First-generation baker’s yeast expression systems have a long, safe track record for biopharmaceutical manufacturing since the first regulatory approval of Novo Nordisk’s human insulin (Novolin®) in 1987. This followed extensive human use of this yeast for food and beverage manufacture since records began. Since then, over 40 therapeutic proteins manufactured by S. cerevisiae have been approved for biopharmaceutical applications by regulatory authorities such as the US Food and Drug Administration (FDA), making it the most widely approved yeast system for FDA-approved biologics. These products include a range of insulin analogues and GLP-1 receptor antagonist precursors for treating diabetes, and virus-like particle (VLP) vaccines for hepatitis B and human papillomavirus (HPV), which have been used to treat billions of patients since they were first approved.
Patents compromised use
However, the use of the 1st generation S. cerevisiae systems for biopharmaceuticals has declined recently despite it still having significant advantages over other systems when looking at overall net production. This may be due to greater access to Pichia pastoris (see below) at a time when patents limited access to the “industrial” versions of first-generation baker’s yeast systems, fuelled by promises of higher yields from some Pichia systems.
Highly characterised genome
Notably, S. cerevisiae is the model eukaryote and is by far the most highly characterised of all eukaryotes, if not all organisms. This is partly due to the wide distribution of laboratory-adapted strains worldwide in the 1950s and 1960s and the large yeast community using them for research in many aspects of biology. These early strains became the chassis for microbial production of therapeutic proteins for the first-generation baker’s yeast systems from the 1980s onwards. Publication of the first eukaryotic genome in 1996 for one of these strains, S288c, further facilitated the engineering of these first-generation systems for biologics manufacture. These strains were considered advantageous over many bacterial systems at the time due to their ability to secrete fully-folded active biopharmaceutical proteins into the culture supernatant with relatively low levels of host cell proteins compared to other systems.
Established safety, virus, toxin and animal derived ingredients free
Because the metabolism of S. cerevisiae is naturally well-adapted to the conversion of fruit sugars to antimicrobial ethanol and acetic acid for low-pH, it can obtain nutrients and compete against other microbes without the need to secrete the quantity or diversity of enzymes secreted by many other fungi. It has GRAS status and allows rapid growth on low-cost, fully-defined media, free from animal-derived ingredients. Additionally, S. cerevisiae does not have any known lytic viruses or pathogenic prions, so it presents neither the safety nor the production issues these agents present for CHO or E. coli. This, combined with its long, safe history of human use, makes it an excellent starting point with low regulatory risk for biopharmaceutical manufacture because the desired therapeutic protein is less likely to be modified by host cell proteins in the culture supernatant. Therefore, the secreted products can be highly homogeneous, without undesirable post-translational modifications that may cause immunogenicity risks, and high recoveries can be achieved during downstream processing.
Plasmids and promoters now free of patent restrictions
The advantages of S. cerevisiae over other systems include the high levels of homologous recombination, making genetic engineering relatively straightforward; a wide variety of episomal plasmids, allowing modulation of copy number; numerous constitutive and inducible promoters, allowing modulation of expression. The best of these, including the whole 2-micron plasmid systems and genomic mutations to enhance production quality or quality, became protected by patents or hidden as trade secrets, slowing wider adoption and development of first-generation baker’s yeast systems. Unlike the original dissemination of strains, this effectively limited broader system development to improve therapeutic protein production, leaving only inferior plasmids and strains for newcomers until patents expired.
Unsuited to N-linked glycosylation
Issues with the first-generation system include the various post-translational modifications, such as proteolytic cleavage and glycosylation patterns that are not the same as in human proteins. While these were solved for the FDA-approved biologics described above, they have limited successful approvals for many other therapeutic products, especially those requiring human-like N-linked glycosylation for efficient protein-folding, protease-resistance or biological activity. When the initial fermentation titres were too low for commercially viable manufacture, mutagenesis or rational engineering was successfully applied to improve yield. However, this typically required extensive strain development programs and was only applied to a selection of products for large-scale manufacture, e.g. insulins and albumin. Because the bottlenecks limiting the production of each protein product tend to be different, there was no guarantee these strains would perform well for different proteins, and in many cases, the engineering used to improve yields also created undesirable phenotypes, such as slower growth or increased lysis.
First Generation S. cerevisiae has fallen out of fashion
Consequently, the first-generation S. cerevisiae systems have fallen out of fashion and are less widely considered when selecting expression systems for new product development. This has been fuelled by a disproportionate emphasis on product yield, compared to product quality, favouring systems like E. coli and Pichia without fully considering DSP requirements, regulatory approval and overall process economics. The first-generation baker’s yeast system has excelled on net titre for many products after downstream processing and purification. However, the second-generation baker’s yeast systems developed using full genome optimisation offer even greater advantages, especially for products naturally suited to yeast expression and biologics that are difficult-to-express with other systems. The primary advantage is far greater flexibility for phenotypic improvements, because genetic diversity and evolution drive the optimisation process, rather than single-step modifications of a single fixed genomic chassis.
