An aromatic amino acid auxotrophic mutant of Bordetella bronchiseptica is attenuated and immunogenic in a mouse model of infection

Authors


*Corresponding author. Tel.: +61 (242) 213439; Fax: +61 (242) 214135, E-mail address: mwalker@uow.edu.au

Abstract

We have constructed an aromatic amino acid auxotrophic mutant of Bordetella bronchiseptica, harbouring mutations in aroA and trpE to investigate the use of such a strain as a live-attenuated vaccine. B. bronchiseptica aroA trpE was unable to grow in minimal medium without aromatic supplementation. Compared to the parental wild-type strain, the mutant displayed significantly reduced abilities to invade and survive within the mouse macrophage-like cell line J774A.1 in vitro and in the murine respiratory tract following experimental intranasal infection. Mice vaccinated with B. bronchiseptica aroA trpE displayed significant dose-dependent increases in B. bronchiseptica-specific antibody responses, and exhibited increases in the number of B. bronchiseptica-reactive spleen cells in lymphoproliferation assays. Immunised animals were protected against lung colonisation after challenge with the wild-type parental strain. With such a broad host range displayed by B. bronchiseptica, the attenuated strain constructed in this study may not only be used for the prevention of B. bronchiseptica-associated disease, but also for the potential delivery of heterologous antigen.

1Introduction

Bordetella bronchiseptica is a Gram-negative respiratory pathogen of a variety of animals. This bacterium causes kennel cough in dogs, bronchopneumonia in rabbits and guinea pigs, and has been associated with the disease atrophic rhinitis in swine [1]. B. bronchiseptica is closely related to Bordetella pertussis, the aetiologic agent of whooping cough [2]. B. bronchiseptica will rarely infect humans, although infections of immunocompromised patients have been reported [3–6]. The ability of Bordetellae to colonise and establish a respiratory tract infection is dependent upon the production of numerous virulence factors including filamentous haemagglutinin, fimbriae, pertactin, adenylate cyclase-haemolysin, dermonecrotic toxin and tracheal cytotoxin [7] which in turn are responsible for adherence [8–10], tissue damage [11] and may also play a role in altering host defence mechanisms [12,13]. Genes encoding the majority of these virulence and colonisation factors are coordinately controlled at the transcriptional level by the Bordetella virulence gene (bvg) locus [14,15]. This locus encodes a two-component regulatory system that enables Bordetella species to respond to a wide range of external stimuli and to alternate between distinct phenotypic phases [16]. The Bvg+ or virulent phase is characterised by the expression of the above-mentioned bvg-activated virulence factors, whilst the Bvg or avirulent phase is characterised by the loss of virulence factor expression and the induction of bvg-repressed genes, which include the genes encoding urease [17,18], flagella biosynthesis [19,20] and acid phosphatase [21]. The Bvg phase has also been shown to be advantageous for survival during times of nutrient deprivation such as those encountered in the external environment between hosts [22]. This has led to the proposal of an environmental reservoir for B. bronchiseptica and has implications for disease transmission and management practices.

In bacteria, the synthesis of aromatic amino acids begins with a common, linear biochemical pathway consisting of seven enzymatic reactions known as the pre-chorismate pathway. The construction of auxotrophic mutants has been a popular method for the generation of rationally attenuated strains exhibiting live vaccine potential. Disruption of the pre-chorismate pathway through mutation of aroA, a gene encoding 5-enolpyruvylshikimate 3-phosphate synthase, has been a common strategy in a number of rational attenuation studies [23]. Although avirulent, these strains are able to stimulate a protective immune response against a wild-type challenge with the parental strain. This interest in live vaccines can be attributed to the superior protection afforded by such vaccines, the type of immunity enlisted and the ease of preparation and administration of vaccine strains [24].

Currently it has not been demonstrated that B. bronchiseptica can be attenuated through disruption of aromatic amino acid biosynthesis. In this report, we describe the cloning and sequencing of the B. bronchiseptica aroA and trpE genes and the construction of a B. bronchiseptica aromatic amino acid auxotroph. We demonstrate that this strain is attenuated in both in vitro studies and a mouse infection model and promotes a significant B. bronchiseptica-specific immune response when used as a live intranasal vaccine strain. Mice vaccinated with B. bronchiseptica aroA trpE are protected from challenge with the wild-type parental strain.

