Sequence Diversity of N. meningitidis fHBP
• JID 2009:200 (1 August) • 379
M A J O R A R T I C L E
Sequence Diversity of the Factor H Binding Protein
Vaccine Candidate in Epidemiologically Relevant
Strains of Serogroup B Neisseria meningitidis
Ellen Murphy,1Lubomira Andrew,1Kwok-Leung Lee,1Deborah A. Dilts,1Lorna Nunez,1Pamela S. Fink,1
Karita Ambrose,1Ray Borrow,4Jamie Findlow,4Muhamed-Kheir Taha,5Ala-Eddine Deghmane,5Paula Kriz,6
Martin Musilek,6Jitka Kalmusova,6Dominique A. Caugant,7Torill Alvestad,7Leonard W. Mayer,2Claudio T. Sacchi,2,8
Xin Wang,2Diana Martin,9Anne von Gottberg,10Mignon du Plessis,10Keith P. Klugman,3,10Annaliesa S. Anderson,1
Kathrin U. Jansen,1Gary W. Zlotnick,1and Susan K. Hoiseth1
1Wyeth Vaccines Research, Pearl River, New York;
University, Atlanta, Georgia,
Bacterial Infections Unit, Paris, France;
and Meningeal Pathogens Research Unit, National Institute for Communicable Diseases, Johannesburg, South Africa
2Centers for Disease Control and Prevention, and
4Health Protection Agency, Manchester Royal Infirmary, Manchester, United Kingdom;
6National Institute of Public Health, Prague, Czech Republic;
8Instituto Adolfo Lutz, Sao Paulo, Brazil;
3Rollins School of Public Health, Emory
5Institut Pasteur, Invasive
7Norwegian Institute of Public Health, Oslo,
9Institute of Environmental Science and Research, Porirua, New Zealand; and
dergoing clinical trials in candidate vaccines against invasive meningococcal serogroup B disease. We report an
extensive survey and phylogenetic analysis of the diversity of fhbp genes and predicted protein sequences ininvasive
clinical isolates obtained in the period 2000–2006.
Nucleotide sequences of fhbp genes were obtained from 1837 invasive N. meningitidis serogroup B
(MnB) strains from the United States, Europe, New Zealand, and South Africa. Multilocus sequencetyping(MLST)
analysis was performed on a subset of the strains.
Every strain contained the fhbp gene. All sequences fell into 1 of 2 subfamilies (A or B), with 60%–
75% amino acid identity between subfamilies and at least 83% identity within each subfamily. One fHBP sequence
may have arisen via inter-subfamily recombination. Subfamily B sequences were found in 70% of the isolates, and
subfamily A sequences were found in 30%. Multiple fHBP variants were detected in each of the common MLST
clonal complexes. All major MLST complexes include strains in both subfamily A and subfamily B.
The diversity of strains observed underscores the importance of studying the distribution of the
vaccine antigen itself rather than relying on common epidemiological surrogates such as MLST.
Recombinant forms of Neisseria meningitidis human factor H binding protein (fHBP) are un-
Invasive disease caused by Neisseria meningitidis is a
rapidly progressing, disseminated infection with a case
fatality rate of ∼10% , and 10%–20% of survivors
experience serious permanent sequelae (eg, neurolog-
ical impairment, digit, limb, or hearing loss). Vaccines
for meningococcal serogroups A, C, Y, and W135,
Received 18 November 2008; accepted 11 March 2009; electronically published
17 June 2009.
Presented in part: International Pathogenic Neisseria Conference, Rotterdam,
The Netherlands, 7–12 September 2008 (abstracts O45 and P137).
Financial support: Wyeth Vaccines Research, Institut Pasteur (M-K.T and A-
E.D.), and the Internal Grant Agency, Ministry of Health of the Czech Republic
(Grant 1A8688-3 to P. K.).
