Modulating DNA Repair Pathways to Diversify Genomic Alterations in Saccharomyces cerevisiae

Abstract

Nuclease based genome editing systems have emerged as powerful tools to drive genomic alterations and enhance genome evolution via precise engineering in the various human and microbial cells. However, error-prone DNA repair has not been well studied previously to generate diverse genomic alterations and novel phenotypes. Here, we systematically investigated the potential interplay between DNA double strand break (DSB) repair and genome editing tools, and found that modulating the DSB end resection proteins could significantly improve mutational efficiency and diversity without exogenous DNA template in yeast. Deleting SAE2, EXO1, or FUN30, or overexpressing MRE11-H125N (nuclease-dead allele of MRE11), for DSB end resection markedly increased the efficiency of CRISPR/SpCas9 (more than 22-fold) and CRISPR/AsCpf1 (more than 30-fold)-induced mutagenesis. Deleting SAE2 or overexpressing MRE11-H125N substantially diversified CRISPR/SpCas9 or AsCpf1-induced mutation 2-3-fold at URA3 locus, and 3-5-fold at ADE2 locus. Thus, the error-prone DNA repair protein was employed to develop a novel mutagenic genome editing (mGE) strategy, which can increase the mutation numbers and effectively improve the ethanol/glycerol ratio of Saccharomyces cerevisiae through modulating the expression of FPS1 and GPD1. This study highlighted the feasibility of potentially reshaping the capability of genome editing by regulating the different DSB repair proteins and can thus expand the application of genome editing in diversifying gene expression and enhancing genome evolution. IMPORTANCE Most of the published papers about nuclease-assisted genome editing focused on precision engineering in human cells. However, the topic of inducing mutagenesis via error-prone repair has often been ignored in yeast. In this study, we reported that perturbing DNA repair, especially modifications of the various DSB end resection-related proteins, could greatly improve the mutational efficiency and diversity, and thus functionally reshape the capability of the different genome editing tools without requiring an exogenous DNA template in yeast. Specifically, mutagenic genome editing (mGE) was developed based on CRISPR/AsCpf1 and MRE11-H125N overexpression, and used to generate promoters of different strengths more efficiently. Thus, this work provides a novel method to diversify gene expression and enhance genome evolution.

Keywords: DSB repair; diversified mutation; gene expression; genome editing; mutational efficiency.

FIG 1

Evaluation and comparison of the mutational landscape generated by the different editing tools. (A) Diagram of the plasmids containing programmable nucleases and guide modules for the different genome editing tools. SpCas9, AsCpf1, and SpCas9 N863A were fused with C-terminal SV40 nuclear localization signal sequence and thereafter expressed under the control of GAL1 promoter and the CYC1 terminator. TALEN modules with 12 bp target sequences (left and right), the nuclease FokIs at C termini, and SV40 at N termini were also expressed under the control of GAL1 promoters and the CYC1 terminators. The gRNA containing self-cleaving hepatitis delta virus (HDV) ribozyme, 20 bp specific target sequence for URA3 (URA3-1, URA3-2, or URA3-3), and the structural components were expressed under the control of snoRNA SNR52 promoter and the terminator of yeast SUP4. The crRNA containing HDV ribozyme, crRNA scaffold, and 23 bp specific target sequence for URA3 was expressed under the control of SNR52 promoter and SUP4 terminator. (B) The cell viability of BY4741a after different genome editing at URA3 locus. (C) Mutational efficiency of BY4741a after the different genome editing at URA3 locus. (D) The ratios of the deletion and insertion generated by the different genome editing tools at URA3 locus in BY4741a. The different repair outcomes were measured by PCR amplification of the target URA3 locus from randomly selected 100 individual colonies in SD + 5-FOA plate (Fig. S2), followed by DNA sequencing. The ratio was calculated based on the number of inserted or deleted mutations at the target URA3 loci. (E) Frequency for the size of inserted (+) or deleted (–) nucleotides at the vicinity of the target loci generated by the different genome editing tools. (F) The top four mutational types were generated by four different editing tools. The data of (D), (E), and (F) were calculated from DNA sequencing results of above analyzed 100-individual colonies. Error bars were derived from two different biological triplicates. Control: BY4741a harboring pRS423; CRISPR/SpCas9: BY4741a harboring p423-gRNA(URA3-1)-SpCas9; CRISPR/AsCpf1: BY4741a harboring p423-crRNA(URA3)-AsCpf1; TALENs: BY4741a harboring p423-GAL-L12R12; CRISPR/SpCas9 N863A: BY4741a harboring p423-gRNA1-gRNA2-SpCas9 N863A.