2Materials and methods

2.1Bacterial strains and culture conditions

Wild-type B. bronchiseptica strain K8744/1b was isolated from a pig with a natural infection and was provided by P. Blackall. B. bronchiseptica strains were grown at 37°C on Bordet-Gengou (BG) agar (Difco Laboratories) containing 1% glycerol w/v and 10% v/v defibrinated horse blood. To test the aromatic amino acid phenotype, B. bronchiseptica strains were grown in minimal Stainer Scholte (SS-X) [25] broth or on agar containing 1.5% w/v Noble agar (Sigma). SS-C medium consisted of a modified form of SS-X where 2.5 g l−1 NaCl was replaced with 10 g l−1 Mg2SO4. When added to media, aromatic amino acid supplementation (aamix) consisted of a combination of tyrosine, tryptophan and phenylalanine at a final concentration of 40 μg ml−1, and 2,3-dihydroxybenzoate and p-aminobenzoic acid at 10 μg ml−1. Motility testing agar was prepared as previously described [20]. Escherichia coli strains were routinely grown on Z agar [26] or in Luria Bertani (LB) broth [27]. For transfection with λ phage, E. coli 294 Rif was grown in LB broth containing 0.4% w/v maltose and 10 mM MgCl2. When required, antibiotics were added to the medium at the following concentrations: ampicillin (Ap) 100 μg ml−1; kanamycin (Km) 50 μg ml−1; cephalexin (Cp) 50 μg ml−1. Isopropyl-β-d-galactopyranoside (IPTG) 0.04 mM and 5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside (X-Gal) 40 μg ml−1 were used when required.

2.2DNA manipulations

Restriction endonucleases, T4 DNA ligase and Klenow enzyme were purchased from Roche. Plasmid DNA was extracted by alkaline lysis using the Pharmacia Flexi-prep Kit or the Qiagen Spin-prep system. Standard molecular cloning, transformation and electrophoresis techniques were used [27]. Chromosomal DNA was extracted using a genomic DNA purification kit (Qiagen). Polymerase chain reaction (PCR) amplification of the aroA gene from B. bronchiseptica K8744/1b chromosomal DNA was performed with Pfu polymerase (Stratagene) using two oligonucleotide primers (Bresatec, Australia), one containing a flanking EcoRI restriction site (underlined), 5′-GAATTCATGAGCGGATTGGCATAT-3′ and 5′-CGCGTACACGTCGAAATAAT-3′ which were designed using the B. pertussis aroA gene sequence [28]. Reactions were carried out in a final volume of 50 μl (8 μM MgCl2) using a Hybaid thermocycler for 35 amplifications (95°C, 40 s; 52°C, 30 s; 74°C, 90 s). The resulting 1.3 kb fragment was cloned into pCR-Script (Stratagene) to produce pJMC6. Cosmid libraries were generated by partially digesting B. bronchiseptica K8744/1b chromosomal DNA with Sau3A to give DNA in the size range of 40–50 kb, followed by subsequent ligation into the BamHI site of the cosmid vector pHC79. Packaging of recombinant cosmid DNA into λ phage was performed using Max-Plax™ Lambda Packaging Extracts (Epicentre Technologies) and used to infect E. coli 294 Rif. Colony hybridisation of cosmid clones was performed by transferring bacteria from agar plates onto nylon membranes (Hybond-N; Amersham), which were processed according to the method of [27]. Recombinant cosmids were screened using [α-32P]dATP-labelled, PCR-amplified DNA as a probe. Alternatively, DNA was purified from agarose gels using a Bresa-Clean extraction kit (Bresatec, Australia) then labelled with [α-32P]dATP using a nick translation kit (Gibco-BRL). DNA sequence analysis was performed using a Ready Reaction Dye Terminator Cycle Sequencing kit (Applied Biosystems) and electrophoresed using an ABI Prism 377 DNA sequencer. Sequence data were assembled into contiguous sequences using AutoAssembler software (Applied Biosystems). Homology searches using the BLAST algorithm were run on the Australian National Genomic Information Service (ANGIS) at the University of Sydney.

