Microbial strains and growth conditions
Saccharomyces cerevisiae VL6-48 (ATCC MYA-3666: MAT(alpha) his3–(Delta)200 trp1–(Delta)1 ura3-52 lys2 ade2-1 met14 cir (^{0})) was grown in YPD medium or complete minimal medium lacking histidine (Teknova) supplemented with 60 mg (hbox {L}^{-1}) adenine sulfate. Complete minimal media used for spheroplast transformation contained 1 M sorbitol. E. coli (Epi300, Epicentre) was grown in Luria Broth (LB) supplemented with appropriate antibiotics (chloramphenicol (25 mg (hbox {L}^{-1})) or ampicillin (50 mg (hbox {L}^{-1})) or gentamicin (20 mg (hbox {l}^{-1}))). P. tricornutum (Culture Collection of Algae and Protozoa CCAP 1055/1) or the P. tricornutum histidine auxotroph42 were grown in L1 medium without silica, with or without histidine (200 (hbox {mg,L}^{-1})), supplemented with appropriate antibiotics (Zeocin (50 mg (hbox {L}^{-1})) or nourseothricin (150 mg (hbox {L}^{-1}))), at 18(^{circ })C under cool white fluorescent lights (75 (upmu)E (hbox {m}^{-2} hbox {s}^{-1})) and a photoperiod of 16 h light:8 h dark. L1 media supplemented with nourseothricin contained half the normal amount of aquil salts. Long-term cultures of P. tricornutum with RBD expression plasmids were grown in L1 media supplemented with nourseothricin (for the wild-type strain) or with no added antibiotic (for the histidine auxotroph strain) to a density of 20 million cells (hbox {mL}^{-1}) and passaged by diluting 1:10 into fresh L1 media. Cultures were examined for contamination by visual light microscopy (Supplementary Figure S8).
Plasmid design and construction
All plasmids (Supplementary Table 1) were constructed using a modified yeast assembly protocol56,57. We cloned versions of the SARS-CoV-2 spike protein gene into the E. coli/P. tricornutum shuttle plasmid pPtGE3133. Primers are listed in Supplementary Table 5. We obtained SARS-CoV-2 expression plasmids (Krammer Lab, Icahn School of Medicine, Mount Sinai) that served as templates for PCR amplification of the human codon-optimized spike and receptor-binding domain coding regions for cloning into P. tricornutum expression plasmids. We also ordered synthetic constructs corresponding to the full-length spike protein gene and RBD from IDT-DNA that were codon-optimized for P. tricornutum. The nine initial constructs are listed in Supplementary Table 1 and a representative schematic can be seen in Fig. 1a. The constructs differed in the following ways: codon-optimization for human or P. tricornutum, full-length or RBD of the spike protein, version 1 or version 2 of the P. tricornutum HASP1 promoter (originating from two homologous P. tricornutum chromosomes, NBCI accession number XM_002181840.1, UNIPROT accession number PHATRDRAFT_47612). We also cloned a version of the full-length construct with two proline stabilizing mutations and a mutation of the furin cleavage site (RRAR) to an alanine, in addition to a stabilizing trimerization motif. The constructs were made with a promoter from the P. tricornutum 40SRPS8 (40S ribosomal protein S8) gene or from the HASP1 gene. All constructs contained the 40SRPS8 terminator downstream of the spike or RBD coding sequence. All constructs also included a histidine marker (PRA-PH/CH) expression cassette from the pPtPRAPHCH plasmid42 for selection and maintenance in a P. tricornutum histidine auxotroph strain. Plasmids encoding the P. tricornutum codon-optimized versions had the HASP1 promoter and the HASP1 secretory signal peptide, while plasmids encoding the human codon-optimized versions used the 40SRPS8 promoter and the native spike protein secretory signal. The plasmid constructs were assembled in yeast by co-transforming linear DNA fragments of the pPtGE31 plasmid backbone, the PRA-PH/CH expression cassette, the HASP1 or 40SRPS8 promoter, and the spike or RBD protein gene. Resultant yeast colonies were pooled, DNA extracted, and transformed into E. coli. Single E. coli colonies were grown and plasmid DNA isolated using standard plasmid mini-prep kits. Correct assembly was confirmed by restriction enzyme digests, by whole-plasmid sequencing using a MinION sequencer from Oxford Nanopore Technologies, and by Sanger sequencing of the spike and RBD expression cassettes at the London Regional Genomics Centre.
