Strains and cultivations. Wild-type Komagataella phaffii (NRRL Y-11430) was modified to express variants of P[4], P[6], or P[8] under control of the AOX1 promoter using a commercial vector (pPICZ A, Thermo Fisher Scientific). Modified vectors containing markers for selection on G418 (Thermo Fisher Scientific) or Nourseothricin (Gold Biotechnology) were also used to generate a single strain expressing all three serotypes.
Strains for initial characterization were grown in 3 mL culture in 24-well deep well plates (25 °C, 600 rpm) using complex medium (BMGY-Buffered Glycerol Complex Medium, Teknova) supplemented to 4% (v/v) glycerol. After 24 h of biomass accumulation, cells were pelleted and resuspended in BMMY (Buffered Methanol Complex Medium, Teknova) containing 1.5% (v/v) methanol. Samples for RNA-seq and ribosome profiling were collected after 16 h growth in BMMY.
Strains were additionally cultivated in 200 mL culture in 1 L shake flasks to generate material for non-clinical studies and in InSCyT[35] bioreactors for demonstration of end-to-end manufacturing. At both scales, cells were grown in rich defined medium[51] supplemented to 4% (v/v) glycerol for biomass accumulation or 5% (v/v) methanol for production. In the bioreactor, temperature, pH, and dissolved oxygen were maintained at 25⁰C, 6.5, and 25%, respectively.
Transcriptome analysis. RNA was extracted and purified according to the Qiagen RNeasy kit (cat #74104) and RNA quality was analyzed to ensure RNA Quality Number > 7. RNA libraries were prepared using the Roche KAPA HyperPrep kit and sequenced on an Illumina NextSeq to generate 40-nt paired-end reads.
Sequenced reads were aligned and quantified using STAR v2.5.3a[52] and RSEM v1.3.0[53]. Expression was visualized using log2(Fragments per Kilobase of Transcript per Million Mapped Reads + 1) values. Differential gene expression was analyzed using the DESeq2 package in R starting from gene integer counts and including log-fold-change shrinkage. Gene set enrichment analysis (GSEA) was performed in R using the fgsea package.[54] Gene expression data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE159325.
Ribosome footprinting. Cell samples from cultures expressing P[4] or P[6] were transferred to a shake-flask containing 500 mL of pre-warmed fresh BMMY media to achieve an OD600 = 0.5. Cell harvest, cell lysis, and recovery of ribosome footprints were performed as described previously.[55] Footprints were dephosphorylated and cDNA libraries were prepared using the SMARTer smRNA-Seq Kit (Clontech), with an additional rRNA depletion step prior to PCR amplification, all as described previously.[55]
Libraries were sequenced on an Illumina NextSeq to generate 75-nt paired-end reads. The reverse reads were of low quality as expected from the library preparation kit and were not used in analysis. For the forward reads, the first 3 nucleotides and poly(A) sequences were removed from raw FASTQ reads using Cutadapt v.1.16[56] with options –u 3 -a A{100} –m 12. The trimmed reads were then aligned to a target consisting of K. phaffii rRNA sequences with bwa v0.7.16a. Unaligned reads were extracted with samtools v1.5 and FASTQ files were regenerated with bedtools v.2.26.0 for further analysis. Reads were then aligned to the wild-type K. phaffii genome plus the P[4] or P[6] expression cassettes using STAR v.2.5.3a[52] with the --alignIntronMax 600 option. Secondary alignments were removed using samtools v.1.5 and the primary alignment bam files were used as input to plastid v.0.4.8.[57] Plastid was run according to the author’s documentation and the commands used are available in the file plastid_processing.sh, which can be found in the Github repository jrbrady / PlastidCommands. The read sizes considered ranged between 26 and 31 nt, and an offset of 13 nt was used for all lengths except 26 nt, for which a 12 nt offset was used. Ribosome profiling data have been deposited in NCBI’s Gene Expression Omnibus and are accessible through GEO Series accession number GSE159336.
Experimental design. The goal of the animal study was to compare engineered antigens manufactured in K. phaffii to the original antigens manufactured in E. coli. We studied nine groups of five animals each, including a group for each of the original antigens (3), a group for each of the engineered antigens (3), one group of the original P[4] antigen manufactured in K. phaffii, one group receiving a placebo (adjuvant + buffer), and one group receiving trivalent rotavirus vaccine (Table S3).
