Expression fusion immunogen by live attenuated Escherichia coli against enterotoxins infection in mice.

Previous epidemiological studies have shown that enterotoxins from enterotoxigenic Escherichia coli (ETEC) appear to be the most important causes of neonatal piglet and porcine post-weaning diarrhoea (PWD). Thus, it is necessary to develop an effective vaccine against ETEC infection. In the present study, the Kil cassette was inserted into the pseudogene yaiT by homologous recombination to create an attenuated E. coli double selection platform O142(yaiT-Kil). After that, PRPL-Kil was replaced with a fusion gene (LTA1-STa13 -STb-LTA2-LTB-STa13 -STb) to establish oral vaccines O142(yaiT::LTA1-STa13 -STb-LTA2-LTB-STa13 -STb) (ER-T). Subsequently, BALB/c mice were orally immunized with ER-T. Results showed that serum IgG and faecal sIgA responded against all ETEC enterotoxins and induced F41 antibody in BALB/c mice by orogastrically inoculation with recombinant E. coli ER-T. Moreover, the determination of cellular immune response demonstrated that the stimulation index (SI) was significantly higher in immunized mice than in control mice, and a clear trend in the helper T-cell (Th) response was Th2-cell (IL-4) exceed Th1-cell (IFN-γ).Our results indicated that recombinant E. coli ER-T provides effective protection against ETEC infection.

Introduction

Enterotoxigenic Escherichia coli (ETEC) is a bacterial pathogen responsible for severe diarrhoea diseases in animals and humans. Enterotoxigenic Escherichia coli leads to high morbidity and mortality both for neonatal and post‐weaning pigs (Harvey et al., 2006). ETEC diarrhoea causes slow growth, weight loss and death, which result in considerable economic losses for hog producers worldwide (Trevisi et al., 2015). The major virulence factors of these bacteria are bacterial fimbriae/non‐fimbrial adhesins and enterotoxins (Nataro and Kaper, 1998). Fimbriae/non‐fimbrial adhesions mediate the attachment of bacteria to host intestinal villus and facilitate bacterial colonization. Then, signal peptide guides enterotoxins through the cell membrane, and causes epithelial cell chloride‐ion secretion and preventing sodium chloride absorption, exacerbated secretory diarrhoea by simultaneous fluid movement into the lumen (Field et al., 1978). In the past decade, ETEC infection has been prevented using antibiotic agents (Smith et al., 2010). Recently, multi‐resistance has been reported with increasing frequency in several countries worldwide, as hogs have frequently been treated as a group with excessive mass medication (Wang et al., 2010). Moreover, antibiotics could select resistant Escherichia coli to transfer their resistance plasmids to other bacteria that may include pathogens in the faecal flora (Nijsten et al., 1996). Thus, the objective of this study is the prevention of ETEC invasion by alternative methods (Bischoff et al., 2002). Immunization remains an effective approach for preventing infectious diseases. However, there is no broadly effective vaccine available for swine ETEC diarrhoea in China. An effective porcine ETEC vaccine should include all the enterotoxins antigens to lead to anti‐heat‐labile enterotoxin (anti‐LT) and anti‐heat‐stable enterotoxins (anti‐STa and anti‐STb) immunity(Liu et al., 2014).

Recent studies showed that most commercial vaccines are administered by injection, stress response in newborn piglets is induced by repeated injection. Nevertheless, activation of secreted intestinal anti‐ETEC responses is impossible to achieve by parenteral administration (Lasaro et al., 2005). Therefore, stimulating a protective immune response by colonization in the intestinal mucosa without causing inflammations is important for an ETEC vaccine. The gastrointestinal tract (GIT) is the animal's largest immunological organ, with a daily production of more than 60% of antibodies (Tang and Li, 2009). Mucosal immune activity plays a major role in neutralizing ETEC upon entry into the body (Kotton and Hohmann, 2004). Oral vaccination avoids the use of syringes, which evokes both local and systemic immune responses, and production of secreted immunoglobulin A (IgA) blocks bacterial attachment to the intestinal epithelial cells (Jertborn et al., 1998). Furthermore, attenuated strains express more heterologous antigens simultaneously, and safely deliver multiple expressed antigens at mucosal sites (Charles and Dougan, 1990). In addition, oral bacterial vaccine vectors are stable in storage, simple to administer and inexpensive to manufacture (Ascon et al., 1998).

In this study, we have established a recombinant E. coli strain in pseudogene positions yaiT expressing LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion immunogen. Using an oral vaccine, we have gained detailed insight into the immune responses in mouse models. The statistics show that oral immunization with ER‐T can elicit more potent systemic and mucosal immune response in mice.

Results

Establish double selection platform and assess its performance

PCR was used to identify O142 (yaiT‐Kil) using primers T1 and T4. The products showed that the E. coli O142: ΔSTa (yaiT gene, 800 bp), O142:△STa/pKil‐donor (yaiT gene, 800 bp; pKil‐donor, 3800 bp) and O142 (yaiT‐Kil) (yaiT with Kil cassette flanked, 3800 bp) were of expected size (Fig. 1, panel B). The double selection platforms were plated on MacConkey agar 18 h later, the platform strains grew normally at 30°C. Platform‐expressed Kil gene caused cells to die at 43°C (Fig. 1, panel C).

Figure 1

The schematic outline of the recombinant strategy for constructing the double selection platform.

Panel A. Construction of the double selection platform O142(yaiT::PRPL‐Kil) by homologous recombination. The pKil‐donor plasmid and the pACBSCE plasmid are co‐transformed into the Attenuated E. coli O142:△STa. L‐arabinose induction promotes expression of the I‐SceI recombinase system and the λ‐Red endonuclease. I‐SceI generates a linear DNA fragment from the pKil‐donor plasmid that is a substrate for recombination with the pseudogene mediated by the λ‐Red system.

