Ethics approval was not obtained from the CDC Institutional Review Board (IRB) for clinical specimens reported in this study because the specimens were sent to CDC for detection of meningitis etiology as part of reference lab functions.

CDC IRB approval was not obtained for sodC testing of DNA extractions from Brazil and UK carriage specimens because those extractions were not tested by CDC researchers. DNA extractions of carriage specimens were not considered human specimens by the CDC, Emory University, and Children's Hospital of Atlanta (CHOA) IRBs for this study, as they do not meet the definition of a living human subject. IRB approval was obtained by the institutions who collected the biological carriage specimens from the human subjects: (1) the National Ethics Research Committee and by the Regional Ethics Committee of the Hospital Materno Infantil, Secretary of Health of Goias State, Brazil; (2) the Emory University IRB; or (3) the National Health Service Research Ethics Committee (08/H1001/52), sponsored by the Royal Liverpool and Broadgreen University Hospitals Trust.

Control isolates used for assay design and optimization are defined in Table S1.

626 cell lysates were used to determine the sensitivity of the sodC assay (Table 1 and Table S2), including lysates prepared from a temporally and geographically dispersed convenience sample of isolates from the CDC Meningitis Laboratory strain collection (received 1993–2008, n = 106) and all isolates from a US carriage study (n = 520) [20], [21] known to be Nm by SASG [1], [2], rt-PCR serogrouping [5], NH strips (bioMérieux® sa), and Cystine Trypticase Agar (CTA) sugars (Remel) [1], [2]. To further confirm identification, multilocus sequence typing (MLST) was performed on all U.S. carriage study and ctrA-negative NG isolates.

The specificity of the sodC assay for detecting only meningococci was determined using cell lysates from a total of 244 non-Nm isolates (Table 2).

CSF specimens from pediatric meningitis patients were cultured as soon as possible after collection. Specimens that were culture-negative were sent to CDC on ice for detection of meningitis etiology by the Marmara University School of Medicine in Istanbul, Turkey, and came from patients who met the case definition for purulent meningitis [leukocytosis (>100 cells/mm³) and either elevated protein (>100 mg/dl) or decreased glucose (<40 mg/dl)]. After DNA extraction and real-time PCR testing of all specimens for ctrA of Nm [5], lytA of S. pneumoniae [22], and bexA and/or bcs2 of Hi, the subset of specimens chosen to test the sodC assay (n = 120) were either (1) positive for ctrA (n = 12) or (2) ctrA⁻ lytA⁻ bexA/bcs2⁻ (n = 108).

37 U.S. clinical specimens were referred to CDC for detection or confirmation of bacterial meningitis etiology from January to June 2009, including CSF (n = 21), whole blood (n = 6), serum (n = 6), and tissues (n = 4) (Table 3).

The carriage specimens were obtained from three carriage studies. 1. In Goiania, Brazil, nasopharyngeal (NP) swabs (n = 223) were obtained from 154 children (ranging from 2–163 months of age) attending two daycare centers, 59 adult contacts of the attendees, and 10 daycare workers. The specimens were placed into skim milk-tryptone-glucose-glycerine (STGG) transport medium [23] and sent immediately to the Applied Microbiology Laboratory of Federal University of Goiás in Brazil for processing. The vials were then kept frozen during transport to CDC, where DNA extractions and rt-PCR were performed.

2. 291 posterior NP swab specimens were collected from a random sample of children 6 to 59 months of age who presented to the Emergency Department at CHOA at Egleston from March to August, 2009 [24]. Each swab specimen was immediately placed into 1 ml STGG [23]. Specimens were transported at room temperature to the clinical microbiology laboratory within 12 hours of collection for storage at −80°C until processing. An aliquot of the STGG from each specimen was transported on dry ice to the CDC Meningitis Laboratory, where DNA extractions and rt-PCR were performed.

