Development of a scalable single process for producing SARS-CoV-2 RBD monomer and dimer vaccine antigens

We have developed a single process for producing two key COVID-19 vaccine antigens: SARS-CoV-2 receptor binding domain (RBD) monomer and dimer. These antigens are featured in various COVID-19 vaccine formats, including SOBERANA 01 and the licensed SOBERANA 02, and SOBERANA Plus. Our approach involves expressing RBD (319-541)-His6 in Chinese hamster ovary (CHO)-K1 cells, generating and characterizing oligoclones, and selecting the best RBD-producing clones. Critical parameters such as copper supplementation in the culture medium and cell viability influenced the yield of RBD dimer. The purification of RBD involved standard immobilized metal ion affinity chromatography (IMAC), ion exchange chromatography, and size exclusion chromatography. Our findings suggest that copper can improve IMAC performance. Efficient RBD production was achieved using small-scale bioreactor cell culture (2 L). The two RBD forms - monomeric and dimeric RBD - were also produced on a large scale (500 L). This study represents the first large-scale application of perfusion culture for the production of RBD antigens. We conducted a thorough analysis of the purified RBD antigens, which encompassed primary structure, protein integrity, N-glycosylation, size, purity, secondary and tertiary structures, isoform composition, hydrophobicity, and long-term stability. Additionally, we investigated RBD-ACE2 interactions, in vitro ACE2 recognition of RBD, and the immunogenicity of RBD antigens in mice. We have determined that both the monomeric and dimeric RBD antigens possess the necessary quality attributes for vaccine production. By enabling the customizable production of both RBD forms, this unified manufacturing process provides the required flexibility to adapt rapidly to the ever-changing demands of emerging SARS-CoV-2 variants and different COVID-19 vaccine platforms.


Further details for generating clones for RBD (319-541)-His6 production
To produce lentiviral particles carrying the RBDsint201 genetic construct, we transfected human embryonic kidney 293T (HEK-293T) cells (ATCC, CRL-3216) with the RBDsint201 genetic construct and auxiliary plasmids pLPI, pLPII, and LP/VSV-G (Invitrogen, USA) using 25 kDa linear polyethyleneimine (Polysciences, USA) as the transfecting agent.The mass ratio of the four plasmids used during transfection was 2:1:1:1.After 72 h of growth at 37°C in a 5% CO2 atmosphere, we collected the resulting cell culture supernatant, and lentiviral particles were concentrated through precipitation with polyethylene glycol (Sigma Aldrich, USA).We measured the viral titers in the final preparation using the DAVIH Ag p24 ELISA kit (LISIDA, Cuba) according to the manufacturer's instructions, and defined 0.65% of the total virus concentration as infectious virus particles.To transduce CHO-K1 cells (ATCC, CCL-61) with the lentiviral particles, we cultured the cells in 96well plates at a density of 5000 cells in 200 µl/well using DMEM/F-12 (Dulbecco's Modified Eagle Medium/Nutrient Mixture F-12) supplemented with 5% fetal bovine serum (FBS) (HyClone, GE Healthcare, USA) for 24 h at 37 ºC in a 5% CO2 atmosphere.Next, we added lentiviral particles to each well at a multiplicity of infection of 800, along with polybrene at 10 µg/ml, and incubated the cells for 8 h.To stop the infection, we added 10 µl of FBS to each well, and allowed the cells to recover overnight under the same conditions.We repeated this procedure twice every 24 h to increase the transduction efficiency.Cells from some wells were grown in DMEM/F-12 supplemented with 10% FBS and blasticidine at 2 µg/ml for 72 h after being transduced three times (referred to as 3xtransduced).After recovering for 72 h, 3x-transduced cells from some wells were subjected to three additional cycles of lentiviral infection every 24 h before being grown in supplemented medium as described above.These cells are referred to as 6x-transduced.To select oligoclonal cell lines that produce high levels of biologically active RBD (319-541)-His6, we diluted the cells from each transduction well and grew them for 10 days across five 96-well plates in DMEM/F-12 supplemented with 10% FBS and blasticidine at 2 µg/ml.We then subjected the cell culture supernatants to ELISA using polyvinyl chloride microtiter plates coated with 10 µg/ml of recombinant ACE2/human Fc fusion protein, which captured the RBD.To detect the RBD, we used anti-RBD monoclonal antibody (mAb) S1 (CIGB, Sancti Spiritus, Cuba) and anti-mouse IgG antibody conjugated to horseradish peroxidase.The oligoclonal cell lines that displayed the highest ELISA signals (absorbance at 490 nm) were selected and transferred to 24-well plates for further expansion, freezing, adaptation to growth in suspension in the absence of FBS, and initial expression studies.To ensure the production of consistent results, we cloned the same cell lines in parallel through limiting dilution in 96-well plates at a density of 0.8 cells/well in DMEM/F-12 supplemented with 5% FBS and blasticidine at 2 µg/ml.After two weeks, we tested the supernatants of wells showing cell growth (presumably clonal) using ELISA as previously described.The clones that exhibited the highest secretion of RBD, as determined by ELISA, were expanded, frozen, and adapted for growth in suspension.To further evaluate their potential, we selected a panel of ten clones from each lentiviral transduction protocol (three rounds of transduction and six transduction cycles) for small-scale (7 ml) expression studies in shaking flasks (25 cm 2 ) (Greiner Bio-One GmbH, Germany).

