INTRODUCTION
Salmonellosis is a foodborne disease that continues to pose a significant health and socioeconomic threat in developing and developed countries and populations. It is estimated that it affects 90 million people leading to 155 000 deaths every year [Reference Majowicz1]. The clinical symptoms include gastroenteritis, enterocolitis, fever, abdominal pain and diarrhoea caused by excessive inflammation. In some cases the infection spreads from the intestine to other body sites, provoking osteomyelitis, pneumonia and meningitis. In the Unites States, salmonellosis causes estimated annual economic losses from US$ 464 millions to US$ 2·3 billion [Reference Majowicz1]. Many strategies to prevent this illness have been evaluated. For example, live-attenuated vaccines have been shown to limit significantly Salmonella infection in human, stimulating humoral and cellular immunity. From the point of view of the recipients, these vaccines are less safe than inactivated preparations. The latter have shown variable results in their ability to elicit protection against infection [Reference Zhang-Barber2].
Currently, Salmonella enterica serovar Enteritidis (S. Enteritidis) is the main aetiological agent of salmonellosis. Poultry-derived products, particularly meat and chicken eggs, are considered a major source of human infection with this pathogen [Reference Thomas3]. Vaccination, along with other intervention strategies, has been successfully used to reduce the prevalence of S. Enteritidis in poultry flocks [Reference Barrow4], resulting in lower S. Enteritidis human infections [Reference Cogan5, Reference Collard6]. During the last decade several DNA vaccines have been tested against a diverse group of pathogens [Reference Ingolotti7] but this strategy has not yet been proven entirely successful. In consequence, new strategies in vaccine design have to be considered.
S. Enteritidis contains exclusive genomic regions that are thought to confer advantages in the colonization process. One of these is genomic island SPI-19, a genomic island which codes for a Type VI Secretion System (T6SS) that is also present in the genome of the Agona, Dublin, Weltevreden, and Gallinarum serovars of S. enterica. In S. Enteritidis, this island includes some components of the T6SS [Reference Blondel8], one of which corresponds to open reading frame (ORF) SEN1002. This codes for denominated haemolysin co-regulated protein (Hcp), a 28 kDa protein that polymerizes into hexameric rings which forms a tubular structure that is assential for T6SS, permitting the bacteria to secrete effector proteins [Reference Cascales9, Reference Basler10]. In Burkholderia mallei, a Gram-negative bacterium that can cause serious diseases in human and animals, Hcp has been shown to exhibit immunogenic properties in mice, horses and humans. Thus Hcp is an interesting candidate to test for a vaccine design [Reference Schell11]. In addition to SPI-19, S. Enteritidis harbours ΦSE14, a genomic island containing 21 genes, including ORF SEN1395 that codes for a hypothetical protein whose function is still unknown. Bioinformatics analysis has shown that this protein has two conserved domains: a peptidoglycan-binding domain (PG_binding_3) and a second domain with lysosomal activity (DUF847) [Reference Santiviago12, Reference Agron13]. Bacterial lysozyme homologues help maintain cell wall structure during growth and division, and play an important role in many macromolecular transportation systems [Reference Agron13]. Thus, modification of cell-wall structure and the transport mechanisms within it could potentially impede bacterial survival.
Taking into account the aforementioned facts, we constructed DNA vaccines based on the ORFs SEN1002 and SEN1395. Both ORFs were tested for their capacity to generate a protective immune response against S. Enteritidis infection. Taking into consideration the low efficiency of the naked DNA vaccine, both vaccines were administered with monophosphoryl lipid A (MPL) as adjuvant. This agonist of the innate recognition receptor known as Toll-like receptor 4 (TLR4) has previously been used as an adjuvant in vaccine preparations against human papillomavirus and hepatitis B virus [Reference Mbow14].
MATERIALS AND METHODS
Animals
Eight-week-old, female BALB/c mice (obtained from the Instituto de Salud Pública, Santiago, Chile) were habituated for 30 days and divided into five groups. The mice received water and food ad libitum and were handled according to the guidelines of the local Institutional Ethics Committee.