Plasmid validation
Plasmid DNA was extracted using the NEB miniprep kit (T1010L) and 400 ng of DNA was used as input for the rapid barcoding kit library prep (SQK-RBK004). Plasmids were then sequenced using R9.4.1 Flongle flow cells (FLO-FLG001) or R9.4.1 minION flow cells until approximately 200(times) coverage was obtained for each barcode based on an expected plasmid size of 20 kilobases. Basecalling was perfomed using Guppy v4.2.2 in high-accuracy mode (Oxford Nanopore Technologies). Reads were filtered by retaining only those near the expected plasmid length. Reads were then assembled using miniasm58. The assembly was then polished using minipolish59 and medaka (Oxford Nanopore Technologies). Polished assemblies were then compared to the expected sequence to determine if any mutations were present.
RNAseq analysis
Total RNA was extracted from 15 mL cultures with an (hbox {A}_{670}) of (sim) 0.6 to 0.7 by first crushing the algal cells in liquid nitrogen as follows. Cultures were centrifuged at 3000(times)g for 15 mins at 4 (^{circ })C. The pellet was resuspended in (sim) 100 to 500 (upmu)L TE pH 8.0 and added dropwise to a mortar (pre-cooled at − 80 (^{circ })C) filled with liquid nitrogen. The frozen droplets were ground into a fine powder with a mortar and pestle, being careful to keep the cells from thawing by adding more liquid nitrogen when necessary. The frozen ground powder was transferred to a new clean 1.5 mL microfuge tube and stored at − 80 (^{circ })C. RNA was extracted from 50 to 100 mg of frozen ground powder with the Monarch Total RNA Miniprep Kit (T2010S) following the plant protocol. The RNA was stored in TE pH 8.0 at (-80, ^{circ })C until use. Quantity and purity were measured by spectrophotometer, and RNA integrity was evaluated using a 1% pre-stained agarose gel run at 100 V for 30 min. RNA integrity was further evaluated using an Agilent Bioanalyzer. rRNA was depleted using the Vazyme Ribo-off plant rRNA depletion kit (N409). Sequencing libraries were then prepared for two different RNAseq experiments; the first for cells harbouring pSS1 and wild-type cells, and the second for clones of pSS1, pSS2 and pSS7. Libraries were sequenced at the London Regional Genomics Center using an Illumina NextSeq high output single end 75 run for the first experiment, and a high output single end 150 run for the second experiment. Reads were trimmed to 75 base pairs and aligned against the ASM15095v2 reference assembly and expected plasmid sequence using hisat260. Coverage was determined using Mosdepth61.
Diagnostic PCR assays
For direct PCR assays, colonies of P. tricornutum transformants were screened for the presence of the RBD gene using a Thermo Scientific Phire Plant Direct PCR Master Mix according to manufacturers instructions. PCR screens were performed using a forward primer located in the HASP1 promoter (DE5241) for P. tricornutum transformed with pSS1 and pSS2, or in the 40SRPS8 promoter (DE4130) for P. tricornutum transformed with pSS7. Reverse primers were positioned inside the RBD domain coding regions (DE5323, DE5326). Screening primers are listed in Supplementary Table 5.
