Browse Subject Areas

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Development of a PCR algorithm to detect and characterize Neisseria meningitidis carriage isolates in the African meningitis belt

  • Kanny Diallo,

    Roles Formal analysis, Writing – review & editing

    Affiliations Centre pour le Développement des Vaccins (CVD), Bamako, Mali, University of Oxford (Department of Zoology), Oxford, United Kingdom

  • Mamadou D. Coulibaly,

    Roles Formal analysis

    Affiliation Centre pour le Développement des Vaccins (CVD), Bamako, Mali

  • Lisa S. Rebbetts,

    Roles Formal analysis, Writing – review & editing

    Affiliation University of Oxford (Department of Zoology), Oxford, United Kingdom

  • Odile B. Harrison,

    Roles Supervision, Writing – review & editing

    Affiliation University of Oxford (Department of Zoology), Oxford, United Kingdom

  • Jay Lucidarme,

    Roles Supervision, Writing – review & editing

    Affiliation Public Health England, (PHE–Vaccine Evaluation Unit), Manchester, United Kingdom

  • Kadidja Gamougam,

    Roles Resources

    Affiliation Centre de Support en Santé Internationale (CSSI), Ndjamena, Chad

  • Yenenesh K. Tekletsion,

    Roles Resources

    Affiliation Armauer Hansen Research Institute (AHRI), Addis Ababa, Ethiopia

  • Akalifa Bugri †,

    † Deceased.

    Roles Resources

    Affiliation Navrongo Health Research Centre (NHRC), Navrongo, Ghana

  • Aliou Toure,

    Roles Resources

    Affiliation Centre pour le Développement des Vaccins (CVD), Bamako, Mali

  • Bassira Issaka,

    Roles Resources

    Affiliation Centre de Recherche Médicale et Sanitaire (CERMES), Niamey, Niger

  • Marietou Dieng,

    Roles Resources

    Affiliation Institut de Recherche pour le Développement (IRD), Dakar, Senegal

  • Caroline Trotter,

    Roles Supervision, Writing – original draft, Writing – review & editing

    Affiliation University of Cambridge (Disease Dynamics Unit -Department of Veterinary Medicine), Cambridge, United Kingdom

  • Jean-Marc Collard,

    Roles Resources, Supervision, Writing – review & editing

    Affiliation Centre de Recherche Médicale et Sanitaire (CERMES), Niamey, Niger

  • Samba O. Sow,

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliation Centre pour le Développement des Vaccins (CVD), Bamako, Mali

  • Xin Wang,

    Roles Supervision, Writing – review & editing

    Affiliation Centers for Disease Control and Prevention, Division of Bacterial Diseases, Atlanta, United States of America

  • Leonard W. Mayer,

    Roles Supervision, Writing – review & editing

    Affiliation Centers for Disease Control and Prevention, Division of Bacterial Diseases, Atlanta, United States of America

  • Ray Borrow,

    Roles Conceptualization, Methodology, Supervision, Writing – review & editing

    Affiliation Public Health England, (PHE–Vaccine Evaluation Unit), Manchester, United Kingdom

  • Brian M. Greenwood,

    Roles Conceptualization, Supervision, Writing – original draft, Writing – review & editing

    Affiliation London School of Hygiene and Tropical Medicine (LSHTM), London, United Kingdom

  • Martin C. J. Maiden,

    Roles Conceptualization, Supervision, Writing – review & editing

    Affiliation University of Oxford (Department of Zoology), Oxford, United Kingdom

  • Olivier Manigart ,

    Roles Conceptualization, Formal analysis, Writing – original draft, Writing – review & editing

    Affiliation London School of Hygiene and Tropical Medicine (LSHTM), London, United Kingdom

  •  [ ... ],
  • for the MenAfriCar Consortium

    Membership of the MenAfriCar Consortium is provided in the Acknowledgments

  • [ view all ]
  • [ view less ]

Development of a PCR algorithm to detect and characterize Neisseria meningitidis carriage isolates in the African meningitis belt

  • Kanny Diallo, 
  • Mamadou D. Coulibaly, 
  • Lisa S. Rebbetts, 
  • Odile B. Harrison, 
  • Jay Lucidarme, 
  • Kadidja Gamougam, 
  • Yenenesh K. Tekletsion, 
  • Akalifa Bugri, 
  • Aliou Toure, 
  • Bassira Issaka


Improved methods for the detection and characterization of carried Neisseria meningitidis isolates are needed. We evaluated a multiplex PCR algorithm for the detection of a variety of carriage strains in the meningitis belt. To further improve the sensitivity and specificity of the existing PCR assays, primers for gel-based PCR assays (sodC, H, Z) and primers/probe for real-time quantitative PCR (qPCR) assays (porA, cnl, sodC, H, E, Z) were modified or created using Primer Express software. Optimized multiplex PCR assays were tested on 247 well-characterised carriage isolates from six countries of the African meningitis belt. The PCR algorithm developed enabled the detection of N. meningitidis species using gel-based and real-time multiplex PCR targeting porA, sodC, cnl and characterization of capsule genes through sequential multiplex PCR assays for genogroups (A, W, X, then B, C, Y and finally H, E and Z). Targeting both porA and sodC genes together allowed the detection of meningococci with a sensitivity of 96% and 89% and a specificity of 78% and 67%, for qPCR and gel-based PCR respectively. The sensitivity and specificity ranges for capsular genogrouping of N. meningitidis are 67% - 100% and 98%-100% respectively for gel-based PCR and 90%-100% and 99%-100% for qPCR. We developed a PCR algorithm that allows simple, rapid and systematic detection and characterisation of most major and minor N. meningitidis capsular groups, including uncommon capsular groups (H, E, Z).


