Salmonella enterica serovar Typhimurium mutants completely lacking the F(0)F(1) ATPase are novel live attenuated vaccine strains.

The F(0)F(1) ATPase plays a central role in both the generation of ATP and the utilisation of ATP for cellular processes such as rotation of bacterial flagella. We have deleted the entire operon encoding the F(0)F(1) ATPase, as well as genes encoding individual F(0) or F(1) subunits, in Salmonella enteric serovar Typhimurium. These mutants were attenuated for virulence, as assessed by bacterial counts in the livers and spleens of intravenously infected mice. The attenuated in vivo growth of the entire atp operon mutant was complemented by the insertion of the atp operon into the malXY pseudogene region. Following clearance of the attenuated mutants from the organs, mice were protected against challenge with the virulent wild type parent strain. We have shown that the F(0)F(1) ATPase is important for bacterial growth in vivo and that atp mutants are effective live attenuated vaccines against Salmonella infection.

Introduction

Salmonella enterica is a diverse pathogen classified into >2400 serovars and is the cause of important infections in both humans and livestock. S. enterica serovar Typhi (S. Typhi) is the causative agent of typhoid fever, a serious systemic disease in humans. It is estimated that there are 22 million cases of typhoid fever annually worldwide, resulting in 200,000 deaths [1,2]. Vaccination against S. Typhi is a potentially attractive method of disease control, but current vaccines have significant drawbacks and there is a need for improved versions [3,4]. S. enterica serovar Typhimurium (S. Typhimurium), a common cause of gastroenteritis (salmonellosis) in humans, has added significance because infection of mice with this serovar generates a systemic infection with important similarities to human typhoid fever. This mouse model has been used extensively to study typhoid-like infections [5,6].

The F0F1 ATPase is a complex of membrane proteins found in eukaryotes and prokaryotes that has been best studied in mitochondria [7,8], chloroplasts [9,10] and Escherichia coli [11-13]. It plays a central role in energy transduction, generating ATP from ADP and Pi substrates via oxidative phosphorylation. The synthesis of ATP is driven by the flow of protons into the cell, generating a proton motive force which energises processes such as motility and active transport [14-17]. In E. coli, the genes encoding the F0F1 ATPase are located in a single operon, atpIBEFHAGDC, transcribed from a promoter upstream of atpI [18-20]. The F0 subunit of the ATPase is a hydrophobic membrane-embedded proton channel encoded by genes atpBEF. The F1 subunit constitutes the catalytic ATPase, encoded by atpHAGDC [19,21]. The first gene in the operon, atpI, has no defined function and does not appear to form part of the F0F1 ATPase complex [22]. This genetic organisation is conserved between E. coli and S. Typhimurium.

A comprehensive identification of genes required for S. Typhimurium infection of mice by our laboratory identified mutation of atpA as an attenuating lesion [23]. A defined atpA deletion mutant was subsequently confirmed to be attenuated for growth in vivo and furthermore was found to offer significant protection against subsequent challenge [23]. Here we present a full analysis of the role of the F0F1 ATPase in S. Typhimurium infection and the potential use of mutants in the atp operon as live attenuated vaccines.

Materials and methods

Bacterial strains and growth conditions

The bacterial strains and plasmids used in this study are shown in Table 1. Bacteria were grown at 37 °C in Luria–Bertani (LB) broth or on LB agar. Media were supplemented with antibiotics where stated, at the following concentrations, kanamycin 50 μg/ml, ampicillin 100 μg/ml and chloramphenicol 25 μg/ml. Minimal medium (used to determine carbon source utilisation) consisted of M9 salts (Sigma Dorset UK) supplemented with 0.1 mM CaCl2, 1 mM MgSO4, 4 μg/ml histidine and the stated carbon source at 0.4% (final w/v).

Table 1

Salmonella strains and plasmids used in this study.

Bacterial strain or plasmid	Description	References
SL1344	Wild type parent for this work, mouse-virulent strain of S. Typhimurium	[43]
LB5010	S. Typhimurium. LT2 galE mutant, r− m+	[44]
SL3261	aroA mutant of SL1344. Attenuated in mice. Well-characterised vaccine strain	[43,45]
SL1344 atp	atpIBEFHAGDC deletion mutant in SL1344	This study
SL1344 F0	atpBEF deletion mutant in SL1344	This study
SL1344 F1	atpHAGDC deletion mutant in SL1344	This study
SL1344 atp (malXYatp operon+)	SL1344 atp complemented by insertion of atp operon into the malXY pseudogene region	This study
SL1344 atp (malXY CmR)	SL1344 atp with chloramphenicol resistance cassette inserted into the malXY pseudogene region	This study
pBADλred	Plasmid expressing exo bet gam genes from bacteriophage lambda	[46]
	Exo bet gam genes as NcoI-HindIII fragment in pBAD/HisA (invitrogen), AmpR
pCP20	Plasmid expressing Flpase genes from Saccharomyces cerevisiae. ClaI-XbaI fragment of pMMC6 in pHSG415, AmpR, CmR	[24]
	Replicates poorly above 37 °C, stably inherited at 30 °C
pBADkanFRT	pBADTOPO (Invitrogen, Paisley, UK) containing kanamycin cassette flanked by FRT sites. KanR, AmpR	[46]

