Engineering human-related complex proteins has great potential in development of effective therapeutics. Paramount to any directed evolution of such biologics is the construction of large libraries in mammalian cells that provide a native expression environment and allow the direct screening / selection of tens of millions of protein variants for desired properties. The common mammalian cell library construction platforms, such as episomal vectors, viral vectors and transposons, have the issues including low transgene stability, random integration and tug of war between the large library size and single variant per cell. Recently, targeted integration with nucleases has been developed whereas the integration efficiency is low, and off-target effects could result in multiple variants in one cell. To address these limitations, we developed a mammalian cell library construction platform based on recombinase-mediated cassette exchange (RMCE) which directs the integration of the transgenes into a single genomic locus, thereby rapidly achieving stable expression and transcriptional normalization. To improve the intrinsically low RMCE efficiency, we developed two complemental approaches: (1) simian virus 40 (SV40) large T-mediated transgene replication in Expi293F™ cells, and (2) nocodazole-mediated CHO cell mitosis phase arrest which not only presents two copies of targeted genomic locus but also facilitates their access due to temporary nuclear membrane disintegration.
Applying these novel approaches, mammalian cell libraries of up to 10 million clones were constructed, and library diversities were validated by deep sequencing and statistical analysis. Furthermore, the utility of the system was illustrated by isolation of GFP variants with enhanced fluorescence, and antibody Fc variants with significantly improved affinities towards specific Fc gamma receptors important for therapeutic antibody effector functions. Collectively, we developed platform technologies for combinatorial library construction in mammalian cells satisficing all desired features including large diversity, single copy, genomic integration, and defined locus. With contribution to the field of protein engineering, our work will facilitate the discovery of biologics and eventually lead to more and effective therapeutic strategies.