The Influence of IS1301 in the Capsule Biosynthesis Locus on Meningococcal Carriage and Disease

Previously we have shown that insertion of IS1301 in the sia/ctr intergenic region (IGR) of serogroup C Neisseria meningitidis (MenC) isolates from Spain confers increased resistance against complement-mediated killing. Here we investigate the significance of IS1301 in the same location in N. meningitidis isolates from the UK. PCR and sequencing was used to screen a collection of more than 1500 meningococcal carriage and disease isolates from the UK for the presence of IS1301 in the IGR. IS1301 was not identified in the IGR among vaccine failure strains but was frequently found in serogroup B isolates (MenB) from clonal complex 269 (cc269). Almost all IS1301 insertions in cc269 were associated with novel polymorphisms, and did not change capsule expression or resistance to human complement. After excluding sequence types (STs) distant from the central genotype within cc269, there was no significant difference for the presence of IS1301 in the IGR of carriage isolates compared to disease isolates. Isolates with insertion of IS1301 in the IGR are not responsible for MenC disease in UK vaccine failures. Novel polymorphisms associated with IS1301 in the IGR of UK MenB isolates do not lead to the resistance phenotype seen for IS1301 in the IGR of MenC isolates.


Introduction
The exclusively human pathogen Neisseria meningitidis causes serious diseases including septicaemia and bacterial meningitis. However, it is primarily a commensal of the human nasopharynx in around 10% of healthy individuals [1,2]. Rarely, the bacterium overcomes host defences, especially the complement system, and causes bloodstream infection, septic shock and/or meningitis. The meningococcus can express one of 13 capsules (classifying isolates into serogroups) that are necessary for avoidance of complementmediated killing [3]. Isolates are also categorised into sequence types (STs) based on sequences of seven housekeeping genes [4]. Most disease is caused by a limited number of genotypes, while other genotypes seldom invade the bloodstream [5]. Clonal complexes (cc) are groupings of STs with similar allelic profiles typically sharing at least 4-5 loci (http://pubmlst.org/neisseria).
Recently we described a polymorphism in three serogroup C ST-11 N. meningitidis isolates (MenC) from Spain that enhances resistance against complement-mediated killing, the principal mechanism of immunity against meningococcal disease [6]. In all isolates, IS1301 was found in the 134 bp intergenic region (IGR) between operons for capsule biosynthesis (sia) and export (ctr) ( Figure 1A), leading to up-regulation of capsule expression and providing a generic mechanism for increased resistance against bactericidal antibodies.
The MenC glycoconjugate vaccine has been very effective, leading to .90% reduction of MenC cases in countries where it has been introduced [7][8][9][10]. So far there has been no evidence of emergence of strains that are resistant to the immune responses elicited by the vaccine. However, there are a few reports of MenC disease developing in subjects who have completed a course of immunisation [11,12], and these could be caused by strains with increased resistance to complement-mediated killing due to insertion of IS1301 in the sia/ctr IGR. Furthermore, increased resistance to complement-mediated killing conferred by IS1301 could promote meningococcal virulence. Here we investigate whether isolates with IS1301 in the IGR are emerging among vaccine failure strains, and compare its presence in UK carriage and disease isolates.

IS1301 in the IGR Is Not Associated with MenC Conjugate
Vaccine Failure in the UK IS1301 in the sia/ctr IGR was originally described in MenC isolates in Spain; isolates had low titre (,8) in serum bactericidal assays (SBA) using sera from individuals immunised with the MenC conjugate vaccine [6]. As an SBA titre of ,8 is a marker of protective immunity [13,14], we examined MenC isolates from individuals from the UK with meningococcal disease who had completed a course of the MenC conjugate vaccine [11,15]. Most isolates were ST-11 (79%; 26 out of 33, with the others were ST-1095 [two isolates], ST-8, ST-2709, ST-467, ST-3463 and ST-6331), the most common cause of MenC disease at the time of the introduction of the vaccine [16]. IS1301 was not present in the IGR of any of the 33 MenC isolates, even though most isolates (32 out of 33) have IS1301 elsewhere in the genome (data not shown). We also examined MenC disease isolates in the UK before the introduction of the MenC conjugate vaccine (1998-99), and found IS1301 in the IGR in less than 1% of the isolates (1 out of 104 isolates, Table 1); IS1301 was inserted at the same location but in the opposite orientation (inverted) to the ST-11 isolates from Spain [6]. Therefore, there is currently no evidence from the UK that IS1301 in the IGR mediates a mechanism for escape from the MenC conjugate vaccine.