Neisseria meningitidis has been responsible for numerous outbreaks of septicaemia and meningitis in countries of the African Sahel and sub-Sahel, a region known as the “African meningitis belt”[1]. The devastating nature of this disease, which predominately affects children and young adults, makes its control an international health priority. N. meningitidis is a common commensal bacterium capable of harmlessly colonizing the upper respiratory tract in a dynamic process referred to as “carriage” [2]. Mucosal carriage is a necessary prerequisite for meningococcal disease. Carriage of N. meningitidis causes disease when it breaches the host epithelial barriers and invades the bloodstream or penetrates the subarachnoidal cavity and infects the meninges [3].

N. meningitidis are classified into 12 serogroups (A, B, C, E, H, I, K, L, W, X, Y and Z), based on the structure and chemical composition of their capsular polysaccharide [4]. The serogroups associated with disease in the African meningitis belt are A, W, X, Y and C. With serogroup A epidemics being eliminated following the introduction of the serogroup A conjugate vaccine MenAfriVac[57], non-A meningococcal serogroups become the major cause of meningococcal disease in sub Sahelian Africa. Recently, large serogroup C epidemics occurred in Niger and Nigeria, [8] demonstrating the epidemic potential of non- serogroup A meningococci. Timely control and prevention of the epidemics have been a challenge due to vaccine shortages, which has resulted in tremendous pressure on hospital facilities and public health authorities. Serogroup B, a common cause of meningococcal disease in many high-income countries [9], is rarely detected in the meningitis belt. Serogroups E, I, H, K, L, Z are less commonly identified in invasive disease. Queries on the more than 38,633 records on website have demonstrated that serogroups E and Z are found both in invasive and carriage specimens, serogroups K, L, H are only found in carriage samples but very rarely, and serogroup I was not found at the time of writing. Furthermore, previous carriage studies in Europe have identified a high proportion of N. meningitidis isolates, among which, up to 30–45% did not express their capsule genes [10] and 16–20% lacked regions A, B and C of the capsular polysaccharide synthesis (cps) genes, known as a capsule null (cnl) genotype [11].

Several meningococcal cross-sectional carriage studies have been conducted in the African meningitis belt in recent years, including surveys in Burkina Faso [12, 13], and in seven other countries (Chad, Ethiopia, Ghana, Mali, Niger, Nigeria and Senegal) as part of the MenAfriCar project [14, 15]. Conventional microbiological methods have been used in these studies to detect and characterise meningococci, but these methods are time consuming, expensive and susceptible to errors. Consequently, many African laboratories have been increasingly adopting molecular biology tools to detect and characterise N. meningitidis and its capsular genotypes as they are more sensitive, specific, and less time consuming than bacteriological methods. DNA-based methods such as Polymerase Chain Reaction (PCR) have greatly improved laboratory diagnosis [16]; many studies have developed reliable PCR assays for the detection of N. meningitidis [1719] and characterization of its capsular genogroup using specific genes. However, most studies have focused on the common disease serogroups, A, B, C, W, X and Y [2023]. Fewer studies have investigated PCR for the detection of cnl or serogroups E, H, and Z which are mostly found in carriage [20, 24]. Some research groups have developed gel-based PCR assays for the detection and genogrouping of N. meningitidis whereas others have used real-time quantitative PCR (qPCR) protocols. The choice of one technique over another is often dictated by budgetary considerations, as qPCR is more expensive but less subject to contamination. Evaluation of the impact of meningococcal conjugate vaccines requires sensitive and specific PCR assays that cover all serogroups to accurately and thoroughly assess the epidemiology of N. meningitidis carriage and disease, including the potential emergence of less common groups through strain replacement and/or capsule switching [25].

The present study was conducted within the MenAfriCar project, which collected pharyngeal swabs from over 60,000 subjects in seven countries within the African meningitis belt [15]. Our objectives were to improve N. meningitidis carriage assessment by developing and optimising a simple, accurate and standardised multiplex PCR algorithm for the detection of N. meningitidis species and its capsular genogroups including: (i) the most prevalent genogroups A, B, C, X, Y and W (ii) carriage-associated genogroups such as E, H, and Z; and (iii) unencapsulated capsule null meningococci. Both traditional gel-based PCR and qPCR that can be adapted to African laboratories were validated using carriage isolates collected as part of MenAfriCar studies.

Material and methods

Literature search for available primers/probes

A literature search was undertaken to identify PCR methods with primers and probes published on N. meningitidis detection and characterization. The NCBI PubMed database was used with search terms “Neisseria meningitidis” and “PCR” (date of access: November 2011). The different techniques and detection algorithms were evaluated in order to define the most appropriate ones for our context. A focus was placed on summarising the best gel-based and qPCR assays already existing for detection and characterization of meningococci.

In silico specificity screen—Evaluation of selected primers and probes

A total of 112 primers and 16 probes reported in the literature were chosen for further analysis. Firstly, nucleotide mismatches were identified through the comparison with consensus sequences derived from target genes obtained from the website (accessed on September 2012). Primer and probe sequences were compared to all available allelic variants of the target genes to determine whether they were conserved in the majority of alleles available at that time. Technical checks on all primers and probes were then undertaken using the Primer Express software design guidelines (Applied Biosystem International (ABI), Villon sur Yvette, France): melting point temperatures (Tm) and potential secondary structures were assessed with optimal Tm varying between 58–60°C and 68–70°C for primers and probes respectively. Finally, nucleotide BLAST searches were undertaken using the NCBI BLAST tool to identify possible cross-reactions with other bacteria [26]. Primers and probes with more than 3 base pair (bp) differences compared to the consensus sequences were not used. Secondary structures or self-dimers were minimised, to avoid their potential detrimental effect on the efficiency of the reaction, but their presence was not a systematic criterion for rejection. Sequences were retained if they displayed 100% coverage according to the NCBI BLAST tool. When the result of these evaluations was unsatisfactory, new primers and/or probes were created.