FIG 2

Potential effects of modulating DSB repair proteins on mutational efficiency and cell viability for editing tools. (A) CRISPR/SpCas9, (B) CRISPR/AsCpf1, (C) TALENs, and (D) CRISPR/SpCas9N863A-induced mutational efficiency and cell viability were evaluated in the various strains with DSB repair related genetic modifications via gene deletion, overexpression, or mutations. Notably, the TALENs were performed under mixed sugar conditions (2% glucose with 2% galactose). Blue column: mutational efficiency; yellow column: cell viability. All the different strains contained p423-gRNA(URA3-1)-SpCas9, p423-crRNA(URA3)-AsCpf1, p423-GAL-L12R12, and p423-gRNA1-gRNA2-SpCas9 N863A, respectively, for the different genome editing (Fig. 1A and Table S1). Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple-comparison posttest (***, P < 0.001; **, P < 0.01; *, P < 0.05).

FIG 3

Possible effects of modulating DSB repair proteins on mutational diversity for editing tools. (A) CRISPR/SpCas9, (B) CRISPR/AsCpf1, and (C) TALENs-induced total mutational diversity were evaluated in the various strains with DSB repair-related genetic modifications via inducing gene deletion, overexpression, or mutation. The total number of mutations of the wild-type and genetically modified strains after different genome editing is shown in the left panel, and these numbers were counted by PCR amplification of the URA3 loci from randomly selected 100-individual colonies of each strain with the different modulation and genome editing tools, followed by DNA sequencing. Correspondingly, the sequencing outcomes were further analyzed by classifying the different groups based on mutation types, including >11-nt deletion, 2–11-nt deletion, 1-nt deletion, 1-nt insertion, multiple-nt insertion, and multiple mismatches, and the ratio of each group was calculated as shown in the right panel. Multiple mismatches observed represent the mutagenesis sequence containing both the deletion and insertion. All the strains contained p423-gRNA(URA3-1)-SpCas9, p423-crRNA(URA3)-AsCpf1, p423-GAL-L12R12, and p423-gRNA1-gRNA2-SpCas9 N863A, respectively, for the different genome editing (Fig. 1A and Table S1).

FIG 4

Comparison of the mutational landscapes of CRISPR/AsCpf1 genome editing with/without MRE11-H125N. The mutational landscapes were measured based on DNA sequencing results of URA3 loci from each 100 randomly selected 5-FOA-resistant mutants (edited mutants), as described in Fig. S2. WT: S. cerevisiae BY4741a harboring p423-crRNA(URA3)-AsCpf1; MRE11-H125N mutant strain: S. cerevisiae BY4741a mre11Δ harboring pMRE11-H125N and p423-crRNA(URA3)-AsCpf1 (Table S1 and Data Set S3).

FIG 5

A schematic diagram of the mutagenic genome editing (mGE) strategy for improving the different cellular phenotypes. The mGE tool was developed under the cooperation between MRE11-H125N overexpression and CRISPR/AsCpf1. The initial mre11Δ yeast cell used was harboring pMRE11-H125N and p423-crRNA(X)-AsCpf1 (X: the target gene). In this study, eGFP and FPS1 were selected as the examples and first grown in SD media containing 20 g/L glucose and the auxotrophic compounds for 12 h, and then were diluted into the SD media containing 20 g/L galactose and the auxotrophic compounds for 24 h to initiate editing. The cell population was then iteratively edited to generate the final mutated population for screening the desired phenotypes. If necessary, other genetic modification of important DNA repair proteins could also be selected to potentially replace MRE11-H125N overexpression for building a similar mGE strategy in combination with other genome editing tools. For example, MRE11-H125N overexpression could be replaced by SAE2, EXO1, FUN30, or MSH2 deletion as well as CDC9 overexpression, while CRISPR/AsCpf1 could be effectively replaced by CRISPR/SpCas9 or TALENs as well as other suitable genome editing tools.

