INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The major outer membrane protein PorA of Neisseria meningitidis is of interest, since its antigenic variation is used for the subtyping of meningococci (8, 24, 25). In addition, it is under investigation as a component of experimental vaccines against meningococcal infection (13).

The immunization of mice with outer membrane protein complexes results in bactericidal antibodies mainly directed against PorA (35, 36). Its value as a candidate vaccine is derived from experiments in which monoclonal antibodies directed against subtype-specific epitopes on PorA were effective in bactericidal assays and conferred protection in an animal model. However, PorA is subject to antigenic variation, which is thought to be overcome by including multiple antigenic variants of PorA in the vaccine (43). Already, trials with a hexavalent PorA-based vaccine and vaccines in which PorA is a major component have been performed (15).

PorA is expressed by most of the clinical isolates, but its level of expression varies widely (18, 42). Since the stable expression of this protein in meningococci during disease is a prerequisite for the PorA vaccine to be effective, the genetic mechanism of the variable expression of PorA has to be elucidated. Recently, we reported PorA phase variation at the transcriptional level, mediated by a variable polyguanidine stretch between the 10 and 35 domains of the porA promoter (42). In this study we describe a porA-negative meningococcus isolated from a patient with meningococcal disease. In addition, porA-negative variants were selected in vitro from two of nine different isolates. To elucidate the mechanism involved in the deletion of porA, DNA sequences upstream and downstream of this gene were evaluated.

Sequence analysis of the deletion variants indicated that in these three strains a recombination between regions of homology upstream and downstream of porA of 116 bp has occurred. The rise of porA deletion variants during a meningococcal infection could possibly be a mechanism to evade the host immune defense. Therefore, the protective efficacy of a vaccine on the basis of PorA may be limited.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains, culture conditions, and chromosomal DNA isolation. From the cerebrospinal fluid (CSF) and blood of the same patient N. meningitidis 860183 (C:4:P1.1 [CSF] and C:4:P1.NT [blood]) isolates were collected by the Reference Laboratory for Bacterial Meningitis (RLBM), University of Amsterdam, in 1986. In addition, nine isolates, strains 890456 (B:16:P1.5), 900545 (B:4:P1.15), 900111 (B:15:P1.16), 2996 (B:2b:P1.2), 900181 (B:2b:P1.2), 900619 (B:2b:P1.2), 901569 (C:2a:P1.2), H44/76 (B:15:P1.7,16), and H355 (B:15:P1.15), from the collection of the RLBM were used for in vitro studies. The latter two strains were isolated from patients with meningococcal disease during the epidemic period in Norway in the 1970s and are now used as reference strains.

Colony immunoblotting. Colonies were transferred to nitrocellulose filters (0.45-mm pore size; Schleicher and Schuell, Dassel, Germany) and immunologically stained as described before (18).

Detection of porA by colony hybridization. Colonies were transferred onto a nylon membrane (Hybond-N; Amersham International plc, Little Chalfont, Buckinghamshire, England) by replica plating. The colonies were lysed and denatured, essentially as described by Sambrook et al. (34). Briefly, filters were placed on GB003 gel-blotting paper (Schleicher and Schuell) saturated with 10% sodium dodecyl sulfate for 5 min, transferred to a second sheet of paper saturated with 0.5 M NaOH-1.5 M NaCl, and incubated for 10 min. The filters were neutralized by incubating them on paper saturated with 1.5 M NaCl-0.5 M Tris (pH 8.0) for 5 min. The filters were then transferred to paper saturated with 2× SSPE (1× SSPE is 0.18 M NaCl, 10 mM NaH₂PO₄, and 1 mM EDTA [pH 7.7]) and incubated for 5 min. The filters were allowed to dry at room temperature and finally baked at 80°C for 1 h. porA was detected by hybridization with a digoxigenin (DIG)-labeled PCR product as described for the Southern hybridization.

