Trehalose is a nonspecific protective agent for biomacromolecules. Trehalose-6-phosphate synthase (OtsA)/phosphatase (OtsB), which is encoded by the gene operon otsBA located at -42 of the Escherichia coli genome, is the main enzyme system that catalyzes the synthesis of trehalose in E. coli. We cloned the operon and modified it by directed evolution. Unlike in the previously reported work, we modified the whole operon and screened the positive mutant simultaneously. Thus we believe that the gene complex solves the negative effects between two enzymes if one of them diversifies its structure or functions and finds the form most suitable for trehalose synthesis. It is thus mimics the natural process, in which the functional improvement of organisms is related to alterations in coordinated enzymes. The evolution procedure was carried out in a sequence of error prone PCR, shuffling PCR, and then strict screening of the mutants. After screening of a Library of more than 4000 colonies, about 15 positive colonies were analyzed, resulting in a higher concentration of trehalose than control. One of them, E. coli TS7, shows 12.3-fold higher trehalose synthesis ability than E. coli DH5 alpha. In contrast, we introduced the cDNA sequence of the tps1 gene from Saccharomyces cerevisiae, which has 54% identity with the gene otsA, as one of the templates in shuffling PCR. By hybrid evolution and screening, we obtained 10 positive colonies with higher concentrations of trehalose than control. E. coli TS22 appears to have 5.3-fold higher trehalose synthesis ability than E. coli DH5 alpha and 1.6-fold more than E. coli DEF3(pOTS11). This result demonstrated that coevolution and hybrid evolution, as powerful protocols in protein engineering, are effective in modifying enzyme. It indicates that repeating the process of genomic evolution in nature is feasible.