A system for the continuous directed evolution of biomolecules

Researchers in the laboratory of Professor David Liu have developed a platform that enables the continuous directed evolution of gene-encoded molecules that can be linked to protein production in E. coli. During phage-assisted continuous evolution (“PACE”), successfully evolving genes are transferred from host cell to host cell through a modified M13 filamentous bacteriophage life cycle by linking desired activity to the expression of protein III (pIII; encoded by gene III), and thus to the production of infectious progeny phage. To accomplish this, gene III was deleted from the phage vector and inserted it into an “accessory plasmid” (AP) present in the E. coli host cells. The production of pIII from the AP is dependent upon the activity of the evolving gene(s) on the phage vector and in the absence of this activity is insufficient to support the production of infectious phage. Phage vectors able to induce sufficient production of pIII from the AP will therefore propagate and persist in a continuously diluted lagoon, while phage vectors unable to induce pIII will be washed out over time. Because pIII expression level determines the rate of infectious phage production, the progeny of a phage encoding a mutant gene that results in a higher level of pIII production will infect a larger share of host cells relative to the progeny of a phage encoding a less active gene. Mutants with higher activity will therefore experience a selective advantage until pIII levels are sufficient for maximal infectious phage release.

In principle, PACE is capable of evolving any gene that can be linked to pIII production in E. coli, including activities that influence pIII function at the transcriptional, translational, or post-translational levels. Because a wide variety of functions including DNA binding, RNA binding, protein binding, bond-forming catalysis, and a variety of enzyme activities have been linked to the expression of a reporter protein through n-hybrid and other conditional transcription strategies, PACE can be applied to the evolution of many different activities of interest. To date, the investigators have successfully linked and evolved protein-protein binding, recombinase activity, and RNA polymerase activity to phage infectivity in discrete infection assays by creating variants of the AP that associate each of these activities with pIII production. In the example described in the publication referenced below, the T7 RNA polymerase was evolved to recognize the T3 promoter and to initiate transcripts with nucleotides other than G.

The PACE system can be assembled entirely from a modest collection of commercially available equipment and does not require the manufacture of any specialized components. PACE proceeds at a rate of several dozen rounds of mutation, selection, and gene replication per day, representing roughly a 100-fold increase over most current protein evolution methods.

Overview of the PACE system: (a) PACE in a single lagoon. Newly arrived host cells are infected with selection phage (SP) encoding library members. Functional library members induce production of pIII from the accessory plasmid (AP) and release progeny capable of infecting new host cells. Non-functional library members do not produce pIII and release only non-infectious progeny. Increased mutagenesis is triggered upon lagoon entry through induction of a mutagenesis plasmid (MP). Host cells flow out of the lagoon on average faster than they can replicate, confining the accumulation of mutations to actively replicating SP. (b) Schematic of the PACE apparatus. Host E. coli cells maintained at constant cell density are continuously fed into the lagoon by peristaltic pump along with chemical inducers at a dilution rate of 1.0-3.2 volumes per hour.


Protein engineering by directed evolution has proven useful in conferring desired properties upon a wide variety of biomolecules of interest. However, with current methods, each round of evolution typically requires days or longer, with frequent intervention by the researcher involving discrete cycles of mutagenesis, transformation or in vitro expression, screening or selection, and gene harvesting and manipulation. Since successful evolution is strongly dependent on the total number of rounds performed, the labor- and time-intensive nature of discrete directed evolution cycles limit many laboratory evolution efforts to a modest number of rounds. A more rapid and automated approach to directed evolution is needed.

Intellectual Property Status: Patent(s) Pending