Received 2 August 1999/Returned for modification 12 October
1999/Accepted 22 December 1999
 |
INTRODUCTION |
Neisseria meningitidis
colonizes the nasopharynx, from which it can seed the bloodstream
before crossing the blood-brain barrier (BBB) to cause meningitis. In
contrast, Neisseria gonorrhoeae colonizes and invades the
epithelium of the genitourinary tract, where it can cause a localized
inflammation; bacteremia, though frequent, is asymptomatic and
dissemination is rare. Thus, both species are capable of crossing a
cellular barrier at their port-of-entry but they differ in their
abilities to subsequently disseminate in the blood. The ability to
induce intense and prolonged bacteremia is one of the prerequisites for
a bacterial pathogen to cross the BBB. In contrast, the details of
specific interactions with the cellular components of the BBB remain
unclear. Therefore, in order to understand the mechanisms that allow
N. meningitidis to cross the BBB, it will be necessary to
identify the genes that are involved in bloodstream dissemination
and/or specific interaction with the cellular components of the BBB.
Such genes might be present in both N. meningitidis and
N. gonorrhoeae but differ subtly in sequence or regulation,
or they might be present in only one of the two species.
Results from in vitro models have shown that most of the mechanisms
mediating cellular interactions are common to both N. meningitidis and N. gonorrhoeae. On the other hand,
several determinants have been identified that are specific to N. meningitidis: the polysaccharide capsule (8), the
enzyme rotamase (26), the RTX toxin-like Frp proteins
(29, 30), and a glutathione peroxidase (20). Of
these, the capsule locus is required for systemic dissemination and
bloodstream survival (34), whereas a role in virulence has not been demonstrated for the other genes.
We have recently created a bank of N. meningitidis-specific
sequences after subtractive hybridization between N. meningitidis and N. gonorrhoeae in order to identify
genes which are present only in N. meningitidis and might
therefore account for its differential pathogenesis (32).
Some of the clones mapped closely together, suggesting that they may
have been derived from larger regions of N. meningitidis-specific DNA. One region containing such clones (region 1) corresponds to the locus of capsule synthesis which had
previously been well characterized (8, 12, 13). However, the
significance of the other regions was unknown. We have now investigated
the other regions of N. meningitidis-specific DNA in order
to obtain details on the differences between N. meningitidis and N. gonorrhoeae and to possibly identify mechanisms
responsible for the specificity of N. meningitidis
pathogenesis. Our data identify eight novel DNA islands that are
specifically present in N. meningitidis and absent from
N. gonorrhoeae. Those islands that were conserved among a
representative set of meningococcal strains and/or showed homologies
with known virulence factors were deleted, and the resulting strains
were tested for phenotypes that are associated with crossing the BBB.
The results show that one of the eight islands is required for high
levels of bacteremia.
| |
MATERIALS AND METHODS |
Bacterial strains and growth conditions.
Strains of N. meningitidis were tested that represent the genetic diversity of
this species according to multilocus sequence typing (MLST)
(18). Their MLST assignments were: ST1 (subgroup I, strain
B40), ST2 (subgroup VI, Z6835), ST4 (subgroup IV-1, Z5463, Z2491
[27]), ST5 (subgroup III, Z3524), ST8 (A4 cluster, BZ
10), ST11 (ET-37 complex, serogroup C: FAM18; serogroup W135: ROU
[24]), ST25 (NG G40), ST30 (NG 4/88), ST32 (ET-5
complex, 44/76), ST41 (lineage 3, BZ 198); ST48 (BZ147), ST49 (297-0), ST60 (subgroup IX, 890592), and ST74 (ET-5 complex, MC58
[33]). Additional strains were N. meningitidis 8013, N. gonorrhoeae FA1090 and two
strains of Neisseria lactamica (Z6793 and Z6784). N. meningitidis strains were grown on GC agar (GCB; Difco), with the
addition of Kellogg's defined supplement plus ferric nitrate
(14) for 12 to 20 h at 37°C in a moist atmosphere
containing 5% CO2. Liquid media were GC-PO4
(1.5% Proteose Peptone number 3 [Difco], 0.5% NaCl, 30 mM potassium
phosphate; pH 7.5) and GC-HEPES (like GC-PO4 but the
potassium phosphate was replaced by 30 mM HEPES [pH 7.5]), both
supplemented as for the solid medium. Escherichia coli were grown on Luria-Bertani (LB) agar or in LB liquid medium. Antibiotics used were: ampicillin, 50 µg/ml; kanamycin, 50 µg/ml (N. meningitidis) or 100 µg/ml (E. coli); nalidixic acid,
20 µg/ml; and spectinomycin, 40 µg/ml.
Oligonucleotide primers and PCR conditions.
The sequences of
the primers used to amplify the individual regions and to construct
deletions are listed in Table 1; the other primer sequences are available on request. Template chromosomal DNA was isolated as described elsewhere (27).
The PCR conditions used depended on the length of the desired product.
For products up to 3 kb, the reaction mixture contained template DNA (1 µg ml
1); reaction buffer (10 mM Tris-Cl, pH 8.0; 50 mM
KCl; 1.5 mM MgCl2; 0.01% gelatin); dATP, dCTP, dGTP, and
dTTP (200 µM concentrations of each); dimethyl sulfoxide (5%);
forward and reverse primers (100 nM concentrations of each); and
Taq polymerase. The PCR reactions were incubated 1 min at
94°C, followed by 30 cycles of 1 min at 94°C, 1.5 min at 5°C
below the Tm of the oligonucleotide primers, and
2 min at 72°C, followed by incubation for 5 min at 72°C. For PCR
products between 3 and 8 kb, semi-long-range PCR was performed by using
the Expand Long Template PCR System (Boehringer Mannheim) under the
same conditions except that the mixture contained Buffer 1 and the
polymerase mix (0.75 µl) supplied with the kit. The thermocycling
conditions were 1 min at 94°C, 30 cycles of 45 s at 94°C, 1 min at 65°C, and 3 min at 68°C, and a final incubation for 5 min at
68°C. Template DNA longer than 8 kb was amplified by using the same
kit and conditions except that higher concentrations of dATP, dCTP,
dGTP, and dTTP (350 µM concentrations of each) and oligonucleotide
primers (300 nM) and 2 µl of the polymerase mix were used. Incubation
was for 1 min at 94°C, 30 cycles of 10 s at 94°C, 30 s at
65°C, and 20 min at 68°C, followed by 7 min at 68°C.
Sequencing of the eight regions.
Chromosomal DNA of N. meningitidis Z2491 was restricted by partial Sau3AI
digestion and fragments of 12- to 23-kb size fractionated by gel
electrophoresis were cloned into the BamHI site of the Lambda DASH II (Stratagene) phage vector by using E. coli
XL1-Blue MRA (Stratagene). Details of the following steps were
according to the DIG System Users Guide (Boehringer Mannheim). Plaques
were transferred to nylon membranes (Hybond N; Amersham). N. meningitidis-specific clones (32) were digoxigenin
labeled during PCR amplification and used as probes for plaque
hybridization under stringent conditions. Phages containing hybridizing
sequences were detected colorimetrically, and single plaques were
purified before lysates were prepared. Two microliters of each lysate
was used as a template for long-range PCR with primers in the phage
vector immediately flanking the inserts. Then, 15 µg of PCR product
was randomly sheared by nebulization (no. 4100 Nebuliser; Inhalation
Plastics) for 20 min at 0.7 atm as described elsewhere
(http://bric.postech.ac.kr/resources/rprotocol/rprotocol_partii.html). The sheared fragments were precipitated, end repaired with T4 DNA
polymerase and Klenow DNA polymerase (New England Biolabs), and size
fractionated on a 0.8% agarose gel. Fragments of between 0.4 and
0.6 kb and between 0.8 and 1 kb were separately eluted from the gel by
using the Qiaquick Gel Extraction Kit (Qiagen). dATP overhangs were
added to the fragments and ligated with the TA cloning vector pCR2.1
(Invitrogen). These preparations were transformed into E. coli XL1-Blue by electroporation. A total of 96 recombinant
colonies were picked per transformation and grown in LB medium with
ampicillin, and their inserts were amplified by PCR by using primers
complementary to the flanking vector sequences. The PCR products were
purified and sequenced by using the M13 reverse primer, a dRhodamine
terminator cycle sequencing kit, and ABI Prism 377 DNA sequencers
(Perkin-Elmer Applied Biosystems). Raw data from the ABI sequencer were
prepared for assembly by using the ASP program
(http://www.sanger.ac.uk/Software/Sequencing/ASD/asp/MODULES.shtml), and sequences were assembled with GAP4 from the Staden sequence analysis package (28).