Transfer of DNA to P. tricornutum
via conjugation from E. coli
Conjugations were performed as previously described19,33. Briefly, liquid cultures (250 (upmu)L) of P. tricornutum, adjusted to a density of 1.0 (times 10^{8}) cells (hbox {mL}^{-1}) using counts from a hemocytometer, were plated on (frac{1}{2} times) L1 1% agar plates with or without histidine (200 mg (hbox {L}^{-1})), and grown for four days. L1 media (1.5 mL) was added to the plate and cells were scraped and the concentration was adjusted to 5.0 (times 10^{8}) cells (hbox {mL}^{-1}). E. coli cultures (50 mL) were grown at 37 (^{circ })C to (hbox {A}_{600}) of 0.8–1.0, centrifuged for 10 mins at 3000(times)g and resuspended in 500 (upmu)L of SOC media. Conjugation was initiated by mixing 200 (upmu)L of P. tricornutum and 200 (upmu)L of E. coli cells. The cell mixture was plated on (frac{1}{2} times) L1 5% LB 1% agar plates, incubated for 90 mins at 30 (^{circ })C in the dark, and then moved to 18 (^{circ })C in the light and grown for 2 days. After 2 days, L1 media (1.5 mL) was added to the plates, the cells scraped, and 300 (upmu)L (20%) plated on (frac{1}{2} times) L1 1% agar plates supplemented with Zeocin 50 mg (hbox {L}^{-1}) or nourseothricin 100 mg (hbox {L}^{-1}). Colonies appeared after 7–14 days incubation at 18 (^{circ })C with light.
Measuring growth and eGFP production of P. tricornutum cultures
Phaeodactylum tricornutum cultures were adjusted to an (hbox {A}_{670}) of 0.05 in L1 media made without phosphate, nitrate, or iron. Cultures were then washed by centrifugation for 10 mins at 3000(times)g followed by resuspension in fresh L1 media without phosphate, nitrate, or iron. Phosphate, nitrate, and iron stock solutions were then used to adjust cultures to the follow conditions: full L1 or L1 with 5% phosphate, 5% nitrate, 5% iron, 5% phosphate and 5% iron, 5% phosphate and 5% nitrate, or 5% nitrate and 5% iron. Cultures were grown at 18 (^{circ })C under cool white fluorescent lights (75 (upmu)E (hbox {m}^{-2} hbox {s}^{-1})) and a photoperiod of 16 h light:8 h dark for 28 days, and absorbance at 670 nm ((hbox {A}_{670})) or 750 nm ((hbox {A}_{750})) was measured every 48 h using an Ultrospec 2100 pro UV/vis spectrophotometer. Samples for fluorescence readings and western blots were taken every 4 days by centrifuging 700 (upmu)L of culture at 16,000(times)g for 15 min. Three 200 (upmu)L aliquots of the supernatant were pipetted into a clear bottom 96 well plate for fluorescence readings. Another 44 (upmu)L of supernatant was mixed with 22 (upmu)L of 3(times) SDS sample loading buffer (187.5 mM Tris-HCl (pH 6.8), 6% (w/v) SDS, 30% [v/v] glycerol, 150 mM DTT, 0.03% (w/v) bromophenol blue, 2% [v/v] (beta)-mercaptoethanol) and boiled at 95 (^{circ })C for 10 mins, after which 15 (upmu)L of boiled sample was analyzed by western blot. The pellet was resuspended in 50 (upmu)L of 3(times) SDS sample loading buffer and boiled at 95 (^{circ })C for 10 mins, after which 10 (upmu)L of boiled sample was analyzed by western blot. Fluorescence readings were taken in a Biotek Synergy H1 platereader at an excitation wavelength of 475 nm and emission wavelength of 515 nm. Fluorescence values obtained were subtracted from wildtype autofluorescence in the supernatant. Fluorescence values were converted to eGFP in (upmu)g (hbox {mL}^{-1}) using a standard curve generated using commercially available purified eGFP.
Photobioreactor conditions for growth of P. tricornutum
A 5 L photobioreactor system (Eppendorf Canada) was used for the growth of P. tricornutum at bench scale. Temperature was controlled in the photobioreactor at 18 (^{circ })C. Mixing was achieved with a single pitched-blade impeller at 100 rpm. A constant gas flow of 0.75 VVM was sparged into the reactor with a mix of 0.5% carbon dioxide and 99.5% air. The pH of the culture was controlled at 8.1 using a cascade with carbon dioxide from 0.5 to 5% v/v mix. Light was provided by continuous (24 h/day) full spectrum LED grow lights with 5 bulbs at a light intensity of approximately 50 (upmu)mol (hbox {m}^{-2} hbox {s}^{-1}). Samples were collected daily for optical density, cell count and compositional analysis. A 10% inoculum was used to attain a minimum cell density of 2 million cells (hbox {mL}^{-1}). The inoculum was cultured in an incubator (Innova S44i, Eppendorf, Hamburg, Germany) with a photosynthetic light bank containing LED lighting. Lighting for the inoculum was at an intensity of approximately 65 (upmu)mol (hbox {m}^{-2} hbox {s}^{-1}) with a cycling of 16 h on and 8 h off. Temperature was controlled at 18 (^{circ })C with orbital agitation at 100 rpm. pH was not controlled and no gas was sparged into the inoculum.