Construction of mutants and complementation

Oligo-directed mutagenesis (ODM), an adaptation of ET-cloning, was used to replace the target genes on the Salmonella chromosome with a kanamycin resistance cassette flanked with FRT regions from pBADkanFRT [24,25]. PCR was used to amplify the kanamycin resistance FRT cassette with 5′ and 3′ 60 bp arms homologous to DNA flanking the target genes (see Table 2 for primer sequences). S. Typhimurium LB5010 containing pBADλred was grown in LB broth supplemented with ampicillin to an OD595 of 0.25. Arabinose was added to 0.2% (final w/v) to induce red gene expression. Cultures were grown to OD595 0.5 and electroporated with the purified ODM PCR product described above. Mutant colonies were selected on LB agar plates supplemented with 50 μg/ml kanamycin. The desired allelic replacement of the target genes was confirmed by PCR (see Table 2 for primer sequences). Mutations in S. Typhimurium LB5010 were transduced into SL1344 by bacteriophage P22 as described previously [26] with selection on LB agar plus kanamycin and gene deletions were confirmed to be correct by PCR and sequencing.

Table 2

Primers used in this study.

Name	Sequence 5′–3′	Use
Op ODM For	ttgagaggataaaaaaaaaccagtccgcaatcagactggttttatgctttcaagccggtgtagggataggcttaccttcaagctc	Generation of construct to delete atp operon. Kanamycin FRT cassette sequence underlined
Op ODM Rev	cttaaagaacgttttatacgacacgcggcatacctcgaagtgagcaggagtaaaaacgtgatgacgatgacaagctccccctttcg	Generation of construct to delete atp operon. Kanamycin FRT cassette sequence underlined
F0 ODM For	aagctgctttggcgtaggggcgagctaccgtaacaaattcagacatcagcccctccctcctagggataggcttaccttcaagctc	Generation of construct to delete F0 region, atpBEF. Kanamycin FRT cassette sequence underlined
F0 ODM Rev	ggtgctggtggttcagatactggcgccggctgtaattaacaacaaagggtaaaaggcatcatgacgatgacaagctccccctttcg	Generation of construct to delete F0region, atpBEF. Kanamycin FRT cassette sequence underlined
F1 ODM For	ttgagaggataaaaaaaaaccagtccgcaatcagactggttttatgctttcaagccggtgtagggataggcttaccttcaagctc	Generation of construct to delete F1region, atpHAGDC. Kanamycin FRT cassette sequence underlined
F1 ODM Rev	agctgctaacagcgacatcgtggataaacttgtcgctgaactgtaaggagggaggggctgatgacgatgacaagctccccctttcg	Generation of construct to delete F1region, atpHAGDC. Kanamycin FRT cassette sequence underlined
Op test For	cacaatgtgcagatgccaatgacag	Confirmation of atp operon deletion
Op test Rev	atgtttaatgtgtgatctggtgcac	Confirmation of atp operon deletion.
F0 test For	ggttctgaccgttttcgtctaactg	Confirmation of F0 region, atpBEF deletion
F0 test Rev	atgatatttgcctggcgtcacc	Confirmation of F0 region, atpBEF deletion
F1 test For	aatgacatagtaataatccctcat	Confirmation of F1 region, atpHAGDC deletion.
F1 test Rev	cgcgctcagatcctggacgaagcca	Confirmation of F1 region, atpHAGDC deletion
F0F1comp For	ccgcaggttcagtcggtaaaagatgaaatggttggcctgatgaataccgttcaggcataacgacgcggcttgtgttaaaaatcgac	Generation of construct to insert the atp operon into the malXY pseudogene region. Homologous malX sequence underlined
F0F1comp Rev	ctacgtacaccatgtcccgcgtcggtcaacttcctgtgaaaaatcgaacatatcccttccgcttattatcacttattcaggcg	Generation of construct to insert the atp operon into the malXY pseudogene region. Homologous malY sequence underlined
F0F1 test For	catcgtgagtctggacaactgcat	Confirmation of atp operon insertion into malXY pseudogene region
F0F1 test Rev	ataatcccactacgtacaccatgtc	Confirmation of atp operon insertion into malXY pseudogene

The kanamycin resistance FRT cassette was then excised to leave only a 128 bp FRT scar site. Briefly, electrocompetent mutants of SL1344 were transformed with pCP20 [24] grown at 30 °C and then plated onto LB agar containing 100 μg/ml ampicillin. Single colonies were grown in LB at 39 °C (to prevent replication of pCP20) for 6 h then diluted and plated onto LB agar and incubated overnight at 39 °C. Colonies were screened for loss of ampicillin and kanamycin resistance. Excision of the kanamycin resistance FRT cassette was confirmed by PCR and sequencing to be correct. Southern blot using the FRT scar site region as a probe was also used to confirm that the final mutants were as intended. LPS serotype was confirmed by agglutination with anti-04 serotype antiserum using anti-09 antiserum as a negative control (Remel Europe Ltd./Oxoid Ltd., Basingstoke UK).

