Requirement of Histidine Kinases HP0165 and HP1364 for Acid Resistance in Helicobacter pylori

Abstract

In this study, we investigated a potential requirement of two-component signal transduction systems for acid resistance in Helicobacter pylori. In comparison to a wild-type strain, isogenic strains with null mutations in either HP0165 or HP1364 histidine kinases were impaired in their ability to grow at pH 5.0. The growth of complemented mutant strains was similar to that of the wild-type strain. H. pylori DNA array analyses and transcriptional reporter assays indicated that acid-responsive gene transcription was altered in the HP0165 and HP1364 null mutant strains compared to the parental wild-type strain. These results indicate that intact HP0165 and HP1364 histidine kinases are required for acid resistance in H. pylori.

Within the human stomach, Helicobacter pylori encounters a range of acidic pH conditions. The ability of H. pylori to live in this environment is likely to be dependent on regulation of bacterial gene expression in response to pH. Consistent with this view, changes in the pH of the culture medium result in alterations in H. pylori gene expression during growth in vitro (2, 4, 8, 13, 23, 24, 40).

Two-component signal transduction systems (TCSTSs), comprised of a sensor histidine kinase and a cognate response regulator, are commonly used by bacteria to detect and respond to environmental signals (reviewed in references 18 and 31). The H. pylori genome is predicted to encode three histidine kinases (3, 33), which have been designated ArsS, AtoS (FlgS, FleS), and CrdS (26, 28, 38). Other designations include HP0164/HP0165, HP0244, and HP1364 (the respective gene numbers in H. pylori strain 26695) and JHP0151, JHP0229, and JHP1282 (gene numbers in H. pylori strain J99). A genome sequence analysis of H. pylori strain J99 indicated that arsS (JHP0151) is a 1,326-bp open reading frame (ORF) in this strain (3). In contrast, genome sequence analysis of H. pylori strain 26695 indicated the presence of two smaller arsS ORFs, designated HP0165 (519 bp) and HP0164 (762 bp) (33), due to a single nucleotide deletion (at position +497 from the HP0165 translation initiation site). We have sequenced HP0165/HP0164 in our laboratory's copy of H. pylori strain 26695, and it contains a single arsS ORF, similar to that found in strain J99. Consistent with previous studies (8, 12, 15, 28), we use the H. pylori 26695 nomenclature in the current study, and we use the designation HP0165 when describing arsS.

H. pylori mutant strains containing deletions of individual histidine kinase genes (HP0165, HP0244, and HP1364) are defective in the capacity to colonize the mouse stomach (27). A recent study (28) reported that the transcription of three genes (HP0119, HP1432, and ureA) was upregulated by acidic pH in a wild-type H. pylori strain but not in an isogenic HP0165 mutant strain. Therefore, it was suggested that the HP0165 histidine kinase may function as an acid sensor (28). Binding of the cognate response regulator (HP0166) to sequences upstream from HP0119 and ureA has been demonstrated experimentally (12, 29).

Although there is evidence that the HP0165 histidine kinase is required for acid-responsive expression of several H. pylori genes, the role of TCSTSs in acid-responsive gene expression in H. pylori has not yet been investigated in detail, and a potential role of TCSTSs in H. pylori acid resistance has not yet been investigated. As a first step in investigating these topics, we constructed three mutant derivatives of H. pylori strain J99, in which the genes encoding histidine kinases were individually disrupted. Each gene was PCR amplified from H. pylori strain J99, and the amplicons were cloned into the vector pGEM-T (Promega). Primers for all PCR analyses in this study are listed in Table 1. The cloned genes were disrupted by insertion of a blunt-ended 1.2-kb aph3 cassette from pUC4K (Pharmacia) into unique HindIII, Eco47III, and BglII restriction sites in HP0244, HP1364, and HP0165 sequences, respectively. The resulting plasmids, which are unable to replicate in H. pylori, were then introduced into H. pylori J99 by natural transformation, and kanamycin-resistant transformants were selected. In each case, the appropriate insertion of the aph3 cassette into the desired site in the H. pylori chromosome, resulting from a double-crossover event, was confirmed by PCR analyses (data not shown). The mutant strains were designated J99-244Km, J99-1364Km, and J99-165Km (Table 2).

