![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Previous Article | Next Article 
Antimicrobial Agents and Chemotherapy, July 2004, p. 2524-2530, Vol. 48, No. 7
0066-4804/04/$08.00+0 DOI: 10.1128/AAC.48.7.2524-2530.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Department of Molecular Sciences, Infection Discovery, AstraZeneca R&D Boston, Waltham, Massachusetts 02451
Received 16 December 2003/ Returned for modification 27 February 2004/ Accepted 14 March 2004
 |
ABSTRACT |
|---|
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Benzimidazoles (BIs) were previously synthesized that possess potent anti-H. pylori activities with MICs at which 90% of isolates were inhibited of 
0.5 µg/ml for a diverse panel of 27 H. pylori strains, including several strains resistant to MTZ and CLR (6). In addition, many of these compounds have been characterized as being highly selective for H. pylori, with MICs at which 90% of isolates were inhibited of >64 µg/ml for a diverse collection of 25 aerobic bacterial strains and 18 anaerobic strains (6). These are attractive properties (e.g., potent and selective) for a prospective anti-H. pylori-specific agent to possess.
The cellular target of the BIs that is responsible for this highly potent and selective activity is currently unknown. The study of antimicrobial resistance is one means to gain an understanding of the mode of action and to identify the cellular target of an antimicrobial agent. In addition, resistance can be studied to examine the potential for resistance development throughout the research and development process. The intention of this study was to identify the mechanism of H. pylori resistance to BIs with the idea that it may help identify the molecular target for this structural class of compounds.
|
MATERIALS AND METHODS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Antimicrobial compounds. Ampicillin, kanamycin, piericidin A, and rotenone were obtained from Sigma (St. Louis, Mo.). The BI derivative, compound A, was prepared at AstraZeneca R&D Boston as previously described (Fig. 1) (6).
|
|
Genomic DNA for PCR analysis and transformation was isolated using the Wizard genomic DNA purification kit, plasmid DNA was prepared by alkaline lysis using the Wizard Plus plasmid miniprep kit, and phage DNA was purified using the Wizard Lambda Preps DNA purification system according to the manufacturer's instructions (Promega, Madison, Wis.). Genomic DNA for library construction was prepared from H. pylori using a published modification of the guanidine lysis protocol of Pitcher and coworkers (26, 28). This protocol was used in place of the Wizard genomic DNA purification kit to reduce the amount of chromosomal shearing.
Bacteriological methods.
Mutation frequencies were determined as follows: H. pylori was grown in broth culture to mid-exponential phase (optical density at 600 nm [OD600], 
0.8), centrifuged, and suspended in brucella broth to a concentration of 100 OD600, of which 10 µl (108 cells) was spotted onto BAP for overnight incubation. The following morning, each spot containing approximately 1010 cells was suspended in 1 ml of brucella broth and dilution plated onto BAP containing no compound (viable counts) or compound A at 2 times the MIC (resistant counts). The mutation frequency was calculated by comparing the number of viable colonies on BAP with the number of colonies observed on compound A-containing plates. MICs were determined by plating 5 µl of brucella broth containing 106 H. pylori cells, grown as described above, onto media containing twofold-increasing concentrations of compound A. The minimum concentration defining the threshold limit for growth was reported as the MIC.
Transformation of H. pylori. H. pylori was transformed with DNA as previously described (34). The source of DNA (sheared genomic, restriction endonuclease digested, PCR generated, or plasmid) did not impact the method of transformation. Briefly, H. pylori was grown to an OD600 of 0.5 to 0.8. The cells were then concentrated by centrifugation and suspended to a final concentration of OD600 = 100 in brucella broth. Twenty microliters was spotted onto BAP and incubated for 2 h at 37°C under microaerophilic conditions. Next, approximately 1 µg of DNA was applied directly onto the spotted H. pylori cells (sterile water was used as a negative control) and incubated overnight. The spotted cells were suspended into 1 ml of brucella broth and spread directly onto BAP containing the appropriate antimicrobial compound for selection of transformants. Colonies would generally appear within 3 to 5 days.
