Codon usage variable and properties of wild-type and recoded F9 sequences. (A) The underlying primary codon usage variables, namely CAI and similarity to wild-type sequence employed in the design of recoded F9 constructs. Constructs were designed to have either “high” (1A, 1B, and 1C) or “low” (2A, 2B, and 2C) CAI and “close” (1B and 2B), “medium” (1A and 2A), or “distant” (1C and 2C) similarity to wild-type sequence. Percentages of nucleotides and codons altered in recoded sequences and in silico–predicted minimum free energies (MFE) of resulting mRNA structures were also shown. (B) A biplot of codons of wild-type and recoded F9 sequences showing the segregation of sequences based on codon adaptation indices of constructs. (C) Antigen and mRNA expression data and activity levels of wild-type and recoded F9 sequences in transient transfection experiments. Recoded constructs demonstrated higher levels of F9 mRNA and FIX antigen expression. Data are represented as mean ± standard deviation. WT, wild-type.