Compositional analysis of media
The concentration of dissolved phosphorus and iron was measured by Inductively Coupled Plasma-Optical Emission Spectrometry (ICP-OES). An Agilent 5110 (Agilent, USA) spectrometer ICP-OES equipped with a Seaspray concentric glass nebulizer (Agilent, USA) and an SPS 4 auto sampler was used. Argon (purity higher than 99.995%) supplied by Linde Canada was used to sustain plasma and, as carrier gas. The operating conditions employed for ICP-OES determination of iron were 1200 W RF power, 12 L (hbox {min}^{-1}) plasma flow, 1.0 L (hbox {min}^{-1}) auxiliary flow, 0.7 L (hbox {min}^{-1}) nebulizer flow, with radial view used for determination. The operating conditions employed for ICP-OES determination of phosphorus were 950 W RF power, 12 L (hbox {min}^{-1}) plasma flow, 1.0 L (hbox {min}^{-1}) auxiliary flow, 0.95 L (hbox {min}^{-1}) nebulizer flow, with axial view used for determination. The most sensitive lines free of spectral interference were used to determine emission intensities. The calibration standards were prepared by diluting a phosphorus and iron standards (Agilent, USA) in synthetic seawater and 1% (v/v) nitric acid. The calibration curve for phosphorus was in the range of 0.1–10 ppm and for iron was in the range of 0.05–5 ppm.
Protein extraction and purification
Phaeodactylum tricornutum cultures (5 L) were harvested during stationary growth phase and pelleted at 3000(times)g for 10 mins at 4 (^{circ })C. Cell pellets were resuspended in lysis buffer (50 mM Tris-HCl pH 7.4, 0.5 M NaCl, 10 mM imidazole, 0.1% Tween-20, 1 mM DTT, 1 (times) Protease inhibitor cocktail (Sigma)) and homogenized on an Emulsiflex C3 Homogenizer with 5 passes at 20,000 psi to lyse. Lysates were then sonicated with two 15 s pulses, resting on ice for 30 s between pulses. Sonicated lysates were centrifuged at 20,000(times)g for 30 mins at 4 (^{circ })C to pellet cell debris, and supernatants were collected in a new tube and stored on ice before purification using a 20 mL GE Healthcare HisPrep FF 16/10 Ni-sepharose column as follows.
All samples were run on an AKTA Pure FPLC system at 4 (^{circ })C. Ni-sepharose columns were first washed with 10 column volumes of (hbox {ddH}_{2}mathrm{O}), then equilibrated with 10 column volumes of lysis buffer. Supernatants from lysed cultures were run over equilibrated columns at a flow rate of 5 mL (hbox {min}^{-1}) and flowthrough was collected. Columns were washed with 10 column volumes lysis buffer at 5 mL (hbox {min}^{-1}), followed by 10 column volumes wash buffer (50 mM Tris-HCl pH 7.4, 0.5 M NaCl, 50 mM imidazole, 0.1% Tween-20, 1 mM DTT) at 5 mL (hbox {min}^{-1}). His-tagged proteins were eluted with 4 column volumes elution buffer (50 mM Tris-HCl pH 7.4, 0.5 M NaCl, 250 mM imidazole, 0.1% Tween-20, 1 mM DTT) at 1 mL (hbox {min}^{-1}), collecting 5 mL fractions. Samples (20 (upmu)L) of lysis supernatant, flowthrough, washes, and elution fractions were mixed with 10 (upmu)L of 3(times) SDS sample loading buffer (187.5 mM Tris-HCl (pH 6.8), 6% (w/v) SDS, 30% [v/v] glycerol, 150 mM DTT, 0.03% (w/v) bromophenol blue, 2% [v/v] (beta)-mercaptoethanol) and boiled at 95 (^{circ })C for 5 mins. Boiled samples (15 (upmu)L) were resolved on standard SDS-polyacrylamide gels (15%). Bands were visualized with Coommassie Brilliant Blue and destained with a solution of 40% methanol, 10% acetic acid.