3. A total of 33 NP swabs and 35 nasal washes (NWs) were taken from 24 participants ages 21–57 years during ≤7 visits in a Spring 2009 study conducted in the NIHR Biomedical Research Centre in Microbial Diseases at the Liverpool School of Tropical Medicine, Liverpool, United Kingdom (UK). Swab specimens were collected as previously described [25] with some modifications and placed directly into 1 ml STGG [23], then transported to the laboratory on wet ice for culture and processing. 900 µl of each specimen was frozen at −80°C for subsequent DNA extraction and rt-PCR in Liverpool.

NWs were collected as previously described [26] and the nasal fluid from both nostrils of each study participant was pooled. NWs were collected in this manner from each participant. Once at the lab, NW specimens were centrifuged at 1509× g for seven minutes and the pellet was re-suspended in 1 ml STGG. The processing of the NWs from this point was the same as that for the NP specimens.

Genomic DNA was prepared for use in the various steps of assay design and optimization using the QIAamp DNA Mini Kit (QIAGEN, Valencia, California) using Protocol C then quantified for use in standard curve experiments using a NanoDrop ND-1000 or 8000 spectrophotometer (Nanodrop Technologies, Wilmington, Delaware). Preparation of DNA from bacterial isolates was performed as previously described [5].

For CSF specimens and the Brazilian NP swab eluates, DNA extraction was conducted as previously described for clinical specimens [23]. DNA extractions were performed on the CHOA carriage study NP swab eluates using the MagNA Pure LC instrument and the DNA Isolation Kit III (Bacteria, Fungi) per the manufacturer's instructions (Roche Diagnostics GmbH, Mannheim, Germany). DNA was extracted from the UK NP and NW specimens using the QIAsymphony SP System and the QIAsymphony Virus/Bacteria Midi Kit (QIAGEN Inc., UK) according to the manufacturer's instructions. All extracted DNA was stored at −20°C.

The sodC sequencing templates were prepared by conventional PCR using Expand High Fidelity Enzyme Mix (Roche Diagnostics GmbH) per the manufacturer's instructions and primers (see Table 4) that were designed based on a consensus of sodC from meningococcal strains Z2491 (nts 1521721–1522258), FAM18, and MC58 (respective GenBank accession numbers AL157959.1, AM421808.1, and AE002098.2) [19]. DNA sequencing was performed using the BigDye Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, California) and an ABI PRISM® 3130xl Genetic Analyzer (Applied Biosystems) and a consensus sequence was generated using Lasergene DNAStar v. 7 Program SeqMan.

The sodC consensus sequence was entered into Primer Express 3.0 (Applied Biosystems). Primers and probes were analyzed for homology to other known sequences using the Basic Local Alignment Search Tool (BLAST) [27]. Primers were tested for optimal concentration in triplicate or quadruplicate in combinations of final concentrations of 100, 300, 600, and 900 nM; the probe was tested in triplicate at final concentrations of 50, 100, 200, and 300 nM. The amplified product is located at nt 1427446 in MC58 (GenBank accession number AE002098.2).

A Stratagene Mx3005P (Agilent, La Jolla, California) and QuantiTect SYBR Green Master Mix (QIAGEN) were used to optimize primer concentrations. Cycle parameters were 2 minutes at 50°C, 10 minutes at 95°C, and then 50× (15 seconds at 95°C plus 1 minute at 60°C). Product dissociation curves were generated using one round of the following cycle parameters at the end of the primer optimization run: one minute at 95°C, 30 seconds at 55°C, and 30 seconds at 95°C. Master mixes contained 4.5 µl sterile PCR grade water (Roche Diagnostics), 12.5 µl TaqMan®2× PCR Master Mix (Applied Biosystems), 300 nM forward primer, 600 nM reverse primer, 100 nM probe, and 2 µl template DNA per total reaction volume of 25 µl. With each reaction plate that was run, cell lysates from known Nm served as positive external amplification controls, while no-template controls (NTCs) served as negative external amplification controls.

rt-PCR detection of ctrA in the MD isolates [20] was performed at the Laboratory of the Maryland Department of Health and Mental Hygiene using an Applied Biosystems 7500 or 7700 rt-PCR System (Applied Biosystems, Inc., Foster City, CA).

