Antisense RNA Modulation of Alkyl Hydroperoxide Reductase Levels in Helicobacter pylori Correlates with Organic Peroxide Toxicity but Not Infectivity

Abstract

Much of the gene content of the human gastric pathogen Helicobacter pylori (∼1.7-Mb genome) is considered essential. This view is based on the completeness of metabolic pathways, infrequency of nutritional auxotrophies, and paucity of pathway redundancies typically found in bacteria with larger genomes. Thus, genetic analysis of gene function is often hampered by lethality. In the absence of controllable promoters, often used to titrate gene function, we investigated the feasibility of an antisense RNA interference strategy. To test the antisense approach, we targeted alkyl hydroperoxide reductase (AhpC), one of the most abundant proteins expressed by H. pylori and one whose function is essential for both in vitro growth and gastric colonization. Here, we show that antisense ahpC (as-ahpC) RNA expression from shuttle vector pDH37::as-ahpC achieved an ∼72% knockdown of AhpC protein levels, which correlated with increased susceptibilities to hydrogen peroxide, cumene, and tert-butyl hydroperoxides but not with growth efficiency. Compensatory increases in catalase levels were not observed in the knockdowns. Expression of single-copy antisense constructs (expressed under the urease promoter and containing an fd phage terminator) from the rdxA locus of mouse-colonizing strain X47 achieved a 32% knockdown of AhpC protein levels (relative to wild-type X47 levels), which correlated with increased susceptibility to organic peroxides but not with mouse colonization efficiency. Our studies indicate that high levels of AhpC are not required for in vitro growth or for primary gastric colonization. Perhaps AhpC, like catalase, assumes a greater role in combating exogenous peroxides arising from lifelong chronic inflammation. These studies also demonstrate the utility of antisense RNA interference in the evaluation of gene function in H. pylori.

Helicobacter pylori establishes lifelong infections of the gastric mucosa of humans, causing chronic gastritis, duodenal ulcers, and mucosa-associated lymphoid tissue lymphoma, and is a risk factor for gastric cancer (11, 23). The remarkable ability of H. pylori to survive extreme acid and colonize the gastric niche is facilitated by several adaptations: (i) flagella are enveloped in a lipoprotein sheath which enables motility in strong acid (34, 49), (ii) a pH-gated urease system maintains membrane energetics at low pHs (55), and (iii) a novel acid-sensing chemotaxis system directs mucosal colonization and persistence (12). Additional genes associated with acid survival include the ArsRS two-component acid stress regulatory system (43, 45), various metal response regulators (18, 19), and a robust oxidative defense system that affords protection against inflammation-driven, reactive oxygen and nitrogen species (reviewed in reference 53). Over decades and through a process of constant mutation and selection, H. pylori strains become uniquely adapted to the physiologies of their respective hosts, leading to much of the genetic variation and evolution of genes associated with lifelong persistence (29, 31).

Although the H. pylori genome is less than 1.7 Mb in size, it encodes nearly all of the core metabolic capabilities found in bacteria possessing much larger genomes, suggesting that these organisms lack pathway redundancies and are limited in metabolic and nutritional diversity (9, 51). Thus, many housekeeping genes associated with biosynthetic pathways or central metabolism, which are nonessential in bacteria with larger genomes, are often found to be either essential for in vitro growth or conditionally essential for infection in animal models (9). While uniquely essential genes of central metabolism are of interest in drug discovery, these genes can complicate studies of pathogenic mechanisms. Such genes are often picked up through gene inactivation strategies (allelic replacement mutagenesis) that are commonly used to study gene function (9, 12, 13). In cases where mutations are lethal or enfeeble growth, further studies, particularly animal infectivity studies, are not pursued. For some bacteria, control of gene expression and titration of function can be regulated through the use of controllable promoters, but such tools have yet to be developed for H. pylori.

We have investigated the feasibility of using an antisense RNA approach to knock down protein levels in genes associated with infectivity or essential for viability. For other bacterial pathogens, antisense RNA has been applied to study the oxidative stress defense in slow-growing mycobacteria (56), to control exopolysaccharide production by Lactobacillus rhamnosus (7), to confirm the role of FtsZ in cell division in Borrelia burgdorferi (16), and to titrate targets of antibacterial action in Staphylococcus aureus (57). The emergence of antisense RNA as a major regulatory strategy for posttranscriptional control of microbial gene expression has benefited from studies of small regulatory RNAs, such as MicF, OxyR, RyhB, and DsrA (1, 21). To test the feasibility of an antisense RNA interference approach in H. pylori, we have targeted ahpC, which encodes alkyl hydroperoxide reductase, one of the most abundant proteins produced by H. pylori and one that has been extensively studied (3, 35, 36, 40, 41). Peroxide reduction by AhpC is activated by the thioredoxin/NADPH thioredoxin reductase system and is efficient in reducing both hydrogen and organic peroxides as well as peroxynitrite (3, 8). During extended periods of oxidative stress, AhpC becomes inactivated by peroxide (prevents depletion of NADPH pools) and catalase assumes a greater role in peroxide scavenging (47). Under these conditions, inactive AhpC monomers assemble into high-molecular-weight decamers that participate in chaperone-like refolding of damaged proteins (10, 42). ahpC mutants are hypersensitive to oxygen and organic peroxides and are restricted to atmospheres of less than 2% oxygen for growth (40). While ahpC mutants are unable to colonize mice (41), the relative importance of AhpC function in colonization cannot be distinguished from poor vigor due to enfeebled growth (40, 41).