Further details for small-scale purification of RBD
For IMAC chromatography, the Chelating Sepharose Fast Flow XL column was equilibrated with 20 mM sodium phosphate buffer (pH 7.4) containing 300 mM NaCl, 5 mM imidazole, and 0.1% Tween 20.After equilibration, the column was washed with the same buffer solution, except the imidazole concentration was increased to 10 mM.The RBD was then eluted with PBS (pH 9) containing 200 mM imidazole and 0.1% Tween 20.All IMAC steps were performed at a flow rate of 300 cm/h.For buffer exchange, the Sephadex G-25 column was equilibrated with 50 mM sodium phosphate buffer (pH 6) containing 30 mM NaCl and operated at a flow rate of 150 cm/h.A volume of RBDcontaining intermediate product corresponding to 20% of the column volume was loaded onto the Sephadex G-25 column.
For cation exchange chromatography, the SP Sepharose Fast Flow column was pre-equilibrated with 50 mM sodium phosphate buffer (pH 6) containing 30 mM NaCl.The flow rate was 150-200 cm/h.The bound RBD species were eluted using 50 mM sodium phosphate buffer (pH 6) containing 200 mM NaCl.
For SEC-HPLC, the Superdex 200 pregrade column was pre-equilibrated with PBS and eluted at a flow rate of 30 cm/h.Supplementary Figure S2.On the formation of the RBD (319-541)-His6 dimer assembly.(A) Protein transfer from gel to membrane was confirmed by Ponceau red staining, which visualizes the total protein load.(B) Western blot was revealed with ACE2-Fc and anti-human (γ-chain specific)-HRP.Lane 1, CHO-K1 cell culture supernatant with 50 µM CuSO4; lane 2, same as lane 1 but without CuSO4; lanes 3-6, decreasing amounts (120, 90, 60 and 30 µg) of total intracellular proteins; lane 7, RBD dimer control; lane 8, RBD monomer control; lane 9, blank; lane 10, molecular mass marker.(C) Densitometric analysis of monomeric and dimeric RBD (as detected by Western blot) and β-actin, each relative to total protein in lanes 3-6.The densitometric analysis showed that there was agreement between the area corresponding to β-actin bands and the area corresponding to total protein loading per lane.A similar behavior was observed for the RBD monomer.Despite the increasing amount of intracellular protein loading in the lanes, an RBD dimer band was not detected in any of the samples.S3.Breakthrough curve for RBD adsorption on IMAC using purified RBD (328-533)-His6 (on the left panel) and supernatant from 39-3x culture (on the right panel) at two different residence times (red, 2 min; blue, 6 min).Column, XK16/20 packed with 30 ml of Chelating Sepharose Fast Flow.Equilibration buffer, 20 mM sodium phosphate, 500 mM NaCl, and 5 mM imidazole, pH 7.4.The dynamic binding capacity of Ni(II)-loaded IMAC gel matrix was found to be 51 mg/ml (at 2 min) and 58 mg/ml (at 6 min) when purified RBD was used as the sample to load the column.However, when the cell culture supernatant containing RBD was used as the sample, the dynamic binding capacity was reduced by 100-fold to 0.43 mg/ml (at 2 min) and 0.55 mg/ml (at 6 min).lanes 2, 4, 6 and 8) and G25-mediated buffer exchange (lanes 3, 5, 7 and 9).The copper sulfate concentration for each experimental condition is shown at the bottom of the figure.C) SEC-HPLC of IMAC elution samples with copper at different concentrations.SDS-PAGE analysis (lanes 2, 4, and 5) and SEC-HPLC profiling revealed that when copper sulfate concentrations exceeded 50 µM, larger proteins with a molecular mass of over 97 000 Mr were eluted from the IMAC gel matrix.This could be attributed to the capture of highly oxidized and aggregated RBD species or other non-specific protein species present in the original loaded material.These findings support the use of 50 µM as the appropriate molarity of copper sulfate for conditioning the cell culture supernatant containing RBD prior to IMAC purification, ensuring that larger, unwanted protein species are not eluted and recovered during the purification process.In fact, using 50 µM copper sulfate, we were able to obtain RBD with a higher purity level (76.87%) compared to when the material was adjusted to concentrations of 100 µM (53.24%) or 200 µM (44.09%) of copper sulfate (data not shown).