2.3Cloning of the B. bronchiseptica aroA and trpE genes and construction of a B. bronchiseptica aroA trpE mutant

The 1.3 kb aroA fragment of pJMC6 was cloned as an EcoRI fragment into pGP704Sal forming pJMC9. pGP704Sal is a derivative of the suicide plasmid pGP704 that was digested with BglII and XbaI, Klenow treated, then re-ligated resulting in a plasmid lacking the following restriction sites from its polylinker: BglII, PstI, SalI, BamHI, HindIII and XbaI. When pJMC9 was subsequently digested with SalI and re-ligated, it resulted in a 90 bp deletion within the aroA gene fragment producing pJMC10. A 2.2 kb kanamycin resistance cassette was excised from pHP45ΩKm with EcoRI, Klenow treated and ligated into the now unique SalI site of pJMC10, which had also been Klenow treated, to create pJMC13. After conjugation [29] between S17-1λpir containing pJMC13 and B. bronchiseptica K8744/1b, co-integrates were selected for SS-X agar plates containing aamix, Km and Cp. Aro mutants were identified by their reduced, but not abolished ability to grow on SS-X medium lacking aamix. Southern blot analysis of this Aro transconjugant (CMJ25) demonstrated that aroA had been disrupted through integration of pJMC13 into the chromosome. The labelled 1.3 kb PCR-amplified aroA gene fragment reacted with a 5.2 kb fragment in ClaI-digested chromosomal DNA of K8744/1b. In the mutant strain CMJ25, the probe reacted with a band of 11.2 kb which corresponds to the ClaI fragment containing integrated pJMC13 (data not shown). CMJ25 was subjected to rounds of mutagenesis using the gentamicin-resistant mini-transposon mini-Tn5/Gm[30]. Out of 3000 gentamicin-resistant colonies screened, one was identified as being unable to grow in minimal SS-X medium and was designated CMJ60. To identify the gene disrupted by the insertion of the mini-transposon, a cosmid library of CMJ60 chromosomal DNA was produced and a recombinant cosmid containing the transposon (designated pJCOS60) was selected via gentamicin resistance. Suitable fragments for DNA sequencing containing the gentamicin transposon were identified via Southern blotting using the radiolabelled gentamicin gene as a probe. Sequencing revealed that the transposon had disrupted the trpE gene by inserting at base 901 of the trpE open reading frame (ORF) (GenBank accession number AF266751).

2.4Intracellular survival of bacteria within the mouse macrophage-like cell line J774A.1

B. bronchiseptica strains were tested for their ability to survive in the macrophage-like cell line J774A.1 (ATCC TIB 67). Cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 4.5 g l−1 glucose (Sigma), 10% v/v foetal calf serum (FCS), 5 mM glutamine and 1.5 g l−1 NaHCO3 in an atmosphere containing 5% CO2 at 37°C. J774A.1 cells were seeded at a concentration of approximately 5×104 per well in 24-well tissue culture plates (Inter Med NUNC). Invasion assays were then performed as previously described [31].

2.5Intranasal infection and vaccination of mice

Female BALB/c mice (Animal Resource Centre, Perth, Australia) at ages of 6 weeks were used in these studies. To confirm that animals from each lot were Bordetella free, four individual mice were killed and an aliquot of homogenised lung tissue was plated onto BG agar. For infection and challenge studies, bacteria were grown to mid-log phase in supplemented or unsupplemented SS-X broth, washed and resuspended in the appropriate volume of phosphate-buffered saline (PBS; KH2PO4 2.3 mM, K2HPO4 8.3 mM, NaCl 145 mM pH 7.2). Mice were anaesthetised using an intraperitoneal injection containing xylazine (0.25 mg) and ketamine (1.25 mg) and a 25 μl aliquot of the bacterial suspension containing 105 or 106 cells was administered intranasally. Four mice per time point were killed at various times after infection (day 0). Under aseptic conditions, lungs were removed and homogenised in PBS. Appropriate dilutions were then plated onto supplemented or unsupplemented SS-X agar and colony-forming units determined after incubation for 3 days. At least four mice per treatment were also used to investigate the immune response following vaccination and challenge studies. Blood was collected by cardiac puncture and serum separated. Spleens were collected aseptically and lung washes were collected by pertracheal cannulation and lavaging 0.7 ml of PBS containing 2 mM phenylmethylsulfonyl fluoride. For lymphoproliferation assays spleen cells were adjusted to 5×106 cells ml−1 in complete medium and 50 μl was added in triplicate to a 96-well flat bottom cell culture plate (Interpath) containing 200 μl of complete media alone or antigen (heat-killed K8744/1b cells at 64 ng per well). Lymphoproliferation analyses were then essentially performed as previously described [32].