Southern hybridization. DNA fragments were electrophoresed on a 0.6% agarose gel and transferred to a nylon membrane (Zeta Probe; Bio-Rad) (34). The porA- and IS1106-specific probes were made by PCR amplification with primers PorA5 and P22 and IS1 and IS2, respectively. Probes were randomly primed and labeled with DIG (Boehringer, Mannheim, Germany) according to the instructions supplied by the manufacturer. After hybridization the probes were detected with anti-DIG antibodies conjugated to alkaline phosphatase and by staining according to the instructions supplied by Boehringer.

Oligonucleotide synthesis. The oligonucleotides used in this study are shown in Table 1 and Fig. 1. Oligonucleotides were synthesized by Perkin-Elmer Nederland B.V., Gouda, The Netherlands.

View larger version (8K): [in this window] [in a new window]   FIG. 1.   Schematic representation of the locations of the PCR primers used in this study.

Detection of porA by PCR. The presence of the porA gene was assessed by PCR with primers P21 and P22 as described previously (24). The PCR products were analyzed on agarose (1%) gels with the Tris-acetate-EDTA buffer system (34).

Strategy for determination of the sequences of the regions upstream and downstream of porA. Primer PorA11, homologous to a sequence downstream of porA and the IS1106 region, was designed according to the sequence obtained after inverse PCR of the EcoRI restriction enzyme fragment that hybridized with the IS1106-specific probe as well as with the porA gene probe. The EcoRI restriction fragment containing the 3' part of porA of the chromosomal DNA from strain H355, H44/76, or 860183 was used as the template in a PCR with primers IS1 and PorA11. After reamplification with IS41 and PorA111 as primers, amplicons were inserted into the BamHI/EcoRI-linearized vector pUC18 or pUC19 (Invitrogen Corporation, Carlsbad, Calif.). Subclones were made by exonuclease III digestion according to the company's protocol (Promega) and subsequently sequenced.

Fluorescence-based sequencing and analysis. Subclones were sequenced with the fluorescent dye-labeled universal primer 21M13 in a PCR-based sequence reaction by using Taq polymerase (Perkin-Elmer) and the reaction mixture supplied by Amersham according to the instructions of Applied Biosystems Incorporated (Foster City, Calif.). The sequences were analyzed on an automatic sequenator (model 370A; Applied Biosystems Incorporated). Sequences were analyzed with computer programs included in the program package PC/GENE (19a). The sequences were aligned with the CLUSTAL program by the method developed by Higgins and Sharp (17).

Nucleotide sequence accession numbers. The nucleotide sequence data will appear in the EMBL, GenBank, and DDBJ nucleotide sequence databases under accession no. AF117212 (porA upstream region of strain H355), AF117213 (porA upstream region of strain H44/76), AF117214 (porA upstream region of strain 860183), AF117215 (porA downstream region of strain H355), AF117216 (porA downstream region of strain 860183), AF117217 (porA downstream region of strain H44/76), AF117218 (porA locus after porA deletion of strain H355), AF117219 (porA locus after porA deletion of strain H44/76), and AF117220 (porA locus after porA deletion of strain 860183).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Identification of porA-negative meningococcal clinical isolates. During the routine characterization of clinical meningococcal isolates in the RLBM we identified a group C nonsubtypeable meningococcus (strain 860183) from a blood culture while the CSF isolate from the same patient appeared to be subtype P1.1. Both isolates had the same serogroup and type and were identical according to their outer membrane profile, except for the presence of PorA. Both isolates were subjected to colony immunoblotting with a PorA-specific antibody and colony hybridization with the porA-specific probe. All colonies from the CSF isolate appeared to be porA positive and PorA positive. In contrast 4% of colonies from the blood isolate were porA positive and PorA negative and 96% were porA negative and PorA negative. The occurrence of porA-negative and PorA-negative colonies indicated the deletion of porA.

