Scalable protein production by Komagataella phaffii enabled by ARS plasmids and carbon source-based selection

Abstract

Background: Most recombinant Komagataella phaffii (Pichia pastoris) strains for protein production are generated by genomic integration of expression cassettes. The clonal variability in gene copy numbers, integration loci and consequently product titers limit the aptitude for high throughput applications in drug discovery, enzyme engineering or most comparative analyses of genetic elements such as promoters or secretion signals. Circular episomal plasmids with an autonomously replicating sequence (ARS), an alternative which would alleviate some of these limitations, are inherently unstable in K. phaffii. Permanent selection pressure, mostly enabled by antibiotic resistance or auxotrophy markers, is crucial for plasmid maintenance and hardly scalable for production. The establishment and use of extrachromosomal ARS plasmids with key genes of the glycerol metabolism (glycerol kinase 1, GUT1, and triosephosphate isomerase 1, TPI1) as selection markers was investigated to obtain a system with high transformation rates that can be directly used for scalable production processes in lab scale bioreactors.

Results: In micro-scale deep-well plate experiments, ARS plasmids employing the Ashbya gossypii TEF1 (transcription elongation factor 1) promoter to regulate transcription of the marker gene were found to deliver high transformation efficiencies and the best performances with the reporter protein (CalB, lipase B of Candida antarctica) for both, the GUT1- and TPI1-based, marker systems. The GUT1 marker-bearing strain surpassed the reference strain with integrated expression cassette by 46% upon re-evaluation in shake flask cultures regarding CalB production, while the TPI1 system was slightly less productive compared to the control. In 5 L bioreactor methanol-free fed-batch cultivations, the episomal production system employing the GUT1 marker led to 100% increased CalB activity in the culture supernatant compared to integration construct.

Conclusions: For the first time, a scalable and methanol-independent expression system for recombinant protein production for K. phaffii using episomal expression vectors was demonstrated. Expression of the GUT1 selection marker gene of the new ARS plasmids was refined by employing the TEF1 promoter of A. gossypii. Additionally, the antibiotic-free marker toolbox for K. phaffii was expanded by the TPI1 marker system, which proved to be similarly suited for the use in episomal plasmids as well as integrative expression constructs for the purpose of recombinant protein production.

Keywords: GUT1; Komagataella phaffii; Pichia pastoris; TPI1; Carbon source marker; Episomal ARS plasmid; Lipase CalB; Methanol-free bioprocesses.

Fig. 1

CalB reporter activities from transformants of episomal plasmid variants cultivated in micro-scale deep-well plates. Seven transformants of various ARS plasmids with different combinations of selection marker and its promoter, together with 7 colonies of the respective integrative references, were cultivated in 96 deep-well plates containing 250 µl BMG1. After 60 h of incubation at 28 °C and 320 rpm, 250 µl of BMG0.5 were added, followed by 50 µl BMG2.5 after 72 and 84 h to allow de-repression of P_DC and consequent reporter gene expression. Cultivations were harvested after 108 h of cultivation by centrifugation and CalB activities in the supernatants were evaluated

Fig. 2

CalB reporter activities from 250 mL shake flask cultivations. Five transformants of ARS plasmids with P_AgTEF1 driving the GUT1 or TPI1 marker expression were used to directly inoculate baffled 250 mL shake flasks containing 50 mL of BMG1. The respective average, integrative controls were cultivated in triplicates. After 60 h, 72 h and 84 h GUT1 plasmid cultivations and after 100 h, 112 h and 124 h TPI1 plasmid cultivations were fed with 500 µL of 50% glycerol. After 108 h the GUT1 plasmid cultivations and after 148 h the TPI1 plasmid cultivations were harvested and CalB activities were measured in the supernatants

Fig. 3

Time evolution of CalB titers during the fed-batch phase of the cultures. Results plotted are the average of two biological replicates

Fig. 4

Re-screening of individual cells from bioreactor cultivation. Sample 8 from the bioreactor cultivations were plated on BMG1 agar plates to obtain single colonies. A 96 deep-well plate containing 250 µl BMG1 medium was inoculated with one single colony per well. After 60 h of incubation at 28 °C and 320 rpm, 250 µl of BMG0.5 were added, followed by 50 µl BMG2.5 after 72 h and 84 h to allow de-repression of P_DC and consequent reporter gene expression. Cultivations were harvested after 108 h of cultivation by centrifugation and CalB activities in the supernatants were evaluated. Each bar represents a single well of the deep-well plate

Fig. 5

Vector map of a general ARS plasmid carrying a carbon marker cassette. Besides an origin of replication and an Ampicillin resistance cassette for cloning in E. coli, the episomal plasmids used in this study harbor the PARS1 sequence for the replication in K. phaffii, a reporter cassette with the de-repressable P_DC promoter driving CalB expression, whose secretion is facilitated by the D-alpha signal sequence, a deletion variant of the S. cerevisiae mating factor alpha signal peptide. Selective pressure in K. phaffii is provided by either GUT1 or TPI1 genes with one of the 5 tested promoters placed upstream