Materials and methods

Testing centres

Testing was performed in two Australian centres: the Queensland Paediatric Infectious Diseases Laboratory located in Brisbane, Queensland (QPID-Bris) and the Neisseria Reference Laboratory, South Eastern Area Laboratory Services in Sydney, New South Wales (SEALS-Syd).

N gonorrhoeae-positive clinical samples

A total of 73 N gonorrhoeae-positive samples was used in the study, comprising 27 culture-positive samples (table 1) and 46 samples analysed by PCR only (table 2). All cultured isolates were grown and identified by standard methods,12 and national recommendations for the detection of gonorrhoea by nucleic acid amplification assays were followed for the detection of gonococcal-specific nucleic acids that included the use of supplemental assays.13

For the 27 culture-positive samples, NG-MAST analysis was performed on both a DNA extract of the neat clinical specimen (the non-cultured clinical sample) as well as the corresponding gonococcal isolate and the results were compared. These samples included one cervical swab, four rectal swabs, one throat swab and six urethral swabs (samples 1–12; table 1) collected from 12 patients (11 men and one woman) as part of a previous study at QPID-Bris.14 At SEALS-Syd, 15 urine samples from 15 male patients positive by gonococcal PCR were prospectively analysed in parallel with corresponding urethral swab isolates (samples 13–27; table 1).

A further 46 clinical sample DNA extracts that were positive by three different gonococcal PCR assays in a previous study15 were retrospectively analysed using NG-MAST at QPID-Bris to assess further the performance of the typing system across a range of sample types (samples 1–46; table 2). These included seven cervical swabs, nine rectal swabs, six throat swabs, three urethral swabs, 20 urine samples and one vaginal swab collected from 42 patients (34 men and eight women). Paired clinical sample DNA extracts were tested for four male patients, including paired urethral swab and urine samples for two patients and paired rectal swab and urine samples for two patients.

The cycle threshold (Ct) values for the 46 specimens when originally tested in a previous study15 by a diagnostic PCR (NGduplex) ranged from 22 to 40 cycles (table 2). Ct values are a semiquantitative marker and are indirectly proportional to the DNA load, with a difference of 3.3 cycles representing approximately a one log difference in DNA load. Therefore these specimens represented a broad range of gonococcal DNA concentrations. For all 46 specimens bacterial culture had not been performed and therefore corresponding gonococcal isolates were not available.

N gonorrhoeae-negative clinical samples and non-gonococcal Neisseria species

DNA extracts of 100 clinical samples, comprising 50 urine (from 15 male and 35 female patients) and 50 throat swab samples (from 27 male and 23 female patients), which were negative for N gonorrhoeae by three different PCR assays in a previous study,15 were tested to assess further the specificity of the NG-MAST method. The analytical specificity was also investigated by testing DNA of 44 non-gonoococcal Neisseria isolates, including Neisseria canis (one isolate), Neisseria cinerea (three), Neisseria flavescens (one), Neisseria lactamica (four), Neisseria meningitidis (10), Neisseria mucosa (four), Neisseria polysacchareae (one), Neisseria sicca (three), Neisseria subflava (11) and Moraxella catarrhalis (six).

NG-MAST analysis

NG-MAST analysis was performed as previously described.9 Briefly, porB and tbpB sequences were amplified by PCR and then submitted for DNA sequencing. The porB and tbpB sequences were then trimmed to 490 bp and 390 bp, respectively, starting at particular sites of sequence conservation (TTGAA at the pre-loop 3 to loop 6 for porB and CGTCTGAA for tbpB) and were then analysed using the NG-MAST website (http://www.ng-mast.net/). Only specimens providing positive PCR amplification for both porB and tbpB sequences were submitted for DNA sequencing and analysed by NG-MAST.

