Standardizing cassette‐based deep mutagenesis by Golden Gate assembly

Abstract

Protocols for the construction of large, deeply mutagenized protein encoding libraries via Golden Gate assembly of synthetic DNA cassettes employ disparate, system‐specific methodology. Here we present a standardized Golden Gate method for building user‐defined libraries. We demonstrate that a 25 μL reaction, using 40 fmol of input DNA, can generate a library on the order of 1 × 106 members and that reaction volume or input DNA concentration can be scaled up with no losses in transformation efficiency. Such libraries can be constructed from dsDNA cassettes generated either by degenerate oligonucleotides or oligo pools. We demonstrate its real‐world effectiveness by building custom, user‐defined libraries on the order of 104–107 unique protein encoding variants for two orthogonal protein engineering systems. We include a detailed protocol and provide several general‐use destination vectors.

GGAssembler: precise and economical design and synthesis of combinatorial mutation libraries

“…The fold-change differences ranged between 0.58-1.76 and 0.47-2.67 in vHH1 and vHH2, respectively ( Figure 6A, B ). Mutation frequencies exhibited an even narrower range, with fold-change ranging 0.67-1.26 and 0.72-1.32 in vHH1 and vHH2, respectively ( Figure 6C, D ); this range is considered to be near-uniform(Daffern et al 2023).…”

“…New approaches to probe the mutational tolerance of proteins experimentally and computationally fuel interest in methods for efficient synthesis of combinatorial mutation libraries 3,15,25,26,[31][32][33][34][35][36][37] . GGAssembler exploits the fact that desired diversity in active sites is typically clustered in several contiguous epitopes (e.g., antibody CDRs or enzyme active-site loops) that are separated by immutable regions.…”

“…New approaches to probe the mutational tolerance of proteins experimentally and computationally fuel interest in methods for efficient synthesis of combinatorial mutation libraries (Tretyachenko et al 2020;Wrenbeck et al 2016;Kirby et al 2021;Daffern et al 2023;Jacobs et al 2015;Shimko, Fordyce, and Orenstein 2020;Püllmann et al 2019;Yamamoto et al 2020;Alejaldre, Pelletier, and Quaglia 2021;Sidore et al 2020;Plesa et al 2018;Öling et al 2022;Lund et al 2024). GGAssembler exploits the fact that desired diversity in active sites is typically clustered in several contiguous epitopes (e.g., antibody CDRs or enzyme active-site loops) that are separated by immutable regions.…”

“…We transformed the libraries into bacterial cells and sequenced 12 and 14 colonies for vHH1 and vHH2, respectively. 75% and 93% of the resulting sequences were in frame and matched desired designs for vHH1 and vHH2, respectively, on par with the current state of the art (Daffern et al 2023;Choi et al 2022). In both errors observed in vHH1, a pair of overhangs misligated, resulting in a short product (Supplementary Table 1).…”

“…In principle, these developments allow the assembly of dozens of fragments for combinatorial gene synthesis(Potapov, Ong, Kucera, et al 2018; Pryor et al 2020). Despite much progress(Püllmann et al 2019; Pryor et al 2020; Kirby et al 2021; Daffern et al 2023), however, methods for economical and accurate construction of variant libraries from dozens of fragments have not yet been demonstrated.…”