Presence of IS1301 in the IGR of UK Meningococcal Disease and Carriage Isolates
Enhanced resistance to human sera may promote the virulence of N. meningitidis during infection, and therefore insertion of IS1301 in the IGR might be more frequent among meningococcal disease compared to carriage isolates. We examined 741 carriage and 345 disease isolates of different serogroups collected between 1998 and 2000 in the South East of England (Table 1, [17,18]). In this collection, insertion of IS1301 was significantly more frequent in the IGR of disease (10%; 33 out of 345) compared to carriage isolates (1.5%; 11 out of 741, p,0.0001, Table 1). Approximately 50% of the carriage isolates were unencapsulated (non-groupable), and the presence/absence of IS1301 in the IGR should not affect the phenotype of these isolates. To exclude this as a potential confounding factor, we restricted our analysis to serogroupable isolates. There was still a significant difference for the presence of IS1301 in the IGR in disease (31 out of 336) compared to carriage isolates (11 out of 365, p = 0.0004) even after excluding unencapsulated isolates.
Insertion of IS1301 was most frequently found in MenB isolates belonging to clonal complex ST-269 (cc269) with 24 of 55 having IS1301 in the IGR (Table 2). Cc269 causes a significant proportion of meningococcal cases in UK [16,19], and in the collection, cc269 isolates were more commonly found among disease isolates (16%; 55 out of 345) compared to carriage isolates (3%; 22 out of 741, p,0.0001). To account for this potential bias, we analysed only cc269 isolates, and still found a significant difference between the presence of IS1301 in the IGR of disease isolates (44%; 24 out of 55) compared to carriage isolates (14%; 3 out of 22, p = 0.017).  [28] as well as the insertion of IS1301 in the IGR. The two most common polymorphic IGR with IS1301, IS-8 and iIS+69, are also illustrated. B: Isogenic strains with and without the polymorphisms associated with IS1301 were constructed by tagging the nmb76 gene downstream of the ctr operon with a kanamycin resistance (kan R ) cassette. C: The 8 bp deletion associated with IS1301 in the IGR; i) wild-type sequence, and ii) to iv) shows the 8 bp deletion with associated base pair changes (indicated in bold and lower case). * Refers to the most common type of IS-8, which was used for phenotypic analysis. The sequence denotes the base pairs from the insertion site of IS1301 up to the siaA translational start (underlined). doi:10.1371/journal.pone.0009413.g001

IGRs with IS1301 in MenB Isolates Are Associated with Polymorphisms
Strikingly, only one of 44 IGRs with IS1301 identified among isolates from South East England was identical to that in Spanish MenC ST-11 isolates (Table 3). Most commonly IS1301 was associated with an 8 bp deletion upstream of siaA (IS-8) or the IGR contained an inverted IS1301 with an internal 69 bp duplication (iIS+69); these two polymorphisms account for 66% (29 out of 44) of the changes seen. The 8 bp deletion was accompanied by three other sequence changes, one of which represented .80% of IS-8 (type (ii) in Figure 1C) and hence was chosen for further study. The 8 bp deletion is located 14 nucleotides upstream of siaA translational start, and has been reported not to influence siaA translation [20]. The 69 bp duplication within the inverted IS1301 duplicates bases 194 to 264 corresponding to nucleotides 580 to 649 of IS1301 in the original orientation [21].
Polymorphic sia/ctr IGRs with IS1301 Lead to Increased Transcription of the Genes for Capsule Transport but Not Capsule Biosynthesis Insertion of IS1301 in the sia/ctr IGR of MenC isolates leads to increased expression of the sia and ctr operons [6]. To investigate the effect of IS1301 with associated polymorphisms, we constructed isogenic strains by introducing the two common types of IGRs with IS1301, IS-8 and iIS+69, into a MenB cc269 strain. An antibiotic resistance marker was inserted into nmb76 in strains containing these polymorphisms ( Figure 1B), and genomic DNA was used to transform a MenB cc269 disease isolate (D157) without IGR changes; transformants were selected with (IS-8+ and iIS+69+) and without (IS-8-and iIS+69-) changes in the IGR. Compared with the effect of IS1301 in MenC [6], both IS-8+ and iIS+69+ led to a similar two-fold increase in the expression of ctrA as determined by qrtRT-PCR ( Figure 2A). However, no upregulation in siaA expression was detected in these strains. The expression of siaA and ctrA in strains without IS1301 in the IGR was similar to wild-type.
To verify the findings for siaA expression obtained by qrtRT-PCR, we constructed lacZ translational sia fusions in MenB with IGRs without IS1301, and with the original IS1301, IS-8, or iIS+69. The construct with IS1301 in the IGR showed a .30-fold increase in  reporter activity, whereas fusions with the IGRs found in MenB isolates showed similar levels of b-galactosidase activity as the IGR without changes ( Figure 2B), consistent with the qrtRT-PCR results.

