Article Text
PDF
Abstract
Objective To determine in Neisseria gonorrhoeae (NG) isolates from different geographical areas whether monitoring of major determinants involved in chromosomal antimicrobial resistance correlated with phenotypes and could constitute complementary tools for surveillance.
Methods Real-time multiplex PCR assays targeting penA, mtrR, penB, ponA, gyrA and parC determinants were applied to 169 NG extracts. Minimum inhibitory concentrations for penicillin and ciprofloxacin were determined by E tests, and β-lactamase production was analysed using nitrocefin discs.
Results A total of 169 NGs were examined, 110 from New Caledonia, 44 from Madagascar and 15 from Cambodia. Despite the heterogeneity in the number of isolates tested, the susceptibility trends observed in the different geographic areas studied showed a good fit with the multigene genotypes. In addition, features related to a specific geographical diversity were found: (1) a high prevalence of strains harbouring the porB1a allele and showing reduced penicillin susceptibility in Madagascar and Cambodia (39% and 40% respectively); (2) almost all strains from Cambodia were resistant to the drugs tested (11/15 and 14/15 resistant to penicillin and ciprofloxacin respectively); and (3) identification of novel penB and mtrR genotypes associated with a moderately decreased penicillin susceptibility in New Caledonia (mtrR novel genotype in 47% of intermediate vs 14% of susceptible isolates).
Conclusions Showing a good correlation with phenotypic trends of susceptibility, multiplex real-time PCR assays could be used successfully for prospective epidemiological studies notably by characterising mtrR and penB determinants for their fundamental and complementary roles in increasing the antibiotic resistance. These molecular tools could also provide useful alternative surveillance tools for non-viable strains.
- Antibiotic sensitivity
- gonorrhoea
- molecular typing
- screening
Statistics from Altmetric.com
Neisseria gonorrhoeae (NG) is the causative agent of gonorrhoea, one of the major sexually transmitted infections in many countries. Its remarkable ability to evolve resistance mechanisms threatens the use of most antimicrobial agents so that attention in an effective surveillance to limit the spread of multiresistant strains has become an absolute necessity.1
Penicillin resistance arises either as a consequence of antibiotic-degrading enzymes production2 or from simultaneous genetic variations in endogenous genes and regulatory regions resulting in an increasing resistance to structurally unrelated groups of drugs.3 4
Reduced affinity of penicillin-binding proteins (encoded by penA and ponA), overexpression of the MtrC–MtrD–MtrE efflux pump due to mutation in the promoter of mtrR and decreased permeation of the outer membrane induced by mutations in the porin PorB1b are the commonly described mechanisms of chromosomally mediated resistance.5–12 These mutations, referred to as penA, ponA, mtrR and penB determinants, have been widely observed in clinical isolates displaying decreased susceptibility. In vitro transformation experiments demonstrated that their phenotypic expression was leading to a gradual increase in minimum inhibitory concentration (MIC) when successive variations accumulate.6 11 12
Resistance to fluoroquinolones in NG involves mutations in the quinolone resistance-determining regions (QRDRs) of gyrA and parC.13–15 Mutations in gyrA are considered as the first and most frequent, leading to either a less susceptible or a resistant phenotype, whereas parC changes constitute later steps always leading to complete resistance.16 17 Efflux activation or decreased porin permeability has also been suspected of being involved in resistance to quinolones.16 18
Using real-time PCR with hybridisation probes, we genotyped these chromosomal determinants in NG isolates from different geographic areas (New Caledonia, Oceania; Madagascar, Africa and Cambodia, Asia). We then evaluated the correlation between these genetic variations and the susceptibility to the corresponding antibiotics in order to evaluate the predictive value of these molecular tools and their possible use in surveillance as a way to forecast forthcoming resistance and adapt recommendations.
