For the rapid production of influenza vaccine antigens in unlimited quantities, a transition from conventional egg-based production to cell-based and recombinant systems is required. The need for higher-yield, lower-cost, and faster production processes is critical to provide adequate supplies of influenza vaccine to counter global pandemic threats. In this study, recombinant hemagglutinin proteins of influenza virus were expressed in the microalga Schizochytrium sp., an established, fermentable organism grown in large scale for the manufacture of polyunsaturated fatty acids for animal and human health applications. Schizochytrium was capable of exporting the full-length membrane-bound proteins in a secreted form suitable for vaccine formulation. One recombinant hemagglutinin (rHA) protein derived from A/Puerto Rico/8/34 (H1N1) influenza virus was evaluated as a vaccine in a murine challenge model. Protective immunity from lethal challenge with homologous virus was elicited by a single dose of 1.7, 5 or 15 µg rHA with or without adjuvant at survival rates between 80–100%. Full protection (100%) was established at all dose levels with or without adjuvant when mice were given a second vaccination. These data demonstrate the potential of Schizochytrium sp. as a platform for the production of recombinant antigens useful for vaccination against influenza.
Citation: Bayne A-CV, Boltz D, Owen C, Betz Y, Maia G, Azadi P, et al. (2013) Vaccination against Influenza with Recombinant Hemagglutinin Expressed by Schizochytrium sp. Confers Protective Immunity. PLoS ONE 8(4): e61790. doi:10.1371/journal.pone.0061790
Editor: Paul G. Thomas, St. Jude Children's Research Hospital, United States of America
Received: November 30, 2012; Accepted: March 13, 2013; Published: April 23, 2013
Copyright: © 2013 Bayne et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This research was funded in part by the National Institutes of Health (NIH/NCRR)-funded grant entitled “Integrated Technology Resource for Biomedical Glycomics” (grant no. 8P41GM103490) to the Complex Carbohydrate Research Center and by Royal DSM. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have read the journal's policy and have the following conflicts: DSM owns or has applied for several patents related to aspects of this study. The following are the most relevant - US Patent numbers 7,732,170; 7,001,772; 8,003,772 and US Patent Application numbers 12/980,321; 12/980,322; 12/724,403; 12/980,319. The authors confirm that this intellectual property does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials. The following authors are currently employed by DSM. a. Anne-Cécile Bayne b. Carole Owen c. Yelena Betz d. Goncalo Maia e. Ross Zirkle f. J. Casey Lippmeier None of the individuals above have any other relevant declarations. The submitted manuscript does not describe any marketed, commercial, or in-development product or process. The employment of these individuals by DSM does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Influenza is an infectious disease caused by a few ever evolving quasi-species of the family Orthomyxoviridae against which human vaccination was first reported in 1937 . Traditionally, influenza vaccines are created from inactivated or attenuated preparations of live virus cultured in chicken eggs. This approach suffers from several drawbacks. Most notably, it is a labor intensive process requiring 1 or 2 eggs per vaccine dose and no less than 6 months to scale-up for commercial manufacturing . For any given influenza strain included in the annual trivalent vaccine, virus production usually requires artificial re-assortant strains and adaption regimens for growth in eggs. However, these manipulations contribute to the six-month timeline of influenza vaccine production and often result in vaccine antigens which no longer represent a perfect match to those of the parent strain, resulting in a mismatched vaccine. It is recognized that a transition from egg-based production systems to flexible cell-based and recombinant systems is desirable to continue long-term expansion of influenza vaccination programs and to better respond to sudden pandemics. To address these concerns, several groups have produced vaccines using cell-based systems, either by infection of cultured cells with live virus or by expression of influenza proteins from recombinant hosts including; vertebrate-derived cell lines , insect cell lines , yeast , filamentous fungi , higher plants , and bacteria . Recombinant subunit vaccines are of particular interest as they can be used to eliminate many of the complications associated with currently available influenza vaccines (inactivated, split, and live-attenuated virus vaccines), having the potential to reduce scale-up periods to 12 weeks or less; half the time of egg-based systems . Of the two influenza envelope glycoproteins, neuraminidase (NA) and hemagglutinin (HA), the latter elicits the highest proportion of virus-neutralizing antibodies which correlate to protection , . Thus, HA has been the most popular target for recombinant expression using these alternative host cells, expression system platforms, and antigen delivery scaffolds . Of the several influenza subunit vaccines being investigated, the most developed is manufactured using baculovirus-transfected insect cells, and has successfully completed Phase III clinical trials for consequent USFDA approval . However, the acceptance of subunit vaccines composed of influenza envelope proteins has been confounded by limitations in antigen expression and presentation, glycosylation, and immune responses. To explore alternative approaches for the production of functional influenza antigens, this report investigates the expression and secretion of rHA using a novel, well-defined, commercially feasible, microalgal-based expression system.
Influenza HA binding to terminal sialic acids of host glycoproteins is required for viral entry into the host cell. HA is a type I fusogenic, membrane glycoprotein with an N-terminal signal sequence, a hydrophobic transmembrane anchor domain near its C-terminus, and a short cytoplasmic tail. The HA protein is synthesized as a precursor polypeptide (HA0), which folds and self-associates as non-covalently linked homotrimers in the endoplasmic reticulum, prior to transport through the Golgi apparatus to the plasma membrane. Each HA0 polypeptide is activated through cleavage by a host-encoded protease in the secretory pathway. The resulting polypeptides, HA1 and HA2, are linked by a single disulfide bond . The HA protein is subject to other post-translational modifications before it is exported in association with the viral envelope. N-glycosylation sites in both the HA1 and HA2 vary in location and frequencies among individual strains of influenza, but some are conserved and play roles in immune evasion, virion export, receptor binding and protein folding –. HA is also acylated at a cysteine residue in the C-terminal tail which facilitates fusion pore formation of infectious virions . In general, characteristics such as these influence the choice of transgenic host for expression of a vaccine antigen. Ideally, the required features of a target antigen will align with the known traits of a given host system. Preserving the critical, qualitative attributes of the HA protein may result in a more vigorous, multi-faceted immune response targeting multiple functional epitopes with enhanced neutralization potential.