TABLE 1.

Oligonucleotide primers used in this study

Primera	Plasmid	Primer sequenceb	Locationc
1364-1	pRdx1364WT, pRdxA1364MUT	5′-CACGGCTCTAGATGGCTTCTAAGATCGCTAG-3′	1 to 20
1364-2	p1364	5′-CACGGCTCTAGATGGCTATCGCTTTAACCCCAC-3′	−146 to −124
1364-3	pRdx1364WT, pRdxA1364MUT	5′-GTACCCGGGTTATCCTTGAAATTGAACGC-3′	1012 to 1032
1364REV	p1364	5′-TCTTCGCTTAAAAGCCATTCTTC-3′	1082 to 1105
HP0165-1	pRdxA165WT, pRdxA165MUT	5′-CACGGCTCTAGATGCGTTTCTCTATCTTTTTTAAAGTTGTCG-3′	1 to 31
HP0165-2	pRdxA165WT, p165	5′-GTACCCGGGTTAACCCCACAATTGGCTTTTTCTGTTCC-3′	1297 to 1326
HP0165-4	pRdxA165MUT	5′-GTACCCGGGTCACGCTTTTATCCCCTTGAGCGAAGG-3′	499 to 526
165FOR	p165	5′-GCGATGCGAATATCTTTAGGGTAG-3′	−318 to −295
244-1	p244	5′-CGACAAGCTCCTCAAAGAACAAC-3′	−222 to −199
244-2	p244	5′-GCACATTGAAAAAATCGGTGCC-3′	1439 to 1461
73-1	pUreAcat	5′-GAAGGGCAGTTGTGCTAAACACC-3′	−858 to −835
73-2	pUreAcat, pRdxA-1	5′-GGAAGTGTGAACCGATTTGAACC-3′	410 to 433
PADureA-1	pRdxA-1	5′-GTAGCTCAGTTGGTAGAGCATACCTTGACATGG-3′	−336 to −303
954-1	pRdxA-1	5′-GGGGCAAGGATAGGATACGTTC-3′	−667 to −645
954-3	pRdxA-1	5′-CGTAGGGAACTTTAGGATCTTCAAAAACCC-3′	842 to 872
1432-3	p1432cat	5′-TTTCTGCCCTTAACCTCGCTC-3′	−475 to −454
1432-4	p1432cat	5′-CCGTGTTGTTCTTCATGGTGTG-3′	509 to 531

^aPrimers were used for the construction of the plasmids indicated.

^bRestriction sites (XbaI [TCTAGA] and SmaI [CCCGGG]) are underlined.

^cLocations refer to positions relative to the first nucleotide of the initiation codon. The target gene (HP number) for each primer is indicated in the primer designation.

TABLE 2.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Reference or source
H. pylori
J99	Wild type	Alm et al. (3)
J99-244Km	HP0244::aph3	This study
J99-1364Km	HP1364::aph3	This study
J99-165Km	HP0165::aph3	This study
J99-1364Km-WT	J99-1364Km, rdxA::HP1364, encodes wild-type HP1364 protein	This study
J99-1364Km-MUT	J99-1364Km, rdxA::HP1364Δ462-779, encodes HP1364Δ154-260 protein	This study
J1364Km-vector	J99-1364Km, rdxA::vector control insert	This study
J99-165Km-WT	J99-165Km, rdxA::HP0165, encodes wild-type HP0165 protein	This study
J99-165Km-MUT	J99-165Km, rdxA::HP0165Δ527-1326, encodes HP0165Δ175-442 protein	This study
J99-165Km-vector	J99-165Km, rdxA::vector control insert	This study
Plasmid
pGEM-T Easy	Ampr, PCR cloning vector	Promega
pUC4K	Kanr, Ampr	Pharmacia
p244	pGEMT::HP0244	This study
p165	pGEMT::HP0165	This study
p1364	pGEMT::HP1364	This study
p244Km36	p244::aph3	This study
p1364Km1	p1364::aph3	This study
p165Km1	p165::aph3	This study
p1364Δ2	HP1364Δ462-779, encodes HP1364Δ154-260 protein	This study
pRdxA-1	pGEMT::rdxA, PureureAB, Ampr	This study
pRdxA1364WT	pRdxA-1::HP1364, encodes wild-type HP1364 protein	This study
pRdxA1364MUT	pRdxA-1::HP1364Δ462-779, encodes HP1364Δ154-260 protein	This study
pRdxA165WT	pRdxA-1::HP0165, encodes wild-type HP0165 protein	This study
pRdxA165MUT	pRdxA-1::HP165Δ527-1326, encodes HP0165Δ175-442 protein	This study
pUreAcat	ureA::cat, pGEMT	This study
p1432cat	HP1432::cat, pGEMT	This study
pCM7	Ampr, promoterless cat	ATCC