Immunization of guinea pigs. Groups of 5 guinea pigs (Hartley, Sisi Al strain) weighing 400–500 g (National Institute of Agricultural Technology, INTA, laboratory animal production facility) were immunized intramuscularly three times at two-week intervals with 200 µl of aqueous vaccine containing 12 µg of the recombinant antigen (Table S3) adjuvanted with 2% Alhydrogel (Brenntag Biosector, Frederikssund, Denmark) (aluminum content 100 µg/dose). A group of animals immunized with a trivalent RVA vaccine formulated with a mix of HRVA Wa, DS1 and ST3 at 107 FFU/dose in Alhydrogel was used as positive control. A group of 5 animals receiving buffer in Alhydrogel was used as negative control. Blood samples were collected prior to each immunization as well as at 7, 14, and 28 days after the third immunization. All guinea pig experiments were conducted in compliance with the guidelines of the Institutional Animal Welfare Committee of the National Institute of Agricultural Technology, INTA, CICUAE protocol # 55.2018.
Neutralization and protein specific ELISA assays. Human rotavirus (HRVA) Wa GP[8], DS1 G2P[4] and ST3 G4P[6] adapted for cell culture were grown in MA104 cells using standard procedures. Viral titer was determined by cell culture immuno-fluorescence (CCIF) assay and expressed as FFU/ml as previously described.[58]
Virus neutralizing antibody titers (VNtAb) in serum samples were determined by a modified fluorescence focus reduction neutralization assay.[59] Serial two-fold dilutions of each serum sample were mixed with an equal volume of virus to a final concentration of 100 FFU. Serum-virus mixes were incubated at 37°C, 5% CO2 for 1 h, and then transferred onto confluent monolayers of MA-104 cells in 96 well plates. Plates were incubated at 37°C, 5% CO2 for 48 h. The monolayers were fixed with 70% acetone and the presence of residual virus in each well was evidenced by immuno-staining using a 2KD1 nanobody to rotavirus VP6 protein labeled with Alexafluor 488.[60] VNtAb of each serum sample is expressed as the inverse of the maximal dilution showing 60% fluorescence focus reduction.
Homotypic and heterotypic IgG antibody responses to vaccination were monitored using standard P2-VP8 antigen ELISA techniques. Individual serum samples were evaluated against the homologous antigen while the heterologous response was evaluated in serum pooled at every time point (T0 to T6). Thermo Fisher Maxisorp® high- binding 96-well ELISA plates were coated with each antigen diluted in carbonate buffer pH 9.6 for 24 hours at 4 °C. Plates were washed with wash buffer (PBS with 0.05% TweenTM 20), blocked with 200 µL of blocking buffer (PBS with 10% Skim milk), and incubated for one to four hours at 37 °C. Serum samples were diluted in serial four-fold dilutions, starting at 1:160, and 100 µL of each dilution was transferred to the assay plate, incubated for one hour at 37 °C and washed as described above. Plates were incubated for one hour at 37 °C with peroxidase labeled goat anti-guinea pig IgG antibody (Jackson) diluted 1:3000 in wash buffer. After four washes, 100 µL per well of detection reagent (H2O2, ABTS in citrate buffer, pH 5) was allowed to develop for 15 minutes in the dark at room temperature, and stopped with 50 µL of 5% SDS. Absorbance was measured at 405 nm with a SpectraMax® M2 microplate reader (Molecular Devices, LLC). The VNtAb of each serum sample was expressed as the inverse of the maximal dilution showing 60% fluorescence focus reduction. Serum that yielded no signal for any technique at 1:160 dilution was assigned a titer of 1:80.
Protein purification. Protein purification for non-clinical studies and end-to-end manufacturing was carried out on the purification module of the InSCyT system as described previously[35]. All columns were equilibrated in the appropriate buffer prior to each run. Product-containing supernatant was adjusted to pH 4.5 using 100 mM citric acid. The adjusted supernatant was loaded into a pre-packed CMM HyperCel column (1-mL or 5-mL) (Pall Corporation, Port Washington, NY), re-equilibrated with 20 mM sodium citrate pH 4.5, washed with 20 mM sodium phosphate pH 6.0, and eluted with 20 mM sodium phosphate pH 7.0, 100 mM NaCl. Eluate from column 1 above 15 mAU was flowed through a 1-mL pre-packed HyperCel STAR AX column (Pall Corporation, Port Washington, NY). Flow-through from column 2 above 15 mAU was collected.
