Glycoconjugate Vaccine Containing Escherichia coli O157:H7 O-Antigen Linked with Maltose-Binding Protein Elicits Humoral and Cellular Responses

Glycoconjugate is one of the most efficacious and safest vaccines against bacterial pathogens. Previous studies of glycoconjugates against pathogen E. coli O157:H7 focused more on the humoral responses they elicited. However, little was known about their cellular responses. In this study, we exploited a novel approach based on bacterial protein N-linked glycosylation system to produce glycoconjugate containing Escherichia coli O157:H7 O-antigen linked with maltose-binding protein and examined its humoral and cellular responses in BALB/c mice. The transfer of E. coli O157:H7 O-antigen to MBP was confirmed by western blot and MALDI-TOF MS. Mice injected with glycoconjugate O-Ag-MBP elicited serum bactericidal antibodies including anti-E. coli O157:H7 O-antigen IgG and IgM. Interestingly, O-Ag-MBP also stimulated the secretion of anti-E. coli O157:H7 O-antigen IgA in intestine. In addition, O-Ag-MBP stimulated cellular responses by recruiting Th1-biased CD4+ T cells, CD8+ T cells. Meanwhile, O-Ag-MBP induced the upregulation of Th1-related IFN-γ and downregulation of Th2-related IL-4, and the upregulation of IFN-γ was stimulated by MBP in a dose-dependent manner. MBP showed TLR4 agonist-like properties to activate Th1 cells as carrier protein of O-Ag-MBP. Thus, glycoconjugate vaccine E. coli O157:H7-specific O-Ag-MBP produced by bacterial protein N-linked glycosylation system was able to elicit both humoral and Th1-biased cellular responses.


Introduction
Enterhemorrhagic Escherichia coli (EHEC) O157:H7, as a severe enteric pathogen in human, generally causes bloody diarrhea, hemorrhagic colitis and hemolytic uremic syndrome (HUS) [1,2]. Conventional antibiotic therapy increases the incidence of HUS due to the release of Shiga toxin into intestinal mucosa by antibiotic-mediated bacteriolysis [3,4]. There is an urgent need for vaccines to prevent Escherichia coli (E. coli) O157:H7 infection. Several potential vaccines have been tested in a number of preclinical or clinical studies [5].
Glycoconjugate (O-antigen covalently linked to carrier protein) is considered one of the most efficacious and safest vaccines against bacterial pathogens. O-antigen (O-Ag) plus core oligosaccharides and lipid A form an intact lipopolysaccharide (LPS) of gramnegative bacteria. O-Ag is exposed on the very outer surface of the bacterial cell, and, as a consequence, is a target for recognition by host antibodies. Even though, O-Ag is T-lymphocyte independent antigen like other bacterial polysaccharides. O-Ag alone cannot induce a high serum IgG and sustained T cell memory [6].
Carrier protein contributes a lot to glycoconjugate's immunological efficacy. When O-Ag is conjugated to carrier protein, this glycoconjugate could elicit high serum IgG antibodies to the O-Ag component and induce memory B cells and long-lived T cell memory [7]. The common used carrier proteins in glycoconjugate vaccines include tetanus toxoid, diphtheria toxoid, CRM197, rEPA (Pseudomonas aeruginosa exotoxin A), KLH (keyhole limpet hemocyanin) and flagellin [8]. Currently, exploration of new carrier proteins focuses on T cell epitopes, Toll-like receptors (TLRs), etc. [9]. Recent studies showed that E. coli maltosebinding protein (MBP) had TLR4 agonist-like properties to induce the activation of NF-kB signaling pathway and secretion of proinflammatory cytokines like IL-1b, IL-6, IL-8, TNF-a and IL-12p70 [10]. Some bioactive proteins such as MUC1 and BCG, after fused to MBP, showed better immunity against tumors [11,12], indicating that MBP could be ideal carrier protein in antitumor vaccines. However, little was known about the potential application of MBP against pathogenic bacteria and its immunological enhancement as carrier protein towards glycoconjugates.
Previous studies showed that glycoconjugates including O-Ag-rEPA [13] and O-Ag-Stx [14] against E. coli O157:H7 were able to induce IgG and IgM with serum bactericidal activity. Even though, little was known about the expressions of T-cell differentiation and corresponding secretion of cytokines in response to these glycoconjugates. There is increasing evidence for the role of cellular responses in protection, cellular responses are recognized as more persistent than antibodies [15].
Traditional chemical technologies to produce glycoconjugates are complex and go through multi-step processes. In recent years, via bacterial protein N-linked glycosylation system from Campylobacter jejuni (C. jejuni), next generation of glycoconjugate vaccines is booming rapidly. Oligosaccharyltransferase (OTase) PglB, the key enzyme of C. jejuni, is able to transfer glycan with N-acetyl at the reducing end sugar to asparagine (N) residue of consensus sequence D/E-X-N-Y-S/T (X, Y?proline) [16,17]. Currently, using this convenient in vivo method, O-Ag from a variety of strains including E. coli O121 [18], Yersinia enterocolitica O9 [19] and Francisella tularensis, [20] were coupled to diverse carrier proteins. In this study, we evaluated whether O-Ag from E. coli O157:H7 was a suitable substrate for PglB and explored to use this in vivo protein glycosylation platform to generate a novel glycoconjugate O-Ag-MBP against E. coli O157:H7.
In this study, we provided a novel way to produce anti-E. coli O157:H7 glycoconjugate O-Ag-MBP and showed a comprehensive profile of immune system induced by O-Ag-MBP. Meanwhile, we explored the potential role of MBP as a novel carrier protein of glycoconjugate vaccines against pathogenic bacteria to enhance the immunological efficacy of O-Ag-MBP.

