes were reversed or absent. The codon usage in Entamoeba species significantly differed from that in other amitochondrial protist, Giardia lamblia and Trichomonas vaginalis. Two sequences encoding ribosomal protein S10 and S27 showed significantly smaller codon biases than the rest of E. histolytica sequences, suggesting that these ribosomal proteins might be under specific functional constraint of codon usage. The differences in the A + U content of the coding sequences and in the codon usage between the mammalian E. histolytica and E. dispar and the reptilian E. invadens suggested that the reptilian Entamoeba species were distantly related to the mammalian species. These results may aid in elucidating pressures that facilitate changes in the patterns of the genetic codon usage.