Bacterial strains and growth conditions
Virulent S. Enteritidis PT1 and the non-virulent mutant S. Enteritidis ΔaroA::Kan were used. AroA is implicated in the chorismic acid biosynthesis pathway, a central metabolic node [Reference Arrach15, Reference Hao16]. This mutant was used as a control to DNA-vaccinated mice. Cells were grown under aerobic conditions in Luria–Bertani (LB ) broth for 18 h at 37 °C. E. coli BL21 (DE3) pLys (Novagen, USA) and E. coli strain DH5α (Invitrogen, USA) were grown at 37 °C in TB, supplemented with 50 μg/ml kanamycin as required.
Plasmid construction
The coding regions for the mature SEN1002 and SEN1395 proteins were amplified from the genomic DNA of S. Enteritidis P125109 by polymerase chain reaction (PCR) using primers (Table 1) [Reference Thomson17]. The vector pVAX-3xFlag (pVF) was used for the construction of the recombinant plasmids, pVF1002 and pVF1395. These plasmids were used to transform the E. coli BL21 (DE3) pLys strain to express the recombinant proteins, rF1002 and rF1395. For immunization assays the ORFs SEN1002 and SEN1395 were cloned in the vector pVAX1 (Invitrogen) (pV) using the above-mentioned primers for PCR amplification. The recombinant plasmids pV1002 and pV1395 were used for the transformation of E. coli DH5α cells. Large-scale plasmid DNA isolation was performed using the Endofree Plasmid Giga kit (Qiagen, USA), according to the manufacturer's instructions.
Restriction endonuclease cleavage sites are underlined.
Expression and purification of recombinant proteins
Each recombinant protein, rF1002 and rF1395, was expressed in transformed E. coli BL21 (DE3) pLys by induction with 1 mm isopropyl-d-thiogalactopyranoside (IPTG) and then purified using the ANTI-FLAG® M2 Magnetic Beads kit (Sigma-Aldrich, USA), following the manufacturer's instructions. Evaluation and detection of expression of the recombinant proteins was done using Western blot. Immunodetection of both recombinant proteins (rF1002 and rF1395) was performed by the use of monoclonal anti-Flag antibodies produced in mice (Sigma-Aldrich) as the primary antibody.
Mucosal immunization
Groups of 11 mice were anaesthetized and immunized with pV1002, pV1395 and pV alone as a negative control, at days 1, 16 and 31. The plasmids were co-administered intranasally (50 μg DNA/mouse) with 3·75 μg of the Sigma Adjuvant System (Sigma-Aldrich), which contains 3-deacylated monophosphoryl lipid A (MPL® adjuvant, Corixa, USA), in a total volume of 15 μl for each preparation. Another group of mice was immunized with PBS co-administered with MPL. Mice in the positive control group were vaccinated orally with 1 × 109 colony-forming units (c.f.u.) of the S. Enteritidis ΔaroA::Kan-attenuated mutant strain in 30 μl PBS at time zero.
ELISA detection of antigen-specific antibodies in serum and mucosa
Intestinal (INT), intranasal (IN) and bronchoalveolar (BAL) lavages were performed as described previously [Reference Robinson18–Reference Olive20]. The amount of total specific IgA present in INT, IN, and BAL lavages was determined by ELISA. Antibody titres were estimated as the reciprocals of the last sample dilution giving an absorbance (A 450) value above the cut-off. To compensate for potential variations in the efficiency of recovery of secretory antibodies between animals, the results were normalized according to the total IgA content of the sample. Thus, results were expressed as ELISA units (EU), i.e. the endpoint titre of antigen-specific IgA divided by the total concentration in micrograms of the IgA present in the sample. To establish the IgA standard curve, plates were coated with anti-mouse IgA (Sigma) and further incubated with serial dilution of purified mouse IgA (Sigma). As secondary antibody, HRP-conjugated goat anti-mouse IgA (ICN Biomedicals, USA) was used; plates were developed as described above. For calculation purposes samples negative for SOD-specific IgA were assigned an arbitrary titre of the lowest dilution measured [Reference Retamal-Díaz21]. Moreover, mice were bled and sera obtained (six mice per sample) at days 0, 15, 30 and 45, counting from the first immunization. The presence of IgG, IgG1 and IgG2a isotypes with specificity for SEN1002 or SEN1395 in sera was also determined by ELISA. To this end, 2·5 μg/ml of each recombinant protein, rF1002 and rF1395, diluted in carbonate-bicarbonate buffer (pH 9·6), was used to coat the wells of a polystyrene plate (100 μl/well; Nunc-Immuno plate with MaxiSorp surface). After overnight incubation at 4 °C, the plates were blocked with 0·8% gelatin in TBS for 1 h at 37 °C. The plates were then incubated for 3 h at room temperature with 1:100 dilutions of experimental sera in PBS. Isotype-specific rabbit anti-mouse HRP conjugates (ICN Biomedicals) were added at an appropriate dilution. Finally, after 30 min incubation, 200 μl of substrate solution (Sigma-Aldrich) was added to each well. The cut-off value for the assay was calculated as the mean specific OD450 plus standard deviation (s.d.) for 10 sera from non-immunized mice assayed at a dilution of 1:50. The results of total IgG, IgG1, IgG2a and sIgA are expressed as the mean OD450 at a dilution of 1:100.
