Combined Real-Time PCR and Pyrosequencing Strategy for Objective, Sensitive, Specific, and High-Throughput Identification of Reduced Susceptibility to Penicillins in Neisseria meningitidis

Abstract

A segment of penA in Neisseria meningitidis strains (n = 127), including two nucleotide sites closely associated to reduced susceptibility to penicillins, was amplified and pyrosequenced. All results were in concordance with Sanger sequencing, and a high correlation between alterations in the two Penⁱ-specific sites and reduced susceptibility to penicillins was identified.

During recent decades, Neisseria meningitidis isolates with reduced susceptibility to penicillins have increasingly been reported worldwide (11, 15, 19, 22). These intermediate resistant isolates, Penⁱ, have previously been defined by MICs of >0.064 to ≤1.0 μg/ml of penicillin G by using Etest (6). However, the precision, standardization, and quality assurance of phenotypic antibiotic susceptibility testing of N. meningitidis may be suboptimal (21).

The Penⁱ phenotype is mainly due to alterations in penicillin-binding protein 2, encoded by penA (2, 13, 15, 18). penA of susceptible isolates, i.e., wild-type penA, is highly conserved. However, penA genes of Penⁱ isolates, i.e., penA mosaic alleles, are fairly variable and highly divergent from wild-type penA (2, 8, 14, 15, 18).

A few genetic methods for detection of reduced susceptibility to penicillins in N. meningitidis have previously been developed (1, 14, 16). In an earlier study (18), two polymorphic nucleotide sites, which were altered in all isolates displaying reduced susceptibility to penicillins and penA mosaic alleles but not in any penicillin-susceptible wild-type penA isolates, were identified. Alterations in these two Penⁱ-specific sites, i.e., C₁₅₁₂→G and C₁₅₂₉→T (Fig. 1), encoding the F504L and A510V amino acid alterations, are present in all Penⁱ isolates (2, 14-16, 18), and these alterations are also included in the proposed definition of Penⁱ isolates (15).

FIG. 1.

Multiple-sequence alignment of the 17 different partial penA sequences, which corresponds to nucleotides 1460 to 1600 of MC58 penA (17), that were identified among the N. meningitidis isolates included for development of the pyrosequencing strategy (n = 79). The positions of the used PCR primers (penA-F, 5′-GTGCGGTAGATGGTTTCGA-3′; penA-R, 5′-TTACCGCCACAATCACACG-3′) and pyrosequencing primers (penA-s1-3, 5′-GTAGCGATGTGTTTGT-3′; penA-s1-4, GTGGCAACGTGTTTGT) are shown. Accordingly, due to genetic heterogeneity different sequencing primers were used for penA wild-type alleles and penA mosaic alleles. The two boxes indicate the two Penⁱ-specific sites that are altered only in penA mosaic alleles and, accordingly, N. meningitidis isolates with reduced susceptibility to penicillins. WT, wild-type penA sequence; M, penA mosaic allele; B, primer biotinylated at the 5′ end.

The aim was to investigate whether real-time PCR and pyrosequencing (12) of a short segment of penA, spanning the two Penⁱ-specific sites, could be used as a rapid, objective, and high-throughput method for detection of N. meningitidis isolates with reduced susceptibility to penicillins.

Sixty N. meningitidis isolates (invasive [n = 55] and carrier [n = 5] isolates) collected in Sweden from 1996 to 2004 were examined. Furthermore, 17 N. meningitidis strains previously analyzed in a European antibiotic susceptibility study (21) and two reference strains, i.e., MC58 (17) and OR173/87 (7), were investigated. All isolates were previously penA sequenced using conventional Sanger sequencing (18, 21). Moreover, for evaluation of the developed method from a clinical perspective, all Swedish invasive N. meningitidis isolates (one per patient) from 2005 (n = 48) and five cerebrospinal fluid samples, previously Sanger sequenced (18), were included.

DNA was isolated using the MagNA Pure systems (Roche Diagnostics GmbH, Mannheim, Germany). A 141-bp segment of penA (Fig. 1) was amplified using the LightCycler system (Roche Diagnostics GmbH) with SYBR Green I fluorescence melting curve analysis. The amplicons were purified using a vacuum prep tool and vacuum prep worktable (Biotage AB, Uppsala, Sweden) and subsequently pyrosequenced in a PSQ 96 MA instrument (Biotage AB). The DNA isolation, PCR, and pyrosequencing were performed according to the manufacturer's instructions.