Reprints or correspondence: Susan K. Hoiseth, Ph.D., Wyeth Vaccines Research,
401 N. Middletown Rd., Pearl River, NY 10965 (firstname.lastname@example.org).
The Journal of Infectious Diseases
? 2009 by the Infectious Diseases Society of America. All rights reserved.
which were initially based on the serogroup-defining
polysaccharides and more recently on conjugates of the
Potential Conflicts of Interest: E.M., L.A., K.L.L., D.A.D., L.N., P.S.F., K.A., A.S.A.,
K.U.J., G.W.Z., and S.K.H. are employees of Wyeth. R.B., J.F., M.K.T., A.-E.D., P.K.,
M.M., J.K., D.A.C., T.A., L.W.M., C.T.S., X.W., D.M., A.v.G., M.d.P., and K.P.K.
received funding from Wyeth for provision of the strain collections, associated
strain information, and some of the sequence data reported here. R.B. has received
assistance to attend scientific meetings from Wyeth, Sanofi Pasteur, and Baxter
Bioscience; has served as an ad hoc consultant for Wyeth, GlaxoSmithKline,
Novartis, Sanofi Pasteur, and Baxter Bioscience; and has performed other contract
research funded by Wyeth, Novartis Vaccines, Baxter Bioscience, GlaxoSmithKline,
Sanofi Pasteur, Alexion Pharmaceuticals, Emergent Europe, and Merck. J.F. has
received assistance to attend scientific meetings from Wyeth and Novartis
Vaccines. L.W.M. and X.W. have received funding (under Cooperative Research
and Development Agreements) from Novartis Vaccines and Wyeth. K.P.K. has
consulted for Wyeth and Novartis and has received other research funding from
Wyeth. D.A.C. has served as an ad hoc consultant for Wyeth, Novartis, and Sanofi
Pasteur. P.K. has served as an ad hoc consultant for Wyeth. A.V.G. has consulted
for Norvartis. D.M. has received contract funding from Novartis. M.-K.T has also
received funding from Chiron and Sanofi Pasteur for contract research and has
consulted for Wyeth, GlaxoSmithKline, Sanofi Pasteur, and Novartis.
380 • JID 2009:200 (1 August) • Murphy et al
polysaccharides with protein carriers, have been available for
several years. In countries in which these vaccines, particularly
meningococcal serogroup C conjugate vaccines, have been
adopted, disease caused by strains of these serogroups has sig-
nificantly decreased . However, a polysaccharide-based vac-
cine is not feasible for serogroup B (MnB) strains, because of
the similarity of the MnB capsular polysaccharide, a homo-
polymer of (a2r8) sialic acid,to thepolysaccharidecomponent
of human neural cell adhesion molecule [3, 4]. Therefore, the
focus of attention for MnB vaccines has shifted to surface-
exposed protein antigens that are capable of eliciting protective
Factor H binding protein (fHBP), an ∼28 kD lipoprotein,
was first identified as a protective antigen through biochemical
fractionation of a soluble outer membrane preparationthathad
been shown to elicit abroadlycross-reactive,PorA-independent
bactericidal response against a variety of heterogeneous sero-
group B strains [5–7]. This lipoprotein, called LP2086 by
Fletcher et al. , was also identified by Masignani et al. ,
who referred to it as GNA1870. It has recently been shown to
bind complement factor H , with specificity for factor H
from humans or higher primates . Accordingly, the protein
has been renamed fHBP. Recruitment of factor H to the surface
of the bacterium, thereby inhibitingthealternativecomplement
pathway, may be an important survival mechanism for path-
ogenic Neisseria . Individuals carrying a genetic polymor-
phism in a presumed regulatory region for factor H have been
reported to have elevated serum levels of factor H and an
increased risk for meningococcal disease , and strains in
which fhbp has been deleted are more susceptible to comple-
ment-mediated killing . Thus, in addition to its ability to
elicit a bactericidal response in humans , fHBP plays an
important role in virulence for this organism.
Early analysis of fHBP sequences from 63 strains that rep-
resented a variety of N. meningitidis lineages identified 39
unique protein variants (differing by at least 1 amino acid) that
we classified into 2 distinct groups, subfamilies A and B, with
183% amino acid identity within a subfamily but only 60%–
75% amino acid identity between the subfamilies . Bacte-
ricidal antibody responses were found to be mainly specific to
each fHBP subfamily . In a similar analysis, Masignani et
al.  observed 24 unique variants among 78 strains and clas-
sified fHBP into 3 groups of variants: group 1 (corresponding
to subfamily B) and groups 2 and 3 (together corresponding
to subfamily A).
