Pyrosequencing assay to rapidly detect clarithromycin resistance mutations in Canadian Helicobacter pylori isolates

1Children’s Hospital of Eastern Ontario, University of Ottawa, Ottawa, Ontario; 2EpigenDx Inc, Worcester, Massachusetts, USA; 3Dalhousie University; 4Queen Elizabeth II Health Sciences Centre, Halifax, Nova Scotia; 5University of Alberta, Edmonton, Alberta Correspondence: Dr Robert Slinger, Infectious Diseases/Medical Microbiology, Children’s Hospital of Eastern Ontario, 401 Smyth Road, Ottawa, Ontario, K1H 8L1. Telephone 613-737-7600 ext 2651, fax 613-738-4832, e-mail slinger@cheo.on.ca Received for publication July 31, 2008. Accepted December 2, 2008 Clarithromycin is a component of first-line therapy regimens for Helicobacter pylori infections. However, the prevalence of pretreatment resistance to this antibiotic appears to be increasing, potentially due to the widespread use of macrolide therapy for respiratory tract infections. Resistance rates vary widely by region, from 0% to 50% worldwide (1). This in vivo resistance is clinically important because it has been shown to greatly increase the risk of treatment failure when clarithromycin-containing regimens are used (2,3). However, despite the clinical importance and increasing prevalence of clarithromycin resistance in H pylori strains, susceptibility is not generally determined before treatment. This is partially due to the fact that conventional clarithromycin antimicrobial susceptibility testing for H pylori by agar dilution or E test (AB Biodisk, Sweden) is a relatively slow process that takes 72 h to perform. Clarithromycin acts by binding to the 23S ribosomal RNA (rRNA) gene, thereby blocking protein synthesis. Resistance to this agent is most commonly caused by point mutations at positions 2142 or 2143 in either one or both copies of the 23S rRNA gene, which then leads to altered macrolide binding (1). Because of the difficulties involved in performing traditional susceptibility testing with bacterial cultures, and because resistance is highly associated with mutations in these two adjacent positions, rapid genotypic susceptibility determination using molecular diagnostics has become an attractive and feasible option. Rapid detection and sequencing of mutations in the 23S gene from DNA extracted from H pylori cultures may potentially lead to improved antibiotic selection for patients infected with H pylori, and also provide epidemiological information original artiCle

regarding the types of mutation(s) present in a population.DNA sequencing is considered the 'gold standard' for mutation detection, but is rarely used in the clinical microbiology laboratory setting due to issues of time, cost and lack of familiarity with techniques.
Recently, pyrosequencing, a rapid method of sequencing relatively short DNA targets, has been used in a number of microbiological applications (4)(5)(6)(7)(8).This is a novel method of sequencing that potentially has advantages over conventional sequencing when only a short target sequence is required because it can be less costly and more rapid.
Pyrosequencing is a polymerase chain reaction (PCR)-based method of DNA sequencing, in which primers must be biotin labelled at the 5′ end for capturing the PCR product with streptavidin-coated beads.Once captured, single-stranded DNA is prepared by denaturation with sodium hydroxide.A pyrosequencing primer is then added.This primer anneals to the single-stranded DNA bound to the beads.DNA synthesis then begins at the 3′ end of the primer and proceeds in a manner in which the addition of a nucleotide to the DNA strand results in the generation of light via a series of enzymatic reactions.This is known as 'sequencing by synthesis'.
On incoporation of a deoxynucleotide, the primer is extended by DNA polymerase and an inorganic phosphate (PPi) is released.ATP sulfurylase is then used to convert the PPi to ATP.The reaction also contains the substrate luciferin and a luciferase enzyme that uses the ATP to produce light.This is a stoichiometric reaction, and the amount of light produced is proportional to the number of PPi molecules produced and hence the number of incorporated nucleotides.Any unincorporated nucleotide is degraded with apyrase and the next nucleotide is then added, and the series of reactions is repeated.
Nucleotides are added at a rate of approximately 1 nucleotide/min.Following the addition of each nucleotide, the light emitted from each individual reaction well is detected by a camera and visualized in real time on a computer monitor.Software then analyzes the peak heights and determines the sequence.When more than one copy of a gene is present, as in the case of H pylori 23S rRNA genes, pyrosequencing has been shown to be able to clearly determine the number of copies of each variant present (9).
In Canada, the mutations responsible for clarithromycin resistance in H pylori strains are not well described; consequently, we decided to develop a pyrosequencing assay to provide this information.A previous publication (10) described the pyrosequencing of a small number of clarithromycin-resistant H pylori strains.We designed new PCR primers for a short 23S target region in H pylori to allow for a rapid PCR protocol, and used these with the published sequencing primer.To increase assay speed, a primer designed to bind directly adjacent to position 2142 of the H pylori 23S gene was used, thereby expediting the sequencing of only the minimum essential number of bases.We then tested a collection of Canadian H pylori isolates to investigate which mutations were most prevalent.

