LTB was obtained after cloning the nucleotide sequence of the eltB gene (encoding the B subunit of the LT from H10407 strain) into the pET22b vector. The eltB sequence (450 bp) was amplified by PCR using the forward primer 5′-CAGAAGGCGGAATGACATATGAATAA-3′ and reverse primer 5′-TTAAAAGCATGTCTCTCGAGAGAACC-3′ (the underlined sequences indicate the NdeI and XhoI restriction sites, respectively). The insert generated by the PCR reaction was digested and ligated into the corresponding NdeI and XhoI restriction sites of the pET22b expression vector (Novagen-Merck Bioscience, MA, USA), generating the recombinant plasmid pET22bLTB, which was subsequently used to transform chemically competent E. coli DH5α. Transformants were analyzed by digestion with the enzymes NdeI and XhoI. The recombinant DNA fragment was sequenced and compared to data reported for the LT sequence (GenBank database GI: 408994). A chemically competent E. coli BL21 (DE3) strain was transformed with the pET22bLTB plasmid and denominated as E. coli BL-LTB.

The LT and LTB adjuvants were obtained from E. coli strains harboring the plasmid pML19 (47, 48) or pET22bLTB, respectively. The purification of LT was based on a previously described method (47). The recombinant E. coli strains were cultivated in Luria-Bertani (LB) medium or Terrific Broth (TB) supplemented with 100 µg/ml of ampicillin at 37°C in an orbital shaker (200 rpm). While LTB expression was obtained after 0.75 mM IPTG induction of the E. coli BL-LTB strain cultures for 6 h, LT production was achieved by constitutive expression of the eltAB gene in overnight cultures of a pML19-harboring E. coli strain. The cell pellets were suspended, and the bacterial cells were disrupted by high pressure in a homogenizer (APLAB-10 model, ARTEPEÇAS, Brazil). The LT and LTB contained in the bacterial cell homogenates were purified by affinity chromatography on immobilized d-galactose columns (Pierce, Waltham, MA, USA) previously equilibrated with TEAN buffer (50 mM Tris, 1 mM EDTA, 3 mM NaN₃, and 0.2 M NaCl, pH 7.4) in an Akta FPLC (Amersham Pharmacia Biotech, Buckinghamshire, UK). The elution step was performed with TEAN supplemented with 0.3 M d-galactose. The proteins were monitored by 15% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The concentrations of the purified proteins were determined by absorbance measurements in a spectrophotometer (Gene Quant spectrophotometer GE Amersham Biosciences, Buckinghamshire, UK) as previously described (49).

The pE1D2 plasmid (50), harboring gene encoding the ectodomain of the envelope glycoprotein from the DENV2 New Guinea C (NGC) strain, was kindly provided by Dr. Ada Alves and used as template for PCR reactions. The sequences of the forward and reverse primers used in PCR reactions were 5′-ACATGCGAGGATCCGGAATGTCATACTCTAT-3′ (the underlined sequence indicates the BamHI restriction site) and 5′-GCCTTCTACTCGAGTTACGATAGAACTTCCTTTCTTA-3′ (the underlined sequence indicates the XhoI restriction site), respectively. The PCR product was introduced into pET28a(+) expression vector (Novagen, Germany), generating the recombinant plasmid pDEDIII. The E. coli BL21 (DE3) strain harboring the plasmid pDEDIII was cultivated in LB medium containing 50 µg/ml kanamycin. The culture at an OD_600nm of 0.5 was induced by IPTG to a final concentration of 0.5 mM and incubated for 4 h. The bacterial pellet was suspended in buffer A (0.1 M Na₂PO₄, 0.5 M NaCl, pH 6.8) and lysed in the APLAB-10 homogenizer (ARTEPEÇAS, Brazil). The inclusion bodies were solubilized in buffer A supplemented with 8 M urea, and the recombinant protein was refolded in buffer A. The soluble protein was submitted to nickel affinity chromatography using a HisTrap™ HP column (GE Healthcare Life Sciences, Buckinghamshire, UK). The purified EDIII was evaluated by 15% SDS-PAGE and measured by spectrophotometry.

The endotoxin was removed using a Detoxi-gel Endotoxin Removing Gel (Pierce, Waltham, MA, USA), and any residual endotoxin was detected by the Limulus Amebocyte Lysate (LAL) QCL 1000 kit (Lonza, Basel, Switzerland). Following treatment for LPS removal, the residual endotoxin contents in the LT, LTB, and EDIII samples were less than 0.1 EU/μg of protein, corresponding to the levels allowed for preclinical research (51).