2.6Immunological techniques

The antibody response to B. bronchiseptica was measured by an enzyme-linked immunosorbent assay (ELISA) using whole B. bronchiseptica K8744/1b cells (in PBS containing 10% v/v methanol) as antigen. Immunoassay plates (96 flat bottom wells, Greiner) were coated with 50 μl of antigen (OD560 of 0.05) by centrifugation at 420×g for 10 min. Unbound antigen was removed and plates were allowed to air dry completely before being blocked with 50 μl of PBS containing 0.05% v/v Tween 20 and 1% w/v bovine serum albumin (Sigma) for 1 h at 37°C. After washing three times with PBS containing 0.05% Tween 20 (PBS-T20), plates were incubated for 2 h at 37°C with serial dilutions of individual mouse serum or lung wash. The plates were again washed, and incubated for 1 h at 37°C with 50 μl of horseradish peroxidase-labelled goat anti-mouse IgG (Kirkegaard and Perry Laboratories) or IgA (Sigma) diluted 1:2000 in PBS-T20. After washing, 50 μl of orthophenylenediamine substrate solution (Sigma) was added to each well and the reaction was allowed to proceed at room temperature for 15 min. The reaction was stopped by the addition of 2 M H2SO4 and plates were read at 490 nm with a SpectraMax 250 microplate spectrophotometer (Molecular Devices). Titres for anti-B. bronchiseptica antibodies were determined using the SoftMax programme (Molecular Devices).

3Results

3.1Nucleotide sequence analysis of the B. bronchiseptica aroA and trpE operons

To clone the area surrounding the aroA gene, a cosmid library of B. bronchiseptica chromosomal DNA was generated and the recombinant clones were screened using radiolabelled, PCR-amplified aroA gene fragment of K8744/1b as a probe. A positive clone containing cosmid pCOS24 was identified. A 4.5 kb ClaI fragment within cosmid pCOS24 was found to contain the aroA gene through Southern blotting analysis (data not shown) and DNA sequence analysis revealed the presence of open reading frames with significant homology to the predicted amino acid sequences of the pheA (chorismate mutase/prephenate dehydratase), tyrA (prephenylate dehydrogenase), aroA (5-enolpyruvylshikimate-3-phosphate synthase), cmk (monophosphate kinase) and rpsA (ribosomal protein S1) (GenBank accession number AF182427). Promoter or terminator sequences were not identified between these genes, which were transcribed in the same direction, suggesting that they are organised into a putative operon. The translational stop signal (bold) of tyrA overlaps with the translational start signal (underlined) of aroAATG A, indicating that these genes may be translationally coupled. This gene arrangement is the same as that seen in the genome sequence of B. bronchiseptica (strain RB50) currently being assembled at the Sanger Centre (http://www.sanger.ac.uk/projects/B_bronchiseptica). From these data the ORFs surrounding this region were determined and their relative positions can be seen in Fig. 1a (unfilled arrows). The predicted amino acid sequences from these ORFs were compared to the non-redundant protein database at the National Centre for Biotechnology Information (NCBI) using the BLASTX algorithm and identity was assigned based on the highest homology. Putative promoter sequences consisting of −35 and −10 regions were found 169 bp and 144 bp upstream of the gyrA start codon respectively and inverted repeat terminator sequences were found 41 bp downstream of the ihfA stop codon, indicating that the full operon containing aroA may include gyrA, serC, pheA, tyrA, aroA, cmk, rpsA and ihfB.

Figure 1.