Schizochytrium sp. is a robustly fermentable eukaryote, best known for its ability to accumulate large amounts of cellular triglyceride containing nutritionally important polyunsaturated fatty acids (PUFA); most notably, docosahexaenoic acid (DHA). The majority of Schizochytrium-derived triglyceride oil is used to fortify foods with DHA. Schizochytrium biomass has also found use in animal applications, most commonly as a blend in poultry or aquaculture feeds to increase levels of DHA in eggs or farmed seafood respectively. Ten tons per annum of triglyceride oil have been produced using Schizochytrium, after culture in volumes greater than 150,000 liters of fully defined media , . The organism is highly efficient at consuming nutrients in the presence of low dissolved oxygen levels, with a reported peak rate of 9 doublings per day, which translates to more than 200 g/l of dry biomass accumulated in less than 4 days . This productivity further translates to lower materials, energy and labor costs per unit mass produced in fermentors as compared to many other eukaryotic, cell based systems. A proprietary genome sequence and transformation system has been developed for Schizochytrium , which in prior studies has been used to create targeted gene disruptions via homologous recombination . Several dominant, selectable markers have also been used for Schizochytrium transformation ,  enabling a number of metabolic engineering projects centered on carbon metabolism and fatty acid biosynthesis –. Schizochytrium is not photosynthetic and is thus cultivated using fermentation methods. It is also characterized by bi-flagellate zoospores and by thin, non-retractable, filopodia-like extensions of the cytoplasm. These extensions complex to form net-like structures which are associated with an ectoplasmic net composed largely of exported polysaccharide , . As a recombinant host expression system these Schizochytrium characteristics; cell wall flexibility and motility, may contribute to the export or extrusion of complex membrane bound molecules such as influenza HA.
In the present study, properties of recombinant HA derived from the classic strain, H1N1 influenza A/Puerto Rico/8/34 (PR8), were characterized, including glycosylation and proteolytic cleavage. Mice vaccinated with the PR8 rHA antigen were protected against lethal challenge with homologous virus. Additionally, we demonstrated expression of recombinant hemagglutinins derived from a number of influenza strains using Schizochytrium.
Materials and Methods
Strains, Vector Constructs, and Transformations
All vectors were constructed either in whole or in parts by Blue Heron Biotechnologies (Bothell, WA) or DNA 2.0 (Menlo Park, CA) and supplied as purified frozen DNA. All HA protein coding regions were codon optimized for expression in Schizochytrium sp. ATCC 20888 using the proprietary algorithms of either DNA 2.0 (Menlo Park, CA, USA) or Blue Heron Biotechnologies (Bothell, WA, USA). Accession numbers for each of the vector constructs harboring various HA genes are listed in Table 1. Vector elements including promoters, terminators and selectable markers have been described elsewhere . Cultures of Schizochytrium sp. ATCC 20888 were transformed with supercoiled vector by particle bombardment as previously described  and as follows. Schizochytrium cultures were grown from cryostocks in 50 ml volumes of “M50–20” media consisting of 20 g/l glucose, 12.5 g/l NaCl, 2.5 g/l MgSO4·7 H2O, 0.4 g/l KH2PO4, 0.5 g/l KCl, 0.05 g/l CaCl2·2 H2O, 0.5% modified PB26 metals and 0.1% modified PB26 vitamins (v/v). Modified PB26 vitamins consisted of 0.5 µg/ml biotin, 100 µg/ml thiamine, and 500 µg/ml cyanocobalamin. Modified PB26 metals were adjusted to pH 8.0 and consisted of 6 g/l Na2-EDTA, 0.29 g/l FeSO4·7 H2O, 6.84 g/l H3BO3, 0.86 g/l MnCl2·4 H2O, 60 mg/l ZnCl2, 26 mg/ml CoCl2·6 H2O, 5 mg/l Na2MoO4·2 H2O, 2 mg/l CuSO4·5 H2O, and 52 mg/l NiSO4·6 H2O. PB26 stock solutions were filter sterilized separately and added to the broth after autoclaving. Glucose, KH2PO4, and CaCl2·2 H2O were each autoclaved separately from the broth ingredients. All media components and antibiotics were purchased from Sigma (St. Louis, MO). Cultures of Schizochytrium sp. were grown to log-scale before transformation with ∼5 µg of plasmid coated on M10 tungsten particles delivered via biolistic particle bombarder (BioRad, Hercules, CA, USA) using a 1100-psi rated burst disk and with the culture dish on the second shelf slot. Primary transformants were selected on solid M2B media  containing 15 g/l agar (VWR, West Chester, PA) and 50 µg/ml paromomycin. All primary transformants were manually transferred to grid-plates containing paromomycin before being used to inoculate a 50 ml culture of M50–20 for preparation of 40% glycerol cryostocks. Primary transformants were further verified by polymerase chain reactions directed to vector elements and performed on gDNA extracted as previously described .
Shake Flask and Fermentation Culture Conditions for HA Expression
Supernatants containing recombinant HA proteins were prepared from fermentors or baffled shake flask cultures. Shake flask cultures contained 50 ml of protein-free “SPFM” media, inoculated from a colony or M50–20 culture aliquot. SPFM media contained 13.62 g/l Na2SO4, 0.72 g/l K2SO4, 0.56 g/l KCl, 2.27 g/l MgSO4·7 H2O, 3 g/l (NH4)2 SO4, 0.19 g/l CaCl2·2 H2O, 3 g/l Na-glutamate, 21.4 ml/l 2-(N-morpholino)ethanesulfonic acid (MES) (100 mM, pH 6), 0.4 g/l KH2PO4, 50 g/l glucose, 2 ml/l of SSFM-PB26 trace metals, and 1 ml/l of SSFM-PB26 vitamins. SSFM-PB26 vitamin stock contained 0.16 g/l vitamin B12, 9.75 g/l thiamine, and 3.33 g/l Ca-pantothenate. SSFM-PB26 trace metals stock was set to pH 2.5 with HCl and contained 1 g/l citric acid, 5.15 g/l FeSO4·7 H2O, 1.55 g/l MnCl2·4 H2O, 0.965 g/l ZnSO4·7 H2O, 0.02 g/l CoCl2·6 H2O, 0.02 g/l Na2MoO4·2 H2O, 1.035 g/l CuSO4·5 H2O, and 1.035 g/l NiSO4·6 H2O. Trace metals, vitamins, glucose, KH2PO4, and CaCl2·2 H2O were all filter sterilized and added to the media after the remaining ingredients were autoclaved. Flask cultures were grown for 2–5 days at 27°C with constant, 200 rpm agitation before cell-free supernatants (CFS) were collected by centrifugation at 5000 g.
