Induction of dnaK through Its Native Heat Shock Promoter Is Necessary for Intramacrophagic Replication of Brucella suis

Abstract

The heat shock protein DnaK is essential for intramacrophagic replication of Brucella suis. The replacement of the stress-inducible, native dnaK promoter of B. suis by the promoter of the constitutively expressed bla gene resulted in temperature-independent synthesis of DnaK. In contrast to a dnaK null mutant, this strain grew at 37°C, with a thermal cutoff at 39°C. However, the constitutive dnaK mutant, which showed high sensitivity to H₂O₂-mediated stress, failed to multiply in murine macrophage-like cells and was rapidly eliminated in a mouse model of infection, adding strong arguments to our hypothesis that stress-mediated and heat shock promoter-dependent induction of dnaK is a crucial event in the intracellular replication of B. suis.

We have described the important role of the heat shock protein and molecular chaperone DnaK in intramacrophagic growth of Brucella suis and its acid-induced expression (12). Brucella spp., facultative intracellular, gram-negative bacteria which are the causative agents of brucellosis in humans and animals (6), resist intracellular killing and replicate within the phagosome of macrophages (3, 11, 16). It has been observed recently that this phagosome is acidic (pH 4.0 to 4.5), suggesting a stressful environment, and that early acidification is essential for intracellular multiplication of B. suis (19). Work on the identification of proteins induced under stress conditions in Brucella abortus and Brucella melitensis confirmed our results (17, 20, 23). From previous data (12), we concluded that DnaK from B. suis may play an essential role as part of protein repair systems in protecting the bacteria from the environment encountered in the phagosome. Another hypothesis is that DnaK may be directly involved in folding and proper localization of virulence factors, as intracellular multiplication is abolished in the null mutant.

The earlier work was done with a dnaK null mutant, allowing us to conclude only that the chaperone participated in intracellular multiplication of B. suis (12). On the other hand, dnaJ, which is located downstream of dnaK and forms an operon with the latter, is not involved in resistance to acid stress and intracellular multiplication of the pathogen (12). As a dnaJ knockout mutant of B. suis behaves like the wild-type strain with respect to these properties, we concluded that dnaJ was of no relevance in the context of the work presented here. Western blot analysis showed induction of dnaK during heat and acid shock and under intramacrophagic growth conditions (12). We therefore hypothesized that low-level, constitutive expression of dnaK was not sufficient for B. suis to resist macrophage attack. To verify this hypothesis, we replaced the original dnaK promoter on the B. suis chromosome by the constitutive promoter of the β-lactamase gene bla_TEM (P_bla) from pUC18, and we studied the phenotype of the mutant obtained. P_bla was chosen, as we and others (13, 14) have observed that bla is expressed in brucellae, conferring ampicillin resistance to strains bearing this gene on a plasmid. Furthermore, bla and its promoter region are well characterized, in contrast to promoters of constitutively expressed Brucella genes.

Replacement of the native heat shock promoter of dnaK by the β-lactamase promoter