Physical map of the chromosomal regions containing the (a) aroA gene and (b) trpE genes of B. bronchiseptica. Shaded regions correspond to the DNA sequences determined during this study. The sequences surrounding these regions was obtained from the B. bronchiseptica genome sequencing project at the Sanger Centre (http://www.sanger.ac.uk/Projects/B_bronchiseptica). Putative genes identified based on amino acid sequence homology with other known genes include: ubiG (3-demethyubiquinone-9 3-methyltransferase), ompA (outer membrane protein A), gyrA (DNA gyrase), serC (3-phosphoserine aminotransferase), pheA (chorismate mutase/prephenate dehydratase), tyrA (prephenylate dehydrogenase), aroA (5-enolpyruvylshikimate-3-phosphate synthase), cmk (monophosphate kinase), rpsA (ribosomal protein S1), ihfB (integration host factor B), ubiF (2-octaprenyl-3-methyl-6-methoxy-1,4-benoquinol hydroxylase), mltA (murein hydrolase A), apaG (conserved unknown function), rpe (ribulose phosphate 3-epimerase), pgp (2-phosphoglycolate phosphatase), trpE (anthranilate synthase a subunit), trpG (para-aminobenzoate synthase glutamine amidotransferase component II), trpD (phosphoribosyltransferase), trpC (indoleglycerol phosphate synthase), ygfF (ribitol dehydrogenase), and phnD (2-hydroxychromene-2-coboxylate dehydrogenase). Scale is given in kb.

Analysis of the region surrounding the trpE gene was again undertaken by examining the genome sequence of B. bronchiseptica (RB50) currently being assembled at the Sanger Centre (http://www.sanger.ac.uk). The putative identity of the ORFs was determined through deduced amino acid sequence homology (Fig. 1b). Upstream of the trpE start codon, putative promoter sequences consisting of −35 (235 bp upstream) and −10 regions (211 bp upstream) were identified as well as a potential trpL leader peptide that begins 190 bp upstream of the trpE start codon and encodes 15 amino acids which includes a methionine start codon and four tryptophan codons which may act as a possible attenuation mechanism for the trp operon. Included in the operon with trpE are genes that code for other tryptophan biosynthetic enzymes, trpG, trpD and trpC. Potential inverted repeat terminator sequences are located 41 bp downstream of the trpC stop codon.

3.2Phenotypic analysis of B. bronchiseptica aroA trpE

Growth of the B. bronchiseptica aroA trpE mutant strain CMJ60 was examined over time and found to be unable to replicate in SS-X broth containing Km and Gm, in the absence of aamix supplementation. However in the presence of aamix in the growth medium, the growth rate of CMJ60 was similar to that of the wild-type strain (Fig. 2).

Figure 2.

Growth of the parental B. bronchiseptica strain K8744/1b compared to mutant CMJ60. K8744/1b (triangles) and mutant CMJ60 (squares) were cultured in SS-X broth at 37°C (unfilled triangles or squares) or in SS-X broth supplemented with aromatic amino acids at 37°C (filled triangles or squares). CMJ60 was cultured in the presence of Km and Gm respectively.

In Bordetella, the expression of the majority of virulence and colonisation factors is coordinately controlled at the transcriptional level by the bvg locus. To ensure that strains K8744/1b and CMJ60 had not undergone phase variation (a permanent change in phenotype that is due to small deletions in the bvg locus), they were tested for the ability to undergo phenotypic modulation. Both strains produced a zone of haemolysis on BG agar supplemented with aromatic amino acids at 37°C, indicating that they were expressing the bvg-activated, adenylate cyclase/haemolysin under these conditions. Motility assays with K8744/1b and CMJ60 showed a non-motile phenotype when grown in supplemented SS-X at 37°C, as expected. When grown under modulating conditions in SS-C minimal medium at 20°C, both strains displayed a motile phenotype (data not shown). These data suggest that both strains were capable of phenotypic modulation and had not undergone genetic mutations in the bvg locus.