Supernatants containing PR8 HA obtained using 10 l fermentors were derived from cultures grown with modified SPFM media containing the same salts as above, with the exception of 6 g/l (NH4)2 SO4, 6 g/l Na-glutamate, and omitted MES. All vessels were autoclaved for 2 hours at 120°C and to these, 50 g/l glucose, 0.8 g/l KH2PO4, 0.19 g/l CaCl2·2 H2O, 4 mg/l citric acid, 20.6 mg/l FeSO4·7 H2O, 6.2 mg/l MnCl2·4 H2O, 3.9 mg/l ZnSO4·7 H2O, 0.08 mg/l CoCl2·6 H2O, 0.08 mg/l Na2 MoO4·2 H2O, 4.14 mg/l CuSO4·5 H2O, and 4.14 g/l NiSO4·6 H2O were added. These additions were autoclaved for 50 minutes at 120°C. Finally, a vitamin stock solution containing 0.32 g/l vitamin B12, 19.5 g/l thiamine, and 6.66 g/l Ca-pantothenate was filter sterilized and 10 ml of this stock was added to the media. The fermentation was provided with a constant airflow of 8 l/minute supplied throughout. Dissolved oxygen was set at 20%. Vessel agitation was allowed to adjust from a range of 357–900 rpm as needed to maintain the dissolved oxygen set point. Temperature was maintained at 22.5°C. The pH was set at 6.75 and maintained through control loop additions of either 4 N NaOH or 3 N H2SO4. Broth was collected at 36 hours and centrifuged at 5000 g.
HA assays were performed as described previously with modifications . Briefly, 50 µl of doubling dilutions of CFS in PBS were prepared in a 96-well microtiter plate. Equal volume of an approximate 1% solution of chicken red blood cells (Fitzgerald Industries, Acton, MA) in PBS, pH 7.4, was then added to each well followed by incubation at room temperature for 30 minutes. The degree of agglutination was then analyzed visually. The hemagglutination activity unit (HAU) is defined as the highest dilution that causes visible hemagglutination in the well. A recombinant HA protein derived from Influenza A/Vietnam/1203/2004 (H5 N1) (Protein Sciences Corporation, Meriden, CT, USA) was used as positive control for HA activity.
Protein Analysis and Immunoblotting
Schizochytrium cultures were transferred to 50 ml conical tubes and centrifuged at 3000 g or 4500 g for 15 min. The supernatants resulting from this centrifugation were in some cases further centrifuged at 100,000 g for 1 hour. CFS or other protein samples were separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) on a NuPAGE® Novex® 12% bis-tris gel (Invitrogen, Carlsbad, CA) under reducing conditions with 3-(N-morpholino) propanesulfonic acid - sodium dodecylsulfate (MOPS-SDS) running buffer, unless indicated otherwise in the text. The proteins were then stained with SimplyBlue Safe Stain (Invitrogen, Carlsbad, CA, USA) or transferred onto polyvinylidene fluoride membrane and probed for the presence of HA protein with anti-PR8 virus antiserum from rabbit, catalog number 11684-RP01 (Sino Biological Inc., Beijing, China), followed by anti-rabbit IgG (Fc) secondary antibody coupled to alkaline phosphatase (Promega Corporation, Madison, WI). The membrane was then treated with 5-bromo-4-chloro-3-indoyl-phosphate/nitroblue tetrazolium solution according to the manufacturer's instructions (KPL, Gaithersburg, MD). Protein sequences were confirmed by mass spectrometry (Finnigan LTQ linear ion trap) according to the protocol developed by the Mass Spectrometry Laboratory for Protein Sequencing at the Lerner Research Institute (Cleveland Clinic, Ohio)  and by Edman degradation using a 494 Procise Protein Sequencer/140 C Analyzer (Applied Biosystems, Inc., Carlsbad, CA, USA) according to the protocol developed by Iowa State University Protein Facility .
A CFS fraction from a Schizochytrium culture containing PR8 HA was run on NuPAGE® Novex® 12% bis-tris gel (Invitrogen, Carlsbad, CA, USA) under non-reducing conditions with MOPS buffer and the resulting gel was stained with SimplyBlue Safe Stain. The HA protein was excised from the gel and alternately treated with destain solution (DS) comprised of 40 mM ammonium bicarbonate, and 100% acetonitrile, until the color turned clear. Destained gel was re-swelled in 10 mM dithiothreitol (DTT) in 40 mM ammonium bicarbonate at 55°C for 1 hour. The DTT solution was exchanged with 55 mM iodoacetamide and incubated in the dark for 45 minutes. Incubation was followed by two washes with DS before being dehydrated in acetonitrile and subjected to in-gel deglycosylation . The dehydrated gel sections were hydrated at 4°C with 50 mM sodium phosphate, pH 7.5 containing PNGase F (New England Biolabs, Ipswich, MA, USA) and then incubated at 37°C overnight to release N-glycans. Free glycans were extracted from the gel in series with 20% acetonitrile in 5% formic acid, 50% acetonitrile in 5% formic acid, and finally 80% acetonitrile in 5% formic acid. The sample solutions were dried, combined for passage through a C18 sep-pak cartridge (Resprep, Bellefonte, PA, USA), eluted in 5% acetic acid to remove contaminants, and dried again by lyophilization. Released N-linked oligosaccharides were permethylated  and analyzed by nanospray ionization mass spectrometry (NSI-MSn) using a LTQ Orbitrap XL mass spectrometer (Thermo Fisher, Waltham, MA, USA). Specifically, permethylated glycans were dissolved in 1 mM NaOH in 50% methanol and infused directly into the instrument at a constant 0.5 µl/min flow rate. A full Fourier transform MS spectrum was collected at a resolution of 30,000. The capillary temperature was set at 210°C and MS analysis was performed in the positive ion mode. For total ion mapping (automated MS/MS analysis), m/z range 800 to 2000 was scanned using ion trap MS mode in successive 2.8 mass unit windows that overlapped the preceding window by 2 mass units.
Cells of Schizochytrium expressing HA were cultivated under the fermentation conditions described above and centrifuged for 15 min at 3000 g. The CFS was then diluted 1∶1 in a solution containing 20 mM sodium phosphate pH 7.0, 5% glycerol, 1 mM EDTA, and 0.2% Triton™ X-100. The mixture was filtered (0.45 µm pore size) and transferred to an anion-exchange spin column (Thermo Scientific Pierce Biotech, Fisher Scientific, Rockford, IL USA) previously equilibrated with 20 mM sodium phosphate pH 7.0, 5% glycerol, and 0.1% Triton™ X-100. The loaded sample was centrifuged according to the manufacturer’s instructions. The resulting flow-through was concentrated and buffer was exchanged with phosphate-buffered saline (PBS), pH 7.4 by ultrafiltration using Amicon centrifugal devices (Millipore, Billerica, MA).
