Toxicity and Immunogenicity of Enterotoxigenic Escherichia coli Heat-Labile and Heat-Stable Toxoid Fusion 3xSTaA14Q-LTS63K/R192G/L211A in a Murine Model

Diarrhea is the second leading cause of death to young children. Enterotoxigenic Escherichia coli (ETEC) are the most common bacteria causing diarrhea. Adhesins and enterotoxins are the virulence determinants in ETEC diarrhea. Adhesins mediate bacterial attachment and colonization, and enterotoxins including heat-labile (LT) and heat-stable type Ib toxin (STa) disrupt fluid homeostasis in host cells that leads to fluid hyper-secretion and diarrhea. Thus, adhesins and enterotoxins have been primarily targeted in ETEC vaccine development. A recent study reported toxoid fusions with STa toxoid (STaP13F) fused at the N- or C-terminus, or inside the A subunit of LTR192G elicited neutralizing antitoxin antibodies, and suggested application of toxoid fusions in ETEC vaccine development (Liu et al., Infect. Immun. 79:4002-4009, 2011). In this study, we generated a different STa toxoid (STaA14Q) and a triple-mutant LT toxoid (LTS63K/R192G/L211A, tmLT), constructed a toxoid fusion (3xSTaA14Q-tmLT) that carried 3 copies of STaA14Q for further facilitation of anti-STa immunogenicity, and assessed antigen safety and immunogenicity in a murine model to explore its potential for ETEC vaccine development. Mice immunized with this fusion antigen showed no adverse effects, and developed antitoxin antibodies particularly through the IP route. Anti-LT antibodies were detected and were shown neutralizing against CT in vitro. Anti-STa antibodies were also detected in the immunized mice, and serum from the IP immunized mice neutralized STa toxin in vitro. Data from this study indicated that toxoid fusion 3xSTaA14Q-tmLT is safe and can induce neutralizing antitoxin antibodies, and provided helpful information for vaccine development against ETEC diarrhea.


Introduction
Diarrhea is the second leading cause of death to young children who live in developing countries [1], and continues to be a major threat to global health [2]. Enterotoxigenic Escherichia coli (ETEC), E. coli strains producing enterotoxins, are the most common bacteria that cause diarrhea, and are responsible for 300,000 -500,000 deaths of young children annually [2,3]. In addition, ETEC strains are the most common cause of diarrhea to children and adults travelling to ETEC endemic countries or regions, military personnel deployed at these areas, and immunocompromised patients [2,[4][5][6]. These ETEC strains produce various bacterial adhesins and one or more enterotoxins. Bacterial adhesins mediate ETEC initial attachment to host epithelial cells and subsequent colonization at host small intestines, and 23 different adhesins including colonization factor antigens (CFAs) and coli surface antigens (CSs) were characterized among ETEC strains [7]. Enterotoxins including heat-labile toxin (LT) and heat-stable toxin type Ib (STa) disrupt fluid homeostasis in host small intestinal epithelial cells to cause hyper-secretion of electrolyterich fluid through activation of intracellular adenylate cyclase (by LT) or guanylate cyclase (by STa), that leads to diarrhea [8]. Since being identified as virulence determinants in ETECassociated diarrhea, adhesins and toxins have been primarily targeted in anti-adhesin and antitoxin vaccine development. It is believed that anti-adhesin vaccines inducing immunity to block attachment and colonization of ETEC at host small intestines, and antitoxin vaccines inducing antitoxin immunity to neutralize LT and STa enterotoxicity should provide effective protection against ETEC diarrhea [9,10]. Unfortunately, there are no effective vaccines currently available to protect against ETEC diarrhea [10], despite the facts that the association between E. coli and children diarrhea was discovered over 100 years ago [11], that the disease mechanism of ETECassociated diarrhea has been well studied [8,10], and that ETEC strains have been identified the leading bacteria that cause diarrhea [2].
Developing broadly effective ETEC vaccines is hampered by challenges including heterogeneity of ETEC adhesins and potent toxicity of enterotoxins. As different ETEC strains produce immunologically heterogeneous adhesins, experimental vaccines targeting on one adhesin provide protection against only ETEC expressing the same or homologous adhesin, but not strains expressing heterogeneous adhesins. The potent enterotoxicity of LT and STa pre-excludes both toxins from being considered as antigens in developing safe vaccines. Moreover, STa, a 19amino-acid peptide, is poorly immunogenic, thus itself cannot be used as a vaccine component [10,12,13]. In addition, STa shares no genetic or antigenic homology with LT; therefore, anti-LT immunity is not cross protective against STa toxin. Indeed, early experimental vaccines using LT antigens (the nontoxic LT B subunit) were found protective against only ETEC strains expressing LT toxin but not against strains expressing STa toxin [14,15]. Now it becomes acknowledged that an effective antitoxin vaccine should include both LT and STa antigens to induce anti-LT and anti-STa immunity. To be included as safe vaccine components, however, LT and STa would first have their toxicity eliminated or reduced, and only STa and LT derivatives with toxicity reduced or eliminated can be considered safe antigens; second, STa must also have its immunogenicity facilitated to stimulate anti-STa immune responses [16,17]. STa peptides were found to induce anti-STa antibodies when genetically fused or chemically conjugated to strongly immunogenic carrier proteins, such as BSA or detoxified LT peptides [15,[17][18][19][20][21]. Recently, studies demonstrated that some full-length non-toxic STa molecules can be genetically fused to a detoxified LT toxoid (LT R192G ) and resultant LT-STa toxoid fusions were found safe and elicited neutralizing antibodies against both toxins, and suggested that LT-STa toxoid fusions can be potentially used for developing antitoxin vaccines against ETEC diarrhea [17,22,23].
