FX174 Attenuation by Whole Genome Codon Deoptimization

James T. Van Leuven; Martina M. Ederer; Katelyn Burleigh; LuAnn Scott; Randall A. Hughes; Vlad Codrea; Andrew D. Ellington; Holly Wichman; Craig Miller

We replaced ΦX174 codons with less preferred codons without consideration for how alterations might affect other features in the genome. As mentioned in the introduction, many such features may be under selection. To investigate unintended consequences of codon deoptimization, we calculated numerous genome characteristics to see if any correlate with the fitness decreases observed in deoptimized fragments (supplementary table S1). We included many different measures of codon usage bias (CAI, tAI, ITE, etc), codon pair bias (CPB), frequency of Shine-Dalgarno motifs, mRNA folding stability, as well as simply the number of changes made. The best predictor of fitness is the folding stability of the codon deoptimized mRNA (R 2 =0.67, p.adj=0.02), which performed significantly better than the best measure of codon usage bias which was CAI using the Xia2007 method ( Δ AIC=15.5). This correlation is easily observed when mRNA stability values are plotted against change in fitness ( fig. 6b ). We were interested to see if the correlation between genomic features like CAI and fitness held up even when features were optimized, so we replaced ΦX174 codons with codons frequently used in E. coli (supplementary table S2), expecting to observe beneficial fitness effects.