Positive controls

Positive controls consisted of purified DNA from genogroups: A (isolate name: Z2491); W (A22); X (860060); Y (71/94); B (EG329, BZ198); C (L93/4286); Z (0084/93); E (297–0); H (H/ASH/87); and capsule null (0083/93) [27] as well as two genogroup E isolates (M07746, M22445). Reference strains for genogroups A and X, were purchased from the American Type Culture Collection (ATCC). Another confirmed capsule null isolate from the first Malian MenAfriCar cross-sectional study (sample 01-010664-XS1) (18) was used as a cnl positive control.

Carriage isolates

N. meningitidis isolates were collected during the first MenAfriCar cross-sectional studies as previously described [15]. Briefly, pharyngeal swabs were obtained from participants in the rural and urban sites of six centres in Chad, Ethiopia, Ghana, Mali, Niger and Senegal. Suspected N. meningitidis were subjected to oxidase test, Gram stain, biochemical tests and sero-agglutination [28]. Boiled cell suspensions to release high concentration of DNA were prepared in each center as part of the MenAfriCar routine protocol from bacteria grown overnight on blood agar plates [15]. Extracted DNA samples from all oxidase positive bacteria were also sent to the University of Oxford [14] for molecular characterization. A subset of characterized samples was chosen to constitute a representative panel of the most prevalent N. meningitidis genogroups as well as non N. meningitidis from each country. Next, the selected boiled cell suspension extracted DNA samples were sent to the Center for Vaccine Development in Mali (CVD-Mali) where PCR validation was conducted. A total of 247 extracted DNA samples were received, with 42, 38, 18, 46, 44 and 59 samples from Chad, Ethiopia, Ghana, Niger, Mali and Senegal, respectively.

Characterisation of N. meningitidis at the University of Oxford, UK

Neisseria species were analysed using the rplF assay [29]. All samples were also tested with a capsule null (cnl) sequencing assay [11] or sequencing of a fragment of the ctrA gene allowing the characterisation of genogroup A, E, K, L, X and Z. Samples that were positive for N. meningitidis and negative for cnl were tested by a qPCR multiplex assay to determine genogroups A, B, C, W, X and Y [23]. These results were considered as the reference or gold standard for this study since the assays described above were already validated and published.

Novel gel-based PCR assay design and optimization in Bamako, Mali

New primers were designed for sodC, genogroup Z, and H using the online tools Primer 3 and Primer quest (Table 1)[30, 31]. A ready-to-load Master Mix (rlMMx) from Solis Biodyne, stable at room temperature and therefore very convenient for transport and testing in Africa, was employed. rlMMx with different concentrations of MgCl2 were tested to find the optimal concentration for the assay. Thermal cycling conditions were the same for all tests: 1 cycle of 3 minutes (min) at 95°C; 35 cycles of 30 seconds (s) at 95°C; 30 s at 55°C; 30 s at 72°C. This was followed by 1 cycle of 10 min at 72°C and a final conservation step at 4°C.

Novel quantitative real-time PCR (qPCR) assay design and optimization in Bamako, Mali

Although our assay was developed to determine whether a specimen is positive or negative for meningococci and their genogroups, we keep the term “qPCR” to comply with the MIQE guidelines [32]. Quantification of meningococci would not add any information in our context since we were working with DNA samples extracted from bacterial culture. Primers and probes were designed for the N. meningitidis species-specific gene porA and for the non-species-specific cnl, and a modified species-specific sodC probe was designed based on a previously published method [18]. Genogroups were determined using the 2nd (A, W, X) and 3rd (B, C, Y) qPCR multiplexes developed by Wang et al. [23]. Primers and probes for these genogroups were kindly provided by the Centers for Disease Control and Prevention (CDC). Primers and probes were also designed for genogroups E, H, and Z using the Primer Express software from ABI. Minor Groove Binding (MGB) hydrolysis probes were designed, as opposed to Black Hole Quencher (BHQ) hydrolysis probes which are more difficult to manufacture. The sodC probe was modified in Primer Express to make it 12 base pairs shorter and accommodate an MGB quencher (Table 2). All primers/probe were tested at the following concentrations in a 25 μl solution: 900nM/250nM; 600nM/200nM; 300nM/100nM and 100nM/50nM respectively. Optimal concentrations in a TaqMan Gene Expression MasterMix (reference: 4369016 from ABI) were identified for each primers/probe set and used subsequently in the different assays. Cycling conditions were similar for all tests: 1 cycle of 2 min at 50°C, 1 cycle of 10 min at 95°C followed by 50 cycles of 15 s at 95°C and 1 min at 60°C. The ABI 7500 fast cycler was used to perform the reaction and the results analysed using the 7500 Fast software (see for detailed information).