3.3Level of attenuation of B. bronchiseptica aroA trpE

The ability of the parental wild-type strain K8744/1b and the aroA trpE strain CMJ60 to survive within the macrophage-like cell line J774A.1 was investigated. Compared to the parental strain, CMJ60 displayed significantly reduced abilities to invade and survive within the mouse macrophage-like cell line J774A.1 after both 2 h (P<0.05) and 24 h (P<0.01) infection (Fig. 3). The level of attenuation of CMJ60 following experimental intranasal infection of mice was also examined. Mice were infected with a sub-lethal dose of either K8744/1b or CMJ60 and the number of viable bacteria present in the lungs was determined at different time intervals (Fig. 4a). Compared to K8744/1b, the mutant strain CMJ60 was significantly attenuated in its ability to survive in the murine respiratory tract. CMJ60 failed to replicate in the murine respiratory tract and was cleared more rapidly than the wild-type strain K8744/1b. Total clearance from the lung was achieved by day 35 for CMJ60 and day 45 for K8744/1b. To examine the in vivo stability of the mutations in CMJ60, all bacterial colonies isolated from CMJ60-infected mice on day 16 were examined and shown to be phenotypically identical to the inocula (i.e. haemolytic, auxotrophic, Kmr, Gmr and Cpr; data not shown) suggesting 100% in vivo stability of the vaccine strain for at least 16 days post-immunisation.

Figure 3.

Intracellular survival of the parental B. bronchiseptica strain K8744/1b (filled) and the aroA trpE mutant CMJ60 (unfilled) at 2 and 24 h in the macrophage-like cell line J774A.1. Results represent the mean colony-forming units per well of triplicate wells for the respective strains. Error bars represent standard deviations. Variation between K8744/1b and CMJ60 was considered significant at P<0.05 (*).

Figure 4.

Clearance from the murine lung of the parental B. bronchiseptica strain K8744/1b (filled squares) and the aroA trpE mutant CMJ60 (unfilled triangles) following intranasal infection of BALB/c mice (a). Results represent the mean colony-forming units per lung of four mice per time point for the respective strains. Error bars represent standard deviations. Variation between K8744/1b and CMJ60 was considered significant at P<0.05 on days 2, 4 and 9. B. bronchiseptica-specific antibody responses in the serum of mice infected with K8744/1b (b) or CMJ60 (c). Graphs represent specific IgG (solid bar), IgM (dotted bar) and IgA (cross hatched bar) anti-B. bronchiseptica titres determined by ELISA. Results represent the titres of pooled sera for four mice per time point.

The serum obtained during the course of the murine B. bronchiseptica K8744/1b and CMJ60 infection was assayed for specific antibodies by ELISA against whole cell B. bronchiseptica K8744/1b as antigen. For the wild-type infection, specific IgM was found to steadily increase from day 2 until day 18, after which titres started to decline. Specific IgG was not seen until day 9 with titres increasing up to day 26 where they leveled out. Specific IgA antibodies were only detected at low levels, late in the course of infection (Fig. 4b). However, for the CMJ60 infection specific IgM only increased slightly up until day 10 at which point titres started to decline. Only slight increases in specific IgG titres were seen from day 16 to day 35 at which point levels started to decline. A small increase in specific IgA was seen on day 35 but was again undetectable by day 45 (Fig. 4c).

3.4Immunological response following booster vaccinations with B. bronchiseptica aroA trpE

The anti-B. bronchiseptica immunological response of mice was examined in non-immunised mice, in mice intranasally immunised with 1×105 CMJ60 viable cells on day 0, in mice intranasally immunised and boosted with 1×105 CMJ60 viable cells on day 0 and day 14, and in mice immunised and boosted with 1×105 CMJ60 viable cells on day 0, day 14 and day 28. Non-immunised mice did not exhibit significant anti-B. bronchiseptica immune responses. In contrast 10 days after the last booster immunisation, the serum anti-B. bronchiseptica IgG titre increased by over 10-fold following a single boost at day 14 (P<0.05) and over 100-fold following booster immunisations at days 14 and 28 (P<0.05). A moderate increase in serum IgA levels was also observed (Fig. 5a). The anti-B. bronchiseptica IgG titre was significantly increased in lung lavages following a single immunisation (P<0.05). Anti-B. bronchiseptica IgG and IgA titres were significantly increased in lung lavages following two (P<0.05) or three doses (P<0.05) (Fig. 5b). Compared with non-immunised mice, vaccinated animals exhibited a significant increase (P<0.05) in the number of B. bronchiseptica-reactive spleen cells in lymphoproliferation assays only after the second boost (Fig. 5c).