The cytotonic activity of the LT derivatives on the adrenal Y1 cells was evaluated as previously described (6). The Y1 cells were seeded in 96-well plates (5 × 10⁴ cells/well) and exposed to 1 µg of LT or LTB, or to phosphate-buffered saline (PBS) as negative control, diluted in Dulbecco’s modified Eagle’s medium (DMEM, Vitrocell Embriolife, Brazil) supplemented with 2% fetal bovine serum (FBS, Vitrocell Embriolife, Brazil). After 8 h of incubation, the cells were observed for the cytotonic effect. The GM1-ELISA was performed as previously described (52) to evaluate LT binding to the cognate ganglioside GM1. Microtiter plates (Polysorp, Thermo Fisher-Nunc, Roskilde, Denmark) were coated with GM1 ganglioside (1 µg/ml) (Sigma-Aldrich, St. Louis, MO, USA) diluted in PBS (pH 7.4) and incubated overnight at room temperature in a humid chamber. Plates were blocked with PBS containing 0.1% BSA. Then, serially twofold diluted LT samples were applied in duplicate wells. To detect LT derivatives, the plates were sequentially incubated with anti-LT serum (titer equal to 10⁵) and horseradish peroxidase (HRP)-conjugated anti-mouse IgG antibodies (Sigma-Aldrich) diluted to 1:1,000 and 1:3,000, respectively, in PBS containing 0.05% Tween-20 and 0.1% BSA. The chromogenic reactions developed with ortho-phenylenediamine dihydrochloride (Sigma-Aldrich) and H₂O₂ were stopped with 1 M H₂SO₄ and measured at A_492nm.

This assay was performed as previously described (53). African Green Monkey Kidney Epithelial Cells (Vero Line), which express receptors for E glycoprotein, were cultured in minimum essential medium (MEM) supplemented with 2% FBS and incubated at 37°C in an atmosphere containing 5% CO₂ for 24 h. Cells were incubated with 100 µg/ml of EDIII protein in MEM at 37°C in 5% CO₂ for 30 min. After washing with PBS with 2% FBS, Vero cells were fixed with 4% paraformaldehyde (Sigma-Aldrich, St. Louis, MO, USA) for 10 min, treated with anti-EDIII sera harvested from mice immunized with EDIII plus Freund’s adjuvant for 60 min and subsequently with goat anti-mouse IgG conjugated to FITC (Invitrogen, Paisley, UK) for further 60 min. The FITC-stained Vero cells were detected by BD LSRFortessa flow cytometry (BD Bioscience, San Jose, CA, USA), and the data were analyzed using the FlowJo v10 program.

All experiments involving mice were approved by the Committee on the Ethical Use of Laboratory Animals of the Institute of Biomedical Sciences (CEUA 198) at the University of São Paulo (USP), in accordance with the guidelines for the care and use of laboratory animals adopted by the National Council of Animal Experimentation (CONCEA). The mice were obtained from the Isogenic Mouse Breeding Facility of Department of Parasitology, Institute of Biomedical Sciences—USP. Female BALB/c mice (6–8 weeks old) were immunized via the subcutaneous (s.c.) route at days 0, 14, and 28 of the immunization protocol with the vaccine formulations. Groups of five BALB/c mice were treated with 100 µl of PBS alone as a negative control or PBS containing 10 µg of EDIII with or without one of the following adjuvants: 1.0 µg of LT, 3.2 µg of LTB, or 12.5 µg of Al(OH)₃ (Rehydragel, Reheis, NJ, USA). Serum samples were harvested 2 weeks after each vaccine dose and stored at −20°C until immunological analyses were conducted.

ELISA for anti-EDIII antibodies was performed as previously described (53). Microtiter plates (Maxisorp Thermo Fisher-Nunc, Roskilde, Denmark) were coated with EDIII (1 µg/ml) diluted in PBS pH 7.4 and incubated overnight at 4°C. For ELISA using denaturing conditions, the EDIII protein was submitted to heat treatment at 100°C for 15 min and used to coat the plates. Plates were blocked with PBS containing 3% gelatin for 2 h at 37°C. The serum samples harvested from mice immunized with EDIII alone or coadministered with each adjuvant were applied to duplicate wells and serially twofold diluted. The serum IgGs were measured with HRP-conjugated anti-mouse IgG (1:3,000), IgG1 (1:10,000), or IgG2a (1:3,000) antibodies (Sigma-Aldrich, St. Louis, MO, USA). Absorbance at 492 nm was obtained as previously described and used to calculate the antibody titers, defined as the highest sample dilution able to generate an A_492nm of 0.2 above the preimmune sera.