In this study, we set out to determine how much AhpC function is required for growth and for mouse colonization. To facilitate these studies, we expressed various antisense ahpC (as-ahpC) RNA constructs from a plasmid and from a single copy in the rdxA chromosomal locus and demonstrated knockdowns in protein levels of up to 72%. Depressed levels of AhpC correlated with increased susceptibilities to hydrogen and organic peroxides; yet, AhpC levels 25% of wild-type (WT) levels had no enfeebling effect on in vitro growth. Knockdowns of 32% did not alter colonization efficiency for mice, suggesting that high levels of AhpC may not be required during the initial colonization, when oxidative stress would be minimal.

MATERIALS AND METHODS

Bacterial growth conditions.

Helicobacter pylori strains KE26695 (51), Hp1061 (20), and X47-2AL (herein referred to as X47) (22) were grown from −70°C frozen stocks on brucella agar containing 7.5% newborn calf serum (NCS; Sigma), 10 μg/ml vancomycin, 5 μg/ml trimethoprim, and 4 μg/ml amphotericin B at 37°C under humid, microaerophilic conditions (83% N₂, 10% CO₂, 7% O₂). When necessary, brucella agar was supplemented with 20 μg/ml kanamycin (Km). Liquid cultures were started at an optical density at 600 nm (OD₆₀₀) of 0.1 in 20 ml of brucella broth with supplements. The flasks were agitated at 195 rpm on a Labnet 30 microtiter dish shaker (Labnet International Inc.) under microaerophilic conditions. Escherichia coli DH5α and BL21 CodonPlus(DE3)-RIL cells (Stratagene) were grown in Luria-Bertani (LB) broth at 37°C under normal atmospheric conditions and, when necessary, supplemented with 100 μg/ml ampicillin (Amp), 20 μg/ml chloramphenicol (Cm), or 25 μg/ml Km unless otherwise described.

Genetic manipulation and PCR.

Basic DNA manipulations were performed as described before (46). PCR was performed using Roche Expand high-fidelity polymerase according to the manufacturer's protocol. Screening of mutants was typically performed using HotStarTaq (QIAGEN) according to the manufacturer's instructions. Plasmids and oligonucleotides are listed in Tables 1 and 2, respectively. DNA was subjected to 1% agarose gel electrophoresis, and when necessary, excised bands were cleaned with a QIAquick gel purification kit (QIAGEN).

TABLE 1.

Plasmids used in this study

Plasmid	Characteristics	Source or reference
pBlueScript II SK+ (pBSK)	Cloning vector, Ampr	Stratagene
pBSK::as-ahpC100	100 bp as-ahpC cloned in pBSK; Ampr	This study
pBSK::as-ahpC250	250 bp as-ahpC cloned in pBSK; Ampr	This study
pBSK::as-ahpC	Full length as-ahpC cloned in pBSK; Ampr	This study
pET-29b	Six-His-tagged C-terminal overexpression vector; Kmr	Novagen
pET-29b::ahpC	ahpCHp cloned in pET-29b; Kmr	This study
pDH37	H. pylori/E. coli shuttle vector; Kmr	25
pDH37::as-ahpC100	100 bp as-ahpC cloned in pDH37; Kmr	This study
pDH37::as-ahpC250	250 bp as-ahpC cloned in pDH37; Kmr	This study
pDH37::as-ahpC	Full length as-ahpC cloned in pDH37; Kmr	This study
pBC KS+	Cloning vector; Cmr	Stratagene
pHP1	aphA3 cassette; Kmr	30
pFD	fd bacteriophage terminator cloned in pBC; Cmr	This study
pRDX-K+	rdxA-replacement suicide vector; Cmr; Kmr	This study
pAS1563-2	Full length as-ahpC driven by PureA and terminated by fd bacteriophage terminator cloned in pRDX-K+	This study

TABLE 2.

Oligonucleotides used in this study

Primer	Sequence (5′ to 3′)a
AHPCA2F	TTCCCGGGATGTTAGTTACAAAACTTGC
AHPCA2R	GAGAATTCAGAAAATTCCATTAAGCTT
AHPCAF	TTCCCGGGAATAACGATGAAACAAGA
AHPCAR	ACGAATTCATCGTAACTCCTTAAGTG
AAHPCFBHI	CAGGGATCCAATAACGATGAAACAAGA
AAHPC100BPE	CAGGAATTCCATGGTGAAGTTTGCCCAGCA
AAHPC250BPE	CAGGAATTCGAAGAAGCGATCGCTTTGAG
AHPCFBamHI	GCGGATCCGATGTTAGTTACAAAACTTGCCCCA
AHPCRXhoI	CCGCTCGAGAAGCTTAATGGAAT
RDXAIFSacI	GTAGAGCTCGCATTCGTGGGATGAGCTA
RDXAIRXbaI	GCATCTAGACTTGCAAGAATGGCGCTCG
RDXAIIFXhoI	CGACTCGAGGTGGCAGAAGCGAGTCA
RDXAIIRKpnI	CTGGGTACCGCTCAATCTGACAACCCAC
PUREAFSpeI	CGTAACTAGTAGTCGTGGCCACCATTATCA
PUREARBamHI	GTAGGATCCATCAAGGTTGGATGTAATTG
AAHPCBFPstI	ATGCCTGCAGGCTGATTGAGTGGAAAGCATA
AAHPCB670R	GATACTCTGCAACGCCTTGGT

^aRestriction sites are underlined in the oligonucleotide sequences.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE).