Construction and characterization of an H. pylori genomic library. BglI-digested H. pylori genomic DNA (9 to 23 kb) was ligated into LambdaGEM-11 BamHI arms and packaged with Packagene extract (Promega). Phage buffer (20 mM Tris-HCl [pH 7.4], 100 mM NaCl, 10 mM MgSO4) and chloroform were added to the packaged DNA library and gently mixed by inverting the tube six times. The chloroform was then allowed to settle to the bottom of the tube. The resulting phage lysate, minus the chloroform, was subsequently used for transduction experiments.
Preparation of the plating bacteria was done according to the method of Sambrook et al. (30). Briefly, E. coli strain ER1793 was prepared for transduction by first growing it in 50 ml of sterile LB supplemented with 0.2% maltose (to induce the 
receptor) overnight at 37°C on a rotary shaker (250 rpm). The cells were concentrated by centrifugation (4,000 x g; 10 min) and suspended to an OD600 of 2 in 10 mM MgSO4. Cells could be stored at 4°C from several days to 3 weeks.
The packaged phage was titrated to determine the appropriate dilution to generate well-separated plaques on a lawn of ER1793. For each plate of plaques, 100 µl of the appropriately diluted phage was added to 100 µl of the prepared ER1793 cells described above. The phage were allowed to adsorb to the cells for 30 min at 37°C. Three milliliters of molten (47°C) LB top agar (autoclaved 1.0 g of Bacto-tryptone, 0.8 g of NaCl, 0.6 g of agarose in 100 ml of H2O) was added to a 5-ml tube, gently vortexed, and immediately poured onto 1-day-old LB agar plates prewarmed to 37°C. Once the top agar hardened, the plates were incubated at 37°C overnight (8 to 10 h). The following morning, well-separated plaques were picked and plated onto fresh lawns of ER1793 using a 36-square grid. Ten such 36-grid plates were prepared to comprise a 360-clone H. pylori library (10 pools of 36). To prepare lysate from each of the 10 pools, the plates were overlaid with 5 ml of SM buffer (sterile 0.01% gelatin, 50 mM Tris-HCl [pH 7.5], 100 mM NaCl, 8 mM MgSO4). Following this, the top agarose was scraped with a glass spreader (avoiding the bottom agar) and transferred to a conical 50-ml centrifuge tube. The agarose was broken up with a spatula and incubated at room temperature for 30 min with periodic shaking. The tube was then centrifuged at 10,000 x g for 15 min at 4°C, and the supernatant was transferred to a fresh tube and gently mixed with chloroform to 0.3% (vol/vol) by inverting the tube six times. One additional centrifugation step was added to settle out the remaining agarose and the chloroform. The recombinant phage DNA was purified from the lysate medium using Promega's Wizard Lambda Preps DNA purification system. The purified phage DNA was used for restriction analysis, transformation of H. pylori, and DNA sequencing.
|
RESULTS |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Genomic DNA was isolated from a compound A-resistant mutant of strain J99 and from its isogenic parent (wild type [WT] J99). The isolated DNAs were transformed into a highly competent H. pylori laboratory strain, AH244. A negative (water) control was included in this experiment to compare the DNA treatments to the spontaneous rate of resistance. Greater than 1,000 H. pylori colonies were observed following transformation with 1 µg of DNA isolated from the resistant strain, while only 3 and 5 colonies were observed after transformation of an equal number of cells with 1 µg of DNA from WT J99 and from the negative control, respectively. The latter treatments that resulted in only a few colonies reflected the spontaneous resistance rate determined for H. pylori strain AH244. The results from this transformation experiment clearly indicated that there was a transformable genetic basis to compound A resistance in H. pylori.
