TEXT

Top Abstract Text References

The attenuation of bacterial pathogens by auxotrophic mutations was first demonstrated by Bacon et al. with "Bacterium typhosa" 50 years ago (1). Over 2 decades after these original observations, Hoiseth and Stocker (13) showed that a Salmonella enterica serovar Tyhpimirium aroA (5-enolpyruvylshikimate 3-phosphate [EPSP] synthase) mutant was both attenuated and an excellent live vaccine in the mouse typhoid model. This enzyme is part of the aro pathway, which leads, through shikimic acid, to chorismic acid. Chorismate is a branching point from which separate pathways lead to the aromatic amino acids, to para-aminobenzoic acid and hence folic acid, to vitamin K, to ubiquinone and hence the electron transport systems, and to dihydroxybenzoic acid, which is the first step in the biosynthesis of the siderophore enterochelin (18, 25). The shikimate pathway occurs in prokaryotes, yeasts and filamentous fungi, apicomplexan parasites (16), and the plastids of plants and algae. The aro pathway is not, however, present in vertebrates, meaning that these animals must obtain the essential products derived from chorismic acid from their diet and that the intermediates of this pathway are not available to complement the requirements of an auxotroph. Mutations in the aroC gene (7), encoding chorismate synthase, the final enzyme in this pathway, which catalyzes the conversion of 5-enolpyruvylshikimate 3-phosphate (EPSP) to chorismic acid, and also in aroD and aroB are equally attenuating, confirming the key role of the aro pathway for the virulence of Salmonella (11).

Brucella spp. are gram-negative bacteria responsible of animal brucellosis in a variety of mammalian hosts. A major characteristic of this intracellular pathogen is its ability to survive and replicate in the macrophages of the host, where it remains enclosed in phagocytic compartments. Little is known about the genes implicated in the virulence of Brucella. In a previous study, we used signature-tagged mutagenesis (STM) to identify genes required for the intracellular growth and survival of Brucella in a macrophage infection model (10). Several mutants which multiplied poorly or not at all in THP1 (human macrophage-like cells) and HeLa cells were isolated. In one of them, the transposon interrupted a gene that was highly similar to known aroC genes. In the present work, we characterized this gene and the consequence of its mutation on the virulence of Brucella suis.

Cloning and sequencing of the B. suis aroC gene. The genomic DNA from an aroC::Tn5 mutant identified by STM was extracted, and a 3.5-kb EcoRV genomic fragment containing the mini-Tn5 Km2 transposon was cloned in pUC18-SmaI-BAP (Amersham Pharmacia) to transform Escherichia coli DH5. Sequences flanking the transposon were determined using transposon primers P6 and P7 (12) and the direct and reverse primers of pUC18, showing that the transposon was inserted between bp 201 and 215 of the coding sequence and that it had created a deletion of the 14 bases. A digoxigenin-labeled probe was generated by PCR using primers P7 and M13 direct as described previously (10). This allowed us to identify a clone from a B. suis cosmid library in pSuperCos containing genomic inserts of approximately 45 kb (15). The complete sequence of aroC was obtained directly by sequencing the cosmid DNA by primer walking. The aroC gene sequence is 1,038 bp long with a G+C content of 61%. Approximately 500 bp of upstream sequence and 600 bp of downstream sequence were determined. No consensus sequence for Brucella promoters has been described; however a putative ribosome-binding site is located 6 bp upstream from the initiation ATG codon in the aroC gene. There are no detectable open reading frames (ORFs) in the 500 bp upstream. A possible stem-loop is found immediately downstream of the stop codon and an ORF encoding a putative riboflavin biosynthesis protein was detected 150 bp downstream, suggesting that in Brucella aroC is not part of an operon (data not shown). The sequence encodes a single protein of 345 amino acids with an estimated molecular mass of 36.6 kDa. The deduced amino acid sequence has considerable homology with chorismate synthase sequences from other bacteria (55% identity and 79% homology with AroC of both E. coli and S. enterica serovar Typhimurium) (5).

