Salmonella enterica Serovar Typhimurium HtrA: regulation of expression and role of the chaperone and protease activities during infection

The ability to adapt to adverse environments within and outside of hosts is essential for the success of most bacterial pathogens. Facultative intracellular pathogens such as Salmonella enterica serovar Typhimurium (S. Typhimurium) possess multiple stress-response systems to deal with the myriad of stresses that the organism may encounter (Foster & Spector, 1995; Rowley et al., 2006). Perturbations to the cell envelope are dealt with by a group of stress-response systems known collectively as envelope (or extracytoplasmic) stress responses (ESRs) (Rowley et al., 2006). The ESR controlled by the alternative sigma factor, RpoE (sigma factor E, σ^E), is dispensable in S. Typhimurium under standard laboratory growth conditions but is indispensable in vivo (Humphreys et al., 1999; Testerman et al., 2002). We have been investigating the role and importance of σ^E-regulated genes to S. Typhimurium in different environments (Humphreys et al., 1999, 2003; Kenyon et al., 2002; Lewis et al., 2008; Rowley et al., 2005).

The first σ^E-regulated gene found to be important for the virulence of a bacterial pathogen (S. Typhimurium) was htrA (Baumler et al., 1994; Johnson et al., 1991). HtrA (also known as DegP or Protease Do) is a serine protease active in the periplasm. In Escherichia coli HtrA is essential for growth above 42 °C (Lipinska et al., 1989). This is not the case in Salmonella, but S. Typhimurium htrA mutants exhibit increased sensitivity to killing by reactive oxygen species and reduced survival and/or replication within macrophages (Baumler et al., 1994; Humphreys et al., 1999; Johnson et al., 1991), and are highly attenuated in the murine model of systemic Salmonella infection (Chatfield et al., 1992; Humphreys et al., 1999). Since the seminal studies in Salmonella, HtrA has been found to play a role in the virulence of a wide range of Gram-negative and Gram-positive bacterial pathogens (Bringer et al., 2005; Cortes et al., 2002; Humphreys et al., 1999; Ibrahim et al., 2004; Johnson et al., 1991; Jones et al., 2001; Pallen & Wren, 1997; Pedersen et al., 2001; Rigoulay et al., 2005; Williams et al., 2000).

The expression of htrA has been best studied in E. coli. In this organism htrA expression is governed by a single promoter that is recognized by σ^E but is also positively regulated by the CpxAR two-component regulator (Danese et al., 1995; Lipinska et al., 1988; Pogliano et al., 1997). Several CpxR-P binding sites have been mapped upstream of the htrA promoter (Danese et al., 1995; Pogliano et al., 1997). These studies were performed using laboratory-adapted E. coli. Most virulent strains of E. coli are extracellular pathogens; in contrast, S. Typhimurium and other S. enterica serovars are predominantly intracellular pathogens which can survive and prosper in the hostile interior of macrophages. Survival of S. Typhimurium within macrophages requires HtrA (Baumler et al., 1994; Humphreys et al., 1999; Johnson et al., 1991). As the lifestyles of S. Typhimurium and E. coli are different, their requirement for HtrA and the regulation of htrA may also differ.

The main role of HtrA was thought to be degradation of damaged and misfolded proteins which accumulate in the periplasm when the bacteria are exposed to certain stresses such as high temperature. However, it was recently discovered that at low temperature HtrA can also act as a chaperone (Spiess et al., 1999). The chaperone and protease activity are reciprocally regulated such that the chaperone activity predominates at low temperature whereas the protease activity is dominant at high temperature (Spiess et al., 1999). Mature HtrA consists of three domains, an N-terminal protease domain and two C-terminal PDZ domains (PDZ1 and PDZ2). The protease activity of HtrA can be abolished by the change of a single serine residue at the active site to alanine (HtrA S210A). HtrA variants that lack protease activity (HtrA S210A) or either of the PDZ domains (HtrA ΔPDZ1 and HtrA ΔPDZ2), but not a variant lacking the entire protease domain, retain chaperone activity in vitro (Spiess et al., 1999). The individual contributions of the chaperone and protease domains of HtrA to S. Typhimurium survival and virulence have not been determined previously and were investigated in this study. We also analysed the regulation and expression of htrA in S. Typhimurium and show that it differs considerably from that reported in E. coli.