Ni-sepharose elution fractions containing RBD protein were pooled and dialyzed at 4 (^{circ })C for 5 h in 1 L of IEX loading buffer (50 mM HEPES pH 8.0, 10 mM NaCl, 1 mM DTT), then overnight in 2 L of fresh IEX loading buffer. The dialyzed sample was then loaded onto a 5 mL SP HP HiTrap column at 0.5 mL (hbox {min}^{-1}) and flowthrough was collected. Columns were washed with 10 column volumes IEX wash buffer (50 mM HEPES pH 8.0, 25 mM NaCl, 1 mM DTT) at 2 mL (hbox {min}^{-1}). Bound proteins were eluted with IEX elution buffer (50 mM HEPES pH 8.0, 500 mM NaCl, 1 mM DTT) at 1 mL (hbox {min}^{-1}), collecting 1 mL fractions. Samples were analyzed on SDS-polyacrylamide gels as above.
IEX elution fractions containing RBD protein were pooled, concentrated using Pierce centrifugal protein concentrators (10 kDa cutoff), and loaded onto a Superdex 200 Increase 10/300 GL column (24 mL bed volume) followed by IEX loading buffer at 0.5 mL (hbox {min}^{-1}). Flowthrough was collected in 0.5 mL fractions. Samples were analyzed on SDS-polyacrylamide gels as above. Elution fractions containing RBD protein were pooled, then concentrated and buffer exchanged with PBS using Pierce centrifugal protein concentrators (10 kDa cutoff).
Western blots
Samples were resolved on standard SDS-polyacrylamide gels (15%) and electroblotted to a polyvinylidene difluoride (PVDF) membrane using a Trans-Blot Turbo Transfer System (BioRad, Hercules, CA, USA). Mammalian expressed RBD (HEK293-RBD) was purchased from Sino Biological, 40592-V08H. Membranes were incubated for 1 hour in blocking solution (3% bovine serum albumin (BSA), 0.1% Tween-20, 1(times) TBS) before adding anti-RBD primary antibody (Sino Biological, 40592-T62) at a 1:1000 final dilution, or anti-GFP primary antibody (Invitrogen, A-6455) at a 1:2500 final dilution, or anti-6His primary antibody (Invitrogen, MA1-21315) at a 1:1000 final dilution. Membranes were incubated overnight at 4 (^{circ })C, washed for 3 (times) 10 mins in washing solution (1% BSA, 0.1% Tween-20, 1(times) TBS), then incubated with anti-rabbit (Sigma, GENA9340) or anti-mouse (Amersham, NA931) horseradish peroxidase-linked secondary antibody for 2 h at 1:5000 final dilution in washing solution. Membranes were then washed in 1(times) TBS with 0.1% Tween-20 for 3 (times) 10 mins, followed by one wash for 10 mins in 1(times) TBS. Blots were developed using Clarity ECL western blotting Substrate (BioRad) following the manufacturer’s instructions and imaged with a ChemiDoc XRS+ System (Bio-Rad).
Mass spectrometry
Protein samples were resolved on 8% (full length spike) or 15% (RBD) SDS-PAGE gels. Bands were visualized with Coommassie Brilliant Blue and destained with a solution of 40% methanol, 10% acetic acid. Bands were excised and placed in 1.5 mL microfuge tubes with 500 (upmu)L of 1% acetic acid. Mass spectrometry analysis and peptide identification by ms/ms was performed at the SPARC BioCentre, Sick Kids Hospital, University of Toronto. Peptide data was downloaded from the SPARC BioCentre server and analyzed by the Scaffold software package using a FASTA file of the P. tricornutum and SARS-CoV-2 proteomes. The protein threshold was set at 90% and the peptide threshold at a 1% false discovery rate.
