A fragment of C-CPE (GenBank accession no. M98037.1) was cloned into pET16b (Novagen, Darmstadt, Germany) to prepare pET16b-C-CPE, as described previously [10] to be used for ELISA. A fragment of Stx2B (GenBank accession no. NP_050540.1) was cloned into pCold I (Takara Bio, Shiga, Japan), as described previously [1] to be used for ELISA. The Stx2B-C-CPE fragment was obtained from pCold I-Stx2B-C-CPE [1] and cloned into pET9a (Takara Bio). Briefly, Stx2B-C-CPE DNA was amplified through polymerase chain reaction (PCR) by using specific primers (forward primer, 5${}^{\prime }$-GGGCATATGAAGAAGATGTTTATGGCGGTT-3${}^{\prime }$ [NdeI site underlined]; reverse primer, 5${}^{\prime }$-GGGGATCCTTAAAATTTTTGAAATAATATTGAAT-3${}^{\prime }$ [BamHI site underlined]). The Stx2B-C-CPE PCR product was digested with NdeI (New England Biolabs Japan, Sumidaku, Japan) and BamHI (New England Biolabs Japan) and inserted into pET9a by using T4 DNA ligase (New England Biolabs Japan) to produce pET9a-Stx2B-C-CPE (Supplementary Fig. 1).

Recombinant histidine-tagged C-CPE and Stx2B proteins were prepared as described previously [1]. To obtain recombinant non-tagged Stx2B-C-CPE protein, the pET9a-Stx2B-C-CPE plasmid was transformed into E. coli strain BL21(DE3) (Toyobo, Osaka, Japan) and cultured in Terrific broth (Invitrogen, Carlsbad, CA, USA) supplemented with 50 $\mu$g/mL kanamycin (Nacalai Tesque, Kyoto, Japan) at 37 °C or 25 °C for 6 h or 24 h without isopropyl β-D-thiogalactopyranoside (IPTG) induction because of lacking of lac operon. After centrifugation at 17,800 $\times$g at 4 °C for 5 min, the supernatant was filtered by using Millex-GV filter unit (membrane, polyvinylidene; pore size, 0.22 $\mu$m; MilliporeSigma, Burlington, MA, USA). The filtered supernatant was mixed with a saturated solution of ammonium sulfate (Wako Pure Chemical, Osaka, Japan) at various ratios from 1:1 to 1:0.2, stirred for 10 min at room temperature (25 °C), and then allowed to stand for 60 min at room temperature (25 °C). The crude purified protein was obtained by centrifugation at 10,000 $\times$g at 4 °C for 20 min. After buffer exchange to 20 mM Tris-HCl, pH 8.0, by using a PD-10 column (GE Healthcare, Wakesha, WI, USA), proteins were separated through anion exchange chromatography by using an AKTAprime Plus system (GE Healthcare) with a HiTrap Q HP column (Cytiva, Marlborough, MA, USA) under the following conditions: buffer, 20 mM Tris-HCl, pH 8.0, to 20 mM Tris-HCl–1 M NaCl, pH 8.0; flow rate, 1 mL/min; equilibration volume, 5 mL; sample volume, 3.5 mL; wash volume, 5 mL; elution volume per fraction, 1 mL. Fractions 10 and 11 were pooled and loaded onto a PD-10 column for buffer exchange to phosphate-buffered saline (PBS). The concentration of the purified protein was measured by using a BCA Protein Assay kit (Life Technologies, Carlsbad, CA, USA). The purity of the eluted protein was confirmed by using the NuPAGE electrophoresis system (Life Technologies) followed by staining with Coomassie Brilliant Blue.