Study design and statistical analyses

Initially, 16 positive controls, representing all target genogroups (A, B, C, W, X, Y, H, E, Z, cnl), were used for the optimization of both gel-based and qPCR in monoplex and appropriate multiplexes. Optimal qualitative results from monoplex and multiplex assays were evaluated for gel-based PCR by visualizing the quality and strength of the bands, whereas optimal conditions were defined for the qPCR assay using quantitative Ct values. Then, 44 Malian isolates were analysed using all primers and probes in monoplex and multiplex assays in parallel to further evaluate the optimized PCR conditions for multiplex assays and ensure the multiplex assays were comparable to monoplex assays. Later, as more samples were prospectively characterized at Oxford University from other MenAfriCar countries, samples from Chad, Ethiopia, Ghana, Niger and Senegal were tested with the optimized multiplex assays (Fig 1). Sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV) (see definitions in Table 3) were calculated for primers/probe triplets (forward primer, reverse primer and probe for each qPCR) using the results obtained from Oxford University as a gold standard.

Fig 1. Summary of the methodology employed in the study and samples tested.

Table 3. Definition of sensitivity, specificity, positive predictive value and negative predictive value.


Literature search and evaluation of primers and probes

Eleven papers describing 112 primers and 16 probes for gel-based or qPCR methods to detect and characterize meningococci were identified (S1 Table). Only papers that described the primer and probe sequences were used. Gel-based PCR assays targeting different genes were identified for N. meningitidis detection: 1 assay for crgA, 2 for porA, 2 for ctrA and 1 for sodC. Similarly, qPCR assays for characterization of genogroups were analyzed: 3 for A, 3 for W, 7 for X, 5 for B, 4 for C, 3 for Y, 1 for H, 2 for E, and 1 for Z.

For the gel-based PCR assays: the same primers were often chosen for detection and genogrouping in different papers. The most frequent were for characterization of genogroups A, W, Y, B, C [22], and X [33]. For the qPCR assays: primers/probe triplets (forward primer, reverse primer and probe) were retrieved for porA (1 triplet), ctrA (2 triplets), and sodC (1 triplet) for the detection of N. meningitidis and for genogroups A (1 triplet), W (2 triplets), X (1 triplet), B (3 triplets), C (3 triplets), Y (2 triplets). In one of the most recent comprehensive assays, two multiplex PCR for genogroups A, W, X and B, C, Y were used successively [23].

Nineteen (14 primers and 5 probes–see Tables 1 and 2) out of the 68 primers and probes sequences evaluated within the PubMLST database had 1 to 3 mismatches, compared with the consensus sequences, which was considered acceptable. The Tm and dimer analyses generally gave acceptable results for gel-based and qPCR, except for H and Z that had to be redesigned to fit the cycling conditions in the gel-based PCR assay (Tables A, B, C, D, E, F, G and H in S2 Table). A final list of 47 primers and 12 probes was evaluated for our detection and characterization of the PCR algorithm (Tables 1 and 2). Primers and probes regions are represented schematically on their respective loci in Panels A, B, C and D in S1 Fig.

Optimisation of the four multiplex assays (1- porA, sodC, cnl; 2- A, W, X; 3- Y, B, C; 4- H, E, Z)

Sixteen positive controls from different genogroups were used to optimize both the gel-based and the real-time PCR assays.

For gel-based assay, the optimal MgCl2 concentration for the assay was determined to be 10.4mM and this concentration was retained for the remaining experiments. Each reaction consisted of a 30μl solution with 0.5μl of each primer, 4μl of rlMMx (monoplex) or 6μl of rlMMx (multiplex) and molecular biology grade water (ddH2O).

Table 2 presents the optimal concentrations for the selected primers and probes for real-time PCR (qPCR). The qPCR reaction consisted of a 25μl solution with 12.5 μl of 2X Gene Expression Master Mix and ddH2O.

Evaluation of sensitivity, specificity, PPV and NPV compared with our gold standard (S3 Table).

Gel-based assay (Table A in S3 Table):

Sixteen positive controls and the samples from Mali (n = 44) were used to compare monoplex and multiplex PCR. All meningococci identified at the University of Oxford were correctly detected and characterized by monoplex gel-based PCR resulting in a sensitivity of 100% for all genes except 75% for genogroup Y (see S3 Table). The porA and sodC sensitivities were decreased to 82.14% and 96.43% respectively when working in multiplex whereas it was 100% for cnl and all genogroups. The specificity of porA assay increased from 77.78% in monoplex to 88.89% in multiplex, whereas specificity of sodC decreased from 29.64% in monoplex to 25.93% in multiplex. Specificity was 100% for cnl and all genogroups except X and Z with 96.55% and 96.30% respectively in monoplex and 100% for cnl and all genogroups except X and B with 96.55% and 96.15% respectively in multiplex (Supplemental Table 3).

Real-time PCR assay (Table B in S3 Table):

Starting with 100ng of DNA, eight 2-fold serial dilutions were performed to determine the Ct limit for categorizing a sample as positive. The last dilution, at a concentration of 0.79ng, systematically had a Ct lower or equal to 30. This is roughly equivalent to 40,000 genome copies which is far lower than the expected number of colony forming units after culture. The Ct limit was determined to be 30 cycles for crude DNA isolates. For this assay, a sample with a Ct lower or equal to 30 was considered positive and samples above 30 were considered negative (Fig 2).

Fig 2. qPCR standard curves.

Eight two-fold dilutions of DNA from each positive control were performed and tested in a qPCR assay with appropriate primers and probes. The Ct observed for each dilution is plotted against the DNA concentration demonstrating a decrease in Ct with increased DNA concentration and detection for all genogroups with a threshold of 30. sodC was empirically tested later and demonstrated the same threshold (data not shown).

The sensitivity was 100% for all PCR assays evaluated in this study both in monoplex and multiplex. Specificity was poor for sodC with 25.93% in monoplex and 29.63% in multiplex. Specificity for porA was slightly decreased from 81.48% to 74.07% in multiplex. All cnl and genogroups had a specificity of 100% both in monoplex and multiplex except genogroup X with 96.55% in monoplex. No statistical differences were observed between the new MGB sodC probe and the original BHQ sodC probe; however, the new MGB probe detected N. meningitidis at lower Ct values (data not shown) and was included in the final assay.

