Automated multiplex genome-scale engineering in yeast

Abstract

Genome-scale engineering is indispensable in understanding and engineering microorganisms, but the current tools are mainly limited to bacterial systems. Here we report an automated platform for multiplex genome-scale engineering in Saccharomyces cerevisiae, an important eukaryotic model and widely used microbial cell factory. Standardized genetic parts encoding overexpression and knockdown mutations of >90% yeast genes are created in a single step from a full-length cDNA library. With the aid of CRISPR-Cas, these genetic parts are iteratively integrated into the repetitive genomic sequences in a modular manner using robotic automation. This system allows functional mapping and multiplex optimization on a genome scale for diverse phenotypes including cellulase expression, isobutanol production, glycerol utilization and acetic acid tolerance, and may greatly accelerate future genome-scale engineering endeavours in yeast.

Figure 1. Scheme of automated genome-scale engineering in yeast.

(a) Design of a general method to create multiplex genomic mutations in yeast. Gene modulation parts encoding various genetic modifications were flanked by homologous δ sequences for iterative and multiplex integration into repetitive genomic sequences. To enable CRISPR-Cas for efficient and selection-free δ integration, a P_ADH2-Cas9 expression cassette was integrated into the CAD strain (a S. cerevisiae strain with a constitutive RNAi pathway5) to create the CAD-Cas9 strain. (b) Construction of a genome-wide modulation part library. Full-length-enriched cDNA library was directionally cloned after a constitutive promoter P_TEF1, and resulted in genetic overexpression or knockdown for sense and anti-sense configurations, respectively, when transformed into the CAD strain.

Figure 2. Genome-wide screening and gene–trait profiling using the gene modulation collection.

(a) Scheme of cell surface display of a cellulase (Trichoderma reesei endoglucanase II, EGII) in yeast. (b) Genetic mutations conferring improved EGII-display levels estimated using fluorescence immunostaining of a His-tag attached to EGII. Overexpression and knockdown targets are listed as white and grey bars, respectively. (c) Correlation of norvaline (NVA) resistance, valine (VAL) concentration and isobutanol (iBuOH) production. (d) A knockdown mutant with increased isobutanol production. (e) Frequencies of individual genetic part in strain libraries before and after serial transfer in synthetic media containing glycerol as the sole carbon source. Calculated from the NGS data, frequency=(read count of individual part+0.5)/(total read count of all parts of the same modulation mode) (see Methods). Dots denote individual modulation part as red (overexpression part, OE), blue (knockdown part, KD), orange (KD for cit1) and blue (KD for afg1). (f) Genetic mutants conferring enhanced cell growth on glycerol. The same colour code is used as in d. C denotes the control strain (CAD-EGII in b and CAD in d,f) harbouring an empty pRS416 plasmid. Averages are defined as centre values, and error bars represent the mean±s.d. from biological replicates (n=3).

Figure 3. CRISPR-assisted δ integration of a GFP reporter.

(a) Workflow for iterative and multiplex integration. A screening/selection step can be included after plasmid curation to enrich mutants with desirable traits. (b) Donor plasmid of CRISPR-assisted δ integration, consisting of a 2 μ high-copy replication origin, CRISPR RNAs to direct Cas9 to the δ.a site in both the donor and genome, the selection markers (KanMX and URA3) for plasmid maintenance and curation, and an integration donor cassette flanked by homologous δ sequences (δ1 and δ2). (c) Genomic accumulation of GFP donors. Flow cytometry histogram overlays depict GFP fluorescence of cell populations after one (red) or two (blue) rounds of integration with the GFP donor plasmid in CAD-Cas9. Negative control (grey shade, NoGFP) was treated with an empty pRS426 plasmid in CAD-Cas9. CAD treated with the GFP donor in two rounds of integration was denoted as NoCas9 (black). In c, representative histograms are presented from biological triplicates.

Figure 4. Automated yeast strain engineering using iBioFAB.

(a) Hardware layout of iBioFAB (adapted with permission from ref. . Copyright (2017) American Chemical Society). (b) Process modules and unit operations in the workflow of yeast engineering. Any subset of all the unit operations enabled by iBioFAB can be programmed in custom-designed sequences to perform necessary process modules for creating and screening strain libraries in an iterative and automated manner. (c) Process flow diagram. Unit operations used in the yeast engineering workflow are marked in blue in b,c.

Figure 5. Screening of HAc-resistant yeast strains.

(a) Improved HAc tolerance in a strain library after the first round of integration (R1, solid line) compared with the wild-type control strain (dashed line) under various levels of HAc stress. (b) Ethanol fermentation with 1.1% (v/v) HAc by the parent and a mutant yeast strain isolated using iBioFAB. In a, error bars represent the mean±s.d. from technical replicates (n=16), where the cell libraries after plasmid curation were combined before being inoculated into 16 microtitre wells of selection media for each HAc concentration. In b, error bars represent the mean±s.d. from biological rplicates (n=3).