Members of the fHBP family are currently being evaluated
clinically in 2 separate MnB candidate vaccines, 1 of which
contains both a subfamily A and subfamily B protein, and the
other of which contains only a subfamily B protein (in com-
bination with other antigens). Measuring vaccine performance
and demonstrating the sensitivity of disease-causing strains to
vaccine-induced antibody is more complex for protein-based
vaccines than for capsule-based vaccines. To lay the ground-
work to address these issues, we have undertaken a compre-
hensive survey to examine the distribution and diversity of
fHBP in an epidemiologically relevant, systematic collection of
recent MnB clinical isolates.
MATERIALS AND METHODS
the public health laboratories of the United States, Norway,
France, the Czech Republic, and from the Health Protection
Agency in Manchester (HPA), which covers England, Wales,
and Northern Ireland (table 1). For the United States, strains
were from the Active Bacterial Core Surveillance sites ,
which collectively cover ∼13% of the US population. Available
isolates for the period 2000–2005 (180% of reported cases for
those years) were included in the study. FortheEuropeancoun-
tries, strains were collected in a systematic way by order of date
received at the reference laboratory. Every seventh (Czech Re-
public) or every eighth (HPA, France, and Norway)isolatefrom
2001–2006 was included. The isolates from Europe that were
included inthis collectionthusrepresent∼13%ofinvasiveMnB
isolates from the respective reference laboratories for theperiod
covered by this study. The starting collections from these ref-
erence laboratories are estimated to cover 80%–85% of all in-
vasive MnB isolates in France  and 50%–70% of those in
the Czech Republic. For the HPA collection and Norway, cov-
erage is estimated to be 95%. A total of 9 strainsfromallsources
were not viable upon receipt at Wyeth and were excluded from
We also sequenced fhbp from 574 additional invasive MnB
strains from various collections and years, for a total of 1837
strains (table 1). These additional strains were mainly from the
United States, HPA, Norway, Czech Republic, New Zealand,
and South Africa. They were included in our analysis of se-
quence diversity but were not included in analyses of variant
frequency unless otherwise noted (eg, isolates from South Af-
rica, for which all available isolates for 2005 were evaluated).
Polymerase chain reaction (PCR) and sequencing.
strategy for sequencing fhbp employed an initial PCR reaction
with primers designed to recognize conserved regionsupstream
of and partially overlapping the leader peptide of fhbp and
within 45 nucleotides downstream of the termination codon
GGTAAATTATCGTG-3?, respectively). PCR templates were
prepared by boiling several N. meningitidis colonies from choc-
olate agar plates in 100 mL of distilled H2O for 5 min, then
diluting 1:4 in distilled H2O. PCR reactions (50 mL total vol-
ume) contained 4 mL DNA, Premix Taq-EX TAKARA enzyme
cocktail (TAKARA Bio USA), and primers (0.4 mmol/L each).
The amplification steps were 95?C for 5 min, 33 cycles of 95?C
A set of 1263 invasive MnB strainswasobtainedfrom
Neisseria meningitidis Serogroup B Strains
2000 2001 2002 2003 2004 2005 2006
Systematically collected strain collection
Centers for Disease Control and Prevention, Atlanta, Georgia 2000–2005
England, Wales, Northern Ireland
Health Protection Agency, Manchester
Neisseria Unit, Institut Pasteur, Paris
Norwegian Institute of Public Health, Oslo
National Institute of Public Health, Prague
Centers for Disease Prevention and Control, Atlanta, Georgia 1996–1999 (mainly); 2006
England, Wales, Northern Island
Health Protection Agency, Manchester
National Institute of Public Health, Prague
Norwegian Institute of Public Health, Oslo
Institute of Environmental Science and Research, Porirua
National Institute for Communicable Diseases, Johannesburg 2005d
Chile (1 strain); Netherlands (5 strains); France (2 strains)
aIncludes 57 strains sequenced by Fletcher et al. .
bIncludes strains from 2001–2006 not included in the 1263 strain set and strains from 1984–2000 and 2007.
cIncludes additional strains from 2001–2006 not included in the 1263 strain set
dThese strains represent all available serogroup B strains collected in 2005 .
382 • JID 2009:200 (1 August) • Murphy et al
for 50 s, 59?C for 50 s, and 72?C for 50 s, followed by an
extension step at 72?C for 7 min. Amplified DNA was purified
using AMPure magnetic beads (Agencourt) and resuspended
in 80 mL 10 mmol/L Tris-acetate buffer (pH 8.0).
