Protein production dynamics and physiological adaptation of recombinant Komagataella phaffii at near-zero growth rates

Abstract

Background: Specific productivity (q_P) in yeast correlates with growth, typically peaking at intermediate or maximum specific growth rates (μ). Understanding the factors limiting productivity at extremely low μ might reveal decoupling strategies, but knowledge of production dynamics and physiology in such conditions is scarce. Retentostats, a type of continuous cultivation, enable the well-controlled transition to near-zero µ through the combined retention of biomass and limited substrate supply. Recombinant Komagataella phaffii (syn Pichia pastoris) secreting a bivalent single domain antibody (VHH) was cultivated in aerobic, glucose-limited retentostats to investigate recombinant protein production dynamics and broaden our understanding of relevant physiological adaptations at near-zero growth conditions.

Results: By the end of the retentostat cultivation, doubling times of approx. two months were reached, corresponding to µ = 0.00047 h^-1. Despite these extremely slow growth rates, the proportion of viable cells remained high, and de novo synthesis and secretion of the VHH were observed. The average q_P at the end of the retentostat was estimated at 0.019 mg g^-1 h^-1. Transcriptomics indicated that genes involved in protein biosynthesis were only moderately downregulated towards zero growth, while secretory pathway genes were mostly regulated in a manner seemingly detrimental to protein secretion. Adaptation to near-zero growth conditions of recombinant K. phaffii resulted in significant changes in the total protein, RNA, DNA and lipid content, and lipidomics revealed a complex adaptation pattern regarding the lipid class composition. The higher abundance of storage lipids as well as storage carbohydrates indicates that the cells are preparing for long-term survival.

Conclusions: In conclusion, retentostat cultivation proved to be a valuable tool to identify potential engineering targets to decouple growth and protein production and gain important insights into the physiological adaptation of K. phaffii to near-zero growth conditions.

Fig. 1

VHH production and growth parameters estimated from chemostat cultivations and corresponding model-based predictions for retentostat cultures. Aerobic, glucose-limited chemostats for parameter estimation were operated at 7 different dilution rate setpoints (D = μ) ranging from 0.015 to 0.170 h⁻¹. Shown are A the biomass-specific VHH secretion rate (q_P) and B the maintenance energy requirement (m_S) as well as the maximum theoretical biomass yield (Y_XS^max) in relation to μ. The dynamics of m_S and Y_XS^max were estimated by linear regression analysis on moving windows of the specific glucose uptake rate (q_S) as determined from three different consecutive dilution rates and at least two individual cultivations per dilution rate setpoint. C Model-based predictions for retentostat cultivations of the biomass built-up (Cx), the glucose concentration in the feed (C_S), µ, q_S, as well as D the VHH concentration in the supernatant (C_P) and q_P were calculated based on the average m_S and Y_XS^max values returned from the three sets of q_S-vs-μ relations for μ ≤ 0.100 h⁻¹ from (B)

Fig. 2

Aerobic, glucose-limited retentostat cultures of VHH-secreting K. phaffii. Retentostat cultures were initiated from chemostat cultures operated at D = µ = 0.025 h⁻¹ at time-point zero. Shown are A the total (C_X) and viable (C_Xv) biomass accumulation profiles and the glucose concentration (C_S) in the feed throughout the retentostat phase as well as corresponding model-based predictions. B Estimates of the specific glucose uptake rate (q_S), μ and the average m_S calculated based on non-linear regression analysis. C VHH titer (C_P) in supernatants as well as corresponding model predictions based on extrapolation from chemostat data. D Ratio of the VHH-titer measured in supernatants (C_{P_R}) and filtrates (C_{P_F}) at the respective sampling point. E Representative SDS-PAGE of retentostat supernatants and VHH standard. Protein bands were visualized by silver-staining. F Relationship between q_P and µ as well as corresponding model predictions based on extrapolation from chemostat data

Fig. 3

Biomass composition of recombinant K. phaffii over a wide range of µ. Samples were taken from separate chemostat cultures operated at D = μ = 0.1 h⁻¹, retentostat cultures in the chemostat phase (D = µ = 0.025 h⁻¹; day 0), and throughout the retentostat phase (μ < 0.025 h⁻¹). A Macromolecular biomass composition; and B percentage of storage carbohydrate content and other carbohydrates. C Amino acid content measured in hydrolysates of whole cells. GlX represents the sum of Glu and Gln and AsX the sum of Asp and Asn. Tyrosine was not measured. D Lipid class content, including ceramide (CER), diacylglycerol (DAG), lyso-phosphatidylcholine (LPC), phosphatidylcholine (PC), phosphatidylethanolamine (PE), phosphatidylglycerol (PG), phosphatidylinositol (PI), phosphatidylserine (PS), sterol ester (SE), sterol (ST), triacylglycerol (TAG). All biomass composition data for the main sampling points is also provided in Additional file 2

Fig. 4

k-means cluster analysis, corresponding enriched GO terms, and regulatory trends of secretory pathway genes. A Samples for RNA-Seq analysis were taken from separate chemostat cultures operated at D = μ = 0.1 h⁻¹, retentostat cultures in the chemostat phase (D = µ = 0.025 h⁻¹; day 0), and throughout the retentostat phase (μ < 0.025 h⁻¹; days 6, 14 and 28). Genes that were differentially expressed compared to the highest μ of 0.10 h⁻¹ (FC ≥ 2; adj. p-value ≤ 0.01) in at least one comparison were grouped into four clusters by k-means clustering and analyzed for enriched GO terms. A comprehensive list of all enriched GO terms of the categories biological process (BP), molecular function (MF), and cellular component (CC) identified in the analysis is provided in Additional file 2. Enriched GO terms for Clusters 1 and 2 overlapped to a high degree. The main terms shared between these two clusters as well as all enriched GO terms from Clusters 3 and 4 were listed, respectively. B Total numbers of genes allocated to biological processes related to the secretory pathway and relative numbers of regulated genes. Color intensities reflect the degree of regulation of the respective group