Figure 5.

Levels of B. bronchiseptica-specific IgG (filled bars) and IgA (unfilled bars) in mouse serum (a) and lung lavage (b) samples in non-immunised mice (zero doses), mice intranasally immunised with CMJ60 (one dose), mice intranasally immunised then boosted with CMJ60 on day 0 and day 14 (two doses), and mice immunised then boosted with CMJ60 on day 0, day 14 and day 28 (three doses). Results represent the titres of four mice per treatment. Error bars represent standard deviations. Variation between non-immunised and vaccinated mice was considered significant at P<0.05 (*). c: Levels of B. bronchiseptica-reactive spleen cells in non-immunised mice (filled bars), mice intranasally immunised with CMJ60 (unfilled bar, one dose) mice intranasally immunised then boosted with CMJ60 on day 0 and day 14 (unfilled bar, two doses), and mice immunised then boosted with CMJ60 on day 0, day 14 and day 28 (unfilled bar, three doses). Error bars represent standard deviations. Variation between non-immunised and vaccinated mice was considered significant at P<0.05 (*).

3.5Protection afforded to mice following vaccination with B. bronchiseptica aroA trpE

The level of protection afforded following intranasal immunisation and boosting of mice with 1×105 CMJ60 viable cells on day 0, day 14 and day 28 was examined. Non-immunised and vaccinated mice were challenged with 1×106 wild-type K8744/1b viable cells on day 38. Clearance of the challenge strain from the lungs of mice was monitored by determining the number of viable bacteria present in the lungs at various time intervals. Only prototrophic B. bronchiseptica colonies were isolated post-challenge, screened for the ability to grow without aromatic amino acid supplementation on minimal SS-X medium. This thereby enabled isolation and quantification of challenge bacteria (K8744/1b) only, even in the presence of the vaccine strain. Compared with non-vaccinated mice, a significant reduction in the ability of the wild-type K8744/1b strain to colonise the lungs of CMJ60 vaccinated mice was observed. In non-vaccinated mice, the wild-type infection showed a typical profile with bacterial numbers increasing for 4 days after initial inoculation, then being slowly cleared from the lung over the next 23 days. The wild-type infection of mice vaccinated with CMJ60 failed to replicate within the lung and was rapidly cleared to low levels 8 days after initial inoculation (Fig. 6).

Figure 6.

Clearance from the murine lung of the parental B. bronchiseptica strain K8744/1b in non-immunised mice (unfilled diamonds) and mice immunised then boosted with CMJ60 on day 0, day 14 and day 28 (filled squares). Results represent the mean colony-forming units per lung of four mice per time point. Error bars represent standard deviations. Variation between non-immunised and vaccinated mice was considered significant at P<0.05 4 h post-challenge and on days 2 and 4.

4Discussion

This paper describes the cloning and sequencing of the aroA and trpE genes as well as the construction of an aromatic amino acid auxotrophic mutant strain of B. bronchiseptica. The aroA gene of B. bronchiseptica is situated in close proximity and in the same orientation as a number of other genes which most likely form part of an operon in the order of gyrA, serC, pheA, tyrA, aroA, cmk, rpsA and ihfB. A very similar gene organisation is seen in Pseudomonas[33]. This mixed function supraoperon contains the genes SerC-(pdxF)-aroQ·pheA-HisHb-tyrAc-aroF-cmk-rpsA and this linkage has been shown to be well conserved through wide phylogenetic distances. The genes pheA and tyrA encode proteins with a role in aromatic amino acid biosynthesis and are involved in the conversion of chorismate to phenylalanine and tyrosine respectively. The pheA and tyrA genes in Lactococcus lactis are organised into an operon with aroA[34]. As the translational stop signal of tyrA overlaps the translational start signal of aroA, this may indicate that these genes are translationally coupled. In other organisms, genes involved in aromatic amino acid biosynthesis that are translationally coupled include trpB with trpA[35] and trpE with trpD[36]. In such cases it was shown that efficient translation of the gene directly downstream is dependent on the translation of the end of the gene upstream [37]. This may have implications for the regulation of aroA in B. bronchiseptica as it would therefore be directly affected by the mechanisms controlling the transcription and translation of tyrA. Downstream from aroA, the gene cmk encodes cytidine monophosphate kinase that is involved in the conversion of CMP and dCMP to CTP and dCTP respectively, whereas rpsA encodes the ribosomal protein S1. The genes flanking the B. bronchiseptica trpE gene were found in a DNA fragment with typical tryptophan biosynthetic gene arrangement containing trpG, trpD and trpC. The presence of a putative trpL leader sequence upstream of trpE may indicate that B. bronchiseptica uses an attenuation mechanism to regulate expression of this operon.