The B. suis aroC gene complements Salmonella and E. coli aroC mutants. A 1,162-bp fragment containing the complete aroC gene and 60 bp upstream and downstream was amplified by PCR using primers Aro1 (5' GGC CGG TAA AAG AAA CTG GT 3') and Aro2 (5' ATT ATT TTC AGG CGC GGC CA 3') and cloned in the pGEMT-Easy vector (Promega). An ApaI-SacI fragment containing the gene was subcloned into pBBR1MCS (17), a low-copy-number, broad-host-range vector which replicates in Brucella. These two plasmids containing the B. suis aroC gene were used to transform an S. enterica serovar Typhimurium LT2 aroC strain and an E. coli BRD049 aroC strain by electroporation. Transformants were assayed for growth on M9 minimal medium (27) and M9 supplemented with para-aminobenzoic acid, dihydroxybenzoic acid (100 µg/ml), phenylalanine, tryptophan, and tyrosine (40 µg/ml). The B. suis aroC gene, either on a high- (pGEM) or low-copy-number (pBBR1MCS) vector restored the ability of both Salmonella and E. coli aroC mutants to grow on minimal medium. This shows that the Brucella gene has the same function as the S. enterica serovar Typhimurium and E. coli genes and that it is aroC. We cannot say whether expression was driven by the natural promoter or by a plasmid promoter.

B. suis aroC mutant multiplies slowly in infected cells. The B. suis aroC mutant was identified as attenuated in a STM screen. We monitored the growth of the mutant in both human macrophages and HeLa cells. As shown in Fig. 1, the aroC mutant enters HeLa cells and THP1 macrophages at levels similar to those of the wild type but over the next 48 h multiplies very slowly in the cells, reaching levels of between 1.5 and 2 log units less than those of the wild type. The presence of pBBR1MCS-aroC restored the capacity of the mutant to multiply in infected cells.

View larger version (25K): [in this window] [in a new window]   FIG. 1.   The B. suis aroC mutant is attenuated in cell infection models. (a) HeLa cells. (b) THP1 macrophages. Cells were infected as previously described with multiplicities of infection of 100 for HeLa cells and 20 for THP1 cells (10). Data points are the means of three wells with the standard deviations. These data sets are from one of three independent experiments with similar results.

B. suis aroC mutant is attenuated in BALB/c mice. A mouse model allowed us to confirm the participation of aroC in the virulence of Brucella. Groups of BALB/c mice were infected intraperitoneally (i.p.) with 5 × 10⁵ wild-type, aroC mutant, or complemented aroC mutant B. suis cells. The bacterial counts in spleens and the weights of spleens were determined at 1, 7, 35, and 56 days postinfection. A multivariate statistical analysis and a profile analysis using PROC GLM (1996 version; Statistical Analysis System Institute Inc.) yielded significant differences (P < 0.006) for pairwise comparisons of individual profiles. The wild-type strain multiplied over 1,000-fold over the first week of infection and then persisted at very high levels with induction of splenomegaly (Table 1) for the duration of the experiment (8 weeks). The aroC mutant colonized spleens and appeared to multiply very slowly over the first week postinfection. After the first week, the spleens of the mice became enlarged and the mice slowly eliminated the bacteria. At 5 weeks postinfection, there was large mouse-to-mouse variation, with some animals still colonized and others clear, but at 8 weeks the spleens of all four mice were clear (limit of detection, 1 CFU) of Brucella. The complemented mutant multiplied as rapidly as the wild type during the first week but then was unable to maintain the high levels of colonization of the wild type at later points (here again there was large variation between animals). It is not clear why complementation was not complete; the plasmid was stable since all the bacteria recovered from the spleens were chloramphenicol resistant, and the DNA sequence argues against a polar effect on a downstream gene. A possibility is that expression of aroC was not optimal in vivo.

View this table: [in this window] [in a new window]   TABLE 1.   Spleen weights of mice infected i.p. with wild type B. suis S1, the B. suis aroC strain, and the B. suis aroC complemented strain

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Kinetic survival in BALB/c mice of wild-type B. suis versus that of the aroC mutant and aroC complemented strain. Mice were infected i.p. with a dose of 5 × 105 CFU. At different times postinfection the bacterial load was determined by plating appropriate dilutions of homogenized spleens on agar. Each point represents the geometric mean with standard deviations for four mice.

Nucleotide sequence accession number. The aroC sequence obtained in this study has been assigned GenBank accession no. AF276655.