A Corbett Rotor-Gene 6000 (QIAGEN) and Corbett software 24 series 1.7 were employed in the UK for rt-PCR and data analysis on extractions of NWs and NP eluates. In every UK sodC reaction, 2.5 µl of extracted DNA was used.

Throughout this study, C_t values≤35 were considered positive; C_t values in the range of 36–40 equivocal; and C_t values>40 negative. Extracted DNA from equivocal clinical specimens was diluted 1∶4 and 1∶10 in an attempt to dilute possible inhibitors of the reaction. Dilutions were then re-tested in duplicate. If the average C_t of the diluted specimen fell below 35, that specimen was considered positive. If the average C_t of the diluted specimen remained in the 35–40 range, that specimen was considered negative.

MLST was performed as described by Maiden et al. [28]. Sequences were assembled and alleles were determined using Staden [29] v. 1.5.3 and Sequence Typing Analysis Retrieval System (STARS) v. 1.2a or the Meningococcus Genome Informatics Platform (MGIP) (http://mgip.biology.gatech.edu.) [30]. Sequence types (STs) and clonal complexes were assigned by querying http://pubmlst.org/neisseria or http://neisseria.org/nm/typing.

The GenBank accession numbers for the Nm sodC sequences generated in this study are listed in Table S3.

A consensus nucleotide sequence was built based on 2× coverage of sodC sequenced from each of 14 geographically and temporally dispersed groupable and NG meningococci (Table S1). Over the 439–473 out of 560 sodC nucleotides sequenced in these isolates, these meningococcal isolates were 99–100% identical to each other and 81% identical to an Hi sodC consensus that was built using two sequences from GenBank (accession numbers M84012 and AF549211). Given that sodC was likely acquired by Nm via horizontal transfer from Hi [18], primers and a probe that differed by at least three nucleotides per oligo between Nm and Hi sodC were chosen (Figure S1, Table 4).

BLAST results showed that the primers had no homology over 78% nucleotide identity with any genes but meningococcal sodC. The only notable homology found for the probe was a two-nucleotide difference with H. parainfluenzae sodC; the primers, however, showed no homology to this gene.

Standard curves were generated by testing genomic DNA from four invasive meningococcal isolates with the sodC assay (data not shown). LLDs at a C_t value of 35 were found to be 39, 70, 101, and 82 genomes per reaction, yielding an average of 73 genomes detected per reaction. The average reaction efficiency was 100% and the average R² value was 0.9925.

Invasive (i.e., isolated from CSF, blood, joint fluid, or autopsy tissue, n = 76) and non-invasive (i.e., including carriage isolates and those from sputum or oral swab specimens, n = 30) known Nm isolates from the CDC Meningitis Laboratory Strain Collection were tested with the sodC assay. All isolates were positive for sodC, including 26 ctrA-negative NG isolates, with a median C_t value of 19.7, mean of 19.9±1.9, and range of 16.4 to 26.0 (Table 1 and Table S2).

518/520 (99.6%) of meningococcal carriage isolates from two U.S. carriage studies [20], [21] were positive for sodC (median C_t 16.9, mean 17.0±1.5, and range 13.5 to 29.3), while ctrA detected only 368/520 (70.8%) of these isolates (median C_t 19.0, mean 19.2±2.7, and range 13.5 to 34.0) (Table 1 and Table S2). The two sodC-negative carried Nm were SASG NG, ctrA-negative but were confirmed to be Nm ST-1117 and ST-4788, both cc1117; both were isolated from the same study participant at different time points. Therefore, 176/178 (98.9%) SASG NG, ctrA-negative invasive and non-invasive Nm isolates were sodC-positive. Four sodC-negative carriage study isolates that were identified as Nm by conventional methods were re-investigated and shown to actually be non-Nm.