Sequencing primers were slightly internal to the initial PCR
primers and were hybridized to conserved regions in the N-
terminus of fhbp or to subfamily-specific C-terminal regions.
Conserved N-terminal and internal primerswere5?-TATGACT-
CGGA-3?, respectively. Subfamily-specific C-terminal and in-
ternal primers were 5?-TACTGTTTGCCGGCGATG-3?and
5?-GAATGCTTTGCCGTGATACTCGGCT-3?, respectively, for
subfamily A and were 5?-TTCGGACGGCATTTTCACAATGG-
subfamily B. All primer reactions were run for all strains; only
those corresponding to the correct subfamily were successful.
Sequencing reactions contained 4 mL PCR product, 2 mL 5?
buffer (ABI PRISM BigDye Terminator V.3.1 Cycle Sequencing
Ready Reaction Kit), 4 mL ABI BDT_V3 polymerase, and 1 mL
primer (3.2 pmol/mL) in a reaction volume of 20 mL.
Plates were handled on a Rapid Plate 96 channel Qiagen Bio
Robot and with a Multi-probe II 4 tip system. Reactions were
heated to 96?C for 30 s and then cycled at 96?C for 10 s, 50?C
for 5 s, and 60?C for 4 min for 25 cycles. Electrophoresis was
performed on an ABI3730 DNA sequencer (Applied Biosys-
tems). Sequences were examined to verifydouble-strandedcov-
erage with Sequencher 4.0, and sequences were subsequently
trimmed to match the length of the mature protein (lipid at-
tachment cysteine to termination codon). fhbpvariantnumbers
were assigned sequentially to new sequences with use of a no-
menclature that indicated subfamily, protein, and nucleotide
variant (eg, A22_003 indicates the third nucleotide sequence
variant of subfamily A protein variant number 22). This no-
menclature replaces that used in our other publications .
Initially, all strains were sequenced twice from independent
PCR reactions. After sequencing ∼100 strains with no discrep-
ancies between duplicates, only those strains containing new
fhbp sequences were confirmed by repeat PCR and sequencing.
Sequences were deposited in GenBank with accession numbers
FJ184079-FJ184274. Multilocus sequence typing (MLST) was
performed according to the protocols at the Neisseria MLST
Web site (http://pubmlst.org/neisseria/) .
Protein sequences were aligned
with ClustalW 1.83 . Distances were calculated by the
neighbor-joining method  in ClustalW, with correction for
multiple substitutions . Consensus sequences were calcu-
lated using the EMBOSS utility “con” . Each variant had
equal weight in calculation of the consensus, with no adjust-
ment for variant frequency. Trees were displayed using MEGA
software, version 4.0 . Independent subfamily A and sub-
family B alignments were generated byseparatingthealignment
of all sequences, without realignment or gap removal, to pre-
serve a common numbering system. Networks were generated
using Splitstree, version 4.0 , with default parameters.
For construction of minimum spanning trees, the aligned
unique protein sequences were reduced to the set of 77 (sub-
family A) or 101 (subfamily B) variable positions, and the
remaining data were converted to a numerical matrix. Gen-
eration of minimum spanning trees and construction of clus-
ters of fHBP sequences were performed with Bionumerics, ver-
sion 5.1 (Applied Maths), with use of a categorical coefficient.
Linkage priorities were assigned as follows: (1) maximumnum-
ber of single-locus variants, (2) maximum number of single-
and double-locus variants, and (3) maximum number of se-
quences belonging to a single type, with hypothetical types
Summary of fhbp sequencing.
strains examined. All strains encoded a full-length protein ex-
cept for 1 (0030/01), which carried a premature stop codon at
nucleotide 366. There were 218 unique nucleotide sequences
encoding 173 unique protein variants of mature fHBP. Two
variants carried a substitution in the termination codon that
resulted in a 3 amino acid extension at the C-terminus (B82
and B112). Phylogenetic analysis indicated that all genes fell
into 1 of 2 groups (figure 1A), which had previously been
named subfamilies A and B . There were 74 unique fHBP
subfamily A and 99 unique subfamily B protein sequences, of
which 41 and 67, respectively, were represented by just 1 strain.