Immunizations and Animal Model
Female BALB/c mice (6–8 weeks old) were purchased from Charles River Laboratories (Wilmington, MA, USA). The purified HA antigen was stored at ≤ −65°C before being formulated for injection. Vaccine and vehicle control formulations were prepared immediately prior to injection such that each dose (15, 5, 1.7 and 0 µg of HA) could be administered i.m. in 100 µl volumes bilaterally (50 µl/quadricep). Each dose was diluted in PBS to the appropriate concentration, or for corresponding adjuvant cohorts, to a concentration 2 fold higher for emulsification with Addavax™ (Invivogen, San Diego, CA, USA) at a 1∶1 ratio (v:v) according to the manufacturer’s instructions. Placebo (PBS) and adjuvant were also mixed at a ratio of 1∶1. The study design and experimental groups are detailed in the results section. Precisely, groups 4 and 11 were immunized with the rHA-Addavax mixture. Groups 1–3 were given single dose immunizations of descending mass. Groups 7–9 were given identical initial doses which were then followed by a second dose 3 weeks later. Group 11 was established to test mice subjected to all of the above variables combined. Groups 5, 6, and 10 were given volume-matched placebo injections of a single PBS dose, a single PBS-Addavax dose, or two doses of PBS respectively.
All mice were lethally challenged 21 days after the final vaccination. Mice were anesthetized i.p. with ketamine (100 mg/kg)/xylazine (10 mg/kg) and intranasally inoculated with 3000 TCID50 (5 MLD50) of infectious PR8 virus. The mice were then weighed and monitored daily for signs of disease. Mice were humanely euthanized with CO2 if their body weight dropped to 75% of baseline weights. All studies were conducted under applicable laws and guidelines and were approved by the IIT Research Institute Animal Care and Use Committee.
Serum samples were treated with receptor destroying enzyme (RDE) (Denka Seiken, Tokyo, Japan) overnight at 37°C and diluted 1∶10 with phosphate buffered saline (PBS). RDE treated samples were serially diluted into eight two-fold dilutions (1∶10 to 1∶20,480) in PBS and tested in a hemagglutination inhibition (HI) assay with 0.5% chicken red blood cells. The reciprocal of the highest dilution of serum or plasma that completely inhibited hemagglutination was determined to be the HI titer. For computational purposes, titers of <1∶10 were assigned a value of 1∶5. Titer values were log2 transformed prior to conducting statistical analysis and expressed as geometric mean titers (GMT).
Bronchoalveolar Lavage Virus Titers
On days 3 and 7 after inoculation with PR8 influenza virus, 1–3 mice in each experimental and control group were sacrificed. Bronchoalveolar lavage (BAL) fluids were collected in two 1 ml washes with PBS. BAL samples were stored at ≤ −65°C until live virus titer determination by TCID50 assay in Madin-Darby Canine Kidney (MDCK) cells. MDCK cells were inoculated with serial dilutions of BAL fluid and incubated at 37°C for 48 hours. A hemagglutination assay was performed and virus titers (log10 TCID50/ml) were calculated by the Reed and Muench method .
Comparison of survival between groups of mice was performed with the log rank test on the Kaplan–Meier survival data. Comparison of the antibody and viral titer data was performed using analysis of variance for multiple comparisons with Systat (Systat, Chicago, IL). Error bars represent standard deviation, and statistical significance was defined as P<0.05.
Expression of Viral Glycoproteins by Schizochytrium sp
To determine if Schizochytrium is capable of expressing intact viral glycoproteins, the full-length reading frames of two H1, one H5 and one influenza B hemagglutinins were codon optimized and incorporated within a Schizochytrium expression vector in a context that permitted expression under control of the Schizochytrium EF-1 promoter and PFA3 terminator. The strains from which the HA protein sequences were derived as well as the accession numbers of the vector constructs used to express them, are listed in Table 1. The constructs were specifically designed to bias for ectopic integration events. No fewer than ten transgenic clones resulting from transformation with each vector were cultured for two days under standard conditions for Schizochytrium and cell free supernatant (CFS) preparations were screened by hemagglutination assay for recombinant protein activity. Results indicated that the proteins were expressed and exported to the extracellular space in an active form. Among all transformants, hemagglutination activities ranged from 16 to 512 HAU/50 µl of CFS (Table 1).
Production of HA from Influenza PR8 by Schizochytrium sp
Of the various hemagglutinins screened by HA assay, the one derived from PR8 virus (H1 serotype) was selected for further characterization. Transformants of the vector pCL0143 were screened as described above and by immunoblot analysis. A range of hemagglutination activities were observed among the pCL0143 transformants (0–512 HAU). The strain generating the highest activity was designated 143-9. CFS prepared from a culture of this clone was analyzed by immunoblot under reducing conditions and showed the presence of two reactive bands corresponding to the expected molecular weights of HA1 and HA2 (Fig. 1A, lane “E”). The identity of the bands was confirmed by peptide sequence analysis (data not shown). Immunoblot analysis of the CFS under non-reducing conditions resulted in a single reactive band corresponding to the disulfide-bonded HA heterodimer (Fig. 1B). The migration of these bands agreed well with the predicted masses of HA1 (36.7 kDa), HA2 (25.1 kDa) and of the HA heterodimer (63.4 kDa). Centrifugation of the CFS from a shake flask culture at 100,000 g for 60 minutes indicated that the hemagglutinin is mostly associated with the insoluble fraction. The soluble fraction displayed 16 HAU whereas the insoluble fraction yielded 256 HAU. N-terminal sequence analysis also confirmed that the signal peptide is cleaved as would be expected for export via the secretory pathway (data not shown).
SDS-PAGE analysis of cell-free supernatants under reducing (A and D) and non-reducing (B and C) conditions and corresponding immunoblots with anti-Influenza A virus antibody are shown. Panels “I” are gels stained with Coomassie blue and panels “II” are the corresponding immunoblots. (A) Analysis of cell-free supernatants from Schizochytrium control “C” and experimental “E” strains showing the presence of HA subunits 1 and 2 under reducing conditions HA1 and HA2). (B) Analysis of cell-free supernatants from Schizochytrium control “C” and experimental “E” strains showing the presence of HA under non-reducing conditions. (C) SDS-PAGE analysis (non-reducing) of the Schizochytrium HA fraction used for the mouse challenge study. (D) Analysis of cell-free supernatants from Schizochytrium control “C” and mutant experimental “MutE” strains showing the presence of an uncleaved form of HA (HA0) under reducing conditions.