Bacterial strains, plasmids and culture conditions.
The bacterial strains and plasmids used in this study are shown in Table 1. Bacteria were grown in Luria–Bertani (LB) medium or on LB agar (LA) plates (Miller, 1972). When required, the media were supplemented with 100 µg ampicillin ml^–1, 40 µg chloramphenicol ml^–1 or 50 µg kanamycin ml^–1.

Table 1. Bacterial strains and plasmids used in this study

Detection of HtrA in S. Typhimurium by Western blotting.
Bacteria were grown on LA plates overnight at 30 °C or 42 °C, harvested into PBS, pelleted and resuspended in reducing SDS-PAGE sample buffer. Proteins (from ∼8x10⁷ c.f.u.) were separated by SDS-PAGE and transferred to PVDF membranes by electroblotting. HtrA was visualized by incubating the membranes with a rabbit polyclonal antiserum specific for HtrA followed by anti-rabbit immunoglobulin horseradish peroxidase conjugate and 4-chloro-1-naphthol substrate. Densitometry on bands was performed with ImageQuant software (GE Healthcare).

Isolation of RNA and S1-nuclease mapping.
RNA was isolated and S1 mapping was carried out essentially as previously described (Kormanec, 2001; Miticka et al., 2003; Skovierova et al., 2006). Briefly, an overnight culture was diluted 500-fold into fresh LB and incubated at 37 °C with aeration to the exponential (OD₆₀₀ 0.5) and stationary phase (OD₆₀₀ 1.7). Heat-shock- and cold-shock-stressed cells were grown to exponential phase and subjected to 30 min at 45 °C or 60 min at 10 °C respectively. For artificial rpoE expression, S. Typhimurium SL1344 containing pAC-rpoEST4 or pAC7 was grown in LB with chloramphenicol to early exponential phase (OD₆₀₀ 0.24) and expression of rpoE was induced for 3 h with 0.2 % arabinose.

At the appropriate time point the S. Typhimurium culture was chilled, washed with diethylpyrocarbonate (DEPC)-treated ice-cold 0.15 M NaCl, and total RNA was prepared as described previously (Kormanec, 2001). High-resolution S1-nuclease mapping was performed according to Kormanec (2001). RNA samples (40 µg) were hybridized to 0.02 pmol of appropriate DNA probe labelled at one 5' end with [γ-³²P]ATP and treated with 120 U S1 nuclease (Promega). The S1 probe (Fig. 2A) was a 447 bp DNA fragment prepared by PCR amplification from the S. Typhimurium SL1344 chromosomal DNA as template using the 5' end-labelled reverse primer htrASt1 (5'-GCCAAACCTAAACTCAGAGCCAG-3') from the htrA coding region and the non-labelled forward primer htrASt2 (5'-AAGCTTGTCGCTTAACGACTTTCG-3') from the dgt coding region. Oligonucleotide htrASt1 was labelled at the 5' end with [γ-³²P]ATP (ICN; 4500 Ci mmol^–1, 166.5 TBq mmol^–1) and T4 polynucleotide kinase (Promega). The protected DNA fragments were analysed on DNA sequencing gels together with G+A and T+C chemical sequencing ladders derived from the end-labelled fragment (Maxam & Gilbert, 1980). All S1-nuclease mapping experiments were performed several times using independent sets of RNA with similar results.

(44K): Fig. 2. Transcriptional analysis of the htrA gene in S. Typhimurium. (A) Genetic organization of the S. Typhimurium htrA region. The genes are indicated by thick arrows. The thin line below the map represents the DNA fragment (5' end-labelled at the end marked with an asterisk) used as a probe for S1-nuclease mapping. The positions of the identified promoters are indicated by thick bent arrows. (B) High-resolution S1-nuclease mapping of the TSPs of the S. Typhimurium promoters directing expression of htrA. See Methods for details. RNA was isolated from S. Typhimurium SL1344 containing pAC-rpoEST4 (lane 1), or pAC7 (lane 2) grown to early exponential phase in LB medium and induced with arabinose. (C) Mapping of the TSPs for promoters directing expression of S. Typhimurium htrA by primer extension analysis. The primer extension analysis was performed using RNA isolated from the indicated S. Typhimurium strain grown under different conditions in LB medium (see Methods). Lanes: 1, exponential phase; 2, heat shock; 3, cold shock; 4, stationary phase. Thin horizontal arrows indicate the positions of RNA-protected fragments; thick bent arrows indicate the nucleotides corresponding to the TSP of the indicated promoters. Before assigning the TSP, 1.5 nucleotides were subtracted from the length of the protected fragment to account for the difference in the 3' ends resulting from primer extension products and the chemical sequencing reactions.