Evaluation of the optimized multiplex assay using the pool of African carriage isolates (Table 4)

The best performing assay was tested in multiplex on 203 isolates from five other countries of the MenAfriCar study with the gel-based and qPCR assays making a total of 247 samples. Again, the comparison was made with the gold standard results from Oxford: for the detection by gel-based PCR assay, the best sensitivity was obtained with sodC (80.92%—with a specificity of 69.47%) and the best specificity with porA (93.68% with a sensitivity of 70.39%). Sensitivity for cnl and all genogroups ranged from 66.67% (genogroup X) to 100% for genogroup B and E, whereas specificity ranged from 97.99% (genogroup X) and 100% (genogroups C, Y, H and Z).

For the qPCR assay, the best sensitivity and specificity was obtained with porA (96.05% and 91.58% respectively) showing no advantage of using sodC (94.74% and 77.89% respectively). For cnl and all genogroups, sensitivity ranged between 90% (genogroup A and Y) and 100%, whereas specificity ranged between 99.33% (genogroup X) and 100% (Table 4).

Table 4. Statistical analysis results for the multiplex gel-based and qPCR assays.

The benefit of evaluating results from sodC and porA together, instead of individually from the same multiplex, was analysed by comparing each primers/probe sets taken individually or together. Sensitivity and NPV were improved by using both genes together and considering positive any samples that were positive for at least one of the targets for the gel-based PCR assay in comparison with porA or sodC alone from the same multiplex whereas the specificity and PPV were decreased. For the qPCR assay all statistical results for sodC were lower or equal to those for porA. Inclusion of sodC did not provide additional value (Table 4).

From the panel of 247 samples from all six different countries, 152 were characterized as N. meningitidis at the University of Oxford using rplF sequencing [29]. Of these, 88.82% (135) isolates were correctly identified as meningococci by the gel-based PCR assay and 96.05% (146) by the qPCR assay. All genogroups, but H and Z, were identified in the panel tested: the most commonly represented was the cnl with 43.9%/48.4% (gel based/qPCR). Some genogroups considered as rare in the region were detected, such as B (2%) and E (0.66%) by both assays.


Conventional bacteriological methods that are usually routinely used to detect and characterize meningococci are complex, long and involve several overnight cultures and chemical tests which might generate error [15]. Developing a simple and rapid PCR algorithm that detects and characterizes carriage of meningococci is important, especially in the meningitis belt where recent vaccinations with conjugate vaccines were carried out and carriage is usually low [14]. Therefore, it is important to correctly detect N. meningitidis in a context where transmission is high especially in young children and where the period of carriage is short [34].

In this study, after an in silico evaluation of the existing primers and probes, we evaluated a multiplex PCR algorithm by comparison with techniques that were validated and published in a reference laboratory at the University of Oxford (24, 30). Our multiplex PCR algorithm was later used for a meningococcal carriage study in The Gambia and improved the detection of carried meningococci by more than twice (239%) [35].To our knowledge, this is the first time that such an extensive evaluation was made for a carriage study in the meningitis belt, using the most advanced tools and following recommended advanced guidelines for qualitative techniques [32].

To optimise the sensitivity of our algorithm for the detection of meningococci, we targeted both porA and sodC genes simultaneously. The published sodC gel-based PCR primers were re-designed due to too many mismatches against the consensus sequence [18]. Similarly, primers and probe of qPCR targeting porA available from the literature [17] yielded too many mismatches and were re-designed entirely. Designing qPCR primers and probes for cnl identification was not an easy task as the cnl sequence is short and variable, but the design of two complementary forward primers (Table 2), allowed the complete coverage of the cnl sequences that were present in pubMLST at that time. Recent comparison of primers/probes sequences used in this study with a larger number of sequences now available in the pubMLST identified additional point mutations in newly published alleles (data not shown). Although most alleles are still covered by the primers and probes developed here, new variations within these primers/probes sequences may lead to decreased sensitivity and we advise a complete new in silico evaluation of primers and probes before using these. Since pubMLST is exponentially growing, we recommend our methodology to be applied–or, at a minimum, blasting of primers and probes to make them more robust–when developing or implementing new molecular diagnosis tools for N. meningitidis as well as other pathogens. The variability of the targeted genes between meningococcal strains might explain the sensitivity and specificity variations observed between monoplex and multiplex assays and future studies will be needed to investigate this further for the detection of N. meningitidis.

A major strength of the developed algorithm is that the first multiplex assay (porA, sodC and cnl) allows the simultaneous detection of meningococci and the presence/absence of their capsule locus. Since cnl genotype is most common among meningococcal strains in carriage [11, 14], this assay improves the rapidity of N. meningitidis characterization. The sequential use of the following three multiplex assays for genogroup characterization, from the most frequent groups (A, W, and X followed by B, C, and Y) to the carriage uncommon genogroups (E, H and Z) in the meningitis belt, further reduces the time required for meningococcal characterization. The primers/probe sequences used for these genogroup assays could also be modified to fit the variable epidemiology of meningococci by region and over time.

