FIGURE 5 Identifying the most productive enzyme homolog for each step in a step-wise manner by using the normalized screening. (a) Schematic representation of the workflow for step-wise identifying the most productive enzyme homologs. (b) The most active Nampt homolog, SuNampt, was first identified. “Mr” represents Nampt from Meiothermus ruber , and “Mr ” represents Nampt from Meiothermus rufus . (c) By mixing with SuNampt, the most productive Prs homolog was proved to be MjPrs. (d) By combining with SuNampt and Mjprs, the most active Rbk homolog (i.e., OkRbk) was identified. The left histograms in (b), (c), and (d) show the results of the complementation fluorescence values for Nampt homologs, Prs homologs, and Rbk homologs, respectively. Error bars represent the standard deviation of three biological replicates. The (×) symbols represent that the RT/Fvalues of these homologs were not calculated. The right histograms in (b), (c), and (d) show the results of RT/F values of different NMN metabolic pathway enzymes. Error bars represent propagated error. See Table S2 for more detailed information on NMN synthesis reactions.
3.4 Testing and optimizing NMN biosynthesis
To confirm the results obtained using the normalized screening procedure, the 30 homologs were expressed in E. coli BL21(DE3), purified (Figure S4), and tested for their activities of NMN biosynthesis. It was noticed that the performance of our normalized screening procedure in evaluating the homologs with middle activity, such as CbNampt, PcPrs, and ThRbk, was unsatisfactory. However, there was no occurrence of the false positives using this procedure. As shown in Figure 6, the activity of SuNampt, MjPrs, and OkRbk was highest in the respective reaction system. In addition, the combination of OkRbk, MjPrs, and SuNampt improved NMN production by 2.9-fold to 295 mg/L from the best initial enzyme set, which consisted of EcRbk, PcPrs and MrNampt (Figure 2c). These results indicated that the normalized screening procedure provided a quick and easy way to screen active enzyme homologs. Particularly, it was well suitable for preliminary screening of dozens of enzyme homologs to find candidate homologs with high catalytic activities.