Primer extension analysis.
The primer extension analysis was performed as follows: 50 µg total RNA was dissolved in 60 µl hybridization buffer (40 mM PIPES pH 6.4, 1 mM EDTA, 0.4 M NaCl, 80 %, v/v, formamide) at 65 °C, denatured together with 0.5 pmol 5'-labelled primer htrASt1 for 5 min at 95 °C, and annealed for 4 h at 45 °C. DNA samples were ethanol-precipitated, dissolved in 9 µl water and the following components were added: 0.75 µl RNasin (Promega), 3 µl 5x AMV-RT buffer (Promega), 0.75 µl 5 mM each of dATP, dGTP, dTTP, dCTP, and 0.75 µl 4 mg ml^–1 actinomycin, and the mixture was incubated for 2 min at 42 °C. The primer extension was initiated by adding 1.3 µl (26 U) AMV-RT (Finnzymes) and incubated for 2 h at 42 °C. The reaction was terminated with 25 µl RNase mix (100 µg ml^–1 DNase-free RNase A, 30 µg sonicated salmon sperm DNA ml^–1, 10 mM TE, pH 8) and incubated for 30 min at 37 °C. After addition of 20 µl 1 M NaCl, the mixture was extracted with alkaline phenol/chloroform and DNA precipitated with ethanol. The pellet was dissolved in 5 µl loading buffer (80 %, v/v, formamide, 10 mM NaOH, 1 mM EDTA, 0.05 % xylene cyanol, 0.05 % bromphenol blue), heated for 2 min at 95 °C, and an aliquot was loaded on a 6 % denaturing gel and separated together with the G+A and T+C sequencing ladders derived from the 447 bp end-labelled DNA fragment corresponding to the S1 probe. All primer extension experiments were performed several times using independent sets of RNA with similar results.

Construction of plasmids for the expression of HtrA variants in S. Typhimurium.
Spiess et al. (1999) constructed plasmids that encoded variants of E. coli HtrA lacking particular functions or regions of HtrA. These plasmids were kindly sent to us by Michael Ehrmann (University of Cardiff, UK). We wished to use the different htrA alleles to investigate the role of different HtrA domains and functions in survival of S. Typhimurium in vivo. The plasmid that carries these htrA variants, pCS19, is a high-copy-number vector. Expression of htrA is controlled by the trc promoter, which is repressed by LacI^Q, also encoded on the plasmid. The htrA plasmids were not directly suitable for our purpose because they would not be expressed in vivo and would not be regulated as htrA is normally in S. Typhimurium. To construct plasmids that could be used in S. Typhimurium we fused the genes for the htrA variants to the htrA promoter region of S. Typhimurium and cloned the genes into the low-copy-number vector pWSK29. The htrA promoter of S. Typhimurium SL1344 was amplified using oligonucleotides htrAFWproCL (5'-AGCCGCTATTAAGCTTGTCGCTTAACGAC-3') and htrAREVproCL (5'-TGTGGTTTTCGCCATGTGTTTCAATCTCGATTAAC-3'). All PCRs were carried out using the BIO-X-ACT high-fidelity PCR mix (Bioline). The individual htrA variants were amplified using the primers htrAFWCL (5'-CGAGATTGAAACACATGGCGAAAACCACATTAGC-3') and htrArv1bamhi (5'-CGGGATCCTTACTGCATTAACAGGTAGATGGTG-3'). The primers were designed such that the amplified promoter region fragment and amplified htrA variants fragments possessed overlapping regions of homology. To splice the molecules together, the PCR products were added in a ratio of 1 : 1 and another PCR was carried out using the primers htrAFWproCL and htrArv1bamhi. The final PCR products possessed HindIII and BamHI sites at their 5' and 3' ends, respectively. This facilitated their cloning into the multiple cloning site of the low-copy-number vector pWSK29. The resulting plasmids, phtrA1–phtrA5 (Table 1) were electroporated into the S. Typhimurium htrA mutant GVB1343. Expression of the different HtrA variants was checked by Western blotting using anti-HtrA antibodies (data not shown).