Anti-EDIII IgG antibody fractions were obtained from serum samples harvested from mice submitted to EDIII-based immunization regimens. First, the recombinant 6×-His-tagged EDIII was immobilized in a column filled with a nickel-coupled resin. After serum delipidation, as previously described (54), the sera were submitted to affinity chromatography on EDIII-immobilized resin using 0.05 M PBS, pH 7.4. The anti-EDIII antibodies were eluted in 0.1 M glycine buffer (pH 4), neutralized with 1 M Tris–HCl (pH 9.0), and dialyzed in PBS (pH 7.4). Subsequently, the anti-EDIII antibodies were subjected to affinity chromatography to protein G sepharose according to the manufacturer’s instructions (GE Healthcare, Buckinghamshire, UK), and the IgG fraction was obtained following the elution step described above. The purified EDIII-specific IgG antibodies were measured by BCA assay (Pierce, Waltham, MA, USA) and monitored by 12.5% SDS-PAGE.

The antibody affinity index was measured using a modified ELISA protocol with ammonium thiocyanate (47). Briefly, the plates were coated with EDIII as described above. The sera collected from mice from different immunization groups were tested at dilutions corresponding to an OD_492nm of 0.8, while purified anti-EDIII IgG antibodies were evaluated at an OD_492nm of 0.5. After incubation with sera or purified antibodies, different concentrations of sodium thiocyanate were added to wells and incubated for 15 min. The plates were washed and incubated with HRP-conjugated anti-mouse IgG antibody. The percentage of antibodies bound to EDIII was determined: OD_492nm in the presence of ammonium thiocyanate × 100/OD_492nm in the absence of ammonium thiocyanate. The values obtained with the serum samples represent the antibody avidity, which under our conditions depends on the affinity and valence of the immunoglobulins since the parameter of structural arrangement was standardized by the use of the antigen immobilized on the solid support. The analyses of the purified IgG antibodies showing constant Ig valence generated average affinity values for each immunization group.

To perform the PRNT, the DENV2 NGC strain, kindly provided by Laura Helen Vera Gonzales Gil (Fiocruz, Recife, Brazil), was purified, and the Vero cells were cultivated as previously described (35). Briefly, the NGC strain was propagated in the Aedes albopictus cell line C6/36 cultured in Leibovitz medium containing 5% FBS and grown at 28°C for 7 days. The supernatant of C6/36 infected with the NGC strain was collected and titrated for the determination of PFU/ml. For the neutralization test, Vero cells were maintained in DMEM supplemented with 10% FBS and seeded in 24-well plates (1 × 10⁵ cells/well) 24 h before infection. Serum samples from immunized mice were inactivated for complement proteins for 30 min at 56°C and serially twofold diluted. Purified EDIII-specific IgGs were tested at different concentrations. C6/36 cell supernatants containing 500 PFU of DENV/ml were incubated with each dilution of anti-EDIII sera or purified IgG antibodies for 1 h at 37°C. Vero cells were washed with serum-free DMEM and infected in duplicate wells with 200 µl of the neutralization mixture for 1 h at 37°C. The viral suspension was removed, and the cells were overlaid with 1 ml of complete DMEM [2.5% FBS and 1% carboxymethylcellulose (Synth, São Paulo, Brazil)]. Following incubation of the plates at 37°C for 7 days, the cells were fixed for 15 min with 4% paraformaldehyde and stained with crystal violet for 10 min. The percentage of plaque reduction compared to the positive control (DENV not exposed to antibodies) was calculated. Neutralizing antibody titers were expressed as the serum dilution, or IgG concentration, yielding 50% plaque number reduction (PRNT₅₀).