We then analyzed the growth of the wild-type strain and the isogenic mutants in sulfite-free brucella broth containing 5% fetal bovine serum (BB-FBS), adjusted to pH 7.0 or pH 5.0 by the addition of hydrochloric acid. The growth of wild-type H. pylori strain J99 was similar under pH 7.0 and pH 5.0 conditions (Fig. 1A). The growth of strain J99-244Km was similar to the growth of the wild-type strain (Fig. 1A). In contrast, the growth of J99-165Km and J99-1364Km at pH 5.0 was impaired compared to the growth of the wild-type strain at pH 5.0 (Fig. 1B and C) (P ≤ 0.005 at 20-h and 32-h time points).

FIG. 1.

Growth of H. pylori J99 and isogenic histidine kinase null mutants. H. pylori strains were grown in BB-FBS (pH 7.0) and were then inoculated into fresh BB-FBS (initial pH of 5.0 or 7.0). OD₆₀₀ measurements were taken at the indicated time points. Growth of wild-type strain J99 and the isogenic histidine kinase mutants (J99-244Km [A], J99-165Km [B], and J99-1364Km [C]) are shown. Colony counts of cultures were performed in parallel and were consistent with the OD₆₀₀ readings (data not shown). These graphs depict analyses of triplicate cultures from single experiments (means ± SDs). The results shown are representative of three separate experiments.

The requirement of an intact HP0165 gene for H. pylori acid resistance is consistent with a previous report suggesting that HP0165 functions as an acid sensor (28), but a requirement of HP1364 for acid resistance was unexpected. To verify that the alterations in acid resistance were due specifically to mutations in HP0165 or HP1364, we sought to complement these mutations by inserting intact copies of HP0165 or HP1364 into the rdxA locus. Disruption of rdxA results in a metronidazole-resistant phenotype (17, 19). As a first step, we constructed a plasmid containing a 1.5-kb fragment of rdxA by cloning the appropriate PCR amplicon (i.e., HP0954, amplified with primers 954-1 and 954-3 from H. pylori strain J99) into pGEM-T. A 428-bp fragment of plasmid pAD-1 (15) containing the ureA promoter, ribosomal binding site, and initiation codon, as well as XbaI and SmaI restriction sites downstream from the ureA ATG start site, was then PCR amplified (primers PADureA-1 and 73-2) and inserted into the BbsI site of rdxA. The resulting plasmid was designated pRdxA-1. The complete HP1364 gene was PCR amplified using genomic DNA from H. pylori strain J99 as a template and cloned into the XbaI and SmaI sites of pRdxA-1. This suicide plasmid was then transformed into H. pylori strain J99-1364Km. Metronidazole-resistant transformants were selected, and PCR methods were used to verify that HP1364 had been inserted into the rdxA locus. To obtain a construct encoding a mutant form of HP1364 (HP1364Δ154-260), the cloned HP1364 gene was digested with Eco47III/NsiI to release nucleotide sequences coding for amino acids 154 to 260, and the mutated HP1364 gene was PCR amplified and cloned into pRdxA-1. This suicide plasmid was then transformed into J99-1364Km. Similar approaches were used to construct plasmids containing an intact HP0165 gene or a truncated HP0165 encoding HP0165Δ175-442. These plasmids were then introduced into the J99-165Km mutant strain, and metronidazole-resistant transformants were selected.