Increased Transcription of Capsule Transport Genes Does Not Change the Amount of Capsule or Resistance against Complement-Mediated Killing
It is not known whether increased expression of only capsule transport genes is sufficient to increase capsule expression. FACS analysis of the amount of capsule expressed by the strains IS-8+, IS-8-, iIS+69+, iIS+69-, and D157 showed that strains with polymorphic IGRs with IS1301 do not express increased amounts of capsule compared to isogenic strains without the polymorphic IGRs ( Figure 3A). Although there was no detectable difference in the amount of capsule, the novel polymorphic IGRs with IS1301 could still influence resistance against complement-mediated killing. However, SBA titres were similar for strains with and without the polymorphic IGRs with IS1301 (50% killing of all strains was seen at a 1/512 dilution of human immune sera, not shown). This was confirmed by human serum assays, in which  there was no difference in the survival of isogenic strains with and without IS1301 with associated polymorphisms ( Figure 3B).

Presence of IS1301 in MenB cc269 Carriage and Disease Isolates
Phenotypic analyses suggest that MenB IGR polymorphisms might have no or little impact on meningococcal pathogenesis. This is in contrast to our observation that the polymorphic IGRs with IS1301 were more frequent in disease compared to carriage isolates in the survey from South East England. IS1301 was predominantly found among MenB cc269 isolates, of which only a limited number were examined (55 disease and 22 carriage isolates). We examined a larger number of MenB cc269 with 241 carriage isolates and 421 disease isolates from seven regions of the UK. Consistent with the South East England isolates, insertion of IS1301 in the IGR was significantly more frequent among disease isolates (35%; 148 out of 421) compared to carriage isolates (26%; 63 out of 241, p,0.05). All IS1301 insertions in cc269 were polymorphic and the majority (.80%; 178 out of 211) were IS-8 or iIS+69. Again when excluding non-groupable carriage isolates from the analysis, there was still a significant difference between the presence of IS1301 in the IGR of disease isolates (35%) compared to carriage isolates (26%; 49 out of 187, p,0.05).
Next we extended our analysis to examine potential variation within cc269. Some STs in the clonal complex are relatively distant from the ancestral ST-269 and share only four out of seven MLST loci with ST-269 ( Figure 4); most of these STs (e.g. 1161 and 1163) cluster together around the subgroup-founder ST-275. IS1301 in the IGR, or indeed anywhere in the genome, is rare among these STs, present in only one out of 217 isolates ( Figure 5). Isolates which share four out of seven loci with ST-269 are significantly more prevalent in carriage (48%, 91 of 187) compared with disease isolates (31%, 129 of 421). This is distinct from other isolates in cc269 (that share five or more loci with ST-269) of which approximately 50% have IS1301 in the IGR, with a similar prevalence among carriage and disease isolates ( Figure 5). Therefore we excluded from our analysis all STs sharing only four out of seven loci with ST-269. Strikingly, there was no longer a significant difference between the presence of IS1301 in the IGR of disease (51%; 148 out of 292) compared to carriage isolates (49%; 47 out of 96, p = 0.80).