Materials and methods
NG isolates
A total of 169 NG isolates from men or women were studied. They were isolated at the bacteriology laboratories of (1) Institut Pasteur of New Caledonia (110 isolates collected from October 2006 to October 2007); (2) Institut Pasteur in Antananarivo, Madagascar (47 isolates collected from March 2004 to March 2006); (3) Institut Pasteur in Phnom Penh, Cambodia (15 isolates collected over 2006). Most were obtained from persons attending a sexually transmitted disease clinic and collected from cervical or urethral swabs. Four Malagasy isolates were obtained from systematic control in risk populations. None was isolated twice from the same patient. The species identification was confirmed by Gram staining, oxidase activity and specific characteristics on API NH (BioMérieux) after culture on chocolate agar supplemented with PolyVitex plus colistin, vancomycin, amphotericin B and trimethoprim. After antibiotic susceptibility testing, Cambodian and New Caledonian isolates were suspended into 2SP transport medium and stored at −20°C until DNA extraction as described previously.19 Malagasy isolates were stored in skimmed milk at −80°C for a subsequent recultivation before DNA extraction.
Ten NG reference strains (see table 1) were used as controls for antimicrobial susceptibility testing.
Reference strains used in this study (kindly provided by J Tapsall): genotypes; penicillin and ciprofloxacin susceptibility profiles
Antimicrobial susceptibility determination
MICs for penicillin G and ciprofloxacin were determined using the E test diffusion method (AB Biodisk) on chocolate agar plus PolyVitex by reading the intercept of the antibiotic gradient strip and the zone of inhibition. Breakpoint criteria defined by the CLSI20 for NG were followed: for penicillin—susceptible strain, MIC ≤0.06 μg/ml, intermediate 0.125–1 μg/ml and resistant strain, MIC of ≥2 μg/ml. β-Lactamase production was analysed using nitrocefin discs (BioMérieux), and strains testing positive were referred to as penicillinase-producing NG (PPNG). For ciprofloxacin, isolates were considered as susceptible when MICs were ≤0.06 μg/ml, as intermediate when MICs were 0.125–0.5 μg/ml and as resistant when MICs were ≥1.0 μg/ml. MICs for New Caledonian and Cambodian isolates were determined at the Institut Pasteur in New Caledonia, whereas MICs for Malagasy isolates were determined in Madagascar.
DNA extraction and genotyping assays
Genomic DNA was isolated from 0.2 ml of a thawed bacterial suspension in 2-sucrose phosphate using the QIAamp DNA minikit (Qiagen) according to the manufacturer's instructions.
All primers and probes were designed using LightCycler Probe Design software V.2.0 (Roche Diagnostics), and synthesised by Sigma-Proligo (Singapore Pty). The additional aspartic acid codon (D345a) in penA, the single-base substitution in ponA (L421P), the −A deletion within mtrR promoter and porB1b characterisation (penB determinant) were identified using two duplex PCR described previously,19 21 on a LightCycler 2.0. Genotyping for +T insert within mtrR promoter was achieved with newly designed probes combined with those previously described for genotyping of G45 codon using mtrR-F and mtrRcod-R primers.19 Duplex genotyping of gyrA and parC QRDRs was also described recently.22
Each PCR mixture contained 1× Roche LightCycler FastStart DNA Master HybProbe mix, 4 mM of MgCl2, 0.5 μM of each primer (1.0 μM for penA/ponA duplex and penB assays), 0.2 μM of each probe, 2 μl of template DNA and PCR grade water to a final volume of 20 μl. The amplification parameters were as follows: an initial enzyme activation step at 95°C for 10 min, followed by 50 sequential cycles of denaturation at 95°C for 8 s, primers and probes annealing at 55°C for 8 s and elongation at 72°C for 30 s (50 s for mtrR +Tinsert/G45D duplex PCR). Subsequently, the melting curves analysis allowed the identification of the different genotypes after a compensation colour had been activated as described previously.19 22 Controls in every experiment included a blank capillary, wild-type DNA and mutated genotype DNA of each known sequence. All genotyping assays were developed and implemented at the Institut Pasteur in New Caledonia.
Sequencing of the mtrR promoter region
A panel of 10 DNAs, representative of either the −A deletion or the +T insertion in the mtrR promoter according to their melting temperature, was amplified with primers mtrR-F and mtrR-R19, purified using the MinElute PCR Purification kit (Qiagen) and sequenced by Waikato University (Hamilton, New Zealand) using the same primers.