Construction of a genomic library of DNA isolated from a BI-resistant mutant of H. pylori strain J99. Given the results described above, a DNA library was constructed to screen for the compound A resistance gene(s). BglII was chosen to cleave the DNA into larger fragments (the J99 genome is predicted to have 231 BglII sites) for library construction. Three fractions from the BglII-restricted DNA, corresponding to 3.0 to 3.5 kb (F1), 6.5 to 9.5 kb (F2), and 9.5 to 23 kb (F3), were purified after electrophoretic separation on an agarose gel. Approximately 0.5 µg of each DNA fraction was used to transform WT AH244 cells. This was followed by selection for resistance on BAP containing 1 µg of compound A/ml. The results from this transformation experiment indicated that fraction F3 (9.5 to 23 kb) contained the resistance gene, since it was the only fraction that generated compound A-resistant colonies above the spontaneous resistance rate.
GEM-11 (lambda-based vector) was chosen as the vehicle of choice to construct the genomic library, since it accepts inserts between 9.5 and 23 kb in size and the BamHI-generated cloning arms of GEM-11 are compatible with the BglII-generated inserts in fraction F3. The BamHI arms were ligated with F3 DNA, packaged, and titrated onto lawns of ER1793 to determine the most appropriate working dilution of the packaged ligation reaction mixture. Phadnis and coworkers showed previously that ER1793 was by far the best E. coli host (out of eight tested) for cloning H. pylori DNA into GEM-11 (27).
The formula of Clarke and Carbon, N = ln(1 – P)/ln(1 – 1/n), where N = the number of individual recombinants, P = the probability of including any DNA sequence of a random library, and n = the size of the genome relative to a single cloned fragment, was used to determine how many clones should be screened to be >99% confident that the constructed library was complete (7). In this study, there were 62 BglII fragments in the sequenced H. pylori strain J99 genome between 8,869 and 23,373 bp in size. These 62 fragments were predicted to comprise 868 kb, or roughly 52% of the genome. Using the above formula, 99% confidence would require 283 clones. Our strategy included 360 clones, giving a confidence level of 99.7%.
Initially, 10 pools each comprising 36 individually amplified clones were prepared and analyzed by agarose gel electrophoresis. The pooled DNA was digested with SacI, which cuts the GEM-11 arms away from the insert. As expected, the arms were liberated and observed as heavily staining bands at approximately 9 and 20 kb. The cloned DNA inserts from each pool were observed as unique profiles of smaller SacI fragments ranging in size from approximately 2 to 20 kb. This suggested that the library was random and contained many unique DNA inserts.
Screening for the compound A resistance gene(s). In the first round of screening the 10 pools of 36 clones, 2 pools (numbers 4 and 7) were identified as positive after transformation of WT AH244. Pool 7 was chosen for further analysis. Pool 4 is discussed below in the section describing the sequencing of additional spontaneous mutants. Pool 7 was further divided into six pools of six clones, and the DNA was transformed into WT AH244 for selection on 1 µg of compound A/ml. Pool 3 contained a positive clone in this experiment, and it was therefore divided into the six individual clones for transformation. A single clone, 17, was identified as containing the genetic determinant(s) responsible for resistance.
Characterization of clone 17. Clone 17 DNA was prepared for sequencing and subcloning. The DNA sequence was determined for the junctions of the cloned insert by using primers specific for the T7 and SP6 promoters at the vector arm junctions (Table 2). The sequenced junctions demonstrated that the clone spanned from nucleotides 1310501 to 1327589 on the J99 chromosome (2; http://scriabin.astrazeneca-boston.com/hpylori/). The genes contained on clone 17 were as follows: frr (ribosome releasing factor), pyrE (orotate phosphoribosyltransferase), JHP1179 (H. pylori-specific gene), JHP1180 (gene of unknown function), and the entire nuo operon of 14 open reading frames (JHP1181 to -1194) encoding NADH oxidoreductase (Fig. 2A).