Protein extracts were prepared as follows. Each culture was normalized to an OD₆₀₀ of 0.5 and centrifuged for 4 min at 6,000 rpm. The bacterial pellet was washed twice with phosphate-buffered saline (PBS; pH 7.4). The final pellet was then lysed with 100 μl loading buffer (25 μl 4× NuPAGE LDS [Invitrogen], 10 μl β-mercaptoethanol, and 65 μl double-distilled water [ddH₂O]), boiled for 10 min, and then centrifuged at 13,000 rpm for 5 min to remove cellular debris. Each sample (15 μl for E. coli extracts or 12.5 μl for H. pylori extracts) was run on a 4 to 12% NuPAGE gel (Invitrogen) at 150 V for approximately 1.5 h and then visualized with standard Coomassie blue.

Densitometry.

Densitometry was performed using Gel Pro v3.1.00.00 (Media Cybernetics Inc.) or TotalLab v2005 1D gel analysis (Nonlinear Dynamics Ltd.) software. Black and white scans were taken at 1,200 dpi using an Epson Expression 1680 scanner and imported into the densitometry software. After normalization of protein loads, with the WT level set at 100%, bands of interest were selected and the densities were compared.

Cloning of as-ahpC.

The full-length ahpC gene (hp1563) was PCR amplified using primers AHPCA2F and AHPCA2R from H. pylori KE26695 chromosomal DNA. Similarly, a 300-bp upstream region containing the native promoter of ahpC was PCR amplified with primers AHPCAF and AHPCAR. After EcoRI digestion, the two fragments were ligated together with T4 ligase and used as a template in a PCR with the flanking primers AHPCAF and AHPCA2F to place the antisense gene under its promoter. After gel purification (QIAGEN) and treatment with SmaI, the amplicon was cloned into similarly cut pBlueScript SK+ (pBSK), yielding pBSK::as-ahpC, and transformed into E. coli DH5α. The construct was confirmed by restriction digestion and DNA sequencing (DALGEN; Halifax, Nova Scotia, Canada).

as-ahpC constructs.

pBSK::as-ahpC was used as a template with primer AAHPCFBHI to amplify 100- and 250-bp as-ahpC constructs with primers AAHPC100BPE and AAHPC250BPE, respectively, by PCR. The resulting amplicons were gel purified and partially digested with EcoRI. After a full BamHI digestion, the DNAs were cloned into an EcoRI- and BamHI-digested pBSK plasmid, yielding pBSK::as-ahpC100 and pBSK::as-ahpC250, and transformed into E. coli DH5α. All constructs were confirmed with restriction digestion and DNA sequencing.

Testing as-ahpC constructs in E. coli.

The ahpC gene from H. pylori KE26695 was PCR amplified using primers AHPCFBamHI and AHPCRXhoI, digested with BamHI and XhoI, and cloned into a similarly digested pET-29 plasmid (Novagen), yielding pET-29b::ahpC containing a hexa-histidine tag (His₆). After confirmation with restriction digestion and DNA sequencing, pET-29b::ahpC was transformed into E. coli BL21 CodonPlus(DE3)-RIL cells (Stratagene). The resulting strain was made CaCl₂ competent and transformed with either pBSK, pBSK::as-ahpC100, pBSK::as-ahpC250, or pBSK::as-ahpC and selected on Amp (150 μg/ml), Cm, and Km (50 μg/ml). The resulting colonies were grown overnight in 5 ml of LB broth containing the same antibiotic selection. The next day, a 1:100 dilution of the overnight culture was used to inoculate 5 ml of LB containing the antibiotic mix and incubated with shaking (150 rpm) at 37°C. After the cultures reached an OD₆₀₀ of ∼0.6, 1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) was added and the cultures were put back in the shaker for an additional hour. Protein extracts were prepared and run on an SDS-PAGE gel as described above, except 15 μl was loaded per lane.

Introduction of as-ahpC constructs into H. pylori.

The full-length as-ahpC construct was digested from pBSK::as-ahpC with SmaI. The fragment was cloned into an NruI-digested pDH37 plasmid, an E. coli/H. pylori shuttle vector obtained from Rainer Haas (identical to pHel3) (25), yielding pDH37::as-ahpC. The 100- and 250-bp as-ahpC constructs were cut from pBSK::as-ahpC100 and pBSK::as-ahpC250 with BamHI and partial EcoRI digestion, and their ends were filled in with T4 polymerase (NEB). After gel purification, the fragments were cloned into a SacI-digested/T4 polymerase-treated pDH37 plasmid, yielding pDH37::as-ahpC100 and pDH37::as-ahpC250. All three plasmids were transformed into E. coli DH5α. Each plasmid was transformed into Hp1061, which stably maintains the pDH37 shuttle vectors (20). Natural transformation was performed as previously described (54), and colonies were selected with Km 20. The resulting colonies were expanded, and the protein profiles were viewed following SDS-PAGE.

Construction of the pRDX-K+ vector.

Approximately 580-bp flanking regions upstream and downstream of rdxA (hp0954) were amplified by PCR, using oligonucleotides RDXAIFSacI/RDXAIRXbaI and RDXAIIFXhoI/RDXAIIRKpnI, respectively, which were sequentially cloned into pBC KS+ as previously described (12) to create pRDX+. pRDX+ was confirmed by restriction digestion and DNA sequencing. A nonpolar Km resistance cassette from Campylobacter coli (aphA3) was excised from the E. coli/H. pylori shuttle plasmid pHP1 (30) with EcoRI, and following T4 polymerase treatment, the fragment was blunt end cloned into the EcoRV site of pRDX+, yielding pRDX-K+. The forward orientation (with the same direction as that of the rdxA flanking regions) was confirmed by restriction analysis and DNA sequencing.

Insertion of as-ahpC into the chromosome of H. pylori.

