Background
Zika virus (ZIKV) is an emerging arthropod-borne virus that belongs to the family
Flaviviridae and the genus Flavivirus. It was first discovered in a sentinel rhesus monkey in the Zika forest of Uganda in 1947 [
1]. ZIKV is classified into two lineages: African and Asian strains [
2] but they share > 95% amino acid identity with a single serotype unlike the closely related flavivirus, dengue virus (DENV) that is composed of 4 different serotypes [
3]. Since its first detection in Brazil in 2015 [
4], ZIKV has spread rapidly throughout the Americas, and over 170,000 laboratory cases had been confirmed in 48 countries worldwide by the end of 2016 [
5]. ZIKV is mainly transmitted via mosquitoes of the genus
Aedes [
6]
, but also through sexual contact [
7] and vertical transmission [
8,
9]. Most of ZIKV infections are asymptomatic or mild [
10], but in some cases can lead to neurological complications including Guillain-Barré syndrome (GBS) in adults and microcephaly in foetuses [
8,
9,
11].
ZIKV is an enveloped virus with a ~ 10.7 Kb single-stranded, positive-sense RNA genome which encodes a single viral polyprotein that is processed post-translationally into three structural proteins (capsid [C], pre-membrane [prM] and envelope [Env]) and seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B and NS5) [
12,
13]. Recent cryo-electron microscopy analysis of ZIKV revealed that the overall structure is nearly identical to other closely related flaviviruses such as dengue (DENV) [
14,
15]. Env is involved in receptor binding, fusion and viral entry into target cells and is a primary target for neutralizing antibodies during ZIKV infection [
16,
17]. prM interacts with the nascent Env protein during the viral assembly to assist correct folding and prevent premature fusion to the endoplasmic reticulum membrane. It does this by covering the fusion loop in Env before eventually releasing the prM to form mature virions [
18]. The expression and purification of soluble recombinant Asian-lineage ZIKV Env proteins have been described using
E. coli as the host cell and involve expression as inclusion bodies and re-folding in vitro [
19]. Expression using
Drosophilia cells has also been described [
20]. In addition, expression of African ZIKV Env proteins in mammalian cells such as HEK293T cells has been described more recently and the N-glycosylation on ZIKV Env was found to be important for expression and secretion of Env [
21‐
23]. Many commercially available ZIKV Env proteins are baculovirus-derived and expressed in insect cells based on sequences from the African strain, with a purity of > 85% for ELISA applications, but they may not be easily affordable in Zika endemic areas in developing countries. Protein expression in a mammalian system is more challenging than in
E. coli or in insect cells due to lower yield and increased cost. Use of a mammalian platform however, ensures correct post-translational modifications [
24,
25] including glycosylation at N154 of ZIKV Env which is an important determinant of ZIKV virulence and neuro-invasive potential [
26]. Previous studies have investigated the use of fusion partners such as maltose binding protein (MBP) fused to dengue EDIII or rat CD4 fused to the CD5 immunoglobulin domain to enhance protein solubility and secretion but there is no data available that describes the use of fusion partners for Zika Env [
27‐
29].
We have recently developed adenovirus based (ChAdOx1 ZIKV) vaccines, which induce strong immune response in mice and provide protection in a mice ZIKV challenge model [
30]. An Asian-lineage Env antigen would be ideal for ELISA assays to monitor the humoral immune response in ZIKV patients and in human volunteers in upcoming clinical trials; as well as to monitor responses in ZIKV pre-clinical models including mice and non-human primates. Such assays could also have a potential use for sero-diagnosis of Zika infection, as these strains of Asian-American lineage are currently causing outbreaks in the Americas.