Analytical assays for product characterization. Biolayer inferometry was performed using the Octet Red96 with streptavidin (SA) biosensors (Pall ForteBio, Fremont, CA), which were hydrated for 15 min in kinetics buffer prior to each run. Kinetics buffer consisting of 1X PBS pH 7.2, 0.5% BSA, and 0.05% tween 20 was used for all dilutions, baseline, and disassociation steps. Monoclonal antibodies specific to P[4], P[6], and P[8], respectively, were biotinylated using the EZ-Link Sulfo-NHS-LC-biotinylation kit according to manufacturer instructions (Thermo Fisher Scientific) and used in the assay at concentrations of 1.25, 1.7, and 1 µg/mL, respectively. Triplicate samples were loaded in a 96-well black microplate (Greiner Bio-One) at starting concentrations of 2, 14, or 5 µg/mL, respectively. Seven 1:1 serial dilutions and a reference well of kinetics buffer were analyzed for each sample. Association and disassociation were measured at 1000 rpm for 300 and 600 sec, respectively. Binding affinity was calculated using the Octet Data Analysis software v10.0 (Pall ForteBio), using reference subtraction, baseline alignment, inter-step correction, Savitzky-Golay filtering, and a global 1:1 binding model.
Wet cell weight (WCW) was determined as described previously.[35] Sample concentrations were determined by absorbance at A280 nm. Sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) was carried out under reducing and non-reducing conditions using Novex 12% Tris-Glycine Gels or Novex 16% Tricine Gels (Thermo Fisher Scientific, Waltham, MA) according to the manufacturer’s recommended protocol and stained using Instant Blue Protein Stain (Thermo Fisher Scientific, Waltham, MA). Isoelectric focusing was carried out as described previously.[35] Samples were analyzed for host cell-protein content using the Pichia pastoris 1st generation HCP ELISA kit from Cygnus Technologies (Southport, NC) according to the manufacturer’s recommended protocol. Samples were analyzed for residual host-cell DNA using the Quant-iT dsDNA High-Sensitivity Assay Kit (Invitrogen) according to the manufacturer’s protocol except the standard curve was reduced to 0–20 ng.
High performance liquid chromatography (HPLC) analysis was performed on an Agilent 1260 HPLC system equipped with a diode array detector and controlled using OpenLab CDS software (Agilent Technologies, Santa Clara, CA). Antigen concentration was determined using a PLRP-S column (2.1 × 150 mm, 300 Å, 3 µm) operated at 0.5 mL/min and 60⁰C (Agilent Technologies, Santa Clara, CA). Buffer A was 0.1% (v/v) TFA in water and buffer B was 0.1% (v/v) TFA, 0.5% (v/v) water in ACN. A gradient of 37–42% B was performed over 30 minutes; total method run time was 45 minutes. Sample injection volumes were 10 µL. Data analysis was completed using OpenLab CDS Data Analysis (Agilent Technologies, Santa Clara, CA).
Intact mass analysis was performed on a 6230B time of flight (TOF) LC-MS with a 1220 series HPLC (Agilent Technologies, Santa Clara, CA). Mobile phase A consisted of water with 0.1% formic acid, and mobile phase B was acetonitrile with 0.1% formic acid. About 20 pmol of each sample was injected, bound to a ZORBAX 300SB C3 column (Agilent Technologies), desalted, and subjected to electrospray ionization. The LC gradient consisted of 20–70% B over 1 min at a flow rate of 1.5 mL/min. Elution of proteins was monitored using the absorbance signal at 214 nm. 50 µL of isopropanol was injected after each sample to control sample carry-over. The typical electrospray ionization parameters consisted of: 290 °C gas temperature, 4000 V Vcap, 2000 V nozzle, and 275 V fragmentor voltage. Mass spectra were collected from 700–2800 m/z at a scan rate of 1 spectra/sec. MS spectra were processed using MassHunter Qualitative Analysis software (v B.07.00, Agilent Technologies) with deconvolution range of 10–50 kDa, using 1 Da mass step.