To perform promoter exchange, three distinct sets of cloning steps were performed in Escherichia coli DH5α (Life Technologies, Cergy Pontoise, France). (i) A 580-bp DNA fragment containing the promoter region and the sequence coding for the N-terminal 36 amino acids of the bla protein was isolated from pUC18 (24) following digestion by Sau3AI, and cloned into the promoter selection vector pKK232-8 (Amersham Pharmacia Biotech, Orsay, France). Activation of the reporter gene cat, leading to chloramphenicol resistance, determined the orientation of the bla promoter. We then inserted a blunted kanamycin resistance gene isolated from pUC4K (Amersham Pharmacia Biotech, Orsay, France) in the SmaI site upstream of the bla promoter and in the opposite direction of transcription. (ii) Based on the sequence from B. ovis described earlier (4), a 1-kb DNA fragment containing 330 bp upstream of the start codon of dnaK from B. suis and the first 670 bp of the same gene was obtained by PCR with the primers 5′-GCGGGGTGAAAATGTGCCGC-3′ (positions 108 to 127), and 5′-ACCGCCAAGGAACGTGTCGC-3′ (1113 to 1094) and cloned blunt-ended into pUC18 (Amersham Pharmacia Biotech). This DNA was then digested by the restriction enzymes AgeI and NdeI, resulting in the excision of a 60-bp fragment from positions 330 to 390 of the published sequence carrying the promoter region of the dnaK/dnaJ operon (4, 21). (iii) In a final step, the AgeI-NdeI gap was blunted and filled with a 1.8-kb DNA fragment containing the kanamycin resistance gene and the bla promoter from step 1 in the proper orientation, i.e., the bla promoter supporting transcription of dnaK. For allelic exchange between this modified 5′ region of dnaK and the wild-type gene of B. suis, we recloned the final construct from pUC18 into the brucella suicide vector pCVD442 and performed electroporation of B. suis 1330 and selection for homologous recombination on sucrose-containing tryptic soy (TS) agar plates as described previously (7). Southern blot analysis (22) of the resulting dnaK_mut mutant with a dnaK probe revealed the correct insertion of the Kan^r-P_bla sequence into the expected site of the chromosomal dnaK gene (data not shown).

Constitutive, temperature-independent expression of dnaK and growth of B. suis at various temperatures

The expression of dnaK in the mutant strain at 30, 37, and 46°C compared to that in wild-type B. suis 1330 was analyzed by Western blotting with monoclonal anti-DnaK antibody as described previously (5, 12). Briefly, strains were grown to an optical density at 600 nm (OD₆₀₀) of 0.5 at 30°C, concentrated fourfold in preheated TS broth, and incubated for 30 min at the appropriate temperatures. Bacteria were immediately heat killed at 65°C for 45 min, washed once in phosphate-buffered saline (PBS), and resuspended in sodium dodecyl sulfate-polyacrylamide gel electrophoresis sample buffer. Total cell lysates equivalent to 5 × 10⁸ bacteria per lane were loaded onto a sodium dodecyl sulfate-12.5% polyacrylamide gel. Total proteins were visualized by Coomassie staining after transfer onto a PolyScreen membrane (NEN) prior to antibody incubation and detection of DnaK by enhanced chemiluminescence (Amersham). As determined with SigmaGel software (SPSS Science, Chicago, Ill.), B. suis DnaK was induced approximately 1.7-fold between 30 and 37°C and another 2-fold between 37 and 46°C in the wild type, whereas the amounts of DnaK produced by the mutant containing the bla promoter remained constant (Fig. 1). The dnaK null mutant, obtained by gene knockout (12), did not produce any DnaK (Fig. 1A, lane 7). This result provided evidence that promoter replacement led to the noninducible expression of dnaK in this B. suis mutant.

FIG. 1.

(A) Western blot analysis of dnaK expression in total cell lysates of wild-type B. suis (lanes 1, 3, and 5) and the dnaK_mut strain (lanes 2, 4, and 6) with monoclonal anti-DnaK antibody. Growth temperatures were 30°C (lanes 1, 2, and 7), 37°C (lanes 3 and 4), and 46°C (lanes 5 and 6). The dnaK null mutant was used as a control (lane 7). (B) Quantification of DnaK in panel A, using SigmaGel software, relative to the amount of DnaK present in the wild-type lysate at 30°C (lane 1).

The absence of DnaK in the null mutant of B. suis described earlier resulted in temperature sensitivity for growth at 37°C and above (12). We compared this observation to the behavior of the dnaK_mut strain at different temperatures: As we observed that the strain grew well on agar plates at 37 but not at 42°C (data not shown), we performed growth curve experiments in TS broth at 37, 38, and 39°C. B. suis 1330 (wild type) and the dnaK_mut strain were grown to stationary phase, diluted to an OD₆₀₀ of 0.03, and grown at the appropriate temperatures for 56 h. We observed lower growth rates for the mutant than for the wild-type at 37 and 38°C. At both temperatures, wild-type B. suis entered stationary phase at about 15 h postdilution, whereas the dnaK_mut strain reached a similar OD₆₀₀ only at 56 h (Fig. 2). The dnaK null mutant (12), in contrast, did not show any growth at 37°C. The temperature limit for normal growth of B. suis dnaK_mut was 39°C. Constitutive, low-level production of DnaK therefore appeared to be sufficient for growth at temperatures of up to 38°C. In contrast, heat stress conditions such as growth at 39°C and above necessitate higher-level synthesis of DnaK, which is not possible in the dnaK_mut strain (Fig. 1, lane 6).