FIG 6

Mutational diversity of fluorescence expression resulted from the synthetic promoter editing carried out via mGE. (A) An enhanced GFP reporter driven by the synthetic minimal promoter Pmini was integrated to the PDC1 locus of S. cerevisiae BY4741a, and then the mGE tools were introduced to generate pdc1::eGFP (CT) and MRE11-H125N, pdc1::eGFP (MT), respectively. The schematic diagram of the promoter architecture is illustrated at the bottom and the detailed sequencing information is shown in Fig. S10. (B) The distribution of the fluorescence was examined by flow cytometry, and then fluorescence diversity was characterized as coefficient variation (CV). Three random mutants from six different phenotypic diversity ranges were thereafter sorted out from iterative mGE-based edited populations (MT-Edit, R1–R6) (C) and analyzed by DNA sequencing to align their genetic variations (D). R1 and R2 appeared to have the same mutated sequences. (E) Flow cytometry histograms of the six mutants (R1–R6) and the two controls. Gray histograms represent the negative controls (no GFP expression), and black histograms represent positive controls (unedited GFP expression). The histograms of the six mutants are portrayed with green, purple, blue, yellow, sapphirine, and red, respectively. (F) The relative eGFP expression profile of the six mutants with mutated synthetic promoters at the genomes or plasmids. For genome scale, the originally sorted mutants of R1–R6 and the strain BY4741a mre11Δ, pdc1::eGFP (control) were effectively used to measure the fluorescences of eGFP expression. The fluorescence of BY4741a mre11Δ, pdc1::eGFP was first normalized to 100% and others were calculated as the values related to that of BY4741a mre11Δ, pdc1::eGFP. For the plasmid scale, S. cerevisiae BY4741a harboring the plasmid p315-Pmini(Original)-GFP, p315-Pmini(R1)-GFP, p315-Pmini(R2)-GFP, p315-Pmini(R3)-GFP, p315-Pmini(R4)-GFP, p315-Pmini(R5)-GFP, or p315-Pmini(R6)-GFP) with the enhanced GFP reporters driven by original (control) or mutated (R1–R6) synthetic minimal promoters, respectively, was used to measure the fluorescence of eGFP expression. The fluorescence of S. cerevisiae BY4741a with p315-Pmini(Original)-GFP was normalized to 100%, and others were then calculated as the values related to that of S. cerevisiae BY4741a with p315-Pmini(Original)-GFP. The error bars were derived from three different biological triplicates.

FIG 7

Diversification of the gene expression of FPS1 and GPD1 for regulating glycerol and ethanol production in yeast via mGE. (A) To alter the glycerol pathway, three different guides (gRNA1, gRNA2, and gRNA3) were designed at the native promoter of FPS1 and GPD1 separately. (B) The distribution of glycerol productivity of MT-Edit and CT-Edit with the different guides of FPS1 and GPD1 was examined in MT-Edit and CT-Edit populations (Data Set S4). BY4741 or MRE11-H125N mutant strain harboring blank plasmid pRS423 was then used as the control. The diversity of glycerol productivity was characterized as coefficient variation (CV). (C) The statistical analysis of the mutation types of MT-Edit and CT-Edit with the different guides of FPS1 and GPD1. After three rounds of editing via mGE, the genomic DNA of each mutant library was next obtained and then used for amplifying the target locus to characterize the mutation types via amplicon sequencing (Novogene, Tianjin, China). (D) Glycerol and ethanol production profile of the two mutants with mutated pFPS1 or pGPD1 promoters. The mutants were sorted out from mGE-based edited populations (MT-Edit/gRNA2). Fermentation capacities were later evaluated at 30°C in 100-mL Erlenmeyer flasks containing 50 mL SD medium with 100 g/L glucose at 220 rpm. In panels B and D, data represent the mean and standard error of duplicate cultures under each condition. Statistical analysis was performed using one-way ANOVA followed by Tukey’s multiple-comparison posttest (***, P < 0.001; **, P < 0.01; *, P < 0.05). In addition, statistical analysis in B was performed using two-way ANOVA (with the strains and gNRA as the factors) followed by Tukey’s multiple-comparison posttest for selecting the best gRNA for multiplex editing. (E) The mutational landscape of the two different mutants of FPS1-M and GPD1-M.