Endotoxin was removed by using Acrodisc units containing Mustang E membrane (Pall, Port Washington, NY, USA), which is a positively charged polyethersulfone membrane with high binding capacity for endotoxins, in accordance with the manufacturer’s instructions or by using Triton X-114 phase separation, which is a liquid-liquid extraction method [11]. For the phase separation technique, 500 $\mu$L of Stx2B-C-CPE was mixed with 5 $\mu$L of Triton X-114 (MP Biomedicals, Santa Ana, CA, USA) by vortexing, incubated on ice for 5 min, mixed by vortexing, and then incubated at 37 °C for 5 min. After centrifugation at 3000 $\times$g at 25 °C for 3 min, the supernatant was collected. The endotoxin removal procedure was repeated to obtain endotoxin-free protein. To remove the remaining detergent, the buffer was exchanged to PBS by using a PD-10 column. Endotoxin activity was measured by using the ToxinSensor chromogenic limulus amebocyte lysate endotoxin assay kit (GenScript Biotech, Piscataway, NJ, USA) in accordance with the manufacturer’s instructions.

Female BALB/c mice (age, 7 wk) were purchased from CLEA Japan (Tokyo, Japan) and used for experiments after pre-breeding for 1 wk. All experiments were approved by the Animal Care and Use Committee of the National Institutes of Biomedical Innovation, Health and Nutrition (Approval No. DSR01-1R2) and conducted in accordance with their guidelines. Mice were subcutaneously immunized with PBS (vehicle control), Stx2B-C-CPE, endotoxin-free Stx2B-C-CPE without or with aluminum adjuvant (BIKEN, Osaka, Japan), endotoxin-free Stx2B-C-CPE without or with Alcaligenes lipid A (Peptide Institute, Osaka, Japan) [9], or endotoxin-free Stx2B-C-CPE without or with both aluminum adjuvant and Alcaligenes lipid A once weekly for 2 consecutive weeks. Serum was collected 1 week after the last immunization. The immunization schedule followed the previous study [1].

We coated 96-well immunoplates (Thermo Fisher Scientific, Waltham, MA, USA) with Stx2B or C-CPE (0.5 $\mu$g per well) in PBS at 4 °C overnight. After the coating solution was removed, the plates were treated with 1% (w/v) BSA in PBS for 2 h at room temperature. The plates then underwent the following four steps, each of which was preceded by a wash with 0.05% (v/v) Tween-20 in PBS: (i) serial dilutions of serum were added to the wells, and the plates were incubated for 2 h at room temperature; (ii) goat anti-mouse IgG conjugated with horseradish peroxidase (Southern Biotech, Birmingham, AL, USA) was added, and the plates were incubated for 1 h at room temperature; (iii) tetramethylbenzidine peroxidase substrate was added to wells, and the plates were incubated for 2 min at room temperature; and (iv) 0.5 M HCl (Nacalai Tesque) was added to the sample wells, and the absorbance at 450 nm was measured in an iMark microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).

Statistical significance was evaluated through one-way ANOVA for comparison of multiple groups and the unpaired t-test for two groups by using Prism 7 (GraphPad Software, La Jolla, CA, USA). A p value less than 0.05 was considered to be significant.

We previously reported the fusion protein Stx2B-C-CPE as a bivalent vaccine to induce protective immunity against Clostridium perfringens and Shiga toxin (Stx)-producing Escherichia coli (STEC) infections and thus prevent food poisoning [1]. To establish an antigen preparation method for commercial vaccine production, we cloned a target fusion gene without any tag or other protein-modifying sequences into pET9a vector (Supplementary Fig. 1) and then examined the experimental conditions for protein expression. A strong signal due to Stx2B-C-CPE was detected at 24 h rather than 6 h in the cell pellet of E. coli strain BL21 (DE3) cultured in Terrific broth at both 37 °C and 25 °C (Fig. 1A). Because this fusion protein retains the signal peptide present in native Stx2B, Stx2B-C-CPE was secreted from E. coli cells and detected in the culture supernatant (Fig. 1A). The amount of impurities, such as the bands at 28–98 kDa, appeared to be less when the cells were grown at 25 °C than at 37 °C (Fig. 1A). We therefore considered that the supernatant obtained after culturing at 25 °C for 24 h is suitable for the next step in our protein purification process.

Fig. 1.