For the peptide microarray assay, slides were manufactured using in situ synthesis of 330,000 (CIM330K) sequences of non-natural randomized peptides as previously described (40). Statistical analyses of microarray data were performed using JMP 12 software (Statistical Discovery Software from SAS) by importing data analysis from Excel and image-processed data from GenePix Pro-6.0 (Molecular Devices, Sunnyvale, CA, USA). Raw intensities were normalized to each slide by dividing all values per array by the median of that array (median normalization). Poor quality spots were excluded from analysis by flagging them as “absent” upon visual inspection. Values less than 0.01 were set to 0.01, and values from duplicate arrays were averaged and used in the analysis. Important peptides were determined by Student’s two-tailed t-test, with Benjamini and Hochberg multiple test correction applied to a p value < 0.05. For principal component analysis, we used significant peptides for each vaccine approach and an analysis using JMP 12 software (39). The prediction of recognized epitopes was based on lists of informative peptides compared to the EDIII sequence from the DENV2 strain NGC (AHG97599.1) using GuiTope (55). The amino acid substitution matrix was provided by the peptide library and protein sequence. An inversion weight of 1 was selected. Library subtracted scores were returned using the mean of 10 random samplings of the peptide library and a minimum score cutoff of 8.0. The results were graphed using GraphPad Prism v5 using a moving average of 15.

This assay was adapted from a protocol previously described (56, 57). Vero cells (1 × 10⁵ cells/well, on 96-well plates) were cultured in MEM supplemented with 3% FBS and incubated at 37°C in an atmosphere containing 5% CO₂ for 18 h. Intact or heat-denatured EDIII protein, as well as peptide 47 (369-AEPPFGDSYIIIGVEPGQLK-388), custom-made by GeneScript (Piscataway, NJ, USA), at different molar concentrations, was incubated with the DENV2 strain NGC (multiplicity of infection—equal to 1.0). The mixtures were added to Vero cells at 37°C in 5% CO₂ for 1 h. The mixtures were removed, and MEM supplemented with 2% FBS was added to each well. Following 18 h of incubation, the culture plates were washed with PBS and treated with 50 µl/well of trypsin (2.5 mg/ml) plus EDTA at 37°C in 5% CO₂ for 10 min to remove the cells. Vero cells were resuspended with PBS containing 10% FBS and fixed and permeabilized with Cytofix/Citoperm kit (BD Bioscience, San Jose, CA, USA) for 10 min. The cells were incubated for 1 h with mAb 4G2, which recognizes the fusion loop at the extremity of the domain II of the envelope protein (ATCC-HB112), and subsequently labeled with anti-mouse IgG conjugated with Alexa 488 (BD Biosciences) for 30 min. The stained cells were measured by flow cytometry using a BD LSRFortessa (BD Bioscience, San Jose, CA, USA) instrument, and the data were analyzed with FlowJo v10 software to determine the amount of DENV-positive Vero cells. Viral replication analysis was carried out at 24 h post-infection based on a protocol previously described (58). RNA was extracted from 100 µl of infected Vero cell, using a guanidine isothiocyanate phenol method (Trizol LS, Invitrogen), according to the manufacturer’s instructions. The quantitative real-time polymerase-chain-reaction was carried out with a set of primers and probes with FAM as dye reporter for the probe. Primers/probes used in assay for DENV were previously described (58). The assay was performed using the AgPath-IDTM One-Step RT-PCR reagents (Applied Biosystems). Succinctly, we used 5 µl of the extracted RNA in 1 µl of the mix from primers/probe (10 pM/μl) and 19 µl of the reagent mix from AgPath-IDTM One-Step RT-PCR kit following the manufacturer’s instructions.

The amino acid sequence composing the main predicted epitope was indicated in a previously published structural model of the quaternary DENV2 envelope glycoprotein (PDB accession number 1OKE) using PyMol v1.8.4.0.

The data are represented as arithmetic means ± SD and analyzed for variance (ANOVA) with a subsequent Bonferroni’s multiple comparison test using GraphPad Prism v5 software. For the data generated by the PRNT and affinity assay, the statistical analyses were calculated from two independent experiments, while all other results were analyzed from three independent experiments.