Sequences that were 100% identical to those available in the public
domain (Sanger Center;
http://www.sanger.ac.uk/Projects/N_meningitidis/) at that time were
accepted as correct, whereas all discrepancies were resequenced as
follows using PCR products from the chromosomal DNA of strain Z2491.
Fragments of approximately 5 kb were amplified by semi-long-range PCR
by using primers designed from the sequences of the phage inserts. The
PCR products were purified using the Qiaquick PCR Purification Kit
(Qiagen) and sequenced from both strands with appropriate primers as
described above. Additional smaller PCR products from bacterial
chromosomes were sequenced from other strains of N. meningitidis using the same strategy.
Region 2 was sequenced from both strands by using a different strategy
than for the other regions. Primers were designed according to the
sequences of the clones isolated by Tinsley and Nassif (32),
and products were obtained from chromosomal DNA by semi-long-range PCR.
Sequence walking was used to complete the sequences of each of these products.
Analysis of nucleotide sequences.
Open reading frames (ORFs)
were recognized by using the Codon Use program written by Conrad
Halling, which supplies a graphical output for a sliding window of the
codon adaptation index in all six frames. The permitted start codons
were ATG and GTG, and the permitted stop codons were TAA, TAG, and TGA.
Homology searches at the nucleotide level were performed by using
BLASTN (http://www.ncbi.nlm.nih.gov/cgi-bin/BLAST/nph-blast?) and
at the amino acid level by using PSI-BLAST
(http://www.ncbi.nlm.nih.gov/cgi-bin/BLAST/nph-psi_blast). Repetitive
nucleotide sequences were detected by using Miropeats (22).
DNA dot blot and Southern hybridization.
DNA dot blot
hybridization was performed according to the DIG System Users Guide
(Boehringer) by spotting 1 µl containing 100 ng of denatured
chromosomal DNA from each strain onto nylon membranes (Hybond N;
Amersham). For Southern hybridization analysis, chromosomal DNA was
digested with restriction endonucleases and separated by conventional
electrophoresis or by pulsed-field gel electrophoresis (PFGE) and then
transferred to nylon membranes. The DNA dot blots were hybridized with
digoxigenin-labeled probes obtained by PCR amplification of each ORF,
and Southern blots were hybridized with labeled probes corresponding to
ORFs or entire regions. For the analysis of the distribution of the
regions among meningococcal strains, hybridizations were performed at
37°C in 5× SSC (1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate)
containing 50 mM sodium phosphate, 7% sodium dodecyl sulfate
(SDS), 2% Blocking Reagent (Boehringer-Mannheim), 0.1%
N-lauroylsarcosine, and 50% formamide, and the last washing
step was with 0.5× SSC-0.1% SDS at 50°C in order to allow
approximately 30% mismatch. Positive hybridization signals were
detected by chemiluminescence. For verification of the mutants by
Southern blotting, probes were labeled with
[
-32P]dCTP. Hybridization in 500 mM sodium phosphate
(pH 7.2) containing 7% SDS and 1 mM EDTA and washing in 40 mM sodium
phosphate (pH 7.2), containing 1% SDS and 1 mM EDTA were performed at
65°C.
PCR analysis of DNA islands in different N. meningitidis strains.
The sizes of the eight islands were
determined by PCR amplification with primers Reg2-for to Reg9-rev
(Table 1) that are complementary to the 5' and 3' flanking sequences,
respectively, in both N. meningitidis and N. gonorrhoeae. Semi-long-range PCR was used for all regions. Region
3 containing Pnm1 was amplified by eight sets of long-range PCRs. The
locations of each island were confirmed by additional PCR reactions by
using each forward primer (Reg2-for, Reg3-for, etc.) with a reverse
primer in the leftmost ORF and each reverse primer (Reg2-rev, Reg3-rev,
etc.) with a forward primer in the rightmost ORF of the corresponding island. Additionally, each ORF was separately amplified from strains that reacted in dot blot hybridization by primers specific to its 5'
and 3' ends, and the sizes were confirmed by gel electrophoresis.
Production of deletion mutants in the N. meningitidis-specific islands.
Briefly, deletions were
produced by transforming N. meningitidis ROU, MC58, and
Z5463 with plasmid DNA into which sequences flanking an island had been
cloned, such that they were separated by an antibiotic resistance
cassette in place of the N. meningitidis-specific island.
After transformation of N. meningitidis, DNA from a single colony expressing the selected antibiotic resistance was used to
retransform the same strain, and a mixture of several hundred transformants was pooled and used for biological assays.
Figure 1B outlines the procedure for the
production of the plasmids used for the deletions, taking region 8 as
its example. The PCR products of the flanking regions (approximately 1 kb) were amplified using the primers shown in Table 1. The primers were
designed so that both PCR products contained either a BamHI or a EcoRI restriction site at the internal end and one of a
variety of different restriction sites at the external end. The PCR
products were digested with BamHI or EcoRI,
purified, and ligated. The ligation products were reamplified by using
the "external" primers, cleaved at the ends with the appropriate
enzyme(s) and cloned into pBluescript KS(+) (Stratagene). A resistance
cassette (see below) flanked by two neisserial uptake sequences was
then cloned into the BamHI or the EcoRI site.
View larger version (25K):
[in this window]
[in a new window]
|
FIG. 1.
Production of a cassette, flanked by neisserial uptake
sequences, and deletion of the N. meningitidis-specific
islands. (A) For production of the cassette, an inverted repeat
containing two neisserial uptake sequences (arrows) and an internal
BglII site (shaded) was constructed by ligating synthetic
oligonucleotides T1 to T4 (Table 1). This molecule was cloned into
pBluescript, and an  spectinomycin resistance cassette was cloned
into the BglII site to yield a cassette that may be excised
with BamHI. (B) Deletion of region 8. For replacement of the
N. meningitidis-specific islands with the cassette, the
flanking regions of the islands were PCR amplified by using the
oligonucleotides in Table 1. These were ligated together at an internal
restriction enzyme site (EcoRI), reamplified, and cloned
into pBluescript. The construct was reopened at the EcoRI
site, and the cassette was inserted between the flanking sequences.
This plasmid was then used to transform N. meningitidis and
replace the chromosomal island with the resistance cassette by
homologous recombination.
|
|
Construction of the resistance cassette.