In this study, we generated a different STa molecule, STa A14Q , and a less toxic triple-mutant LT, LT S63K/R192G/L211A (tmLT), to construct a different toxoid fusion antigen. STa A14Q was selected because its analogue, porcine-type pSTa A13Q not only has toxicity more reduced but also maintains an antigenic topology more similar to native STa toxin compared to toxoids pSTa N11K (an analogue of STa N12K ) and pSTa P12F (an analogue of STa P13F ) [17]. Therefore, this STa A14Q toxoid, after being genetically fused to an LT toxoid, is expected to elicit stronger neutralizing antibodies against STa toxin. Attempting to further facilitate anti-STa immunogenicity, we genetically fused three copies of STa A14Q at the N-terminus, the C-terminus, and inside the LT A subunit of tmLT. This constructed toxoid fusion was evaluated in a murine model for safety and immunogenicity, and potential application in ETEC vaccine development.

Materials and Methods
Bacterial strains and plasmids E. coli strains and plasmids used in this study are listed in Table 1. Three previously constructed LT R192G -STa P13F toxoid fusion recombinant strains: 8751, 8752 and 8753 [22], were used as templates to construct the new fusion strain. Since the eltAB genes (coding LT AB toxin) and the estA gene (coding heat-stable toxin 1b; Figure 1A) were of human ETEC prototype strain H10407, the eltAB, LT, estA, and STa in this study are of the human-type. E. coli TOP10 (Invitrogen, Carlsbad, CA) and BL21 (GE Healthcare, Piscataway, NJ) were used as host strains. Vector pBR322 (Promega, Madison, WI) was used to clone the mutated STa and LT genes, and pET28α (Novagen, Madison, WI) was used to clone and express the toxoid fusion gene. Strain 8955, a BL21 strain with vector pET28α [23], was used as the negative control. Recombinant E. coli strains were cultured in Luria Broth (LB) supplemented with ampicillin (100 µg/ml) or kanamycin (30 µg/ ml).
To construct the toxoid fusion gene carrying 3 copies of the STa A14Q gene and 1 copy of the tmLT gene, we genetically fused together 3 pieces of DNA fragments, which were resulted from 3 SOEs, to form a single open reading frame ( Figure 1B).
Briefly, the first piece of fragment consisted of nucleotides coding the first copy of STa A14Q and the first 63 amino acids of the LT A1 peptide. That came from an overlap of two PCR products: one by primer 1 (T7-F, 5'-taatacgactcactataggg-'3) paired with primer 3 (hSTa14Q-R), and the other by primer 2 (STa14Q-F) paired with primer 4 (LT63-R), with plasmid of 8752 (STa 13 -gly-pro-LT 192 /pET28α) as the template. The second piece of fragment contained nucleotides coding the second copy of STa A14Q and the LT A peptide coding the 64 -211 amino acids. This fragment was generated by the overlap of another two PCR products: one by primer 5 (LT63-F) with primer 3 (hSTa14Q-R) and the other by primer 2 (hSTa14Q-F) with primer 6 (LT211-R), with plasmid 8753 (LT 192A1 -gly-pro-STa 13 -LT A2-B /pET28α) as the DNA template. The third fragment included the LT A peptide coding the 212-240 amino acids, the LT B peptide, and a third copy of STa A14Q (with a stop codon). The third fragment was resulted from the overlap of PCR products using primer 7 (LT211-F) and primer 3 (hSTa14Q-R), and primer 2 (hSTa14Q-F) paired with primer 8 (T7-R, 5'tgctagttattggtcaggggt-'3), with plasmid 8751 (LT 192 -L-STa 13 / pET28α) as the template. The second and third fragments were connected first in another SOE PCR, and the resultant fragment was further overlapped to the first fragment in a final SOE PCR to generate the single-open-reading-frame toxoid fusion gene. Since the DNA templates (p8751,8752,p8753) used in PCRs derived from strain 8543 that had nucleotides coding the LT 192 residue mutated, and PCR primers carried mutated nucleotides coding the LT 63 and LT 211 residues, overlapping these 3 pieces of fragments resulted in an entire chimeric toxoid fusion gene designated as 3xSTa A14Q -LT S63K/ R192GL211A ( Figure 1B). All PCRs and SOEs were carried out with pfu DNA polymerase (Strategene/Agilent Technologies, Santa Clara, CA) as described previously [17,22]. The final overlapped fragment was further amplified with PCR primers T7-F and T7-R, digested with NheI and EagI restriction enzymes, cloned into pET28α vector, and expressed as a 6xHis-tagged protein in E. coli TOP10 or E. coli BL21, by following standard protocols [17,22,24].