FIG. 2.

Growth rates in TS broth of the B. suis wild-type strain at 37°C (•) and 39°C (▪) and of the dnaK_mut strain at 37°C (○), 38°C (▵), and 39°C (□). The absence of growth of the dnaK null mutant at 37°C is indicated (▾). Growth curves are from one representative experiment out of three performed under identical conditions.

Sensitivity of the dnaK_mut strain and of the dnaK null mutant to acid pH and oxidative stress

Additional phenotypic characteristics that might be linked to a modification of dnaK expression, such as sensitivity to acid pH and to oxidative stress, were investigated according to previously published protocols (8, 12). Briefly, for pH sensitivity experiments B. suis 1330 and the dnaK_mut strain were grown to stationary phase (OD₆₀₀ = 1.5) in TS broth, and the number of CFU per milliliter was determined prior to dilution 1:50 in TS broth adjusted to pH 4.5. At 0, 6, 24, 48, and 54 h postdilution, culture samples were diluted appropriately and plated on TS agar for colony enumeration. In contrast to the very high sensitivity of the dnaK null mutant to acid pH, as expressed by a 5,000-fold reduction in viable bacteria after 24 h of incubation (12), viability of the dnaK_mut strain decreased only 1.8-fold over the same period compared to the wild-type and 5-fold following incubation for 54 h. The mutant constitutively expressing low levels of DnaK therefore resisted acid pH significantly better than the null mutant but did not reach the wild-type level of resistance. We concluded from these data that an acid pH stress could not by itself account for the rapid elimination of the dnaK_mut strain in the macrophage.

To investigate the sensitivity of the different strains to oxidative killing, strains were grown to stationary phase and rediluted to an OD₆₀₀ of 0.25 each. Each strain (75 μl) was plated on TS agar, and sterile paper disks saturated with 10 μl of H₂O₂ at concentrations of 1.8, 2.9, and 4.4 M were layered on top prior to incubation at 30°C for 2 days and measurement of inhibition zone diameters (8). Interestingly, both the dnaK null mutant and the dnaK_mut strain were significantly more sensitive to the three concentrations of H₂O₂ than the parental strain (Fig. 3) (P < 0.001). The sensitivities of both mutants were comparable, with inhibition zone diameters that were not significantly different from one another. The results therefore showed (i) that DnaK participated in the resistance of B. suis to oxidative stress and (ii) that noninducible dnaK expression was not sufficient for protection from oxidative stress.

FIG. 3.

Sensitivities of wild-type B. suis 1330 (open bars), the dnaK_mut strain (hatched bars), and the dnaK null mutant (solid bars) to various concentrations of H₂O₂. Six disk assays were performed per strain per concentration. Error bars indicate standard deviations.

The dnaK_mut strain is rapidly eliminated in J774 murine macrophages and in BALB/c mice