The resistance
cassette contains the omega fragment (23), which encodes
resistance to spectinomycin due to aadA (aminoglycoside adenyltransferase) and interrupts both translation and transcription, flanked on each side in inverted orientation by the neisserial uptake
sequence GCCGTCTGAA. Due to the degree of secondary
structure in the omega fragment, the construct was made without using
PCR technology, as outlined in Fig. 1A. An artificial inverted repeat of uptake sequences was produced as follows. The oligonucleotides T1
and T2 (Table 1) were mixed (25 µM, 30 µl in T4 DNA ligase buffer),
heated at 75°C for 10 min, and allowed to cool to room temperature
for 1 h. This results in a double-stranded DNA containing one copy
of the neisserial uptake sequence, with a 5' overhang of GATC
(compatible with BamHI) and a 3' overhang of TGCA. A similar hybridization of oligonucleotides T3 and T4 also produced a DNA with
the same overhangs. After mixing and ligation, only the desired internal join (TGCA) can be ligated due to the 5' phosphate groups. The
resulting mixture was cloned into the BamHI site in
pBluescript KS(+), and the correct construct (plasmid pT1) was detected
by its possession of a BglII site (present in
oligonucleotide T1).
The omega fragment was excised from plasmid pHP45 by using
BamHI and inserted into the unique BglII site of
pT1, resulting in plasmid pT11, where the omega cassette is flanked
by neisserial uptake sequences and can be excised by digestion with
BamHI. We also constructed plasmid pT11E, where the
fragment can be excised by digestion with EcoRI. To this
end, the BamHI fragment of pT11 was ligated to the
adaptors BE1 and BE2 and cloned into the EcoRI site of
pBluescript KS(+).
Transformation of N. meningitidis.
After overnight
growth on GC agar plates, bacteria were resuspended in GC-HEPES
containing 1 mM K2HPO4 and Kellogg supplement 1 (14) to an optical density at 600 nm (OD600) of
0.1. MgCl2 was added to a final concentration of 10 mM, and
transforming DNA (0.5 to 5 µg of linearized plasmid or 0.5 µg of
chromosome) was added to 500 µl of the suspension. After incubation
without shaking at 37°C for 30 min, 4.5 ml of GC-PO4 was
added, and the mixture was incubated with shaking for a further 2 h. The bacteria were plated onto GC agar containing 40 µg of
spectinomycin per ml for selection of transformants.
Complement-dependent serum bactericidal assay.
Antiserum
against strains ROU and MC58 was obtained by immunizing rabbits with a
mixture of paraformaldehyde-treated and sonicated bacteria with
Freund's adjuvant and boosting three times using Freund's incomplete
adjuvant. After overnight growth on GC agar plates, bacteria were
resuspended to a final concentration of 106 CFU/ml in
phosphate-buffered saline (PBS) containing 5 mM MgCl2 and
0.25 mM CaCl2 (PBSB). Then, 10 µl of bacterial suspension was mixed with 360 µl of antiserum (decomplemented by heating at
56°C for 30 min) diluted in PBSB and 40 µl of freshly thawed guinea
pig serum (Gibco-BRL) as a complement source. Killing was measured by
determining colony counts after 45 min incubation at 37°C, and the
survival was compared to bacteria incubated in the absence of
antiserum. In each experiment, a positive killing control used a
serum-sensitive polyphosphate kinase mutant (ppk) (31), while as a negative control the ppk
bacteria were incubated with decomplemented guinea pig serum in
the presence of opsonizing antiserum.
Adhesion and invasion assays.
Bacteria from GC agar plates
were grown in RPMI (Gibco-BRL) containing 10% heat-inactivated fetal
calf serum (FCS; Gibco-BRL) with gentle shaking for 2 h to an
OD600 of 0.1. Then, 1 ml of a 100-fold dilution
was added to a confluent monolayer of human umbilical vein
endothelial cells (HUVEC) in 2-cm2 tissue culture wells
(Costar). One well was used per N. meningitidis strain to
test adherence, and two wells were used to assay invasion.
Adhesion.
After incubation for 1 h at 37°C in 5%
CO2, the supernatant was removed (nonadherent bacteria) and
the monolayer was washed three times with RPMI. The adherent bacteria
were released by adding 1 ml of PBS-1% saponin and scraping the
bottom of the wells with a micropipette tip. The numbers of adherent
and nonadherent bacteria were determined after plating them on
supplemented GC agar. Adherence was calculated as the number of
adherent bacteria divided by the total number of adherent plus
nonadherent bacteria.
Invasion.
The wells were washed every hour as described
above for 6 h and then filled with RPMI-FCS containing 150 µg of
gentamicin per ml. After 1 h of incubation to kill external
bacteria, internalized bacteria were harvested and enumerated as
described above. Invasion was calculated as the number of internalized
bacteria divided by the total bacteria at 1 h after infection.
Infant rat model of meningococcal infection.
Bacteria grown
on GC agar plates for 14 h were resuspended in pyrogen-free 0.9%
NaCl to an OD600 of 0.06. Four- to five-day-old Lewis rats
(IFFA Credo, L'Arbresle, France) anesthetised with diethyl ether were
injected intraperitoneally with 100 µl of bacterial suspension. Half
of each litter (usually 10 animals) was injected with the parental
strain and half was injected with the mutant strain. Samples of blood
(5 µl) were taken from an incision in the tail after 1, 3, 6, 9, and
24 h. The blood samples were diluted in GC-PO4 and
plated on GC agar for enumeration.
Nucleotide sequence accession numbers. The DNA sequences
described in this work have been deposited in the EMBL database under the following accession numbers: for region 2 of N. meningitidis Z2491 (fhaB and fhaC homologues
and genes of unknown function) and flanking genes, AJ391255; for region
3 (prophage Pnm1 and gpx/A) and flanking genes,
AJ391256; for region 4 (genes of unknown function) and flanking genes,
AJ391257; for region 5 (restriction/modification system) and
flanking gene, AJ391258; for region 6 (pseudogene with homology
to siderophore receptor genes) and flanking genes, AJ391259; for
region 7 (homology to type I secretion system) and flanking genes,
AJ391260; for region 8 (fhuA and dsbA homologues)
and flanking genes, AJ391261; for region 9 (cluster of putative ORFs
and insertion element IS4351N2) and flanking genes, AJ391262; for
strain FAM18 hlyD gene (putative component of type I
secretion system), AJ391263; for strain FAM18 tolC gene
(putative component of type I secretion system), AJ391264; for strain
NG 4/88 tolC gene (putative component of type I secretion
system), AJ391265; for strain 297-0 tolC pseudogene
(putative component of type I secretion system), AJ391266; for strain
FAM18 fhuA gene for putative siderophore receptor, AJ391267;
for strain NG 4/88 fhuA gene for putative siderophore receptor, AJ391268; for strain MC58 fhuA gene for putative siderophore receptor, AJ391269; for strain ROU fhuA gene for putative siderophore receptor, AJ391270; for strain BZ 10 fhuA pseudogene for putative siderophore receptor,
AJ391271; for strain BZ 147 fhuA pseudogene for putative
siderophore receptor, AJ391272; for strain BZ 198 fhuA
pseudogene for putative siderophore receptor, AJ391273; for
strain B40 fhuA pseudogene for putative siderophore
receptor, AJ391274; for strain Z3524 fhuA pseudogene for
putative siderophore receptor, AJ391275; for strain 297-0 fhuA pseudogene for putative siderophore
receptor, AJ391276; for strain 44/76 fhuA pseudogene,
AJ391277; for strain FAM18 dsbA gene for putative disulfide
oxidoreductase, AJ391278; for strain BZ10 dsbA
gene for putative disulfide oxidoreductase, AJ391279; for
strain NG 4/88 dsbA gene for putative disulfide oxidoreductase, AJ391280; for strain 44/76 DNA for region 6 (rsi1 pseudogene) and flanking fnr and
dinP genes (partial), AJ391281; for strain BZ 198 DNA for
region 6, insertion sequence, partial rsi1 pseudogene and flanking fnr and
dinP genes (partial), AJ391282; for strain 297-0 partial fnr gene for putative ferredoxin-NADP+
reductase and partial dinP gene for putative DNA-damage
inducible protein P, AJ391283; and for strain FAM18 DNA for
region 2 (fhaB and fhaC homologues and genes of
unknown function) and flanking genes, AJ391284.
| |
RESULTS |
Identification of N. meningitidis-specific regions in
strain Z2491.