Pairwise identities within a subfamily ranged from 83% to99%;
between subfamilies, pairwise identities ranged from 60% to
75%. A consensus alignment is shown in figure 2. Evaluation
of individual residue variation revealed that 111 amino acids
(44% of all residues) were invariant among all proteins of both
subfamilies (coded red in figure 2). Within subfamily B, 166
residues were invariant; subfamily A is somewhat less diverse,
with 192 positions completely conserved. At 32 positions, the
residues were subfamily-defining; that is, the residues were
identical within and characteristic for either subfamily A or B
(coded yellow in figure 2).
Despite the clear existence of 2 subfamilies, 1 variant, A62
(found in 2 genetically and geographically unrelated strains,
0167/03 [Czech Republic, ST-5932] and 21626 [France, ST-
2688]), was an apparent result of recombination between sub-
families. The N-terminus of A62_001 was 100% identical to
B09 from nucleotides 1–556, and its C-terminus was 100%
identical to A22 from nucleotides 507–768. The 2 A62 genes
differed by 2 nucleotides. Most of the diversity between the 2
families is found in the C-terminal domain, and therefore, the
recombinant variant was assigned to subfamily A. Such inter-
subfamily recombination events evidently occur very rarely, be-
fhbp was detected in all 1837
Sequence Diversity of N. meningitidis fHBP • JID 2009:200 (1 August) • 383
was excluded from the alignment. The tree was bootstrapped 500 times and drawn using MEGA software. B, SplitsTree analysis of subfamily A fHBP
variants. C, SplitsTree analysis of subfamily B variants. The trees in panels B and C are based on full-length alignments of subfamily A and subfamily
B variants, respectively. D, SplitsTree analysis of the N-terminal domain (residues 1–139 ) of all subfamily A and B variants. Representative variants
are labeled. Subdomain assignments are described in the Results. In the nomenclature of Masignani et al. , variant 2 is equivalent to subfamily A
variants carrying the N2 domain (N2C1 and N2C2); variant 3 is equivalent to subfamily A variants carrying the N1 domain (N1C1 and N1C2); and
variant 1 corresponds to subfamily B variants.
A, Neighbor-joining tree generated based on the ClustalW alignment of 172 unique fHBP protein sequences. The truncated B110 sequence
cause only these 2 recombinants were identified among 11800
Network analysis of fHBP sequences.
work analysis provides a better description of sequence rela-
tionships than standard tree representations when, as in the
case of N. meningitidis, evolution most likely proceeds via hor-
izontal transfer and recombination rather than linear, branched
speciation events . Network analysis using SplitsTree, ver-
sion 4.0, subdivided subfamily A into 4 groups with multiple
and equally likely paths among them (figure 1B). Visual ex-
amination of the aligned sequences in each group further re-
vealed the existence of 2 N-terminal domain types (N1 and
N2) and 2 C-terminal domain types (C1 and C2), which have
apparently recombined in all 4 possible combinations. The 2
groups with the N1 domain, N1C1 and N1C2, are together
equivalent to variant 3 in Masignani et al. , whereas the N2-
containing variants, N2C1 and N2C2, together correspond to
variant 2 in Masignani et al. . The average pairwise distance
of all N1-containing variants versus those with the N2 domain
was 0.15266, whereas the distance between variants carrying
C1 (ie, N1C1 and N2C1) versus those with C2 (N1C2 and
N2C2) was 0.11529, indicating that, within subfamily A, more
diversity is contributed by the N-terminal domains than by the
C-terminal domains. However, the genetic distance between
either of the N1- or N2-containing subfamily A groups and
subfamily B (0. 47303 and 0.39271, respectively) was 2.5–3
times greater than the distance between the 2 subgroups of
subfamily A, reaffirming the essentially bifurcated nature of the
fHBP phylogenetic tree (figure 1A).
N1 and N2 were distinguished by 20 “signature” residues
(truncated variant B110 excluded), 73 subfamily A variants (hybrid A62 excluded), and each of the 4 groups of subfamily A variants N1C1 (
N1C2 (), N2C1 (), and N2C2 (). A57, which carried an N3 domain, was included only in the complete subfamily A alignment.n p 23n p 21n p 14
Residues are numbered according to the full-length structure of the B01 variant ; asterisks (*) indicate the position of the KDN residues in some
variants, which are absent in the reference sequence B01. Residues are color-coded according to the type of variation exhibited at each position.