Cytokine ELISAs
Spleen cell suspensions from six immunized or control mice were prepared in RPMI 1640 medium and seeded at 4 × 105 cells/well in flat-bottomed 24-well plates (Nunc, Denmark). Cells were stimulated in vitro with rF1002, rF1395 (2 μg/ml for each one), or medium alone, and incubated at 37 °C under 5% CO2. Supernatants were collected after 48 h of culture and stored at −80 °C. Levels of IL-4 and IFN-γ in culture supernatants were measured by using Ready-SET-Go! (eBioscience, USA) mouse IFN-γ and IL-4 according to the manufacturer's instructions. Assays were performed in triplicate and linear regression equations obtained from the absorbance values of the standards was used to obtain the concentration of cytokine in the samples.
Lymphocyte proliferation
Four weeks after the last immunization, six mice from each group were euthanized, and their spleens were removed under aseptic conditions. Single-cell suspensions were prepared from the spleens according to standard procedures [Reference Oñate22]. RPMI 1640 medium (Sigma), supplemented with 10% heat-inactivated fetal calf serum (Gibco BRL, USA) and 1% of antibiotic-antimitotic solution (Invitrogen) was used for culturing the splenocytes. Viable splenocytes at a cell density of 4 × 105 were seeded into flat-bottomed 96-well plates and kept at 37 °C in a 5% CO2 atmosphere in the presence of purified recombinant proteins rF1002 and rF1395 at concentrations of 0·08, 0·4, 2 and 10 μg/ml. Cells were cultured for a period of 72 h. Their proliferative activity was evaluated measuring the level of lactate dehydrogenase released in medium after induced cell lysis. CytoTox One 96 kit (Promega Corporation, USA) was used according to the manufacturer's instructions. Cell proliferation data were expressed as the Stimulation index (SI) of triplicate cultures from a cell pool from each group.
Protection experiments
Six weeks following the last vaccination of the groups immunized with pV1002, pV1395, control and PBS, and 10 weeks for the group immunized with the mutant S. Enteritidis ΔaroA::Kan [Reference Heithoff23], five mice from each group were challenged through oral gavage with 106 c.f.u./mouse of the virulent strain PT1 of S. Enteritidis, a 100% lethal dose by day 9 in this animal model [Reference Araya24]. Seven days later, the infected mice were euthanized and their spleens removed. Dilutions of spleen homogenated were plated out to determine the number of Salmonella c.f.u. per spleen. Units of protection were calculated by subtracting the mean log10 c.f.u. for the experimental group from the mean log10 c.f.u. for the corresponding control group PBS plus MPL.
Statistical analysis
The data were analysed by ANOVA test. A P value of ⩽0·05 was considered statistically significant.
RESULTS
Humoral immune response elicited by pV1002 and pV1395 immunization
To test whether vaccination with recombinant plasmids pV1002 and pV1395 could induce the expression of the corresponding fusion protein in vivo and induce humoral response, groups of BALB/c mice were immunized intranasally with three doses of pV1002 and pV1395 in experimental groups, while pV alone (mock) and PBS were administered in negative control groups. On days 0, 15, 30 and 45, mice were bled to obtain sera. The serum levels of IgG, IgG1 and IgG2a against rF1002 and rF1395 were determined by ELISA. Western blot analysis revealed a single clear band for each recombinat protein purified, with a molecular weight according to expectations: 24 kDa for rF1002 and 19 kDa for rF1395 (Fig. 1). The serum levels of anti-rF1002 IgG and anti-rF1395 IgG increased in a time-dependent fashion, being significantly higher than control groups on days 30 and 45 (Fig. 2a, b ). In addition, an IgG2a/IgG1 ratio sixfold higher than control groups (P < 0·001) was observed in the serum from mice immunized with the DNA vaccine SEN1002 and SEN1395 (Fig. 3). Four weeks after the last immunization the mice were euthanized. INT, IN and BAL lavages were performed to measure the levels of specific IgA. In both experimental groups the levels of anti-rF1002 sIgA and anti-rF1395 sIgA found in INT, IN and BAL lavages were significantly higher than in the control groups (P < 0·001) (Fig. 4a–c ). Taken together, the data suggest that both recombinant proteins were expressed in vivo and retained a high immunogenicity. In addition, IN vaccination induced detectable mucosal and systemic humoral immunity.