For complementation of SL1344 atp, lacking the entire atp operon, PCR was used to amplify the entire atp operon from SL1344 fused to a chloramphenicol resistance cassette, from pACYC184. This was inserted into the malXY pseudogene region on the Salmonella chromosome using ODM with selection on chloramphenicol. Insertion of the atp operon into malXY was confirmed by PCR and sequencing of the mutated malXY junction and by Southern blotting using atpG as the probe. In addition to the complemented strain, SL1344 atp (malXY atp operon+), a complementation control strain was also generated, SL1344 atp (malXY CmR). For this control strain a chloramphenicol resistance cassette was inserted into the malXY pseudogene region of SL1344 atp to ensure the insertion into the pseudogene had no phenotypic effects.

Growth in vitro and succinate utilisation

Cultures in 5 ml of LB broth were incubated overnight with shaking (180 rpm) at 37 °C. Cultures were diluted 1:100,000 into 100 ml of pre-warmed LB broth, and incubated with shaking at 37 °C. Growth was measured by viable count on LB agar plates. Exponential generation times were calculated from growth rates between 4 and 6 h.

To assess the ability to utilise succinate as a sole carbon source wild type and the various atp mutants were grown in M9 minimal medium supplemented with 0.4% (w/v) of sodium succinate. Growth was assessed by OD595 after 24 and 48 h.

Mouse typhoid model

Inocula were prepared from overnight cultures grown statically in LB broth at 37 °C. Cultures were centrifuged and bacteria were re-suspended in phosphate buffered saline (pH 7.4) to the required concentration. Seven to nine week-old female BALB/c mice (Harlan, Oxon, UK) were inoculated with 200 μl of bacteria suspension via intravenous injection, or they were lightly anaesthetised with halothane and inoculated by oral gavage. Doses of bacteria given were confirmed by viable counts in LB agar. Gene knock-out mice lacking gp91phox or IFNγR1 on a C57/BL6j background where originally purchased from Jackson Laboratory (Bar Marbour, ME) and maintained as homozygous matings at the Wellcome Trust Sanger Institute. C57/BL6j age- and sex-matched control mice were purchased from Harlan (Oxon, UK).

At pre-determined time points postinfection animals were killed, spleens and livers removed and homogenised in 5 ml of sterile water in a Stomacher® 80 Lab System (Seward). Bacterial numbers were enumerated via serial dilutions and plating in LB agar. When required, blood was collected via cardiac puncture under terminal anaesthesia.

For vaccination studies, animals were immunised intravenously with 105 CFU or orally with 109 CFU. At these doses, immunising strains did not induce clinical signs, were completely cleared with all mice surviving the infection. At 13 weeks postimmunisation clearance of the bacteria was confirmed by viable counts from spleens and livers. Mice were subsequently re-challenged either intravenously with 104 CFU, or orally with 108 CFU of SL1344. Age-matched unimmunised mice were included for comparison. Viable counts in the target organs were enumerated as detailed above. All work was licensed by the UK Home Office.

Histolopathological analysis

For histopathological analysis, a portion of spleen was fixed in 10% buffered formalin then embedded in paraffin wax. Four 3 μm sections were cut approximately 20–30 μM apart then stained with Haematoxylin and Eosin (H&E). Spleen sections were examined microscopically.