Discussion
The presence of IS1301 in the sia/ctr IGR of MenC ST-11 enhances capsule expression and resistance against bactericidal antibodies elicited by the MenC conjugate vaccine [6]. Although the MenC vaccine has been successfully introduced across Europe (with .90% efficacy [8,10]), a small number of vaccinated individuals still develop MenC disease [11,12]. As isolates with IS1301 in the IGR are resistant to complement, it was possible . Genetic relatedness between STs within cc269. Neighbor-net constructed from the concatenation of the seven MLST housekeeping genes of N. meningitidis [4]. The major groupings (indicated with green, black, and yellow circles) represent the number of allelic differences (1, 2 and 3 respectively) from the core ST-269 allelic profile. The tree was made using the Split-Tree4 software [29]. doi:10.1371/journal.pone.0009413.g004 that they might lead to such vaccine failure. However, MenC isolates from people immunised with the MenC vaccine in the UK had no evidence of IS1301 in the IGR. It is possible that these isolates have other polymorphisms that contribute to their ability to be virulent in the vaccinees, or that the cases have arisen through waning host immunity rather than bacterial factors [11].
Our initial study of over 1,000 isolates from South East England showed a highly significant association of IS1301 in the IGR among disease versus carriage isolates. Closer examination of the dataset showed that IS1301 in the IGR was largely restricted to certain isolates, especially those in cc269. This is an important clonal complex, which accounts for a substantial proportion of UK disease cases [16,19,22]. Therefore, we examined the frequency of the IS1301 in the IGR in a larger collection of cc269 carriage (n = 241) and disease (n = 421) isolates across England and Wales, and still observed a significant association of IS1301 in the IGR in disease compared to carriage isolates.
Closer inspection of the IGRs with IS1301 in MenB revealed that they are associated with novel polymorphisms. The most frequent polymorphisms, IS-8 and iIS+69 (which account for .80% of changes), abrogate the up-regulation of capsule expression associated with insertion of IS1301 in MenC and do not confer increased resistance to complement-mediated killing. Nearly all IS1301 insertions in the original orientation [6] (99%) identified in this study were associated with an 8 bp deletion upstream of the siaA start. This indicates that this small deletion influences the functional consequences of IS1301 insertion, even though, on its own, it does not change transcriptional activity [20]. IS1301 is predicted to have two open reading frames (ORFs) [21]. It is not clear whether the 69 bp internal duplication (which is in the second ORF of IS1301) or the orientation of the insertion sequence is responsible for the lack of sia up-regulation in isolates containing the iIS+69 polymorphism.
The mechanisms by which IS1301 insertion leads to upregulation of sia and ctr expression are not understood. Indeed, surprisingly little is known about the genetic regulation of this key virulence determinant, or the transcription factors that govern expression. IS1301 could provide novel transcriptional start sites, prevent binding of transcription factors, or change their interaction with RNA polymerase. It is therefore interesting that both iIS+69 and IS-8 lead to upregulation of ctr expression as in the ST-11 isolates. This occurs regardless of the orientation of IS1301 or its precise sequence, suggesting that increased ctr transcription upon IS1301 insertion may result from changes in spacing due to the additional sequence in the promoter. However, sia expression was not elevated in either iIS+69 or in IS-8. The reasons for this are unknown and are under further investigation.
The association of polymorphisms with IS1301 in cc269 MenB suggests that increased expression of capsule might lead to a significant loss of fitness in these strains, in contrast to MenC ST-11 isolates. The reason for this apparent difference between the two groups of isolates is not obvious. However, it is interesting to note that isolates belonging to ST-11 are rarely found in carriage studies, and when they are present in the nasopharynx, are more likely to express a polysaccharide capsule than other carriage isolates [14].
Isolates from certain STs such as 275, 1161 and 1163 were remarkable as less than 2% have IS1301 in the IGR or anywhere in their genome. While these isolates have been assigned to cc269, they only share four out of seven MLST loci with ST-269 and appear to constitute a divergent cluster, centred around ST-275. Presence of IS1301 in the genome and sequence comparison of vaccine candidates among cc269 isolates supports that these isolates form or are part of a clonal complex that is distinct from cc269 [22].
Attempts are being made to identify virulence determinants by comparing the frequency of genetic traits between disease and carriage isolates [23,24]. In this study we were only alerted to potential confounding effects by the lack of increased complement resistance conferred by the common IGR changes found in MenB isolates. This demonstrates that even minor associated polymorphisms or sequence variation within genetic traits might have a major influence on their effect. Our findings also indicate that meaningful epidemiological comparisons of disease and carriage populations can only be obtained by matching for individual STs, and not at the level of clonal complex, especially when the traits being examined do not confer an observable phenotype. This is a major challenge, as it requires extensive collections that include sufficient numbers of relevant STs. Such large-scale prospective studies are needed to serve as a valuable resource.