Expression and detection of the 3xSTa A14Q -tmLT toxoid fusion protein
Expression of the 6xHis-tagged fusion protein by E. coli TOP10 and BL21 recombinant strains was examined in a standard sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The toxoid fusion recombinant strain was grown at 37 °C in 500 ml LB medium supplemented with kanamycin (30 µg/ml), and was induced with isopropyl-1thio-β-D-galactoside (IPTG; 100 µM) for 4 h after culture optical density (OD) reached 0.5. Bacteria were pelleted with centrifugation and were resuspended with 5 ml bacterial protein extraction reagent (B-PER, in phosphate buffer; Pierce, Rockford, IL) for total protein extraction (largely from inclusion body, in denatured buffer), followed by extraction of 6xHistagged proteins to a purity of greater than 90% using nickel affinity chromatography with Ni-nitrilotriacetic acid (Ni-NTA) agarose by following the manufacturer's protocol (QIAGen, Valencia, CA). Extracted 6xHis-tagged proteins were refolded using a Pierce ® Protein Refolding kit with #5 and #9 buffers (Thermo Scientific, Rochester, NY). Refolded proteins were dialyzed 24 h at 4 °C in a series of guanidine-HCl (Sigma, St. Louis, MO) -PBS buffers with guanidine concentrations gradually reduced from 6M to 1M by following manufacture's recommendation (Thermo Scientific), then concentrated using Spectra/Por ® molecularporous membrane tubing (Spectrum Laboratories Inc., Rancho Dominquez, CA) and polyethylene glycol compound (PEG; Sigma) at 4 °C overnight. Ten microliter 6xHis-tagged protein extracts were analyzed in a 12% SDS-PAGE gel and immuno-blot assay. Rabbit anti-CT (1:3300; Sigma) and anti-STa antisera (1:5000; Robertson laboratory) were used to detect the fusion protein. IRDyelabeled goat anti-rabbit IgG (1:5000; LI-COR, Lincoln, NE) were used as the secondary antibody. Bound proteins were detected using a LI-COR Odyssey premium infrared gel imaging system (LI-COR). In addition, ELISAs using fusion proteins as the coating antigen and anti-CT and anti-STa antiserum as antibodies were carried out to assess the toxoid fusion for LT and STa antigenicity, and GM 1 ELISA to examine its binding activity to GM 1 receptors.

Toxicity detection of the STa toxoid, LT toxoids, and the 3xSTa A14Q -tmLT toxoid fusion protein
Toxicity of the expressed STa A14Q , LT S63K , LT S63K/R192G , LT R192G/ L211A and LT S63K/R192G/L211A proteins (in pBR322 vector and E. coli TOP10 cells), with prototype H10407 and STa recombinant strain 8325 as positives, as well as the 3xSTa A14Q -tmLT fusion protein (200 ng or 2 µg) with purified STa and CT toxins as references, was assessed in T84 cells using EIA cGMP and cAMP kits (Assay Designs, Ann Arbor, MI). As described previously [17,22], 150 µl overnight-grown culture supernatants (from an equal amount of cells based on culture optical density values) of each toxoid mutant strain and the positive control strain, the extracted refolded fusion protein (200 ng or 2 µg in 150 µl PBS), 10 ng cholera toxin (CT, Sigma; in 150 µl PBS) which is highly homologous to LT structurally and functionally, or 2 ng purified STa toxin (from Robertson laboratory; in 150 µl PBS), were added to each microplate well (in duplicate) that had 1x10 5 T-84 cells seeded. After 2 h incubation in a CO 2 incubator, wells were washed and T-84 cells were lysed. The lyses were collected and used in ELISAs to measure intracellular cAMP or cGMP levels (pmol/ml) by following the manufacturer's protocol. Mice were held at a horizontal position for 1 min after each drop so that antigen could have a prolong stay at nostril area. The same dose was given at each booster on day 14 and day 28. Mice were anaesthetized with CO 2 and exsanguinated at day 37. Serum and fecal samples (1 g feces resuspended in 5 ml fecal reconstitution buffer; 1:6 dilution) [22,25] collected before and 7 days after each immunization were stored at -80°C until use. After exsanguination mice were necropsied, and intestines were collected and minced with a pair of surgical scissors in PBS (1 g in 2.5 ml; 1:3.5 dilution) supplemented with protease inhibitor phenylmethylsulfonyl fluoride (0.2 mg/ ml). Minced products were vortexed for 1 min and centrifuged 3 min at 13,000 rpm to collect supernatants. Collected supernatants were referred as 'intestinal wash samples' in this study. Animal studies were in compliance with the Animal Welfare Act by following The 1996 National Research Council guidelines [26], and were approved and supervised by a state veterinarian and the South Dakota State University's Institutional Animal Care and Use Committee.