In this study, we have designed various constructs (prM-Env, Env only and truncations of Env) with or without a rat CD4 fusion tag at the C-terminus to optimise large-scale production of soluble recombinant Asian-lineage ZIKV Env proteins. Production of Env is intended for both Zika disease diagnosis and monitoring of humoral responses in both mice and humans vaccinated with Env-based vaccines. The correct expression of Env-CD4 and Env proteins was verified by western blot using a mouse pan-flavivirus monoclonal antibody, a ZIKV Env monoclonal antibody, and by ELISA assays using sera from mice immunized with a ChAdOx1 ZIKV vaccine. The recombinant Env-CD4 was able to bind antibodies from patient sera with Zika infection and distinguished from human sera without previous exposure to flaviviral infections. This serological data was also correlated with ZIKV neutralisation capacity in vitro.
Discussion
Given the increasing importance of accurate sero-diagnosis of ZIKV infection and the necessity of monitoring humoral immune responses following vaccination with ZIKV vaccines in preclinical and clinical development, we have produced an Asian-lineage ZIKV Env recombinant protein with an optimized secretory profile in mammalian cells, to facilitate large-scale production. The rat CD4 fusion tag fused at the Env protein C-terminus increased secretion of Env-CD4, in-keeping with data for immunoglobulin CD48 [
29] and the addition of prM also enhanced its secretion [
32,
33]. The silver stained PAGE- profiles of purified ZIKV Env-CD4 and Env proteins indicates improved purity when compared with the commercially available African ZIKV Env protein produced in an insect system. Purified Env-CD4 and Env proteins were shown to bind to a pan-flavivirus antibody, to serum antibodies from mice immunized with adenovirus vectored ChAdOx1 ZIKV prME_ΔTM vaccines [
30], and to a ZIKV Env monoclonal antibody, therefore confirming that the purified proteins are correctly expressed. A portion of ZIKV Env-CD4 was found to undergo auto-cleavage of CD4 and release of ZIKV Env and CD4 peptides under reducing conditions, but this was limited under the non-reducing conditions, which suggests that Env-CD4 may retain its intact structure under non-reducing conditions including in ELISA assays. Our novel ZIKV Env-CD4 and Env proteins are based on the Asian-lineage strain which is currently circulating in the Americas [
4,
34,
35]. ELISA assays with these proteins using sera from mice immunized with adenovirus vectored ChAdOx1 ZIKV prME_ΔTM indicated that antibodies elicited by this vaccine are capable of binding Env-CD4 and Env, to a similar extent as binding to a commercial African lineage ZIKV Env.
The 3C protease site in Env-CD4 allows cleavage of CD4 following secretion and purification, however the presence of the CD4 fusion tag had no effect on the binding of ZIKV Env specific antibodies as shown by WB and ELISA. This suggests we can produce either ZIKV Env-CD4 or Env recombinant proteins for future use in sero-diagnosis of ZIKV and in future clinical trials or in further vaccine efficacy assessments. The ELISA results demonstrated that Env-CD4 was recognised by antibodies in ZIKV-exposed patients and with higher ODs than sera from healthy donors. The mean endpoint reciprocal titers for Env-CD4 were significantly higher than those for commercial Env which suggests that Env-CD4 proteins based on Asian-lineage strains may have more specific binding towards the Asian-American strain of ZIKV specific antibodies in Mexican patients. Alternatively, lower purity or stability of commercial Env may account for such differences. As the UK population is largely naive to flavivirus except those vaccinated against yellow fever virus and tick-borne encephalitis virus [
36], the sero-diagnosis of ZIKV and the evaluation of immunogenicity during upcoming clinical trials in the UK is feasible. However, the difficulty in sero-diagnosis in flaviviral endemic areas, such as areas with dengue virus prevalence, makes the diagnosis more challenging due to antibody cross-reactivity [
37‐
40], thus perhaps requiring ZIKV neutralisation assays to provide a correct sero-diagnosis which is still considered to be the “gold standard” [
41‐
43]. To assess ZIKV antibody neutralisation activity in vitro and correlate the endpoint reciprocal antibody titers with the level of neutralising antibodies, a Zika plaque reduction and neutralisation test (PRNT) was performed which suggested that the reciprocal titers from the ELISA assay could be correlated with neutralising titers (ND50). This may be relevant during human clinical trials to monitor the levels of neutralising antibodies against ZIKV Env. It was previously shown in some studies that anti-dengue titers from the ELISA did not always correlate with neutralizing antibody titers unlike other flaviviruses, and this is yet to be determined for ZIKV [
44]. 1 out of 10 control sera (BD4), had a detectable ND50 titer of 15.5. It is possible that this patient had been previously exposed to ZIKV or other flaviviruses or vaccinated against the yellow fever virus, but as these were blood donor samples, the information is not available. It would be of interest to further determine the cross-reactivity of the ZIKV Env in patients exposed to a different set of flavivirus (YFV, TBEV, JEV, DENV, etc.), with the main challenge of sourcing retrospective serum collection of both flavivirus vaccinated or flavivirus exposed individuals.