At SEALS-Syd, the tbpB PCR reaction mix (total volume 50 μl) comprised 25 μl of AmpliTaq Gold PCR Master mix (Applied Biosystems, Australia), 0.1 μM of primers tbpB-F(CGTTGTCGGCAGCGCGAAAAC) and tbpB-R(TTCATCGGTGCGCTCGCCTTG)9 and 10 μl of DNA extract. PCR cycling was performed on a geneAmp PCR System 2400 conventional thermocycler (Applied Biosystems) with the following cycling conditions: 95°C for 4 minutes, followed by 25 cycles of 95°C for 30 s, 62°C for 30 s and 72°C for 1 minute, followed by a final extension step of 72°C for 10 minutes. The porB PCR used a similar reaction mix but with primers POR-F (CAAGAAGACCTCGGCAA) and POR-R (CCGACAACCACTTGGT)9 and was thermocycled at 95°C for 5 minutes, followed by 30 cycles of 95°C for 30 s, 58°C for 30 s and 72°C for 1 minute, followed by extension of 72°C for 10 minutes. Some samples, which were not successfully amplified in the initial porB reaction, were subsequently re-tested in the porB PCR using 35 cycles in the PCR. PCR products were visualised by agarose gel electrophoresis, purified and then sequenced using the ABI PRISM BigDye sequencing kit (Applied Biosystems).

At QPID-Bris, the porB and tbpB PCR reactions were performed using a standard reaction mix and cycling conditions. Each reaction mix (total volume 25 μl) contained 12.5 μl of 2 × QuantiTect SYBR Green PCR Master Mix (Qiagen, Australia), 0.4 μM of forward and reverse primers (tbpB-F and tbpB-R for tbpB; POR-F2, CAAAGGCCAAGAAGACCTCGGCAA and POR-R2, TTCCGCACCGACAACCACTTGGT for porB) and 5.0 μl of DNA extract. Thermocycling was performed on a RotorGene 6000 real-time PCR instrument (Corbett Life Science, Australia) with the following cycling conditions; 95°C for 15 minutes, followed by 45 cycles of 95°C for 30 s, 60°C for 30 s and 72°C for 1 minute. PCR product formation was indicted by the presence of real-time PCR amplification curves. PCR products were sent for automated fluorescent sequencing at the Australian Genome Research Facility (http://www.agrf.org.au/).

For all specimens returning previously unidentified (new) NG-MAST sequence types, the respective porB and tbpB sequences were further analysed by Genbank nucleotide Blast search (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

N gonorrhoeae and N meningitidis isolation data

The data base of SEALS-Syd was used to determine the total numbers of isolates of N gonorrhoeae and N meningitidis reported from cultures collected from male urethral and anorectal, and female endocervical swabs in the period 2006–8 inclusive.

Results

N gonorrhoeae-positive clinical samples

For all 27 culture-positive specimens, porB and tbpB sequences were successfully amplified by PCR from both the DNA extract of the neat clinical specimen and the corresponding gonococcal isolate. In addition, the porB and tbpB sequences obtained from the DNA extract of the neat clinical specimen were identical to those obtained using the corresponding gonococcal isolate at both QPID-Bris and SEALS-Syd. A total of 21 different NG-MAST sequence types, including three previously unidentified sequence types, was observed in the 27 culture-positive specimens (table 1).

For the 46 N gonorrhoeae PCR-positive clinical sample DNA extracts, porB and tbpB sequences were successfully amplified by PCR from 41 specimens. A total of 24 different NG-MAST sequence types, including seven previously unidentified sequence types, was observed (table 2). For the four male patients with paired clinical samples, the same NG-MAST type was observed in paired samples. For five of the 46 N gonorrhoeae PCR-positive specimens, including three throat swabs and two rectal swabs, an NG-MAST type could not be obtained due to failure to amplify porB (one throat swab), tbpB (one throat swab) or both sequences (one throat and two rectal swabs) from these specimens. When originally tested by the diagnostic PCR assay, these three throat swabs and one of the rectal swab specimens provided Ct values in the highest observed range (36–40 cycles) and therefore were considered to contain the lowest concentrations of gonococcal DNA (table 2).