In this study, we have developed an algorithm that: (i) has a high sensitivity for the detection of meningococci; and, (ii) allows rapid characterization of all genogroups with high specificity. A strength of gel-based assays is that the assays joined two commonly used genes to detect N. meningitidis: porA and sodC in multiplex, which improved the sensitivity from 70.39% for porA and 80.92% for sodC alone to 88.82% (see Table 4) for meningococcal detection. However, no improvement in sensitivity and specificity was observed when adding sodC in the multiplex qPCR assay and we advise using only porA/cnl for the qPCR. This might be due to the modification of the published sodC probe to accommodate a MGB quencher in order to harmonize the quenchers in the first multiplex (porA, sodC, cnl). Indeed, “BHQ” probes are difficult to obtain from oligonucleotide suppliers, because of their unusual design [23]. A subsequent comparison of 44 Malian samples tested with both BHQ and MGB probes showed no differences in detection of isolates. However, the efficacy of the shortened MGB sodC probe should be further investigated.

For genogroup characterization, multiplex qPCR that has already been evaluated were validated in gel-based PCR platform. The statistical results varied from one genogroup to the other (Table 4). For the two multiplexes (A, W, and X; B, C, and Y), published primers/probes were used and their statistical results were acceptable for all genogroups. The genogroup X primers set had the lowest sensitivity and PPV in the gel-based assay and may require redesign. Overall, the qPCR assay yielded better results than the gel-based PCR and should be the preferred assay for genogroup characterization.

Our study used the laboratory results obtained from Oxford University as reference standard for the statistical analysis. The sensitivity, specificity, PPV and NPV may be affected by the transport conditions when DNA samples were shipped from MenAfriCar countries and how the samples were processed at Oxford University. This may have affected our comparisons between PCR tests realized in Bamako and analyses in Oxford. In addition, no qPCR assay for the rare genogroups H, E, and Z was available in Oxford, making the direct comparison difficult. The results generated in this study were compared to the ctrA sequencing results when available and may explain some of the observed discrepancies; indeed, the comparison of two different techniques could explain the low PPV for group E in both assays.

The quality of the template may have also affected the detection of meningococci as most studies use DNA purified by kits. Boiled cell suspensions were purposely used to make our assays as inexpensive as possible and the good statistics obtained with those crude samples demonstrate the efficacy of the assays. Stringent rules were applied to determine whether a test was positive or not: by qPCR, a CT threshold of 30 was chosen based on preliminary results obtained (Fig 2), whereas the CDC/WHO reference guidelines recommend a threshold of 35 which, with the reagents and conditions used in our study, would have led to a considerable number of false positives.

In conclusion, this algorithm allows the detection of N. meningitidis with high sensitivity for samples from 6 different countries of the meningitis belt, it significantly reduces the number of tests required for meningococcal detection and allows very specific characterization of samples for carriage studies in comparison with conventional bacteriological techniques.

Supporting information

S1 Fig. Primers and probes positioned on appropriate genes.

Gel-based (green arrows) and rt (black arrows) PCR primers and probes for each multiplex (A-multiplex 1, B-multiplex 2, C-multiplex 3 and D-multiplex 4); the gene arrow pointing toward the 3’ end of the gene. Both forward primers (available in the literature and newly designed) for porA are shown on the gene (A). Figure adapted from Harisson OB and al. 2013.


S1 Table. List of papers that were selected for further analyses.


S2 Table. Primer and probe sequences general technical analysis.

Results of the BLAST analyses performed for each primer and probe using the NCBI BLAST tool. All BLAST results shown were done in the nucleotide collection database, against N. meningitidis (taxon id 487). The Tm and possible dimers were obtained using the primer express design tool.


S3 Table.

A. Statistical results for gel-based PCR assay. Statistical analysis results of the gel-based PCR assay. Sensitivity, Specificity, Positive Predictive Value (PPV) and Negative Predictive Value (NPV) were assessed for each primers pair in the multiplex and monoplex assay pooling results from test on positive controls (n = 16) and the Malian samples (n = 44) and presented as percentages (%). B. Statistical results for the qPCR assay. Statistical analysis results of the qPCR assay. Sensitivity, Specificity, Positive Predictive Value (PPV) and Negative Predictive Value (NPV) were assessed for each primers pair in the multiplex and monoplex assay pooling results from test on positive controls (n = 16) and the Malian samples (n = 44) and presented as percentages (%).


S1 Dataset. Database used for the analyses for the Chadian Site.


S2 Dataset. Database used for the analyses for the Ethiopian Site.


S3 Dataset. Database used for the analyses for the Ghanaian Site.


S4 Dataset. Database used for the analyses for the Malian Site.


S5 Dataset. Database used for the analyses for the Niger Site.


S6 Dataset. Database used for the analyses for the Senegalese Site.



K.D. Is a Wellcome Trust Tropical Fellow. O.B.H. and M.C.J.M. were supported by the Wellcome Trust (grant 087622). This research made use of the Neisseria Sequence Typing database website ( developed by Keith A. Jolley and Martin C. J. Maiden, hosted by the University of Oxford, and supported by the Wellcome Trust (grant 104992).

The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention.

^Institutions and individual members of the MenAfriCar consortium

Armauer Hansen Research Institute, Addis Ababa, Ethiopia: Oumer Ali, Abraham Aseffa (PI), Ahmed Bedru, Tsehaynesh Lema, Tesfaye Moti, Yenenesh Tekletsion, Alemayehu Worku, Haimanot Guebre Xabher (deceased), Lawrence Yamuah.

Centre de Recherche Médicale et Sanitaire (CERMES), Niamey, Niger (Member of the International Network of Pasteur Institutes): Rahamatou Moustapha Boukary, Jean-Marc Collard (PI), Ibrahim Dan Dano, Ibrahim Habiboulaye, Bassira Issaka, Jean-François Jusot, Sani Ousmane, Issoufa Rabe.