Twenty-eight subtractive clones from N. meningitidis serogroup A, subgroup IV-1 strain Z2491 have been
described that have no homologies in N. gonorrhoeae
(32). DNA probes corresponding to these clones were used to
screen a Lambda Dash II library containing 12- to 23-kb DNA fragments
of strain Z2491. Other clones comprising "region 1" were not
investigated further because they correspond to the well-characterized
locus of capsule production.
A minimal set of Lambda Dash II recombinant phages were
identified whose inserts hybridized with 25 of the probes
(all except B305, B333, and E103), and these inserts were
sequenced. Comparison of these N. meningitidis
sequences with that of N. gonorrhoeae strain FA1090
(University of Oklahoma; http://dna1.chem.ou.edu/gono.html) identified
the flanking ends that were homologous in both species and, hence, the
extent of eight regions of N. meningitidis-specific sequences. The eight regions were designated regions 2 to 9 and are
shown in Fig. 2,
3, and 4.
They contain 85 ORFs, most with codon usage typical of N. meningitidis, of which 43 are homologous to previously described
ORFs in other species (Table 2). ORFs with no significant homology to genes of known function
were named according to the region (rtw
[region two], rth [region three], rfo
[region four], rfi [region five], rsi
[region six], rse [region seven], rei
[region eight], and rni [region nine]) plus a sequential number corresponding to the position of the ORF (rtw1,
rth52, etc.).
View larger version (29K):
[in this window]
[in a new window]
|
FIG. 2.
Relative genetic locations of N. meningitidis-specific regions 1 to 9 in the serogroup A subgroup
IV-1 strain Z2491. For the six smaller regions, insets indicate the
size of ORFs (open rectangles) and the orientation of transcription
(internal arrows). Flanking sequences that are homologous in N. meningitidis and N. gonorrhoeae are indicated by black
bars, and the numbers show the percent homology between N. meningitidis and N. gonorrhoeae. c106 and c154 in
regions 6, 7, and 8 refer to 106- and 154-bp Correia elements,
respectively, whereas IS4351N2 in region 9 is an IS element
with flanking inverted repeats (open arrowheads). The PCR probes used
for hybridization are indicated by gray bars above and below the map.
The results from dot blot hybridization and PCR analyses with diverse
strains of N. meningitidis and N. lactamica are
summarized at the top and bottom of the figure by "+" and " ".
*, positive hybridization due to other copies of IS4351N2
present at other locations in the chromosome. The numbers indicate the
sizes (in kilobases) of the PCR products obtained after amplification
of the regions with primers complementary to the 5' and 3' flanking
sequences. An explanation of the genetic designations is in Table 2.
 , duplication in hlyD of region 7;  , frameshift in
 rsi1 of region 6.
|
|
View larger version (13K):
[in this window]
[in a new window]
|
FIG. 3.
Contents of region 2. Region 2 is located
within a homologue of the E. coli cvaB gene that is a
pseudogene in N. gonorrhoeae inactivated by the insertion of
a Correia element (4). In N. meningitidis, region
2 has replaced the Correia element plus 693 bp of the cvaB
gene. Symbols are as described for Fig. 2. Repetitive stretches within
the region from fhaB to rtw4 are indicated by R1
to R4. A stop codon in the cvaA gene is indicated by an
asterisk. c106, a 106-bp Correia element.
|
|
View larger version (10K):
[in this window]
[in a new window]
|
FIG. 4.
Region 3 and the Pnm1 prophage. Details are as
given in Fig. 2. The numbers 1 to 52 above the ORFs indicate genes
rth1 to rth52. The transcription orientation is
from left to right for rth2 to rth52. Dark gray
ORFs are homologous to both phage genes and to clustered genes in
H. influenzae. ORFs shaded in light gray are only homologous
to clustered genes in H. influenzae. ORFs in hatched boxes
are only homologous to phage proteins (rth17,
rth40, and rth43). Blocks of ORFs that yielded
identical results in dot blot analysis are delineated by paired
vertical lines above the map. +/, positive result with one of two
strains of N. lactamica.
|
|
The N. meningitidis-specific regions are imported DNA
islands in N. gonorrhoeae and N. meningitidis.
The GC contents of most of the ORFs in regions 2 to 3 and 6 to 9 were
close to the 51% value typical of N. meningitidis. However, rei1 of region 8 has a GC content of 40%, and regions 4 and
5 have average GC contents of only 33%. The ends of the eight regions are defined by longer stretches of flanking DNA that are 80 to 96%
homologous between the two species. Except for region 2 (Fig. 3), the
regions begin and end in intergenic stretches. No repeat structures
similar to those flanking pathogenicity islands in Enterobacteriaceae were found, except possibly for region 9. In region 9, 39 bp at the 5' end are repeated 180 bp to the left of the
3' end of the N. meningitidis-specific sequence.
Furthermore, the right end of region 9 includes a putative IS element,
IS4351N2, containing a transposase that might have been
involved in the insertion of this region.
The comparison between N. meningitidis and N. gonorrhoeae shows that the regions do not correspond to simple
insertions but rather replace alternative sequences, 44 bp to over 12 kb in size, in N. gonorrhoeae FA1090 (Table
3). Thus, the N. meningitidis specific regions correspond to DNA islands that are species specific. The following data suggest that some or all of these islands may have
arisen by import from other species via recombination in the homologous
flanking DNA, similar to the mechanism deduced for three small islands
described elsewhere (39). First, DNA uptake sequences (DUS)
(GCCGTCTGAA) (9, 10) were found in all the
N. meningitidis islands except regions 4 and 5, where they
were found in their immediate borders. The presence of a DUS has been
associated with import of DNA into N. meningitidis from
Haemophilus influenzae (15) and would facilitate
the import of such islands from unrelated bacteria. Second, regions of
70 to 88% homology over 75 to 1,100 bp were found at one of the two borders in regions 4 and 6 to 8, whereas the homology between N. meningitidis and N. gonorrhoeae is normally at least
90%. These borders of low homology might represent the remnants of the
recombination with foreign DNA (39). Taken together, all of
these results are compatible with the import of these specific
sequences from different foreign species into N. meningitidis and/or N. gonorrhoeae and justify using
the term DNA island for the eight regions.
Conservation of the eight islands among diverse N. meningitidis.
Twelve variant capsular polysaccharides are
expressed by different N. meningitidis, suggesting the
existence of at least as many variants of region 1. The data presented
above for regions 2 to 9 were based on the sequence analysis of one
serogroup A strain of N. meningitidis and do not indicate
whether the eight islands are generally present at the same location in
diverse strains of N. meningitidis or whether their genetic
content is constant. To address this issue, three kinds of experiments
were performed with 13 representative N. meningitidis
strains, N. gonorrhoeae FA1090, and two commensal N. lactamica strains. The 13 N. meningitidis strains were
chosen to represent the genetic diversity of this species according to
MLST (18) and include members of the eight hypervirulent
clonal groupings called subgroups I, III, IV-1, and VI, ET-5 and ET-37
complex, the A4 cluster, and lineage 3, as well as unrelated endemic
strains (serogroup A subgroup IX, STs 25, 30, 48, and 49). (i) To
confirm the localization of each island, PCR amplifications were
performed by using one primer in each flanking sequence plus a matching
reverse primer in the neighboring gene within the island. (ii)
The length of each region was determined after PCR amplification by
using primers located in the flanking sequences. (iii) Dot blot and
Southern hybridizations against chromosomal DNA were performed using
PCR products amplified from each of the 85 ORFs as probes.
