WT S. enterica serovar Typhimurium 14028 was purchased from American Type Culture Collection (ATCC), Manassas, VA. The ΔlppAB and ΔlppAB ΔmsbB mutants were generated in our laboratory (Sha et al., 2004; Fadl et al., 2005a,b). The Salmonella strains were grown either in Luria-Bertani (LB) or on MSB medium, the latter consisted of LB medium with no NaCl but supplemented with 2 mM MgSO₄ and 2 mM CaCl₂ (Murray et al., 2001). Bacterial cells from the exponential growth phase at 37°C were harvested and used for both animal studies (immunization and challenge) and 2D gel electrophoresis analysis. All of the Salmonella strains were periodically examined on Salmonella-Shigella (SS) agar plates (Difco, Detroit, MI) for purity.

Six-to-eight week old C57BL/6J female mice were obtained from the Jackson Laboratory (Bar Harbor, ME). Animals were first fasted for at least 4 h and then orally dosed with 100 μl of bacterial suspension with a mouse stomach tube (Sigma, St. Lois, MO). The bacterial doses were prepared in 5% sodium bicarbonate solution to overcome stomach acidity. Briefly, mice (n = 10 per group) were orally infected with 2.0 × 10⁶ colony forming units (CFU)/100 μl of WT S. typhimurium (our calculated LD₅₀ was 2.0 × 10⁵ CFU) or its ΔlppAB or ΔlppAB ΔmsbB mutants. Blood from the surviving animals from four independent experiments was collected by retro-orbital bleeding on days 0, 1, 3, 7, 14, and 21 post infection (p.i.). Total IgG and its isotype antibody titers to the whole bacterial cells on days 14 and 21 p.i. were evaluated by ELISA in serially diluted serum as we previously described (van Lier et al., 2014). Briefly, ELISA plates were first coated with the whole bacterial cells overnight at 4°C. For preparing the bacterial cells, the WT S. typhimurium was grown at 37°C until an OD_600nm of 0.8 was reached. The culture was then resuspended to a concentration of 5 × 10⁹ CFU/ml and used to coat the plates treated with poly-L-lysine (10 μg/ml). A serial dilution (1:5) of serum was made to evaluate antibody titers, and a positive antibody titer was defined as the inverse of the highest serum dilution giving an absorbance reading of ≥0.2. Antibody classes and IgG isotypes were also examined by using specific isotype secondary antibodies as we previously described (van Lier et al., 2014).

Likewise, the serum cytokine levels were examined on days 1, 3, 7, 14, and 21 p.i. by using a mouse 6-plex assay kit (Bio-Rad Laboratories Inc., Hercules, CA; Tiner et al., 2015). The fecal matter was collected on day 21 p.i. and suspended in phosphate-buffered saline (PBS) containing 0.15 mg soybean trypsin inhibitor and 25 mM EDTA (ethylenedinitrilo tetraacetic acid) (Sigma-Aldrich, St. Louis, MO). To measure the level of IgA, the ELISA plates were first coated with anti-IgA antibody (without biotinylation), and then the serial dilutions of fecal suspension were added to the plates. After 2 h incubation at room temperature followed by serial washes, the biotinylated anti-IgA antibody with the peroxidase-labeled streptavidin were applied. After further incubation, washes, and addition of the substrate, the plates were read under OD_450mm and the values were normalized by their corresponding total protein concentrations (μg/ml) in the fecal suspensions (Cong et al., 1998; Cao et al., 2012). Four independent experiments were performed and the pooled data (each sample was run in duplicate) were presented with statistical analysis. All of the animal studies were performed under an approved Institutional Animal Care and Use Committee protocol.

C57BL/6J mice (n = 10 per group) were orally immunized/infected with 3 × 10³ CFU/100 μl of WT S. typhimurium and its ΔlppAB or ΔlppAB ΔmsbB mutants. After 36 days, mice were challenged via the oral route with 2 × 10⁶ CFU/100 μl of WT S. typhimurium. Our goal was to discern efficacy of the vaccine strains at a low dose of 3 × 10³ CFU.

