Mining Data From Plasma Cell Differentiation Identified Novel Genes for Engineering of a Yeast Antibody Factory

Abstract

Saccharomyces cerevisiae is a common platform for production of therapeutic proteins, but it is not intrinsically suited for the manufacturing of antibodies. Antibodies are naturally produced by plasma cells (PCs) and studies conducted on PC differentiation provide a comprehensive blueprint for the cellular transformations needed to create an antibody factory. In this study we mined transcriptomics data from PC differentiation to improve antibody secretion by S. cerevisiae. Through data exploration, we identified several new target genes. We tested the effects of 14 genetic modifications belonging to different cellular processes on protein production. Four of the tested genes resulted in improved antibody expression. The ER stress sensor IRE1 increased the final titer by 1.8-fold and smaller effects were observed with PSA1, GOT1, and HUT1 increasing antibody titers by 1. 6-, 1. 4-, and 1.4-fold. When testing combinations of these genes, the highest increases were observed when co-expressing IRE1 with PSA1, or IRE1 with PSA1 and HUT1, resulting in 3.8- and 3.1-fold higher antibody titers. In contrast, strains expressing IRE1 alone or in combination with the other genes produced similar or lower levels of recombinantly expressed endogenous yeast acid phosphatase compared to the controls. Using a genetic UPR responsive GFP reporter construct, we show that IRE1 acts through constitutive activation of the unfolded protein response. Moreover, the positive effect of IRE1 expression was transferable to other antibody molecules. We demonstrate how data exploration from an evolutionary distant, but highly specialized cell type can pinpoint new genetic targets and provide a novel concept for rationalized cell engineering.

Keywords: Saccharomyces cerevisiae; antibody; plasma cell differentiation; synthetic biology; transcriptomics.

FIGURE 1

Overview of the yeast-relevant GO-terms displaying enrichment in plasma cell differentiation. Some of the enriched GO-terms of biological processes are visualized in (A) for early phases of plasma cell differentiation until the PB phenotype and (B) for plasmacytic phase of differentiation from PBs into long-lived plasma cells. Up-regulated processes are colored red and down-regulated blue. Size of the node reflects the number of genes identified for that process and the color intensity reflects the statistical significance of the enrichment. The showed networks are mainly descriptive and not directly comparable, as they were created with different settings of GO-levels, annotation requirements and network connectivity. Figure was created with Cytoscape and the ClueGO-app.

FIGURE 2

Effects of the tested genes on IgG secretion and expression. Mean antibody titers are displayed for each genetic modification for overexpression (A) and deletion strains (B). Overexpressed or deleted genes are indicated below the charts. Overexpression and corresponding control strains were derived from YEK66. Strains were grown in SD medium lacking leucine. Deletion strains are derived from YEK18 and were grown in complete SD medium. Data points of biological replicates together with mean and standard error are shown. N = 4–8. (C) Immunoblot analysis of cell extracts after 16 h of expression.

FIGURE 3

Effects of combined expression of IRE1, GOT1, HUT1, and PSA1 genes on IgG secretion and growth. Mean antibody titers and final cell density are displayed for each genetic modification for strains expressing combinations of two (A,B), three (C,D), and four genes (E,F). Overexpressed genes are indicated below the charts. Overexpression and corresponding control strains were derived from YEK66 and contain two, three, or four low copy number plasmids. Control strains contain corresponding number of empty plasmids. Strains were grown in SD medium lacking the appropriate amino acids (leucine, histidine, and tryptophan) and uracil. Data points of biological replicates together with mean and standard error are shown. N = 6–9.

FIGURE 4

Effects of the selected genes on IgG secretion and growth. Mean antibody titers (A) and final cell density (B) are displayed for each strain expressing one or selected combinations of two, three, and four genes. Overexpressed genes are indicated below the charts. Overexpression and corresponding control strains were derived from YEK66 and all strains contain four low copy number plasmids. Empty plasmids are derived from pRS plasmid series and are labeled 413, 414, 415, and 416. Strains were grown in SD medium lacking leucine, histidine, tryptophan, and uracil. Data points of biological replicates together with mean and standard error are shown. N = 8–12.

FIGURE 5

Expression of IRE1 increases secretion of various IgGs. Two full-length and two scFv-Fc antibodies were co-expressed with IRE1 or a control plasmid. Mean antibody titers are displayed for each strain expressing scFv-Fc antibodies (pAX512 and pAX529) or full-length antibodies (pAX535 and pAX538). Overexpression and corresponding control strains were derived from W303α and all strains contain two low copy number plasmids. Antibody expression is under control of P_GAL1. Strains were grown in SD medium lacking leucine and uracil. Data points of biological replicates together with mean and standard error are shown. N = 4.

FIGURE 6

IRE1, GOT1, HUT1, and PSA1 enhance secretion of IgG but not of overexpressed endogenous AP. Antibody titers (A), secreted AP activity (B), and final cell density (C) are displayed for each strain expressing one or selected combinations of two, or all four genes. Overexpressed genes are indicated below the charts. All strains were derived from YMH14 and contain four low copy number plasmids. Expression of antibody and AP are under control of P_GAL1. Expression strains were grown in SD medium lacking leucine, histidine, tryptophan, and uracil. Protein expression was induced with 0.5 and 2.0% galactose. Data points of biological replicates together with mean and standard error are shown. N = 4.

FIGURE 7

Expression of UPR sensor and UPR specific transcription factor Hac1p enhances antibody production. Mean antibody titers (A), final cell density (B), and UPR activation (C) are displayed for strains constitutively expressing active spliced HAC1ⁱ and IRE1. For each expression construct the corresponding control plasmid was included. All strains were derived from YEK66. Strains were grown in SD medium lacking leucine. Data points of biological replicates together with mean and standard error are shown. N = 4–8.