Panel B. PCR analysis of chromosomal DNA from double selection platform O142(yaiT::PRPL‐Kil) by using the primers T1 and T4. N: PCR negative control; M: molecular size marker; 1: PCR product of E. coli O142: △STa; 2: PCR product of O142: △STa/pKil‐donor; 3: PCR product of O142(yaiT::PRPL‐Kil).

Panel C. Inversion screen test of the double selection platform. a: O142(yaiT::PRPL‐Kil) incubated at 43°C, cannot grow on MacConkey agar plates; b: O142(yaiT::PRPL‐Kil) incubated at 30°C, grown normally.

Expression of the fusion protein by recombinant E. coli

ER‐T was verified by PCR using the primers T1 with P2. The product of ER‐T was of expected size (3000 bp) for LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion gene, meanwhile, there were no bands in O142:△STa, O142(yaiT‐Kil) (Fig. 2, panel C). After sequencing, we found that fusion genes (LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb cassette) were correct insertions of the E. coli O142:△STa (data not shown). In addition, Western blot analysis of LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion protein, and the expected sizes (17 kDa and 35 kDa) were observed (Fig. 2, panel D).

Figure 2

The schematic outline of the recombinant strategy for constructing the recombinant E. coli O142(yaiT::LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb) for oral vaccine candidate.

Panel A. The full‐length porcine LT192 operon was used conjugate with LT192, STa13, STb for generating LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion antigen and the native LT promoter was retained which expressed without induction. PCR primers P1 and P2 amplified the entire LT cassette including the native LT promoter and terminator. Primers P2 paired with P4 amplified the STa13‐6 × His‐terminator chimeric gene. Primers A1 and A7 mutated the LT gene for LT192.

Panel B. Construction of the recombinant E. coli O142(yaiT::LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb) according to Gene Doctoring method. The pL‐S‐donor plasmid and the recombineering plasmid pACBSCE are co‐transformed into the double selection platform. Arabinose induction promotes expression of the λ‐Red gene products and I‐SceI. I‐SceI cleaves the pL‐S‐donor plasmid resulting in generation of the linear DNA fragment for λ‐Red mediated recombination to generate the recombinant E. coli O142(yaiT:: LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb).

Panel C. PCR reaction for the verifying of the recombinant E. coli strain by using the primers yaiT‐L‐arm and P2. M: molecular size marker; N: PCR negative control; 1: PCR product of E. coli O142: △STa; 2: PCR product of O142(yaiT::PRPL‐Kil); 3: PCR product of .

Panel D. Detection of the LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion protein in the Western blot assay. anti‐His (ZSGB‐BIO Co.; 1:500) were used as primary antibody and secondary antibody by (HRP)‐conjugated goat anti‐mouse IgG (ZSGB‐BIO Co.; 1:2000), development following the manufacturer's instructions of ECL Plus Reagent Kit (7 Sea Biotech, China) M: Protein marker; 1: E. coli O142: △STa as the negative control; 2: E. coli .

Feasibility analysis

Mice were orally inoculated with ER‐T, after 4 h gut/carcass mass ratios (G/C = 0.065 ± 0.006) remained normal, E. coli O142 and E. coli 344‐C induced a significant increase of intestinal fluid accumulation in mice model (G/C = 0.125 ± 0.005, G/C = 0.107 ± 0.003), indicating that the toxicity of ER‐T had been reduced or eliminated (Fig. 3, panel A). The results of in vitro cytotoxicity assay showed that supernatant of ER‐T cannot cause ZYM‐DIEC02 cells to die, but supernatant of E. coli O142 (STa), 274‐A (LT) and 344‐C (STb) induce cell death (Fig. 3, panel B). The tolerance test indicated that ER‐T tolerates well gastric acid pH 2.5 to 4.5, intestinal juice, and bile 0.05–0.3% (Fig. 3, panels C–E). During the first 3 days of the rearing period, the feed intake decrease. Over the whole period, the feed intake and weight were similar in each group (Fig. 3, panel F). The stability curves of ER‐T in the growth phases were similar and no significant differences between them were found, indicating that insertion of the chimeric gene did not impact the growth of the bacterial cell (Fig. 3, panel G). Large doses of ER‐T did not affect the weight gain of mice (Fig. 3, panel H). We checked the stability of ER‐T, and the results indicated that all the colonies analysed presented the expected bands. Sequencing showed that chimeric gene LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb is not lost after 100 generations (data not shown). EM images revealed that insertion of the LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion gene did not affect the E. coli O142: ΔSTa fimbriae expression (Fig. 3, panel I). The colonization assay demonstrated that ER‐T had a higher binding capacity in the intestinal tract of mice (Fig. 3, panel J).

Figure 3

Feasibility of recombinant E. coli for oral vaccine candidate.

Panel A. Suckling mice assay for identifying the toxicity of . The toxicity of was eliminated.

Panel B. ZYM‐DIEC02 cells inoculated with supernatant of E. coli. a: cells inoculated with supernatant of O142 showed significant cell death; b: cells inoculated with supernatant of 274‐A showed significant cell death; c: cells inoculated with supernatant of 344‐C showed significant cell death; d: cells inoculated with supernatant of grew normally; e: non‐treated cells grew normally.

Panel C. The gastric acid tolerance of , from using the plate method to enumerate the amounts of that survived in different pH levels of gastric acid.

Panel D. The intestinal juice tolerance of , from using the plate method to enumerate the amounts of .

Panel E. The bile tolerance of , from using the plate method to enumerate the amounts of that survived in different pH concentrations of bile.

Panel F. The daily appetite of mice orally fed with .

Panel G. The growth curve of . Fed was determined by the number of CFU at the time points indicated.

Panel H. The change of avoirdupois in mice orally fed with .

Panel I. Electron microscope observations of the structure of fimbriae of E. coli O142, O142: △STa, , (magnification 30 000×).