Clone 143-9 was also cultured in a 10 liter fermentor overnight in protein-free media according to the conditions described herein. Fermentation-derived CFS was assayed for HA activity which was notably improved (4096 HAU) as compared to CFS derived from shake flask cultures (512 HAU). PR8 HA antigen from fermentation culture supernatants of 143-9 were partially purified (approximately 20% of the total protein) for glycosylation analysis and for formulation as a vaccine as described in the Materials and Methods section. This material was also analyzed by non-reducing SDS-PAGE and by immunoblot (Fig. 1C).
To verify that the protein was cleaved at the appropriate location, a point mutant variant of the HA gene, that specifically changed the arginine residue of the conserved R∧G hemagglutinin cleavage site to a serine, was engineered and inserted in the same backbone vector as was used for pCL0143 construction. This new vector was named pCL0154 and was used to transform Schizochytrium. The resulting transgenic strains also exported HA to the extracellular medium, as indicated by activity assays performed on isolated CFS. Immunoblot analysis of one of those strains (“MutE”) indicated that the protein migrated as a single band in reducing conditions (Fig. 1D). This finding suggests that Schizochytrium sp. possesses an endogenous protease capable of mediating HA0 cleavage; a property shared with susceptible hosts of influenza viruses.
Characterization of PR8 HA Glycosylation
Influenza hemagglutinin is a glycosylated protein which binds terminal sialic acid residues of host protein glycoforms. Mass spectrometry (NSI-MSn) of the released and permethylated N-glycans was used to determine the glycoform structures associated with purified rHA, and to determine how they might differ from glycoforms made by native hosts of influenza (Figure 2). A series of ions were detected which correspond to oligosaccharide Man5–9GlcNAc2 structures. The identities of glycan signals observed in the full MS spectrum were confirmed by subjecting the signals to fragmentation in MS/MS analysis. Typical high mass fragmentation patterns were detected for each of the glycans, with predominant fragmentation at the bisecting GlcNAc residues. A trace amount (less than 1%) of hyper-mannosylation was also detected at 1312, 1407 and 1516 (Z = 2) corresponding to (GlcNAc)2(Mannose)9(Hex)1–3, respectively. The precise identity of the terminal hexose found among these structures could not be inferred by these analyses because permethylated hexoses all share the same mass (MW 204). No fucose or xylose residues were detected among any glycoforms. Sialic acid residues were also not detected in this analysis and a survey of a proprietary, Schizochytrium genome database found no genes related to sialic acid biosynthesis.
Given weights include addition of sodium, and account for permethylation. *Observed during total ion mapping (TIM), m/z value refers to the 2 mass unit window in which the noted MS/MS loss was observed. Hexose (Hex), mannose (Man), and N-acetylglucosamine (GlcNAc) residues are represented by gray circles, open circles, and open squares, respectively.
Immunization and Infectious Challenge of Mice
Mice were separated into 11 groups according to the schedule presented in Table 2. Three groups were given single-dose immunizations without adjuvant, and one group was given a single dose immunization with adjuvant. Similarly, three groups were given two doses (3 weeks apart) and one was given two doses (3 weeks apart) with adjuvant. Adjuvant was used with two different groups according to differing rationales. Because this study represents the first parenteral use of a Schizochytrium-derived preparation, adjuvant use with the highest, two-dose group was intended to provide an upper range of measurable responses (group 11). Adjuvant effects could be more clearly tested by establishment of parallel groups receiving the lowest rHA dosage with and without adjuvant (groups 4 and 3 respectively). All remaining groups served as placebo controls. Mice were monitored for survival and weighed for 14 days after lethal challenge with PR8 (Figures 3 and 4 respectively).
Mice were given a single vaccination (A) or two vaccinations 21 days apart (B), challenged and their survival monitored for 14 days. N = 10 mice per group.
Mice were given a single vaccination (A) or two doses 21 days apart (B), challenged and their body weight monitored for 14 days. N = 10 mice per group.
Among the groups receiving single dose vaccinations without adjuvant, protection correlated with dose level; 80% survival with 1.7 µg rHA, 90% with 5 µg rHA, and 100% with 15 µg rHA (groups 3, 2 and 1, respectively). All surviving mice from these groups initially lost some weight (as low as ∼23% of their initial body mass) but eventually recovered, with the exception of three animals; one from group 2 and one from group 3 died on day 9, with the last death in group 3 on day 13. Mice in groups receiving 2 doses of rHA without adjuvant all survived regardless of dose level (1.7, 5 or 15 µg). All mice vaccinated with rHA+adjuvant (groups 4 and 11) also survived challenge. As expected, mice receiving any of the placebo formulations died between days 7 and 11 (groups 5, 6 and 10). Mice given 1.7 µg of vaccine with adjuvant (group 4) realized slight gains in average weight during the 14 days post-challenge. Similarly, all mice from the groups receiving two doses of vaccine either gained or retained weight after virus challenge (groups 7–9 and 11).
Infectious virus titers (log10 TCID50/ml) from samples of bronchoalveolar lavage (BAL) recovered on days 3 and 7 post-challenge are shown in Figure 5. Titers from day 3 were measureable in all groups. Groups vaccinated with rHA+adjuvant (4 and 11) had virus titers several logs lower than groups vaccinated with rHA alone or placebo. Small sample sizes (group 9, n = 1) precluded a statistical evaluation for day 3 BAL samples across the study groups. For all groups, virus titers appeared lower at day 7 post-challenge than at day 3 post-challenge. By day 7, virus had been completely cleared in the groups receiving two doses of 5 and 1.7 µg of HA, but not in the 15 µg group (significant variation among individual titer values). Virus could not be detected in cohorts receiving either of the adjuvant-formulated vaccines (groups 4 and 11) at days 3 and 7 post-challenge, indicating that the mice in these groups were unable to establish a persistent, replicating viral infection in the lungs. Two vaccinations (with or without adjuvant) or a single vaccination (with adjuvant) significantly decreased viral titers by day 7 when compared to controls (p<0.05; ANOVA).