The complemented mutant strains were then tested for their ability to grow in pH 5.0 medium. Complementation of J99-1364Km by the introduction of an intact HP1364 gene into the rdxA locus restored the acid resistance of this mutant strain (Fig. 2A). J99-1364Km derivatives containing either an incomplete fragment of HP1364 (i.e., HP1364Δ154-260) or a vector control sequence (a 428-bp fragment of pAD-1 containing the ureA insert) in the rdxA locus were impaired in their ability to grow at pH 5.0 compared to the wild-type strain. Complementation of J99-165Km by the introduction of an intact HP0165 gene, but not the truncated HP0165 gene, into the rdxA locus restored the acid resistance of this mutant strain (Fig. 2B). These data indicated that intact HP1364 and HP0165 genes are required for H. pylori acid resistance.

FIG. 2.

Growth of complemented H. pylori HP0165 and HP1364 mutant strains. Panel A compares growth of wild-type H. pylori J99, an HP1364 null mutant strain (J99-1364Km), and complemented derivatives of the J99-1364Km mutant strain. The complemented mutants contain DNA sequences encoding wild-type HP1364 (1364 WT) or HP1364Δ154-260 (1364MUT) or vector control sequences in the rdxA locus of H. pylori J99-1364Km. Panel B compares growth of H. pylori J99-165Km and complemented mutant strains. Complemented mutants contain DNA sequences encoding wild-type HP0165 (165WT) or HP0165Δ175-442 (165MUT) or vector control sequences in the rdxA locus of H. pylori J99-165Km. These graphs depict analyses of triplicate cultures from single experiments (means ± SDs). The results shown are representative of three separate experiments.

To investigate whether intact histidine kinase genes were required for acid-responsive gene expression, we analyzed wild-type strain J99 and the isogenic mutant strains using H. pylori DNA macroarray analysis. H. pylori strains were grown in pH 7.0 BB-FBS to an optical density at 600 nm (OD₆₀₀) of 0.4 and then were inoculated into fresh BB-FBS (pH 5.0 or pH 7.0) and cultured in wells of a 24-well plate (Becton Dickinson Laboratories) with shaking for 75 min. Bacteria were harvested for RNA isolation using the acid guanidinium-phenol-chloroform method (TRIzol; Gibco BRL). Contaminating DNA was removed from the samples by treatment with RQ1 (RNA-qualified) RNase-free DNase (Promega), and RNA was purified using an RNeasy kit (QIAGEN). Total RNA (1 μg) was used to generate ³³P-labeled cDNA using H. pylori-specific cDNA labeling primers (Sigma-Genosys) as described previously (15). The labeled cDNA was hybridized to nylon Panorama H. pylori DNA arrays (Sigma-Genosys) using the manufacturer's protocols. Hybridized arrays were analyzed with a phosphorimager (Fuji, Inc.), and the hybridization signals were quantified with ImageQuant (Fuji, Inc.) and Array Vision (Imaging Research, Inc.) software. Four separate array hybridizations, using RNA from independent cultures, were performed for each growth condition examined. The hybridization signal for each gene was normalized to the total signal of that particular array, and a mean normalized signal for each gene was then calculated based on values obtained from all four arrays. For each gene, a pH 5/pH 7 expression ratio was calculated by dividing the mean normalized signal from pH 5.0 growth by the corresponding mean normalized signal from pH 7.0 growth. In an analysis of wild-type strain J99, the pH 5/pH 7 ratio (mean ± standard deviation [SD]) for all arrayed genes was 1.01 ± 0.29. Genes with pH 5/pH 7 expression ratios >2 SDs higher or lower than the mean ratio were regarded as putative acid-responsive genes.