Supplementary Tables
Supplementary Table S1.Mean values of relevant variables characterizing each stage of the perfusion process described in Fig. 5 (primary text).Xv, concentration of viable cells; D, dilution rate; CSPR, cell-specific perfusion rate; qRBD, RBD (319-541)-His6 specific production rate; qGlc specific glucosa uptake rate; qO2, specific oxygen uptake rate; VP, volumetric productivity; dRBD, RBD (319-541)-His6 dimer; d, days.The values are averaged over six to eight replicates, and the standard deviations are also shown.Supplementary Table S3.Detection of RBD (319-541)-His6 disulfide bonds by ESI-MS, in six independent 500-liter batches.The "code" column refers to the cysteine residues forming the disulfide bond, and the "theor m/z" column refers to the theoretical mass-to-charge ratio below which the experimental value was found for each batch of RBD.Nd, not determined.

Figure S4 .
Effect of adding copper sulfate to clarified cell supernatant containing RBD (328-533)-His6 on subsequent IMAC separation.(A) Yield of RBD eluted from the IMAC column after application of monomeric RBD in cell supernatant supplemented with different concentrations of copper sulfate (50, 100 and 200 µM).It is worth noting that 50 µM of CuSO4 is the concentration used in the culture medium during the culture of the CHO cells producing RBD (319-541)-His6.Notably, the addition of copper sulfate resulted in an increase in RBD recovery, indicating that this approach can enhance the efficiency of the IMAC separation process.(B) SDS-PAGE of IMAC elution samples ( Figure S5.Purification of RBD (319-541)-His6 from culture supernatant (clone 39-3x).(A) RBD-containing culture supernatant was conditioned by adding 50 µM CuSO4 and subjected to IMAC chromatography.This was followed by a desalting step with Sephadex G25 and then SP cation exchange chromatography.The purification samples were analyzed by analytical SEC-HPLC (B) and SDS-PAGE (C).After cation exchange chromatography, the species of interest (RBD monomer [D] and RBD dimer [E]) were separated by SEC chromatography.O, oligomer; D, dimer; M, monomer.Supplementary Figure S6.Proportion of RBD (319-541)-His6 oligomers, dimers, and monomers prior to the final purification step.Size exclusion-HPLC profile of RBD-containing clarified supernatant separated by cation exchange from three different stages of the perfusion process: A, stage 1; B, stage 2; and C, stage 3.The materials from stage 1 and 3 were free of copper ions.O, peak of RBD oligomers; D, peak of RBD dimers; M, peak of RBD monomers.Supplementary Figure S9.Surface plasmon resonance (SPR) of RBD (319-541)-His6 species interacting with the ACE2 receptor immobilized on a protein A sensor chip.(A) SPR sensorgrams showing one replicate experiment of the RBD monomer dissolved in phosphate-buffered saline (PBS), pH 7.2.(B) SPR sensorgrams showing one replicate of the RBD dimer in PBS, pH 7.2.Each colored curve represents the indicated protein concentration on the right side.The black curves show the fit of the data to a 1:1 Langmuir binding model, yielding chi-squared values below 1.66.

Table S2 .
ESI-MS spectral peak assignments for independent batches of trypsindigested RBD (319-541)-His6.To facilitate identification, RBD samples were treated with Nethylmaleimide to ensure detection of native S-S bonds.The samples were also N-deglycosylated to improve the efficiency of proteolytic digestion and confirm the sequence of protein regions containing N-glycosylation.Tryptic digestion peptides were assigned based on agreement between theoretical and experimental monoisotopic m/z values.Nd, not determined.

Table S5 .
The glycan structures of NP-HPLC chromatogram peaks were assigned using Glucose Unit (GU) measurements.The GU values for each peak were determined by calibrating with a glucose homopolymer ladder.The abundance of each specific glycan structure was quantified using the % peak area.RT, retention time.