Mice were euthanized on day 3 (white bars) and day 7 (gray bars) and brochoalveolar lavage (BAL) washes were collected. Titers are given as the averages of three BAL washes except for those harvested from group 8 on day 3 and 7 and from groups 5 and 9 on day 3, for which only 2 BAL were available. Only one BAL was harvested from group 9 on day 3. Groups 1–6 were given single vaccinations. Groups 1–8 were given two vaccinations (three weeks apart) of the same mass with each injection. Groups 1, 7, and 11 were given 15 µg rHA, groups 2 and 8 were given 5 µg rHA, and groups 3, 4, and 9 were given 1.7 µg HA. Groups 5, 6, and 10 were given placebo injections. Groups 4, 5, and 11 were given vaccinations formulated with adjuvant. Results are shown as mean ± SD virus titers in (BAL) washes. *Statistically significant decreases in titers as compared with placebo are indicated. Limit of detection is 101.5 TCID50/ml.
Serological Response: Hemagglutination Inhibition
To evaluate the presence of HA-specific antibodies, the hemagglutination inhibition (HI) activity of collected sera was determined (Table 3). The highest HI titers were found in the groups receiving adjuvant-formulated vaccine (groups 4 and 11). HI titers increased significantly after the second injection (for applicable groups). HI titers increased in all groups after infectious challenge, with the exception of group 11, for which HI titers dropped moderately after challenge (HI GMT 1225 to 787 from day 42 to 56). For those groups receiving 2 vaccinations, day 42 HI sero-response rates were between 25 and 100%. Among the single-dose groups at day 42, the majority of mice (49/52) had no detectable HI titer (≥40), with the animals receiving 1.7 µg of rHA+adjuvant having the highest sero-response rate (12.5% 2/16). Interestingly, in the absence of detectable HI titers, mice were protected from lethality.
A novel system for the production of transgenic proteins was used to express several hemagglutinin genes of influenza virus. This investigation describes the first use of the Schizochytrium platform for the expression of recombinant proteins for commercial pharmaceutical or veterinary applications. The four HA proteins reported in this study are derived from strains of Influenza A and B, and Influenza A hemagglutinin subtypes H1 and H5. The hemagglutination assay was used for simple and rapid detection of HA expression for all initial screens. Factors associated with variation in HA activity levels included HA subtype (H1 vs. H5), individual HA transformants, and culture conditions (fermentors vs. flasks). We chose to characterize one HA derived from influenza PR8 in greater depth due to the availability of an acceptable animal model for challenge with homologous virus and because of the broad use of this isolate in both laboratory and industrial contexts.
The expression of full-length HA as an extracellular protein without the aid of the matrix protein, viral vectors, or any agent which might facilitate the budding of virus-like particles for which other platforms are well-known, is a notable result . In some systems, for efficient, extracellular expression of HA, the C-terminal membrane spanning domain is necessarily truncated , –. The ability of Schizochytrium to facilitate export of membrane-bound HA however, obviates the need for truncated forms. One potential mechanism for this may involve filamentous extensions from the main cell body of many other Thraustochytriales under certain conditions. These filopodia-like structures are produced by unique organelles called bothrosomes which assemble actin polymers and extrude them through the cell wall, while retaining a coating of outer membrane lipids and embedded membrane proteins . The role these structures play in HA export is a subject for continuing investigation.
Although cleavage of HA0 to HA1 and HA2 subunits is not required to elicit immunity or hemagglutination, it is required for membrane fusion and influenza infectivity . Therefore cleavage of HA may yield an antigen offering more conformational epitopes for the generation of functional neutralizing antibodies. HA cleavage, presumably by an endogenous Schizochytrium protease, was characterized by the introduction of a point mutation in a region of the PR8 HA gene encoding the cleavage recognition site. This mutation prevented HA cleavage and verified that the event occurs at exactly the same site as in native influenza hosts, suggesting that the antigenic conformation of HA from Schizochytrium might closely mimic that of HA produced by natural hosts. This feature is apparently not shared by certain other systems used to express transgenic influenza HA, such as fungi and higher plants.
Native PR8 HA possesses 6 glycosylation sites and like all influenza strains, glycoforms found on envelope proteins are generally devoid of terminal sialic acid residues by the action of neuraminidase. It is thought that the desialylation of host glycoproteins prevents virus aggregation and re-infection of an infected host cell, thereby improving the efficiency of the infection cycle . Although the NA of PR8 has been successfully expressed using Schizochytrium (both independently and as a co-expression product with HA, data not shown) its presence was not required for expression of HA, since Schizochytrium does not synthesize or attach terminal sialyl residues to protein glycoforms which could induce HA aggregation. We detected only trace amounts of hypermannose glycans of the type found on yeast proteins, and a majority of high-mannose glycoforms. These latter glycoforms may be advantageous for vaccine design as other studies have demonstrated that similar glycan structures expose the HA molecule to neutralizing antibodies and receptor ligands to a greater degree, and that HA-based vaccines become more protective with correspondingly simpler HA glycostructures , . Similar high-mannose glycans are also found among glycoproteins expressed by lepidopteran expression systems, with the exception that these glycoforms usually also contain a core fucose residue , . Our analyses did not reveal any fucose residues; or xylose residues commonly associated with plant and green algal glycans . However, the high-mannose structures which were observed, such as (GlcNAc)2(Man)5–6, are assembled in part by α1,3 and α1,6 linkages which have been demonstrated to be immunogenic in mammals . The glycan (GlcNAc)2(Man)6 is also found at Asn 65 of the HA1 subunit from mammalian cells infected with influenza  although it is unclear to what degree these α1,3 mannoses may promote immunity to viral infection when associated with vaccine antigens.
To determine whether PR8 HA produced by Schizochytrium could serve as an effective vaccine, the BALB/C murine challenge model was used. The results show that HA, as expressed from Schizochytrium, enabled mice to mount an immune response which protected against subsequent lethal challenge with live homologous influenza virus. Health and survival of groups given two rHA vaccinations (100%) or rHA with a squalene adjuvant (100%) was unequivocal. Robust HI titers were observed in all vaccinated mice by day 56 (14 days post-challenge) although HI titers on day 42 were not clearly correlated with protection, justifying additional investigation of immune mechanisms in this model. All mice receiving two doses or adjuvant-formulated vaccine gained weight over the course of the study, including cohorts receiving the lowest dose of HA. No mice among any placebo group survived viral challenge, while 80% of mice given a single injection of the lowest dose of HA without adjuvant survived the infection. Taken together the results from this murine model indicate that recombinant influenza HA, as expressed by Schizochytrium, functions as an effective vaccine against lethal influenza virus challenge, comparing favorably with proof of concept studies reported for other transgenic expression systems , , . Further studies will be required to evaluate the feasibility of Schizochytrium-produced influenza vaccine candidates purified and formulated for advanced non-clinical and eventual clinical development.
