Generation of research cell bank
To generate a research cell bank (RCB), P. pastoris X33 strain was transformed with expression plasmid pPICZαA containing RBD219-N1C1 coding DNA, and one transformed colony with high expression of recombinant RBD219-N1C1 protein (Chen et al. 2021) was selected and streaked on yeast extract peptone dextrose (YPD) plates containing 100 μg/mL Zeocin to make single colonies. The plates were incubated at 30 °C for approximately 3 days until single colonies were observed. Subsequently, 200-mL plant-derived phytone YPD medium was inoculated with a single colony from the respective plate and incubated at 30 °C with constant shaking (225 rpm) until the OD600 reached 9.3. Finally, the cell culture was mixed with plant-derived glycerol to a final concentration of 20% and aseptically aliquoted (1 mL each) into 1.2-mL cryovials. For long-term storage, the cryovials were stored at −80 °C.
One vial of the SARS-CoV-2 RBD219-N1C1 RCB was used to inoculate a 0.5-L buffered minimal glycerol (BMG) medium in a 2-L baffled shake flask. The shake flask culture was grown at 30 °C and 225 rpm until an OD600 of 5–10. For 1-L fermentations, this seed culture (20–40 mL) was inoculated into 0.4 L of sterile basal-salt medium (BSM) (pH 5.0; BSM: 18.2 g/L potassium sulfate, 14.9 g/L magnesium sulfate heptahydrate, 4.13 g/L potassium hydroxide, 0.93 g/L calcium sulfate dehydrate, 26.7 mL/L of 85% phosphoric acid, and 40 g/L glycerol) or low-salt medium (LSM) (pH 5.0; LSM: 4.55 g/L potassium sulfate, 3.73 g/L magnesium sulfate heptahydrate, 1.03 g/L potassium hydroxide, 0.23 g/L calcium sulfate dehydrate, 10.9 mL/L of 85% phosphoric, and 40 g/L glycerol) to a starting cell density (OD600) of 0.5. Fermentation was conducted using a Biostat Qplus bioreactor with a 1-L vessel (Sartorius Stedim, Guxhagen, Germany). For 5-L runs, the seed culture (125–250 mL) was inoculated into 2.5 L of LSM, and fermentation was conducted in a CelliGen 310 bioreactor with a 7.5-L vessel (Eppendorf, New York, USA), controlled by the Eppendorf Bio Command software. Cell expansion was continued at 30 °C with a dissolved oxygen (DO) set point of 30%. After 19 ± 2 h of growth, a dissolved oxygen spike was observed on the trend chart, which indicates glycerol depletion. A fed-batch was initiated with 50% glycerol at a feed rate of 15 mL/L/h for 6 h to further expand biomass. During the last hour of the fed-batch phase, pH was adjusted to 6.5 using 14% NH4OH, while the temperature was adjusted to 25 °C. When a glycerol fed-batch was not included in the fermentation process, the pH and temperature were adjusted to the desired value during the first hour of induction. After the fed-batch phase, methanol induction was initiated; the total induction time was approximately 68–72 h. Biomass was removed by centrifugation at 12,227×g for 30 min at 4 °C before the supernatant was filtered through 0.45-μm polyethersulfone (PES) filters stored at −80 °C until purification.
Purification overview of three processes
The fermentation supernatant (FS) was removed from −80 °C and thawed at 22 °C for 4–6 h. Three purification processes were performed with 1-L FS aliquots (Fig. 1b). In process 1, the RBD219-N1C1 protein was captured from the FS using hydrophobic interaction chromatography (HIC), concentrated by ultrafiltration/diafiltration (UFDF), and polished using size exclusion chromatography (SEC). In process 2, the RBD219-N1C1 protein was captured using HIC, buffer-exchanged (UFDF), and polished using anion-exchange chromatography (AEX). Finally, in process 3, the FS was buffer-exchanged using UFDF before the target protein was captured using cation-exchange chromatography (CEX), buffer-exchanged (UFDF), and polished using AEX.