Residues conserved between both subfamilies are colored in red and are 100% identical among all variants; orange indicates residues that differ in
just 1 variant. Subfamily-defining residues (ie, residues that differ between subfamily A and subfamily B but are invariant within each subfamily) are
shown in yellow. Subdomain signature residues (shown in green) are those that define the 4 subgroups of subfamily A (figure 1B); these are noted
as “X” in the consensus line for all subfamily A variants. Variable residues (black on white) are those that varied in more than a single variant without
conforming to any pattern. Many of these are conserved in the majority of sequences, and thus display identical consensus residues across all 6
groups. “X” indicates that there is no consensus residue. Only 1 position in the alignment (180) contained as many as 5 different residues, all other
positions were limited to a maximum of 4.
fHBP consensus alignment and evaluation of residue conservation. Alignment of the consensus sequences of 97 subfamily B variants
),n p 14
Sequence Diversity of N. meningitidis fHBP • JID 2009:200 (1 August) • 385
(systematically collected strains and strains from South Africa). Trees were generated with Bionumerics with use of a categorical coefficient and are based
on 77 variable positions in subfamily A (A) and 101 variable positions in subfamily B (B). The size of each circle is proportional to the number of strains
carrying that variant. Each color in the circles represents a different MLST clonal complex. Groups based on fHBP protein sequence (connected by colored
background) were formed where the neighboring distance was ?5 amino acid differences, with formation of hypothetical types allowed. Connecting lines
are proportional to distance.
Minimum spanning trees generated based on variable amino acid positions in the protein alignment of fHBP variants in 1317 strains
(marked in green in figure 2), including the KDN insertion
after residue 67, that were 100% diagnostic for either N1 or
N2. A third form of the N terminal domain, N3 (found in only
1 strain [variant A57]), had the signature residues of N1 except
for the 3 most C-terminal signature residues, where it carried
the N2 signature Ile85, Arg86 and Gln87. In the C-terminal
domain, 8 residues were 100% diagnostic for either C1 or C2
in subfamily A (figure 2).
Network analysis of subfamily B did not yield obvious sub-
groups similar to that of subfamily A (figure 1C). However,
consideration of just the N-terminal domains (residues 1–139
) of all variants of both subfamilies (figure 1D) indicated
a close relationship between subfamily A and B variants. The
majority of subfamily B variants (84% of subfamily B strains
in the systematically collected set) possessed an N-terminal do-
main (N6) with all signature residues of the A subfamily N2
domain. Variation at the nonsignature residues resulted in the
independent branching of N2 and N6. This analysis also re-
vealed 7 subfamily B variants with N-terminal domains that,
like N3, appeared to be a result of recombination between an
N1- and N2-like domain between residues 79 and 85 (N4,
exemplified by B44), as well as 4 subfamily B variants that
could have arisen by recombination of an N1- and N2-like
domain between residues 55 and 63 (N5).
Minimum spanning trees constructed from the variable po-
sitions in the subfamily A and subfamily B protein alignments
are shown in figure 3. Complexes were allowed to form from
neighboring sequences containing up to 5 amino acid differ-
ences. In both trees, it is evident that a few “founder” sequences
and closely related variants (with 1–2 amino acid differences)
make up the majority of the fHBP population. In subfamily B,
31 variants (39 strains) were not included within any complex;
26 of these occurred in just 1 strain, and 5 others occurred in
that were not included within any complex.
Geographic distribution and frequency of fHBP variants.
Overall, 71% of the systematically collected strains and 70% of
the larger strain set were subfamily B variants, the remainder
being subfamily A variants. Within each individual country,the
percentage of strains carrying subfamily A variants rangedfrom
23% to 35%, and the percentage carrying subfamily B variants
ranged from 65% to 77% (table 2). The distribution of strains
in the United States is affected by an ongoing epidemic in
Oregon that is not representative of disease elsewhere in the
country. If Oregon is excluded from the analysis of strains from
the United States, the subfamily distribution for the other US
sites is 45% subfamily A and 55% subfamily B. In South Africa
(where all available isolates from 2005 were evaluated), sub-
386 • JID 2009:200 (1 August) • Murphy et al
Table 2.Frequency of the 10 Most Common fHBP Variants in the Systematically Collected Strain Set, by Geographic Location
Common variants, percentage of strains
ABB24 B16B44 B03 B09A22A19A12 A05 A07
England, Wales, Northern Ireland
Not including Oregona
aStatistics for the United States are impacted by the epidemic in Oregon .