Representative examples of the Southern and dot blot analyses are
shown in Fig. 5. Finally, the regions
were sequenced at least in part from some of the 13 N. meningitidis strains.
View larger version (55K):
[in this window]
[in a new window]
|
FIG. 5.
Southern analysis of pulsed-field gels of
SpeI-digested chromosomal DNA (A to E) and DNA dot blot
analysis (F to H) of regions 2 and 3. The numeration of bacterial
strains is the same in all panels: 1, subgroup IV-1; 2, subgroup I; 3, subgroup III; 4, subgroup IX; 5, subgroup VI; 6, A4 cluster; 7, lineage
3; 8, ST30; 9, ST49; 10, ST48; 11, ST25; 12, ET-5 complex (44/76); 13, ET-5 complex (MC58); 14, ET-37 complex (FAM18); 15, ET-37 complex
(ROU); 16, N. gonorrhoeae FA1090; 17, N. lactamica Z6793; 18, N. lactamica Z6784. (A to C)
Southern analysis of region 3 with probes corresponding to
rth18 (A), rth33 (B), and gpxA (C).
Results identical to those shown in panel A were obtained with probes
for rth17 and rth19. Hybridization with probes
for rth20 and rth21 resulted in a similar pattern
except for lanes 12 and 13, where the upper signal (marked by an
asterisk) disappeared. The same results as in panel B were obtained
with probes for rth1, rth4, rth28, and
rth30. The other ORFs of Pnm1 were not tested by Southern
analysis, but similar results can be expected according to dot blot
analysis except for rth26, which is also present in strains
of the ET-5 complex, ST25, and ST48. (D and E) Southern analysis of
region 2 with probes for fhaC (D) and rtw7 (E).
The same results as in panel D were obtained with probes for
fhaB (5'-part), rtw2, and rtw4. With a
probe for rtw5, the lower signals in lanes 12 and 13 (marked
by an asterisk) disappeared. The result of hybridization with a probe
for rtw8 was the same as with rtw7 (E). Positions
of the molecular size markers are shown in kilobases. (F to H) DNA dot
blot analysis with probes for the 5' end (F), the central part (G), and
the 3' end (H) of fhaB.
|
|
As expected, none of the probes hybridized with the N. gonorrhoeae strain in dot blots. The other results are summarized
in Fig. 2, 3, and 4. They show that, when present, the islands are flanked by the same sequences in all N. meningitidis strains
tested and are therefore located at the same chromosomal locations.
Furthermore, regions 2, 5, 7, 8, and 9 were present in all
representative N. meningitidis strains (Fig. 2 and 3), with
little apparent variation in content, except that the central and 3'
end of fhaB in region 2 and rei1 in region 8 did
not hybridize with some strains. On the other hand, region 4 was absent
in about half of the strains tested and region 6 showed considerable
interstrain variation (Fig. 2) due to replacement by an IS element
and/or deletions (data not shown). Probes from regions 5 and 6 and two
probes from region 3 hybridized with the N. lactamica
strains, showing that these genes are not totally specific to N. meningitidis.
Data presented below show that region 3 consists of the gpxA
gene encoding glutathione peroxidase (19, 20) plus an
integrated 39.3-kb prophage that we have called Pnm1 (prophage in
N. meningitidis 1). Pnm1 was present only in serogroup A
strains of the epidemic subgroups I, III, IV-1, and VI (Fig. 4) and was
lacking in serogroup A subgroup IX, which is unrelated to the epidemic
subgroups of serogroup A (1). Furthermore, Pnm1 was also
lacking in all other N. meningitidis strains tested. From
N. meningitidis strains outside the epidemic serogroup A
subgroups, primers specific for the flanking norZ and
mccF genes amplified a 1.4-kb product whose sequence
contained only gpxA plus small, flanking intergenic regions. Within Pnm1, only probes from rth17 to rth21
hybridized with N. meningitidis strains outside the epidemic
serogroup A subgroups. Homology searches in the Sanger Centre genome
project for serogroup A strain Z2491
(http://www.sanger.ac.uk/Projects/N_meningitidis/) revealed that
these five genes had homologues within a second prophage containing at
least 22 ORFs. Furthermore, Southern blots revealed that
rth17 to rth21 hybridized with two pulsed-field DNA fragments in strains possessing Pnm1 (Fig. 5). Thus, the
hybridization of rth17 to rth21 with all strains
of N. meningitidis probably reflects the universal presence
of a second prophage.
As with rth17 to rth21, some of the region 2 genes (fhaC, fhaB, rtw2, and
rtw4 to rtw8) also hybridized with two
pulsed-field DNA fragments in Southern blots from strains of the ET-5
complex, lineage 3, ST25, and ST48 and are therefore present twice in
these bacteria (Fig. 5).
Together, these data show that the N. meningitidis-specific
islands 2, 5, 7, 8, and 9 are fairly well conserved among
representative strains of N. meningitidis and that Pnm1 is
conserved among epidemic serogroup A strains.
Sequence analysis of the N. meningitidis-specific
islands.
The possible functions of the ORFs in the eight islands
were investigated by protein homology searches and computer analysis of
the sequences (Table 2). The following description concentrates on
those homologies that might be relevant to the virulence potential of
N. meningitidis.
The two ORFs at the left of region 2 are homologous to and
approximately the same size as genes in Bordetella pertussis
encoding the filamentous hemagglutinin (FHA) precursor, FhaB, and the
accessory protein, FhaC, involved in the secretion of FHA
(17). In B. pertussis, FhaB (367 kDa) is
processed during secretion, yielding the 220 kDa FHA protein (5,
6). FHA is thought to be one of the major adhesins of B. pertussis and is also a component of several vaccines against
whooping cough. The homology between FhaB in N. meningitidis
and B. pertussis is largely restricted to the N-terminal 900 amino acids, including the NPNGI(S/T) sequence involved in secretion
and the signal peptide cleavage site (HAQ), at the end of an
unusually long signal peptide. The central and 3' end of the
fhaB gene are variable between strains according to the dot
blot hybridization results (Fig. 3). In addition, the sequencing of
region 2 was complicated by the presence of direct repeats (>95%
similarity) within fhaB, rtw2, and
rtw4 that were designated R1 (431 bp), R2 (271 bp), R3 (160 bp), and R4 (185 bp) (Fig. 3). An interesting hypothesis is that
recombination between these repeats could provide for variation in the
C-terminal portion of the FhaB protein, a feature that is common among
virulence factors in pathogenic neisseriae.
Integrated prophages are known to encode virulence factors in other
bacteria (3, 35). Prophage Pnm1 in region 3 contains 52 ORFs, of which 16 are homologous to Mu and related phages (including genes for the transposase and repressor protein). A total of 13 of
these 16 ORFs, as well as 7 others, also show homologies to a group of
genes from H. influenzae in section 140 to 143 of the genome
(7) (Table 2), suggesting that homologous prophages have
integrated into both serogroup A N. meningitidis and
H. influenzae. In N. meningitidis Z2491, Pnm1 is
flanked by direct repeats of AACT (21 bp downstream of norZ
and 153 bp upstream of gpxA) which are present only once in
nonepidemic serogroup A strains. Pnm1 is the first example of a
neisserial prophage, although N. meningitidis bacteriophages
have previously been reported (2).
Region 7 contains the pseudogene hlyD and the
tolC gene (Fig. 2) in N. meningitidis Z2491.
