PCRless library mutagenesis via oligonucleotide recombination in yeast

Abstract

The directed evolution of biomolecules with new functions is largely performed in vitro, with PCR mutagenesis followed by high-throughput assays for desired activities. As synthetic biology creates impetus for generating biomolecules that function in living cells, new technologies are needed for performing mutagenesis and selection for directed evolution in vivo. Homologous recombination, routinely exploited for targeted gene alteration, is an attractive tool for in vivo library mutagenesis, yet surprisingly is not routinely used for this purpose. Here, we report the design and characterization of a yeast-based system for library mutagenesis of protein loops via oligonucleotide recombination. In this system, a linear vector is co-transformed with single-stranded mutagenic oligonucleotides. Using repair of nonsense codons engineered in three different active-site loops in the selectable marker TRP1 as a model system, we first optimized the recombination efficiency. Single-loop recombination was highly efficient, averaging 5%, or 4.0×10(5) recombinants. Multiple loops could be simultaneously mutagenized, although the efficiencies dropped to 0.2%, or 6.0×10(3) recombinants, for two loops and 0.01% efficiency, or 1.5×10(2) recombinants, for three loops. Finally, the utility of this system for directed evolution was tested explicitly by selecting functional variants from a mock library of 1:10(6) wild-type:nonsense codons. Sequencing showed that oligonucleotide recombination readily covered this large library, mutating not only the target codon but also encoded silent mutations on either side of the library cassette. Together these results establish oligonucleotide recombination as a simple and powerful library mutagenesis technique and advance efforts to engineer the cell for fully in vivo directed evolution.

Figure 1

Oligonucleotide recombination via yeast HR. Oligonucleotide recombination provides a general method for generating targeted libraries of DNA mutants in vivo. A linearized vector expressing the target gene (gray) and linear DNA oligonucleotides (green, yellow, blue) are co-transformed into yeast. HR between the target gene and DNA oligonucleotides yields libraries of the mutated target gene.

Figure 2

Optimization of experimental parameters for high rates of recombination. A: Linear or circular trp1-Arg44* vector (1 μg) is co-transformed with 5 μg of the ss oligonucleotide ARG44Fix (1:1000 mol vector: oligonucleotide) into the Δtrp1 S. cerevisiae strain ATCC4017202. Co-transformation of linear vector and oligonucleotide yields the greatest percentage TRP1 recombinants. B: Linear trp1-Arg44* vector (1 μg) and varying amounts of ss and ds oligonucleotides ARG44Fix and ARG44Fix_ds, respectively, are co-transformed into the Δtrp1 S. cerevisiae strain ATCC4017202 and recombinant colonies are scored by plating on SC (Ura⁻) and SC (Ura⁻Trp⁻) selective plates. Co-transformation of 1:1000 mol vector:oligonucleotide yields the greatest percentage TRP1 recombinants. The data shown are the mean ± the standard error of at least three separate experiments. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 3

Efficiency of oligonucleotide recombination at multiple loops. Linear vectors carrying trp1-Arg44*, trp1-Arg78*, trp1-Ser201*, trp1-Arg44*Arg78*, trp1-Arg44*Ser201*, and trp1-Arg44*Arg78*Ser201* were co-transformed into the Δtrp1 S. cerevisiae strain ATCC4017202 with the appropriate oligonucleotide (see Table I). A: Oligonucleotide used in this study to fix nonsense mutations in residues Arg44, Arg78, and Ser201 in yPRAI. Sequences of target codons are highlighted. B–C: Simultaneous mutagenesis of multiple loops exhibits multiplicative efficiency. (B) Schematic representation of one (a–c), two (d–e), or three (f) oligonucleotide mutagenesis. (C) The data shown are mean ± standard error for percent recombination of at least three separate experiments. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]

Figure 4

Mock library mutagenesis of trp1-Ser201* vector. A wt codon was enriched from a mock library of 10⁶ inactive variants in a single step. A: Vector NP2282, carrying a trp1-Ser201* allele, is targeted with a 1:10⁶ mix of fixing to nonsense ssDNA oligonucleotides (oligonucleotides SER201LibraryFix and SER201LibraryOpal, respectively). B: Eighteen viable colonies were analyzed by sequencing (see text), and 13 high quality recombinant sequences were aligned using the clustalW server. The vector sequence is shown at the top. All sequenced colonies carried the fixing codon (TCC) at position 201 (framed) as well as the encoded silent mutations upstream (A>G) and downstream (T>C) (highlighted). Library results suggest that the library size is fully covered and therefore may allow for larger library experiments. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]