Laboratory evolution has generated many biomolecules with desired properties, but a single round of mutation, gene expression, screening or selection, and replication typically requires days or longer with frequent human intervention1. Because evolutionary success is dependent on the total number of rounds performed2, a means of performing laboratory evolution continuously and rapidly could dramatically enhance its effectiveness3. Although researchers have accelerated individual steps in the evolutionary cycle4,5,6,7,8,9, the only previous example of continuous directed evolution was the landmark study of Wright and Joyce10, who continuously evolved RNA ligase ribozymes with an in vitro replication cycle that unfortunately cannot be easily adapted to other biomolecules. Here we describe a system that enables the continuous directed evolution of gene-encoded molecules that can be linked to protein production in Escherichia coli. During phage-assisted continuous evolution (PACE), evolving genes are transferred from host cell to host cell through a modified bacteriophage life cycle in a manner that is dependent on the activity of interest. Dozens of rounds of evolution can occur in a single day of PACE without human intervention. Using PACE, we evolved T7 RNA polymerase (RNAP) variants that recognize a distinct promoter, initiate transcripts with ATP instead of GTP, and initiate transcripts with CTP. In one example, PACE executed 200 rounds of protein evolution over the course of 8 days. Starting from undetectable activity levels in two of these cases, enzymes with each of the three target activities emerged in less than 1 week of PACE. In all three cases, PACE-evolved polymerase activities exceeded or were comparable to that of the wild-type T7 RNAP on its wild-type promoter, representing improvements of up to several hundred-fold. By greatly accelerating laboratory evolution, PACE may provide solutions to otherwise intractable directed evolution problems and address novel questions about molecular evolution.
This work was supported by National Institutes of Health/National Institute of General Medical Sciences R01 GM065400 and by HHMI. K.M.E. acknowledges graduate research fellowships from the Hertz Foundation and the National Science Foundation. J.C.C. was supported by the Harvard Chemical Biology Graduate Program. We thank B. Dorr for assistance with phage generation modelling, E. Curtis for suggestions and V. D’Souza for plasmid pT7-911Q.
The file contains Supplementary Results, Supplementary Figures 1-12 with legends, Supplementary Tables 1-3 and additional references.