Bacterial Strains and Growth Conditions
MenC conjugate vaccine failure strains were recovered from 33 patients in the UK who had completed a course of immunisation yet contracted meningococcal disease [11]. All disease isolates were obtained from sterile sites of patients from the UK with invasive meningococcal disease from 1998 to 2007. Carriage isolates were collected in November and December in 1999 and 2001 among 15-to 19-year old students in seven centres in the UK: Bangor, Cardiff, Glasgow, Nottingham, Plymouth, Stockport and Oxfordshire [18,25]. A MenB ST-269 disease isolate without changes in the IGR, D157 (Meningococcal Reference Unit number M01.241107), was used for construction of isogenic strains. Escherichia coli was used for construction of the nmb76 mutant and translational fusions.
N. meningitidis was grown on Brain Heart Infusion (BHI) medium with 5% Levanthal's supplement at 37uC in 5% CO2; kanamycin (kan) was used at 100 mg/ml and erythromycin at 2 mg/ml. E. coli was grown on Luria-Bertani (LB) broth with 50 mg/ml of kan or 200 mg/ml of erythromycin, as required.

Molecular Methods
The presence of IS1301 in the IGR and the genome was detected as described previously [6]. Primers used for PCR of the IGR were used for DNA sequencing. To make isogenic strains, isolates were tagged with an antibiotic marker as shown in Figure 1B. Nmb76 was inactivated by in vitro Tn5 transposition. Primers NG1494-for with a BamHI site (59-CTATGGGATCC-GATTTTGACTTATGCACACACCA-39), and NG1495-rev with an NheI site (59-GATCGCTAGCGGAAGCTATTT-TCTCTTTATTACCG-39) were used to amplify nmb76 from MC58. The PCR product was purified, digested with BamHI and NheI, and ligated into pTrcHis (Invitrogen) lacking a kan resistance cassette. The resulting vector was subjected to in vitro mutagenesis (EPICENTRE); PCR and DNA sequencing verified insertion of Tn5 at nucleotide 722 of nmb76. The nmb76 construct was transformed into D157 (generating D157nmb76), a strain (M99.240027) with an intact IS1301 in the IGR with an 8 bp deletion (IS-8), and a strain (M99.240335) with an inverted IS1301 with an internal 69 bp duplication (iIS+69, Figure 1AB). Genomic DNA from the tagged strains was used to transform D157 to kan resistance. Depending on the site of the cross-overs, transformants did (IS-8+ and iIS+69+) or did not (IS-8-and iIS+69-) contain IS1301 in the IGR.
For RNA isolation, bacteria were grown in liquid BHI to midlogarithmic phase and RNA was extracted from 2610 8 cells with the RNeasy kit (QIAGEN). Transcript levels of ctrA and siaA were compared to levels of gdh [6]. Results were analysed by comparative quantitation (Rotor-Gene Analysis Software, version 6.0, Corbett Research). RNA was isolated on two different occasions and transcript levels determined twice for each sample.
For b-galactosidase assays, 10 9 cells were inoculated into 10 ml of BHI and grown to mid-logarithmic phase. Cells were centrifuged and washed twice with phosphate buffered saline (PBS). b-galactosidase assays were performed as described previously [27]. Miller

FACS Analysis
Strains were grown overnight on BHI plates and 2610 9 cells were fixed in 3% paraformaldehyde for 1 hr, and washed three times with PBS. Bacteria (2610 8 cells) were incubated with 1/10 dilution of serogroup B anticapsular antibody (NIBSC, code 95/ 750) for 30 min at 37uC. Cells were washed twice in PBS/0.1% Tween 20 (PBS-T), resuspended in PBS containing a FITCconjugated donkey anti-mouse polyclonal antibody (1:200 dilution; Jackson Immuno-Research Laboratories), and incubated for 30 min at 4uC in the dark. Samples were washed with PBS-T and fluorescence was measured using an analyzer (FACSCalibur; Becton Dickson), recording at least 10 4 events. Results were calculated as mean fluorescence index (MFI) [6], which is the geometric mean multiplied by the percentage of positive cells.
Serum bactericidal assay (SBA) and human serum assay. Bacteria were grown overnight on solid medium. For SBAs, 10 4 cells were incubated for 1 hr at 37uC with serial dilutions of heatinactivated sera in the presence of 1/12 dilution of baby rabbit complement (Pelfreeze). Cells were plated on solid media and the SBA titres expressed as the reciprocal of the final dilution giving (50% killing compared to controls without complement or without sera. For serum assays, 10 5 cells were incubated in DMEM with serial dilutions of serum from healthy volunteers for 1 hr at 37uC. Controls were incubated with DMEM media only or with heatinactivated sera. Survival was calculated by counting the number of bacteria in the input and following incubation in human serum. Significant differences were examined with Student's t test.

Ethics Statement
Samples were obtained following informed written consent, and procedures were approved by the Riverside Research Ethics Committee (ref. No. 05/Q0401/126).