Panel J. Colonization efficacy of in intestinal tracts of mice. Mean values are shown, and error bars represent standard deviations.

Systemic and mucosal immunogenicity elicited by recombinant E. coli

After oral immunization recombinant, bacteria can enhance the serum IgG and faecal IgA. High levels of anti‐F41, anti‐LTA, anti‐LTB, anti‐STa and anti‐STb IgG antibodies were detected after 7, 14, 14, 21 and 21 days in the serum samples respectively (Fig. 4A–E). Forty‐two days after immunization, ER‐T (Fig. 4F–I) specific IgG antibodies were revealed in milk, spleens, mesenteric lymph nodes and intestinal mucus samples. By contrast, in the control groups there was no enhancement of specific IgG antibody responses. Similar findings were revealed on the secreted IgA antibody responses in faeces of mice after being administered orally with ER‐T (Fig. 5A–I). These results show that the recombinant E. coli has the capability to elicit both systemic and mucosal antibody responses.

Figure 4

ELISA analysis of sera IgG from mice inoculated intragastrically with .

A. IgG‐specific of anti‐LTA in the serum.

B. IgG‐specific of anti‐LTB in the serum.

C. IgG‐specific of anti‐STa in the serum.

D. IgG‐specific of anti‐STb in the serum.

E. IgG‐specific of anti‐F41 in the serum.

F. IgG‐specific in the milk samples.

G. IgG‐specific in the spleen samples.

H. IgG‐specific in the mesenteric lymph node samples.

I. IgG‐specific in the intestinal mucus samples. Because LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb insertion causes systemic antibody responses have a slight advantage. Error bars represent standard deviations, and mean values are shown. *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 5

ELISA analysis of faecal IgA from mice inoculated intragastrically with .

A. IgA‐specific of anti‐LTA in the faecal.

B. IgA‐specific of anti‐LTB in the faecal.

C. IgA‐specific of anti‐STa in the faecal.

D. IgA‐specific of anti‐STb in the faecal.

E. IgA‐specific of anti‐F41 in the faecal.

F. IgA‐specific in the spleen samples.

G. IgA‐specific in the milk samples.

H. IgA‐specific in the mesenteric lymph node samples.

I. IgA‐specific in the intestinal mucus samples. mucosal antibody responses are stronger by inserting foreign genes into pseudogenes. Error bars represent standard deviations, and mean values are shown. *P < 0.05; **P < 0.01; ***P < 0.001.

Lymphocyte proliferation responses and cytokines assay

The results indicated that the splenic and mesenteric lymphocytes from mice immunized orally with ER‐T showed significant responses compared with the PBS group by an MTT assay. For splenic lymphocyte proliferation, splenocyte, mesenteric lymphocyte from different groups of immunized mice were cultured in vitro. The higher level of proliferation was demonstrated in ER‐T group for STa, LT192‐STa13, STb, LT192‐STb, LT with the mean SI of 2.38, 2.34, 2.40, 2.11 and 2.31 (Fig. 6, panel A). In addition, the group of ER‐T (Fig. 6, panel B) induced higher mesenteric lymphocyte proliferation responses (SI of STa 2.59, LT192‐STa13 2.85, STb 2.58, LT192‐STb 2.93, LT 2.79) than control and O142:△STa.

Figure 6

Characterization of lymphocyte and cytokines proliferation responses in mice by orally administered .

A. Splenocyte samples stimulated with different purified antigens by MTT method.

B. Mesenteric lymphocyte samples stimulated with different purified antigens by MTT method.

C. The levels of IFN‐γ concentration in the culture supernatant were measured by ELISA.

D. The levels of IL‐4 concentration in the culture supernatant were measured by ELISA. Error bars represent standard deviations, and mean values are shown. *P < 0.05; **P < 0.01; ***P < 0.001.

Mice vaccinated with ER‐T had splenic lymphocytes IFN‐γ (mean 43.55 ± 12.63), and neither were different from those found in splenic lymphocytes from both control groups (Fig. 6, panel C). In contrast, the groups vaccinated with ER‐T (mean 313.89 ± 33.21) had notably higher splenic lymphocytes IL‐4 than the groups control (Fig. 6, panel D). When the IL‐4 to IFN‐γ ratio was analysed, it was found that the oral immunization groups had 7 times more IL‐4 than IFN‐γ, which is a clear indicator of a polarized Th2 immune response.

Toxin‐neutralizing ability in vitro

We used ZYM‐DIEC02 cells to detect the neutralizing efficacy of serum, intestinal mucus, splenocyte lysate and mesenteric lymphocyte lysate from orally inoculated mice. Results showed that the samples (from the immunized mice) had effective neutralization LT toxin, in serum (1:32), splenocyte lysate (1:16), intestinal mucus (1:32) and mesenteric lymphocytes lysate (1:32) (Fig. 7, panel A–D). Likewise, the immunized mice exhibited higher STa toxin‐neutralizing activity in serum (1:16), splenocyte lysate (1:16), intestinal mucus (1:32) and mesenteric lymphocytes lysate (1:32) (Fig. 7, panel E–H). Moreover, samples from the immunized mice showed the STb toxin neutralization potential in serum (1:16), splenocyte lysate (1:16), intestinal mucus (1:32) and mesenteric lymphocytes lysate (1:32) (Fig. 7, panel I–L).

Figure 7

In vitro neutralization assays of samples from group neutralized to LT, STa and STb toxin.

A–D. Serum, splenocyte lysate, intestinal mucus and mesenteric lymphocyte lysate from immunized mice showed neutralization efficiency to LT toxin when compared with that from control mice.

E–H. Serum, splenocyte lysate, intestinal mucus and mesenteric lymphocyte lysate from immunized mice showed neutralization efficiency to STa toxin when compared with that from control mice.