Ethics statement
All animal studies were approved by the Ethics Committee of Shandong University School of Medicine (No. 001 in 2011 for Animal Ethics Approval) and all efforts were made to minimize sufferings.
Bacterial strains and growth conditions E. coli DH5a was used for cloning of plasmids. E. coli CLM24 was used for glycoconjugate production experiments. CLM24 derived from E. coli W3110 by knocking out of waal gene (performed in this study). E. coli strains were grown in Luria-Bertani (LB) broth unless otherwise noted. For glycoconjugate production experiments, E. coli strains were grown in Terrific Broth (TB) broth. Ampicillin at 100 mg/ml, chloramphenicol at 25 mg/ml and spectinomycin at 50 mg/ml were used for plasmids selection as needed.

LPS sliver-stained SDS-PAGE
The LPS profile of the proteinase K-digested whole cells was examined by sliver-stained SDS-PAGE. Briefly, pellets from 1 ml E. coli broth with OD 600 of 1.0 was resuspended in 50 ml 16loading buffer and boiled for 6 min. After cooling to room temperature, 80-100 mg Proteinase K (Thermo Scientific) was added and then incubated at 60uC for 2 hrs. 10 ml sample was load onto 12% SDS-PAGE, and then sliver stained as previously described [21].

Expression and purification of O-Ag-MBP
CLM24 transferred with pBAD24-MBP-GT-66-His, pACT3-PglB and pYES1L-O-Ag was grown in 50 ml LB broth at 37uC for 16 hrs, with shaking. Cultures were then inoculated 1/100 into 1 L TB broth and further grown at 37uC with shaking until OD 600 reached 0.6. Subsequently, 0.1% (w/v) L-arabinose and 50 mM IPTG were added to induce the expression of MBP and PglB, respectively. After another incubation at 28uC for 6 hrs, 0.1% (w/ v) L-arabinose was added again for further induction of MBP.
After a total 20-hr induction at 28uC, cells were harvested by centrifugation at 10,000 g for 10 min and the periplasmic component of the cells were extracted by lysozyme treatment (20 mM Tris-HCl (pH 7.5), 20% (w/v) sucrose, 1 mM EDTA, 1 mg/ml lysozyme, at 4uC for 1 hr). Then, after centrifugation at 10,000 g for 30 min, cell debris was removed and the supernatant was loaded to a Ni-NTA column filled with 3 ml Ni-NTA agarose (GE Healthcare), which was pre-equilibrated with wash buffer (10 mM imidazole, 0.5 M NaCl, 20 mM Na 2 HPO 4 /NaH 2 PO 4 buffer, pH 7.4). Subsequently, the column was washed with wash buffer again and then eluted with elution buffer (250 mM imidazole, 0.5 M NaCl, 20 mM Na 2 HPO 4 /NaH 2 PO 4 buffer, pH 7.4). Fraction containing the purified glycoconjugate was collected and then desalted using centrifugal filter (Amicon Ultra-15, Milipore) against PBS (PH 7.4).

