Proteogenomics analysis of CUG codon translation in the human pathogen Candida albicans

Abstract

Background Yeasts of the CTG-clade lineage, which includes the human-infecting Candida albicans, Candida parapsilosis and Candida tropicalis species, are characterized by an altered genetic code. Instead of translating CUG codons as leucine, as happens in most eukaryotes, these yeasts, whose ancestors are thought to have lost the relevant leucine-tRNA gene, translate CUG codons as serine using a serine-tRNA with a mutated anticodon, tRNASerCAG . Previously reported experiments have suggested that 3–5% of the CTG-clade CUG codons are mistranslated as leucine due to mischarging of the tRNASerCAG . The mistranslation was suggested to result in variable surface proteins explaining fast host adaptation and pathogenicity.

Results In this study, we reassess this potential mistranslation by high-resolution mass spectrometry-based proteogenomics of multiple CTG-clade yeasts, including various C. albicans strains, isolated from colonized and from infected human body sites, and C. albicans grown in yeast and hyphal forms. Our data do not support a bias towards CUG codon mistranslation as leucine. Instead, our data suggest that (i) CUG codons are mistranslated at a frequency corresponding to the normal extent of ribosomal mistranslation with no preference for specific amino acids, (ii) CUG codons are as unambiguous (or ambiguous) as the related CUU leucine and UCC serine codons, (iii) tRNA anticodon loop variation across the CTG-clade yeasts does not result in any difference of the mistranslation level, and (iv) CUG codon unambiguity is independent of C. albicans’ strain pathogenicity or growth form.

Conclusions Our findings imply that C. albicans does not decode CUG ambiguously. This suggests that the proposed misleucylation of the tRNASerCAG might be as prevalent as every other misacylation or mistranslation event and, if at all, be just one of many reasons causing phenotypic diversity.