The B. bronchiseptica aroA trpE mutant constructed in this study was found to be unable to grow in the absence of aromatic amino acid supplementation. The mutant was attenuated both in the ability to invade/survive in the macrophage-like cell line J774A.1 and in the murine lung with the attenuating mutations introduced being genetically stable for at least 16 days in vivo. As the strain was unable to replicate in the lungs of mice, it was cleared at a faster rate than that of the wild-type and failed to elicit strong humoral responses. However, upon administration of boosting doses this attenuated strain promoted the elicitation of efficient anti-B. bronchiseptica cellular and humoral immune responses, which were able to protect mice against challenge with a virulent strain.

Natural B. bronchiseptica infections are characterised by efficient establishment, long-term persistence and the absence of chronic disease. It has recently been suggested that this is achieved via modulation of host immunity. Mechanisms employed by B. bronchiseptica to modulate host immunity have been elaborated [13,38] and it is believed that these mechanisms and others yet to be characterised, facilitate persistent infection in the respiratory tract by B. bronchiseptica during the infection process. Gueirard and Guiso [39] have shown that the systemic antibody response induced by wild-type B. bronchiseptica infection is mainly composed of IgG2a immunoglobulins and is largely non-specific. These authors have suggested that B. bronchiseptica infection primarily induces a Th1-type T-cell response. Yuk et al. [13] have shown that type III secreted proteins produced by B. bronchiseptica can down-regulate the humoral response in the host. However, administering booster immunisations of the vaccine constructed in this study may have allowed us to overcome the suppression of the adaptive response and evidence for this is seen through the induction of a serum IgG response and local immunity in the respiratory tract in the form of a secretory IgA response.

Previously, other researchers have shown that killed B. bronchiseptica whole cell vaccines produced from either dog or pig isolates demonstrate cross-protection in a murine model [40]. Shimizu and Ishikawa [41] intranasally administered an undefined attenuated B. bronchiseptica mutant that was protective in a Guinea pig model of infection. Bey et al. [42] vaccinated dogs by the intranasal route with a live avirulent B. bronchiseptica vaccine strain, and correlated rises in serum IgG and salivary IgA levels with the development of resistance to infection. Finally, Gueirard and Guiso [39] demonstrated that purified adenylate cyclase protected vaccinated mice against virulent challenge. However, killed whole cell vaccines do not necessarily engender immunity at the site of infection, avirulent mutant strains will not express important antigenic virulence factors, undefined attenuated vaccines may revert to virulence whilst purified component vaccines will ultimately prove too costly for widespread use in veterinary situations. Therefore, the generation of defined attenuated B. bronchiseptica vaccine strains, such as the prototype presented here, represent a significant advance towards the development of a new generation vaccine against this pathogen. The resulting strains may be utilised for both the prevention of B. bronchiseptica-associated disease and the delivery of heterologous antigens to a wide range of mammalian species.

Acknowledgements

J.D.M. is the recipient of an Australian Postgraduate Research Award (industry). B. bronchiseptica K8744/1b was kindly provided by P. Blackall (Queensland Department of Primary Industries). This work was in part supported by the BMBF grant ‘PathoGenoMik’ project ‘Vergleichende funktionelle Genomanalyse von Bordetella-Arten’ (601III4-1/1VIIZV-21/AG Guzman) to C.A.G.

Ancillary