Synonymous Codon Usage Bias in the Chloroplast Genomes of 13 Oil-Tea Camellia Samples from South China

Synonymous codon usage (SCU) bias in oil-tea camellia cpDNAs was determined by examining 13 South Chinese oil-tea camellia samples and performing bioinformatics analysis using GenBank sequence information, revealing conserved bias among the samples. GC content at the third position (GC3) was the lowest, with a preference for A or T, suggesting weak SCU bias. The GC contents at the first two codon positions (GC1 and GC2) were extremely significantly correlated with one another but not with the expected number of codons (ENC). GC3 was not correlated with GC1 and GC2 but was extremely significantly correlated with ENC. Of the 30 high-frequency codons, 15, 14, 1 and 0 codons had U, A, G and C at the third position, respectively. The points for most genes were distributed above the neutrality plot diagonal. The points for 20 genes, accounting for 37.74% of all coding sequences (CDSs), were distributed on or near the ENC plot standard curve, and the ENC ratio ranged from -0.05-0.05. However, those of the other genes were under the standard curve, with higher ENC ratios. The points for most genes were distributed in the lower part of the PR2 plot, especially the bottom right corner. Twenty-eight highly expressed codons were screened and 11, 9, 7 and 1 codons had U, A, C and G as the third base, respectively. Twenty optimal codons were screened by comparing high-frequency codons and 11, 8, 0 and 1 codons had U, A, C and G as the third base, respectively. All samples were divided into six clades (r(2) = 0.9190, d = 0.5395) according to a relative synonymous codon usage (RSCU)-based phylogenetic tree. Camellia gauchowensis, C. vietnamensis, an undetermined oil-tea camellia species from Hainan province, and C. osmantha belonged to the same clade; the genetic relationships between C. gauchowensis, C. vietnamensis and the undetermined species were the closest. In summary, SCU bias is influenced by selection, while the influence of mutation cannot be ignored. As the SCU bias differed between species, this feature can be used to identify plant species and infer their genetic relationships. For example, C. vietnamensis and C. gauchowensis can be merged into one species, and the undetermined species can be considered C. vietnamensis. The results described here provide a basis for studying cpDNA gene expression and the development of cpDNA genetic engineering.