The recombinant forms of LT, LTB and type 2 DENV EDIII were purified by affinity chromatography (Figures S1 and S2 in Supplementary Material). Both LT and LTB bind to host cell receptors, but as expected, only LT had cytotonic effects to Y1 cells (Figures S1C,D in Supplementary Material). The recombinant EDIII also preserved the receptor-binding function of the native viral protein (Figure S2B in Supplementary Material). Mouse groups were immunized with purified EDIII and vaccine formulations containing EDIII admixed with alum, LT, or LTB. Mice received three s.c. doses at intervals of 2 weeks and were bled one day before each dose and 2 weeks after the last vaccine dose (Figure 1A). Significant increases in the EDIII-specific IgG antibody levels were detected in serum samples of mice immunized with vaccines containing LT and LTB, but not with alum compared to EDIII alone (Figures 1B,C). The antigen-specific IgG1/IgG2a ratios in vaccinated mice were similar, indicating that the adjuvants were not able to modulate the IgG subclass responses induced by the antigen alone. Although anti-EDIII IgG1 and IgG2a titers detected in mice immunized with LT and LTB enhanced in comparison with the non-adjuvanted EDIII, significant differences were only found regarding to IgG1 levels (Figure 1D). In addition, EDIII-specific antibodies in serum samples collected from vaccinated mice were capable of recognizing both conformational and linear epitopes of the antigen, as demonstrated by reactivity with both intact and heat-denatured antigen. Serum samples from mice immunized with LT or LTB showed significantly higher reactivity with both conformational and linear epitopes than those collected from mice immunized with non-adjuvanted EDIII or EDIII admixed with alum (Figures 1E,F).

To evaluate the role of adjuvants in the modulation of the effector activities of antigen-directed antibodies induced by subunit vaccines, EDIII-specific antibodies raised in different mouse groups were purified by affinity chromatography (Figure S3 in Supplementary Material). The purified anti-EDIII antibodies reacted with the native antigen but did not recognize the heat-denatured antigen, suggesting that the purified antibodies are specifically directed toward conformational epitopes (Figure 2A; data not shown). No difference in antigen affinity was detected in EDIII-specific antibodies raised in the different immunization groups (Figure 2B). Anti-EDIII antibodies, either in serum or purified fractions from vaccinated mice showed virus-neutralization activity in vitro; however, anti-EDIII antibodies elicited with LT or LTB were more efficient than those raised with antigen alone or admixed to alum (Figures 2C,D; Table S1 in Supplementary Material). The mice immunized with LTB coadministered to EDIII elicited antigen-specific sera with twofold higher neutralizing capability than LT-adjuvanted vaccination (Figure 2C; Table S1 in Supplementary Material). The purified anti-EDIII IgG from both LT and LTB-adjuvanted mice yielded neutralization titers (PRNT₅₀) of 1.6 ng/ml and thereby revealed strong virus-neutralizing effects (Figure 2D).

To gain a better understanding of the enhanced virus-neutralization activity of the anti-EDIII antibodies raised in mice immunized with different vaccine formulations, we determined the immunosignatures of purified EDIII-specific antibodies raised in different immunization groups (Figure 3). By profiling the antibodies with random linear peptide sequences spotted on microarray slides, we observed that antibodies from mice immunized with EDIII and LT or LTB had distinct immunological signatures for those raised in mice immunized without adjuvant or with alum (Figures 3A,B). In addition, fewer peptides were recognized with high binding intensity by EDIII-specific IgGs adjuvanted with LT or LTB, suggesting an increased specificity of these antibodies (Figures 3A,C). Analysis of peptide-binding profiles indicated that antibodies raised in mice immunized with LT and LTB reacted with 119 peptides sequences each, while antibodies raised in mice immunized only with antigen, or with alum admixed with the antigen, reacted with 605 and 603 peptides, respectively. Among these peptides, only 10 were solely recognized by antibodies raised in mice immunized with LT or LTB and 88 peptides among all groups (Figure 3C).

Prediction of epitopes present in the EDIII antigen and recognized by the different set of purified IgG antibodies showed that mice immunized with both LT derivatives reacted with a specific sequence comprising the EF loop, F beta-sheet, and FG loop between amino acid residues 371–384 (Figures 3D and 4A,B). To further investigate the biological relevance of this finding, we tested a 21-mer synthetic peptide encompassing the complete predicted epitope of the EDIII antigen (Figure 4B). As shown in Figure 4C and Figure S4 in Supplementary Material, the peptide (EDIII_369–388) efficiently blocked the infectivity of DENV2 NGC strain in Vero cells. A similar inhibition of DENV infectious activity was achieved with the purified EDIII antigen, whereas the heat-denatured antigen lost the virus-infection inhibitory effect.