In conclusion, we have achieved large-scale production of an Asian-lineage ZIKV Env recombinant protein in mammalian cells, and by attaching a CD4 fusion tag at the end of ZIKV Env, we have increased the protein yield up to 10 times. The resulting recombinant proteins were recognised by anti-sera from ChAdOx1 ZIKV prME_ΔTM vaccinated mice and sera from patients with confirmed exposure to ZIKV. In addition, our ZIKV proteins are Asian-lineage based and presented high purity, whereas commercial ZIKV Env proteins and ELISA kits are based on an African strain. Besides further applications (basic research, diagnoses), our fused or single Env proteins might be useful to evaluate the humoral immune response to ZIKV in upcoming clinical trials of novel ZIKV vaccines with the added value of the correlation of neutralising capacity against ZIKV Env.
Materials and methods
Expression and purification of recombinant ZIKV envelope proteins (Env-CD4 and Env)
The pre-membrane (prM) and envelope (Env) gene sequence for Zika virus were synthesized from Geneart as a DNA fragment based on the sequence of an isolate from the 2013 French Polynesian ZIKV outbreak (GenBank: AHZ13508.1). The ZIKV synthetic gene was codon optimized for efficient expression in mammalian cells. The protein expression and purification was carried out at the Oxford Protein Production Facility (OPPF- UK), Harwell, UK. Briefly, in order to select constructs with best expression profiles in mammalian cells, constructs encoding different lengths of ZIKV prM-Env (amino acid residues 1–589, 1–595, 1–603 and 1–612) and ZIKV Env (amino acid residues 187–589, 187–595, 187–603 and 187–612) were amplified by PCR using
Phusion™
Flash High Fidelity polymerase and cloned into pOPINTTGneo or pOPINTTGneo-3C-CD4 expression vectors via In-Fusion™ cloning and transformed into OmniMaxII T1-phage resistant cells (Invitrogen) in 96-tube format as previously described [
45]. The plasmids were extracted using a QIAgen BioRobot 8000 with the Wizard SV96 minprep kit (Promega). The content of each plasmid was verified by PCR screening using pOPIN forward primer (a standard T7 forward primer) and a gene specific reverse primer and confirmed by DNA sequencing. For both constructs a 6x histidine tag was incorporated at the C-terminus. The expression screening was carried out in HEK293T cells according to the protocol by Nettleship et al. [
24]. Briefly, 1 ml HEK 293 T cells were grown for 24 h in 24-well plates to reach ~ 70% confluency and transfected with each pOPIN construct using the GeneJuice™ transfection reagent (Novagen). To analyse the secretion of proteins into the media, the culture supernatant was harvested 3 days post transfection and analysed by SDS-PAGE using pre-cast Novex ® 4–12% gel in NuPAGE ® MOPS running buffer (Invitrogen). The His-tagged proteins were detected by Western blotting using 6x- His tag antibody (His.H8, Thermo Scientific). Based on the expression screen, the constructs encoding the Zika prM-Env (amino acid residues 1–589) in pOPINTTGneo and pOPINTTGneo-3C-CD4 were chosen for large scale protein expression using the Expi293™ Expression System Kit according to the manufacturer’s protocol. The supernatant containing the Zika prME was harvested after 96 h. Secreted protein was purified by automated immobilized metal affinity chromatography followed by gel filtration chromatography on ÄKTAxpress unit using the method of Nettleship et al. [
24,
25]. Briefly, 200 mL of sample was loaded onto a 5 ml HisTrap FF column (GE Healthcare) before washing with 50 mL of 50 mM Tris, pH 7.5, 500 mM NaCl, 30 mM imidazole. Elution from the HisTrap FF column was performed with 50 mM Tris, pH 7.5, 500 mM NaCl, 500 mM imidazole and the eluted sample was injected directly onto a HiLoad 16/600 Superdex 200 column. Size exclusion chromatography was performed using 20 mM Tris pH 7.5, 200 mM NaCl. Resulting fractions were analysed by SDS-PAGE using InstantBlue protein stain (coomassie stain) and concentrated using an Amicon 400 concentrator with 30 kDa molecular weight cut-off membrane for subsequent immunological studies.