METHODS
Forty-two H pylori isolates were studied, the majority of which were from patients in Nova Scotia.Clarithromycin susceptibility of these isolates was determined using an E-test method (11) that included 24 clarithromycin-resistant and 18 clarithromycinsusceptible H pylori strains.A clarithromycin-susceptible H pylori strain (American Type Culture Collection reference 43504) was used as a control.
DNA was extracted from isolated colonies of H pylori in pure culture using the PrepMan Ultra reagent (Applied BioSystems, USA) according to the manufacturer's protocol.The DNA extracts were stored frozen at -80°C until PCR was performed.DNA concentrations were estimated by measuring optical densities.DNA extracts were diluted with water and a DNA concentration of 50 pg/µL was used in the PCR reaction.
Primers were designed using the Primer 3 program (http:// frodo.wi.mit.edu/primer3) to amplify a short 23S target region that included the binding site for the previously described pyrosequencing primer and positions 2142 and 2143.A short amplicon was chosen to allow for efficient rapid-cycle PCR, and for the sequencing of only the two target positions of interest.
A 56 bp fragment of the 23S gene was amplified in the PCR reaction using commercially synthesized primers (Integrated DNA Technologies Inc, USA): the forward primer HpF-GAGGTGAAAATTCCTCCTACCC and the reverse primer HpR Biotin-AGTAAAGGTCCACGGGGTCT.
The rapid PCR protocol consisted of 45 cycles of 94°C denaturation, 58°C annealing and 72°C extension, with a programmed transition rate of 9.9°C/s with no holds.After PCR, samples were transferred to a high-resolution melting instrument (HR-1, Idaho Technology, USA) for melting curve analysis.The derivative curve of the measured fluorescence plotted against temperature was examined to determine whether amplification was successful.
The pyrosequencing assay was optimized using the PSQ96MA system (Biotage, Sweden).Reagents were mixed together in preparation for the sequencing reaction according to the pyrosequencing sample preparation protocol.A forward pyrosequencing reaction was performed using the sequencing primer 5′-GCGGCAAGACGG-3′, at a concentration of 0.5 µmol/L.

RESULTS
All of the 42 H pylori isolates were successfully amplified and sequenced using the procedures described.Representative pyrograms for clarithromycin-resistant and -susceptible strains are presented in Figure 1.Sequencing results and clarithromycin minimum inhibitory concentration (MIC) range and susceptibility status for the isolates are shown in Table 1.
Mutations at one of the two positions studied were present in 20 of 24 (83%) clarithromycin-resistant isolates.Seventeen isolates had identical mutations in both copies of the 23S gene: 13 with double adenine (A) to guanine (G) (A2143G) mutations and four with double A2142G mutations.Three resistant isolates had a mutation in only one copy of the 23S gene (all were A2143A/G).Thus, an A to G mutation in either one or both copies of the 23S gene at position 2143 was present in 16 of 24 (66%) resistant isolates.
There were no mutations in 17 of 18 (94%) susceptible isolates.A single-copy mutation was detected in one susceptible isolate (A2142A/G).
Procedure times were short.Rapid-cycle PCR and amplicon detection were completed in approximately 30 min.Pyrosequencing procedures were completed in approximately 20 min.