Anti-Salmonella antibodies assayed by ELISA

Sonicated SL1344 was used as the ELISA capture antigen to assay anti-Salmonella antibodies following vaccination. This was diluted in carbonate coating buffer (1.59 g/l sodium carbonate, 2.93 g/l sodium bicarbonate, pH 8.2) to 1 × 106 bacteria/ml, based on the viable count of the original culture. 100 μl of this antigen solution was used to coat the wells of an ELISA plate (Immunoplates, Nunc, Thermofisher Scientific, Lutterworth, UK) through overnight incubation at 4 °C. Plates were washed with washing buffer (PBS containing 0.05%, w/v, Tween 20) then wells were blocked with 300 μl/well of blocking buffer (PBS containing 1% bovine serum albumin) for 2 h. Serial fivefold dilutions of heat-inactivated mouse serum were prepared in blocking buffer and 100 μl were added to washed plates. Sera from normal mice and known positive sera were included on each plate as negative and positive controls. Plates were incubated for 2 h at room temperature. Total antibody was detected using 100 μl/well of biotinylated goat anti-mouse immunoglobulins (Dako, Ely, UK) diluted 1:1000 in blocking buffer. Subtypes IgG1 and IgG2a were detected using 100 μl/well of biotinylated rat anti-mouse IgG1 or IgG2a antibodies (BD Bioscience, Oxford, UK) diluted 1:500 in blocking buffer. Plates were incubated with secondary antibody for 1 h at room temperature and then washed three times in wash buffer. Then 100 μl/well of streptavidin (BD Bioscience, Oxford, UK), diluted 1:100 in blocking buffer, was added and plates were incubated in the dark for 30 minutes. Plates were then washed and developed with 100 μl TMB substrate solution (BD Bioscience, Oxford, UK) and the reaction stopped with the addition of 50 μl/well of 5N sulphuric acid. Absorbance was read at 450 nm. Data presented are from dilutions of 1:12,500 for total Ig and 1:2500 for Ig subclasses.

Macrophage infection in vitro

RAW 264.7 cells were seeded into 96 well plates at a density of 2 × 105 cells/well in RPMI medium (Sigma Dorset, UK) supplemented with 10% FCS and 2 mM l-glutamate. Plates were seeded the evening before infection and incubated throughout at 37 °C with 5% CO2. For the bacterial inoculum, overnight cultures were diluted 1:10 into fresh LB broth and incubated for 2 hr at 37 °C with shaking. Bacteria were collected by centrifugation, re-suspended in PBS and diluted in tissue-culture medium to the required concentration. Bacteria were added to host cells and incubated at 37 °C 5% CO2 for 2 h. The monolayer was washed twice in pre-warmed PBS and medium containing 50 μg/ml gentamicin was added to kill extracellular bacteria. Following incubation for 1 h host cells were washed twice with PBS and medium containing 10 μg/ml gentamicin was added for the remainder of the experiment. Intracellular bacteria were enumerated by serial dilution and plating on LB agar following lysis of host cells with 0.5% Triton 100×. Following the manufacturer's instructions, the Cytotox96 assay kit (Promega, Southampton, UK) was used to determine the relative viability of host cells after infection.

Statistical analysis

Statistical analysis was performed using Student's t-test or one-way ANOVA with Bonferroni correction. P ≤ 0.05 was considered significant.

Results

Growth of S. Typhimurium atp mutants in vitro

Deletion mutants were generated in SL1344 that lacked the entire atp operon or the F0 or F1 components only. When grown in LB broth the various atp mutants all had similar generation times in comparison with SL1344. These were 29.72 (±0.78) min for SL1344, 32.22 (±1.90) min for SL1344 F0, 33.12 (±1.06) min for SL1344 F1 and 29.24 (±0.85) min for SL1344 atp (all mean ± SEM from 3 replicates). However, final viable bacterial counts of overnight cultures were consistently lower in the various atp mutants compared to SL1344. The viable counts in 24 hr cultures were log10 9.69 CFU (±0.08) for SL1344, log10 9.19 CFU (±0.04) for SL1344 F0, log10 9.21 CFU (±0.16) for SL1344 F1 and log10 9.29 CFU (±0.09) for SL1344 atp (all mean ± SEM from 3 replicates), although these differences were only statistically significant between SL1344 and SL1344 F0.

As seen with mutations in the atp operon in E. coli [27], Bacillus subtilis [28] and S. Typhimurium [29] all our atp mutants were unable to utilise succinate as a sole carbon or energy source. The three atp mutants showed no growth after 24 or 48 h, as measured by OD595. The atp mutants had OD595 readings of 0.001 (±0.001) for SL1344 atp, 0.0015 (±0.0005) for SL1344 F0 and 0.0015 (±0.0005) for SL1344 F1 at 48hrs, whereas SL1344 showed visible growth at both 24 and 48 h, with OD595 readings of 0.0335 (±0.01) and 0.374 (±0.07) respectively (all mean ± SEM from 3 replicates).

SL1344 atp mutants are attenuated in a mouse model of typhoid fever

Previous studies have shown that individual gene deletions or transposon insertions in the atp operon attenuate S. Typhimurium in both mice and chickens [23,29,30] but attenuation following deletion of the whole operon or individual subunits has not been tested.

To assess the level of attenuation caused by the deletion of the F0 or F1 subunits, or the entire atp operon, BALB/c mice were infected intravenously with 105 CFU of SL1344, SL1344 F0, SL1344 F1 or SL1344 atp. Bacterial loads in the spleens and livers were enumerated at the time points shown (Fig. 1). In both spleens and livers, bacterial counts were significantly lower in mice infected with the various atp mutants in comparison with those infected with SL1344. The three atp mutants showed little net bacterial growth between days 1 and 3 postinfection whereas bacterial loads in mice infected with SL1344 increased by nearly 3 logs over the same period. By day 7 the various atp mutants showed no significant bacterial growth, with counts similar to those at day 3, whereas mice infected with SL1344 would have been dead by this time point.