Antitoxin antibody titration
Anti-LT and anti-STa IgG and IgA antibodies in serum, fecal suspension and intestine wash samples of each mouse were examined in ELISAs. ELISAs were conducted similarly as described previously [17,22,23], but with modification. In ELISA to titrate anti-LT antibodies, 25 ng CT (an LT homologue which has been commonly used as coating antigen to titrate anti-LT antibodies) in 100 µl antigen coating buffer (0.015 M Na 2 CO 3 , 0.035 M NaHCO 3 , pH9.6) were coated to each well of an Immulon 2HB plate (Thermo Scientific, Rochester, NY) for 1 h at 37 °C and followed by overnight at 4 °C. Coated wells were washed with PBST (0.05% Tween-20), and uncoated sites were blocked with 10% non-fat milk-PBST (150 µl per well) for 1 h at 37 °C. After washing (5x with PBST), each well was incubated with serum (initially diluted 1:200 in 5% milk-PBST), fecal suspension (1:20 dilution in 5% milk-PBST), or intestine wash sample (1:25 dilution in 5% milk-PBST), in a binary dilution, for 1 h at 37 °C. Wells were washed (5x with PBST) and incubated with HRP-conjugated goat anti-mouse IgG (1:5000 dilution; Sigma) or IgA (1:1000 dilution; Sigma), 100 µl per well, for 1 h at 37 °C; then washed again (3x with PBST and 2x with PBS), and incubated with TMB Microwell Peroxidase Substrate System (2-C) (KPL, Gaithersburg, MD), 100 µl per well, for 20 min at room temperature. Optical density (OD) was measured with a plate reader at 405 nm wavelength. OD readings, greater than 0.4 after subtraction of background readings, were calculated for antibody titers at a scale of log 10 , as described previously [17,22].
Samples in anti-LT antibody titration ELISAs were examined in triplicate, and in anti-STa antibody titration ELISAs were in duplicate due to limited supply of the STa-ovalbumin conjugates. Each antibody titration ELISA was repeated at least once. Weekly collected serum and fecal samples, of individual mouse or pooled from the group, were also used to measure anti-LT and anti-STa antibody titers.

Antitoxin antibody neutralization assays
Mouse serum, fecal suspension and intestine wash samples were examined for antibody neutralization activities against CT and STa toxins using T-84 cells and EIA cAMP and cGMP kits (Assay Design). Since neutralizing antitoxin antibodies prevent CT or STa from stimulating intracellular cyclic GMP or AMP levels in T-84 cells, antibody neutralization activities against CT (or LT) and STa toxins can be assessed with cyclic GMP and cAMP ELISA kits. As described previously [17], serum (30 µl serum in 120 µl DMEM/F12 medium; 1:33.3 dilution in a final volume of 1 ml), fecal (30 µl 1:6 diluted fecal suspension in 120 µl DMEM/F12 medium, a 1:200 final dilution; or 150 µl 1:6 diluted fecal suspension, a 1:40 dilution) and intestine washes (30 µl 1:2.5 diluted washes in 120 µl DMEM/F12 medium; 1:83.3 final dilution), pooled from each group or from individual mouse, were incubated with 2 ng STa toxin or 10 ng CT (diluted in 150-µl DMEM/F12 medium). After 1 h at room temperature, the mixture (300 µl in total) was added to T-84 cells (with 700 µl culture medium), and incubated 2 h at 37 °C in a CO 2 incubator. Cells were washed and lysed, and cell lysates were collected and measured for intracellular cAMP or cGMP levels (pmol/ml) using cAMP and cGMP ELISA kits, respectively.

Statistical analysis
Data were analyzed by using the SAS for windows version 8 (SAS Institute, Cary, N.C.), adjusting for multiple comparison by Bonferroni. Results were expressed as means ± standard errors. Student's t-test was used to compare the different treatment groups. Calculated p values of < 0.05 were regarded as significant when treatments were compared at two-tailed distribution and two-sample equal or unequal variance.