The cellular immune response elicited by pV1002 and pV1395 immunization
Next, we evaluated the proliferative response and cytokine profiles following in vitro stimulation of splenic cells with rF1002 or rF1395. Four weeks following the final immunization, the splenic cells of mice that were vaccinated with pV1002 exhibited an antigen-specific response against different concentrations of rF1002 protein (Fig. 5a ). This response was significantly different from the response obtained in the mock and PBS vaccinated groups (P < 0·001). Similarly, splenic cells from mice immunized with pV1395 that were stimulated with different concentrations of rF1395 (P < 0·001) showed a different response (Fig. 5b ). These results suggest that the pV1002 and pV1395 DNA vaccines are able to stimulate antigen-specific cell-mediated immunity. To further characterize the functional phenotype of the antigen-specific T-cell response, we evaluated the levels of IFN-γ and IL-4. Splenic cells, stimulated with rF1002 or rF1395 from mice vaccinated with pV1002 or pV1395, respectively, showed a significantly higher level of IFN-γ secretion (P < 0·001) (Fig. 6a ). There was no significant difference in the levels of IL-4 secretion between the experimental and control groups (Fig. 6b ).
Efficacy of protection against a virulent strain of S. Enteritidis conveyed by pV1002 and pV1395 immunization
Mirroring previous studies [Reference Araya24], 6 weeks following the final immunization of mice immunized with pV1002, pV1395, pV alone and PBS, and 11 weeks in the case of the group immunized with the mutant S. Enteritidis ΔaroA::Kan, all mice were infected orally with 106 c.f.u./mouse of a virulent S. Enteritidis PT1 strain [Reference Araya24]. In this study, as in former studies, no mouse in the positive control group died after challenge. Although clear signs of illness following a lethal infection in the PBS and mock groups were observed, instead of looking for death in mice, we investigated the effectiveness of DNA vaccines in generating a protective immune response against the presence of S. Enteritidis in key organs. Thus, 7 days after challenge, the level of infection in each mouse was evaluated by determining the number of c.f.u. in the spleen. As shown in Table 2, while a considerable number of Salmonella colonies was detected in the spleens of the negative control groups, the number of Salmonella colonies was significantly reduced in the spleen of mice vaccinated with the plasmids pV1002 or pV1395. Therefore both DNA vaccines provided a significant degree of protection compared to the mock and PBS groups. The greatest protection was conferred by immunization with the attenuated S. Enteritidis ΔaroA::Kan mutant (P < 0·05).
* Mice were challenged orally with 106 c.f.u. of strain PT1 7 days prior to euthanization.
† P < 0·05 compared to the control groups.
DISCUSSION
We have designed two DNA vaccines against S. Enteritidis, which is currently the main aetiological agent of salmonellosis worldwide [Reference Thomas3]. The immune response stimulated by the two DNA vaccines was evaluated. Each vaccine included only one ORF of S. Enteritidis. SEN1002 is an ORF located in SPI-19, which codes for Hcp, a protein with immunological properties [Reference Schell11]. SEN1395 codes for a protein which belongs to a new superfamily of lysozymes. It is located in the genomic island ΦSE14, which is exclusive of S. Enteritidis [Reference Santiviago12, Reference Agron13]. These vaccines were designated pV1002 (containing SEN1002) and pV1395 (containing SEN1395).
Protection against pathogens that invade mucosal surfaces is frequently associated with the secretion of local antibodies [Reference Kim25]. In this study, the presence of sIgA specific to the recombinant proteins rF1002 and rF1395 was used as one of the parameters to evaluate mucosal immunity induced by the DNA vaccines. sIgA acts against the passage of pathogens through epithelial and M cell barriers, agglutinating them in the intestinal lumen, inhibiting their motility and evading contact with cell surface receptors used in adhesion processes [Reference Michetti26–Reference Borges28]. Quantification of the sIgA detected in the INT, IN and BAL lavages indicated that both DNA vaccines effectively induce mucosal immunity.