Analysis of virulence.
For all in vivo studies strains were grown statically overnight at 37 °C, centrifuged, washed and resuspended to the appropriate concentration in sterile PBS. Groups of four or five female BALB/c mice (6–8 weeks old, Harlan) were inoculated intraperitoneally with ∼4–9x10³ c.f.u. (in 200 µl) of one of the S. Typhimurium strains. Mice were killed 4 days after infection, the livers and spleens were removed and the number of c.f.u. present was determined.

Statistical analysis.
Statistical significance of datasets was analysed by ANOVA.

Identification of S. Typhimurium htrA promoters
A single promoter, which is recognized by σ^E, controls expression of htrA in E. coli. In a screen for S. Typhimurium σ^E-regulated genes using an E. coli two-plasmid system we previously identified a σ^E-dependent promoter upstream of htrA. S1-nuclease mapping identified a transcriptional start point (TSP) which was located 6 bp downstream of the sequence highly similar to the σ^E consensus sequence, GGAACTT-N₁₅-GTCTAA (Skovierova et al., 2006). However, there is still considerable HtrA production in a S. Typhimurium rpoE mutant (Fig. 1). At 30 °C there is no significant difference (P>0.05) in the amount of HtrA produced by the wild-type (WT) strain and the rpoE mutant. Growth at 42 °C significantly (P<0.05) increased expression of HtrA in the WT but not the rpoE mutant (Fig. 1). This suggests that there may be more than one promoter controlling htrA expression in S. Typhimurium. To investigate if this is the case we performed transcription mapping of the upstream region of S. Typhimurium htrA.

(20K): Fig. 1. Effect of rpoE expression on HtrA expression in S. Typhimurium. (A) Western blot (WB) of whole-cell lysates from S. Typhimurium strains probed with a specific anti-HtrA polyclonal rabbit serum. Bacteria were grown overnight on LA plates at 30 °C or 42 °C. Lanes: 1, S. Typhimurium WT 30 °C; 2, S. Typhimurium rpoE 30 °C; 3, S. Typhimurium WT 42 °C; 4, S. Typhimurium rpoE 42 °C. (B) Densitometry on WB bands. Each bar represents the mean±SD of three independent biological experiments. # indicates that the WT 42 °C mean is significantly higher (P<0.05) than that of any of the other groups. None of the other groups was significantly different from any of the others.

Expression of rpoE in SL1344 pAC-rpoEST4 was artificially induced with arabinose, RNA was isolated and subjected to S1-nuclease mapping (Fig. 2B). In addition to the σ^E-dependent promoter we previously identified, two other RNA-protected fragments were identified in the upstream region, suggesting the presence of other, σ^E-independent, promoters upstream of htrA (Fig. 2B). Although S1-nuclease mapping is more reliable for determining TSPs, for some A/T-rich RNA : DNA hybrids S1 nuclease can sometimes partially end-nibble, resulting in two or more closely migrating RNA-protected fragments. This may influence the localization of the TSP of the promoter (Kormanec, 2001). Therefore, in order to corroborate these results and to overcome the problems with S1 mapping, we performed primer extension analysis using 5'-labelled primer htrASt1 and RNA isolated from S. Typhimurium SL1344 and its isogenic rpoE mutant S. Typhimurium GVB311 from different growth phases and under stress conditions that have been previously shown to induce expression of σ^E-dependent promoters (Miticka et al., 2003). As shown in Fig. 2(C), the positions of all the promoters were confirmed and there were only single extension products corresponding to a single TSP for the corresponding promoters. The longest extension product corresponded to the most upstream promoter, htrAp1, with a TSP at G, 215 nt upstream of the most likely translation initiation codon ATG; the other extension products corresponded to the htrAp2 and htrAp3 promoters with TSPs at G or A respectively, 47 or 40 nt upstream of the ATG initiation codon (Fig. 2). The htrAp3 promoter was induced at the stationary phase and under cold shock, conditions previously shown to induce σ^E-dependent promoters in S. Typhimurium (Miticka et al., 2003). When RNA was prepared from the S. Typhimurium rpoE mutant grown under the same conditions, no RNA-protected fragments corresponding to the htrAp3 promoter were identified, while the intensities of the RNA-protected fragments corresponding to the htrAp1 and htrAp2 promoters were not affected (Fig. 2C). These results clearly indicated that the htrAp3 promoter is dependent in vivo upon σ^E in S. Typhimurium. As shown in Fig. 3, sequences upstream of the TSP of the two upstream σ^E-independent promoters, htrAp1 and htrAp2, exhibit partial similarity to the consensus sequence TTGACA-N_16–18-TATAAT for promoters recognized by the principal sigma factor σ⁷⁰ (Pribnow, 1975); their transcription may be therefore governed by RNA polymerase containing σ⁷⁰. Both promoters were expressed in S. Typhimurium WT and rpoE strains under all conditions. However, they were slightly induced under heat shock and their expression was reduced under cold shock and at stationary phase (Fig. 2C). The heat shock induction correlates with the increased production of HtrA at 42 °C vs 30 °C in both WT and rpoE S. Typhimurium (Fig. 1).