STa A14Q and tmLT (LT S63K/R192G/L211A ) showed significant reduction in toxicity
Cloning of the mutated STa or LT genes in each mutant strain was verified from DNA sequencing, and expression and secretion of the STa and LT proteins were confirmed from STa competitive ELISA and GM1 ELISA, respectively. In vitro toxicity assays indicated that the mutated STa and LT had toxicity eliminated or reduced. Cyclic GMP ELISA showed the cGMP level (to measure STa toxicity) in T-84 cells incubated with culture supernatant of STa A14Q mutant strain 8407 (Table  1)  The 3xSTa A14Q -tmLT toxoid fusion protein was expressed DNA sequencing verified that the toxoid gene consisted of 3 copies of the mutated STa gene (estA) which carried no precursor sequences and only the mutated STa at the 3'end of the fusion gene carried nucleotides coding the stop codon, and one copy of a triple mutated LT S63K/R192G/L211A ( Figure 1B). It was also revealed that the entire toxoid fusion was a single open reading frame. Expression of this 3xSTa A14Q -tmLT fusion protein in strain 9164 was confirmed in Western blot. A protein at the size close to 50 KDa, equivalent to the expected size of the monomeric 3xSTa A14Q -tmLT fusion, was detected with rabbit anti-CT ( Figure 1C) and anti-STa antiserum ( Figure 1D). No proteins of 50 KDa from the E. coli BL21 host strain 8955 were detected. Additionally, ELISAs indicated that this fusion protein reacted to anti-CT antiserum, with an average OD value of 0.58 (1.69 in wells coated with CT; 0.079 in negative control wells). The GM 1 ELISA showed this fusion protein had significant reduction in binding to GM 1 receptor, compared to CT (OD values 0.19 vs. 1.33, p < 0.001).

The 3xSTa A14Q -tmLT toxoid fusion protein was shown to be safe in vitro
This 3xSTa A14Q -tmLT toxoid fusion protein did not stimulate an increase of cAMP or cGMP in T-84 cells (Figure 2). The cAMP levels in T-84 cells incubated with 200 ng and 2 µg toxoid fusion protein were 1.17 and 1.35 pmole/ml, respectively. These levels were similar to the cAMP level in T-84 cells incubated with cell culture medium alone, but significantly lower compared to the cAMP in the T-84 cells incubated with 10 ng CT (15 pmole/ml; p=0.003, 0.003) ( Figure  2A). The cGMP levels in T-84 cells incubated with 200 ng and 2 µg purified toxoid fusion protein were 0.365 and 0.585 pmole/ml, respectively. These levels were not different from the cGMP in T-84 cells treated with cell culture medium (0.38 pmol/ ml), but differed significantly from the cGMP in T-84 cells treated with 2 ng STa toxin (46.5 pmole/ml; p<0.001, <0.001) ( Figure 2B). Moreover, mice tolerated this fusion protein well, as the immunized mice exhibited no apparent adverse effects.

Toxoid fusion protein 3xSTa A14Q -tmLT was immunogenic
Mice IP or IN immunized with the toxoid fusion protein (with adjuvants) had anti-STa antibodies detected. The IP immunized mice had anti-STa IgG antibody detected in the serum samples of all 10 immunized mice and in intestine wash samples of 9 (out 10) immunized mice, and anti-STa IgA antibodies in the intestine wash samples of 5 immunized mice ( Figure 3A). Anti-STa IgA antibodies were also examined from the fecal suspension samples, but the ODs were lower than 0.4, the cutoff point; therefore, ELISA data were not included in final analyses. In the IN groups, anti-STa IgG antibodies were detected in the serum (7 out 10 immunized mice) and washes (7 out 10), and anti-STa IgA antibodies only in the intestinal washes (4 out 10) of the mice immunized with the toxoid fusion antigen with CT adjuvant ( Figure 3B). When weekly-collected serum samples were examined, anti-STa IgG antibodies were detected after the IP primary immunization and after the first IN booster immunization.
Anti-LT antibodies were detected in the immunized (with adjuvants) mice ( Figure 4). In the IP immunization groups, anti-LT IgG antibodies in the serum and intestinal wash samples, and anti-LT IgA antibodies in the intestinal wash samples were detected from the immunized mice, but not the control mice ( Figure 4A). In the IN immunized mice, anti-LT antibodies were detected in mice immunized with either the toxoid fusion antigen (with CT adjuvant) or CT adjuvant alone, but not among mice immunized with the toxoid fusion protein alone or the buffer. Mice IN immunized with the toxoid fusion with CT adjuvant or CT adjuvant alone had anti-LT IgG and IgA antibodies detected in the serum samples ( Figure 4B), anti-LT IgA antibodies in the fecal suspension sample ( Figure 4C), and anti-LT IgG and IgA in the intestinal wash samples ( Figure 4D). Analyses of the weekly collected samples (pooled) showed that IP immunized mice has anti-LT IgG antibodies detected in serum samples after the IP primary immunization and that antibody induction was boosted after each booster immunization. Mice IN immunized with the toxoid fusion (with CT adjuvant) or CT adjuvant had anti-LT antibodies detected in serum samples after the primary injection, and anti-LT IgA detected in the serum samples and fecal suspension after the 1 st booster injection.