Isolation of porA-negative variants in vitro. Meningococci reveal phenotypic changes, depending upon growth phase and growth rate (31, 32). The pellicle growth of meningococci has been shown to yield phenotypic variants (31). Two of nine isolates (H44/76 [B:15:P1.7,16] and H355 [B:15:P1.15]) tested yielded porA-negative variants after pellicle growth. The proportion of porA-negative variants on the culture plates of strain H44/76 was 3% after six cycles of culturing and reculturing of the pellicle. A similar proportion of porA-negative variants was obtained with strain H355 after 16 cycles of culturing and reculturing of the pellicle. PorA phase variants, which usually appear with a frequency of 10⁴ to 10³, were not observed in these experiments.

Characterization of porA-negative variants. Chromosomal DNA digested by EcoRI of the porA-negative variants and their porA-positive counterparts were assessed by Southern hybridization. The porA gene has an EcoRI restriction site dividing the gene roughly in half. The bands of both restriction fragments were absent in the porA deletion variants after hybridization with the porA probe, containing the complete porA gene, including its promoter (Fig. 1). This means that the deletion extends beyond the size of the probe (1.5 kb) (Fig. 2).

View larger version (37K): [in this window] [in a new window]   FIG. 2.   Southern hybridization analysis of EcoRI-digested chromosomal DNA of PorA-positive and PorA-negative variants of N. meningitidis H44/76, H355, and 860183. +, PorA-positive variants; , PorA-negative variants; M, molecular weight markers, in kilobase pairs.

Sequence upstream of porA. The sequencing of the porA upstream region of each of the three strains revealed a highly repetitive DNA sequence (Fig. 3). No significant open reading frames were found within the 2- to 2.5-kb region. The porA upstream sequence is preceded by the 3'-terminal part of the gene coding for the elongation factor Tu (EF-Tu) (Fig. 3A). The porA upstream sequence of strain 860183 contains a cluster of 22 palindromic sequences with RS3 (14) core sequences (ATTCCC-N₈-GGGAAT) in an inverted orientation (19, 37) and 10 RS3 core sequences (ATTCCC) in the direct orientation. The cluster is flanked by two neisserial repetitive sequences first described by Correia and colleagues (5) in the direct orientation. The pairs of inverted RS3 core sequences are actually inverted repeats of 8 bp; two nucleotides of the inner core sequence are also inverted. In addition, the nucleotides at the fourth position on either side of the inverted repeat are conserved and complementary (Fig. 4). The pairs of inverted RS3 core sequences were termed dRS3 by Morelli et al. (27) to distinguish them from the RS3 sequences originally described by Haas and Meyer (14). In this report we will refer to cases of single RS3 core sequences as cRS3 sequences. The cluster of dRS3 sequences can be subdivided into three repeats of 518 bp, containing six dRS3 sequences (dRS3 a to f) and three cRS3 sequences (cRS3 k to m) in the direct orientation (Fig. 3B). The three 518-bp repeats are followed by another, truncated repeat (Fig. 3A). The dRS3 a sequence in repeat A is part of a larger inverted imperfect repeat of 23 bp. Repeats B to D show minor sequence variations just upstream of dRS3 a, destroying the 23-bp inverted repeat. In addition, a direct repeat of 14 bp (TTTCCGATAAATTC) is found (Fig. 3B).

View larger version (29K): [in this window] [in a new window]   FIG. 3.   (A) Schematic representation of the structure of the chromosome upstream of porA of N. meningitidis 860183, H355, and H44/76. Open arrows indicate the 518-bp repeat. The letters in the open arrows refer to the different dRS3 repeats and cRS3 sequences. Hatched arrows indicate the Correia elements (5). (B) Sequences of the different repetitive elements in the porA upstream region. The different dRS3 and cRS3 sequences are indicated. Open arrows indicate inverted repeats in the 518-bp repeat. Black arrows indicate the 14-bp direct repeat.

View larger version (25K): [in this window] [in a new window]   FIG. 4.   Homology between the different dRS3 palindromic sequences. The conserved nucleotides are indicated in bold.

View larger version (23K): [in this window] [in a new window]   FIG. 5.   Putative recombination sites in the porA upstream region. The different dRS3 and cRS3 sequences are indicated. The capital letters in the designations refer to the different 518-bp repeats shown in Fig. 3A.