The detectable levels of antibodies specific for both recombinant proteins demonstrated a humoral response for every immunization. Following Salmonella infection a cellular immune response is generally considered sufficient for immunity against this pathogen. However, it has been demonstrated that antibodies against Salmonella are also secreted, and therefore it must be part of the overall immunity elicited [Reference McSorley29–Reference Mastroeni31]. Specifically, antibodies neutralize pathogens in the extracellular medium and stimulate phagocytosis (through opsonization), and antigen presentation to reactive T lymphocytes, which become effector T cells [Reference Igietseme32]. The levels of specific IgG produced between days 0 and 45 showed that a humoral response was activated by immunization with pV1002 or pV1395. For further analysis of the humoral response observed, we quantified the levels of the isotypes IgG1 and IgG2a secreted against the plasmids. In previous studies it has been shown that IgG1 is predominantly released following priming of a Th2-mediated immune response. Likewise, IgG2a secretion is related to a Th1 response [Reference Mosmann33, Reference Abbas34]. The higher binding affinity of IgG2a with FcγRIV (activation receptor) and its reduced affinity with FcγRIIB (inhibitory receptor) ensure that IgG2a functions as a highly effective opsonin that induces the clearance of pathogens through effector cell activation [Reference Nimmerjahn35]. The high levels of IgG2a, sixfold higher than IgG1, indicates that cellular response activity is present, which is adequate enough to protect against S. Enteritidis infection. The high levels of IgG and IgG2a could be related to the presence of the adjuvant MPL, an agonist of TLR4. A study performed using Neisseria meningitis and agonists of TLR3, TLR4, TLR7 and TLR9, showed that their activity is characterized by a predominantly IgG2a response, based on analysis of IgG2a/IgG1 ratios [Reference Fransen36].
During infection with Salmonella, CD4+ and CD8+ T cells undergo clonal expansion and are sensitized and acquire effector functions that are crucial for pathogen clearance [Reference Mittrücker37, Reference Kirby38]. We found a lymphoproliferative response of splenocytes in both experimental groups. Thus, from these antigen-specific T-cellular responses it is established that pV1002 and pV1395 DNA vaccines most likely induce immunological memory. Moreover, T cells from the immunized mice synthesized high levels of IFN-γ in the presence of recombinant proteins rF1002 and rF1395; there is a significant body of evidence highlighting the importance of IFN-γ for the effective clearance of intracellular pathogens [Reference Nimmerjahn35, Reference Akdis39]. It is of interest that the levels of IL-4 were similar in all groups. Altogether, our results indicate that pV1002 and pV1395 DNA vaccines, administered with the MPL adjuvant, are capable of stimulating a Th1 response, which most likely includes the generation of immunological memory. Notably, this is the same qualitative response as that generated by an aroA mutant of S. Typhimurium in BALB/c mice [Reference Harrison40].
Protective immunity, as assessed through protection assays, is the most important parameter for determining vaccine efficacy. Both pV1002 and pV1395 DNA vaccines induced a protective immune response against infection with a virulent strain of S. Enteritidis. There is no information in the literature regarding a DNA vaccine against S. Enteritidis but a DNA vaccine carrying the sopB gene, codifying a protein SPI-1 of S. Typhimurium, which administered with live-attenuated S. Typhimurium shows excellent protection [Reference Nagarajan41]. Our work constitutes a ‘proof of principle’ trial to demonstrate that our vaccine provides protection against an experimental oral infection in mice. This vaccine could prevent infection in other animals, such as poultry and cattle, if orally inoculated, but this possibility needs to be tested in the future. Our results are promising, taking into account that this is the first report involving DNA vaccines using these ORFs. Hereinafter, further studies incorporating more variables could be conducted to establish if the vaccines are effective against higher doses of infection.
ACKNOWLEDGEMENTS
This work was supported by grant ADI-08/2006 from PBCT (CONICYT, Chile) and the World Bank. A.O. was supported by grant 1130093 from FONDECYT; C.A.S. was supported by grant 1110172 from FONDECYT and I.C. was supported by grant 1100092 from FONDECYT.
We are indebted to Susan Bueno for the generous gift of the S. Enteritidis PT1 virulent strain.
DECLARATION OF INTEREST
None.