Protein and carbohydrate quantification of O-Ag-MBP
The protein quantity of O-Ag-MBP was measured using Bradford method. The carbohydrate quantity of O-Ag-MBP was measured using phenol-sulfate method. Briefly, 80 ml sample diluent, 40 ml 5% (v/v) aqueous phenol and 200 ml concentrated H 2 SO 4 were consecutively mixed in an Eppendorf tube. After reaction at 30uC for 30 min, read against glucose standards at OD 490 .

Western blot analysis
Samples were separated by 12% SDS-PAGE gels and transferred onto polyvinylidene fluoride (PVDF) membranes. The PVDF membranes were probed with one of the following: anti-MBP antibody (TransGen Biotech, China), anti-66-His antibody (Beyotime, China) and O157:H7 antiserum. In case of anti-MBP antibodies and anti-66-His antibodies, HRP-rabbit anti-mouse IgG (Invitrogen) was used as the secondary antibody. In case of O157:H7 antiserum, HRP-goat anti-rabbit IgG (Invitrogen) was used as the secondary antibody. charge (m/z) ratio of 3,500 to 80,000 using a MALDI-TOF mass spectrometer (AXIMA Confidence, SHIMAZU) in linear mode and power 130.
LPS extraction from E. coli O157:H7 LPS was extracted from E. coli O157:H7 using the hot phenolwater method as previously described [22]. Briefly, 10 g dried E. coli O157:H7 cells were extracted with 100 ml 50% (v/v) aqueous phenol at 65uC for 20 min. After centrifugation at 10,000 g for 30 min, the cell debris was discarded and the top aqueous solution was dialyzed against ddH 2 O to remove phenol. Then the solution was lyophilized and then dissolved in 10 ml ddH 2 O. Subsequently, the solution was consecutively treated with DNase I, RNase A and Proteinase K (Thermo Scientific) according to the manufacturer's instructions. After ultracentrifugation at 110,000 g for 4 hrs, the precipitated gels were dissolved in ddH 2 O and lyophilized to obtain pure LPS.
BALB/c mice immunization 6-week-old female BALB/c mice were divided into three groups, each group contained 8 mice. Mice of each group were injected subcutaneously for once immunization with one of the following: 750 ml of 30 mg (protein quantity) O-Ag-MBP in PBS (pH 7.4) +750 ml adjuvant, 750 ml of 30 mg (protein quantity) MBP in PBS+750 ml adjuvant and 750 ml PBS+750 ml adjuvant, respectively. Mice were injected three times of a two-weekinterval. The first immunization used Freund's Complete Adjuvant (FCA) (Sigma) and the second and third immunizations used Freund's Incomplete Adjuvant (FIA) (Sigma). Seven days after the third immunization, the blood were taken from the tail vein of mice and centrifuged at 3,000 g for 30 min to obtain sera. 100 mg feces were taken from mice and immediately mixed with 1 ml PBS (pH 7.4), the supernatants were collected after centrifugation at 10,000 g for 30 min. The mice were then killed and the splenic mononuclear cells were isolated using mice lymphocyte separation liquid (Dakewe Biotech Co., China) and counting.