I–L. Serum, splenocyte lysate, intestinal mucus and mesenteric lymphocyte lysate from immunized mice showed neutralization efficiency to STb toxin when compared with that from control mice. The ratios on the figure are the dilution gradient of toxins, the immune mice produced antibodies that protected ZYM‐DIEC02 cells (cell death < 50%), the control group did not protect ZYM‐DIEC02 cells.

Toxin‐neutralizing activity in vivo

Recombinant E. coli ER‐T used as immunoprophylactic had the ability to induce neutralizing antibodies against STa and STb. When comparing the samples of ER‐T for the neutralizing efficiency to STa toxins, results indicated that STa toxin antibodies were in serum (1:5), splenocyte lysate (1:5), intestinal mucus (1:2.5) and mesenteric lymphocyte lysate (1:5) (Fig. 8, panel A–D). Likewise, the samples from the immunized mice group of ER‐T showed the neutralization ability to STb toxin in serum of (1:2.5), splenocyte lysate (1:7.5), intestinal mucus (1:2.5) and mesenteric lymphocyte lysate (1:2.5) (Fig. 8, panel E–H).

Figure 8

Enterotoxins neutralization assay with suckling mice, samples from mix with serially diluted toxins STa and STb respectively.

A. Serum neutralization of STa toxin activity compared with control and E. coli O142: △STa mice.

B. Splenocyte lysate neutralization of STa toxin activity compared with control and E. coli O142: △STa mice.

C. Intestinal mucus neutralization of STa toxin activity compared with control and E. coli O142: △STa mice.

D. Mesenteric lymphocyte lysate neutralization of STa toxin activity compared with control and E. coli O142: △STa mice.

E. Serum neutralization of STb toxin activity compared with control and E. coli O142: △STa mice.

F. Splenocyte lysate neutralization of STb toxin activity compared with control and E. coli O142: △STa mice.

G. Intestinal mucus neutralization of STb toxin activity compared with control and E. coli O142: △STa mice.

H. Mesenteric lymphocyte lysate neutralization of STb toxin activity compared with control and E. coli O142: △STa mice. Samples from control and E. coli O142: △STa suckling mice resisted 1:15 dilution of the toxin, resistant to 1:5 dilution of toxins. Experimental results show that recombinant has certain ability to protect suckling mice. Error bars represent standard deviations, and mean values are shown.

Protection efficacy of maternal antibody

The protective efficacy of suckling mice intaking milk from the immunized pregnant mice challenged with STa or STb toxin was investigated. After being administered orally with STa toxin (1:15 diluted) the ER‐T G/C ratios, outlined below (ER‐T G/C = 0.083 ± 0.005), significantly below that of the mice of the control group (Fig. 9, panel A). Likewise, after challenge with STb toxin (1:15 diluted), ER‐T G/C ratios outlined below (ER‐T G/C = 0.080 ± 0.005), which is lower than control (1:15 diluted G/C = 0.104 ± 0.005) and O142:△STa (1:15 diluted G/C = 0.094 ± 0.006) (Fig. 9, panel B).

Figure 9

In vivo neutralization assays using milk‐immunized suckling mice challenge with STa and STb toxin in serial dilutions respectively.

A. The group of G/C ratios challenged with STa toxin negative value at 1:20 dilution below than that of control.

B. The group of G/C ratios challenged with STb toxin negative value at 1:15 dilution below than that of control. Error bars represent standard deviations, and mean values are shown.

Discussion

Porcine ETEC‐associated diarrhoea, especially PWD and neonatal piglet diarrhoea, remains a major problem for swine producers around the world, but an effective vaccine against neonatal piglet diarrhoea and PWD is lacking (Ruan and Zhang, 2013). Due to the vast majority of piglet diarrhoea being caused by ETEC expressing K88ac or F18 fimbriae in North America (Zhang et al., 2007), abundantly expressed CFA/I, 987P, K99, F18 and/or K88ac fimbriae vaccines were developed. However, in China, pregnant sows are often vaccinated with inactivated vaccines containing fimbriae antigens to protect their offspring from intestinal infection with ETEC. In a recent epidemiological survey, the fimbriae were not frequently associated with ETEC for suckling pigs with diarrhoea in China. Likewise, low proportions of adhesin‐positive E. coli strains have been reported in previous studies in the Netherlands (Guinee and Jansen, 1979), Japan (Nakazawa et al., 1987) and Sweden (Soderlind et al., 1988). Therefore, efficient ETEC vaccines should work against enterotoxins (Ruan and Zhang, 2013).

Oral subunit or killed vaccine antigen delivery through the digestive tract is subject to digestive degradation. Therefore, oral vaccination must maintain the integrity of the antigen during delivery to the intestinal tract. Recently, extensive work on a live vehicle vaccine system presented evidence that recombinant attenuated E. coli is an efficient carrier for antigens by the oral route (Liu et al., 2014). In our study, we build oral vaccine strains obtained from attenuated wild‐type ETEC strain. Moreover, recombinant bacteria achieve ephemeral colonization in the host by F41 fimbriae to overcome host defences, such as the gastrointestinal mucus barrier and intestinal peristalsis cleaning mechanism (Pacheco et al., 2012). Orally administered ER‐T can deliver antigens to the immune system for a prolonged period. In addition, we demonstrated that a single oral dose of recombinant E. coli was without significant reactogenicity or toxicity at dosages up to 109 CFU. Recombinant E. coli containing LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb operon have been stably maintained for over 100 generations in wild‐type E. coli. Another aspect of this research is that it is the first report of oral polyvalent vaccine fusion of all swine ETEC enterotoxins simultaneously. Enterotoxins are an important virulence factor of swine ETEC strains and have been frequently identified in pigs with ETEC diarrhoea disease (Zhang et al., 2006). The most significant contributors to piglet ETEC diarrhoea were LT, STa and STb.