In a parallel experiment, mice (n = 5 per group) were orally immunized with a high dose of 1 × 10⁸ CFU (~1000 LD₅₀ of WT S. typhimurium) of ΔlppAB or the ΔlppAB ΔmsbB mutant and all survived. After 21 days, the immunized animals along with naïve control mice (n = 5) were challenged with 1 × 10⁸ CFU of WT S. typhimurium via the oral route. Organs from these challenged animals were excised on either day 7 (for unimmunized mice that were infected with WT S. typhimurium before succumbing to infection) or day 21 (for immunized and challenged animals, all survived). The organs were fixed in 10% neutral buffered formalin (Sha et al., 2008; Agar et al., 2009) and tissues processed and sectioned at 5 μm. The samples were mounted on slides and stained with hematoxylin and eosin (H&E). Tissue lesions were scored on the basis of a severity scale, which correlated with estimates of lesion distribution and the extent of tissue involvement, as we previously described (Sha et al., 2008; Agar et al., 2009). The histopathological evaluation of the tissue sections was performed in a blinded fashion.

Mice (n = 5) were orally immunized with 1 × 10⁸ CFU of ΔlppAB or the ΔlppAB ΔmsbB mutant, and spleens were harvested on day 21 post immunization. We chose a higher dosage of the mutants for vaccination as Salmonella vaccines are administered at a dosage range of 10⁷-10¹⁰ CFU in humans, e.g., S. typhi Ty21a (WHO, 2014) or other typhoid vaccine candidates that have undergone clinical trials (Tacket et al., 2000; Tacket and Levine, 2007). T-cells from the mutant-immunized mice were isolated, and their ability to proliferate as well as to produce cytokines was evaluated upon co-culture with the γ-irradiated and heat-killed WT S. typhimurium-stimulated antigen-presenting cells (APCs; pulsed). T-cells incubated with γ-irradiated naïve APCs (without the bacterial stimulation; un-pulsed) served as controls (Sha et al., 2013; van Lier et al., 2014). After 72 h of incubation, 1μCi of [³H] thymidine was added to a set of each co-culture well, and the cells harvested 16 h later using a semi-automated sample harvester, FilterMate Harvester (PerkinElmer, Waltham, MA), followed by the measurement of radioactive counts (TopCount NXT, PerkinElmer). Likewise, from another set of the T-cell cultures, a portion of the supernatant was collected at 48 and 72 h to measure cytokine/chemokine production by using a mouse 6-plex assay kit (Bio-Rad Laboratories Inc.).

Bacterial cells (mutants and WT S. typhimurium) were lysed in 8 M urea, 4% CHAPS [3-(3-cholamidopropyl)-dimethylammonio-1-propanesulfonate], and 40 mM Tris-HCl (pH 8.0), followed by 10% (v/v) trichloroacetic acid (TCA) precipitation to remove salts and to concentrate proteins. The TCA precipitated proteins were re-dissolved in the lysis solution and subjected to 2D gel electrophoresis as we previously described (Chopra et al., 2006).

For each vaccine strain, the above prepared samples were run in triplicate gels, and the WT S. typhimurium sample-containing gels were set as a reference during analysis by using Progenesis Workstation (Nonlinear_Dynamics, Durham, NC) at the Protein Chemistry Laboratory, UTMB. The normalized volume (NV) of each spot was calculated by using the total volume normalization method in which each spot on a gel image was expressed relative to the total volume of all spots on that image, and then normalized to the total volume of all spots on the reference gel image (Berth et al., 2007). The detected spots with NVs of ≤60 were filtered out. The NVs of the corresponding spots from gels containing samples of WT S. typhimurium and its two mutants were compared, and a fold-change of ≥2 was considered as differentially expressed/produced proteins. Some of these well separated protein spots (a total of 61) were picked robotically, trypsin-digested, and peptides identified by Matrix-Assisted Laser Desorption Ionization Time-of-Flight Mass Spectrometry (MALDI TOF-MS) at the Protein Chemistry Laboratory, UTMB.

Two-way analysis of variance (ANOVA) with the Tukey's post-hoc test or the multiple Student's t-test with the Holm-sidak post-hoc test correction was used for data analysis. We used Kaplan–Meier survival estimates or Fisher exact test for animal studies, and p ≤ 0.05 were considered significant for all of the statistical tests used.