An 81-bp PstI/EcoRI fragment of pHel2 (25) containing the fd bacteriophage transcriptional terminator was cloned into a PstI/EcoRI-digested pBC KS+, yielding pFD. The new antisense RNA was constructed using the strong ureA promoter (P_ureA) to drive expression. P_ureA was PCR amplified with primers PUREAFSpeI/PUREARBamHI, digested with SpeI and BamHI, and cloned into a similarly cut pFD plasmid, yielding pFD::P_ureA. Secondly, the full-length ahpC gene was PCR amplified to include the 5′ untranslated region with primers AAHPCBFPstI/AAHPCB670R. This full-length ahpC amplicon was treated with T4 polymerase and digested with PstI. The digested ahpC amplicon was cloned into pFD::P_ureA, yielding pBC::AS1563-2. The P_ureA-as-ahpC fd terminator was digested from pBC::AS1563-2 with SpeI and EcoRI and cloned into pRDX-K+, yielding pAS1563-2. All constructs were confirmed by restriction digestion and DNA sequencing.

A SacI/KpnI digestion of pAS1563-2 excised an ∼3.5-kb fragment (containing the rdxA-flanking regions, the as-ahpC construct, and the aphA3 cassette) that was naturally transformed into X47. Colonies were selected on Km 20 and were designated X47rdxA::AS1563-1 and X47rdxA::AS1563-2. Allelic replacement of rdxA with the antisense construct was verified with PCR. As a control, a 2.6-kb SacI/KpnI fragment of pRDX-K+ (containing the rdxA-flanking regions and the aphA3 cassette) was naturally transformed into H. pylori X47 and designated X47rdxA::aphA3. Protein profiles were viewed following SDS-PAGE.

In vitro growth curves.

Broth cultures were prepared as described above and grown to OD₆₀₀s of 0.6 to 0.7 (14 to 15 h). The cultures were diluted to an OD₆₀₀ of 0.1 in 10 ml brucella broth supplemented with 10% NCS, and OD₆₀₀s were measured at 0, 4, 8, 12, 24, and 32 h with a SpectraMax M2 spectrophotometer (Molecular Devices). All experiments were performed in triplicate, and the means and standard deviations were plotted using KaleidaGraph software.

Peroxide challenge.

H. pylori strains were grown for 16 h, normalized to an OD₆₀₀ of 0.1, and plated for confluent growth on fresh brucella agar. After 2 h of recovery in a microaerophilic incubator, 7.5-mm disks saturated with 20 μl of various concentrations of hydrogen peroxide diluted in sterile ddH₂O, tert-butyl hydroperoxide diluted in sterile ddH₂O, or cumene hydroperoxide diluted in dimethyl sulfoxide (DMSO) were placed on the plates. After 3 days, zones of inhibition were measured around the disc. All experiments were performed in triplicate, and means and standard deviations were determined.

Kill curve.

Fourteen-hour broth cultures of X47, X47rdxA::aphA3, and X47rdxA::AS1563-2 (OD₆₀₀, ∼0.6) were normalized to an OD₆₀₀ of 0.1 in 4 ml of brucella broth containing 10% NCS. Each sample was subjected to 500 μM tert-butyl hydroperoxide, and 100-μl aliquots were removed at 0, 5, 10, 15, and 25 min, washed in PBS, and plated in triplicate on brucella agar after 10-fold serial dilutions. Colonies were scored after 4 days of microaerophilic incubation and means and standard deviations computed.

Catalase assay.

Catalase activity was determined spectrophotometrically (4) with cell extracts prepared from broth-grown H. pylori bacteria (20). Briefly, catalase activity was followed by a decrease in absorbance at 240 nm (ɛ = 43.48 M⁻¹ cm⁻¹), and units are expressed in μmol H₂O₂ oxidized per min per mg of protein. Protein was estimated by the Bradford method (Bio-Rad), using bovine serum albumin as the standard. Each catalase activity reported represents the means and standard deviations for at least three determinations.

Mouse infections.

CJB/6J mice (Jackson Laboratories, Bar Harbor, ME) were maintained in the University of Virginia School of Medicine animal quarters. Animals were bred under isolator conditions and placed in conventional housing for at least 4 weeks prior to experimentation. All animals were given access to water and commercial chow throughout the course of the experiments. Animals were sacrificed using CO₂ asphyxiation. Animal protocols were approved by the animal ethics committee of the University of Virginia. Broth cultures of X47, X47rdxA::aphA3, and X47rdxA::AS1563-2 were grown for 15 h to OD₆₀₀s of 0.6 to 0.7. Approximately 10⁸ cells were gently centrifuged (2,000 rpm for 8 min), resuspended in 200 μl PBS, and injected orally into 6- to 8-week-old CJB/6J mice on three separate days over a 5-day period. After 3 weeks, the stomachs of the mice were isolated and homogenized, and the contents were 10-fold serially diluted and plated on H. pylori selective media as previously described (26). Colonies were scored after 3 days of incubation at 37°C in a microaerobic incubator. Infectivity is reported as the number of animals colonized of the total infected, and microbial load in the stomach was determined from the means and standard deviations for the numbers of CFU/g stomach material (triplicate plating) from the infected animals of each group.

