OUP user menu

Lack of antigenic diversification of major outer membrane proteins during clonal waves of Neisseria meningitidis serogroup A colonization and disease

Charlotte A. Huber, Valentin Pflüger, Abdul-Wahab M. Hamid, Abudulai A. Forgor, Abraham Hodgson, Ali Sié, Thomas Junghanss, Gerd Pluschke
DOI: http://dx.doi.org/10.1111/2049-632X.12000 4-10 First published online: 1 February 2013


In particular in the ‘meningitis belt’ of sub-Saharan Africa, epidemic meningococcal meningitis is a severe public health problem. In the past decades, serogroup A lineages have been the dominant etiologic agents, but also other serogroups have caused outbreaks. A comprehensive vaccine based on subcapsular outer membrane proteins (OMPs) is not available. Here, we have investigated whether meningococcal populations overcome herd immunity by changing antigenic properties of their OMPs. Meningococcal isolates were collected in the context of longitudinal studies in Ghana between 2002 and 2008 and in Burkina Faso between 2006 and 2007. Serogroup A strains isolated during two clonal waves of colonization and disease showed no diversification in the genes encoding their PorA, PorB, and FetA proteins. However, we detected occasional allelic exchange of opa genes, as well as wide variation in the number of intragenic tandem repeats, showing that phase variation of Opa protein expression is a frequent event. Altogether we observed a remarkable antigenic stability of the PorA, PorB and FetA proteins over years. Our results indicate that while herd immunity may be responsible for the disappearance of meningococcal clones over time, it is not a strong driving force for antigenic diversification of the major OMPs analyzed here.

  • Neisseria meningitidis
  • outer membrane protein
  • herd immunity


Neisseria meningitidis, a Gram-negative diplococcus, is an obligate commensal of humans, usually carried in the upper airways of healthy individuals. The bacterium is transmitted to close contacts by the spread of respiratory secretions. Occasionally, N. meningitidis strains might penetrate the mucosal membrane, multiply in the bloodstream, and gain access to the cerebrospinal fluid. The overall incidence rate of meningococcal disease in Europe and North America is 1–3 per 100 000 person-years. The highest incidence rates are found in countries of sub-Saharan Africa, in the so-called ‘meningitis belt’, which extends from Ethiopia to Senegal. In these countries, annual incidence rates of as high as 1000 per 100 000 person-years are recorded during the most severe epidemics (Caugant, 2008). The savannah climate of the meningitis belt with an annual precipitation index of 300–1100 mm and extremely dry and hot winter seasons seems to be optimal for transmission of invasive meningococcal disease. Epidemics classically occur in the dry season, between December and April. They nearly always start in the early part of the dry season and then stop abruptly at the onset of rains. High temperature, low absolute humidity and the harmattan (a dusty wind that blows from the Sahara) at the end of the dry season may favor the occurrence of meningococcal disease by damaging the local mucosal defenses (Greenwood, 1999).

Based on different capsular polysaccharide structures, 13 serogroups of N. meningitidis can be distinguished. The vast majority of invasive meningococcal disease is caused by six of these serogroups (A, B, C, W-135, X, and Y) (Stephens et al., 2007). Most of the large epidemics in Africa in the past 100 years were associated with serogroup A (Caugant, 2008). Until recently, reactive immunization campaigns with polysaccharide vaccines initiated after the onset of meningitis epidemics have been used to limit epidemics. Now a monovalent conjugate vaccine against serogroup A meningococci, the most common cause of large epidemics, is being introduced stepwise in the countries of the African meningitis belt. In contrast to the unconjugated polysaccharides, conjugate vaccines will not only protect the vaccinated from invasive disease, but are also expected to block transmission of the meningococci expressing the vaccine serogroup.

Lack of a stable nasopharyngeal population of apathogenic meningococci may contribute to the vulnerability of the local population of the meningitis belt to major epidemics (Leimkugel et al., 2007, b). These are caused by hypervirulent clones of N. meningitidis, which may travel through the entire meningitis belt, but seem to remain only for a few years associated with a particular population (Leimkugel et al., 2007, b). Herd immunity may lead to the disappearance of individual clonal complexes, but new clones with a different antigenic make-up may subsequently spread. Clonal waves of colonization and disease are therefore a characteristic feature of the epidemiology of meningococcal meningitis in the African meningitis belt (Leimkugel et al., 2007, b). Neisseria meningitidis serogroup A strains of the clonal complex 5 that are associated with the sequence types (STs) 5, 7, and 2859 have been responsible for outbreaks in the last two decades (Nicolas et al., 2005; Teyssou & Muros-Le, 2007; Sie et al., 2008). Molecular typing approaches based on DNA sequencing allow meningococci to be distinguished and tracked (Jolley et al., 2007). During the last decade, multi-locus sequence typing (MLST) based on the identification of the alleles of fragments of seven meningococcal housekeeping genes has developed into the ‘gold standard’ for typing these bacterial pathogens (Maiden et al., 1998; Brehony et al., 2007).

