Representational difference analysis of Neisseria meningitidis identifies sequences that are specific for the hyper-virulent lineage III clone


  • Aldert Bart,

    Corresponding author
    1. Department of Medical Microbiology, Academic Medical Center, University of Amsterdam, P.O. Box 22660, 1105 AZ Amsterdam, The Netherlands
    Search for more papers by this author
  • Jacob Dankert,

    1. Department of Medical Microbiology, Academic Medical Center, University of Amsterdam, P.O. Box 22660, 1105 AZ Amsterdam, The Netherlands
    2. Reference Laboratory for Bacterial Meningitis, University of Amsterdam/RIVM, P.O. Box 22700, 1100 DE Amsterdam, The Netherlands
    Search for more papers by this author
  • Arie van der Ende

    1. Department of Medical Microbiology, Academic Medical Center, University of Amsterdam, P.O. Box 22660, 1105 AZ Amsterdam, The Netherlands
    Search for more papers by this author

*Corresponding author. Tel.: +31 (20) 5664863; Fax: +31 (20) 6979271, E-mail address:


Neisseria meningitidis may cause meningitis and septicemia. Since the early 1980s, an increased incidence of meningococcal disease has been caused by the lineage III clone in many countries in Europe and in New Zealand. We hypothesized that lineage III meningococci have specific DNA sequences, providing an opportunity to facilitate epidemiological studies by detecting lineage III isolates rapidly. Applying representational difference analysis on one lineage III tester strain and two non-lineage III driver strains, we identified three lineage III-specific sequences, probably part of a single locus encoding a restriction modification system. A PCR based on one of these sequences identified lineage III meningococcal isolates with a sensitivity of 100% and a specificity of 93%, which is superior to the serological identification of lineage III isolates.


Neisseria meningitidis or the meningococcus, an encapsulated Gram-negative bacterium, causes world-wide meningococcal disease. Acute meningococcal disease displays various clinical manifestations, meningitis, sepsis or both [1]. Since 1980, the incidence of meningococcal disease in The Netherlands has increased more than three-fold. This increase is mainly due to genotypically related serogroup B isolates, designated lineage III strains [2]. Recently, lineage III strains have been isolated in many other Western European countries, in Chile and in New Zealand [3]. In the latter country, lineage III causes an ongoing epidemic [4]. Since lineage III strains are particularly associated with an increased incidence of disease, this lineage is termed a ‘hyper-virulent lineage’[5].

Originally, strains of lineage III were distinguished serologically, since most strains expressed an antigenically distinct PorA outer membrane protein, the P1.4 subtype [6]. However, our studies have shown that non-lineage III strains of the P1.4 subtype were already present in The Netherlands more than 20 years prior to the increasing frequency of cases in the 1980s due to lineage III P1.4 strains [6,7]. On the other hand, the occurrence of lineage III may be underestimated since antigenic variation, possibly selected for by the host immune response, hampers the recognition of lineage III strains by the currently used serological methods. Indeed, eight out of 45 lineage III strains isolated in The Netherlands in the period 1984–1990 were not recognized by the P1.4-specific antibody [8]. This result provided circumstantial evidence for the exchange of porA genes between lineage III strains and other clones. Therefore, we hypothesized that the identification of lineage III-specific DNA sequences would provide a valuable tool for rapid detection of isolates of this clone for epidemiological purposes. The presence of such sequences was investigated with representational difference analysis (RDA) [9]. In short, this method preferentially amplifies DNA fragments that are present in one DNA population (tester), but absent or polymorphic in another DNA population (driver). Recently, RDA has been used for the identification of strain-specific or species-specific sequences in Neisseria species [10,11].

2Materials and methods

2.1Strains, plasmids and growth conditions

All N. meningitidis strains used in this study were isolates from patients with meningococcal disease collected by the Netherlands Reference Laboratory for Bacterial Meningitis (AMC, Amsterdam, The Netherlands and the RIVM, Bilthoven, The Netherlands). Meningococci were grown on heated blood (chocolate) agar plates at 37°C in a humidified atmosphere containing 5% CO2.

Lineage III strains used are strain 800615, isolated in 1980 as the first known lineage III strain being of electrophoretic type (ET) 24, and strain 882066, which is the reference strain for the P1.4 subtype. Strain 3532 (ET-30) and strain 830248 (ET-46) were used as non-lineage III strains. ET-30 and ET-46 belong to lineage IV, most closely related to lineage III [2], containing isolates causing endemic disease in the period 1958–1986 [6].