STa is composed of only 19 amino acids (2 KDa), so the immunogenicity is poor, and it can cause diarrhoea unless inactivated or modified STa peptides pSTa (A13Q) reduction toxicity (Zhang et al., 2010). Recent studies indicate that STa (A13Q) fuses to LT became immunogenic (Zhang et al., 2013), resulting in LT‐STb fusion antigen can stimulate body to produce antibodies against LT and STa toxins (Liu et al., 2011).

STb positive isolates from diarrhoeic piglets were more prevalent than STa in Canada (Harel et al., 1991), Poland (Osek and Truszczynski, 1992) and Spain (Blanco et al., 1997) respectively. STb consisting of 40 amino acids (5.2 KDa) was too short for induced anti‐STb immunity in hosts (Sears and Kaper, 1996). Moreover, a study demonstrated that STb coupled to an appropriate carrier molecule had the advantages of reducing toxicity and inducing neutralizing antibodies (Dubreuil et al., 1996).

LT consisted of a toxic A1 subunit (LTA1), non‐toxic A2 subunit (LTA2) and five identical polypeptide chains of B subunit (LTB) (Spangler, 1992) and was capable of enhancing systemic and mucosal antibodies. Insertion of exogenous genes between LTA1 and LTA2 shows that reduce the toxicity of LTA1 and preserve the adjuvant function of LTA (Zhang and Sack, 2015). LTB is a potent immunogen and possess adjuvant properties (Millar et al., 2001; Sanchez and Holmgren, 2005), and binding to specific receptors on the host cell membrane then toxin entry into host cells (Salimian et al., 2010). In addition, LT facilitates the immunogenicity of STa and STb in vaccines immunization against bacterial infections (Norton et al., 2012). Therefore, using STa or STb fusion at the C terminus of LTB subunits would generate quadruple or quintuple anti‐ST fusion antigens (Ruan et al., 2011). Thus, LT plays a key role in the prevention of enteric infection. Moreover, previous vaccines were mainly targeted at LT‐STa or against LT‐STb, so diarrhoea caused by enterotoxins (LTB, STa, STb) could not be completely prevented (Zhang et al., 2013; Ruan et al., 2014), to prevent enterotoxin‐induced diarrhoea in piglets, two vaccines need to be used simultaneously. In this study, increasing immunogenicity of enterotoxins (STa, STb) a mature STb peptide fused to a full‐length of porcine STaA13Q toxoid generated STa13‐STb fusion antigen. However, the preliminary results of this laboratory show that STa13 can poison its own cells after four times tandem repeat (STa13‐STa13‐STa13‐STa13) (data not published). Therefore, to reduce the toxicity a point insertion was used. Connecting STa13‐STb to the 3′ end of the full‐length porcine LTR192G operon, constitutes LTB‐STa13‐STb. Between LTA1 and LTA2 inserted STa13‐STb then form LTA1‐STaA13Q‐STb‐LTA2. Finally, LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb is formed. Since the native promoter of LT is preserved, the expression of the fusion protein does not require an inducer. Fusion antigen LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb was bicistronic like LT, LTA1‐STa13‐STb and LTB‐STa13‐STb were expressed in the cytosol of the bacteria. The A2 region of LT A subunit complexes with LT B subunit forming a structure of A:B5 when secreted to the external of live bacteria under the guidance of signal peptide. When SDS‐PAGE carried out, subunits A and B were separated by heating. Previously, we inserted STa13‐STb‐His after LTA1 and STa13‐STb‐His downstream of LTB, respectively, therefore two bands in WB experiment 17 and 35 kDa.

In addition, previous LT pathogenesis studies have indicated that the LTB amino acids 82‐97 were binding regions of the GM1 receptor; we also retained GM1 binding activity as suggested. Data from this study indicated that ER‐T expression fusion gene LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb elicits specific immune response anti‐LT, anti‐STa and anti‐STb in a mouse model by the oral route. We chose the location of pseudogene yaiT gene, inserted by IS element, in addition far away from the promoter and terminator. Homologous recombination systems were used to insert exogenous genes on pseudogenes. Then, the influence on ER‐T of biological activity and growth was tested (Echols et al., 2002; Balakirev and Ayala, 2003).

IgA production exceeds all the other immunoglobulins at the mucosal surface. Thus, IgA plays a significant role in defending against the invasion of pathogens. Our results indicated faecal ER‐T sIgA levels against LTA (day 21, P < 0.001), LTB (day 14, P < 0.001), STa (day 21, P < 0.05), STb (day 21, P < 0.01) and F41 (day 21, P < 0.01). Likewise, ELISA titres of IgG display statistically significant differences in spleens, milk, mesenteric lymph nodes and intestinal mucus of mice immunized with ER‐T. In addition, serum IgG and faecal sIgA titres from mice after immunization with the vaccine ER‐T were conspicuously higher than those of the control group in both experiments (P < 0.05). This indicates oral administration could trigger immune responses not only mucosa but also systemically. The lymphocyte proliferation assay indicated that SI ratios of splenocytes and mesenteric lymphocytes cell proliferative responses were significantly higher in the ER‐T immunized groups than in controls (P < 0.05; Fig. 6, panels A and B). For this study, within the ER‐T groups of mice immunized by oral administration, results suggested that LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion antigen can induce sufficient cellular immune responses. Immunized mice had higher levels of IL‐4 than IFN‐γ. The IFN‐γ effects are partially blocked by IL‐4 in infectious diseases (Myers et al., 1992), suggesting that recombinant strains induced Th2‐preferred responses. in vitro and in vivo neutralization assays were performed for evaluating the sufficient neutralizing antibody titres by ER‐T. Results showed that after LT, STa and STb enterotoxin incubation with serum, splenocyte lysate, intestinal mucus and mesenteric lymphocyte lysate samples from the immunized mice had the capability of preventing infection of ZYM‐DIEC02 cells in vitro. Furthermore, these samples obviously neutralized the biological activity for STa and STb enterotoxins in suckling mice.