Meningococci are naturally competent for transformation by exogenous DNA, and high rates of recombination have been observed (Jolley et al., 2005). However, most genetic exchange appears to takes place between closely related meningococci, and it has been suggested that recombination may be primarily a mechanism for genome repair that will only occasionally result in generation of diversity (Caugant & Maiden, 2009). Nevertheless, N. meningitidis is genetically and antigenically highly diverse (http://pubmlst.org/neisseria/). On the other hand, hyper-invasive lineages seem to be surprisingly stable over decades and during global spread (Caugant, 2008). This may facilitate development of a comprehensive protein-based vaccine, effective against a broad range of hyper-virulent meningococci. Cross-reactivity could be achieved by targeting antigenically invariant subcapsular structures or by combining a cocktail of vaccine antigens selected on the basis of molecular epidemiological studies. A number of meningococcal surface structures are thought to play a role in mucosal colonization, hematogenous spread and crossing of the blood brain barrier (Hill et al., 2010). To escape immune surveillance, meningococci have developed a range of mechanisms to change surface components. We assume that development of herd immunity is responsible for the complete disappearance of meningococcal clones after a few years of colonization of populations in the African meningitis belt (Leimkugel et al., 2007a, b). Here we have investigated whether meningococcal populations escape from immune detection by varying their outer membrane proteins (OMPs) PorA, PorB, FetA, and Opa. The serogroup A ST7 and ST 2859 meningococci analyzed have been collected between March 2002 and April 2008 in the course of longitudinal meningococcal colonization and disease surveys in Ghana and Burkina Faso (Leimkugel et al., 2007, b; Sie et al., 2008).

Materials and methods

Bacterial isolates

The N. meningitidis isolates investigated in this study had been collected in the Kassena-Nankana District (KND) and the neighboring district of Bawku in Ghana and in the Nouna Health District (NHD) in the Kossi region of Burkina Faso. Case strains were isolated from the cerebrospinal fluid of meningitis patients, and carriage strains were isolated from throat swabs collected in the context of longitudinal carriage surveys. Isolation and characterization of strains has been described previously (Gagneux, 2000; Leimkugel et al., 2007a, b; Sie et al., 2008). For the analysis of genetic diversification, serogroup A ST7 and ST2859 strains isolated from cases and carriers at different times during clonal colonization and disease waves were selected from our strain collection. Included were nine ST7 carriage and seven ST7 case isolates collected between March 2002 and March 2005 in the KND of Ghana. In the case of ST2859 strains, we analyzed six case and seven carriage isolates collected between March 2006 and March 2007 in the NHD of Burkina Faso, three case and 15 carriage isolates from the KND, as well as three case isolates collected between March 2007 and April 2008 in the neighboring district of Bawku.

Genetic analysis

DNA was extracted from bacterial pellets using the Wizard® Genomic DNA Purification Kit (Promega AG, Duebendorf, Switzerland). The DNA concentration was measured using a Nano drop Spectrophotometer (Witec Ag, Littau, Switzerland). PCR was performed using 5 μL of 10× BD buffer and 1 μL of FirePol Taq polymerase, 1.25 mM MgCl2 (Solis BioDyne, Tartu, Estonia), 100 ng of genomic DNA or the equivalent volume of nuclease-free water as a negative control, a 0.2 μM concentration of each forward and reverse primer, and a 0.2 mM concentration of each deoxynucleoside triphosphate in a total volume of 50 μL. PCRs were run in a T Professional Basic PCR machine (Biometra GmbH, Göttingen, Germany). The thermal profile for PCR amplification included an initial denaturation step of 94 °C for 2 min, followed by 35 cycles of 94 °C for 30 s, annealing at 55 °C for 30 s, and elongation at 72 °C for 1 min 20 s up to 2 min. The PCRs were finalized by an extension step at 72 °C for 10 min. PCR products were analyzed on 1% agarose gels by gel electrophoresis using ethidium bromide staining and the AlphaImager illuminator and AlphaImager software (Alpha Innotech, San Leandro, CA). PCR products were purified using a NucleoSpin purification kit (Macherey-Nagel GmbH & Co. KG, Düren, Germany) and subjected to direct sequencing or cloned using a TOPO TA cloning kit (Invitrogen Corp., Carlsbad, CA), transformed into Novablue competent cells (Merck, Darmstadt, Germany), and sequenced after DNA preparation (Macherey-Nagel GmbH & Co. KG, Düren, Germany). Sequencing was performed by Macrogen Inc, Seoul, Korea. For PorA, PorB, and FetA typing, primers were used as previously described (Sacchi et al., 1998; Thompson et al., 2003; Russell et al., 2004). Primers used for amplification and sequencing of the opa genes and their flanking regions are listed in Table 1.