Our previous studies with both ΔlppAB and ΔlppAB ΔmsbB mutants in mice were focused on the i.p. model of infection (Sha et al., 2004; Fadl et al., 2005a,b; Liu et al., 2008). Both the mutant strains displayed significantly decreased virulence phenotype and provided protection to mice upon subsequent WT S. typhimurium lethal challenge. However, their potential as vaccine strains (attenuation and protection) in the oral infection mouse model was not characterized and formed the basis of this study. As shown in Figure 1A, when mice were orally challenged with WT S. typhimurium at the dose of 2.0 × 10⁶ CFU (~10LD₅₀), animals started to die at day 8 p.i., and 90% of them eventually succumbed to infection by day 16 p.i. In sharp contrast, 100% of the mice infected with either the ΔlppAB and or the ΔlppAB ΔmsbB mutant at the same dose of 2.0 × 10⁶ CFU survived up to tested 30 days, and no signs of the disease were observed during the course of infection (Figure 1A).

Specific anti-S. typhimurium antibodies were evaluated on days 14 and 21 p.i. from the above infected groups of mice. Compared to the pre-immune serum, animals in all of the infected groups (WT S. typhimurium and the ΔlppAB and ΔlppAB ΔmsbB mutants) developed high levels (~1:10,000) of Salmonella specific IgG antibodies on day 14 p.i. (Figure 1B). More specifically, all infected mice had similar levels of total IgG and IgG2b isotypes; however, slightly higher levels of IgG1 and IgG2a were observed in the mutant-infected groups of animals when compared to mice in the WT S. typhimurium-infected group (Figure 1B).

On day 21 p.i., the total IgG titers remained unchanged in the ΔlppAB mutant-infected mice, while it slightly dropped in animals infected with WT S. typhimurium when compared to that on day 14 p.i. (Figures 1B,C). In contrast, the total IgG titers in the ΔlppAB ΔmsbB mutant-infected mice continued to mount (>1:100,000) on day 21 p.i. (Figure 1C). Likewise, the IgG isotype titers in all of the infected mice were maintained essentially at the similar levels when comparisons were made for days 14 and 21 p.i., with the exception of the level of IgG2a in WT S. typhimurium-infected animals which dropped back to the level detected in the pre-immune serum (Figure 1C).

The intestinal secretory IgA was also monitored during the course of infection, and a significant level of fecal IgA was detected on day 21 p.i. from all of the infected mice, whether with WT S. typhimurium or its ΔlppAB or ΔlppAB ΔmsbB mutants. However, the highest level of IgA was noted in mice infected with the ΔlppAB ΔmsbB mutant followed by the ΔlppAB mutant, and then the WT S. typhimurium-infected group of animals (Figure 1D).

Serum cytokines were monitored periodically by Bioplex during the course of infection. In general, irrespective of whether the mice were infected with WT S. typhimurium or its designated mutants, the levels of examined cytokines, except for IL-6 and interferon (IFN)-γ, were higher in the serum on day 1 p.i. when compared to the pre-immune serum (Figure 2). The levels of all examined cytokines were also maintained at significantly higher levels during the later stages of infection (i.e., days 14 or 21 p.i.), indicating the mounting of both early (innate) and late (adaptive) immune responses in the infected mice (Figure 2).

Interestingly, a similar trend in IL-1β, IL-10, IL-17A, and TNF-α production was observed in all of the infected groups of mice (i.e., WT S. typhimurium or its ΔlppAB or ΔlppAB ΔmsbB mutants), in which their levels were relatively high on day 1 p.i., but subsequently declined by the end of the first week. However, a resurgence in these cytokine production reaching to their highest levels was noted on either day 14 or 21 p.i. (Figures 2A,C,D,F). On the other hand, the trends for IL-6 and IFN-γ production varied among different groups of mice (Figures 2B,E).

IFN-γ levels gradually mounted during the course of infection in the WT S. typhimurium-infected group of mice and reached the peak level on day 14 p.i. (Figure 2E). In contrast, a substantial level of IFN-γ was only detected on days 14 and 21 p.i. in the ΔlppAB ΔmsbB mutant-infected group of mice. Strikingly, minimal IFN-γ was produced in the group of animals infected with the ΔlppAB mutant throughout the infection course (Figure 2E).

Likewise, the level of IL-6 gradually increased during the course of infection in the WT S. typhimurium-infected group of mice and reached the highest level on day 14 p.i. However, the levels of IL-6 in either of the mutant-infected mice increased somewhat on day 3 p.i. and declined on day 7 p.i., followed by its increase on days 14 or 21 p.i. (Figure 2B).

