TEXT

Top Abstract Text References

Rifabutin and some other rifamycin derivatives inhibit the growth of Helicobacter pylori in vitro at much lower concentrations than rifampin and might be possible candidates for second- or third-line eradication therapy (1, 2). It has been shown that triple therapy containing rifabutin is effective in the eradication of H. pylori after failure of other therapies and in spite of resistance to other antibiotics (10).

The target of all rifamycins is the -subunit of the DNA-dependent RNA polymerase encoded by the rpoB gene (4, 5). Amino acid substitutions resulting from changes in codons 507 to 533 of rpoB in mycobacteria (8, 9), codons 146, 507 to 534, 563 to 572, and 687 in Escherichia coli (4, 6, 11), and codons 473 to 483 and 528 to 530 in Listeria monocytogenes (7) induce resistance.

Resistant mutants of H. pylori ATCC 43504 selected in vitro by serial passage in the presence of rifampin all showed mutations in codons 525 to 545 or codon 586 (2). Recently, we described a clinical isolate of H. pylori that developed resistance during therapy and that harbored a mutation at codon 149 (V149F), corresponding to the V146F mutation observed in E. coli (6). Consequently, we detected a homologous mutation (V176F) in three clinical isolates of Mycobacterium tuberculosis (3).

In this study, we tried a compilation of possible mutations in H. pylori at four different regions of rpoB which are highly conserved between different species that are susceptible to rifamycins. Codon 149 was randomized by site-directed mutagenesis. Arginine 701, as encoded by E. coli codon 687, was replaced with histidine. The levels of resistance of these new mutants were determined and compared to the resistance levels of the panel of H. pylori ATCC 43504 variants with mutations in codons 525 to 545 (cluster I) or in codon 585 (cluster II).

Culture, susceptibility testing, and transformation conditions were described recently (2, 3). H. pylori suspensions adjusted to McFarland 2 were tested with a semiautomatic multipoint inoculator (Multipoint AD; Mast) in the agar dilution assay. Rifabutin and rifampin concentrations ranged from 0.002 to 256 µg/ml. Type strains were included for quality control.

H. pylori 2802A is a clinical isolate from a patient with duodenal ulcers and was found to be competent for natural transformation with our system, in contrast to H. pylori ATCC 43504. DR62a and DR62n are paired isolates obtained before and after treatment with rifabutin (3). Primers for amplification, sequencing, introduction of restriction sites for cloning in pMin1, site-directed mutagenesis (V149X and R701H), and final amplification before transformation are shown in Table 1. All base pair and codon designations correspond to the published H. pylori sequence (12).

PCR products of the wild-type rpoB (bp 54 to 916 and bp 1764 to 2423) were cloned into vector pMin1 using the BglII and XhoI restriction sites introduced by the respective PCR primers (Table 1). Plasmid preparation was done using Wizard Miniprep (Promega).

After the discovery of the mutation encoding V149F in the rpoB gene of a clinical isolate of H. pylori, we further evaluated the capacity of this codon to induce resistance. The rpoB gene of H. pylori 2802A was randomly mutagenized at codon position 149. Basically, the mutagenesis was performed as depicted in the ExSite PCR-based site-directed mutagenesis kit (Stratagene). The mutation of interest was generated by PCR using primers (Table 1) carrying the desired mutation (R701H-Rev) or the randomized codon (V149X/Muta-for). The resulting plasmid was used as a template for PCR with the original primers for the respective linear DNA fragments of the rpoB gene (bp 54 to 916 and bp 1764 to 2423). These were directly transformed into H. pylori 2802A, which was subsequently cultured on selective agar.

For the evaluation of the random mutagenesis, single colonies were picked and expanded on selective agar for MIC determination (n = 80). Subsequently, strains with different levels of resistance and colony size after transformation were chosen for sequencing (n = 40). We were able to detect seven different replacement amino acids at codon 149, associated with different resistance levels. Whether the remaining possible mutations did not induce resistance, were lethal, or were just missed by the chosen panel is unknown. Several replacement amino acids were encoded by different (up to four) triplets. Independent strains carrying the same mutation showed similar resistance levels (Table 2). Only two of the detected codon exchanges could occur in vivo by point mutation. One of these, encoding the V149F mutation, represented the largest portion in our panel of resistant clones. This is in agreement with our findings for the clinical isolate DR62n and for spontaneously resistant mutants after transformation with wild-type DNA.

One mutation encoded by codon 148 (V148I) could be detected in combination with V149F but was associated with a lower rifabutin MIC (2 µg/ml) than that for V149F alone (~32 µg/ml) (Table 2). Isoleucine is located at the homologous position in E. coli, Staphylococcus aureus, and L. monocytogenes.

Transformation of H. pylori 2802A with specific PCR products (bp 1371 to 2106) containing the cluster region of different resistant ATCC 43504 variants (2) resulted in rifabutin MICs similar to those for the parent strains (Table 3). The association of the exchanged amino acid and resistance levels could be confirmed for all mutations between codons 525 and 545 and for two different mutations at codon 586. Differences in the MICs of rifabutin between the donor strain (resistant variants of H. pylori ATCC 43504) and the recipient strain (H. pylori 2802A) of more than 1 dilution remain unexplained. We repeated the transformation experiments in cases of discrepant MICs (for D530N and I586) several times with well-defined fragments. The resulting MICs for the recipient strain were all in the same range. Transformation of the respective mutations into H. pylori 26695 (12) resulted in very similar MICs of rifabutin (Table 3). Additional mutations might have occurred in the donor strains during serial passage.

The mutation R701H was associated with slower growth of the H. pylori 2802A variant on selective agar. The single mutation R701H induced an up to 32-fold elevation of the MIC of rifabutin for H. pylori. Culture on selective agar also induced numerous (50 to 200) larger colonies with high-level resistance (2 to 32 µg/ml), indicating additional mutations in the cluster region. The MIC of rifabutin for the primary (background) culture after transformation (10⁵ to 10⁶ CFU) ranged from 0.065 to 0.25 µg/ml (Table 2). The occurrence of highly resistant large colonies was more pronounced during selection with rifabutin at 0.032 µg/ml than at 0.01 µg/ml. Further studies might show whether there is synergism between different low-level resistance mutations.

All observed mutations, even if they were inducing only slight increases of the rifabutin MIC (525, 530, and 586), had a distinct impact on the MICs of rifampin (>64 µg/ml). E-test assays are used for evaluation of clinical isolates. The rifampin MIC (E-test) for H. pylori 2802A (recipient, wild type) was 0.25 µg/ml. MICs for wild-type isolates generally ranged from 0.032 to 2 µg/ml.

The site-directed mutagenesis technique was shown to be helpful for the evaluation of possible loci for mutational resistance. Our findings might contribute to a better understanding of conformation and function of the RNA polymerase and the interaction with rifamycins.