Cell-based systems, paired with synthetic biology, may be beneficial alternatives to egg-based manufacturing. Synthetic biology provides greater flexibility for transgenic systems via rigorous codon optimization techniques and RNA structure prediction; allowing for not only higher levels of protein expression but greater probability of expression for a broader range of targets. The scalability and yield of systems for expression of influenza proteins are also important to consider. Although Schizochytrium is routinely grown in 250,000 l fermentors, the existing process is designed for production of lipids. On a mass basis, HA yields from Schizochytrium shake flask cultures of 5–20 mg/l have been observed. These levels are comparable or better than reported HA yields from other eukaryotic recombinant systems . The need for purification of unrelated viral material (such as baculovirus) is also obviated by the Schizochytrium platform. Similar to the baculovirus system, Schizochytrium is capable of producing an HA molecule with an intact transmembrane domain which is free of fusion partners or truncations that may reduce immune specificity but without the risks of virus-vectored expression. These traits justify continued development and evaluation of the Schizochytrium platform as an alternative to conventional systems not only for influenza antigens but other recombinant proteins for vaccine and therapeutic applications.
The authors gratefully acknowledge Mayumi Ishihara for technical assistance, Joan Fusco and Julie Nauroth for review of the manuscript, and Joe Pfeifer, Joe Speroni, and Peter Nitze for organizational support.
Conceived and designed the experiments: A-CB DB JCL RZ. Performed the experiments: A-CB DB CO YB GM SA-H. Analyzed the data: A-CB DB PA JCL RZ. Contributed reagents/materials/analysis tools: A-CB DB CO YB GM SA-H PA JCL. Wrote the paper: A-CB DB GM PA JCL.
- 1. Stokes J, Chenoweth AD, Waltz AD, Gladen RG, Shaw D (1937) Results of immunization by means of active virus of human influenza. J Clin Invest 16: 237–243. doi: 10.1172/jci100853
- 2. Montomoli E, Khadang B, Piccirella S, Trombetta C, Mennitto E, et al. (2012) Cell culture-derived influenza vaccines from Vero cells: a new horizon for vaccine production. Expert Rev Vaccines 11: 587–594. doi: 10.1586/erv.12.24
- 3. Genzel Y, Reichl U (2009) Continuous cell lines as a production system for influenza vaccines. Expert Rev Vaccines 8: 1681–1692. doi: 10.1586/erv.09.128
- 4. Haynes J (2009) Influenza virus-like particle vaccines. Expert Rev Vaccines 8: 435–445. doi: 10.1586/erv.09.8
- 5. Athmaram TN, Saraswat S, Santhosh SR, Singh AK, Suryanarayana WS, et al. (2011) Yeast expressed recombinant Hemagglutinin protein of novel H1N1 elicits neutralising antibodies in rabbits and mice. Virol J 8: 524. doi: 10.1186/1743-422x-8-524
- 6. Allgaier S, Taylor RD, Brudnaya Y, Jacobson DJ, Cambareri E, et al. (2009) Vaccine production in Neurospora crassa. Biologicals 37: 128–32. doi: 10.1016/j.biologicals.2009.02.006
- 7. Mett V, Musiychuk K, Bi H, Farrance CE, Horsey A, et al. (2008) A plant-produced influenza subunit vaccine protects ferrets against virus challenge. Influenza Other Respi Viruses 2: 33–40. doi: 10.1111/j.1750-2659.2008.00037.x
- 8. Davis AR, Bos T, Ueda M, Nayak DP, Dowbenko D, et al. (1983) Immune response to human influenza virus hemagglutinin expressed in Escherichia coli.. Gene 21: 273–284. doi: 10.1016/0378-1119(83)90011-2
- 9. Downie JC (1970) Neuraminidase- and hemagglutinin-inhibiting antibodies in serum and nasal secretions of volunteers immunized with attenuated and inactivated influenza B/Eng/13/65 virus vaccines. J Immunol 105: 620–26.
- 10. Johansson BE, Bucher DJ, Kilbourne ED (1989) Purified influenza virus hemagglutinin and neuraminidase are equivalent in stimulation of antibody response but induce contrasting types of immunity to infection. J Virol 63: 1239–1246.
- 11. Cox MM, Hollister JR (2009) FluBlok, a next generation influenza vaccine manufactured in insect cells. Biologicals 37: 182–189. doi: 10.1016/j.biologicals.2009.02.014
- 12. Wiley DC, Skehel JJ (1987) The structure and function of the hemagglutinin membrane glycoprotein of influenza virus. Annu Rev Biochem 56: 365–394. doi: 10.1146/annurev.bi.56.070187.002053
- 13. Mir-Shekari SY, Ashford DA, Harvey DJ, Dwek RA, Schulze IT (1997) The glycosylation of the influenza A virus hemagglutinin by mammalian cells. J Biol Chem 272: 4027–4036. doi: 10.1074/jbc.272.7.4027
- 14. Gallagher PJ, Henneberry JM, Sambrook JF, Gething MJ (1992) Glycosylation requirements for intracellular transport and function of the hemagglutinin of influenza virus. J Virol 66: 7136–7145.
- 15. Gallagher P, Henneberry J, Wilson I, Sambrook J, Gething MJ (1988) Addition of carbohydrate side chains at novel sites on influenza virus hemagglutinin can modulate the folding, transport, and activity of the molecule. J Cell Biol 107: 2059–2073. doi: 10.1083/jcb.107.6.2059
- 16. Wagner R, Herwig A, Azzouz N, Klenk HD (2005) Acylation-mediated membrane anchoring of avian influenza virus hemagglutinin is essential for fusion pore formation and virus infectivity. J Virol 79: 6449–6458. doi: 10.1128/jvi.79.10.6449-6458.2005
- 17. Ratledge C (2004) Fatty acid biosynthesis in microorganisms being used for Single Cell Oil production. Biochimie 86: 807–815. doi: 10.1016/j.biochi.2004.09.017
- 18. Bailey RB, DiMasi D, Hansen JM, Mirrasoul PJ, Ruecker CM, et al.. (2010) Enhanced production of lipids containing polyenoic fatty acids by high density cultures of eukaryotic microbes in fermentors. US Patent No. 7,732,170.
- 19. Barclay W, Weaver C, Metz J (2005) Development of a docosahexaenoic acid production technology using Schizochytrium: a historical perspective. In: Ratledge C, Cohen Z editors. Single Cell Oils. Champaign: AOCS Press 36–52.