View this table:
Table 1

Amplification and sequencing primers for opa genes and their flanking regions

OpaBAmplification primersOpaB2fCAGGACAAGGCGACGAG
OpaDAmplification primersOpaD5fTCTCCGTAGAGGAAATGATGC
OpaJAmplification primersOpaj5fCGCCCCAAACCTGATATAGT
Sequencing primersOpaj5fCGCCCCAAACCTGATATAGT

Design of the primers used for amplification of opaA, opaB, and opaD was based on the serogroup A strain Z2491 genome sequence (Parkhill et al., 2000) and the primer3 program (http://frodo.wi.mit.edu/). The sequences obtained were analyzed making use of the N. meningitidis homepage (http://neisseria.org/nm/) and the Neisseria sequence typing homepage (http://pubmlst.org/neisseria/). The multiple sequence alignment websites Multalin (http://bioinfo.genopole-toulouse.prd.fr/multalin/multalin.html) and clustalw2 (http://www.ebi.ac.uk/Tools/clustalw2/index.html) were used for comparisons between Opa gene sequences and their flanking regions. The emboss transeq tool was used for the translation of the retrieved opa gene sequences (http://www.ebi.ac.uk/Tools/emboss/transeq/).

Results and discussion

Humans are the only natural host for N. meningitidis, and nasopharyngeal carriage rates are known to be much higher than disease rates. In a longitudinal study in the northern Ghana, we have observed waves of colonization and disease with hyper-virulent clones of serogroup A meningococci. About 4 years after their local emergence, these clones disappeared completely (Leimkugel et al., 2007, b). This may be attributed to the development of herd immunity against the colonizing clone in the local population. During their spread from country to country, meningococcal clones undergo microevolution with recombination as major driving force for genetic variation (Achtman, 1995). As a consequence, clones that, based on MLST, appear to be closely related, may differ – due to multiple recombination events – very substantially in their antigenic profiles A. Lamelas and G. Pluschke, Swiss Tropical and Public Health Institute.

Antibodies directed against OMPs have been implicated in the development of natural immunity against meningococci. In particular, antibodies directed against the PorA and PorB proteins seem to provide serosubtype-specific protection (Jordens et al., 2004). This prompted us to investigate whether immune selection pressure building up in the population during colonization waves leads to an antigenic diversification of OMPs in the colonizing meningococcal population.

The strains included in the sequence analysis of OMP-encoding genes have been isolated from cerebrospinal fluid (CSF) of meningitis patients or from the pharynx of healthy carriers during two sequential colonization and disease waves in northern Ghana and during a meningococcal disease outbreak in Burkina Faso. Serogroup A ST7 strains have been isolated from the start (2002) till the end (2005) of a colonization wave in northern Ghana (Leimkugel et al., 2007, b). In the case of serogroup A ST 2859, isolates from an outbreak in Burkina Faso (2006–2007) (Sie et al., 2008) and in northern Ghana (2007–2008) were compared.

It has been proposed that the propensity of N. meningitidis to accumulate mutations increases dramatically toward the end of an epidemic, presumably due to immune pressure, and that with time, variants arise that are founders of new clonal complexes (Achtman, 2004). However, in none of the strains investigated here a single mutational change in the genes encoding the surface proteins PorA, PorB, or FetA was found. All PorA sequences had the VR1, VR2: 20, 9 variable regions. The Por B sequences were invariably of class 3 and had the allele 47. The variable region of Fet A was F3-1. In spite of microevolution of the epidemic clones, as detected by pulsed field gel electrophoresis (Leimkugel et al., 2007a, b), the sequence of the analyzed OMP-encoding genes thus was strikingly stable. No mutations were detected in the variable regions of the surface proteins PorA, PorB, and FetA in any of the strains analyzed. The benefit of accumulating mutations to escape herd immunity may be dispossessed by the fitness cost that such mutations confer (Achtman, 2004). Furthermore, the lack of a genetically diverse pharyngeal flora of N. meningitidis, in the study population (Leimkugel et al., 2007a, b), limiting the extent of horizontal genetic exchange (Achtman, 1995) may contribute to this striking lack of diversification. On the other hand, immune selection may act more strongly on exposed antigens other than the major OMPs analyzed here.