None of 35 non-Nm from various sources and none of 209 non-meningococcal carriage study [20], [24] isolates were detected by sodC (100% specificity) (Table 2). Interestingly, sodC identified one isolate, M16160, as Nm when other standard carriage study tests could not correctly resolve its species. It was originally identified as N. polysaccharea/N. spp. by NH strip, but upon re-investigation, did not grow at room temperature on chocolate agar, indicating that it is not N. polysaccharea [1]. All 7 meningococcal housekeeping genes were readily amplified during MLST, suggesting that M16160 is indeed Nm (ST-7456, cc60), and again showing sodC to be a useful tool for definitive identification of carriage isolates.

The ability of sodC to detect Nm was assessed and was compared to ctrA as the target gene [5] using extracted DNA from 120 Turkish CSF specimens and 37 U.S. clinical specimens. ctrA detected Nm in 21/157 (13.4%) specimens (20 CSF and 1 blood), while sodC was positive for those 21 plus four additional CSF specimens (25/157, 15.9%) (Table 3 and Table S4). Therefore, sodC was 100% (95% confidence interval [CI]: 84–100%) sensitive compared to ctrA at detection of Nm from these clinical specimens, but specificity is difficult to calculate using ctrA as a reference test. The C_t values for the four ctrA-negative, sodC-positive CSF extractions averaged 40.4 for ctrA while their sodC C_t values averaged 32.0.

One ctrA-negative, sodC-positive CSF specimen was Nm culture-positive (1/157, 0.6%); the remaining specimens were either culture-negative (139/157) or culture was not attempted or not reported (17/157). Therefore, sodC was 88% (95% CI: 82–93%) specific compared to culture at detection of Nm from the clinical specimens for which culture was attempted (Table S4), but sensitivity is undetermined due to the low number of culture-positives.

Results from all three carriage studies are summarized in Table 5. None (0/223, 0%) of the Brazilian NP swab eluate extractions tested at CDC were ctrA-positive while 3 (3/223, 1.3%) were sodC-positive, yielding C_t values of 29.1, 26.2, and 25.5. The ctrA-negative, sodC-positive specimens were negative for serogroups A, B, C, W135, X, and Y by rt-PCR. In the 1/3 ctrA-negative, sodC-positive specimen that had a sufficient amount of extraction volume remaining, MLST confirmed the presence of meningococcal DNA (ST-823, cc198). All 23 Hi culture-positive Brazilian specimens were ctrA- and sodC-negative.

All (291/291, 100%) NP swab eluate extractions from the Georgia CHOA carriage study were ctrA-negative and sodC-negative, as expected, since no Nm was cultured. These specimens were, however, culture-positive for N. spp. (n = 3), N. polysaccharea (n = 1), M. catarrhalis (n = 8), H. parainfluenzae (n = 9), H. haemolyticus (n = 1), and Hi (n = 76), further demonstrating the specificity of sodC.

S. aureus (n = 17), alpha-hemolytic streptococci (n = 18), M. catarrhalis (n = 5), diptheroids (n = 10), N. polysaccharea (n = 1), N. cinerea (n = 2), and N. meningitidis (n = 1) were cultured from the 68 UK carriage study specimens; the Nm culture-positive NW was ctrA-positive and sodC-positive. sodC and ctrA were negative for all of the non-Nm culture-positive specimens except 1 ctrA-positive, sodC-positive NW that grew N. cinerea and alpha-hemolytic streptococci. 1/33 (3%) NP swab eluate extractions from the UK carriage study was ctrA-positive, 0/33 were sodC-positive, and 0/33 were Nm culture-positive. 3/35 (8.6%) NW extractions were ctrA-positive, 2/35 (5.7%) were also sodC-positive; of these, one (1/35, 2.9%) ctrA-positive, sodC-positive NW was Nm culture-positive. The ctrA-positive, sodC-negative NP specimen (average ctrA C_t 35.0±0.3, average sodC C_t 37.8±0.8) and the ctrA-positive, sodC-negative NW (average ctrA C_t value of 34.3±0.2 and an average sodC C_t value of 35.1±0.2) were both from patient 8, visit 3; this patient had the Nm-positive culture at visit 1 and the N. cinerea-positive culture at visit 2.