There is a 61-bp insertion containing a stop codon 1,311 bp after the
ATG start codon of the hlyD pseudogene in N. meningitidis strain Z2491, followed by a direct duplication of the
73 bp preceding the insertion. On the other hand, sequencing region 7 from the ET-37 complex strain revealed that this strain contains a
complete hlyD ORF. In the other N. meningitidis
strains, region 7 was the same size as from the ET-37 complex (6.0 kb) (Fig. 2). Furthermore, sequences of the 3' end of hlyD
revealed that none of these other strains contained the insertion from strain Z2491. Thus, with the exception of Z2491, hlyD is
normally not a pseudogene. Considering the homologies of
hlyD and tolC with genes of the type 1 secretion
apparatus in other bacteria, it seems possible that region 7 might be
involved in the virulence of most N. meningitidis strains.
Region 8 contains the fhuA, rei1, and
dsbA genes in Z2491 (Fig. 2). PCR analysis of region 8 revealed size variations in half of the representative N. meningitidis strains. This region was subsequently sequenced from
several strains to determine the basis for this variability. The data
revealed that rei1 is deleted in the ET-5 complex, whereas
in four other strains (ET-37 complex, lineage 3, ST30, and ST48) an IS
element, IS4351N1 (1,076 bp), has been inserted between
dsbA and the Correia element, c154 (4). However,
fhuA and dsbA were present in all strains. The
fhuA gene is a homologue of an E. coli
ferrichrome-iron receptor that is involved in the uptake of
siderophore-bound iron. fhuA was a pseudogene in some of the
strains (subgroup I, subgroup III, A4 cluster, lineage 3, ST48, and
ST49) due to stop codons introduced by base changes, insertions, or
deletions, but the reading frame was intact in strains belonging to the
ET-37 and ET-5 complex, subgroup IV-1, and in ST30. dsbA is
homologous to genes in E. coli and Pseudomonas syringae that encode a disulfide oxidoreductase. The predicted N. meningitidis DsbA protein contains the motif C-X-X-C
(residues 76 to 79), characteristically present at the active sites of
DsbA enzymes. The N. meningitidis DsbA protein also contains
a typical lipoprotein signal peptide and cleavage site between
A18 and C19 (25) within the motif
L-X-A-C. The amino acid following the LXAC motif is S, indicating that
the N. meningitidis DsbA is sorted to the outer membrane.
Region 9 contains five ORFs of unknown function plus an IS element,
IS4351N2, homologous to that inserted in region 8 in some strains. Sequence analysis revealed that the size differences in region
9 (Fig. 2) were due to the lack of IS4351N2 in strains of
the A4 cluster and ET37 complex. The other five genes in region 9 were
present in all bacteria, and therefore region 9 was retained for
further study of its possible role in virulence.
Regions 4 and 5 are probably not important in N. meningitidis virulence. The former region contains two genes of
unknown function and is absent from a large proportion of N. meningitidis strains, and the latter region encodes a
restriction-modification system and is also present in the commensal
N. lactamica. Region 6 encodes a pseudogene having
homologies with a siderophore receptor in Z2491 but is deleted in most
of the strains. Regions 4, 5, and 6 were not investigated further.
Virulence analysis of the N. meningitidis-specific
islands.
Based on the results described above, regions 2, 3, 7, 8, and 9 were chosen for subsequent investigation of their possible roles
in virulence. The entire regions were inactivated by deletion (Fig. 1)
and confirmed by Southern blot (Fig. 6).
These regions had been sequenced from N. meningitidis strain
Z2491, but this strain is not competent for DNA transformation and
could not be used to construct the deletion mutants. Therefore, regions
7, 8, and 9 were deleted within N. meningitidis strains MC58
(ET-5 complex) and ROU (ET-37 complex). These strains were chosen to give consistent results in the tests for the different biological phenomena and have been tested previously by several models related to
bacterial infection. Neither of these strains contained pseudogenes in
any of these regions. Region 2 was deleted only within strain ROU
because the chromosome of MC58 contains at least two copies at distinct
locations of several of the ORFs from region 2 (see above). The
deletions were constructed by using cloned PCR products from the
chromosome of strain MC58, except that region 2 was deleted using
cloned PCR products from strain ROU. The existence of the desired
deletion mutations was confirmed by PCR and Southern blotting. All of
these mutants grew well on GC agar and were as transformable as are
their parents, indicating that they are piliated.
View larger version (65K):
[in this window]
[in a new window]
|
FIG. 6.
Verification of the deletion mutants. Examples are shown
of the methods used to verify the deletion of the regions. Positions of
the molecular size markers are shown in kilobases for each gel. (A)
Deletion of region 2 from strain ROU. Chromosomal DNA from the parental
strain (lanes 1 and 3) and the region 2 deletion mutant (lanes 2 and 4)
was digested with ClaI. Lanes 1 and 2 were probed with a PCR
product corresponding to the gene fhaC; lanes 3 and 4 were
probed with the cassette "omega" used to replace the region. Due to
its size (about 25 kb) region 2 encompassed several ClaI
fragments; similar results were obtained after probing with PCR
products corresponding to several of the other ORFs. (B) Deletion of
region 3 from strain Z5463. Chromosomal DNA from the parental strain
(lanes 1 and 3) and the region 2 deletion mutant (lanes 2 and 4) was
digested with SgfI. Lanes 1 and 2 are the pulsed-field gel
electrophoresis analysis of the deletion. Note the disappearance of a
band at about 194 kb in the mutant and the new band appearing with a
size of about 146 kb; this corresponds to the deletion of about 50 kb
(region 3). Lanes 3 and 4 were probed with a PCR product corresponding
to a portion of the phage transposase gene. (C) Deletion of region 8 from strains MC58 and ROU. Lane 1, MC58 parental strain; lane 2, MC58
region 8 deletion; lane 3, ROU; lane 4, ROU region 8 deletion.
Chromosomal DNA was digested with ClaI, and Southern blots
were probed with a PCR product corresponding to the entire region 8 from strain Z2491.
|
|
First, the deletion mutants of strain ROU were assessed for their
ability to adhere to and to invade HUVEC, a model for interactions with
the endothelial cells of the BBB. No differences were detected between
the parental strain and the deletion mutants (Table
4), indicating that regions 2, 7, 8, and
9 do not affect interactions with endothelial cells.
Second, the deletion mutants were tested for sensitivity to the
bactericidal activity of complement (Table
5). For these bactericidal assays,
meningococci were incubated with 10% guinea pig serum as a complement
source supplemented with different dilutions of antiserum raised
against the homologous parental strain as an antibody source. The
deletion mutants of region 2, 7, 8, or 9 were as resistant to the
bactericidal action of the immune serum as their parents in contrast to
the serum-sensitive ppk mutant control. Thus, these regions
do not seem to be involved in serum resistance.
Finally, the mutants were tested for their ability to multiply in the
bloodstream in an infant rat model. These tests were performed with the
mutants deleted for regions 7, 8, and 9 of strain MC58, which causes
high levels of bacteremia in this model. Region 2 could not be tested
in this strain because ORFs in region 2 are present at two distinct
locations in MC58, the flanking sequences are not known in this strain,
and the sequence of region 2 is highly variable between strains. For
each experiment, half of the animals in each litter were injected
intraperitoneally with the parental strain, and the other half were
injected with the deletion mutant; the numbers of bacteria in the blood
were then quantitated for 24 h. No differences were detected with
deletion mutants of regions 7 or 9 (data not shown), showing that these regions are not important for causing bacteremia. On the other hand,
lower numbers of bacteria in blood samples than with the parental
control were consistently found with the MC58 deletion mutant lacking
region 8 (Fig. 7). Furthermore, only 2 of
15 infant rats died within 48 h after injection of the region 8 deletion mutant, whereas 11 of 14 rats died after injection of the
parental control. These data demonstrate that region 8 is important for bacterial survival in the bloodstream.