Results
Genotyping results
The genotype distribution of the different isolates is summarised in tables 2, 3, together with their corresponding phenotypes. From 110 New Caledonian strains, 53 harboured the D345a codon in penA, whereas 57 were penA wild-type (WT). Regarding mtrR, no −A deletion was observed but 30 out of 110 isolates had modified mtrR promoter sequences harbouring the +T insertion. Analysing the coding region of mtrR, no G45 mutation was found. Genotypes for penB were more diverse: 52 were WT (G101/A102), seven were G101D, a penicillin-susceptible genotype, and 33 presented an A102S mutation from which 32 were also penA mutated. The 18 remaining strains contained the porB1a allele confirmed by a 692 bp specific amplicon on agarose gel. Lastly, 90 isolates were ponA WT and 20 ponA L421P. Among the latter, 18/20 were penA WT. Finally, none of the New Caledonian strains presented modified gyrA or parC QRDRs.
Number of isolates with their corresponding genotypes and penicillin susceptibility from New Caledonia, Madagascar and Cambodia
Number of isolates with their corresponding genotypes and ciprofloxacin susceptibilities identified in isolates from New Caledonia, Madagascar and Cambodia
Out of 47 Malagasies isolates, 37 bore the D345a codon in penA, nine were WT, and one was non-typable. No mtrR mutation was found. For the penB determinant, 21 strains were characterised as WT porB1b, three presented an A102S mutation, and 17 presented a porB1a allele. Three strains were penB non-typable, including the discrimination between alleles porB1a or porB1b. The ponA determinant occurred in 10 isolates, whereas 34 were WT. In the whole Malagasy collection, no change was observed in either gyrA or parC QRDRs.
All the 15 Cambodian strains contained mutated penA and ponA sequences. Regarding penB determinant, G101K/A102D (KD) and G101K/A102N (KN) genotypes were evidenced in five and four isolates respectively. These nine strains harboured all the determinants described as involved in chromosomal penicillin resistance (penA, mtrR, penB and ponA), one strain presented both −A deletion within promoter and a G45D mutation in the coding region of mtrR. The remaining six isolates had a porB1a allele with four WT and two mutated mtrR. Concerning variations within gyrA and parC, only one isolate was characterised as WT for both QRDRs, whereas genotype distribution for the 14 others was as follows: (1) for gyrA (S91F/D95A (8/14), S91F/D95G (3/14) or S91F/D95Y (3/14)), (2) for parC, three strains had a WT QRDR, nine were single mutated (E91G (4); S87N (2); S87R (3)), and three showed the double mutation S87N/E91Q.
Correlations between phenotypes and genotypes
From New Caledonia, 65 isolates were susceptible to penicillin (59%) and 45 strains intermediate (41%) with MICs ranging from 0.064 to 0.25 μg/ml. Out of these 65 susceptible ones, the penA determinant was present in all strains with MIC ≥0.047 μg/ml. All 18 isolates from New Caledonia that had a ponA L421P together with a WT penA genotype were susceptible to penicillin, illustrating the multigenic determinism of chromosomal resistance to penicillin. Similarly, all non-PPNG strains among Malagasy and Cambodian panels with MIC ≥0.032 μg/ml also harboured the penA determinant. The Malagasy panel contained 12 isolates (27%) resistant to penicillin (all were PPNG), 24 (55%) intermediate and 8 (18%) susceptible. The Cambodian panel included 11 isolates (73%) resistant to penicillin (with 10 PPNG) and four (27%) intermediate with high MIC values (ranging from 0.5 and 1.5 μg/ml).