Figure 3 (upper left panel) is a scatterplot depicting expression of 101 putative acid-responsive genes in wild-type strain J99. Many of these genes have been identified in previous studies as acid responsive (data not shown) (8, 24, 40). Only a few of these genes have been reported to be regulated in response to growth phase (32), a factor known to affect H. pylori adaptation to certain stress conditions (25). The diagonal line depicts a pH 5/pH 7 expression ratio of 1.0; therefore, all of the 101 points corresponding to acid-responsive genes are scattered or dispersed away from the diagonal line (Fig. 3, upper left panel). A scatterplot analysis of the same genes in strain J99-244Km appeared similar to the scatterplot analysis of the wild-type strain (Fig. 3). Considerably less scattering was detected in analyses of strains J99-1364Km and J99-165Km (Fig. 3). The variance among pH 5/pH 7 expression ratios for pH-responsive genes in the wild-type strain was relatively high compared to the variance among pH 5/pH 7 ratios for the same genes in the HP0165 and HP1364 mutant strains (variance of 1.2 versus 0.4 and 0.4, respectively). Table 3 lists genes identified as acid responsive in the wild-type strain but not acid responsive in the isogenic J99-1364Km or J99-165Km mutant strains. As shown in Table 3, mutagenesis of HP1364 and mutagenesis of HP0165 abrogated acid-responsive expression of many of the same genes.

FIG. 3.

DNA macroarray analysis of pH-responsive gene expression. One hundred one putative pH-responsive genes were identified based on array analyses of wild-type strain J99, grown at pH 7.0 and pH 5.0. Scatterplots depict expression of these 101 genes in the wild-type strain and three isogenic mutant strains, each grown at pH 7.0 and pH 5.0. Mean normalized signal intensities for the 101 genes are shown. Mean normalized signal intensities for 1,681 genes on the arrays are represented by the diagonal lines. Differential expression of a given gene at pH 5.0 versus pH 7.0 is reflected by deviation of a point from the diagonal line. The variance in pH 5/pH 7 expression ratios was higher in the wild-type strain (top left panel) than in the HP0165 or HP1364 mutant strains (variance of 1.2 versus 0.4 and 0.4, respectively). Coordinates representing expression of ureA and HP1432 are indicated.

TABLE 3.

Acid-responsive genes affected by either HP0165 or HP1364 mutation^a

Acid-induced genes				Acid-repressed genes
Not acid-responsive in J99-165Km		Not acid-responsive in J99-1364Km		Not acid-responsive in J99-165Km		Not acid-responsive in J99-1364Km
Geneb	Indexc	Gene	Index	Gene	Index	Gene	Index
HP0045	2.1	HP0045	2.0	HP0067	0.5	HP0067	0.5
HP0060	1.8	HP0060	2.1	HP0159 (40)	0.4	HP0192	0.4
HP0072 (8, 24, 40)	2.0	HP0072 (8, 24, 40)	1.8	HP0192	0.4	HP0224	0.8
HP0073 (8, 24, 40)	2.0	HP0073 (8, 24, 40)	2.0	HP0224	0.6	HP0259	0.7
HP0093 (40)	2.0	HP0093 (40)	1.7	HP0229 (8, 24, 40)	0.5	HP0288	0.7
HP0109	1.4	HP0109	1.2	HP0259	0.6	HP0362	0.7
HP0110	1.4	HP0110	1.3	HP0288	0.3	HP0454	0.7
HP0118 (8, 24)	2.3	HP0129	1.7	HP0362	0.4	HP0464	0.8
HP0119 (8, 24)	3.9	HP0148 (8, 40)	1.5	HP0365	0.4	HP0512 (40)	0.6
HP0148 (8, 40)	2.4	HP0225	1.4	HP0454	0.6	HP0523	0.7
HP0225	1.9	HP0294 (8, 24, 40)	2.0	HP0464	0.7	HP0555	0.6
HP0294 (8, 24, 40)	1.4	HP0460	1.6	HP0523	0.3	HP0579	0.5
HP0490	2.0	HP0490	2.4	HP0555	0.6	HP0633	0.6
HP0513	1.5	HP0513	1.5	HP0579	0.5	HP0635	0.6
HP0531	1.3	HP0642 (8, 24, 40)	2.5	HP0623	0.3	HP0656	0.5
HP0668	1.3	HP0668	1.3	HP0633	0.3	HP0675	0.4
HP0708	2.1	HP0682 (40)	2.2	HP0635	0.4	HP0700	0.6
HP0723 (40)	1.9	HP0708	1.6	HP0637 (24)	0.6	HP0827 (40)	0.3
HP0756	3.1	HP0723 (40)	2.0	HP0656	0.5	HP0851	0.7
HP0772	2.1	HP0807	1.7	HP0675	0.4	HP0906	0.1
HP0807	1.8	HP0900 (40)	1.4	HP0700	0.6	HP1045 (24)	0.6
HP0900 (40)	1.7	HP0948 (40)	2.9	HP0827 (40)	0.3	HP1071	0.5
HP0948 (40)	1.8	HP1019 (24)	1.5	HP0851	0.6	HP1177 (8, 24, 40)	0.6
HP1001	2.0	HP1104 (40)	1.2	HP0985	0.4	HP1205 (40)	0.7
HP1019 (24)	2.1	HP1203 (24)	2.0	HP1045 (24)	0.5	HP1240	0.5
HP1084	1.8	HP1238 (24, 40)	2.0	HP1071	0.7	HP1256	0.7
HP1192 (8, 40)	2.7	HP1317	1.5	HP1177 (8, 24, 40)	0.4	HP1270	0.6
HP1203 (24)	2.3	HP1329	2.3	HP1240	0.4	HP1367	0.6
HP1238 (24, 40)	2.8	HP1405 (8)	1.6	HP1270	0.4	HP1385	0.4
HP1319	3.8	HP1475	2.1	HP1366	0.4	HP1390	0.4
HP1432 (8, 24)	2.0			HP1367	0.4	HP1501 (8, 24, 40)	0.6
HP1455	2.5			HP1385	0.4	HP1526 (24)	0.6
HP1475	1.9			HP1390	0.6	JHP0164	0.4
				HP1526 (24)	0.7
				JHP0164	0.5