RESULTS

To determine whether as-ahpC RNA can interfere with expression of AhpC, we set up a test system for E. coli in which H. pylori ahpC was placed under tight control (pET-29b::ahpC) and the antisense construct was expressed from its endogenous promoter in pBSK::as-ahpC. As seen in Fig. 1, H. pylori ahpC was strongly expressed following IPTG induction (Fig. 1, lane A) and this expression was not affected by the presence of the pBSK vector control (Fig. 1, lane B). When the antisense constructs were tested in this system, the 100-bp as-ahpC construct (Fig. 1, lane C) and the full-length as-ahpC construct (Fig. 1, lane E) were highly efficient in abolishing AhpC production (nearly 100%), whereas the 250-bp as-ahpC construct was less efficient (Fig. 1, lane D). E. coli protein patterns were unchanged, relative to those for controls, by as-ahpC expression, suggesting that antisense effects were specific to the H. pylori ahpC mRNA. These results indicate that antisense RNA can be used to effectively knock down AhpC protein levels in the E. coli background.

FIG. 1.

as-ahpC expression in E. coli. SDS-PAGE of E. coli BL21 CodonPlus(DE3)-RIL whole-cell lysates with or without IPTG induction. Cells were normalized to an OD₆₀₀ of 0.5 and lysed in 100 μl of loading buffer, and 15 μl was loaded per lane. The arrow depicts H. pylori (Hp) AhpC expression. IPTG (+/−) signifies the absence (−) or presence (+) of IPTG induction during growth of cells. Lane 1, SDS-PAGE standard; column A (lanes 2 and 3), E. coli pET-29b::ahpC-His₆; column B (lanes 4 and 5), E. coli pET-29b::ahpC pBSK; column C (lanes 6 and 7), E. coli pET-29b::ahpC pBSK::as-ahpC100; column D (lanes 8 and 9), E. coli pET-29b::ahpC pBSK::as-ahpC250; column E (lanes 10 and 11), E. coli pET-29b::ahpC pBSK::as-ahpC. The higher molecular weight (MW, in thousands) for AhpC results from the additional amino acids of the histidine tag (His₆).

Antisense RNA knockdown of AhpC expression in H. pylori.

To determine whether as-ahpC RNA might function similarly in H. pylori, the as-ahpC constructs were introduced into plasmid-proficient strain Hp1061 on a shuttle vector (pDH37). As seen in Fig. 2A, AhpC was identified visually as a 23-kDa protein by SDS-PAGE. Compared with WT and vector controls (Fig. 2A, lanes 2 and 3, respectively), both the 100- and the 250-bp constructs showed knockdown levels of 42% and 34%, respectively (Fig. 2A, lanes 4 and 5), while the full-length as-ahpC gene reduced AhpC levels by 72% (Fig. 2A, lane 6). Repetitions of these experiments yielded comparable results, indicating that antisense RNA expression from the shuttle plasmid was relatively stable. Interestingly, the pDH37 vector control expressed ∼15 times more catalase activity than did WT Hp1061 (Fig. 2A shows the catalase band, and Fig. 2B lists catalase-specific activities). In contrast, the antisense RNA-expressing constructs exhibited diminished catalase activities that may reflect the efficiency of antisense knockdown of AhpC levels (Fig. 2B). We found that any DNA sequence cloned into the multiple cloning site of this vector also resulted in lower catalase activity. The plasmid effect on catalase activity was not further investigated.

FIG. 2.

Protein profile and growth of Hp1061 harboring as-ahpC. (A) SDS-PAGE of Hp1061 whole-cell lysates. Cells were normalized to an OD₆₀₀ of 0.5 and lysed in 100 μl of loading buffer, and 12.5 μl was loaded per lane. Lane 1, SDS-PAGE standard; lane 2, WT Hp1061; lane 3, Hp1061 pDH37; lane 4, Hp1061 pDH37::as-ahpC100; lane 5, Hp1061 pDH37::as-ahpC250; lane 6, Hp1061 pDH37::as-ahpC. *, catalase band. (B) Catalase activities of the respective strains. One unit represents μmol H₂O₂ consumed per min per mg of protein. (C) In vitro growth of WT Hp1061 (⧫), Hp1061 pDH37 (▪), Hp1061 pDH37::as-ahpC100 (▴), pDH37::as-ahpC250 (x), and Hp1061 pDH37::as-ahpC (*). H. pylori strains were grown in brucella broth with 10% NCS, and in the case of pDH37-containing strains, 20 μg/ml Km was added, and OD₆₀₀ measurements were taken at the indicated time points.

Since a previous study reported that an ahpC mutant as well as the WT strain were restricted for growth under oxygen tensions of <2% (40), we examined the antisense RNA-expressing strains for any defects in bacterial growth at 7% oxygen. As shown in Fig. 2C, there was no significant difference in growth rates between the as-ahpC-expressing constructs and the controls. These results suggest that a nearly 75% decrease in AhpC levels has no measurable effect on in vitro growth under atmospheres of 7% oxygen. Higher oxygen tensions were not examined in this study.

Knockdown levels of AhpC render H. pylori more susceptible to oxidative stress.

AhpC is involved in the detoxification of organic peroxides (3), and ahpC mutants are hypersensitive to oxidative stress (40). The peroxide sensitivities of the as-ahpC-expressing strains were compared with those of the WT and vector control strains of Hp1061 in a disk diffusion assay. Figure 3 shows that the antisense RNA-expressing strains were significantly more sensitive to hydrogen peroxide and to organic peroxides than were the WT and vector control strains (P < 0.001). As pointed out in Fig. 2A and B, the vector control strain was highly resistant to the effects of hydrogen peroxide due to the elevated catalase level but was essentially WT for susceptibility to the organic peroxides (Fig. 3). The increased sensitivity of the AhpC knockdown strains to hydrogen peroxide is most likely attributed to the low levels of catalase activity as noted in Fig. 2B. Also, in correlation with the relative AhpC levels, the full-length as-ahpC construct (72% knockdown of AhpC) was more sensitive to organic peroxides than the 100- and 250-bp constructs. These results establish a dose response relationship between AhpC levels and susceptibility to organic peroxides that is independent of catalase activity.