UFDF (ultrafiltration and diafiltration)
Two types of devices were used for UFDF, a centrifugal concentrator, and a flat sheet membrane, depending on the target volume. For process 1, Amicon centrifugal concentrator, with a 10 kDa molecular weight cutoff (MWCO) (MilliporeSigma, Burlington, USA) was used to concentrate the HIC elution pool (2050×g at 4 °C). This allowed concentration to the small volume needed for SEC. For process 2, a flat sheet Pellicon XL Cassette with a Biomax 5 membrane (5 kDa MWCO) and a Labscale TFF System (MilliporeSigma, Burlington, USA) were used to concentrate the HIC elution pool 8-fold, followed by diafiltration with 4 diavolumes of 20 mM Tris-HCl (pH 7.5) and 100 mM NaCl. A crossflow was kept at 25 mL/min over a 0.005-m2 membrane area throughout the entire process with an average transmembrane pressure (TMP) of ~15 psi. For process 3, a flat sheet Pellicon 2 Mini Cassette with a Biomax 5 membrane (MilliporeSigma, Burlington, USA) was used for the first UFDF (UFDF-1) to concentrate the FS 4-fold followed by diafiltration with 4 diavolumes of 20 mM sodium citrate (pH 4.2) and 10 mM NaCl. A crossflow was kept constant at 200 mL/min over a 0.1-m2 membrane area throughout the entire process with an average TMP of ~8 psi. For the UFDF-2, the CEX elution pool was concentrated 4-fold followed by diafiltration with 4 diavolumes of 25 mM Tris-HCl (pH 7.2) and 5 mM NaCl using the Pellicon XL Cassette as described for process 2.
Hydrophobic interaction chromatography
In processes 1 and 2, HIC was used to capture RBD219-N1C1 proteins from the FS. Ammonium sulfate salt was added to the FS to a final concentration of 1 M (w/v), and the pH was adjusted to 8.0. The FS was filtered through a 0.45-μm PES filter unit and loaded on a 112-mL Butyl Sepharose High-Performance column (4.4 cm diameter and 7.4 cm bed height) at a 20 mL/min flow rate. The column was washed with 1 M ammonium sulfate in 30 mM Tris-HCl (pH 8.0). Bound proteins were eluted with 0.4 M ammonium sulfate in 30 mM Tris-HCl (pH 8.0).
Size exclusion chromatography
Five milliliters of the concentrated HIC elution pool was loaded on a HiLoad 16/600 Superdex 75 prep-grade column (Cytiva, Marlborough, USA), pre-equilibrated with 20 mM Tris-HCl (pH 7.5) and 150 mM NaCl, and eluted at a flow rate of 1 mL/min. The SEC elution pool was aseptically filtered using a 0.2-μm PES filter in a biosafety cabinet and stored at −80 °C until usage.
Ion exchange chromatography
In process 3, RBD219-N1C1 was captured using CEX. The Pellicon 2 retentate pool in 20 mM sodium citrate (pH 4.2) and 10 mM NaCl was loaded on a 50-mL CM Sepharose Fast Flow column (2.6 cm diameter and 9.3 cm bed height) at a 10 mL/min flow rate. The column was washed with 20 mM sodium citrate (pH 4.2) and 10 mM NaCl. Bound proteins were eluted in 20 mM sodium citrate (pH 6.6) and 10 mM NaCl.
In processes 2 and 3, RBD219-N1C1 was polished using a negative capture AEX. The Pellicon XL retentate pool was loaded on a HiPrep Q Sepharose XL 16/10 column (Cytiva, Marlborough, USA) that was pre-equilibrated with 20 mM Tris-HCl (pH 7.5) and 100 mM NaCl for process 2, and 25 mM Tris-HCl (pH 7.2) and 5 mM NaCl for process 3. The flowthrough from AEX was collected, aseptically filtered using 0.2-μm PES filters in a biosafety cabinet, and stored at −80 °C until usage. NaCl (95 mM) was added to the final purified proteins from process 3 prior to storage in 25 mM Tris-HCl (pH 7.2) and 100 mM NaCl.
Protein yield and purity determination by quantitative SDS-PAGE
In-process samples taken at each purification step were loaded on either 14% Tris-glycine gels or 4–12% Bis-Tris gels to determine the concentration and purity of the various RBD219-N1C1 samples. Purified RBD219-N1C1 proteins of known concentrations were used as standards. After SDS-PAGE, gels were stained with Coomassie blue and scanned with a GS-900 densitometer (Bio-Rad, Hercules, USA). Gel images were processed with Image Lab software (Bio-Rad, Hercules, USA) to create a standard curve and determine protein concentration and purity.