Stx2B-C-CPE antigen preparation. (A) Expression of Stx2B-C-CPE. Stx2B-C-CPE was expressed by using the pET9a system, E. coli B21(DE3) cells, and Terrific broth (Invitrogen). After incubation for 6 h at 37 °C or for 24 h at 25 °C, Stx2B-C-CPE was detected in the centrifuged bacterial pellets (15 $\mu$g of bacterial wet weight/lane) and cultured supernatants (10 $\mu$L) by using NuPAGE electrophoresis. Marker, SeeBlue Plus 2 pre-stained protein standard (Invitrogen). (B) Precipitation of Stx2B-C-CPE. The supernatant of the culture grown for 24 h at 25 °C was mixed at various ratios with saturated ammonium sulfate solution, proteins were pelleted by centrifugation, and the purity of the precipitated sample was assessed by using NuPAGE electrophoresis. Marker, SeeBlue Plus 2 pre-stained protein standard (Invitrogen). (C) Purification of Stx2B-C-CPE. The precipitated proteins were separated into fractions through anion-exchange chromatography (AKTA System with HiTrap Q HP column, GE Healthcare/Cytiva). Red arrows indicate Stx2B-C-CPE (band or fraction). The results shown are representative of two independent experiments.

Because of the large volume of the culture supernatant, we next addressed conditions for concentrating the protein. Addition of the saturated ammonium sulfate solution resulted in Stx2B-C-CPE protein precipitation, and a concentration of 37.5% (supernatant: saturated ammonium sulfate solution, 1:0.6) appeared to be the minimum for efficient protein recovery (Fig. 1B). Next, Stx2B-C-CPE was purified by using anion-exchange chromatography with a NaCl concentration gradient. Stx2B-C-CPE was eluted in fractions 10 and 11 (Fig. 1C). Thus, we were able to optimize the experimental conditions for the expression and purification of tag-free Stx2B-C-CPE.

Because we used an E. coli expression system to prepare Stx2B-C-CPE, contamination of the preparation by bacteria-derived components such as LPS (i.e., endotoxin) is likely. In fact, the Stx2B-C-CPE solution showed high endotoxin activity (Fig. 2A). However, when we used a filtration system to remove the contaminating endotoxin, the amount of Stx2B-C-CPE protein was greatly decreased also, probably because it was adsorbed on the membrane (Supplementary Fig. 2). These findings indicate that the filtration system we used is unsuitable for removing endotoxin from preparations of Stx2B-C-CPE.

Fig. 2.

Removal of contaminating endotoxin from Stx2B-C-CPE. (A) Endotoxin activity in the Stx2B-C-CPE preparation. Endotoxin was extracted from Stx2B-C-CPE by using Triton X-114 phase separation, and the supernatant was collected as the endotoxin-removed sample fraction. This procedure was repeated twice. The endotoxin activity in the Stx2B-C-CPE preparation before and after endotoxin removal was measured by using ToxinSensor chromogenic LAL endotoxin assay kit (GenScript). The endotoxin-free reference value is 0.05 EU/mL. Data are means $\pm$ SEM (n = 3). **p 0.01 (one-way ANOVA). (B) Protein content of the Stx2B-C-CPE preparation. The protein concentration of the Stx2B-C-CPE preparation before and after endotoxin removal was measured by using BCA Protein Assay kit (Life Technologies). Data are means $\pm$ SEM (n = 3). **p 0.01, *p 0.05 (one-way ANOVA). (C) NuPAGE gel electrophoresis of endotoxin-depleted Stx2B-C-CPE. Marker, SeeBlue Plus 2 pre-stained protein standard (Invitrogen). The results shown are representative of two independent experiments.

We then tried Triton X-114 phase separation [11] as a means to remove contaminating endotoxin from preparations of Stx2B-C-CPE. Two rounds of Triton X-114 treatment reduced the endotoxin activity in Stx2B-C-CPE solution to below the limit of detection and below the industry standard of 0.05 EU/mL (Fig. 2A), whereas, although the protein was slightly reduced, Stx2B-C-CPE protein recovery remained high and in excess of 80% (Fig. 2B). Finally, the endotoxin-depleted Stx2B-C-CPE showed a highly purified single band by electrophoresis (Fig. 2C).