Fig. 1

Bacterial counts in spleens and livers following intravenous infection with various atp mutants or SL1344. Female BALB/c mice were infected intravenously with 105 CFU of S. Typhimurium SL1344, SL1344 atp, SL1344 F0 or SL1344 F1. Bacterial numbers in the spleens (A) and livers (B) were enumerated. Data are presented as mean log10 CFU ± SEM (n = 4), representative of two experiments giving similar results. (+) SL1344 infected mice were not included on day 7 as mice would not survive to this time point. *P ≤ 0.05 compared to SL1344.

Intravenous immunisation with Salmonellaatp mutants confers protection against subsequent re-challenge

Following immunisation with the three atp mutants, mice were re-challenged intravenously with SL1344 (Fig. 2). The wild type infection grew rapidly as expected in unimmunised control mice whereas mice immunised with the atp mutants had significantly lower bacterial counts in spleens and livers at days 1 and 4 postinfection. Bacterial counts were comparable between the animals immunised with the different atp mutants and with mice immunised with the well-characterised aroA mutant vaccine strain, SL3261. Therefore SL1344 F0, SL1344 F1 and SL1344 atp were all protective against subsequent challenge. Since all three atp mutants behaved the same in terms of attenuated growth in vivo and protection against subsequent infection, SL1344 atp was selected for further characterisation.

Fig. 2

Protection following intravenous vaccination with various atp mutants or SL3261. Female BALB/c mice were immunised intravenously with 105 CFU of SL1344 atp, SL1344 F0, SL1344 F1 or SL3261. After 13 weeks, following clearance of the immunising strains, immunised and age-matched unimmunised mice were challenged intravenously with 104 CFU of SL1344. Bacterial numbers in the spleens (A) and livers (B) were enumerated. Data are presented as mean log10 CFU ± SEM (n = 4), representative of two experiments giving similar results. (+) Unimmunised mice were not included on day 7 as mice would not survive to this time point in contrast to immunised mice which all did survive until then. *P < 0.05 compared to unimmunised mice.

Complementation of SL1344 atp with the atp operon restores virulence

To confirm that the attenuation of SL1344 atp was specifically due to the deletion of the atp operon, SL1344 atp was complemented by insertion of the whole atp operon fused to a chloramphenicol resistance cassette into the malXY pseudogene region to generate strain SL1344 atp (malXY atp operon+). BALB/c mice were infected intravenously with 105 CFU of SL1344, SL1344 atp, SL1344 atp (malXY atp operon+) and SL1344 atp (malXY CmR). The complemented strain, SL1344 atp (malXY atp operon+) displayed a wild type-like phenotype with increased bacterial loads in livers and spleens relative to SL1344 atp at days 1, 2 and 3 postinfection (Fig. 3). Insertion of the chloramphenicol resistance cassette into the malXY region in strain SL1344 atp (malXY CmR) had no effect on bacterial counts compared to SL1344 atp (Fig. 3).

Fig. 3

Bacterial counts following intravenous infection with SL1344, SL1344 atp, SL1344 atp (malXY atp operon+) or SL1344 atp operon (malXY CmR). Female BALB/c mice were infected intravenously with 105 CFU of SL1344, SL1344 atp, SL1344 atp (malXY atp operon +) or SL1344 atp (malXY CmR). Bacterial numbers in the spleens (A) and livers (B) were enumerated. Data are represented as mean log10 CFU ± SEM (n = 4), representative of two experiments giving similar results. *P ≤ 0.05 significant compared to SL1344.

SL1344 atp is not impaired for infection of macrophages in vitro

Survival and replication of SL1344 and SL1344 atp were assessed in the RAW 264.7 murine macrophage-like cell line. Host cells were infected at MOIs of 1 and 10 and intracellular bacterial counts and macrophage survival were determined at 3 and 24 h postinfection. At both MOIs and at both time points intracellular bacterial viable counts and macrophage survival were similar after infection with SL1344 or SL1344 atp with no statistically significant difference between the two strains (Fig. 4).

Fig. 4

Intracellular survival of SL1344 or SL1344 atp in RAW264.7 cells. Intracellular bacteria (A and C) and macrophage survival (B and D) were determined at 3 and 24 h postinfection, with MOIs of I (A and B) and 10 (C and D). Macrophage survival expressed as percentage of survival compared to uninfected cells. Data are represented as mean ± SEM from three experiments each performed in triplicate and giving similar results.