Western blot analysis was performed to detect RBD219-N1C1 as well as P. pastoris host cell protein (HCP). Five micrograms of purified protein was run on 14% Tris-Glycine gels under non-reducing and reducing conditions to detect RBD219-N1C1 and HCP, respectively. Proteins in gel were transferred to PVDF membranes and blocked with 5% dry milk in PBST (1× PBS with 0.05% Tween-20). RBD219-N1C1 was detected using a rabbit monoclonal antibody against the SARS-CoV-2 spike S1 protein (Sino Biological, Beijing, China; Cat#: 40150-R007) and goat anti-rabbit IgG secondary antibodies conjugated with horseradish peroxidase (Invitrogen, Carlsbad, USA; Cat#: G21234). HCPs were detected using an anti-P. pastoris:HRP conjugate (2G) solution (Cygnus, Southport, USA; Cat#: F641-12). The blots were developed using ECL Prime Substrate System (Cytiva, Marlborough, USA).
Size Exclusion Chromatography-High Performance Liquid Chromatography
Waters® Alliance HPLC Separations Modules and Associated PDA Detectors were operated to analyze the size and purity of purified RBD219-N1C1 proteins. Fifty micrograms of the RBD219-N1C1 protein was injected into a Yarra SEC-3000 column (300 mm × 7.8 mm; catalog #: 00H-4513-K0) and was eluted in 20 mM Tris-HCl (pH 7.5) and 150 mM NaCl, at the flow rate of 0.6 mL/min. The elution of protein was confirmed by detecting the absorbance at 280 nm.
Dynamic light scattering
The size of the purified RBD219-N1C1 proteins was also analyzed by dynamic light scattering (DLS) (Chen et al. 2017, 2020). In short, RBD219-N1C1 was first diluted to 1 mg/mL with TBS, and approximately 40 μL of protein was then loaded into a clear bottom 384-well plate in four replicates to evaluate the hydrodynamic radius and molecular weight using the cumulant fitting on a Wyatt Technology DynaPro Plate Reader II.
Host cell protein quantification by ELISA
Yeast-expressed RBD219-N1C1 is N-glycosylated (Chen et al. 2021). To avoid any cross-reactivity from anti-P. pastoris HCP antibodies that recognize the N-glycans, which could result in an over-estimation of true HCP, we performed quantitative ELISAs with a second-generation anti-Pichia pastoris HCP ELISA Kit (Cygnus, Southport, USA; Cat#: F640) following the manufacturer’s instructions. This kit provides strips pre-coated with anti-P. pastoris HCP antibodies. Serially diluted RBD219-N1C1 was loaded onto the strips (HCP standards range from 0 to 250 ng/mL) in the presence of HRP-conjugated anti-P. pastoris antibodies. The strips were then incubated for approximately 3 h at room temperature followed by 4 washes. Finally, 100 μL of 3,3′,5,5′-tetramethylbenzidine (TMB) solution was added to react with the HRP-conjugated antibodies that were presented in the strip for 30 min prior to the addition of 100 μL of 1 M HCl to stop the reaction. The absorbance of 450 nm was measured in each well of the strip, and a linear standard curve was generated by plotting an “absorbance vs concentration” graph with the HCP standards to further calculate the HCP concentration present in the RBD219-N1C1 proteins.
Endotoxin levels in the purified RBD219-N1C1 samples were measured using the Endosafe Portable Testing System (Charles River Laboratory, Wilmington, USA). The purified protein was diluted 10-fold with Endosafe LAL reagent water, and 25 μL of diluted protein was loaded to each of the four wells of PTS20 Limulus amebocyte lysate Reagent Cartridge for the measurement as described in the literature (Charles River Laboratory, Wilmington, USA) (Jimenez et al. 2010).
In vitro ACE2 binding ELISA
The binding of RBD219-N1C1 to recombinant human ACE2 was evaluated using an ELISA procedure described previously (Chen et al. 2021). In short, 96-well ELISA plates were coated with 100 μL of 2 μg/mL RBD219-N1C1 overnight at 4 °C followed by blocking with PBST/0.1% BSA. One hundred microliters of serially diluted ACE2-hFc (LakePharma, San Carlos, USA; Cat#: 46672) was added to the wells and incubated at room temperature for 2 h, and the binding was detected by adding 100 μL 1:10,000 diluted HRP conjugated anti-human IgG antibodies (GenScript, Piscataway, USA; Cat#: A00166) with a 1-h incubation period at room temperature. Finally, 100 μL TMB substrate was provided to each well to react with HRP and the reaction was terminated with 100 μL of 1 M HCl. Absorbance at 450 nm was measured using an Epoch 2 microplate reader (BioTek, Winooski, USA).