Experimental procedures

Bacterial strains, plasmids and cells

The strains used in this study are listed in Table 1. E. coli O142 was deposited in the Chinese Veterinary Culture Collection Center (Beijing, China) (CVCC accession no. C83920). E. coli C83903 was purchased from the Chinese Institute of Veterinary Drug Control (Beijing, China). Strain LT192‐STa13 fused LT192 (GenBank accession no. CP002732.1) with (GenBank accession no. V00612.1) and Strain LT192‐STb (GenBank accession no. AY028790.1) have been previously described (Liu et al., 2015a,2015b). The plasmids constructed and used are listed in Table 2. Plasmids pDOC‐K, pDOC‐C and pACBSCE were kindly provided by Prof. David J. Lee (School of Biosciences, University of Birmingham, UK).

Table 1

E. coli strains used and constructed in this study

Strain	Relevant properties	Reference or source
O142	Cattle ETEC field isolate, F41/STa	CVCC: C83920
274‐A	Porcine ETEC field isolate, LT/K88, harboring LT as the only toxin	This laboratory
344‐C	Porcine ETEC field isolate, STb/paa, harbor the STb as the only toxin	This laboratory
C83903	Porcine ETEC field isolate, K88/LT/STb/EAST1	China Institute of Veterinary Drug Control
ZYM‐DIEC02	Swine small intestine cell lines	This laboratory
O142: △STa	STa gene deleted in O142	This laboratory
O142 (yaiT: PRPL‐Kil)	PRPL‐Kil cassette inserted into the yaiT gene to construct the double selection platform: grows normally at 30°C but is killed at 43°C by the expression of Kil	This study

Table 2

Plasmids used and constructed in this study

Plasmid	Relevant properties	Reference or source
pDOC‐K	Kanamycin cassette flanked with FRT sites, MCS and I‐SceI sites	Lee et al. (2009)
pDOC‐C	MCS flanked with I‐SceI sites	Lee et al. (2009)
pACBSCE	I‐SceI and λ‐Red protease under control of arabinose promoter, p15A ori	Lee et al. (2009)
pEASY‐Blunt‐Simple	TA cloning vector	Beijing TransGen Biotech, China
pKil	PRPL‐Kil cassette cloned into the MCS1 of pDOC‐K at BamH I and Kpn I sites	This laboratory
pKil‐donor	The left side (yaiT‐L) and right side (yaiT‐R) of the insertion site cloned into pKil	This study
pL‐S	LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb cassette cloned into the MCS of pDOC‐C at Sal I and Xho I	This study
pL‐S‐Donor	yaiT‐L‐arm and yaiT‐R‐arm cloned into pL‐S by BamH I/Sal I, and Xho I/Spe I, respectively	This study

Construction and characterization of the double selection platform

A double selection platform was established using gene doctoring method and λ‐red method (Datsenko and Wanner, 2000; Lee et al., 2009). Restructuring strategy flow chart of E. coli O142: △STa double selection platform is displayed in the online Resource Fig. 1, panel A. Briefly, use primer T1 paired with T2 to clone the left homologous arms from E. coli O142:ΔSTa, use primer T3 paired with T4 to clone the right homologous arms from E. coli O142:ΔSTa (the primers T1, T2, T3 and T4 are listed in Table 3), construct pKil‐donor plasmid, insert the left fragment and right fragment into the pKil plasmid. Transform pACBSCE plasmid and pKil‐donor plasmid into E. coli O142:ΔSTa, subsequently, homologous recombination is induced by L‐arabinose, then generated O142 (yaiT‐Kil) double selection platform. Use PCR primers (T1 and T4) to verify that pKil‐donor fragment was assembled on the platform, the platform screening capability validation by 30 and 43°C, as previously described (Liu et al., 2015a,2015b).

Table 3

PCR primers used for constructing the double selection platform

Primer	Sequence (5′–3′)	Description
yaiT1 (T1)	GGATCCTAACGGAAGCAAGTGGGTTGGTCAG	Anneals to 5′ end of yaiT‐L gene, with BamH I site
yaiT2 (T2)	GTCGACATCAGCCCCCCACCCAGTAGATT	Anneals to 3′ end of yaiT‐L gene, with Sal I site
yaiT3 (T3)	CTCGAGTACAACCTGTACGCCAATACTATCAC	Anneals to 5′ end of yaiT‐R gene, with Xho I site
yaiT4 (T4)	ACTAGTGTTGCGTTGTCGATACGAACTTTG	Anneals to 3′ end of yaiT‐R gene, with Spe I site

Construction and characterization of the recombinant E. coli

Merges chimeric gene produced the LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb by splicing overlap extension (SOE) PCR (Fig. 2, panel A). The first PCR, using LT192‐STa13 (constructed and preserved in our laboratory) as the template and primer P1 paired with P3, generated a fragment including the native LT gene promoter, full‐length LT192 genes and the 3′ end of the STa13 gene (26 bp). The second PCR, using P2 and P4 primer and LT192‐STb (constructed and preserved in our laboratory) as the template, yield the fragment consisted of the 3′ end of the STa13 gene (13 bp), the STb, the 6 × His gene and the native LT gene terminator. In the third step, we used a splicing overlap extension (SOE) PCR, connected the DNA fragment from the first and second PCR step and generated the fusion gene LT192‐STa13‐STb cassette. In the fourth PCR, we used LT192‐STa13‐STb as the template, primer P5 with P6, and generated STaA13Q‐STb‐6 × His gene and A1 gene (1500 bp), named STa‐STb‐6 × His. In the fifth PCR, we use the same template as in the fourth, primer P1 with P7, and generated A1 gene (1500 bp), named A1. In the sixth PCR by primer P1 paired with P8, we used STa‐STb‐6 × His and A1 as the template, and generated the fragment consisting of the LT192 A1 subunit and the STaA13Q‐STb‐6 × His (1700 bp), named LTA1‐STa‐STb‐6 × His. In the seventh PCR, we used LT‐STa13‐STb as the template by primer P9 with P2, and generated a fragment including the LTA2 subunit, the LTB subunit and the STa13‐STb‐6 × His chimeric gene, named A2‐LTB‐STa‐STb‐6 × His. Subsequently, the eighth SOE PCR connected the LTA1‐STa‐STb‐6 × His and A2‐LTB‐STa‐STb‐6 × His fragment and produced fusion genes (LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb cassette). The primers are listed in Table 4.