^aIn an analysis of wild-type H. pylori strain J99, all the genes listed in this table had pH 5/pH 7 expression ratios >2 SDs higher or lower than the mean pH 5/pH 7 ratio of all arrayed genes. In contrast, in an analysis of strain J99-165Km or J99-1364Km, the genes listed in this table were not acid responsive.

^bReferences in parentheses indicate acid-regulated genes identified by Bury-Mone et al. (8), Merrell et al. (24), or Wen et al. (40).

^cAn index value was calculated for each gene by dividing the pH 5/pH 7 expression ratio for the wild-type strain by the pH 5/pH 7 ratio for the indicated mutant strain.

We next constructed transcriptional reporter strains to monitor expression of two acid-responsive genes (ureA [HP0073] and HP1432) (see Fig. 3 and Table 3). These genes were selected because multiple studies have shown them to be acid responsive (8, 28) and because their transcriptional start sites are known (1, 28-30). Acid-responsive expression of both ureA and HP1432 was previously shown to be dependent on the presence of HP0165 (28, 29), but a requirement of HP1364 for acid-responsive expression of ureA has not been reported previously. The target genes were PCR amplified using the appropriate primers (Table 1) and cloned into the vector pGEM-T. A blunt-ended HindIII fragment containing a promoterless chloramphenicol resistance gene (cat [chloramphenicol acetyltransferase]) cassette from plasmid pCM7 (ATCC 37173 [9]) was inserted into unique cloning sites within the target genes (i.e., HindIII for ureA and XcmI for HP1432), and the resultant plasmids were individually introduced into H. pylori J99 by natural transformation. The insertion of the cat cassette into the desired target genes by allelic exchange was confirmed by PCR (data not shown). CAT expression was quantified using a CAT enzyme-linked immunosorbent assay (ELISA) kit (Roche Diagnostics), as described previously (20).