Centre de Support en Santé International (CSSI), N'Djamena, Chad: Doumagoum Moto Daugla (PI), Jean Pierre Gami, Kadidja Gamougam, Lodoum Mbainadji, Nathan Naibei, Maxime Narbé, Jacques Toralta.

Centre pour les Vaccins en Développement, Bamako, Mali: Abdoulaye Berthe, Kanny Diallo, Mahamadou Keita, Uma Onwuchekwa, Samba O Sow (PI), Boubou Tamboura, Awa Traore, Aliou Toure.

Centers for Disease Control, Atlanta, USA: Tom Clark, Xin Wang, Leonard Mayer.

Department of Community Medicine, University of Maiduguri, Maiduguri, Nigeria: Mary Amodu, Omeiza Beida, Galadima Gadzama, Babatunji Omotara (PI), Zailani Sambo, Shuaibu Yahya.

Faculty of Infectious Disease, London School of Hygiene & Tropical Medicine, London, UK: Daniel Chandramohan, Brian M Greenwood (PI), Musa Hassan-King, Olivier Manigart, Maria Nascimento, James M Stuart, Arouna Woukeu.

Princeton University, USA: Nicole E Basta

Public Health England Vaccine Evaluation Unit, Manchester, UK: Xilian Bai, Jay Lucidarme, Ray Borrow, Helen Findlow.

Institut de Recherche pour le Développement, Dakar, Senegal: Serge Alavo, Hubert Bassene, Aldiouma Diallo (PI), Marietou Dieng, Souleymane Doucouré, Jules François Gomis, Assane Ndiaye, Cheikh Sokhna, Jean François Trape.

Navrongo Health Research Centre, Navrongo, Ghana: Bugri Akalifa (deceased), Abudulai Forgor, Abraham Hodgson (PI), Isaac Osei, Stephen L Quaye, John Williams, Peter Wontuo.

University of Bristol, UK: Thomas Irving.

University of Cambridge, UK; Caroline L Trotter

University of Oxford, UK: Julia Bennett, Dorothea Hill, Odile Harrison, Martin C Maiden, Lisa Rebbetts, Eleanor Watkins.