Table 4

PCR primers used for constructing the recombinant E. coli

Primer	Sequence (5′–3′)	Description
P1	GTCGACGGCGCTGATATCACGATTAGCCT	Anneals to the left side of native LT promoter, with Sal I sites, cloned into pDOC‐C
P2	CTCGAGAAGCTTGCCCCCCAGCCTA	Anneals to the right side of native LT terminator, with Xho I sites, cloned into pDOC‐C
P3	CGGGTACCGAGCTCGATAACATCCAGCACACTGAGGATTAC	3′ end of STa13 (26 bp, no stop codon) + 5′ linker (15 bp)
P4	TGCTGGATGTTATCGAGCTCGGTACCCGGGGAT	3′ end of STa13 (13 bp, no stop codon) + 5′ linker (20 bp)
P5	GGTTGTGGAAATTCATCAGGAACAATCACAAACACATTTTACTGCTGTGAACTTTGTTG	3′ end of LT‐A1 (30 bp) + 5′ STa13(29 bp)
P6	CAGATTCTGGGTCTCCTCATTACAAGTATCACCATGATGATGATGATGGTGGCATCCTT	5′ end of LT‐A2 (33 bp) + 6*his (13 bp) + 3′ end of STb (8 bp)
P7	TGTGATTGTTCCTGATGAATTTCCACAACC	3′ end of LT‐A1 (30 bp)
P8	CAGATTCTGGGTCTCCTCATTACAAGTATCACC	5′ end of LT‐A2 (33 bp)
P9	GGTGATACTTGTAATGAGGAGACCCAGAATCTG	5′ end of LT‐A2 (33 bp)

The LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb cassette was inserted into the pDOC‐C plasmid by restriction endonucleases, producing the pL‐S plasmid, employing left side and right side PCR products ligated into pL‐S plasmid and generating yaiT(pL‐S‐Donor) plasmid respectively. Then, for pACBSCE plasmid, pL‐S‐Donor plasmid co‐transformation selection platforms, an analogous protocol was used for homologous recombination as mentioned above, to yield the recombinant E. coli O142(yaiT::LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb), named as ER‐T (Fig. 2, panel B). Recombinant E. coli strains were identified by PCR, using primers L‐arm paired with P2 respectively. The primers are listed in Table 3. The sequences of the respective fusion genes were confirmed by sequencing. Moreover, recombinant E. coli expressions of LTA1‐STa13‐STb‐LTA2‐LTB‐STa13‐STb fusion protein separated by 12% SDS‐PAGE, examined with anti‐His, were used as primary antibody and secondary antibody by (HRP)‐conjugated goat anti‐mouse IgG, as previously described (Liu et al., 2015a,2015b).

Recombinant E. coli strains were further evaluated for feasibility as an oral vaccine by six different assays. First, the residual enterotoxicity in recombinant E. coli assay by suckling mouse and ZYM‐DIEC02 cell. Briefly, 4‐day‐old suckling mice administered orally with Soybean trypsin inhibitor (2 mg ml−1; Solarbio, Beijing, China), and 100 μl of culture supernatant was inoculated intragastrically. The mice were killed after 4 h, G/C (the weight ratio of gut to the remaining carcass) ratios of ≥ 0.090 were confirmed positive for toxicity of STa or STb as previously described (Frantz and Robertson, 1981). In the ZYM‐DIEC02 cell assay, using 100 μl of culture supernatant incubation with ZYM‐DIEC02 cells in 5% CO2 at 37°C for 24 h, showed the change in the cells as previously described (Feng et al., 2013). Second, we analysed the survival characteristics of recombinant E. coli in the simulation environment of the gastrointestinal tract. Briefly, recombinant E. coli was cultured in lysogeny broth (LB) medium, supplemented with gastric acid (1.5, 2.5, 3.5 or 4.5 of pH), and intestinal juice and bile (0.05%, 0.1%, 0.2%, 0.3% and 0.4%). After shaking at 37°C for 8 h, bacteria survival rate was determined by the plate method of enumerated intermittently, as previously described (Liu et al., 2015a,2015b). Third, we measured the influence of oral administration on feed intake and body weight. Briefly, groups of 10 six‐week‐old female BALB/c mice (Liaoning Changsheng Biotechnology Co., Ltd., China), received doses of 109 CFU, 1010 CFU, 1011 CFU and 1012 CFU for 3 days, and were observed as previously described (You et al., 2011). Fourth, the growth curves of the recombinant E. coli were determined. Briefly, isolate single clone of ER‐T from LB agar medium then inoculate into LB medium, at intervals of 2 h, take sample from LB medium, use the plate method to produce growth curves, as previously described (Yang et al., 2015). Growth curves of E. coli O142, O142: ΔSTa, ER‐T. Fifth, analyse the structural stability of ER‐T, continuous subcultures ER‐T for 100 generations, then the 100 generations ER‐T are plated on LB agar medium, randomly selected several colonies, and identified using PCR primers L‐arm paired with P2, as described above (Nguyen et al., 2005).Sixth, recombinant E. coli was examined via electron microscope (EM). Briefly, recombinant E. coli was in static culture for 3 days, placed onto Formvar carbon‐coated copper grids (200 mesh), then stained with 2% potassium‐phosphotungstic acid (pH 6.8) and observed by EM as previously described (Torres et al., 2004). Seventh, recombinant E. coli intestinal colonization was evaluated. Briefly, six‐week‐old BALB/c mice were orally administered nalidixic acid‐resistant (Nal) recombinant E. coli ER‐T 109 CFU. Subsequently, faeces samples were collected at 7 day intervals and incubated on LBNal agar plates, the assessed by PCR as described above (Huang et al., 2013).