The presence or absence of a lineage III-specific sequence was studied in 104 serogroup B meningococcal isolates. These isolates comprised 30 lineage III strains, including all different ETs of this lineage, and 74 non-lineage III strains representing all 29 lineages or groups of ETs with a genetic difference of more than 0.30 that have been isolated in The Netherlands in the period 1958–1990 [2,8].

2.2Representational difference analysis

RDA was performed using primers JBam12 and JBam24, as described by Lisitsyn et al. [9], with the modifications according to Strathdee and Johnson [10].

2.3Cloning of RDA amplicons

Bands obtained after one or two rounds of RDA were excised from the gels and reamplified by PCR as described by Lisitsyn et al. [9]. After cloning in the TA vector pCRII or pCR2.1, they were transformed to Escherichia coli Top10F′ cells as recommended by the manufacturer (Invitrogen, Groningen, The Netherlands). Recombinants were selected by blue/white screening and ampicillin resistance. Plasmid was isolated using the Qiagen-tip 20 kit (Qiagen, Hilden, Germany) from one to five transformants for each excised RDA amplicon, and used for sequencing.

2.4Fluorescence-based sequencing

Recombinant plasmids were used as templates for sequencing with fluorescence dye-labeled universal M13 primers using Amersham Thermosequenase chemistry. Analysis was performed on an Applied Biosystems (Foster City, CA, USA) automatic sequencer (model 373), according to the instructions supplied by the manufacturer. The nucleotide sequence data will appear in the EMBL/GenBank/DDBJ nucleotide sequence databases under the accession numbers AF119281–AF119296.

2.5Southern hybridization

Southern hybridization to Sau3A-digested chromosomal DNA were performed using the digoxigenin (DIG) detection system (Boehringer Mannheim BV, Almere, The Netherlands). Probes used in Southern hybridization were synthesized and DIG-labeled by PCR amplification using the BamJ24 primer, 60 μM DIG-labeled dUTP, 140 μM dTTP, 200 μM dATP, 200 μM dCTP and 200 μM dGTP, and plasmid DNA inserted with RDA amplicons as templates. Hybridization conditions were as recommended by the manufacturer, at a temperature of 68°C.

2.6PCR detection

For the PCR on the tester strain and the two driver strains, with combinations of primers ABM1 (CAA TCA CAT CTC CAC CAT ACA ATA T), ABM2 (ATT TAG CAG GAT TTT TCA CAT ACC A), ABM8 (GAG ATT GTC CAA CTT TGT TTA GAT A), and MC1 (TAG CAC CAT GGG TTT AGA AAA TTT TCA AT), purified chromosomal DNA was used as template. The reaction mixture contained: 1 μM each primer, 200 μM each nucleotide (dATP, dCTP, dGTP, dTTP), 0.01% gelatin, 2.5 mM MgCl2, 50 mM KCl, 10 mM Tris–HCl, pH 8.3, and 2 U Taq polymerase (Perkin Elmer Cetus) in a final volume of 25 μl. Following initial denaturation at 95°C for 5 min, 35 amplification cycles were performed, each cycle comprising 1 min at 95°C, 1 min at 50°C and 2 min at 72°C, with a final incubation at 72°C for 10 min. Identical reaction conditions were used for PCR with primers ABM1 and ABM2 on 104 serogroup B isolates, for which template was generated by boiling a fraction of a frozen stock culture. As a positive control for the template, a separate PCR reaction detecting the porA gene was used with primers P21 and P22 as described [14]. PCR products were visualized on ethidium bromide-stained Tris–borate–EDTA gels.



The lineage III strain 800615 (ET-24) was used as tester for RDA amplifications with strains 3532 (ET-30), and 830248 (ET-46) as drivers. As negative control, RDA amplification with strain 800615 itself as driver was performed. To assess the robustness of RDA to detect differences between ET-24 and other ETs, RDA with strain 882066 (ET-24) as driver was performed. The reactions with the ET-30 and ET-46 driver strains yielded nine and seven discrete bands after one amplification round, respectively (Fig. 1). The control reactions of strain 800615 with the other ET-24 strain and itself as driver failed to yield amplicons (results not shown). A second RDA round on these bands did not yield additional bands (results not shown).

Figure 1.

Amplicons obtained after one round of RDA. Lane 1: amplicons obtained using the ET-30 strain 3532 as driver; lane 3: amplicons obtained using the ET-46 strain 830248 as driver; lane 5: 100-bp marker; lanes 2 and 4: empty.