(46K): Fig. 3. Comparison of the E. coli and S. Typhimurium htrA promoter regions. Bent arrows indicate the TSPs of the htrA promoters. The –10 and –35 boxes of promoters, initiation (Mf) and stop (*) codons of htrA and dgt genes are in bold characters and underlined. The CpxR recognition sequences (De Wulf et al., 2002; Pogliano et al., 1997) are indicated by boxes around the sequence. The nucleotide sequences are parts of the genomic sequences deposited in GenBank/EMBL/DDBJ databases under accession numbers U00096 for E. coli K-12 MG1655 and AE006468 for S. Typhimurium LT2.

Importance of different HtrA functions to S. Typhimurium
HtrA possesses both protease and chaperone activity. In vitro both activities are strongly temperature dependent; the chaperone activity of HtrA predominates at low temperature whereas the protease activity dominates at higher temperatures (Spiess et al., 1999). HtrA has three domains: a protease domain and two PDZ domains (PDZ1 and PDZ2). Spiess et al. (1999) constructed variants of HtrA that lacked protease activity (due to a point mutation, S210A) or that were missing either or both of the protease domains. They found that the chaperone activity of HtrA was largely independent of the PDZ domains. In contrast, absence of either of the PDZ domains greatly reduced but did not abolish protease activity (Spiess et al., 1999). Which of the activities of HtrA are important for Salmonella during infection has not been determined.

We fused the variant htrA genes constructed by Spiess et al. (1999) to the full-length S. Typhimurium htrA promoter (htrAp123) by PCR and cloned the amplicons into the low-copy-number plasmid vector pWSK29 (see Methods). The resulting plasmids are phtrA1 (WT HtrA), phtrA2 (HtrA S210A, protease minus), phtrA3 (HtrA ΔPDZ1), phtrA4 (HtrA ΔPDZ2) and phtrA5 (HtrA Δprotease domain). The plasmids were introduced into the S. Typhimurium htrA mutant GVB1343. It was shown by Spiess et al. (1999) that these variants could complement the temperature-sensitive phenotype of an E. coli htrA mutant to different degrees. In the absence of HtrA E. coli is unable to grow at 42 °C and above (Skorko-Glonek et al., 2007; Spiess et al., 1999). In contrast, S. Typhimurium htrA mutants grow normally at 42 °C (Humphreys et al., 1999; Johnson et al., 1991). However, we have found that at higher temperature, 46 °C, growth of S. Typhimurium htrA mutants is inhibited (Fig. 4). We examined which, if any, of the phtrA plasmids could complement the temperature-sensitive growth defect of S. Typhimurium htrA at 46 °C (Fig. 4). The plasmids encoding WT HtrA and the protease-minus HtrA were able to complement fully the growth inhibition of S. Typhimurium htrA at 46 °C whereas the other HtrA variants could not (Fig. 4).

(24K): Fig. 4. Complementation of the growth defect of S. Typhimurium htrA at high temperature with HtrA derivatives. The S. Typhimurium htrA mutant GVB1343 harbouring a plasmid encoding an htrA derivative was cultured at 46 °C in LB with shaking. Each point represents the mean±SD of quadruplicate samples.