Elicited antibodies neutralized CT and STa toxins in vitro
In vitro antibody neutralization assays indicated that elicited antibodies exhibited neutralization activities against CT and STa toxins (Figures 5&6). Cyclic AMP ELISA showed that cAMP levels in the T-84 cells incubated with 10 ng CT and the serum or washes of the IP immunized mice were 1.45 and 2.26 pmole/ml ( Figure 5A). These levels were significantly lower than the cAMP in T-84 cells incubated with the same amount of CT but serum (8.9 pmole/ml, p=0.002) or washes (9.9 pmole/ml, p=0.027) from the IP control mice. When incubated with the fecal samples of the IP immunized and control mice, the T-84 cells had the cAMP detected at levels of 8.9 and 8.3 pmole/ml, with no statistically significant differences (p=0.59). Elicited antibodies in the serum of the IP immunized mice exhibited neutralizing activities against STa toxin ( Figure 6). The cGMP in the T-84 cells incubated with STa toxin and the serum of the IP mice was 2.15 pmole/ml, that was significantly different from the cGMP in cells incubated with STa toxin and the serum sample from the IP control mice (76.5 pmole/ml; p<0.001) ( Figure 6A). The cGMP in T-84 cells incubated with fecal suspension or intestinal wash samples of the IP immunized mice were lower than those in cells incubated with fecal and washes of the IP control mice, but were not significantly different (p=0.31, 0.11). When the intestinal wash sample pooled from the 5 IP immunized mice that had anti-STa IgA detected was used, the cGMP in T-84 cells was significantly lower than that in the cells incubated with the intestinal washes of the IP control mice (p=0.04). T-84 cells incubated with STa toxin and the serum, fecal suspension and intestinal washes of the mice IN immunized with the toxoid fusion antigen (with CT adjuvant) had lower cGMP levels compared to cells incubated with the toxin and the serum, fecal suspension and intestinal wash samples from the mice IN immunized with CT alone ( Figure 6B), but differences were not statistically significant.
Serum samples from individual IP immunized mice were examined in neutralization assays against STa toxin. The serum of mouse 7E, among those had greater anti-STa IgG antibody titers detected, showed strong neutralizing activity against STa toxin. It was found that even when 5 µl (versus 30 µl used in the pooled samples) serum was used, STa toxin showed significant reduction in stimulation of intracellular cGMP in T-84 cells (p=0.03). When 10 µl serum (1:100 in final dilution) was used, it completely prevented STa toxin from stimulating any increase of intracellular cGMP in T-84 cells (1.79 pmol/ml, a level detected in T-84 cells incubated with cell culture medium alone).

Discussion
As E. coli strains producing LT, STa (heat-stable toxin I), or both toxins are the most common bacteria causing diarrhea, effective prevention against ETEC is urgently needed. Vaccines that induce host anti-adhesin immunity to inhibit ETEC attachment and colonization and antitoxin immunity to neutralize enterotoxicity are expected to provide effective protection against ETEC diarrhea. Although significant progress has been made in developing vaccines against colonization from the most prevalent CFA adhesins and against LT enterotoxicity [9,10,12,13], ETEC vaccine candidates currently under development including the promising ones do not carry STa antigens to induce immunity protecting against STa toxin [10]. As ETEC strains expressing STa toxin, alone or together with LT, are associated with two thirds of ETECassociated diarrhea cases [28,29] and the severe cases [30], only vaccines that also induce anti-STa immunity can provide effective protection against ETEC diarrhea. But the potent toxicity and poor immunogenicity of STa have brought major challenges. It was suggested that STa needs to have its toxicity eliminated or reduced and also anti-STa immunogenicity enhanced in order to be used as an ETEC vaccine antigen [16,17]. Recent studies reported that STa toxoids, such as STa P13F , STa E8A , STa T16Q , STa G17S , and porcine-type pSTa N11K and pSTa P12F which are homologues to human-type STa N12K and STa P13F , when genetically fused to a full-length LT toxoid (LT R192G ), induced anti-STa antibodies in mice, rabbits or pigs [17,22,23]. Moreover, elicited antibodies demonstrated neutralizing activity against STa toxin (and CT as well) in vitro [17,22], and protected suckling piglets against infection from a STa-producing ETEC strain [17]. The current study demonstrated that the STa A14Q had toxicity eliminated and induced anti-STa antibodies after being genetically fused to an LT antigen, and suggested potential application of toxoid fusion antigens in ETEC vaccine development.