Consistent with the macroarray results, CAT expression was upregulated when ureA or HP1432 reporter strains (derived from wild-type strain J99) were grown at pH 5 compared to pH 7 (Fig. 4). We next introduced the ureA- and HP1432-cat reporters into the histidine kinase mutant strains. As expected, the acid induction of ureA-cat and HP1432-cat in the HP0244 mutant strains remained intact and was similar to expression in the wild-type strain. This result indicated that insertion of the cat cassette did not affect acid-responsive regulation of these genes. In accordance with published reports (28, 29), acid induction of both ureA-cat and HP1432-cat was abrogated in the HP0165 mutant strains. Acid induction of ureA-cat, but not HP1432-cat, was abrogated in the HP1364 mutant strains, consistent with the array results. Acid induction of ureA-cat was restored in the HP1364 mutant strain when a wild-type HP1364 gene was introduced into the rdxA locus (Fig. 4B). In contrast, the introduction of mutant HP1364 sequences or vector control sequences into the rdxA locus of J99-1364Km failed to restore acid induction of ureA-cat. These experiments confirmed that an intact HP1364 gene is required for acid-induced upregulation of ureA expression but is not required for acid induction of HP1432.

FIG. 4.

Analysis of gene expression in H. pylori using transcriptional reporters. H. pylori strains harboring HP1432-cat (A) or ureA-cat (B) transcriptional fusions were grown in pH 7.0 BB-FBS and then inoculated into fresh BB-FBS (pH 7.0 or pH 5.0). After 2 h, cultures were assayed for CAT activity using a CAT ELISA kit (Roche Diagnostics). Cell lysates (0.25 μg for ureA-cat and 0.05 μg for HP1432-cat) were added to each ELISA well. CAT units represent picograms of CAT protein per microgram of cell extract. Data are means ± SDs from triplicate cultures in a single experiment. The results shown are representative of three separate experiments. Asterisks indicate levels of CAT expression that were significantly higher when a reporter strain was grown at pH 5.0 than when the strain was grown at pH 7.0 (Student's t test, P < 0.05). Complemented J99-1364Km strains contain DNA sequences encoding wild-type HP1364 (1364WT) or HP1364Δ154-260 (1364MUT) or vector control sequences (vector) in the rdxA locus.

The results of this study indicate that intact HP0165 and HP1364 histidine kinases are required for H. pylori acid resistance. Several previous studies have shown that multiple H. pylori genetic loci are required for growth or survival of H. pylori at low pH (5, 6, 21, 22). The diminished acid resistance of HP0165 and HP1364 mutant strains compared to the wild-type strain may be due to changes in the expression of several genes, including those that encode urease, amidase (HP0294), and formamidase (HP1238) (Table 3). These enzymes, because of their capacity to generate ammonia, are all likely to play a role in acid resistance (7, 35, 37).

The importance of HP0165 in acid resistance is consistent with results of a previous study which suggested that HP0165 is an acid sensor (28, 29). However, the requirement of an intact HP1364 for acid resistance was unexpected. In a recent study, the HP1364-HP1365 TCSTS was reported to mediate copper resistance through the control of CrdA, a copper resistance determinant (38), but no previous studies have reported a role of HP1364 in acid resistance or acid-responsive gene expression. One could postulate that acidic pH may alter the solubility of copper (39) and, as such, changes in copper bioavailability might be sensed by HP1364.

TCSTSs comprising HP0165 and HP1364 may directly mediate acid-responsive expression of certain genes. In support of this view, binding of the HP0165 cognate response regulator (HP0166, ArsR) to DNA sequences upstream from ureA has been demonstrated experimentally (12, 29). A recent study showed that HP1365 (CrdR), the cognate response regulator of HP1364, binds to a 21-bp sequence upstream of crdA (38), but a survey of DNA sequences upstream from ureA did not reveal this putative HP1365 binding site (data not shown). Another possibility is that there may be regulatory cross talk involved in acid-responsive gene regulation in H. pylori. Several transcriptional regulatory systems, including NikR and Fur, have been reported to control gene expression in H. pylori in response to acidic pH (7, 8, 10, 11, 14, 16, 28, 34, 36). Regulation of H. pylori gene expression in response to acidic pH seems to be complex and may involve at least four regulatory systems (HP0165/HP0166 and HP1364/HP1365 TCSTSs, NikR, and Fur). Further studies will be required to understand the relationships among these regulatory systems and their respective roles in acid-responsive gene regulation.