While the Ghanian meningococcal isolates analyzed here have been collected over a longer period of time (between 2002 and 2008), conclusions on the stability of OMPs for the strains from Burkina Faso are less strong, as the analyzed strains were isolated over a period of only 2 years. Another very important point to consider is that, although we did not detect any diversification in the genes encoding PorA, Por B, and FetA, expression levels of PorA and FetA are subject to variation. Antibodies elicited against these proteins in the host population may give selective advantage to meningococcal strains with decreased expression levels of PorA and FetA (Crowe et al., 1989; Carson et al., 2000). PorB expression levels have been described to be constant, however (Abad et al., 2006).

The multiple copies of genes encoding the Opa proteins can be turned on and off during chromosomal replication by slipped-strand mispairing of tandem CTTCT repeats present in their open reading frames. Here, we have used the published genome sequence of the serogroup A ST4 strain Z2491 to design primers specific for the flanking regions of the opaA, opaB, and opaD genes. In contrast to strain Z2491, the ST7 and ST2859 strains analyzed here all contained in addition an opaJ gene, as has been described for other ST5-complex/subgroup III strains (Callaghan et al., 2006). While all PorA, PorB, and FetA had the same alleles irrespective of the time point and location of isolation as well as the ST of the strains examined (Table 2), two of the four opa genes analyzed were subject to some variation. While all the ST7 strains from Ghana and the ST2859 strains from Burkina Faso had the opaA allele 242, all Ghanaian ST2859 strains had the allele 253. No variation was observed within the three individual groups of epidemiologically related isolates. ST7 and ST2859 are MLST single-locus variants, and the ST2859 genocloud may have developed from ST7 meningococci in Africa. The change in the opaA allele in the ST 2859 strains that emerged in Ghana in 2008 may have been due to homologous recombination and immune selection against the 242 allele (Callaghan et al., 2008; Levin & Cornejo, 2009). In the case of the opaJ gene, all ST7 isolates had the allele 213, whereas the ST2859 strains both from Ghana and from Burkina Faso had the allele 127.