DISCUSSION
The pyrosequencing assay developed in the present study was able to detect and differentiate mutations at positions 2142 and 2143 in either one or both copies of the H pylori 23S rRNA gene.A high proportion of resistant isolates had mutations in the assay target region.As reviewed by Gerrits et al (2), the A2143G mutation is linked to intermediate-level clarithromycin resistance (MIC of 2 µg/mL to 64 µg/mL), and the A2142G mutation is associated with high-level clarithromycin resistance (MIC greater than 64 µg/mL).Although the number of strains tested was small, this association appears to exist with our isolates.The MIC range was 2 µg/mL to more than 256 µg/mL for our 13 dual-copy A2143G isolates, with a MIC required to inhibit the growth of 50% of organisms (MIC 50 ) of 16 µg/mL.For the four A2142G mutants, the MIC range was 32 µg/mL to more than 256 µg/mL, with a MIC 50 of greater than 256 µg/mL.
Clarithromycin resistance not due to mutations at positions 2142 or 2143 was seen in 17% of the isolates.However, because this was primarily a feasibility study to determine whether a pyrosequencing assay could be developed and used successfully for genotypic analysis of H pylori, rather than an epidemiological study, the proportion of resistance due to mutations at positions 2142 and 2143 of the 23S gene in Canadian H pylori isolates cannot be accurately estimated from our data.In large studies, resistance due to these two mutations typically accounts for more than 90% of clarithromycin resistance.A prospective Canadian study to establish the pretreatment prevalence of clarithromycin resistance in H pylori is warranted and could help determine whether current empirical treatment regimens need to be modified.
Further investigation is also needed to determine the mechanism of resistance of the four resistant isolates in our collection that lacked mutations at positions 2142 or 2143.Mutations at other 23S rRNA positions are most probably responsible (1,2), and these mutations could possibly be investigated with additional pyrosequencing assays or by conventional sequencing.
Development of pyrosequencing assays for the detection of mutations in H pylori that lead to resistance to certain other antibiotics should also be possible.For example, high-level tetracycline resistance in H pylori is mediated by a triple basepair substitution AGA926-928→TTC in the 16S rRNA gene, and resistance to fluoroquinolones is caused by point mutations in the DNA gyrase-encoding gene gyrA at amino acid positions 87, 88, 91 and 97 (2).
Additional studies are needed to determine whether the PCR/pyrosequencing method can be applied after the direct amplification of DNA from clinical specimens such as stools and gastric biopsies, as is performed with other nonculturebased molecular methods such as PCR with fluorescent probe detection of clarithromycin resistance (12,13).Although pyrosequencing requires an additional step after PCR, it does provide gold standard results because the sequence itself is obtained.This is not the case with fluorescent PCR methods, in which the possibility of false-positive or false-negative results exist.
Zhou et al ( 14) performed a cost comparison as part of a study examining resistance mutations in malaria parasites.They demonstrated pyrosequencing to be less expensive and more rapid than conventional sequencing.If assays are designed in which valuable clinical information can be obtained from short DNA sequences, as in the present study, pyrosequencing may gain wider acceptance for use in clinical microbiology laboratories.

NOTE:
This work was presented as an abstract at the Association of Medical Microbiology and Infectious Disease Canada and Canadian Association for Clinical Microbiology and Infectious Diseases Annual Conference, Halifax, Nova Scotia March 14-18, 2007.

Figure 1 )
Figure 1) Pyrosequencing pyrograms of the point mutations at positions 2142 and 2143 of the Helicobacter pylori 23S ribosomal RNA gene associated with clarithromycin resistance.Adenine (A) bases were dispensed followed by guanine bases in a sequential manner to create distinct pyrograms.The order of bases added is shown on the x-axis and light intensity is on the y-axis.The height of the base peaks is proportional to the number of bases of that type that occur at any given base position.A Clarithromycin-susceptible strain showing A at positions 2142 and 2143.B At position 2143, both alleles of the 23S rRNA gene show an A to guanine (G) mutation (ie, A2143G).C At position 2142, both alleles of the 23S rRNA gene show an A to G mutation

Table 1 Pyrosequencing results and clarithromycin susceptibilities for Canadian Helicobacter pylori isolates 23S gene nucleotide position Isolates, n Clarithromycin susceptibility
A Adenine; G Guanine; MIC Minimum inhibitory concentration; R Resistant; S Susceptible