FIG. 3.

Peroxide susceptibility testing of Hp1061 and antisense RNA-expressing strains. Sterile 7.5-mm discs were saturated with either 100 mM H₂O₂ (prepared in sterile ddH₂O), 5 mM tert-butyl hydroperoxide (tBOOH; prepared in sterile ddH₂O), or 10 mM cumene hydroperoxide (prepared in DMSO) and applied to brucella agar plates seeded to confluence with the indicated strains: WT Hp1061 (black bar), Hp1061 pDH37 (gray bar), Hp1061 pDH37::as-ahpC100 (white bar), Hp1061::as-ahpC250 (hatched bar), and Hp1061 pDH37::as-ahpC (dotted bar). Zones of clearing were measured after 3 days of incubation. No growth inhibition was observed when discs were saturated with ddH₂O or DMSO. Note the increased catalase activity for the vector control and the resistance to 100 mM H₂O₂ but not to organic peroxides. Asterisks represent statistical significance (three separate experiments) based on Student's t test (P < 0.001).

Construction of chromosomal antisense RNA.

One of the limitations with H. pylori strains that are plasmid permissive is that they do not colonize mice (26). In the few strains where plasmids have been employed in animal studies, plasmid stability has been a limiting factor (Rainer Haas, personal communication). Therefore, we designed a version of as-ahpC that could be introduced into the chromosomes of mouse-colonizing strains. In contrast to non-mouse-colonizing strains, mouse-colonizing strains are nearly intractable for invasive genetics (12). Of several mouse-colonizing strains examined, strain X47 was determined to be a little easier to manipulate genetically than the SS1 and SS2000 strains. The 100- and 250-bp and full-length antisense constructs were introduced into the rdxA locus and under the control of the ahpC promoter. SDS-PAGE analysis of protein extracts from these strains showed that AhpC levels were essentially WT (data not shown). While the results suggest that single-copy expression of antisense RNA was less efficient than expression from the multicopy shuttle vector, the poor results might also be attributable to unknown regulatory effects associated with the rdxA locus or to mRNA stability.

In an effort to optimize the expression and stability of the antisense RNA, we redesigned the expression system. First, the ahpC promoter was replaced with the urease promoter (P_ureA), which is considered a strong promoter and is likely to be up-regulated in situ (19, 45). Second, a full-length as-ahpC construct was extended to include 94 bp of the 5′ untranslated region, including the Shine-Dalgarno sequences (35). Finally, to ensure that message read-through (downstream genes) was not causing unwanted secondary-structure problems, missense (antisense), or accelerated degradation, we added the fd bacteriophage transcriptional terminator (the final construct was designated AS1563-2). The AS1563-2 construct was introduced into the chromosome of X47 via the pRDX-K+ vector (Fig. 4 shows a schematic). In contrast to the ahpC promoter-driven antisense system, the AS1563-2 system achieved a 32% knockdown in AhpC levels (Fig. 5A, lane 4) compared to the WT X47 and X47rdxA::aph3 controls (Fig. 5A, lanes 2 and 3, respectively). Consistent with the plasmid-borne knockdowns, AS1563-2 was not enfeebled for in vitro growth (Fig. 5B).

FIG. 4.

Schematic of the construction of AS1563-2 and insertion into the chromosome of H. pylori (Hp). pRDX-K+ contains ∼600-bp upstream and downstream sequences of rdxA that flank a multiple-cloning site (MCS) and a Km resistance cassette (aphA3) that is oriented in the same direction as the rdxA upstream/downstream sequences. The rdxA replacement construct is flanked by SacI and KpnI sites. The AS1563-2 cassette was created by using the ureA promoter (P_ureA) to drive expression of a full-length as-ahpC construct (containing the 94-bp 5′ untranslated region) and is terminated by the fd bacteriophage transcriptional terminator. The AS1563-2 cassette was cloned into the MCS of pRDX-K+ upstream of the Km cassette and in the same orientation. An ∼3.5-kb rdxA::AS1563-2 fragment was excised from pAS1563-2 with SacI and KpnI and used to naturally transform WT H. pylori X47 cells. Following double recombination and replacement of rdxA with the AS1563-2 construct and aphA3 cassette, transformants were selected on Km.

FIG. 5.

Protein profile and growth of H. pylori X47 harboring AS1563-2. (A) Bacterial suspensions were normalized to an OD₆₀₀ of 0.5 and lysed in 100 μl of loading buffer, and 12.5 μl was loaded per lane. The arrow depicts AhpC. Lane 1, SDS-PAGE standard; lane 2, WT X47; lane 3, X47rdxA::aphA3; lane 4, X47rdxA::AS1563-2. (B) In vitro growth of WT X47 (⧫), X47rdxA::aphA (▪), X47rdxA::AS1563-1 (•), and X47rdxA::AS1563-2 (▴). H. pylori strains were grown in brucella broth with 10% NCS, and OD₆₀₀ measurements were taken at indicated time points.

AS1563-2 is more sensitive to peroxide.