Antibody titer assay
Immunoglobulin subclass titer assay of sera and feces were measured by ELISA as described previously [19]. Briefly, the 96well plate (Costar Polystyrene High Binding Plate 3590) was coated with 100 ml of 1 mg/ml extracted E. coli O157:H7 LPS in 0.05 M Na 2 CO 3 (pH 9.8) at 4uC overnight. The coated plate was then washed with 250 ml PBS+ Tween (PBST) (pH 7.4) for three times and then blocked with 2% (w/v) BSA in PBST for 2 hrs at room temperature. After washed with another 250 ml PBST for 3 times, the plate was incubated with 100 ml of sera or feces diluted with PBS at room temperature for 2 hrs. After washed with 250 ml PBST for 5 times, HRP-goat anti-mouse IgG, IgM or IgA (Abcam) was added to the plate at a dilution of 1:20,000 in PBS for 1 hr at room temperature. After washed with another 250 ml PBST for 5 times, 100 ml TMB solution was added to the plate and incubated for 15 min before 100 ml 1 M HCl was added. The plate was then read at OD 450 .
MBP-specific IgG titer assay was measured by ELISA according to immunoglobulin subclass titer assay. The difference is, 5 mg/ml MBP in 0.05 M Na 2 CO 3 (pH 9.8) was used as the coating antigen.
E. coli O157:H7-specific IgG subtypes titer assay was measured by ELISA according to immunoglobulin subclass titer assay. The difference is, HRP-goat anti-mouse IgG1, IgG2a or IgG2b (Abcam) were used as the second antibodies.

Cytokine quantification assay
Sera of mice were analyzed for IFN-c and IL-4 by cytokine ELISA kit (Dakewe Biotech Co., China). ELISA was performed according to the manufacturer's instructions.

ELISPOT assay
IFN-c and IL-4 secretions by the splenic mononuclear cells were analyzed using precoated Enzyme-Linked Immunospot (ELI-SPOT) kit (Dakewe Biotech Co., China). ELISA was performed according to the manufacturer's instructions. Spots were analyzed with a Biosys Bioreader 4000 (Bio-Sys GmbH, Germany).

Bactericidal activity assay
Briefly, 50 ml serum dilutions with PBS (pH 7.4) from 1:20 to 1:1,280 were added to 96-well plate (Costar Polystyrene High Binding plate 3590), then mixed with 50 ml of ,10 4 /ml E. coli O157:H7 bacteria in 1% peptone supplemented with 5% rabbit sera. The plate was incubated at 37uC for 1 hr. After that, 10 ml CCK8 was added and plate was further incubated at 37uC for 4 hrs. The plate was then read at OD 450 .

Statistical analysis
All figures and statistical analyses were generated using the program GraphPad Prism version 5.0. Data were shown as means 6 standard deviation (SD). t test was used for analyses when two groups were compared. One-way ANOVA was used to test for statistical significance of differences between three experimental groups.

Results
Expression of E. coli O157:H7 O-Ag in E. coli W3110 The O-Ag biosynthesis pathway of E. coli O157:H7 is a process involving multiple enzymes. The assembly, transport and polymerization of O-Ag is encoded by rfb gene cluster and the chain length is determined by wzz gene (Figure S1). A recombinant plasmid pYES1L-O-Ag responding for E. coli O157:H7 O-Ag biosynthesis with its native promoter was constructed in this study.
To confirm the expression of E. coli O157:H7 O-Ag in host strain E. coli W3110, the LPS profiles of the strains were compared. The silver-stained LPS profile of proteinase K-digested cell lysates of E. coli W3110 containing pYES1L-O-Ag showed similarity with that of E. coli O157:H7 (Figure 1). Corresponding bands were absent in vector-deficient E. coli W3110 cells (Figure 1). In consideration that O-Ag is dominant in forming heterogeneous LPS profiles, and E. coli W3110 cannot synthesize its own O-Ag due to the inactivation of wbbl gene [23], the similar LPS profiles between E. coli W3110 containing pYES1L-O-Ag and that of E. coli O157:H7 verified the expression of E. coli O157:H7 O-Ag in E. coli W3110 and the chain length was controlled mainly by wzz O157 gene instead of wzz W3110 gene.