NADPH oxidase (phox) and IFNγ contribute to control of SL1344 atp infection

To begin to define the immunological components required to control infection with SL1344 atp and to assess the potential use of SL1344 atp immunisation in immunocompromised individuals, two gene knock-out mouse strains and their respective wild types were infected with SL1344 atp. Following infection with SL1344 atp, gp91phox−/− mice had significantly increased bacterial loads in spleens and livers relative to wild type mice (Fig. 5A) as did mice lacking IFNγR1 (Fig. 5B). These data indicate a significant role for NADPH oxidase and IFNγ in controlling bacterial proliferation following infection with SL1344 atp. Similarly, both immune components were needed for control of SL3261 replication (Fig. 5).

Fig. 5

Bacterial counts in gp91phox knock-out, IFNγR1 knock-out or wild type C57/BL6 mice. Mice were infected intravenously with 105 CFU of SL1344 atp or SL3261. Bacterial numbers in the spleens and livers of gp91phox knock-out or wild type C57/BL6 mice were enumerated on day 4 postinfection (3–5 mice per point). Bacterial counts in the spleens and livers of IFNγR1 knock-out or wild type C57/BL6 mice were enumerated on day 9 postinfection (4–5 mice per point). Data are presented as mean log10 CFU ± SEM. *P < 0.05 knockout mice compared to wild type mice for each bacterial strain.

Intravenous immunisation with SL1344 atp confers protection against subsequent oral re-challenge

SL1344 atp was assessed for its ability to protect against subsequent oral re-challenge (Fig. 6). Again, the wild type challenge grew rapidly, as expected, in unimmunised mice whereas mice immunised with SL1344 atp had significantly reduced bacterial counts in spleens on days 3, 4 and 7 and in livers on days 4 and 7 postinfection (Fig. 6). Similar levels of protection were observed between SL1344 atp and SL3261-immunised mice (Fig. 6). Therefore, SL1344 atp is protective against subsequent oral challenge and this protection is as effective as immunisation with SL3261.

Fig. 6

Protection following intravenous vaccination with SL1344 atp or SL3261. Female BALB/c mice were immunised intravenously with 105 CFU of SL1344 atp or SL3261. After 13 weeks, following clearance of the immunising strains, immunised and age-matched unimmunised mice were challenged orally with 109 CFU SL1344. Bacterial counts in the spleens (A) and livers (B) were enumerated. Data are presented as mean log10 CFU ± SEM (n = 4), representative of two experiments giving similar results. (+) Unimmunised mice were not included on day 14 as mice would not survive to this time point in contrast to immunised mice which all did survive until then. *P < 0.05 immunised compared to unimmunised mice.

SL1344 atp is protective following oral immunisation

SL1344 atp was further assessed for protection following oral immunisation, given that this would be the preferred route of immunisation with a live attenuated vaccine. The wild type infection grew as expected in unimmunised mice whereas those immunised with SL1344 atp had significantly lower bacterial counts in spleens and livers after being re-challenged intravenously (Fig. 7A and B). Little net bacterial growth was observed in challenged SL1344 atp immunised mice, with similar levels of bacteria seen over 14 days. Following oral re-challenge, SL1344 atp immunised mice showed reduced bacterial counts on days 3 and 7 postinfection relative to unimmunised mice (Fig. 7C and D). Furthermore, bacterial numbers following SL1344 atp oral immunisation were comparable to those seen in SL3261-immunised mice regardless of the re-challenge route. The SL1344 atp mutant is therefore protective following oral administration and is as effective as SL3261 as a vaccine.

Fig. 7

Protection following oral vaccination with SL1344 atp or SL3261. Female BALB/c mice were immunised orally with 109 CFU of SL1344 atp or SL3261. After 13 weeks, following clearance of the immunising strains, immunised and age-matched unimmunised mice were challenged either intravenously (A and B) with 104 of CFU SL1344 or orally (C and D) with 109 CFU of SL1344. Bacterial counts in spleens and livers were enumerated. Data are presented as mean log10 CFU ± SEM (n = 4), representative of two experiments giving similar results. (+) Unimmunised mice were not included on days 7 and 14 during intravenous challenge and day 14 during oral challenge as mice would not survive to this time point in contrast to immunised mice which all did survive until these times. *P < 0.05 immunised compared to unimmunised mice.

Antibody responses following SL1344 atp immunisation

Pooled sera from mice immunised intravenously and orally were assayed for antibodies specific for S. Typhimurium. Mice intravenously immunised with SL1344 atp had significantly higher levels of total antibody against S. Typhimurium than unimmunised mice (Fig. 8A). Levels of total antibody in mice intravenously immunised with SL1344 atp were comparable to those elicited in SL3261-immunised mice. Total antibody levels following oral immunisation were lower than those seen in intravenously immunised animals, however SL1344 atp immunised mice showed higher levels of total antibody compared to unimmunised mice although this did reach statistical significance. Compared with SL3261-immunised mice the antibody levels were lower in SL1344 atp immunised mice although this was not statistically significant.