MICs of penicillin G, ampicillin, penicillin V, cefotaxime, and cefuroxime were determined using the Etest method (AB Biodisk, Solna, Sweden) on Mueller-Hinton agar (Becton Dickinson and Company, Sparks, MD) supplemented with 5% sheep blood (SVA, Uppsala, Sweden) at 37°C in 5% CO₂ for 16 to 18 h (21).

Real-time PCR amplification and subsequent pyrosequencing were performed without obstacles. Pyrosequencing allowed rapid (in approximately 1.5 h) correct determination of 40 to 50 nucleotides in 96 different sequences. The software-interpreted results of pyrosequencing were in concordance with Sanger sequencing. High correlation between the Penⁱ-specific alterations and elevated MICs was observed for penicillin G (Fig. 2a) and ampicillin (Fig. 2b); correlations were somewhat lower for penicillin V (Fig. 2c) and cefuroxime (Fig. 2d). For cefotaxime, all isolates were highly susceptible, i.e., MIC of ≤0.012 μg/ml, and hence no correlations could be determined.

FIG. 2.

MICs of penicillin G, ampicillin, penicillin V, and cefuroxime using Etest (20, 21) and presence of altered Penⁱ-specific sites in N. meningitidis isolates (n = 127). Three N. meningitidis serogroup Y isolates did not grow on Mueller-Hinton agar supplemented with 5% sheep blood, and, accordingly, these were analyzed on medium B (Swedish Reference Group for Antibiotics; http://www.srga.org; accessed 12 November 2007). Gray bars indicate wild-type penA genes, and black bars indicate the presence of altered Penⁱ-specific sites. The broken lines in panels a and b indicate the breakpoints suggested for isolates with reduced susceptibility (15, 18).

In this study, a rapid, sensitive, specific, and high-throughput method for genetic detection and screening of reduced susceptibility to β-lactam antibiotics in N. meningitidis was successfully developed and evaluated. High correlation between alterations in the two Penⁱ-specific sites and reduced susceptibility to penicillin G and ampicillin was observed (Fig. 2a and b), as previously shown by sequencing longer segments of penA (3, 15, 16, 18). Our results show MICs of penicillin G of >0.094 μg/ml for Penⁱ isolates, which is in concordance with previous studies (15, 20), and ampicillin MICs of >0.064 μg/ml. Consequently, the two Penⁱ-specific sites proved to be highly reliable markers for penA mosaic alleles and reduced susceptibility to β-lactam antibiotics in N. meningitidis. Alterations in these two Penⁱ-specific sites are present in all Penⁱ isolates (2, 14-16, 18), and these alterations are also included in the proposed definition of Penⁱ isolates (15). By using real-time PCR and pyrosequencing for examination of the two Penⁱ-specific sites, rapid sequence data that are objective, sensitive, specific, portable for comparison between laboratories, and reproducible for detection of reduced susceptibility to β-lactam antibiotics in N. meningitidis isolates are generated. Pyrosequencing has also previously been beneficially used for detection of antibiotic resistance in bacteria (4, 5, 9, 10). Pyrosequencing is accurate and reproducible and enables reading from the first base after the sequencing primer. In comparison with Sanger sequencing, pyrosequencing is also less expensive, time-consuming, and labor-intensive and easier to perform.

Although reports of treatment failures using penicillins for invasive meningococcal infections have been rare and not conclusive, the increasing incidence of Penⁱ isolates worldwide and the possible cross-resistance to other β-lactam antibiotics (15, 20, 22) are alarming. Accordingly, it is of clinical relevance and crucial to continuously phenotypically monitor the antibiotic susceptibility to penicillins and other antibiotics of N. meningitidis. However, in addition, it would be highly beneficial to perform recurrent genotypical surveillances.

In conclusion, the increasing incidence of Penⁱ isolates worldwide, lack of completely objective, precise, and harmonized phenotypic antibiotic susceptibility testing, and increasing number of N. meningitidis specimens being diagnosed using only PCR clearly emphasize the need for alternative, i.e., genetic, methods. Real-time PCR and subsequent pyrosequencing of a short segment of penA spanning two Penⁱ-specific nucleotide sites are an effective method for detection of penA mosaic alleles and, accordingly, reduced susceptibility to penicillins in N. meningitidis. This method enables rapid, sensitive, specific, and high-throughput diagnostics also for culture-negative cases of meningococcal septicemia/meningitis. Pyrosequencing may also be of great interest for use in remote areas, yielding more objective results for interlaboratory comparisons than phenotypic antibiotic susceptibility testing and being less expensive than Sanger sequencing.