Importance of different HtrA activities to S. Typhimurium pathogenesis
S. Typhimurium htrA mutants are attenuated in vivo. The major defect is in growth in systemic tissues such as the liver and spleen (Humphreys et al., 1999; Johnson et al., 1991). To examine which of the different domains and functions of HtrA are important for S. Typhimurium growth in vivo we infected mice intraperitoneally with GVB1343 harbouring the variant htrA plasmids; each group of mice received a different strain. Four days later the mice were killed, the livers and spleens were removed and homogenized and the number of S. Typhimurium present was determined by viable counting. GVB1343 harbouring the pWSK29 plasmid alone was unable to multiply in the tissues to levels above those that were inoculated. The same was the case for GVB1343/phtrA3, GVB1343/phtrA4 and GVB1343/phtrA5 (Fig. 5). In contrast, GVB1343/phtrA1, which produces WT HtrA, was able to multiply in the liver and spleen, with c.f.u. in these organs ∼1000-fold higher than in the infectious dose. The mean c.f.u. in the organs of mice infected with GVB1343 phtrA1 was significantly greater than the mean c.f.u. of all the other groups (P<0.001). Interestingly, GVB1343/phtrA2 (which encodes HtrA S210A) was also able to increase in number relative to the infectious dose (∼12-fold) and this was significantly higher (P<0.05) than all the other groups except the GVB1343 phtrA1 group.

(16K): Fig. 5. Ability of HtrA derivatives to complement the growth defect of S. Typhimurium htrA during infection. Mice were inoculated intraperitoneally with one of the S. Typhimurium htrA strains. They were killed 4 days after infection and the number of c.f.u. present in the livers and spleens was determined. Each bar represents the mean number, ±SEM, of c.f.u. in the liver or spleen from four or five mice. # indicates that the mean of the phtrA1 group is significantly greater (P<0.001) than the mean of the equivalent organ from all other groups; * indicates that the mean of the phtrA2 group is significantly greater (P<0.05) than the mean of the equivalent organ from all other groups except phtrA1.

HtrA is a periplasmic protease/chaperone that is involved in the virulence of many bacterial pathogens. Expression of htrA at the appropriate time during infection is likely to be essential for successful infection by the particular pathogen. As HtrA is important for pathogens with different lifestyles, e.g. extra- and intracellular pathogens, it is likely that regulation of htrA activity will differ amongst pathogens. The regulation of htrA is best characterized in laboratory strains of E. coli (Danese et al., 1995; Lipinska et al., 1988; Pogliano et al., 1997; Rhodius et al., 2006). There is only a single htrA promoter (which is recognized by σ^E) in E. coli but htrA is also positively regulated by the CpxAR two-component regulator (Danese et al., 1995; Lipinska et al., 1988; Pogliano et al., 1997).

Our results show that the situation in S. Typhimurium is different. S. Typhimurium htrA has three promoters and only the most proximal of them, htrAp3, is dependent upon σ^E. The TSP of the S. Typhimurium htrAp3 promoter is in an identical position to its counterpart in E. coli (Fig. 3); however, no other promoters directing htrA were identified in E. coli (Lipinska et al., 1988; Rhodius et al., 2006). A possible explanation for the absence of htrAp1 and htrAp2 in E. coli may be in the change of the critical last T residue in both –10 promoter regions to a C residue (Fig. 3). All the CpxR-binding sequences identified in the E. coli htrA promoter region (Pogliano et al., 1997) were partially conserved in S. Typhimurium. Moreover, we identified another possible CpxR-binding box that was specific for S. Typhimurium (Fig. 3) which had high similarity to the CpxR consensus sequence (De Wulf et al., 2002). The differences in the –10 regions of the htrAp1 and htrAp2 promoters between S. Typhimurium and E. coli are conserved in uropathogenic and enterohaemorrhagic strains of E. coli (serotypes O73 and O157) (data not shown). S. Typhimurium is adapted to live in a modified vacuole within macrophages and other cells whereas pathogenic strains of E. coli are extracellular pathogens. Aiding survival within the macrophages is probably the most important role of HtrA during S. Typhimurium infection (Baumler et al., 1994; Humphreys et al., 1999; Johnson et al., 1991). It is tempting to speculate that the difference in the expression of htrA in S. Typhimurium and E. coli may be due to differences in their lifestyles.

Expression of htrA is upregulated in S. Typhimurium with macrophages (Eriksson et al., 2003; Everest et al., 1995) and other pathogens that reside within a phagosome such as Yersinia enterocolitica (Yamamoto et al., 1997) and Legionella pneumophila (Pedersen et al., 2001). However, there is almost no similarity between the htrA promoter regions of these bacteria and we were unable to locate counterparts of the S. Typhimurium htrA promoters in these promoter regions.

There is only one report regarding the characterization of the htrA promoters in another pathogenic bacterium. In Bartonella henselae htrA expression is directed by two tandem promoters (Resto-Ruiz et al., 2000), although there is almost no similarity to the S. Typhimurium htrA promoter region. B. henselae htrA expression was activated upon invasion of human microvascular endothelial cells (Resto-Ruiz et al., 2000).