MenAfriCar PI and corresponding author: Professor Brian Greenwood


  1. 1. Lapeyssonnie L. [Comparative epidemiologic study of meningococcic cerebrospinal meningitis in temperate regions and in the meningitis belt in Africa. Attempt at synthesis]. Medecine tropicale: revue du Corps de sante colonial. 1968;28(6):709–20. pmid:5739513.
  2. 2. Broome CV. The carrier state: Neisseria meningitidis. The Journal of antimicrobial chemotherapy. 1986;18 Suppl A:25–34. pmid:3091563.
  3. 3. Stephens DS, Greenwood B, Brandtzaeg P. Epidemic meningitis, meningococcaemia, and Neisseria meningitidis. Lancet. 2007;369(9580):2196–210. pmid:17604802.
  4. 4. Harrison OB, Brueggemann AB, Caugant DA, van der Ende A, Frosch M, Gray S, et al. Molecular typing methods for outbreak detection and surveillance of invasive disease caused by Neisseria meningitidis, Haemophilus influenzae and Streptococcus pneumoniae, a review. Microbiology. 2011;157(Pt 8):2181–95. pmid:21622526.
  5. 5. Nicolas P. [Meningococcal meningitis epidemics in sub-Saharan Africa and the meningococcal A conjugate vaccine]. Medecine et sante tropicales. 22(3):246–58. pmid:23186941.
  6. 6. Gamougam K, Daugla DM, Toralta J, Ngadoua C, Fermon F, Page AL, et al. Continuing effectiveness of serogroup A meningococcal conjugate vaccine, Chad, 2013. Emerging infectious diseases. 2015;21(1):115–8. pmid:25536336; PubMed Central PMCID: PMC4285275.
  7. 7. Kristiansen PA, Jorgensen HJ, Caugant DA. Serogroup A meningococcal conjugate vaccines in Africa. Expert review of vaccines. 2015:1–18. pmid:26358167.
  8. 8. Maurice J. Vaccine shortage threatens spread of meningitis in Niger. Lancet. 2015;385(9984):2241. pmid:26088485.
  9. 9. Abio A, Neal KR, Beck CR. An epidemiological review of changes in meningococcal biology during the last 100 years. Pathogens and global health. 2013;107(7):373–80. pmid:24392681; PubMed Central PMCID: PMC4083158.
  10. 10. Caugant DA, Tzanakaki G, Kriz P. Lessons from meningococcal carriage studies. FEMS microbiology reviews. 2007;31(1):52–63. pmid:17233635.
  11. 11. Claus H, Maiden MC, Maag R, Frosch M, Vogel U. Many carried meningococci lack the genes required for capsule synthesis and transport. Microbiology. 2002;148(Pt 6):1813–9. pmid:12055301.
  12. 12. Kristiansen PA, Diomande F, Wei SC, Ouedraogo R, Sangare L, Sanou I, et al. Baseline meningococcal carriage in Burkina Faso before the introduction of a meningococcal serogroup A conjugate vaccine. Clinical and vaccine immunology: CVI. 2011;18(3):435–43. pmid:21228139; PubMed Central PMCID: PMC3067389.
  13. 13. Kristiansen PA, Ba AK, Ouedraogo AS, Sanou I, Ouedraogo R, Sangare L, et al. Persistent low carriage of serogroup A Neisseria meningitidis two years after mass vaccination with the meningococcal conjugate vaccine, MenAfriVac. BMC Infect Dis. 2014;14:663. pmid:25472422; PubMed Central PMCID: PMC4267149.
  14. 14. MenAfriCar. The diversity of meningococcal carriage across the African meningitis belt and the impact of vaccination with a group A meningococcal conjugate vaccine. The Journal of infectious diseases. 2015. pmid:25858956.
  15. 15. MenAfriCar. Meningococcal carriage in the African meningitis belt. Tropical medicine & international health: TM & IH. 2013;18(8):968–78. pmid:23682910; PubMed Central PMCID: PMC3950817.
  16. 16. Guiver M, Borrow R, Marsh J, Gray SJ, Kaczmarski EB, Howells D, et al. Evaluation of the Applied Biosystems automated Taqman polymerase chain reaction system for the detection of meningococcal DNA. FEMS immunology and medical microbiology. 2000;28(2):173–9. pmid:10799809.
  17. 17. Jordens JZ, Heckels JE. A novel porA-based real-time PCR for detection of meningococcal carriage. Journal of medical microbiology. 2005;54(Pt 5):463–6. pmid:15824424.
  18. 18. Dolan Thomas J, Hatcher CP, Satterfield DA, Theodore MJ, Bach MC, Linscott KB, et al. sodC-based real-time PCR for detection of Neisseria meningitidis. PloS one. 2011;6(5):e19361. Epub 2011/05/17. pmid:21573213; PubMed Central PMCID: PMC3088665.
  19. 19. Jordens JZ, Williams JN, Jones GR, Heckels JE. Detection of meningococcal carriage by culture and PCR of throat swabs and mouth gargles. Journal of clinical microbiology. 2002;40(1):75–9. pmid:11773095; PubMed Central PMCID: PMC120086.
  20. 20. Bennett DE, Cafferkey MT. Consecutive use of two multiplex PCR-based assays for simultaneous identification and determination of capsular status of nine common Neisseria meningitidis serogroups associated with invasive disease. Journal of clinical microbiology. 2006;44(3):1127–31. pmid:16517911; PubMed Central PMCID: PMCPMC1393079.
  21. 21. Mothershed EA, Sacchi CT, Whitney AM, Barnett GA, Ajello GW, Schmink S, et al. Use of real-time PCR to resolve slide agglutination discrepancies in serogroup identification of Neisseria meningitidis. Journal of clinical microbiology. 2004;42(1):320–8. Epub 2004/01/13. pmid:14715772; PubMed Central PMCID: PMC321732.
  22. 22. Taha MK. Simultaneous approach for nonculture PCR-based identification and serogroup prediction of Neisseria meningitidis. Journal of clinical microbiology. 2000;38(2):855–7. pmid:10655397.
  23. 23. Wang X, Theodore MJ, Mair R, Trujillo-Lopez E, du Plessis M, Wolter N, et al. Clinical validation of multiplex real-time PCR assays for detection of bacterial meningitis pathogens. Journal of clinical microbiology. 2012;50(3):702–8. pmid:22170919; PubMed Central PMCID: PMC3295090.
  24. 24. Zhu H, Wang Q, Wen L, Xu J, Shao Z, Chen M, et al. Development of a multiplex PCR assay for detection and genogrouping of Neisseria meningitidis. Journal of clinical microbiology. 50(1):46–51. pmid:22090406.
  25. 25. LaForce FM, Konde K, Viviani S, Preziosi MP. The Meningitis Vaccine Project. Vaccine. 2007;25 Suppl 1:A97–100. pmid:17521780.
  26. 26. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ. Basic local alignment search tool. J Mol Biol. 1990;215(3):403–10. pmid:2231712.
  27. 27. Jolley K. PubMLST, Neisseria sequence Typing 2010. Available from:
  28. 28. Popovic TA, G; Facklam R. Laboratory manual for the diagnosis of meningitis caused by Neisseria meningitidis, Streptococcus pneumoniae and Haemophilus influenzae. Geneva: World Health Organisation; 1999.
  29. 29. Bennett JS, Watkins ER, Jolley KA, Harrison OB, Maiden MC. Identifying Neisseria species by use of the 50S ribosomal protein L6 (rplF) gene. Journal of clinical microbiology. 2014;52(5):1375–81. Epub 2014/02/14. pmid:24523465; PubMed Central PMCID: PMC3993661.
  30. 30. tecnologies ID. PrimerQuest. Available from:
  31. 31. Rozen S, Skaletsky H. Primer3 on the WWW for general users and for biologist programmers. Methods in molecular biology (Clifton, NJ. 2000;132:365–86. pmid:10547847.
  32. 32. Bustin SA, Benes V, Garson JA, Hellemans J, Huggett J, Kubista M, et al. The MIQE guidelines: minimum information for publication of quantitative real-time PCR experiments. Clin Chem. 2009;55(4):611–22. Epub 2009/02/28. pmid:19246619.
  33. 33. Yaro S, Drabo A, Ouangraoua S, Kirakoya-Samadoulougou F, Mueller J, Sanou O, et al. Neisseria meningitis serogroup X outbreak in Burkina Faso, 2009–2010. Open Journal of Internal Medicine2012. p. 41–29.
  34. 34. MenAfriCar C. Household transmission of Neisseria meningitidis in the African meningitis belt: a longitudinal cohort study. Lancet Glob Health. 2016;4(12):e989–e95. Epub 2016/11/20. pmid:27855873.
  35. 35. Manigart O, Okeakpu J, Odutola A, Jarju S, Foster-Nyarko E, Diallo K, et al. Alternative Molecular Methods for Improved Detection of Meningococcal Carriage and Measurement of Bacterial Density. Journal of clinical microbiology. 2016. pmid:27582517.