Saturation mutagenesis genome engineering of infective FX174 bacteriophage via unamplified oligo pools and golden gate assembly

Matthew S. Faber; James T. Van Leuven; Martina M. Ederer; Yesol Sapozhnikov; Zoë L. Wilson; Holly A. Wichman; Timothy A. Whitehead; Craig R. Miller

In nicking mutagenesis, desired mutations are introduced by libraries of mutagenic oligonucleotides, which can be sourced from unamplified oligo pools (Medina-Cucurella et al. 2019 ). We synthesized a custom pool of 11,860 oligos that encoded nearly every missense and nonsense mutation in the F and G genes. Each missense or nonsense mutation was encoded by a single oligonucleotide. Application of nicking mutagenesis using this oligo pool for different plasmids (F1, F2, F3, G1, G2) resulted in at least 14-fold excess transformants required for 99.9% theoretical coverage of the desired library (Supplemental Table S3 ). The diversity of the mutant plasmid libraries was validated using deep sequencing on an Illumina MiSeq platform. All plasmid libraries have >99.7% coverage of all possible single mutations (n = 11,821/11,860, minimum counts for a mutation >5, full library statistics are given in Table 1 . CC-BY-NC 4.0 International license is made available under a The copyright holder for this preprint (which was not peer-reviewed) is the author/funder. It . https://doi.org/10.1101/798546 doi: bioRxiv preprint non-susceptible hosts (Agilent XL1-Blue), then plated on susceptible hosts (E. coli C) after a 30-minute recovery in SOC. Viral titers increase rapidly after 30 minutes, suggesting that the XL1-Blue cells do not burst until after this time point (Supplemental Table S5 ). Accordingly, we reasoned that the number of plaques at 30 minutes represents the number of XL1-Blue cells that were transformed with viable genomes. The number of transformants varied considerably between constructs (Table 2) . 19,200 plaques were recovered from the co-transformed G1/ G2 libraries, where only 3,480 total variants were expected in the pool. On the contrary, 460 plaques were observed for the F2 library, when 2,280 total variants were in the starting plasmid pool. The underrepresentation of the F2 library at this scale could be remedied by transforming more competent cells. We expect that differences between fragments in the number of recovered viruses is due to in part to the tolerability of fragments to mutation, as inviable mutants would not form plaques. Estimates from deep mutational scanning of model proteins show that approximately 18-38% of all mutations are nonviable (Firnberg et al. 2014 , Faber et al. 2019 , and this percentage may be larger for structural proteins like F and G.