Fig. 8

Anti-Salmonella antibody levels following vaccination with SL1344 atp or SL3261. Pooled sera from 3 to 4 mice taken 11 weeks after intravenous immunisation with 105 CFU or oral immunisation with 109 CFU of SL1344 atp or SL3261. Total Ig levels (A) following intravenous or oral immunisation. IgG subclasses IgG1 (B) and IgG2a (C) assayed following intravenous or oral immunisation. *P ≤ 0.05 comparing SL1344 atp and SL3261 with unimmunised controls.

The humoral immune response was further characterised with the determination of IgG subclass levels elicited following immunisation with SL1344 atp (Fig. 8B and C). In comparison with unimmunised mice, levels of 1gG21 and IgG2a were higher in SL1344 atp immunised mice regardless of immunisation route and SL1344 atp immunised mice elicited similar levels of both IgG subclasses in comparison to SL3261.

Reduced splenomegaly following intravenous immunisation with SL1344 atp compared to SL3261

To assess the level of splenomegaly induced following intravenous immunisation with SL1344 atp and SL3261, mice were intravenously immunised with 105 CFU and spleen weights were measured along with bacterial viable counts (Fig. 9). In comparison with uninfected age-matched mice, a significant increase in spleen weight was observed in mice immunised with both SL1344 atp and SL3261 on days 7, 14, 21 and 28 postinfection (Fig. 9A). In addition, SL3261-immunised mice also showed a significant increase in spleen weight relative to uninfected age-matched mice on days 3 and 4 postinfection. Spleen weights of mice immunised with SL3261 were significantly increased relative to those immunised with SL1344 atp on days 7, 14 and 21 postinfection (Fig. 9A). The reduced splenomegaly following immunisation with SL1344 atp compared to SL3261, corresponded with lower splenic bacterial counts of SL1344 atp which may contribute to the reduced pathology (Fig. 9A and B). Although spleen weights were similar from day 28 onwards in all immunised mice, bacterial counts in the spleens were significantly greater in mice immunised with SL1344 atp relative to those immunised with SL3261, from days 28 to 56 postinfection. At 63 days postinfection spleen weights of both immunised groups decreased to a similar level as uninfected controls (data not shown). However SL1344 atp immunised mice did not clear bacteria from the spleen until day 77 postinfection, whereas SL3261-immunised animals cleared bacteria at day 63.

Fig. 9

Spleen weight and bacterial counts following vaccination with SL1344 atp or SL3261. Female BALB/c mice were immunised intravenously with 105 CFU of SL1344 atp or SL3261 and age-matched unimmunised controls were included for spleen weight comparisons. Spleens were weighted (A) and bacterial counts in spleens (B) and livers (C) were enumerated. Data are presented as mean grams ± SEM for spleen weight or mean log10 CFU ± SEM for bacterial count (n = 4–8). Data are pooled from two experiments giving similar results. +P ≤ 0.05 uninfected compared with both infected groups, *P ≤ 0.05 SL1344 atp infected mice compared to SL3261 infected mice.

In contrast, both SL3261 and SL1344 atp immunised mice showed no significant change in liver weight compared with unimmunised controls (data not shown). SL3261 and SL1344 atp were both cleared from the livers of immunised mice by day 56 (Fig. 9C).

Histopathological analysis of H&E-stained sections from the spleens of SL3261-immunised mice showed the presence of granulomatous inflammation and areas of pyogranulomatous inflammation with necrosis on day 7 postinfection. In addition SL3261-immunised mice displayed large amounts of lymphoid hyperplasia in conjunction and lymphoid coalescence, resulting in the inability to distinguish red and white pulp areas. These effects were still evident on day 14 postinfection, albeit reduced compared to day 7. At both time points, but especially at day 7, SL1344 atp immunised mice displayed much reduced histopathological effects relative to those immunised with SL3261 (data not shown).

Discussion

We have examined the role of the F0F1 ATPase in S. Typhimurium infection and shown that mutants in this protein complex have potential as live attenuated vaccine strains. The atpA gene has previously been identified by our laboratory as part of a screen of transposon mutants, as being required by S. Typhimurium for infection of mice [23]. A defined atpA mutant was used subsequently to confirm the role of the F0F1 ATPase in infection, and was also able to confer protection against subsequent wild type challenge. This agrees with previous data showing a role for the F0F1 ATPase in Salmonella infections of mice and chickens [29,30]. We have further characterised the role of the F0F1 ATPase by comparison of defined non-polar mutants lacking the entire atp operon or the F0 or F1 subunits in SL1344. This is a significant advance on previous work which used undefined or potentially polar mutations. Likewise, the use of atp mutants as vaccine strains has not been examined in detail. Our mutants were characterised with respect to their growth in vitro and in the mouse model of typhoid fever. All mutants grew as well as SL1344 in LB broth although they reached a slightly lower bacterial cell density at stationary phase. Unlike SL1344, the various atp mutants were unable to utilise succinate when it was supplied as the sole carbon source. This inability to use succinate for growth has been shown before for atp mutants in E. coli, S. Typhimurium and B. subtilis [27-29].