In this study, we selected STa A14Q to be fused to the LT toxoid and to explore this fusion for eliciting neutralizing anti-STa antibodies. This STa A14Q likely maintains the native STa antigenic epitope topology, indicated by results from a previous study showing its analogue pSTa A13Q (porcine-type STa toxoid) was less toxic but maintained an antigenic structure more similar to native pSTa than toxoid pSTa N11K or pST P12F did [17]. That suggests STa A14Q could be a better candidate for constructing a toxoid fusion antigen to elicit antibodies exhibiting greater neutralizing activity against STa toxin, although its STa antigenic topology could potentially alter after being fused to an LT toxoid. Results from this study showed that antibodies in serum sample of the IP immunized mice neutralized STa toxin, demonstrated by complete prevention of STa toxin in stimulation of any increase of cGMP in T-84 cells ( Figure 6A). Future antibody titration using a serial of diluted samples will determine the anti-STa antibody neutralization titer. Assuming that an antigen dose effect would occur, we included 3 copies of STa A14Q in the fusion in order to further facilitate anti-STa immunogenicity in this study. Results from a comparative study, by re-analyzing serum sample from the previous study with the anti-STa antibody titration method used in this study, indicated that mice IP immunized with 3xSTa A14Q -tmLT developed anti-STa IgG (serum) titers more than twice greater than the mice previously IP immunized with the LT R192G -STa P13F fusion, although a lineal dosage association was not observed (data not shown). That indicated that additional copies of STa toxoid enhanced a toxoid fusion in stimulation of anti-STa immunity and provided helpful information in future ETEC toxoid vaccine development.
It was noted that antibodies induced by the current 3xSTa A14Q -tmLT and the previous LT R192G -STa P13F toxoid fusions showed variation regarding neutralizing activities. Strong anti-STa neutralization activity was observed in serum, but poor in fecal or intestinal wash samples, of the mice IP immunized with the current toxoid fusion 3xSTa A14Q -tmLT. Results from the  previous study suggested that antibodies in fecal samples of mice immunized with fusion LT 192 -STa P13F showed fair to moderate neutralization activity against STa [22]. We also observed that individual mice responded variously to the toxoid fusion antigen. Some IP immunized mice had anti-STa antibody titers detected twice greater than others, and serum sample of these mice, even at a dilution of 1:100 or 1:200, strongly neutralized STa toxin as they prevented STa from stimulation of cGMP in T-84 cells. The structure, or more importantly the antigenic epitope topology, of the STa toxoid presented by toxoid fusions would affect induction of neutralizing anti-STa antibodies. But the cause of variation in immune responses among mice immunized with the same antigen is unclear momently. Immunogenicity of STa toxoids, even enhanced after fused to a strongly immunogenic carrier protein, could be still relatively moderate. That could attribute to inconsistence towards interactions between the STa antigenic domain and host cells, and likely induction of host immune responses in individual mice. Future studies to comparatively examine toxoid fusions, instead of STa toxoids (as STa toxoid topology could be altered after being fused to an LT toxoid), for affinity (or reactivity) to antibodies against native STa can help us to identify fusion antigens with maximum anti-STa immunogenicity to induce antibodies strongly neutralizing against native STa. It should be noted that neutralizing activities against STa (and CT) from antibodies elicited by this toxoid fusion was only assessed in vitro, future in vivo studies including infant suckling mouse assays and piglet challenge studies will be needed to better evaluate elicited anti-STa antibodies.
Data from this study indicated that STa A14Q is a good candidate for LT-STa toxoid fusion antigen to elicit neutralizing anti-STa antibodies. But they may also suggest that other modified STa molecules, such as those having mutation at amino acids other than the 8 th , 12 th , 13 th , 14 th , 16 th and 17 th , or mutations at the same position but with alternative amino acid replacements, may also become STa toxoids and elicit neutralizing anti-STa antibodies when fused to an LT toxoid or other protein carriers. Constructing a STa toxoid or STa-LT toxoid fusion library and thoroughly screening these candidates for not only toxicity reduction but also retention of native STa antigenic structure or topology, a project currently carried out by the STa Toxoid Vaccine Consortium group, may help us to identify optimal STa toxoids for genetic fusions or chemical conjugations to elicit strongly protective anti-STa antibodies.
The tmLT was used to construct the toxoid fusion antigen in this study because it was less toxic based on the results from intracellular cAMP ELISA assays in T-84 cells, although the cAMP level in T-84 cells incubated with tmLT was not significantly different than those incubated with LT S63K (p=0. 24) or LT S63K/R192G (p=0.21), but significantly different compared to dmLT (p=0.046). No studies were carried out to directly compare toxicity of 3xSTa A14Q -tmLT with 3xSTa A14Q -dmLT or 3xSTa A14Q -LT R192G in this study. It is likely that none of these fusions retain LT toxicity, especially considering that only the monomeric LT molecule was presented in these fusions. As toxoid fusion 3xSTa A14Q -tmLT did not stimulate an increase of either cAMP or cGMP level in T-84 cells, and mice showed no adverse effects after immunization, this fusion can be considered a safe antigen. Results from IP immunization studies indicated this tmLT maintained LT immunogenicity as neutralizing anti-LT antibodies were developed in mice immunized with the toxoid fusion (with Freund's adjuvant), thus tmLT is suitable to construct genetic fusions or chemical conjugates for inducing anti-LT immunity and enhancing anti-STa immunity. IN immunization studies, however, showed that mice immunized with this toxoid fusion antigen alone did not stimulate detectable immune responses to LT or STa, unless CT adjuvant was included. That could be caused by the fact that this monomeric toxoid fusion antigen no longer formed a pentamer LT B structure and did not bind effectively to host ganglioside (GM) receptors as LT B pentamer or LT holotoxin did. Binding to GM receptors at host epithelial cells is believed to enhance antigens uptake and stimulation of host mucosal immune responses. On the other hand, binding to gangliosides and then retrograde neuronal transportation of IN administrated LT toxoid LT S63K were believed to cause the undesirable Bell's palsy to human volunteers [31]. Since it does not effectively bind to ganglioside GM 1 , this toxoid fusion antigen may not cause Bell's palsy if taken intranasally by human volunteers, but would induce anti-LT and anti-STa immunity at the presence of a proper adjuvant. That may provide helpful information to solve the dilemma in developing IN immunized ETEC vaccines.