View larger version (18K):
[in this window]
[in a new window]
|
FIG. 7.
Bacterial concentrations in sequential blood samples
after the intraperitoneal infection of infant rats. One typical
experiment is shown with average counts and standard deviations.  ,
MC58 parental strain;  , MC58  region 8.
|
|
Virulence determinants may be carried on bacterial prophages
(3, 35). Though region 3, including Pnm1, is only present in epidemic serogroup A strains, we deleted this region in order to investigate its possible role in the pathogenesis of these strains.
We identified a subgroup IV-1 strain, Z5463, that is transformable and
then constructed a region 3 deletion of this strain. Z5463 does not
yield high levels of bacteremia in infant rats, and we could not carry
out this assay. Because of this strain's sensitivity to complement,
bactericidal assays were performed with guinea pig complement in the
absence of opsonizing antiserum. The deletion of region 3 did not
affect the resistance to complement killing; both the wild-type and
mutant strains were killed at concentrations of serum of >1%. Neither
did deletion of region 3 affect the interaction of Z5463 with
monolayers of HUVEC in the adhesion and invasion assay (data not shown).
| |
DISCUSSION |
We investigated genes within eight DNA islands that comprise
approximately 5% of the chromosome (100 kb) of N. meningitidis and are lacking in N. gonorrhoeae (Table
6). Many of the genes in these islands
seem to be irrelevant to the differences in pathogenicity between these
species. However, the islands designated as regions 2, 7, 8, and 9 contained genes homologous to genes from other pathogenic bacteria that
are involved in virulence and were present in diverse, representative
N. meningitidis strains. Deletion mutants lacking region 8 were impaired in their ability to cause bacteremia in an infant rat
model. A ninth island containing genes encoding the synthesis of
capsular polysaccharide is also known to be necessary for dissemination
in the bloodstream. Thus, two of the nine known islands in N. meningitidis are important for bloodstream dissemination, a
prerequisite for causing meningitis. It remains possible that the other
seven islands may play a role in virulence that was not detected by the
assays used here.
Region 8 contains homologues of fhuA and dsbA, as
well as a new gene called rei1. However, rei1 is
deleted within MC58, the N. meningitidis strain that was
used in the infant rat model, leaving only fhuA and
dsbA as candidates for genes involved in bacteremia. FhuA is
a siderophore-iron receptor in E. coli. The fhuA
gene was defective in several of the N. meningitidis
strains, suggesting that it is not important for virulence. The DsbA
homologue is present in diverse N. meningitidis strains and
is probably sorted to the outer membrane. DsbA proteins in other
pathogenic bacteria have been shown to play an indirect role in
virulence. They are involved in the correct folding of proteins
responsible for both the secretion of invasion proteins by S. flexneri (36) and for its intracellular survival
(37), as well as for the formation of type IV fimbriae in
enteropathogenic E. coli (38). The DsbA homologue
in region 8 does not affect piliation because mutants deleted for
region 8 were normally transformable and adhered to endothelial cells
in a pilus-dependent assay as efficiently as their parent.
In several bacterial species (for example, E. coli,
Helicobacter pylori, Salmonella enterica, and
Yersinia enterocolitica), genes encoding increased
pathogenicity are clustered in so-called pathogenicity islands (PAIs)
(11). PAIs are usually large (50 to 200 kb) and often have a
different GC-content from that of the host chromosome. PAIs can be
genetically unstable due to flanking repetitive sequences or IS
elements and, within the Enterobacteriaceae, tRNA loci often
serve as targets for their integration and excision. None of the
N. meningitidis-specific islands in regions 2 to 9 fulfills
all of these criteria, and most fulfill none of them. Thus, the generic
term "DNA island" seems more appropriate for these sequences in the
neisseriae than the term PAI. We note that some of the differences
between these islands and PAIs may reflect the fact that the neisseriae
are naturally transformable and readily undergo genetic exchange via
homologous recombination after DNA transformation, whereas mobile
genetic elements are often more important for the
Enterobacteriaceae.
N. meningitidis probably contains more DNA islands than the
eight described here. For example, upstream of region 2 in N. meningitidis is another small island containing orf1
plus a 106-bp long Correia element (Fig. 3), whereas two small ORFs are
present at that location in N. gonorrhoeae. Similarly, three
other islands in the opcA and opcB regions
were described elsewhere (39), and the regions corresponding
to three N. meningitidis-specific clones (32)
were also not identified here. Finally, region 1 encodes genes involved
in capsular biosynthesis and was not investigated further in this
report. Thus, a considerable proportion of the differences between
N. meningitidis and N. gonorrhoeae may be encoded
by species-specific DNA islands, only some of which have been analyzed here.
The eight islands were present in most of the strains studied, and
their overall organization was conserved, despite the presence of a
number of pseudogenes. The conservation of region 5 (restriction-modification system) can be explained, since functional
restriction-modification systems act as selfish genetic units
(21). Similarly, region 4 of N. meningitidis is
replaced by a restriction-modification system in N. gonorrhoeae FA1090 (see Table 3, hpaIIM and
hphIR), which has no homologue in N. meningitidis
and whose corresponding genes exhibit a low GC content. The presence of
some of the other islands might be accounted for by the selfish operon
theory (16), whereby the integrity of a group of genes
linked by a common function is favored by stochastic mechanisms. On the
other hand, conserved, functional genes are classically the result of
natural selection by the environment, and those in the N. meningitidis-specific islands should therefore be beneficial to
N. meningitidis in those environments which it specifically
inhabits. N. meningitidis and N. gonorrhoeae
differ not only in their characteristic pathogenicities but also in the
anatomical sites they colonize. Thus, these islands might be relevant
to multiplication under the particular biochemical conditions and in
the presence of the microbial competition experienced by the
meningococcus in its natural habitat. Region 2 encodes fhaB
and fhaC homologues, as well as a number of ORFs with no obvious homologies, and is present in all N. meningitidis
strains analyzed. The FHA protein is an adhesin of B. pertussis. Although the N. meningitidis fhaB homologue
is considerably shorter than that in B. pertussis, it could
produce a 224-kDa protein after cleavage of the signal peptide that is
of comparable size to the processed FHA. The conservation of an
amino-terminal asparagine-rich domain plus the conserved NPNGI
motif might reflect a similar secretion mechanism. However, the C
termini differ extensively between B. pertussis and N. meningitidis; such sequence variability might reflect functional
differences or diversifying selection due to the human immune system.
Serum sensitivity and adhesion were not affected by deletion of region
2, indicating that it is not essential for these phenotypes in N. meningitidis. It was not possible to test whether region 2 is
important for blood dissemination because MC58, the strain used for
these tests, possesses two copies of ORFs from region 2 at distinct
locations, and ROU, the other strain in which a deletion of region 2 was introduced, does not replicate in the infant rat for unknown
reasons, although it is a clinical isolate from a case of meningitis
with septicemia. Notwithstanding the lack of sequence information, we
were able to produce a double mutation in the gene fhaB in
the two region 2 loci of another strain, 8013, by the deletion of one
region 2 followed by the insertional inactivation of the
fhaB gene in the other. The double mutant had no alteration
with respect to the in vitro or in vivo assays (data not shown);
however, it is not possible to extrapolate this result to consider the
whole of region 2.
In summary, we describe the characteristics and distribution of eight
DNA islands that are specific to N. meningitidis and show
that genes on one of these islands are important for virulence. These
results will form the basis of additional experiments to develop new
protein vaccine candidates and to define factors that are important for
bacterial invasion of the bloodstream.