Bands obtained after one or two rounds of RDA were excised from gel. Southern blots using probes generated by PCR on the excised amplicons suggested that one band contained different amplicons of similar sizes. Therefore, the amplicons were reamplified and cloned. The insertions of one to five different recombinant plasmids of the nine bands obtained after RDA with driver strain 3532 (ET-30) were further analyzed by sequencing. In total, 21 clones were sequenced, yielding 12 different sequences. Sequencing of 13 clones of the seven bands excised after RDA with driver strain 830248 (ET-46) yielded nine different sequences. Comparison between the amplicon sequences obtained after RDA with both driver strains showed that five amplicons were obtained for both RDA amplifications and these were further analyzed.

In Southern hybridization, two of the five amplicons hybridized with driver chromosomal DNA, and were not further analyzed. The three other amplicons were designated amplicon 222, amplicon 23A1, and amplicon 26A. BlastX searches of the GenBank database revealed a significant similarity between amplicon 222 and genes encoding methyltransferases of restriction modification systems. Amplicon 23D1 showed similarity to a methyltransferase gene as well. Amplicon 26A showed similarity to a gene for a restriction endonuclease. Based upon these findings these three amplicons were most likely located in close proximity on the lineage III chromosome. Therefore, primers were developed based on the three amplicon sequences. The results of PCR reactions with these primers on the ET-24 strains and the driver strains showed that the three amplicons 222, 23D1 and 26A were lineage III-specific, and located in the same chromosomal locus (Fig. 2).

Figure 2.

PCR reactions based on three amplicon sequences. A: Schematic representation of the deduced location of the RDA amplicons 26A, 222, and 23D1, and of primers based on these sequences. Primer ABM8 is based on the amplicon 26A sequence, primers ABM1 and ABM2 are based on the amplicon 222 sequence, and primer MC1 is based on the amplicon 23D1 sequence. B: PCR reactions were performed on the lineage III tester strain, a second lineage III strain, and the two driver strains. Primer pairs and template chromosomal DNA used are indicated. 24: ET-24 strain 800615 (tester strain); 30: ET-30 strain 3532 (driver); 46: ET-46 strain 830248 (driver); 24*: ET-24 strain 882066. As a template control, P21 and P22 were used to amplify the porA gene, present in all four strains. The combinations ABM2/MC1, ABM8/ABM1, and ABM8/MC1 yield products in the lineage III strains but not in the driver strains. This shows that the amplicons are specific and located in close proximity to each other in the order 26A-222-23D1, as indicated in A.

3.2Distribution of the lineage III-specific amplicon 222 among 104 serogroup B strains

A PCR with primers ABM1 and ABM2, based on the amplicon 222 sequence, was performed to assess the presence of amplicon 222 sequences in 104 serogroup B clinical isolates. The PCR results showed that the 30 lineage III strains were positive for amplicon 222, whereas five of the 74 (6.8%) non-lineage III strains were positive.


RDA proved to be a valuable tool to detect lineage III-specific DNA sequences. Three lineage III-specific sequences, located on the same chromosomal locus, were identified. Database homologies suggest that this locus encodes a restriction modification system. Two other amplicons obtained with both driver strains were not specific in Southern hybridization. These amplicons probably reflect restriction fragment length polymorphisms between the tester and driver strains, which can be detected by RDA [9].

The PCR, based on amplicon 222 in order to identify lineage III isolates, had a sensitivity of 100% (30/30), and a specificity of 93% (69/74). The sensitivity of the currently used lineage III marker [4] serosubtype P1.4 is 91% (49/54) [8], because the P1.4 epitope is not expressed or has undergone antigenic variation in the remaining 9%. The latter may be due to three different mechanisms: horizontal gene transfer, partial deletion of the epitope-encoding region and phase variation of expression of the porA gene [7]. The specificity of the monoclonal antibodies as serological markers for lineage III is 83% (39/47) [8]. This is caused by horizontal gene transfer, yielding non-lineage III strains expressing the P1.4 epitope [7].

Antigenic variants of the PorA protein are thought to be selected for by the host immune system [15]. In contrast, amplicon 222 encodes part of a restriction modification system, for which low selective pressure for variation is expected. Also, restriction modification systems behave as selfish genes, resisting loss by their host [16]. These properties of restriction modification systems have been useful in the phage-typing method for many bacteria in the past, and possibly explain the higher sensitivity of the PCR based on amplicon 222 as compared to that of the serosubtyping method in the identification of lineage III isolates. The higher specificity of the PCR-based lineage III identification can be explained by a lower rate of successful horizontal gene transfer of an entire methyltransferase gene, as compared with that of the smaller P1.4 epitope encoding region.

In conclusion, lineage III strains carry specific sequences which are absent in the majority of non-lineage III strains. A PCR based on part of these sequences is superior to the conventional serological test for serosubtyping in identifying lineage III meningococcal isolates.