In the mouse typhoid model, all three atp mutants were significantly attenuated for growth with bacterial counts in the spleens and livers of infected mice much lower than those in the organs of mice infected with SL1344. The three atp mutants had similar bacterial counts in vivo indicating that they were all attenuated to a similar degree and that the two components, F0 and F1, are equally important for growth in vivo with neither subunit contributing to infection independently of the other. This work is the first direct comparison of the relative roles in infection of the two subunits. Our previous demonstration that immunisation with SL1344 atpA conferred protection against subsequent SL1344 challenge [23], prompted comparison of the protective efficacy of the atp mutant strains generated in this study. All three atp mutants protected against SL1344 challenge and did so to a similar degree as the prototype live attenuated vaccine strain SL3261. Given that the three atp mutants behaved similarly in terms of attenuation and protection, SL1344 atp, lacking the genes encoding the entire atp operon was selected for further characterisation. This mutant has the potential advantage of not displaying artefact phenotypes caused by the presence of non-functional F0F1 ATPase components. Importantly, complementation of SL1344 atp with the atp operon restored bacterial growth in vivo to wild type levels confirming the phenotype was due to the specific deletion of the atp operon and not due to secondary mutations. SL1344 atp elicited significant protection against virulent challenge when delivered orally, which is likely to be the preferred route of vaccine administration. In addition it was protective against oral challenge, which is the natural route of infection. Furthermore, mice immunised with SL1344 atp generated an anti-Salmonella antibody response with antibody levels similar to those of SL3261-immunised animals. Of particular note is the production of IgG2a antibodies which are known to play an important role in the rapid clearance of Salmonellae through complement activation and the promotion of phagocytosis by macrophages [31-33].

Immunisation with both SL3261 and SL1344 atp caused splenomegaly as evidenced by increased spleen weights compared to unimmunised controls. However, the increase in spleen weight was significantly reduced in mice immunised with SL1344 atp versus SL3261. This was further examined via histopathological analysis of H&E-stained spleen sections. Consistent with the differences in spleen weights following immunisation, SL1344 atp immunised mice showed reduced inflammation and reactogenicity compared to mice immunised with SL3261. This reduction in splenomegaly following SL1344 atp immunisation may be a potential benefit of immunisation with SL1344 atp.

The ability to infect host macrophages and survive within them is a key process in Salmonella infection and mutants impaired in this property are typically attenuated in the mouse model [34]. The ability of SL1344 atp to infect and grow within RAW cells was not impaired compared to SL1344. The attenuated growth in vivo of SL1344 atp is therefore not due to an inherent defect in the infection of and growth within host macrophages. This agrees with previous data showing various Salmonella atp mutants had no significant deficiency in intracellular survival [29,30]. However, this finding does not exclude the possibility of a defect in this property being manifested specifically in vivo where conditions are likely to be very different from those in vitro.

Understanding the components of the immune system required to control infection and generate protection following immunisation with live attenuated vaccine strains is of interest as it may offer the potential to enhance immunogenicity and reduce reactogenicity. It also has significance for the use of these strains in immunocompromised hosts. Therefore, IFNγR1−/− and gp91 phox −/− counterparts along with their wild type C57BL/6 mice were infected with SL1344 atp. These gene knock-out mice are of particular interest as they represent immune defects found in humans. Genetic deficiencies in the NADPH oxidase system (phox) manifest as chronic granulamatous disease [35], while deficiencies in IFNγ activity lead to increased susceptibility to bacterial and fungal infections, particularly with mycobacteria [36,37]. Both NADPH oxidase and IFNγ were required to control SL1344 atp infection with bacterial counts in livers and spleens significantly higher in the absence of these host defence mechanisms. A similar effect was seen in mice infected with SL3261. These data are perhaps not surprising given the central role of both NADPH oxidase and IFNγ in the control of S. Typhimurium infection in mice [38-40]. However, it is still valuable to compare different vaccine strains in different gene knock-out mice since the mechanisms controlling infection are known to differ between vaccine strains [41,42]. For instance, while IFNγ is required to control infection with SL3261 as shown here and by Vancott et al. [41] it is dispensable for control of infection with a phoP mutant.

In summary, we have investigated the role of the F0F1 ATPase in S. Typhimurium infection and shown that this protein complex makes a significant contribution to bacterial growth in vivo. Furthermore, mutants lacking the atp operon have potential utility as novel live attenuated vaccine strains against Salmonella infection.