In order to assess the capacity of the dnaK_mut strain of B. suis to survive within the host cell, we infected J774 murine macrophage-like cells with the mutant and the wild-type strain. Bacteria were added to adherent cells at a multiplicity of infection of 20 and incubated for 30 min at 37°C in the presence of 5% CO₂, followed by three washes with PBS and incubation of the cells in RPMI medium supplemented with 10% fetal calf serum and gentamicin at a concentration of 30 μg/ml for at least 1 h. At 1.5, 7, 24, and 48 h postinfection, cells were washed once with PBS and lysed in 0.2% Triton X-100. CFU were enumerated by plating of serial dilutions on TS agar plates and incubation at 37°C. Low-level, noninducible expression of dnaK resulted in rapid elimination of the mutant strain in J774 cells, whereas the wild-type strain showed a profile typical of intracellular multiplication (Fig. 4). This is in contrast to the mutant's extracellular growth capacities at 37°C. These results therefore made it clear that production of DnaK necessary and sufficient for B. suis growth at 37°C in culture medium is not sufficient for intracellular replication of the pathogen, and they added support for our previous hypothesis that induction of dnaK is crucial for intramacrophagic replication of B. suis (12). The rapid elimination of B. suis dnaK_mut in macrophages could not easily be explained by an eventual sensitivity of the strain to low pH, encountered in the phagosome at least during the first hours of infection, as there was no obvious correlation between its 40-fold reduction during the first 24 h of infection and the only 2-fold loss in viability at acid pH in vitro during the same period. The high sensitivity of the dnaK_mut strain to oxidative stress, in contrast, suggested that this type of stress may participate in the antimicrobial defense mechanisms of murine macrophages against brucellae.

FIG. 4.

Intracellular growth of B. suis in J774 murine macrophage-like cells. Adherent cells were infected with the wild-type strain (•) or the dnaK_mut strain (▴). Experiments were performed in triplicate. Error bars indicate standard deviations.

With the well-established murine model of infection (9, 10), the fate of the bacteria was monitored in vivo by the enumeration of B. suis 1330 and the dnaK_mut strain in the spleens of BALB/c mice at various times postinfection. An infectious dose of 5 × 10⁴ bacteria was injected intraperitoneally, and residual virulence of the strains was evaluated following killing of four mice per point and strain at 1 day and 1, 4, 5, and 8 weeks postinoculation, homogenization of the spleens, and plating of serial dilutions on TS agar plates. Virulent B. suis multiplied rapidly in the spleen, and numbers reached a maximum at 1 week postinoculation before slightly declining until the end of the experiment (Fig. 5). In contrast, the dnaK_mut strain was very quickly eliminated from the spleen, and only a few bacteria were recovered at 1 week in only one mouse out of four (Fig. 5). The 2-log difference observed at 1 day postinfection in intrasplenic survival of the wild-type versus the dnaK_mut strain was in agreement with the 100-fold reduction of the intramacrophagic survival of the mutant 24 h after infection of J774 cells (Fig. 4). We therefore concluded that the effect observed in the complex in vivo model could be attributed mainly to the murine macrophage. In analogy to what we observed in J774 macrophage-like cells, these results confirmed the necessity of the presence of the native, inducible promoter of dnaK for survival of B. suis in mice.

FIG. 5.

Course of infection in BALB/c mice of wild-type B. suis (•) and the dnaK_mut strain (▴). Growth of the bacteria in spleens was determined at various time points for 8 weeks. Data are means and standard deviations for four mice.

Conclusions.

It was observed previously that intracellular pathogens such as Salmonella enterica serovar Typhimurium and Legionella pneumophila have adapted to the hostile environment of the macrophage by selective induction of a multitude of proteins, among which are many stress proteins (1, 2, 15). Virulence gene transcription in salmonellae is activated after phagosome acidification, illustrating the importance of a coordinated gene activation in response to the microenvironment encountered by the bacterium (18). Based on the previous finding that a dnaK null mutant lost its capacity for intramacrophagic multiplication (12), our aim in this work on the pathogen B. suis was to determine whether basic levels of DnaK were sufficient for intracellular resistance. We concluded that a constitutive but noninducible expression of dnaK in B. suis was not sufficient to allow survival in a macrophage model of infection and in a host organism, despite extracellular growth of the mutant at 37 and 38°C. The increased sensitivity to oxidative stress of the strain producing constitutively low levels of DnaK therefore suggested that the induction of the gene via its native heat shock promoter was an essential event in the resistance of the pathogen not only to high temperature but also to the stressful environment of the macrophage in the host animal.