Sequence downstream of porA. The sequences of the region downstream of porA in all three strains and strain F207 (20) were essentially similar to each other. However, strain 860183 had only two DR2 repeats (20) downstream of IS1106 instead of four (Fig. 6). In H44/76 and H355 the sequences downstream of IS1106 actually also form a cluster comprising 10 dRS3 sequences, four cRS3 sequences (ATTCCC), and one other cRS3 sequence (GGGAAT) and are followed by Correia elements. The orientation of these Correia elements is opposite to that of the Correia elements found upstream of porA.

View larger version (13K): [in this window] [in a new window]   FIG. 6.   Schematic representation of the structure of the chromosome downstream of porA from N. meningitidis F207 (20), H44/76, H355, and 860183. Hatched arrows indicate the Correia elements (5). DR1 and DR2 were originally characterized by Knight et al. (20).

Homology of the porA upstream region with different loci in Neisseria. The Correia element is found in up to 150 to 200 copies throughout the Neisseria genomes, gonococci as well as meningococci (3). In the porA locus it is also found downstream of porA but in an opposite orientation (Fig. 6) (20).

Deletion of porA by recombination. To determine which of the three regions with homology upstream and downstream of porA is involved in the recombination event that leads to the deletion of porA, the porA locus in the porA-negative variants of the three strains was sequenced. The comparison of these sequences with the sequences of the regions upstream and downstream of porA indicates that in all three strains the deletion of porA occurred via a recombination between a homologous 116-bp sequence containing cRS3 l to dRS3 f (Fig. 7). It should be noted that the small deletion observed in the porA upstream region of strain H355, compared to that of strain 860183, was also observed in the sequence after the deletion of porA in strain H355, indicating that this deletion was not caused by PCR amplification.

View larger version (15K): [in this window] [in a new window]   FIG. 7.   Schematic representation of the porA locus after the deletion of porA in N. meningitidis H44/76, 860183, and H355. Hatched arrows indicate the Correia elements (5). The different dRS3 and cRS3 sequences are indicated. The capital letters refer to the different 518-bp repeats shown in Fig. 3A.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The data presented here show that a patient can have an infection with porA-positive and porA-negative variants of a meningococcal strain. In addition, porA-negative variants can be obtained in vitro. About 4 kb of the chromosome was deleted in the porA-negative variants of all three strains which were analyzed in this study. The deletion of porA occurred by the recombination between a 116-bp region of homology upstream and downstream of porA, containing two dRS3 sequences and one cRS3 sequence. The results of the pellicle growth experiments may give the impression that the deletion occurred with a rather high frequency. However, the pellicle growth experiments do not allow an accurate estimation of the frequency of the porA deletion event. Most likely, in these experiments the porA-negative variants are accumulated during repetitions of culturing and reculturing of the pellicle.

The RS3 core sequences in the porA upstream region form a cluster of as many as 22 dRS3 sequences (two RS3 core sequences in the inverted orientation) and 10 cRS3 sequences (single core sequences in the direct orientation). In addition the cluster is flanked by other neisserial repetitive sequences (5). Downstream of porA the DR2 repeats form a cluster of 10 dRS3 sequences and 5 cRS3 sequences, again flanked by a Correia element (opposite to the porA upstream Correia elements). Knight and colleagues (20) noticed that the palindromic RS3 core sequences (ATTCCC-N₈-GGGAAT) in the porA downstream sequence are similar in structure and distribution to the repetitive extragenic palindromes (REP) of 38 bp, initially identified in Salmonella typhimurium and Escherichia coli (9, 16, 23). After analysis of the DNA sequences flanking the opa genes in N. meningitidis, Morelli and coworkers also observed the parallelism between dRS3 sequences and REP sequences (27). In E. coli the REP sequences form clusters, which also contain other repeated elements. They were termed bacterial interspersed mosaic elements (12). The neisserial complex dRS3 clusters were termed neisserial interspersed mosaic elements (27). These large clusters of repetitive sequences may give rise to difficulties during bacterial genome sequencing projects. In these projects, genomic fragments are randomly cloned and sequenced. It may be that large repetitive sequences are missed, when the sequences of the different clones are connected during a computerized process.