When comparing WT S. typhimurium vs. the mutant-infected mice, the former generally had higher levels of IL-1β, IL-6, IL-10, and IFN-γ on day 14 p.i. (Figures 2A–C,E). These cytokine levels significantly dropped to lower than that of both the mutant-infected groups of mice on day 21 p.i. On the other hand, TNF-α was maintained at generally a similar level across all of the infected mice (with WT S. typhimurium or its ΔlppAB or ΔlppAB ΔmsbB mutants), but were elevated to much higher levels in both of the mutant-infected groups of animals at the later stages of infection (days 14 and 21, Figure 2F).

Interestingly, the production of IL-17A was generally higher in the mutant-infected groups of mice; however, the trend was reversed on day 14 p.i., with significantly higher levels of IL-17A in the WT S. typhimurium -infected group of animals (Figure 2D) compared to that of the mutant-infected groups of mice.

Finally, all of the measured cytokine levels were at the comparable levels in the ΔlppAB mutant-infected group of mice during the first week of infection in comparison to the ΔlppAB ΔmsbB mutant-infected group of animals. However, the trend changed at the later stages of infection (e.g., days 14 and 21 p.i.), during which time significantly higher levels of cytokines were observed in the ΔlppAB ΔmsbB mutant-infected group of mice (Figure 2).

To gauge protection provided by immunization of mice with the mutant strains compared to that of the WT bacterium, animals were initially immunized/infected orally with a relatively low dose (3 × 10³ CFU) of the mutant strains (ΔlppAB and ΔlppAB ΔmsbB) or WT S. typhimurium. This low dose ensured animal survival from WT S. typhimurium infection as its LD₅₀ was calculated to be 2.0 × 10⁵ CFU when administered via the oral route. The naïve mice received PBS and served as a control. No animal succumbed to infection during the initial immunization/infection period in all of the groups. On day 36 post immunization, these mice were orally challenged with 2 × 10⁶ CFU of WT S. typhimurium (Figure 3A). All mice vaccinated with either the ΔlppAB or the ΔlppAB ΔmsbB mutant survived, while 60% of the animals were protected in the WT S. typhimurium-vaccinated group. Naïve control mice (80%), on the other hand, succumbed to infection.

In a parallel experiment, mice were orally immunized with a high dose (1 × 10⁸ CFU) of ΔlppAB or the ΔlppAB ΔmsbB mutant, and no animal succumbed to infection during immunization. After 21 days, the immunized mice were challenged with WT S. typhimurium (1 × 10⁸ CFU) via the oral route and all survived. Mouse organs were excised 21 days post challenge for histopathological analysis. The high doses used for both immunization and challenge were to examine safety of the aforementioned mutants and their protective potential. The unimmunized naïve mice challenged with 1 × 10⁸ CFU of WT S. typhimurium via the oral route were used as a positive control for examining histopathological lesions in organs. These tissues were examined on day 7 post challenge as animals showed severe symptoms such as ruffled fur, lethargy, loss in weight, and diarrhea. As shown in Figure 3B, the WT S. typhimurium-challenged naïve mice had pyogranulomas in various peripheral organs, such as the liver and spleen (arrows), and the lungs had inflammatory cell infiltrate (solid triangle). The splenic tissues revealed lymphoid depletion in white pulps and leukocytosis. The depleted cell types were more likely hematopoietic or stromal cells in the red pulps. In addition, focal necrosis and thrombosis were observed in liver and spleen tissues (asterisks) of mice infected with the WT bacterium. However, intestinal tissue sections, e.g., duodenum, jejunum, and ileum, showed no marked lesions except for inflammation on day 7 p.i. (data not shown).

While analyzing tissues from both of the mutant-immunized groups of mice after WT S. typhimurium challenge, the above observed lesions (e.g., granulomas in the liver, arrows) and focal necrosis/thrombosis (asterisk) in the spleens were much milder in nature with largely intact red and white pulps, when compared to tissues from the WT S. typhimurium-infected naïve mice (Figure 3B).

As shown in Figure 4A, T-cells from both the mutant-immunized groups of mice significantly proliferated in response to the stimulation with WT S. typhimurium antigens. However, the extent of proliferation was much higher for the ΔlppAB mutant-vaccinated mice than that of ΔlppAB ΔmsbB mutant-immunized animals.