|
Sequencing of nuoD from additional independent spontaneous compound A-resistant mutants. Thirty-three independent H. pylori strain J99 resistant mutants were isolated on BAP containing 1 µg of compound A/ml. Genomic DNA was prepared from each of the 33 resistant mutants, and specific primers (JHP1184-F and -R) were used to PCR amplify the entire nuoD gene from each of the mutants. DNA sequencing of each PCR product was completed using primers CF-1 thru CR-5 (Table 2) to determine whether nuoD carried the mutation, as was shown for the first isolated compound A-resistant mutant in clone 17. Figure 3 summarizes the sequencing results for all 33 mutants aligned with the WT J99 NuoD sequence. All but one of the mutants had a single change at the C terminus of NuoD (Fig. 3). Twenty-eight of the mutants had the glycine-to-serine change at residue 398 (G-to-A transition; J99A.1), three had a phenylalanine-to-serine change at residue 404 (T-to-C transition; J99A.2), one had a valine-to-methionine change at residue 407 (G-to-A transition; J99A.3), and one did not have a detectable mutation in nuoD (J99A.4). The other positive pool (number 4) identified during the initial library screening process was subjected to PCR and sequence analysis as above. The positive clone in this pool also contained nuoD with a single G-to-A transition at the same position as observed for the J99A.1 class of mutants (Fig. 3).
|
Directed knockout of nuoD to determine essentiality. It was of interest to determine whether nuoD is essential in H. pylori, especially since it has been reported to be nonessential in E. coli (4, 5). The nuoD gene was cloned into pGEM-T after PCR amplification from WT J99 genomic DNA using primers JHP1184-F and -R (Table 2). A 718-bp internal deletion was created in nuoD by PCR using primers CR-2 and CF-4, followed by ligation of a Campylobacter coli kanamycin resistance gene (aphA-3 [23]) in place of the deleted DNA. The constructed knockout plasmid (pSM135) was transformed into WT H. pylori J99 and AH244, and the cells were plated onto selective medium. No kanamycin-resistant colonies were recovered for the J99 transformation, and 36 kanamycin-resistant colonies were recovered for the AH244 transformation. Eight of the 36 AH244 colonies were subcultured, and the genomic DNA was isolated to assess integration of the knockout construct into the WT nuoD gene by PCR analysis. PCR analysis was performed using primers CF-1 and CR-5 to amplify the region on either side of the aphA-3 gene. All eight kanamycin-resistant strains contained an intact WT nuoD (data not shown). These results suggested that nonspecific integration had occurred. Further, PCR analysis of the transformation mixture, prior to plating onto selective media, revealed that site-specific integration into nuoD had occurred in both lots of cells (J99 and AH244), but no viable colonies resulted with this genotype. This provided additional evidence that this recombination event was lethal. The conclusion drawn from the above results was that nuoD (or the nuo operon) is essential in H. pylori. Since nuoD is contained within an operon, we cannot rule out possible polar effects within the operon.
MIC determinations for H. pylori strain J99, the four classes of compound A-resistant mutants, and E. coli strains to compound A, piericidin A, and rotenone. MICs were determined for WT H. pylori strain J99, each of its four classes of compound A-resistant mutants, and E. coli strains ECO523 and ECO524 (ECO523 with TolC– mutation) using compound A, piericidin A, and rotenone. Rotenone and piericidin A are known complex I inhibitors that have been suggested to interact with NuoD in Rhodobacter capsulatus (12, 13, 18). The MICs for the four mutant classes treated with compound A were consistent at 2 µg/ml, while E. coli strain ECO523 and the isogenic TolC– strain were insensitive to the highest concentration used (4 µg/ml) (Table 3). The TolC– strain was used to address whether efflux may play a role in E. coli resistance. Several additional BI-containing compounds were screened against WT J99 and the four classes of resistant mutants, and cross-resistance was consistently observed (data not shown). In the case of rotenone, the three classes of nuoD mutants were hypersensitive, with MICs of <0.125 µg/ml, while the WT and nuoB mutant had the same MIC of 1 µg/ml (Table 3). Both E. coli strains were insensitive to rotenone at 4 µg/ml (Table 3). WT H. pylori strain J99 and the nuoB mutant were sensitive to piericidin A, with an MIC of 0.008 µg/ml, while the three nuoD mutants were two- to fourfold less sensitive (Table 3). Both E. coli strains were insensitive to piericidin A at 4 µg/ml, the highest concentration tested (Table 3). Growing E. coli under microaerophilic conditions in the presence of compound A, piericidin A, and rotenone yielded the same results (data not shown).