This work was supported by a grant from SmithKline Beecham
Biologicals s.a., Rixensart, Belgium.
During the course of this work, sequences corresponding to two of
the regions have been submitted to GenBank by other groups. Region 5 with its flanking rfaE and rfaD genes has been
sequenced in the N. meningitidis serogroup B strain CDC
8201085 (C. M. Kahler and D. S. Stephens, GenBank accession number
AF125564), where it is 98.6% identical to region 5 of Z2491. The
rfi1 and rfi2 genes are designated
nmgII and nmgI, respectively. Region 7 was sequenced in the N. meningitidis serogroup C strain IR1075
(I. Stojiljkovic, GenBank accession number AF121772), where it contains
natC, which is 99.8% identical to tolC of Z2491,
and natD, with 99% identity to hlyD of Z2491
over the first 1,310 bases.
| 1.
|
Bart, A.,
I. G. A. Schuurman,
M. Achtman,
D. A. Caugant,
J. Dankert, and A. van der Ende.
1998.
Random amplified polymorphic DNA (RAPD) genotyping of serogroup A meningococci yields similar results to multilocus enzyme electrophoresis (MEE) and reveals new genotypes.
J. Clin. Microbiol.
36:1746-1749[Abstract/Free Full Text].
|
| 2.
|
Cary, S. G., and D. H. Hunter.
1967.
Isolation of bacteriophages active against Neisseria meningitidis.
J. Virol.
1:538-542[Abstract/Free Full Text].
|
| 3.
|
Cheetham, B. F., and M. E. Katz.
1995.
A role for bacteriophages in the evolution and transfer of bacterial virulence determinants.
Mol. Microbiol.
18:201-208[CrossRef][Medline].
|
| 4.
|
Correia, F. F.,
S. Inouye, and M. Inouye.
1988.
A family of small repeated elements with some transposon-like properties in the genome of Neisseria gonorrhoeae.
J. Biol. Chem.
263:12194-12198[Abstract/Free Full Text].
|
| 5.
|
Delisse-Gathoye, A. M.,
C. Locht,
F. Jacob,
M. Raaschou-Nielsen,
I. X. R. J. Heron,
M. de Wilde, and T. Cabezon.
1990.
Cloning, partial sequence, expression, and antigenic analysis of the filamentous hemagglutinin gene of Bordetella pertussis.
Infect. Immun.
58:2895-2905[Abstract/Free Full Text].
|
| 6.
|
Domenighini, M.,
D. Relman,
C. Capiau,
S. Falkow,
A. Prugnola,
V. Scarlato, and R. Rappuoli.
1990.
Genetic characterization of Bordetella pertussis filamentous haemagglutinin: a protein processed from an unusually large precursor.
Mol. Microbiol.
4:787-800[CrossRef][Medline].
|
| 7.
|
Fleischmann, R. D.,
M. D. Adams,
O. White,
R. A. Clayton,
E. F. Kirkness,
A. R. Kerlavage,
C. J. Bult,
J. F. Tomb,
B. A. Dougherty,
J. M. Merrick,
K. Mckenney,
G. Sutton,
W. Fitzhugh,
C. Fields,
J. D. Gocayne,
J. Scott,
R. Shirley,
L. I. Liu,
A. Glodek,
J. M. Kelley,
J. F. Weidman,
C. A. Phillips,
T. Spriggs,
E. Hedblom,
M. D. Cotton,
T. R. Utterback,
M. C. Hanna,
D. T. Nguyen,
D. M. Saudek,
R. C. Brandon,
L. D. Fine,
J. L. Fritchman,
J. L. Fuhrmann,
N. S. M. Geoghagen,
C. L. Gnehm,
L. A. McDonald,
K. V. Small,
C. M. Fraser,
H. O. Smith, and J. C. Venter.
1995.
Whole-genome random sequencing and assembly of Haemophilus influenzae Rd.
Science
269:496-512[Abstract/Free Full Text].
|
| 8.
|
Frosch, M.,
C. Weisgerber, and T. F. Meyer.
1989.
Molecular characterization and expression in Escherichia coli of the gene complex encoding the polysaccharide capsule of Neisseria meningitidis group B.
Proc. Natl. Acad. Sci. USA
86:1669-1673[Abstract/Free Full Text].
|
| 9.
|
Goodman, S. D., and J. J. Scocca.
1988.
Identification and arrangement of the DNA sequence recognized in specific transformation of Neisseria gonorrhoeae.
Proc. Natl. Acad. Sci. USA
85:6982-6986[Abstract/Free Full Text].
|
| 10.
|
Goodman, S. D., and J. J. Scocca.
1991.
Factors influencing the specific interaction of Neisseria gonorrhoeae with transforming DNA.
J. Bacteriol.
173:5921-5923[Abstract/Free Full Text].
|
| 11.
|
Hacker, J.,
G. Blum-Oehler,
I. Mühldorfer, and H. Tschäpe.
1997.
Pathogenicity islands of virulent bacteria: structure, function and impact on microbial evolution.
Mol. Microbiol.
23:1089-1097[CrossRef][Medline].
|
| 12.
|
Hammerschmidt, S.,
C. Birkholz,
U. Zähringer,
B. D. Robertson,
J. Van Putten,
O. Ebeling, and M. Frosch.
1994.
Contribution of genes from the capsule gene complex (cps) to lipooligosaccharide biosynthesis and serum resistance in Neisseria meningitidis.
Mol. Microbiol.
11:885-896[CrossRef][Medline].
|
| 13.
|
Hammerschmidt, S.,
A. Müller,
H. Sillmann,
M. Müllenhoff,
R. Borrow,
A. Fox,
J. Van Putten,
W. Zollinger,
R. Gerardy-Schahn, and M. Frosch.
1996.
Capsule phase variation in Neisseria meningitidis serogroup B by slipped-strand misspairing in the polysialyltransferase gene (siaD): correlation with bacterial invasion and the outbreak of meningococcal disease.
Mol. Microbiol.
20:1211-1220[Medline].
|
| 14.
|
Kellogg, D. S.,
W. L. Peacock,
W. E. Deacon,
L. Brown, and C. I. Pirkle.
1963.
Neisseria gonorrhoeae. I. Virulence genetically linked to clonal variation.
J. Bacteriol.
85:1274-1279[Abstract/Free Full Text].
|
| 15.
|
Kroll, J. S.,
K. E. Wilks,
J. L. Farrant, and P. R. Langford.
1998.
Natural genetic exchange between Haemophilus and Neisseria: intergeneric transfer of chromosomal genes between major human pathogens.
Proc. Natl. Acad. Sci. USA
95:12381-12385[Abstract/Free Full Text].
|
| 16.
|
Lawrence, J. G., and J. R. Roth.
1996.
Selfish operons: horizontal transfer may drive the evolution of gene clusters.
Genetics
143:1843-1860[Abstract].
|
| 17.
|
Locht, C.,
P. Bertin,
F. D. Menozzi, and G. Renauld.
1993.
The filamentous haemagglutinin, a multifaceted adhesion produced by virulent Bordetella spp.
Mol. Microbiol.
9:653-660[CrossRef][Medline].
|
|
![[Feedback]](/icons/banner/feedbackACT.gif){kind=icon}
![[For Subscribers]](/icons/banner/subscriberACT.gif){kind=icon}
![[Archive]](/icons/banner/archiveACT.gif){kind=icon}
![[Search]](/icons/banner/searchACT.gif){kind=icon}
![[Contents]](/icons/banner/tocACT.gif){kind=icon}
Top
Abstract
Introduction