By specimen type, NG-MAST types (including new sequence types) were obtained for eight of eight (100%) cervical swabs, nine of nine (100%) urethral swabs, 35 of 35 (100%) urine samples, one of one (100%) vaginal swab, 11 of 13 (85%) rectal swabs and four of seven (57%) throat swabs.

The porB and tbpB sequence analysis for new NG-MAST sequence types (samples 2, 6, 12, table 1 and samples 4, 5, 14, 34, 38, 39, 43, table 2) were consistent with N gonorrhoeae sequences on the Genbank database for all specimens with a previously unidentified sequence type, except for one throat swab (sample 43, table 2). For this specimen, the porB sequence showed greatest similarity to N meningitidis (99%; genbank accession AY699312) compared with the closest N gonorrhoeae porB sequence (88%; genbank accession AF015120) and the tbpB sequence showed greatest similarity to N lactamica (97%; genbank accession AM849587) compared with the closest N gonorrhoeae tbpB sequence (86%; genbank accession EU532707).

N gonorrhoeae-negative clinical samples

Seven of 50 throat swab and one of 50 urine N gonorrhoeae-negative samples provided PCR products in both the porB and tbpB assays. For the throat swabs, five of the porB sequences showed greatest similarity to N meningitidis and for the tbpB sequences showed greatest similarities to N meningitidis (three specimens), N lactamica (one specimen) and N polysacchareae (one specimen). All of these sequences provided less than 90% similarity to N gonorrhoeae sequences on the Genbank database. Sequence information was not obtained for the remaining throat swab porB and tbpB reactions. The porB and tbpB sequences from the urine sample did not match with Neisseria species, showing greatest similarity to Enterobacteriaceae species.

Non-gonococcal Neisseria species

Of the 44 non-gonococcal Neisseria strains investigated, 14 isolates provided positive PCR amplification for both the porB and tbpB NG-MAST PCR reactions. These included all four N lactamica strains, the one N polysacchareae strain and nine of the 10 N meningitidis isolates.

N gonorrhoeae and N meningitidis isolation data

SEALS-Syd performs approximately 15 000 cultures for N gonorrhoeae per annum. There were 241 gonococcal isolates from the male urethra of clinic attendees in the period 2006–9, 193 from the anorectum of men, 23 from the female endocervix and one from female anorectal samples. Over the same period there were nine meningococcal isolates from the male urethra, 59 from the male anorectum and a single isolate from women. No data were available on throat swab samples because meningococci are not reported when cultured from this site.

Discussion

Overall, the results show that the NG-MAST system can be readily performed with good sensitivity and specificity on non-cultured urogenital specimens that were positive in a NAAT for GC and when reliable supplemental assays were performed. Notably, gonococcal porB and tbpB sequences were successfully amplified and DNA sequenced from all N gonorrhoeae-positive urogenital specimens used in the study, and the porB and tbpB sequences matched 100% with those of corresponding isolates, when available. Overall, the sequence chromatograms were of good to high quality for all N gonorrhoeae-positive specimens, including from specimens with higher NGduplex Ct values. These data are in contrast to a previous study that found the NG-MAST system had poor sensitivity when applied directly to non-cultured clinical samples. Ng et al11 could only amplify porB sequences in 10 of 38 clinical specimen extracts and tbpB in only 16 of the 38 clinical specimen extracts that had provided positive results by both Roche Amplicor and BD ProbeTec N gonorrhoeae NAAT assays. Differences in study methods could explain the disparity in these findings. For example, the porB and tbpB PCR conditions used at QPID-Bris were optimised to improve assay sensitivity, and included the use of the Qiagen QuantiTect Master mix in combination with modified porB primer sequences that were lengthened at the 5′ end compared with the originally described primers. It is also possible that Ng et al11 may have over-sampled ‘low load’ specimens, which we believe to be at least one of the factors limiting the applicability of NG-MAST typing to the extragenital specimens investigated in this study.