Humoral responses in BALB/c mice immunized with O-Ag-MBP
To evaluate the potential use of O-Ag-MBP as glycoconjugate vaccine, O-Ag-MBP was injected subcutaneously into BALB/c mice to measure their immune responses, with MBP and PBS as controls.
Seven days after the third immunization, sera and feces obtained from mice were analyzed for E. coli O157:H7-specific immunoglobulin subclass titers by ELISA, using extracted E. coli O157:H7 LPS as a coating antigen. As expected, mice immunized with O-Ag-MBP elicited higher serum total IgG compared with mice immunized with MBP (P,0.01) and PBS (P,0.001), respectively ( Figure 4A). Similarly, only O-Ag-MBP group elicited a rise in serum IgM compared with mice immunized with MBP (P,0.05) and PBS (P,0.01), respectively ( Figure 4B). Interestingly, IgA secretions in feces were detected higher in O-Ag-MBP group compared with MBP group (P,0.001) and PBS group (P, 0.001) ( Figure 4C), reflecting that O-Ag-MBP stimulated the secretion of IgA in intestine. O-Ag-MBP group had significant differences compared with PBS and MBP group, both of which also had a detectable background responses towards E. coli O157:H7. One reason might be that the mice accessed to E. coli O157:H7 or E. coli O157:H7-like bacteria before or in the experiments.
We also measured MBP-specific IgG titers using purified MBP from strain 1 as the coating antigen. Approximately the same level of MBP-specific IgG was measured in serum of MBP group and O-Ag-MBP group ( Figure 4D), suggesting that MBP-specific IgG was produced after immunizing with either MBP or O-Ag-MBP.

O-Ag-MBP induces killing of E. coli O157:H7
To evaluate serum bactericidal activity of O-Ag-MBP against E. coli O157:H7, diluted serum samples were incubated with E. coli O157:H7 in rabbit sera and then developed by CCK8, which allows sensitive colorimetric assay for the determination of cell viability. The highest serum dilution fold yielding 50% killing of E. coli O157:H7 in O-Ag-MBP group was almost 320, while in MBP group and in PBS group were almost 20 ( Figure 5). When the serum dilution fold of O-Ag-MBP was 20, almost all of bacteria were killed ( Figure 5). These results suggested that the sera in O-Ag-MBP group showed obvious bactericidal activity against E. coli O157:H7.

Cellular responses in BALB/c mice immunized with O-Ag-MBP
To reveal immunization-associated changes in T lymphocyte subsets, Seven days after the third immunization, splenic mononuclear cells were collected and then stained with cell marker antibodies specific for CD3, CD4 or CD8 T cells, and then measured by flow cytometry. The upregulation in T-cell expression of CD3 + CD4 + surface antigens associated with active immunization were more significant with O-Ag-MBP than with PBS (P,0.01) ( Figure 6A). Meanwhile, the upregulation of CD3 + CD8 + antigens were significant in O-Ag-MBP group than in PBS group (P,0.05) ( Figure 6B). These results indicated that O-Ag-MBP stimulated the proliferation of CD4 + and CD8 + T cells.
T cell responses were further assessed by measurement of cytokines in sera. Immunization of mice with O-Ag-MBP induced T cells that secreted higher concentrations of IFN-c than with PBS (P,0.05) ( Figure 7A) and low concentrations of IL-4 than with PBS (P,0.01) ( Figure 7A), indicating that O-Ag-MBP induced Th1-biased response. By IgG subtypes titer assay ( Figure 7B), we found that IgG2a accounts for the main type of IgG, which also indicating that O-Ag-MBP induced mainly Th1-biased response.

Immunological enhancement of MBP as carrier protein towards O-Ag-MBP
MBP upregulated IFN-c and downregulated IL-4 ( Figure 7A), which was consistent with previous demonstrations that MBP induced Th1-biased responses [10]. To further determine whether MBP as carrier protein of O-Ag-MBP induced Th1 activation, seven days after the third immunization, splenic mononuclear cells were collected and incubated with different doses of MBP (0 mg/ ml, 1 mg/ml or 10 mg/ml) in vitro for 24 hrs, and IFN-c and IL-4 secretions were assessed by ELISPOT. When the cells were stimulated with MBP, the number of IFN-c producing cells increased in both MBP (P,0.01) and O-Ag-MBP (P,0.01) groups in a dose-dependent manner ( Figure 8A). These results demonstrated that MBP could activate Th1 cells no matter MBP was immunized alone (MBP group) or linked to O-Ag (O-Ag-MBP group). On the contrary, when the cells were stimulated with MBP, the number of IL-4 producing cells decreased in both MBP and O-Ag-MBP groups when the doses of MBP was increased from 1 mg/ml to 10 mg/ml ( Figure 8B), but the differences elicited by different doses of MBP was not significant.