CT (2µg) was used as adjuvant for IN immunization in this study. CT was commonly used as an adjuvant in oral, intragastric and epicutaneous immunization [32][33][34][35][36], as well as in IN immunization [37][38][39]. Data from this study suggested that CT is a very effective adjuvant for IN immunization, as mice IN immunized with the toxoid fusion with CT developed strong immune responses to LT and STa toxins, whereas those immunized with the toxoid fusion alone did not. Unfortunately, due to the high homology between CT and LT, we were unable to differentiate immune responses induced by the LT of the fusion toxoid or the CT adjuvant. Future studies using alternative adjuvants, like monophosphoryl lipid A (MPL), flagellin or others, instead of CT, may help us to better assess induction of anti-LT immunity by this toxoid fusion antigen through the IN route. On the other hand, because of their homology, anti-CT or anti-LT immunity is cross protective; therefore, immunity, whether induced by the toxoid fusion or CT adjuvant, can protect against ETEC LT toxin. CT was also used as the coating antigen to titrate anti-LT antibodies or the toxin to examine anti-LT antibody neutralization activity in this study. The structure homology between CT and LT made the commercially available CT commonly used for anti-LT antibody titration. Although LT is recently also commercially available, CT is still preferred for in vitro anti-LT antibody neutralization assays as CT is more potent in stimulating cAMP levels in T-84 cells.
This toxoid fusion antigen was also sublingually (SL) administrated to a group of 10 mice. However, in contrast to early studies that showed mice developed strong systemic and mucosal immunity after being SL immunized with LT or CT antigens [40][41][42][43][44], mice SL immunized with toxoid fusion 3xSTa A14Q -tmLT (with CT adjuvant) had only anti-STa and anti-LT IgA in the intestine washes, and anti-LT IgG in serum detected. In addition, only antibodies in the serum samples were found to neutralize CT toxin (data not shown). However, in this study, SL administration was carried out without anesthesia, which likely resulted in that the given fusion antigen (40 µg) being swallowed by the mice and thus not effectively delivered to sublingual glands. Under anesthesia, mice would have had antigens effectively delivered and had the given antigens a prolonged stay at sublingual areas to induce robust immune responses. Future studies to SL immunize mice under anesthesia will better assess immunogenicity of this toxoid fusion antigen through the SL route.
We modified antibody titration ELISA assays in this study, with more stringent blocking, additional washes and use of different microtiter plates (for anti-LT), to eliminate background readings. In the previous study [22], antibody titration data from intestinal wash samples had to be excluded due to background readings. Stringent blocking and additional wash steps clearly reduced ELISA background readings. But on the other hand, the stringent blocking and washing procedures likely resulted in low or no detection of anti-STa and anti-LT antibodies from the fecal suspension samples, and also the lack of detection of anti-STa antibodies in some immunized mice. The nature of low antibodies presented in fecal samples may directly attributed to a low detection of anti-STa antibodies in the fecal samples of the immunized mice. Future studies performing kinetic ELISA instead of endpoint measurement may help to detect immune responses from the fecal suspension samples. Unexpectedly, anti-STa antibodies were detected from one mouse in the IP control group (Figure 3). It was unclear to us what caused the outcome, especially since anti-LT antibodies were not detected in this mouse.
Knowing only STa A14Q toxoid was studied in this study, we realized that this 3xSTa A14Q -tmLT may not necessarily be the optimal fusion antigen for ETEC vaccine development. Continuous studies to evaluate STa toxoids and STa-LT toxoid fusions in eliciting neutralizing anti-STa antibodies will help us to identify optimal toxoid fusion antigen(s) for ETEC vaccines development. Nevertheless, results from this study indicated that: 1) LT-STa toxoid fusions are proper antigens to induce protective immunity against ETEC toxins; 2) this STa A14Q toxoid is a better candidate for constructing LT-STa toxoid fusions to elicit neutralizing anti-STa antibodies; and 3) additional copies of STa toxoid enhanced the toxoid fusion for anti-STa immunogenicity. That suggested this 3xSTa A14Q -tmLT fusion antigen can be potentially used in vaccine development against ETEC diarrhea. In addition, the approach of fusing multiple copies of a small and poorly immunogenic antigen to further enhance its immunogenicity may be useful in general vaccine development.