The results of the NG-MAST typing of the extragenital specimens, combined with the results of the commensal Neisseria species, show that the non-culture-based NG-MAST system has limited applicability to these sites, particularly throat specimens. The sensitivities of the NG-MAST system were reduced for throat and rectal specimens compared with the urogenital sites, presumably due to the fact that N gonorrhoeae DNA loads are generally lower in extragenital sites.14 However, the results of this study show that amplification of the porB and tbpB sequences of non-gonococcal Neisseria species represents a more significant hurdle with regard to specificity. The key species associated with this problem appear to be N lactamica, N meningitidis and N polysacchareae, which is consistent with previous phylogenetic grouping of Neisseria species.16 Such cross-reactions clearly have the potential to produce incorrect NG-MAST types. Genbank Blast search analysis suggested that the porB and tbpB sequences obtained from a N gonorrhoeae-positive throat swab specimen in this study (sample 43, table 2) were more likely to have been derived from amplification of N meningitidis and N lactamica sequences rather than N gonorrhoeae. To try to circumvent the cross-reaction problems, we attempted to improve the specificity of the porB and tbpB PCR methods by using the available sequence information on the Genbank database to develop new primers targeting alternative regions of the gonococcal porB and tbpB genes. Despite these efforts, we were unable to avoid cross-reaction with the non-gonococcal species (data not shown). For these reasons, we have discontinued attempts to apply the NG-MAST system to extragenital sites.

The N gonorrhoeae and N meningitidis isolation data obtained from SEALS-Syd provided further support for these findings. For example, the nine meningococci cultured and reported from male urethral samples in the 3 years 2006–8 represented 3.6% of the 250 gonococci and meningococci at this site and all were present only as single species when grown on Modified New York City medium. Therefore, a highly specific NAAT for GC, when performed on urine or urethral samples from men and used in conjunction with supplemental assays, would not produce a positive result for GC in the presence of urethral meningococci, and NG-MAST analysis would not be initiated. In contrast, in the same period 59 meningococci were cultured and reported from male anorectal samples, and represented 23.4% of gonococci and meningococci grown from this site. It would thus be more likely that NG-MAST on an anorectal sample would be initiated in the presence of meningococci, given the greater potential for NAAT cross-reaction.

Restricting non-culture-based NG-MAST analysis to urogenital specimens will reduce, but not necessarily prevent, the potential for incorrect NG-MAST types being caused by cross-reaction with non-gonococcal Neisseria species, given the lower prevalence of these species in this site. However, in the absence of isolate-derived NG-MAST data, the issue of ‘new’ or previously unidentified NG-MAST types will probably remain a problem. This is because it would be difficult to determine whether certain sequence data were of gonococcal or non-gonococcal origin, given that N gonorrhoeae-positive samples may also contain other Neisseria species. Genbank Blast searching, as used in this study, could potentially be used to help exclude sequences that are not consistent with gonococcal sequences. However, given the propensity for N gonorrheoae strains to acquire non-gonococcal sequences, and vice versa, via genetic exchange,17–19 it is still possible that a gonococcal isolate may legitimately possess non-gonococcal Neisseria-like sequences. On the whole, these issues highlight the fact that non-culture-based NG-MAST typing of urogenital specimens as proposed in this study would be best used as a tool to complement culture-based surveillance, rather than be used alone.2 Furthermore, we believe that the restriction of non-culture-based NG-MAST analysis to urogenital specimens will not undermine the utility of the approach, given that urogenital specimens comprise the majority of N gonorrheoae-positive samples. It should also be noted that the majority of N gonorrheoae-positive specimens used in the study were predominately from male patients, and that further validation of NG-MAST testing of female genital tract specimens may be required. Furthermore, the potential for cross-reactivity in N gonorrhoeae-negative urethral, cervical and vaginal swabs was not evaluated.

In conclusion, the results show that the NG-MAST system can successfully be applied to non-cultured urogenital samples, and may be used to complement NG-MAST analysis of gonococcal isolates. However, consideration needs to be given to the potential for cross-reaction with non-gonococcal species, and that non-culture-based NG-MAST analysis may not be suitable for extragenital sites, including the throat and rectum, where commensal Neisseria species are common.