Silver stain and Western blot
Silver staining of purified Zika virus envelope proteins (Env-CD4 and Env) from this study and commercially available recombinant Zika virus envelope protein expressed in insect cells (Aalto Bio reagents No. AZ6312) was carried out to assess purity according to the standard protocol using the Pierce Silver Stain Kit (No. 24612). Purified ZIKV envelope proteins were added to Laemmli sample buffer containing 20 mM β-mercaptoethanol. Samples were boiled for 3 mins and loaded on a Mini-PROTEAN TGX protein gel (reducing conditions) with a protein marker (Bio-Rad, Precision Plus Protein™ WesternC™ standards). For samples under non-reducing conditions, non-reducing sample buffer was used instead. For silver stain, PAGE gels were stained with Pierce Silver Stain Kit according to the manufacturer’s protocol. For Western blot, proteins were transferred onto nitrocellulose membranes (Bio-Rad Trans-Blot® TurboT
M). Membranes were blocked with 5% skim milk in 0.1% PBS/T for 1 h and then incubated with one of the following antibodies; (1). 1:500 dilutions of a mouse anti-flavivirus group antigen monoclonal antibody (catalogue no. MAB10216, Millpore) (2). 1:500 dilutions of mice sera following a vaccination with ChAdOx1 ZIKV prME_ΔTM [
30] (3). 1:1000 anti-Zika Env monoclonal antibody (mouse mAb to Zika Env protein, AZ1176, Aalto BioReagents) (4). 1:2000 anti-His antibody (mouse 6x-His Tag Antibody, MA1–21315, ThermoFisher Scientific). After washing with PBS/T, membranes were incubated for 1 h with a HRP-conjugated goat anti-mouse IgG (Bio-Rad Cat. 170–6516). Finally, membranes were washed again using PBS/T and incubated with a chemiluminescent substrate (Clarity™ Western ECL Blotting Substrates, BIO-RAD); signal was detected using a chemiluminescent Western blot imaging system (Image Lab, Bio-Rad).
Immunisation of mice
Mice were vaccinated intramuscularly with a single dose of chimpanzee adenoviral vectored vaccine (ChAdOx1) encoding ZIKV prME_ΔTM or unrelated malarial antigen (cCSP) as control vaccines at dose of 1 × 10
8 infectious units, as described [
30].