Regarding the mtrR determinant, the −A deletion was not found in any New Caledonian or in any Malagasy isolate. In New Caledonian isolates, the most prevalent sequence feature was the +T insert in mtrR promoter present in 27% of the isolates. This genotype was detected in 21/45 (47%) intermediate but only in 9/65 (14%) susceptible strains. Conversely, all non-PPNG from Cambodia presented the −A deletion, their MICs to penicillin ranging from 0.5 to 2.0 μg/ml. For penB, we identified genotypes that correlated with geographical origins: while WT or G101D was predominantly present in susceptible isolates, the A102S genotype was evidenced in 60% (27/45) of the intermediate strains from New Caledonia with MIC around 0.15 μg/ml. This latter genotype, which does not confer a phenotype as strong as KD or KN,19 was rarely found in Madagascar isolates and absent from Cambodian isolates. On the other hand, a high prevalence of isolates with a porB1a allele (ca. 40%) characterised the Cambodian and Malagasy panels and included strains with increased MICs.
All resistant non-PPNG presented a KD or KN penB genotype in agreement with their phenotypes. Both genotypes appeared to be associated with similarly increased MIC. The ponA determinant appeared more widespread than the −A deletion in mtrR or KD and KN genotypes in penB, being present in isolates from all three origins. It was significantly associated with increased resistance only in Cambodian isolates that also harboured determinants penA, mtrR (−A deletion) and penB (KD or KN genotypes).
A good correlation between QRDRs genotypes and fluoroquinolones susceptibility phenotypes was observed, as illustrated in table 3. No QRDR mutation was reported in any susceptible strain over the three areas, whereas mutations in gyrA QRDR were evidenced (except in one Malagasy strain) in every strain resistant to ciprofloxacin. Finally, correlated with increasing MICs, single or double mutations in parC QRDR were highlighted.
Discussion
The aim of this study was to evaluate the pertinence of genotyping specific determinants involved in chromosomally mediated resistance to penicillin and ciprofloxacin by molecular methods, a technique that is not yet routinely used.
Though genotypes cannot always be accurately related to MICs, many studies proved that some key determinants play an unambiguous role in chromosomal resistance to structurally unrelated drugs. The main mechanism described in NG multidrug resistance is the overexpression of an efflux pump due to an −A deletion in the promoter of mtrR, the repressor of this MtrCDE pump. This mutation is required for penB determinant to increase the level of resistance by decreasing the permeation of hydrophobic agents. These genetic variations act in synergy in a complex interplay including penA and ponA determinants, these latter leading to modified penicillin-binding proteins. Modifications in QRDRs of gyrA and parC are involved in fluoroquinolone resistance. These mutations are associated with different levels of resistance6 so their detection should provide useful susceptibility indications.
Overall, the determinants' profiles evidenced in our work were consistent with the mechanisms described in chromosomal resistance to penicillin and fluoroquinolones, with only 4/169 (2.4%) strains that had each one non-typable determinant. The difficulties in genotyping these four isolates from Madagascar likely highlight sequence changes within the targets. It should also be noted in this panel that two PPNG strains have a lower MIC than expected in such strains. Because of the difficulties associated with the strains' preservation in Madagascar, we could not verify these MICs. Excluding PPNG and although all genotypic combinations were not present in our panel, the data observed in this study suggest that monitoring specific mutations in multiple genes could be a good predictive tool of overall resistance in gonococci from different geographical origins. The penA determinant was present in 89% of non-PPNG isolates with penicillin-MIC ≥0.047 μg/ml confirming this variation as the first step in the process of resistance that, alone, only slightly shifts penicillin MIC (see supplementary online material). Similarly, ponA mutation (L421P) as a single determinant or combined with only penA will not or only moderately raise the MIC, confirming that this variation is a late determinant involved in high-level penicillin resistance. In contrast, the presence of specific mtrR (−A deletion) and penB (KD or KN) determinants is clearly associated with an obvious decrease in susceptibility up to resistance when penA and ponA are present. Because variations in PorB1b have an effect only with coresident mtrR mutation, this latter appears to be particularly essential to monitor.
Using these multigene genotyping assays on phenotypically documented isolates from various origins, we confirm that antibiotic resistance in non-PPNG can be predicted with a reasonable accuracy. Additionally, these genotyping assays could also evidence a decreased susceptibility to tetracycline and erythromycin,6 and might also be of use in the surveillance of resistance to cephalosporins, an antibiotic family now widely used to treat gonococci, as suggested by recent studies.23
Non-typable determinants in Malagasy isolates might correspond to naturally occurring porB hybrids24 25 or mosaic penA alleles associated with third-generation cephalosporin resistance.26 Both have already been described in other places24–26 but were never reported from Madagascar, highlighting the need to reinforce surveillance of Malagasy gonococci.