We next examined whether as-ahpC expressed from a single copy (AS1563-2) was more susceptible to peroxide as measured by disk diffusion. Disk diffusion assays (Fig. 6A) show that H. pylori X47 rdxA::AS1563-2 was more sensitive to organic peroxides than were the WT and rdxA::aphA3 controls. However, equivalence was noted in response to hydrogen peroxide, which may be due to the intrinsically high level of catalase in this strain, which is >2-fold that of Hp1061 (data not presented). Figure 6B shows the time-dependent kinetics of killing of AS1563-2 by 500 μM tert-butyl hydroperoxide. The AS1563-2 strain, expressing as-ahpC, exhibited accelerated time-dependent killing compared with the WT X47 and X47rdxA::aph3 controls. No differences in numbers of CFU were observed when the colonies were challenged with ddH₂O. These results show that antisense RNA can be expressed from a single copy in the rdxA locus of H. pylori and that a measurable decrease in AhpC protein results in increased susceptibility to organic peroxide.

FIG. 6.

Growth inhibition of X47, X47rdxA::aphA3, and X47rdxA::AS1563-2. (A) Brucella agar was plated for confluent growth, and a sterile 7.5-mm disc was saturated with either 500 mM H₂O₂ (prepared in sterile ddH₂O), 15 mM tert-butyl hydroperoxide (tBOOH; prepared in sterile ddH₂O), or 50 mM cumene hydroperoxide (prepared in DMSO). Results for WT X47 (black bar), X47rdxA::aphA3 (gray bar), and X47rdxA::AS1563-2 (white bar) are shown. Zones of clearing around the discs were measured after 3 days of incubation. No growth inhibition was observed around discs saturated with sterile ddH₂O or DMSO. Asterisks represent statistical significance based on the Student's t test (P < 0.001). (B) Kill curves of WT X47 (⧫), X47rdxA::aphA3 (▪), and X47rdxA::AS1563-2 (•) subjected to 500 μM tert-butyl hydroperoxide (prepared in ddH₂O). Details of the experimental procedures are described in the text. The results depicted represent the means for three separate experiments.

Knockdown levels of AhpC do not alter mouse colonization efficiency.

Since previous studies had identified AhpC as a key component of the oxidative stress defense system of H. pylori and essential for mouse infectivity (3, 40, 41, 53), we tested whether antisense knockdowns of AhpC levels might correlate with decreased colonization efficiency for mice. In a preliminary study involving two mice, each infected with WT X47, X47rdxA::aphA3, or X47rdxA::AS1563-2, we noted a slight but not statistically significant decrease in colonization efficiency for mice infected with AS1563-2. However, an expanded follow-up study (Table 3) showed that there was no difference in the ability of the as-ahpC-expressing strain to colonize the stomachs of mice compared to what was found for the WT and rdxA::aphA3 controls. To ensure that the progeny of the infection were still producing as-ahpC RNA, colonies were immediately challenged with 15 mM tert-butyl hydroperoxide in a disk diffusion assay and the diameters of the zones of inhibition were unchanged from the values presented in Fig. 6A (data not presented). These results suggest that in vivo levels of organic peroxides must be lower than the susceptibility levels determined in vitro, probably reflecting the paucity of inflammation at or during the 3-week time course of this study.

TABLE 3.

Mouse colonization densities with H. pylori strain X47 and antisense constructs in CJB/6J mice^a

Strain	No. of mice infected/total no. of mice	No. of CFU/g stomach ± SD
Mock (1× PBS)	0/5	0
WT X47	5/5	(1.01 ± 0.13) × 107
X47rdxA::aphA3	5/5	(1.03 ± 0.30) × 107
X47rdxA::AS1563-2	5/5	(1.00 ± 0.05) × 107

^aFive mice each were infected with strain X47, X47rdxA::aphA3, or X47rdxA::AS1563-2, and 3 weeks postinfection, bacterial load was determined (CFU/g stomach) as described in the text.

DISCUSSION

Here, we demonstrate feasibility for an antisense RNA interference strategy, albeit not under the control of an inducible promoter, that can be used to knock down protein levels of selected genes in H. pylori. Our studies showed that antisense RNA expressed from a shuttle vector decreased AhpC protein levels by nearly 75%, which correlated with a substantial increase in susceptibilities to exogenous hydrogen and organic peroxides but did not affect bacterial growth. If high levels of AhpC function were required for in vitro growth, we would have expected a 75% knockdown in function to manifest in poor growth rates. We also showed that antisense RNA expression from the chromosomal rdxA locus of H. pylori, while less efficient than the plasmid-borne system (32% knockdown), also resulted in a significant increase in susceptibility to organic peroxides, but not with deficiencies in either in vitro growth or mouse colonization efficiency. Single-copy antisense RNA production was substantially improved by adding a strong promoter (P_ureA) and a terminator to reduce production of polycistronic RNA. Our studies also found catalase levels to differ widely among strains and to be influenced by plasmid constructs in strain Hp1061, and while variations in catalase activity affected susceptibilities to H₂O₂, they did not affect susceptibilities to organic peroxides. Since catalase mutants are fully infectious for mice (24), we suggest that a primary function for AhpC in H. pylori is in detoxifying organic (lipid) peroxides and peroxynitrites generated by acute inflammation. Since inflammation in the mouse model does not manifest before 8 to 12 weeks postinfection (17, 26), organic peroxide levels would likely remain low during the initial colonization period of our study; thus, 70% or even 25% of WT AhpC function might provide sufficient antioxidant protection. The knockdown effects on chaperone function reported previously for AhpC (10) could not be directly tested in these studies.