We therefore summarize our proposed optimal procedure for preparation of endotoxin-free Stx2B-C-CPE vaccine antigen (Fig. 3) as follows: Step 1, E. coli strain BL21(DE3) cells transformed with pET9a containing Stx2B-C-CPE are cultured in Terrific broth at 25 °C for 24 h; Step 2, the culture supernatant is mixed with saturated ammonium sulfate solution at a ratio of 1:0.6, and the protein is precipitated by centrifugation; Step 3, the protein is purified by using anion exchange chromatography; and Step 4, endotoxin and other contaminants are depleted by using Triton X-114 phase separation.

Fig. 3.

Summary of preparative procedure for endotoxin-free Stx2B-C-CPE antigen.

Although LPS is recognized as one of the bacterial toxins known as endotoxins, it has potential to be applied as an immunological modulator, including vaccine adjuvant [3]. Therefore, we examined the effect of removing endotoxin on the immunogenicity of Stx2B-C-CPE. When mice were subcutaneously immunized with endotoxin-contaminated Stx2B-C-CPE, endotoxin-free Stx2B-C-CPE, or PBS (vehicle control) according to the experimental schedule shown in Fig. 4A, administration of endotoxin-contaminated Stx2B-C-CPE induced high levels of both C-CPE-specific and Stx2B-specific serum IgG antibodies (Fig. 4B). In contrast, the induction levels of these antibodies were dramatically decreased in mice that received endotoxin-free Stx2B-C-CPE (Fig. 4B). These results indicate that the contaminating endotoxin enhanced the production of Stx2B-C-CPE-specific antibodies by acting as an adjuvant. Consequently, the presence of endotoxin in antigen preparations may give rise to the misconception that the protein antigens themselves are highly immunogenic. Therefore, the immunological properties of bacterially expressed recombinant antigens should be assessed very carefully.

Fig. 4.

Endotoxin contamination and immunogenicity of Stx2B-C-CPE. (A) Immunization schedule. (B) Induction of IgG production by immunization with Stx2B-C-CPE. Female Balb/c mice (age, 8 wk) were subcutaneously immunized with 100 $\mu$g of Stx2B-C-CPE, 100 $\mu$g of endotoxin-free Stx2B-C-CPE, or PBS (vehicle control) once weekly for 2 consecutive weeks. At 1 wk after the last immunization, serum was collected, and C-CPE- and Stx2B-specific serum IgG antibodies were detected by ELISA. Data are means $\pm$ SEM (n = 4). The titer was defined as the serum dilution that optimal density is more than 0.1. **p 0.01, *p 0.05 (Two-tailed unpaired t-test). The results shown are representative of two independent experiments.

Endotoxin removal is critical from the viewpoint of safety and quality control, and simultaneously high immunogenicity is required for the induction of appropriate immune responses to achieve protective immunity. Therefore, to apply endotoxin-free Stx2B-C-CPE as a vaccine antigen, it may need to be injected in combination with an adjuvant to enhance its immunogenicity for the induction of antigen-specific antibody production. To address this issue, we used aluminum, an adjuvant in several human vaccines, and lipid A, the active component of LPS. However, both compounds were only weak antigens when used as the sole adjuvant, slightly enhancing Stx2B-specific but not C-CPE-specific IgG production (Fig. 5). In contrast, combined use of aluminum and lipid A adjuvants with endotoxin-free Stx2B-C-CPE strongly promoted both C-CPE- and Stx2B-specific IgG production in immunized mice (Fig. 5). Thus, the immunogenicity of endotoxin-free Stx2B-C-CPE was significantly improved through the use of vaccine adjuvants.

Fig. 5.

Improved IgG production after combining aluminum and lipid A adjuvants with endotoxin-free Stx2B-C-CPE for immunization. Female Balb/c mice (age, 8 wk) were subcutaneously immunized with 10 $\mu$g of endotoxin-free Stx2B-C-CPE without or with 100 $\mu$g of aluminum and/or 1 $\mu$g of lipid A once weekly for 2 consecutive weeks. At 1 wk after the last immunization, serum was collected, and C-CPE- and Stx2B-specific serum IgG antibodies were detected by ELISA. Data are means $\pm$ SEM (n = 4). The titer was defined as the serum dilution that optimal density is more than 0.1. **p 0.01, *p 0.05 (one-way ANOVA). The results shown are representative of two independent experiments.