Oral immunization of mice

A total of 80 six‐week‐old BALB/c mice were split into eight groups (A, B, C, D, G1 female BALB/c mice, G2, G3 and G4 male BALB/c mice), with 10 mice per group. ER‐T was grown in improved MINCA medium for 24 h at 37°C and harvested by centrifugation. Resuspension of ER‐T centrifugal sedimentation in milk arrived at concentration of approximately 1 × 1010 CFU ml−1. Mice in group A and B were orally administered with 109 CFUs of ER‐T with a 1.5‐inch, 20 gauge ball‐tip needle. Mice in group C and D received 109 CFUs of O142: △STa. G1, G2, G3 and G4 as control groups received doses of 0.1 ml of milk. All groups were inoculated on a single day (Time 0); then, identical booster doses were administered on 14 and 28 days later. At 21 days, groups B coupled with G2, D coupled with G4, G1 coupled with G3, as previously described (Liu et al., 2015a,2015b).

ELISA analysis for antibody levels

IgG and IgA antibodies were measured by enzyme‐linked immunosorbent assay (ELISA). On days 0, 7, 14, 21, 28, 35 and 42, mice were killed and samples harvested as previously described (Liu et al., 2014). Briefly, the plates were coated with purified lymphotoxin alpha (LTA) recombinant protein, lymphotoxin beta (LTB) recombinant protein, F41 fimbriae, 4 × STa recombinant protein and MBP‐STb recombinant protein at 50, 50, 200, 100 and 200 ng per well, respectively, then treated samples with serial dilution with PBS as the primary antibodies. HRP‐conjugated goat anti‐mouse IgA (Sigma‐Aldrich, St. Louis, MO, USA) or IgG (ZSGB‐BIO Co., Beijing, China) were used for bound antibodies detection as previously described (Jiang et al., 2014).

Lymphocyte proliferation assay and Cytokine assay

Forty‐two days after the first immunization, mice mesenteric lymphocytes and splenic lymphocytes were collected and measured by MTT assay, as previously described (Liu et al., 2015b). Briefly, the lymphocytes were incubated in 96‐well plates at 5 × 105 lymphocytes cells/well for 100 μl then stimulated in vitro with LT (10 μg ml−1), STa (10 μg ml−1), STb (10 μg ml−1), LT192‐STa13 (10 μg ml−1) proteins and LT192‐STb (10 μg ml−1) proteins for 100 μl respectively. Meanwhile, concanavalin A (con A) (10 μg ml−1) was used as a positive control and a black controls were without proteins. The plates were kept at 37°C in 5% CO2 for 72 h and then pulsed with MTT (10 mg ml−1) per well for 4 h. The absorbance was measured with a spectrophotometer at 570 nm. The stimulation indices (SI) were calculated by the following formula: SIproliferation = ODsample/ODnormal. Th1 cells produce IFN‐γ, and Th2 cells produce IL‐4. We used mice cytokine ELISA‐kits (IFN‐γ and IL‐4 kits; JRDUN, Shanghai, China) following the manufacturer's instructions to investigate the role of IFN‐γ (Th1) and IL‐4 (Th2) cytokines in the spleen of mice.

In vitro neutralization assays

We used ZYM‐DIEC02 cell to assay the neutralizing ability of mice inoculated intragastrically with ER‐T. Serum, intestinal mucus, splenocyte lysate and mesenteric lymphocyte lysate samples were collected from mice. Filter‐sterilized supernatant of enterotoxins LT, STa, STb, then serial dilutions respectively. Further, incubated enterotoxins with equal volume of mice samples at 37°C for 2 h, after that toxin‐antibody mixture was added to the ZYM‐DIEC02 cells culture plate 0.1 ml per well. After 24 h, crystal violet staining was carried out and normal cells were counted, as previously described (Guan et al., 2015).

In vivo neutralization test

To assess the neutralization activity of antibodies, we used suckling mice (suckling mice from G1 coupled with G3, aged 4 days). Serial dilutions of enterotoxins STa were mixed with samples (samples from immunization mice), as above. Suckling mice were administered orally with STa toxin‐antibody mixture 0.1 ml per mice. The in vivo neutralization of STb: fist orally soybean trypsin inhibitor, then received equal volume of STb toxin‐antibody mixture. The negative control used the samples from the control group and received equal volume of STa, STb toxin‐antibody mixture respectively. Four hours after oral inoculation, the mice were killed and samples harested, as previously described (You et al., 2011).

Challenge for maternal antibody in suckling mice

The protective efficacy evaluation of maternal antibody uses suckling mice with milk from pregnant mice. Briefly, four 4‐day‐old mice per group received serial dilutions enterotoxins (STa or STb) 100 μl. After 4 h, the animals were killed and the G/C ratio was calculated as previously described (Norton et al., 2012).

Statistical analysis

The aim of this study was to estimate the recombinant E. coli ER‐T, statistical analysis (ANOVA) was performed by SPSS version 19.0 (SPSS Statistics, Chicago, IL, USA) and Graphpad Prism 5.0 (Graphpad software, San Diego, CA, USA), and are represented by asterisks (*P < 0.05; **P < 0.01; ***P < 0.001), P values of < 0.05 were considered to be statistically significant.

Conflict of interest

None declared.

Ethics statement

The present study did not involve endangered or protected species. The animal study complied with the Animal Welfare Act by following the NIH guidelines (NIH Pub. No. 85‐23, revised 1996), and the protocols were approved and supervised by the Animal Care and Use Committee of Yichun University.