Discussion
The developments of glycoconjugate vaccines were remarkable over the last 30 years [25]. Whereas, the syntheses via traditional chemical strategies always suffer variable batch-to-batch compo-sition, difficult quality control, inconsistent potency and high production cost [16]. An inexpensive and qualified method is desperately required. In this study, we demonstrated the production of E. coli O157:H7 O-Ag conjugate via a novel bacterial protein N-linked glycosylation system. Our glycoconjugate O-Ag-  MBP was homogenous, and no E. coli O157:H7 for polysaccharide extraction and no toxic chemical reagents for chemical linkage are required during its expression and purification, which would be safe as vaccines towards mice or even humans. In addition, we showed that E. coli O157:H7 O-Ag could serve as substrate for C. jejuni OTase PglB, which provided another evidence for the relaxed substrate specificity of PglB.
MBP could serve as an ideal carrier protein of glycoconjugate vaccines for several advantages. Firstly, MBP does not have to carry an extra signal sequence, because MBP has its own signal peptide (known as MalE) to locate MBP to periplasm [26] (PglB-   protein glycosylation reaction takes place in periplasm of E. coli); Secondly, MBP augments the antigenicity of the target protein by facilitating the interactions between the antigen and antigen-presenting cells (APCs) and/or binding and transporting the antigen to secondary lymphoid organs [27]; Lastly, MBP shows potential as a TLR4 agonist to promote Th1-type responses and associated IgG2 antibody, and elicits IFN-c from spleen mononuclear cells. [10]. In this work, we initiated the exploration of MBP as a carrier protein in glycoconjugate vaccines. Our findings demonstrated that the modified MBP carrying four consecutive N-glycosylation motifs GT at C-terminal could be successfully expressed and conjugated with E. coli O157:H7 O-Ag. In addition, O-Ag-MBP also promoted Th1-type responses and associated IgG2a antibody, and elicited IFN-c from spleen mononuclear cells in response to MBP. Therefore, MBP showed TLR4 agonist-like properties to activate Th1 cells as carrier protein of O-Ag-MBP.
In previous work, E. coli O157:H7 O-Ag conjugated with other proteins like rEPA and Stx elicited statistically significant increases in levels of IgG and IgM towards E. coli O157:H7 LPS [13,14]. Meanwhile, little work reported the secretory of IgA (sIgA) in glycoconjugate vaccine. In our work, O-Ag-MBP elicited high titers of serum bactericidal antibodies including IgG and IgM. Meanwhile, it was inspiring to find the expression of sIgA elicited by O-Ag-MBP, which was suggested to prevent effectively the colonization by enteric pathogens [28].
Several articles argued that humoral responses induced by E. coli O157:H7-specific O-Ag-cholera toxin and O-Ag-horse serum albumin failed to prevent following colonization of E. coli O157:H7 no matter parenteral or oral immunization of mice  [29,30]. The reason might be that cellular response was not elicited by these glycoconjugates, such cellular response was always considered to be more effective to clean pathogens [15]. In our work, the cellular response was elicited by O-Ag-MBP, with the simulation of CD8 + T cells, Th1-biased CD4 + T cells. CD8 + cytotoxic T lymphocytes (CTL) offer protection by complementmediated cell lysis, resulting in the recruitment of macrophages, granulocytes, and eosinophils [31,32]. Th1 CD4 + T cells can also stimulate macrophages by secreting IFN-c.
In summary, E. coli O157:H7 O-Ag was a suitable substrate for PglB, and the glycoconjugate O-Ag-MBP elicited both humoral and Th1-biased cellular responses. Furthermore, MBP in O-Ag-MBP contributed a lot to the Th1-biased cellular responses of O-Ag-MBP as a TLR4 agonist, providing the first evidence of the potential role of MBP as a novel carrier protein in glycoconjugate vaccines against pathogens.