A robust cytokine production was also observed in T-cells from both the ΔlppAB and ΔlppAB ΔmsbB mutant-immunized mice in response to stimulation with WT S. typhimurium antigens. At 48 h after stimulation, T-cell production of TNF-α in the ΔlppAB-immunized mice was significantly higher when compared to T-cells from the ΔlppAB ΔmsbB mutant-immunized animals (Figure 4B). On the other hand, IFN-γ and IL-6 production from T-cells in the ΔlppAB ΔmsbB mutant-immunized animals (Figures 4C,D) were much higher when compared to the animals immunized with the ΔlppAB mutant.

At 72 h after stimulation, T-cell production of IFN-γ (Figure 4C) and IL-17A (Figure 4E) were much higher in the ΔlppAB mutant-immunized mice when compared to that in the ΔlppAB ΔmsbB mutant-vaccinated animals. The level of IL-10 was maintained at a similar level in both the mutant- immunized mice at both time points of 48 and 72 h (Figure 4F).

Compared to the un-pulsed controls, IFN-γ and IL-10 production were at higher levels in T-cells from both of the mutant infected mice (Figures 4C,F) at both the time points; however, IL-6 levels were specifically induced in T-cells of ΔlppAB ΔmsbB mutant-immunized mice at 48 h time point (Figure 4D), Likewise, TNF-α was significantly produced only from the T-cells of ΔlppAB mutant-immunized mice upon stimulation with the WT S. typhimurium antigens at a 48 h time point, but it was significantly dropped at 72 h (Figure 4B). In contrast, the level of TNF-α in the supernatant of pulsed T-cells from the ΔlppAB ΔmsbB mutant-immunized mice was slightly increased at 72 h time point and reached statistical significance when compared to the un-pulsed control (Figure 4B).

2D gel electrophoresis and analysis were performed to obtain protein profiling of WT S. typhimurium and that of its ΔlppAB and ΔlppAB ΔmsbB mutants. There were a total of 1726, 1899, and 1835 protein spots on the 2D gels for WT bacteria, and its ΔlppAB and ΔlppAB ΔmsbB mutants, respectively. After manual editing and filtering, a total of 1099 spots correlated across the gels of different groups of bacteria (e.g., WT S. typhimurium, ΔlppAB, and ΔlppAB ΔmsbB mutants; Table S1). Compared to the 2D gels for WT S. typhimurium, 22 spots had higher normalized volumes (NVs) and 47 spots had lower NVs in the ΔlppAB mutant (Table S2). These numbers were 46 (higher) and 14 (lower) for the ΔlppAB ΔmsbB mutant. Likewise, 10 spots on the 2D gel for the ΔlppAB mutant and 43 spots for the ΔlppAB ΔmsbB mutant had relatively higher NVs when these two mutants were compared to each other (Table S2).

A total of 61 spots were picked based on their fold changes and NVs for mass spectrometric analysis. Among these identified proteins, the majority of them belonged to bacterial inner and outer membranes such as structural proteins, receptors, and transporters. In addition, some proteases, metabolic enzymes, translation associated components, as well as cell signaling related molecules (synthases and regulatory proteins) were also observed. Interestingly, the levels of two proteins, outer membrane protein A (OmpA) and periplasmic protein TolB were higher in both the ΔlppAB and ΔlppAB ΔmsbB mutants when compared to that in the WT bacterium (Table 1).

In contrast, the level of seven proteins, including a serine protease, tricarboxylic transport, cAMP-regulatory protein, arginine-binding periplasmic protein 2 precursor, flagellar-associated GTP-binding protein, esterase, and superoxide dismutase, were generally lower in both the ΔlppAB and ΔlppAB ΔmsbB mutants (Table 1). Likewise, when comparison was made between the two mutants, DNA-directed RNA polymerase beta subunit (RpoB) and Rho factor were the two lead proteins produced significantly more with respective 8.1–39.4 and 7.4-fold changes (Table S3), respectively, in the ΔlppAB ΔmsbB mutant. On the other hand, protein chain initiation factor 2 and Lon protease were increased by 7.0 and 4.4-folds, respectively, in the ΔlppAB mutant. A complete list of 61 spots with their protein identification and fold changes are shown in Table S3.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.