AhpC was selected for antisense RNA targeting because (i) it is a very abundant protein, (ii) it is a key component of the oxidative defense system, (iii) ahpC mutants are incapable of growing at oxygen tensions above 2%, and (iv) ahpC mutants are noninfectious for mice (3, 9, 40, 41, 53). In a previous study, WT strains were also reported to be enfeebled for growth in atmospheres of 2% oxygen (40), suggesting (assuming optimal growth conditions) that oxygen tensions in the gastric mucosa are probably greater than 2%, perhaps as high as 6%, which would be nonpermissive for growth of ahpC mutants. Under the conditions of our mouse infection study, AhpC knockdowns were not attenuated in colonization efficiency, despite the likelihood that knockdown efficiencies of >32% may have been achieved in situ as a result of acid stress induction of urease promoters (44). Thus, unlike pH taxis, urease activity, hrcA and hspR mutants, and motility, AhpC would not be considered a colonization factor (12, 26, 49, 55). Such a view is consistent with ahpC mutants in other pathogens, such as Helicobacter hepaticus, Salmonella enterica serovar Typhimurium, Mycobacterium tuberculosis, Porphyromonas gingivalis, and Legionella pneumophila, where no obvious alterations in virulence were observed (28, 33, 38, 48, 50). However, in the case of pathogens that establish lifelong chronic infections, a compelling reason for expressing high levels of AhpC might be in defense against exogenous organic (lipid) peroxides. Perhaps AhpC knockdowns, like catalase mutants (24), might be less fit in establishing persistent, long-term infections in mice or in gerbils, where inflammation is far more acute. Further studies are needed to fully assess the nature of the reactive oxygen and nitrogen intermediates generated during chronic inflammation and whether AhpC knockdowns are enfeebled for persistence.

Preliminary proof of principle studies with as-ahpC in the E. coli DH5α test system showed AhpC knockdowns of nearly 100%, but this efficiency was not achieved with H. pylori. The ability to overexpress antisense transcripts in H. pylori may be restricted by the relative efficiency of the transcriptional machinery, which is optimized for slow-growing bacteria. The few published attempts to overexpress certain genes in H. pylori through gene duplication or changes in expression of regulatory genes seem to support this notion (26, 43). In addition, little is known regarding mRNA stability and turnover in H. pylori.

Antisense RNA has been used to control gene expression in many bacteria (6, 7, 16, 27, 52), and knockdown efficiency and specificity are often functions of the size and structure of the RNA construct (27). In this regard, Ji et al. (27) showed that approximately 4 to 5% of the mRNA length was required for significant inhibition of protein synthesis. We found that the 100-bp as-ahpC construct was more efficient than a 250-bp antisense construct when expressed from the shuttle vector, but neither was as efficient as the full-length as-ahpC construct. This might be caused by different patterns of folding or a secondary structure of RNA that might affect RNA duplex formation or relative susceptibilities to endogenous RNases (2). When the same set of as-ahpC constructs was expressed from a single copy in the chromosome, the knockdown efficiency was poor under the endogenous ahpC promoter. The addition of a strong promoter (P_ureA) as well as additional sequence material in the 5′ untranslated region (Shine-Dalgarno and ATG start site regions) substantially improved antisense RNA efficiency. Sequestering of the Shine-Dalgarno sequence is a common strategy used with small RNAs and inhibitory RNA sequences in abrogating translation of target mRNAs, such as those from rpoS and sodB in E. coli (15, 37) and the classical trp operon of Bacillus subtilis (32). Since the urease promoter is upregulated in vivo in response to acid stress (19, 43, 44, 45), we cannot rule out the possibility that antisense knockdowns of greater than 32% were achieved in mouse infection studies. Finally, the addition of a strong terminator of transcription is important for reducing read-through into downstream genes, which might alter their expression. Future improvements for enhancement of RNA interference efficiency might include the use of smaller constructs that are complementary to the SD region or perhaps expression of multiple copies of the antisense RNA from a common locus.

To evaluate the general utility of antisense RNA against other genes in H. pylori, we examined the effects of antisense RNA on orphan regulator HP1043 and on the global response regulator ArsR (HP0166). Both of these systems have been extensively studied (5, 14, 39, 43-45), and in the case of HP1043, it has been suggested that posttranscriptional control maintains an invariant protein level regardless of the gene copy number (39). Consistent with previous findings, initial antisense constructs generated by inverting the hp1043 gene under its endogenous promoter did not lead to decreased levels of HP1043 protein as determined with antibody raised to the hexa-His-tagged, purified protein (data not presented). In contrast, antisense RNA generated to arsR was able to decrease ArsR protein levels by 40%, as measured by immunoblot analysis, without affecting growth rate in vitro (data not shown). However, we also noted that there is considerable variation in ArsR protein levels among H. pylori strains, with that in X47 being nearly twofold that in KE26695. Studies are in progress to optimize the expression of as-arsR under the urease promoter, which should achieve knockdown efficiencies of greater than 40%. Perhaps under optimized conditions, ArsR knockdowns will display defects in acid tolerance and infectivity that might further define the functions of regulated genes in this system.

In summary, we have developed and validated an antisense RNA interference method that can be applied to the study of essential genes in H. pylori. Our studies show that antisense knockdowns of AhpC protein levels correlated with increased susceptibilities to hydrogen and organic peroxides, confirming earlier findings that AhpC function is a key component of the oxidative stress defense system of H. pylori (3, 40, 53). We also demonstrated, for the first time, the use of antisense RNA interference in a mouse gastric infection model to assess the function of a bacterial gene in pathogenesis. Our studies showed that 25% of AhpC function is sufficient to promote WT growth in vitro, despite an increase in susceptibility to organic peroxides, and that knockdowns of 32% in a mouse-colonizing strain did not affect colonization efficiency. We suggest that the abundant levels of AhpC noted for H. pylori may represent an essential adaptation required to combat lipid peroxides and peroxynitrites produced during the lifelong chronic infection associated with this pathogen.