Abstract: The coding sequences (CDS) of 1-deoxy-D-xylulose-5-phosphate synthase (DXS)from 16 Leguminous plants were selected for analysis. CodonW, CHIPS, CUSP and other programs were used to calculate the base composition of the sequences, the effective number of codons (ENc), the relative synonymous codon usage (RSCU) and other related parameters. These sequences were also analyzed by ENc-plot and PR2-plot. The results showed that there were 19 codons (RSCU>1) with strong preference for DXS genes in Leguminosae, in which optimal codons were CCA and GGA; both the GC content and GC3s were less than 0.5, indicating that A/U was preferentially used as the terminal codon. The ENc values ranged from 48.66 to 53.4, indicating weak the codon preference of the DXS genes. The ENc-plot and PR2-plot analysis displayed that the codon usage of DXS gene preference was mainly affected by mutation pressure. The clustering result based on CDS sequence was closer to the evolutionary classification of plants than the result based on RSCU clustering. Compared with Escherichia coli, Saccharomyces cerevisiae, Arabidopsis thaliana and Nicotiana tabacum were suitable hosts for DXS gene. All these results provided important significance to increase the yield of downstream terpenoids, by optimizing codons to improve the expression efficiency of DXS gene in other hosts.

Key words: Leguminous plants, DXS gene, codon usage bias, cluster analysis