The comparison of the porA upstream sequences of three strains indicated deletions, most likely due to a recombination between RS3 core sequences. The complete 518-bp repeat is only found upstream of porA. Partial homology is found with sequences flanking other genes in Neisseria strains. Regions showing homology with high identity (>90%) are always flanked by dRS3 sequences, again indicating that these are involved in recombination events. Analysis of the porA locus in porA-negative variants indicates that the deletion of porA occurred by a recombination between a region of 116 bp with homology upstream and downstream of porA, possibly between RS3 core sequences. The resemblance between dRS3 sequences and the E. coli REP sequence supports this idea. The REP sequence has also been implicated in chromosomal rearrangements (9, 39). The identification of this sequence at the junctions of tandem duplications supports this notion (38). The strains yielding porA-negative variants either in a patient or in vitro contained the IS1106 element distal to porA. However, there was variation in the number of DR2 repeats, indicating recombination events leading to the deletion or duplication of some of these repeats. Recombination events in the IS1106 region were also indicated by the results of Knight et al. (20). They found truncated forms of the IS1106 region, indicating a recombination between regions within the DR1 repeat. dRS3 sequences were suggested as sequences involved in these recombination events.

The 518-bp repeat contains six dRS3 sequences, and four pairs of these sequences are equally spaced by 76 bp (b-c, c-d, and e-f) to 80 bp (f-a). The two others are also equally spaced by 102 bp (a-b) to 109 bp (d-e). Downstream of porA the spacing between the dRS3 sequences is either 75 or 277 bp. This regular structure might indicate that dRS3 sequences are involved in organizing the DNA suprastructure, as proposed for REP sequences of S. typhimurium and E. coli, as well as in facilitating recombination. It has been shown that the REP sequence binds DNA gyrase (46) and DNA polymerase I (10). The HU protein stimulates the binding of gyrase to these REP sequences (47). These studies with E. coli have led to the proposal that REP sequences are involved in the folding of the bacterial nucleoid into independent supercoiled looped domains (11, 39).

PorA is the important component of group B meningococcal protein-based vaccines, since capsule polysaccharides of group B meningococci are poorly immunogenic. Antibodies against PorA are bactericidal and protective in a mouse model (35, 36). However, meningococci avoid the humoral host immune response by antigenic variation within PorA. Point mutations in the VR1 and VR2 regions of the protein (26, 40), and the replacement of epitopes by recombination and small deletions (2, 26) contributes to PorA antigenic variation. In addition, PorA expression is variable by means of the variable porA promoter (1, 42). The loss of PorA expression can also be due to the insertion of IS1301 (1, 29) or to frame shift mutation (1 and our unpublished data). The function of PorA is unknown. It has been reported that PorA-positive as well as PorA-negative meningococci can be cultured from the nasopharynx (6, 45). Our results show that a porA-negative variant can also be isolated from the blood of a patient with meningococcal disease. During a preliminary survey, two of a group of 57 nonsubtypeable meningococcal isolates appeared to be porA negative (unpublished data). The number of patients infected with porA-negative variants is likely to be underestimated. The isolates from patients infected with subtypeable meningococci were not investigated, but they could well contain porA-negative variants, since patients can be infected with both porA-positive and porA-negative variants of a meningococcus. These findings do not rule out the possibility that PorA has an essential function in the pathogenesis of meningococcal disease. In fact, the coexistence of both porA-positive and porA-negative variants within samples from one patient might indicate that the latter originates from the porA-positive variant during the infection. Our results show that this occurs by a recombination between homologous regions of 116 bp upstream and downstream of porA. The occurrence of porA-negative meningococci in patients together with the ability of PorA phase variation due to promoter variability might be indicative of the limited efficacy of PorA-based vaccines.