HtrA has both protease and chaperone activity and these activities are reciprocally regulated by temperature in vitro (Spiess et al., 1999). Below 28 °C HtrA has little protease activity and the highest activity occurs at 42 °C or above (Spiess et al., 1999). In contrast, HtrA could efficiently refold denatured MalS protein at 28 °C but activity was reduced by two-thirds at 37 °C. Despite the low chaperone activity of HtrA at high temperature the protease-minus HtrA S210A protein was able to complement the temperature-sensitive phenotype of E. coli (Skorko-Glonek et al., 2007; Spiess et al., 1999) and S. Typhimurium (our results) htrA mutants. Skorko-Glonek et al. (2007) have shown that at high temperatures HtrA S210A forms complexes with denatured proteins but does not refold them; rather it prevents the formation of denatured protein aggregates in the periplasm (Skorko-Glonek et al., 2007). HtrA S210A could rescue an E. coli htrA mutant from the lethal effects of the misfolding of the porins OmpF and OmpC (CastilloKeller & Misra, 2003; Misra et al., 2000). This was not accomplished by HtrA S210A correctly folding OmpF or OmpC, suggesting that HtrA S210A was sequestering the misfolded proteins. Thus at normal and high growth temperatures HtrA's role in sequestering degraded proteins may be more important than the protein refolding function.

Our results show that during systemic infection of mice both the protease and chaperone/sequestration (C/S) functions of HtrA are active and important. The fact that the S. Typhimurium htrA mutant expressing WT HtrA grew to much higher numbers (∼60-fold higher c.f.u.) in the livers and spleens of infected mice than the htrA mutant expressing HtrA S210A might indicate that the protease activity is more important than the C/S function. However, it is possible that the two activities of HtrA act synergistically in vivo and this could account for the superior effectiveness of the WT HtrA compared to HtrA S210A at allowing S. Typhimurium to survive in vivo.

HtrA lacking either of the PDZ domains was unable to complement the growth defects of a S. Typhimurium htrA mutant in vitro at high temperature or in the livers and spleens of infected mice. Purified HtrA ΔPDZ1 and HtrA ΔPDZ2 proteins possess chaperone activity equal to WT HtrA and HtrA S210A in vitro (Spiess et al., 1999). Purified HtrA ΔPDZ1 and HtrA ΔPDZ2 also possess a low level of protease activity (Spiess et al., 1999) so it might be expected that these proteins would provide some benefit to the S. Typhimurium htrA mutant. However, HtrA ΔPDZ2 could not complement the temperature-sensitive phenotype of an E. coli htrA mutant and the HtrA ΔPDZ1 protein could partially complement if induced to very high levels (Spiess et al., 1999). HtrA in the periplasm has recently been shown to form large multimeric cage-like structures that are the functional form of the protein in vivo (Jiang et al., 2008; Krojer et al., 2008). Interaction between the PDZ domains on the different HtrA subunits is important for oligomerization (Jiang et al., 2008; Krojer et al., 2008). Therefore it is possible that the HtrA variants that lack either of the PDZ domains do not complement the growth of a S. Typhimurium mutant in vitro or in vivo because they cannot form multimeric HtrA.

We have shown that even in closely related organisms such as E. coli and S. Typhimurium the same gene, in this case htrA, may not be identically regulated. Also we demonstrate that both the protease and C/S activities of HtrA are involved in aiding survival of S. Typhimurium during infection and that both of the PDZ domains of HtrA are required for its function in protecting S. Typhimurium in stressful environments.

This work was supported by Science and Technology Assistance Agency under contract no. APVT-51-012004, VEGA grants 2/6010/26 and 2/0104/09 from the Slovak Academy of Sciences, and Wellcome Trust grant 065027/Z/01/Z. C. L. was funded by a grant from the Wellcome Trust (069099/Z/02/A). J. F. and J. S. were funded by a grant from the BBSRC (PRS12222). We thank Michael Ehrmann for the kind gift of the plasmids encoding the HtrA variants.

Edited by: V. Sperandio

Footnotes

^,†, Jacinta Farn¹ †Present address: SIPBS, University of Strathclyde, Royal College Building, Glasgow G1 1XW, UK. ‡Present address: c/o Department of Microbiology and Immunology, The University of Melbourne, Vic 3052, Australia.