Mouse sera IgG ELISA
Anti-Zika envelope antibody concentrations were measured by a specific IgG enzyme-linked immunosorbent assay (ELISA) to recombinant Zika Env antigens. Briefly, Nunc Maxisorp Immuno ELISA plates were coated with Zika virus envelope antigens (commercial, Env-CD4 and Env) diluted in PBS to a final concentration of 2 μg/ml and left at RT overnight. Plates were washed 6 times with PBS/0.05% Tween (PBS/T) and blocked with 300 μl with Pierce™ protein-free (PBS) Blocking buffer (ThermoFisher) for 2 h at RT. Mice sera were obtained 3 months after single vaccination with ChAdOx1 ZIKV prME_ΔTM and unrelated control (cCSP) [
30]. Mice sera reactive to ZIKV Env or reactive to an unrelated control antigen (cCSP) were added and serially diluted 3-fold down in PBS/T with 50 μl per well as final volume and incubated for 2 h at RT. Following washing 6 times with PBS/T, bound antibodies were detected following a 1 h incubation with 50 μl of alkaline phosphatase-conjugated antibodies specific for whole mouse IgG (Sigma, A3562-5ML). Following a further 6 washes with PBS/T, development was achieved using 100 μl of 4-nitrophenylphosphate diluted in diethanolamine buffer and the absorbance values at OD405 were measured and analysed using a CLARIOstar instrument (BMG Labtech). Serum antibody endpoint titers were defined by an absorbance value three standard deviations greater than the average OD405 of the control (cCSP) at 1:450 sera dilution.
Human sera IgG ELISA
A total of sixteen sera from patients with RT-PCR confirmed Zika virus infections from Mexico were used for an IgG ELISA assay to measure the antibodies against recombinant Zika virus envelope antigens (Env-CD4). As controls, serum samples from ten healthy individuals from a non-endemic area were used, but no information on previous vaccination, travel to flavivirus endemic area or previous disease was available. Briefly, human sera were diluted in Nunc Maxisorp Immuno ELISA plates coated with Zika virus envelope antigens (Env-CD4 or commercial Env) in PBS to a final concentration of 5 μg/ml. Following blocking, the sera was incubated for 1 h, then plates were washed 6 times with PBS/T (0.05%), bound antibodies were detected by using a goat Anti-Human IgG-alkaline phosphatase-conjugated antibody (Sigma, A3187-5ML). Development was performed using 4-nitrophenylphosphate diluted in diethanolamine buffer and absorbance values at OD405 were measured on a CLARIOstar instrument (BMG Labtech). Serum antibody endpoint titers were defined by absorbance value three standard deviations greater than the average OD405 of control sera pool at 1:300 sera dilution.
Zika plaque reduction and neutralisation assay
Sixteen sera from Mexican patients and ten sera from healthy donors were tested in duplicate by a Zika plaque reduction and neutralisation test (PRNT). 60 pfu of the ZIKV Asian strain (PRVABC59) were mixed with doubling dilutions of human sera ranging from 1:10 to 1:160 and incubated at 37 °C for 1 h to allow the serum to neutralise the virus. This mixture was then added to a confluent monolayer of Vero cells that had been seeded onto a 12 well plate at a density of 2 × 105 cells per well. After a 1.5 h incubation at room temperature, the viral inoculum was removed and cells were overlaid with 1 ml of a 1% agarose solution containing 2% HI-FCS. The plates were then incubated for 4 days at 37 °C, 5% CO2, 95% humidity. Cells were fixed and stained with 1 ml of toluidine blue solution (0.1% toluidine blue, 2.7% formaldehyde, 1x PBS) and incubated at RT overnight to allow the stain to soak through the agarose. The plaque number per treatment was compared to the plaque number in the control samples and neutralising antibody titers were expressed as the serum dilution yielding a 50% plaque number reduction (ND50). ND50 titers were calculated for each serum using the Spearman-Karber formula. The cut off titer for positivity was set as 10.
Data analysis and statistics
All data analysis and statistics were performed using prism 7 software (GraphPad, Software, US). P values and R2 values reflect Pearson correlation tests for determining the correlation between the endpoint reciprocal titer values from ELISA assay and ND50 titers from PRNT assay. P value (< 0.0001) between Env-CD4 and commercial Env was determined by pairwise t-tests.
Acknowledgements
We are grateful to Mark Tilley for helpful comments on the manuscript. A. R-S and C.S.R. are Jenner Investigators and Oxford Martin Fellows. This work was supported by the UK Department of Health through Innovate UK “New vaccines for global epidemics: development and manufacture” grant No. 972216.