View this table:
Table 2

Origin and characteristics of serogroup A meningococci and their OMPs analyzed

1396CarriageKNDMarch 2002720, 93-47F3-124214253 12 296 9 2137
1577CarriageKNDMarch 2003720, 93-47F3-124213253 12 296102138
1808CaseKNDJanuary 2004720, 93-47F3-1242 9 2538296 12 2138
1813CaseKNDFebruary 2004720, 93-47F3-1242 9 2538296 12 2138
1812CaseKNDFebruary 2004720, 93-47F3-124210253 9 29610213 9
1822CaseKNDFebruary 2004720, 93-47F3-124214253 12 29610213 9
1838CaseKNDMarch 2004720, 93-47F3-124216253 12 29610213 9
1902CarriageKNDMarch 2004720, 93-47F3-12421425311296 9 213 9
1990CarriageKNDMarch 2004720, 93-47F3-1242 15 2535296 12 21310
1991CarriageKNDNovember 2004720, 93-47F3-1242 15 2535296 12 21310
2008CaseKNDFebruary 2005720, 93-47F3-124211253 9 29613213 9
2009CaseKNDFebruary 2005720, 93-47F3-1242112531029614213 9
2018CarriageKNDMarch 2005720, 93-47F3-12428253829610213 9
2019CarriageKNDMarch 2005720, 93-47F3-124282531129610213 9
2020CarriageKNDMarch 2005720, 93-47F3-124211253829613213 9
2021CarriageKNDMarch 2005720, 93-47F3-124211253829612213 9
2173CaseNHDMarch 2006285920, 93-47F3-12428253829610127 12
2171CaseNHDMarch 2006285920, 93-47F3-124282538296 9 127 12
2172CaseNHDMarch 2006285920, 93-47F3-124282538296 9 127 12
2174CaseNHDMarch 2006285920, 93-47F3-124282538296 9 127 12
2175CaseNHDMarch 2006285920, 93-47F3-124282538296 9 127 12
2176CaseNHDMarch 2006285920, 93-47F3-124282538296 9 12711
2202CarriageNHDMarch 2006285920, 93-47F3-1242 9 25382961012713
2243CarriageNHDMarch 2006285920, 93-47F3-124282538296 9 12711
2365CarriageNHDNovember 2006285920, 93-47F3-124272539296712713
2378CarriageNHDNovember 2006285920, 93-47F3-1242142538296 9 12717
2554CarriageNHDMarch 2007285920, 93-47F3-12421025311296 9 127 9
2560CarriageNHDMarch 2007285920, 93-47F3-124282538296 12 127 12
2587CarriageNHDMarch 2007285920, 93-47F3-124282538296 12 1278
2537CaseBawkuMarch 2007285920, 93-47F3-1253825382961012711
2539CaseBawkuMarch 2007285920, 93-47F3-12538253829610127 12
2541CaseBawkuMarch 2007285920, 93-47F3-12538253829610127 12
2545CaseKNDApril 2007285920, 93-47F3-12538253829610127 12
2502CarriageKNDApril 2007285920, 93-47F3-12538253829610127 12
2620CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2622CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2624CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2626CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2628CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2630CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2632CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2635CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2669CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2671CarriageKNDNovember 2007285920, 93-47F3-12538253829610127 12
2699CarriageKNDApril 2008285920, 93-47F3-1253102537296 12 127 12
2700CarriageKNDApril 2008285920, 93-47F3-1253102537296 12 127 12
2701CarriageKNDApril 2008285920, 93-47F3-1253102537296 12 127 12
2703CarriageKNDApril 2008285920, 93-47F3-1253102537296 12 127 12
2708CaseKNDApril 2008285920, 93-47F3-1253102537296 12 127 12
2707CaseKNDFebruary 2008285920, 93-47F3-125382538296 12 12710
  • The opa genes that were found to be in frame are framed.

The numbers of pentanucleotide repeats within the coding sequences of the opa genes were subject to much wider variation. However, strains isolated at the same location and a similar time point tended to be relatively uniform. For example, all Ghanaian ST2859 isolates from 2007 had 8 opaA, 8 opaB, and 10 opaD repeats, while 5/6 isolates from 2008 had 10 opaA, 7 opaB, and 12 opaD repeats. This variation in numbers of tandem repeats is due to slipped-strand mispairing and results in phase variation involving on/off expression of the respective opa gene (Murphy et al., 1989). Alleles with a number of pentanucleotide repeats are a multiple of three represent functional genes (Table 2). Antigenic variation in the Opa genes has been suggested to mediate immune evasion (Davidsen & Tonjum, 2006; Callaghan et al., 2008). Functional open reading frames were found for all four opa genes, but no more than two functional opa genes were found in any of the strains. A Poisson regression analysis was performed (http://www.fisherstat.com/Pages/default.aspx), and disease and colonization isolates did not differ significantly in the Opa protein expression patterns (Table 2). While the invasive bacteria may represent rare genetic variants of the colonizing meningococcal population, a change in the Opa protein expression pattern thus does not seem to be relevant for an adaptation for multiplication in the blood stream and the CSF.

Control of meningitis epidemics has relied so far on reactive vaccination strategies with polysaccharide vaccines, and a serogroup A conjugate vaccine is currently being introduced in the meningitis belt. In our study, we have observed an apparent lack of diversification of PorA, PorB, and FetA, and a tailor-made vaccine targeting those OMPs may be able to control the epidemic strains investigated. However, to develop a comprehensive vaccine based on subcapsular OMPs, whole-genome analyses of epidemiologically well-defined isolate collections may be needed. Comparative genomics may give insight into the driving forces behind the microevolution of N. meningitidis, allowing a more comprehensive analysis of antigen variability, as well as the identification of new vaccine targets. Results of such analyses may thus help identifying antigens suitable for inclusion into a multivalent protein subunit vaccine (Tan et al., 2010).


This work was supported by the Volkswagen Foundation through the Collaborative Research Grant (SFB) 544 ‘Control of Tropical Infectious Diseases’. This publication made use of the Neisseria Multi-Locus Sequence Typing website (http://pubmlst.org/neisseria/) developed by Keith Jolley and Man-Suen Chan (Jolley et al., 2004, BMC Bioinformatics, 5: 86) and sited at the University of Oxford. The development of this site has been funded by the Wellcome Trust and European Union.


View Abstract