|
|
DISCUSSION |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Dupuis et al. have established R. capsulatus complex I as a model system for studying this multisubunit enzyme (14). Relevant to the present study, they have isolated mutants of R. capsulatus resistant to piericidin A (a potent inhibitor of complex I that binds close to or at the ubiquinone-binding site) that have a single missense mutation in nuoD resulting in an amino acid change (V407M) at the C terminus of the protein (9, 10). The V407M change maps to the same region that the mutations in H. pylori NuoD were observed in the present study (Fig. 3). The piericidin A-resistant mutants were also resistant to rotenone, a commonly used pesticide that has been shown to bind and inhibit complex I at the ubiquinone-binding site (9, 10, 12). In a follow up study, Prieur et al. constructed additional nuoD mutations by site-directed mutagenesis (29). One additional mutant was resistant to rotenone and piericidin A (D412E) (Fig. 3). In contrast, mutations within the analogous region of H. pylori NuoD resulted in hypersensitivity to rotenone but moderate resistance to piericidin A and compound A (Table 3). Based on the experiments in R. capsulatus, the authors proposed that the quinone polar head binds to the NuoD subunit (13, 29). In addition, they suggested that the binding site for the polar head of ubiquinone is located at the interface of NuoD and NuoB. This was, in part, based on photoaffinity labeling of NuoB from membranes of Paracoccus nitrificans and Thermus thermophilus with a 3H-labeled pyridaben analog (31). In the same study, both rotenone and piericidin A prevented binding of the 3H-labeled pyridaben analog (another pesticide) (31). This is compatible with the results of the present study, where compound A-resistant mutants had missense mutations in either the C terminus of nuoD or in nuoB. We speculate that NuoD and NuoB may provide a binding interface for compound A.
Studies with inhibitors have shown that E. coli complex I is extremely insensitive to rotenone, while complex I from R. capsulatus was notably sensitive (12, 18). These findings are consistent with our results showing that two strains of E. coli were insensitive to rotenone (and compound A), while H. pylori was sensitive to both of these compounds (Table 3). It may be that BI selectivity for H. pylori is due to an insensitivity of complex I from most other bacteria to the BI-containing compounds, much like E. coli complex I is insensitive to rotenone. A second possibility to explain selectivity is that complex I is not essential in most other bacteria; therefore, even if inhibited, the bacteria continue to grow. There is significant evidence that complex I is not essential in several organisms, such as E. coli (37), Salmonella enterica serovar Typhimurium (38), and R. capsulatus (15). In these studies, nuo was not essential, likely because there is an additional NADH dehydrogenase (Ndh II) that can compensate for inhibition of complex I. Ndh II and nuo dual knockouts have been demonstrated to be viable but more limited in growth rate (4, 5). In the present study, nuoD (or the nuo operon) was confirmed to be essential for the viability of H. pylori strains AH244 and J99 by knockout mutagenesis. There is no homolog to Ndh II in the H. pylori genome, suggesting that H. pylori is more limited by a lack of alternative cellular machinery (32).
In the present study, we found the mutation frequencies for H. pylori strains J99 and AH244 challenged with compound A to be low, ranging from 3.1 x 10–8 to 1.3 x 10–10. The compound A MICs for the resistant mutants were only moderately higher (4 times higher) than that observed for the WT J99 strain (Table 3). The hypothesis that the BIs may bind at the same site as ubiquinone and known complex I inhibitors, such as rotenone, is attractive based on the results of this study. There are several parallels to be drawn from the literature, especially regarding the R. capsulatus model, including the involvement of NuoD and NuoB. Alternatively, complex I may serve to alter the BIs into active anti-H. pylori agents, much like RdxA has been shown to be involved in converting metronidazole into an active anti-H. pylori drug (19). Complementary biochemical studies are needed to ascertain whether complex I is the target of the BIs in addition to being involved in resistance.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| Clin. Vaccine Immunol. | Clin. Microbiol. Rev. |
|---|---|
| J. Clin. Microbiol. | ALL ASM JOURNALS |