Endoglycosidase assays
Purified protein samples (20 ng algae-RBD or HEK293-RBD) were treated with PNGase F or Endo H (NEB) according to manufacturers instructions following the denaturing reaction conditions protocol in a total reaction volume of 20 (upmu)L. 10 (upmu)L of 3(times) SDS sample loading buffer was then added to each 20 (upmu)L reaction and boiled at 95 (^{circ })C for 5 mins. Boiled samples (15 (upmu)L) were resolved on standard SDS-polyacrylamide gels (15%) and analyzed by western blot.
ACE2-binding assay
ACE2-binding assays were performed on the algae-RBD and HEK293-RBD using a COVID-19 Spike-ACE2 Binding Assay Kit (RayBio, CoV-ACE2S2-2) according to manufacturers instructions.
Gold conjugation of algae-RBD
For the preparation of gold labelled algae-RBD, 40 nm colloidal gold particles (International Point of Care Inc) were adjusted to a pH of 9.45 (± 0.15) using 1.0 M potassium carbonate buffer. The pH-adjusted gold was then diluted in ultrafiltrated (hbox {H}_{2}mathrm{O}) to an optical density (OD) of 1.0 (peak absorbance observed between wavelength 350–540 nm). Purified algae-RBD (0.213 mg/mL) was conjugated by passive adsorption to gold at a final concentration of 6 (upmu)g/mL algae-RBD per mL of OD 1.0 gold. After incubation at ambient temperature (22–25 (^{circ })C) for 45 min, a solution of 10% bovine serum albumin and 1% polyethylene glycol comprising 5% of the total conjugate volume was added and incubated at ambient temperature for 30 mins under gentle mixing. After incubation the gold conjugate solution was centrifuged at 12,000(times g) for 30 mins. Without disturbing the pellet, the resulting supernatant was removed and discarded. The pellet was then resuspended in 20 mM Tris-HCl (pH 9.25) representing 80% of the initial conjugate volume. This process was repeated in two consecutive centrifugation and resuspension steps. A final centrifugation step was performed and the supernatant discarded. The pellet was resuspended in the remaining supernatant. The RBD labelled gold was then stored overnight at 2–8 (^{circ })C to observe aggregation. Following overnight incubation, algae-RBD gold conjugate was briefly sonicated in a water bath to disperse any gold aggregates. A sample of the gold conjugate was then diluted 1/50 in Tris-HCl pH 9.25 and the absorbance peak was measured. The final OD of the gold conjugate was determined to be 12.8.
Preparation and testing of LFA devices
A commercially available qualitative lateral flow COVID-19 serology test (Lumivi) was adapted to evaluate the algae-RBD that was lyophilized onto polyester pads. Nitrocellulose membrane (Millipore) was striped at a test line location with an anti-human IgG antibody (BiosPacific Inc.) and at a control line location with a Goat anti-Rabbit polyclonal antibody (Cedarlane Inc.). Following lamination onto adhesive lined polyester backing cards, the resulting cards were cut into cut into 5.5 mm strips and placed into the Lumivi commercial housing, packaged in foil pouches with desiccant and stored at ambient temperature for subsequent testing. For comparison purposes a SARS-CoV-2 Spike protein RBD domain expressed in a human cell line (DAGC174 Creative Diagnostics) was similarly conjugated to colloidal gold as described previously and evaluated in the adapted lateral flow assay. An 15 (upmu)L aliquot each test sample (plasma) was applied to the device sample well followed by three drops of 0.1M PBS/Tween running buffer (approximately 150 (upmu)L). After a 15-min incubation at ambient temperature the test was visually interpreted for the presence of purple/red lines at both the Test and Control marker areas within the devices read window. The devices were scanned with an optical reader (i-Lynx) and values of 0.055 reflectance units were considered to be visually detectable by untrained operators and are positive. Values below 0.055 reflectance units are scored as negative. A positive control line (indicating proper sample flow within the prototype device) was required before device interpretation could be made.