With the unexpected exception of two Malagasy isolates out of 169 (1.2%), a clear association between mutations in QRDRs and fluoroquinolone resistance is confirmed (table 3 and supplementary online material). Single- then double mutations in gyrA are clearly correlated with a gradual increase in MIC sufficient to reach a resistant status. Single- and double mutations in parC occur only in gyrA mutated isolates and correlate with high-level fluoroquinolone-resistance. Consistent with a recent report,27 the almost total clinical susceptibility to fluoroquinolones within Malagasy panel correlates with the absence of QRDR mutations. However, possible unknown mutations outside the QRDRs could explain the phenotypes in two strains that showed decreased susceptibility to ciprofloxacin without any QRDR change. Though we evidenced no mtrR change in these isolates, an efflux activation mediated by NorM efflux pump18 could possibly explain these increased MICs.
In addition to a good concordance between multigene genotypes and phenotypes, this study evidenced particular genotypes probably related to their geographical origin. Specific higher prevalence of porB1a alleles was observed among our Malagasy (a finding consistent with a previous study24) and Cambodian panels. Among the strains with a porB1a allele, we identified isolates with a modified susceptibility to penicillin. The possible implication of porB1a sequence modifications in this decreased susceptibility is not elucidated. This allele is suspected to confer a survival or transmission advantage in the urogenital tract28 and to be more frequently responsible for invasive infections; their relatively high prevalence deserves consideration.
In agreement with many reports, our Cambodian gonococcus panel reflects the susceptibility patterns and confirms an established resistance to fluoroquinolones and penicillin in this part of Asia.
Genotyping porB1b among our New Caledonian panel evidenced a genotype recently described19 as being associated with a moderate decrease in susceptibility to penicillin with a high prevalence (around 30%), this genotype being rare and absent in Malagasy and Cambodian panels, respectively. Another New Caledonian feature is the high prevalence of a +T insert within mtrR promoter also associated with a moderate decrease in susceptibility to penicillin (Nandi et al, in prep) and not seen among the other panels.
Though our multigene genotyping does not allow determination of susceptibility in individual strains, our results demonstrate that it could allow surveillance to be carried out, notably from non-viable samples and permit therapeutic recommendations to be adapted in a timely manner. Antibiotic susceptibility testing of cultured organism should continue to be performed notably to detect the emergence of new resistance mechanisms. The current limit for a direct use on clinical samples is the high conservation of the target genes between bacterial species, so other organisms present may distort data. However, our assays, which are usually performed on DNA from isolated organisms, can be successfully performed directly on clinical material.29 They could complement culture-based surveillance in places where a growing proportion of gonorrhoea cases are diagnosed using PCR, providing supplementary data for predicting antibiotic susceptibility trends.
Key messages
The reasonable accuracy of antibiotic susceptibility prediction by genotyping chromosomal genes in non-penicillinase-producing Neisseria gonorrhoeae was validated with 169 isolates from three geographical origins.
Specific genotypes could be related to the geographic origins of the isolates studied, notably for the penB and mtrR determinants.
The genotyping of resistance-associated chromosomal genes might prove to be a useful tool in the surveillance of antibiotic resistance in N gonorrhoeae.
Acknowledgments
The authors sincerely thank J Tapsall for sharing WHO reference isolates as part of the WHO-GASP. This study was funded by a grant from the Pasteur Institutes International Network (ACIP N°A/14/2006). Thanks are due to the different staffs of the Bacteriology laboratories and to JF Mackay for editorial contribution.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Data supplement 1 - Online supplement
Footnotes
Funding Institut Pasteur, Direction des Affaires Internationales, Paris, France.
Competing interests None.
Provenance and peer review Not commissioned; externally peer reviewed.
Linked Articles
- Whistlestop tour