Abstract: In order to understand the preference characteristics of codon usage in the midgut expression genes of Pieris rapae, and to provide a theoretical basis for application of genetic engineering technology to achieve the heterologous expression of serine protease from P. rapae, in this paper, 51 457 full-length sequences in the midgut transcriptome of P. rapae were calculated and counted for GC content, effective number of codons(ENC), relative synonymous codon usage(RSCU), codon usage frequency and other parameters to measure the codon usage preference of P. rapae by using bioinformatics softwares including CodonW, CHIP and CUSP. The results showed that the total GC content was 40.43%, and the average GC content of the third nucleotide of the codon (GC3) was 38.16%. The distribution of ENC ranged from 22.81 to 61.00 with an average ENC of 53.12. There were 1160 unigenes with ENC less than 35 which accounted for 2.25% of the total. The overall codon preference of genes was not high, but there were still differences in codon usage preference among different genes. The transcriptome codon has obvious preference for codons ending in A or U. The four codons of GAA, UUU, AAU and UAU were the optimal codons for the midgut expression genes of P. rapae. The codon usage preferences of serine proteases (chymotrypsin and trypsin) involved in digestion in the midgut of P. rapae and six different heterologous expression hosts were compared and analyzed. The results showed that Escherichia coli B was the most difficult host to heterologous expression, followed by insects High 5 cells and mammalian HEK293S cells. The expression difficulty of tobacco, Pichia pastoris and Arabidopsis thaliana was relatively low.

Key words: Pieris rapae, codon preference, serine protease, heterologous expression