YTK Display-and-Secrete: Screening for Optimal Protein Secretion Elements in Saccharomyces cerevisiae

Abstract

Engineering yeast to secrete target proteins requires searching for optimal combinations of promoters and signal peptides so that genes can be composed that give a high expression and efficient secretion. Most methods for this involve laborious, one-by-one assessments or require the use of enzymatic reporter proteins in order to achieve high-throughput capacity. Here, we introduce a novel modular method for the high throughput screening of yeast strains designed to secrete proteins of interest. Our approach integrates combinatorial DNA assembly, yeast surface display, flow cytometry, and nanopore DNA sequencing to facilitate rapid screening. Building on a widely used yeast toolkit (YTK) for modular cloning, our system creates surface display libraries with N- and C-terminal epitope tags by fast DNA assembly and genome-integration into Saccharomyces cerevisiae. Flow cytometry with fluorescently labeled epitope-binding antibodies identifies strains that secrete and display the most full-length protein and can rapidly sort these from low secretors. We demonstrate our system by optimizing the secretion of the enzyme beta-lactamase and several elastin-like polypeptides (ELPs), first identifying strains with modular genetic element combinations that give the best surface display and then validating that removal of the surface-display anchor protein in these strains gives a high target protein secretion. We then show how pooled long read sequencing of sorted cells can determine the effectiveness of numerous combinations of promoters and signal peptides for a target protein in a single experiment. The data sets from this offer new insights into an optimal element choice for efficient protein secretion and could train machine learning models.

Keywords: nanopore sequencing; protein secretion; yeast surface display; yeast toolkit.

1

Illustration of yeast surface display and genetic designs. (A) Diagram of a resulting yeast cell with an integrated genetic construct into a LEU2 locus. A synthetic protein is displayed via Aga2 anchor protein and carries FLAG tag at the N terminus and HA tag at the C terminus. Epitope tags can be detected with fluorescence-conjugated antibodies PE-DAZZLE (anti-FLAG) and FITC-anti-HA. (B,C) Design of genetic parts and the integration vector for yeast surface display (B) and secretion (C). Promoters are from the original MoClo YTK, SP-FLAG, and CDS #1 designed as YTK parts 3a and 3b, CDS #2, and HA-terminator with or without anchor Aga2 designed as modified 4a and 4b parts.

2

Analysis of yeast libraries displaying β-lactamase, ELP2, ELP4, ELP6, ELP8, and ELP10. (A). Flow cytometry analysis of yeast libraries. ≥ 2000 yeast colonies were collected. Cells were treated with FITC-anti-HA and PE-DAZZLE antibodies and analyzed by Attune NxT Flow Cytometer. The negative control is stained AGA1⁺ strain (without surface display), the positive control is pCCW12 MF-3xFLAG-ELP1-CBM-AGA2-HA stained. At least 10,000 events were analyzed for each sample. A dashed line separates negative and positive populations. (B) Results of nanopore sequencing of pooled amplicons from β-lactamase, ELP2, and ELP4 libraries. Heatmaps were built with the data obtained by nanopore sequencing. ≥ 1000 full reads for each library were analyzed. Each square represents a single genotype, and the intensity of the color correlates with the number of reads for each genotype within the analyzed full reads.

3

Sorting of yeast libraries. (A) Illustration of the sorting strategy into three populations based on the PE-DAZZLE fluorescence intensity. (B) Gating strategy for FACS. The sorting was performed with BD FACSAria III. 10⁶ cells were sorted into three populations based on the PE-DAZZLE fluorescence intensity. All the data for the three libraries is plotted as a histogram. Control is pYTK010 MF-3xFLAG-ELP1-CBM without staining (negative control) and with staining (positive control). (C) Flow cytometry analysis of sorted populations after overnight recovery. The recovered cells were stained with the PE-DAZZLE antibody and were analyzed by an Attune NxT Flow Cytometer. At least 10,000 events were analyzed for each sample. Control is pYTK010 MF-3xFLAG-ELP1-CBM without staining (negative control) and with staining (positive control).

4

Heatmaps of different promoter and signal peptide combinations present in each sorted population. Amplicons with barcodes specific to beta-lactamase, ELP2 and ELP4, and each sorted subpopulation were sequenced with nanopore sequencing. Heatmap intensities corelate with the number of reads for each promoter and signal peptide combination identified by sequencing.

5

Validation of the top hits from beta-lactamase populations. The most abundant promoter and signal peptide were assembled into surface display and secretion vectors and integrated into yAK03 (A) and BY4741 (B), respectively. Pale purplelow, purplemedium, and dark purplehigh population. (A) Yeast cells were stained with PE-Dazzle antibodies and fluorescence intensity was measured by flow cytometry. Three independent isolates were analyzed, data is presented as an average of median fluorescence values, n = 3, error bars = standard deviation. Controls are pYTK010 MF-3xFLAG-ELP1-CBM without staining (display + no stain) and with staining (positive control); yeast without surface display (negative control). (B) Beta-lactamase-CBM enzymatic activity assay was performed with a supernatant with 50 μg/mL of nitrocefin solution. Three independent isolates were analyzed, data is presented as an average of OD490 reads, error bars = standard deviation. Negative control is a BY4741 strain, reference is 1 μg/mL of beta-lactamase solution. Values for the supernatant are normalized to rich media optical density.

6

Validation of the top hits from ELP populations. The most abundant promoter and signal peptide combinations were assembled into surface display and secretion vectors and integrated into yAK03 (A,C) and BY4741 (B,D), respectively. Pale pink and orangelow, pink and orangemedium, and dark pink and brownhigh population. (A,C) Yeast cells were stained with PE-Dazzle antibodies and fluorescence intensity was measured by flow cytometry. Three independent isolates were analyzed, data is presented as an average of median fluorescence values, n = 3, error bars = standard deviation. Controls are pYTK010 MF-3xFLAG-ELP1-CBM without staining (display + no stain) and with staining (positive control); yeast without surface display (negative control). (B,D) Levels of ELP-CBM secretion in strains from medium and high populations were assessed by the Western blot. The supernatant of overnight cultures was concentrated and loaded onto two different SDS-PAGE gels. After the transfer each membrane was probed with anti-HA and anti-FLAG primary antibodies for both N and C termini of the secreted protein. A representative Western blot is shown here.

7

Single colony analysis of Mel1 surface display and secretion. (A) Heatmap of median PE fluorescence intensity of single colonies with Mel1 displayed. A single colony was picked and grown in a 96 well plate. Wells A1-A4 are empty, well A5 is the negative control (pYTK010 MF-3xFLAG-ELP1-CBM without staining) and well A6 is the positive control (pYTK010 MF-3xFLAG-ELP1-CBM stained). (B) PE fluorescence of selected strains with Mel1 surface display with medium and high fluorescence intensities: wells H4, E8, C6, H7, H1, and A9. Combinations of a promoter and a signal peptide for these strains were identified by Sanger sequencing. The highest levels of surface display were in wells E5 and F2, but sequencing showed that these strains do not have integration of the full-length genetic construct. Wells H1, B5, and D3 had an identical promoter and signal peptide combinations. (C) Alpha-galactosidase enzymatic activity assay. Supernatant of overnight cultures from 3 independent transformants was mixed with citric buffer and X-alpha-Gal. Images were made after 6 h of the reaction. The color intensity was analyzed with ImageJ as a gray value of a region of interest within each well. Negative control is the supernatant from BY4741. n = 3, error bars = standard deviation.