- 20. Roessler PG, Matthews TD, Ramseier TM, Metz JG (2006) Product and process for transformation of Thraustochytriales. US Patent No. 7,001,772.
- 21. Lippmeier JC, Crawford KS, Owen CB, Rivas AR, Metz JG, et al. (2009) Characterization of both polyunsaturated fatty acid biosynthetic pathways in Schizochytrium sp. Lipids 44: 621–630. doi: 10.1007/s11745-009-3311-9
- 22. Weaver CA, Zirkle R, Doherty DH, Metz JG (2011) Chimeric PUFA polyketide synthase systems and uses thereof. US Patent No. 8,003,772.
- 23. Lippmeier JC, Apt KE (2011) Recombinant thraustochytrids that grow on xylose, and compositions, methods of making, and uses thereof. US Patent Application No.20110195448.
- 24. Lippmeier JC, Apt KE, Hansen JM (2011) Recombinant thraustochytrids that grow on sucrose, and compositions, methods of making, and uses thereof. US Patent Application No.20110195448.
- 25. Perkins FO, Amon JP (1972) The ultrastructure of holdfasts, “rhizoids”. and 'slime tracks” in thraustochytriaceous fungi and Labyrinthula spp. Arch Mikrobiol 84: 95–118. doi: 10.1007/bf00412431
- 26. Porter D (1969) Ultrastructure of Labyrinthula. Protoplasma 69: 1–19. doi: 10.1007/bf01256763
- 27. Apt KE, Lippmeier JC, Simpson D, Wang J, Wynn JP, et al.. (2010) Protein production in microorganisms of the phylum Labyrinthulomycota. US Patent Application No.20100233760.
- 28. Oxford JS, Kelly LS, Davies S, Lambkin R (1999) Antiviral Testing. In: Cann AJ editor. Virus Culture, A Practical Approach. New York: Oxford University Press, 201–238.
- 29. Kinter M, Sherman NE (2000) Protein Sequencing and Identification Using Tandem Mass Spectrometry. New York, NY : John Wiley & Sons.
- 30. Hunkapiller MW (1985) PTH Amino Acid Analysis. ABI User Bulletin no 14. Foster City: Applied Biosystems. 27 p.
- 31. Küster B, Wheeler SF, Hunter AP, Dwek RA, Harvey DJ (1997) Sequencing of N-linked oligosaccharides directly from protein gels: in-gel deglycosylation followed by matrix-assisted laser desorption/ionization mass spectrometry and normal-phase high-performance liquid chromatography. Anal Biochem 250: 82–101. doi: 10.1006/abio.1997.2199
- 32. Anumula KR, Taylor PB (1992) A comprehensive procedure for preparation of partially methylated alditol acetates from glycoprotein carbohydrates. Anal Biochem 203: 101–108. doi: 10.1016/0003-2697(92)90048-c
- 33. Reed LJ, Muench H (1938) A simple method for estimating fifty percent endpoints. Am J Hygiene 27: 493–497.
- 34. Vanlandschoot P, Maertens G, Jou WM, Fiers W (1993) Recombinant secreted haemagglutinin protects mice against a lethal challenge of influenza virus. Vaccine 11: 1185–1187. doi: 10.1016/0264-410x(93)90040-5
- 35. Khurana S, Verma S, Verma N, Crevar CJ, Carter DM, et al. (2010) Properly folded bacterially expressed H1N1 hemagglutinin globular head and ectodomain vaccines protect ferrets against H1N1 pandemic influenza virus. PLoS One 5: e11548. doi: 10.1371/journal.pone.0011548
- 36. Song L, Nakaar V, Kavita U, Price A, Huleatt J, et al. (2008) Efficacious recombinant influenza vaccines produced by high yield bacterial expression: a solution to global pandemic and seasonal needs. PLoS One 3: e2257. doi: 10.1371/journal.pone.0002257
- 37. Shoji Y, Bi H, Musiychuk K, Rhee A, Horsey A, et al. (2009) Plant-derived hemagglutinin protects ferrets against challenge infection with the A/Indonesia/05/05 strain of avian influenza. Vaccine 27: 1087–1092. doi: 10.1016/j.vaccine.2008.11.108
- 38. Preston TM, King CA (2005) Actin-based motility in the net slime mould Labyrinthula:evidence for the role of myosin in gliding movement. J Eukaryot Microbiol 52: 461–475. doi: 10.1111/j.1550-7408.2005.00064.x
- 39. Skehel JJ, Wiley DC (2000) Receptor binding and membrane fusion in virus entry: the influenza hemagglutinin. Annu Rev Biochem 69: 531–69. doi: 10.1146/annurev.biochem.69.1.531
- 40. Wagner R, Matrosovich M, Klenk H-D (2002) Functional balance between haemagglutinin and neuraminidase in influenza virus infections. Rev Med Virol 12: 159–166. doi: 10.1002/rmv.352
- 41. Wang CC, Chen JR, Tseng YC, Hsu CH, Hung YF, et al. (2009) Glycans on influenza hemagglutinin affect receptor binding and immune response. Proc Natl Acad Sci U S A 106: 18137–18142. doi: 10.1073/pnas.0909696106
- 42. Butters TD, Yudkin B, Jacob GS, Jones IM (1998) Structural characterization of the N-linked oligosaccharides derived from HIVgp120 expressed in lepidopteran cells. Glycoconj J 15: 83–88.
- 43. Hanson SR, Culyba EK, Hsu T-L, Wong C-H, Kelly JW, et al. (2009) The core trisaccharide of an N-linked glycoprotein intrinsically accelerates folding and enhances stability. Proc Natl Acad Sci U S A 106: 3131–3136. doi: 10.1073/pnas.0810318105
- 44. Bosch D, Schots A (2010) Plant glycans: friend or foe in vaccine development? Expert Rev Vaccines 9: 835–842. doi: 10.1586/erv.10.83
- 45. Geijtenbeek TBH, Gringhuis SI (2009) Signalling through C-type lectin receptors: shaping immune responses. Nat Rev Immunol 9: 465–479. doi: 10.1038/nri2569
- 46. Ross TM, Mahmood K, Crevar CJ, Schneider-Ohrum K, Heaton PM, et al. (2009) A trivalent virus-like particle vaccine elicits protective immune responses against seasonal influenza strains in mice and ferrets. PLoS One 4: e6032. doi: 10.1371/journal.pone.0006032