for All fhbp Variants
Domain Assignments and Frequency
This table is available in its entirety in the online
version of the Journal of Infectious Diseases
family A strains were isolated more frequently than were sub-
family B strains (31 [57.4%] of 54 strains were subfamily A),
including a high frequency of a variant (A32) that was rare
Within the systematically collected strain pool, 143 fHBP
variants were identified among the 1263 strains; 92 variants
were found just once, and 20 variants were found in 2 strains
each. Only 18 variants were found in ?10 isolates; of these, 10
variants (5 each of subfamily A and subfamily B) accounted
for 79% of the systematically collected strain set (table 2). Fre-
quencies for all variants are given in table 3.
Although the 10 most common variants were almost always
among the most common in each country, their rank order
differed (table 2).IntheUnitedStates,B24accountedfor42.6%
of the isolates from 2000–2005. B24 is associated with the ST-
32 clonal complex, which has been responsible for long-term
hyperendemic disease in parts of the United States (particularly
in Oregon)  and many parts of Europe [17, 28]. Strains
carrying the B24 variant were common in Norway (35%) and
France (20%) but were less frequent in the HPA set(4%),where
the common subfamily B variants were B16 (23.5%) and B44
(22.0%). B44 was isolated less frequently in France, Norway,
and the Czech Republic and was isolated only once in the
United States. No major changes in variant frequency within
countries were observed over the 6 years surveyed (figure 4).
Relationship of MnB genetic lineage and fHBP sequence
Data for common MLST clonal complexes from the
systematically collected strains from the United States, France,
Norway, and the Czech Republic are shown in figure 5. Most
clonal complexes contained both subfamily A and subfamily B
variants. For example, the ST-41/44 clonal complex is highly
diverse and contains a wide diversity of fHBP variants, includ-
ing 107 subfamily A strains (13 different variants) and 129
subfamily B strains (17 variants). Two subfamily A (A12 and
A22) and 2 subfamily B variants (B03 and B16) accounted for
39% and 42%, respectively, of ST-41/44 strains. The major
exception was the ST-32 clonal complex, where 234 (87%) of
270 strains carried the B24 variant. However, despite the high
frequency of B24 within this lineage, 27 other fHBP variants
representing both subfamilies (10 subfamilyA and17subfamily
B) occurred in the remaining 36 strains. Analysis of PorA sub-
typing data for 500 strains revealed a similar diversity of fHBP
variants within most PorA types (data not shown).
Our molecular subtyping analysis confirms that there are 2
phylogenetic groups of fHBP sequences, subfamilies A and B.
Sequence variation is distributed throughout the length of the
proteins and is interspersed with highly conserved residues;
there are no hypervariableregionscharacteristicofothersurface
antigens, such as PorA. The pattern of natural sequence vari-
ation of fHBP suggests that evolution is operating under dif-
ferent constraints within each subfamily. Variation in both the
N- and C-terminal domains of subfamily A is relatively limited,
whereas subfamily B shows more evidence of recombination
in the N-terminaldomainandmoreallowablemutationoverall.
The division of fHBP variants into 2 sequence families par-
allels the functional immune reactivity of these proteins. The
111 residues common to and conserved between both subfam-
ilies map mainly to the interior core structure of fHBP, whereas
32 subfamily-specific residues lie mainly on one surface of the
structure . These residues are likely to be largely responsible
for the subfamily-specific immune response. Because mono-
valent fHBP vaccines elicit bactericidal antibodies that are
Sequence Diversity of N. meningitidis fHBP • JID 2009:200 (1 August) • 387
included. CZ, Czech Republic; FR, France; NO, Norway; UK, United Kingdom; US, United States.
fHBP variants by country and by year in the systematically collected strain set. The 6 most common variants for each subfamily are
tilocus sequence type was determined for 694 of the 727 strains from
the United States, France, Norway, and Czech Republic from the system-
atic collection. Data for 672 strains belonging to clonal complexes con-
taining at least 6 strains (461 subfamily B and 211 subfamily A) are
included in the figure; 22 strains in less common complexes are not
shown. NA indicates sequence types (STs) that were not assignable to
a clonal complex. Each color represents a different fHBP variant; the
most common variants are labeled.
Distribution of fHBP variants within clonal complexes. Mul-
largely subfamily specific [7, 29, 30], a bivalent vaccine is nec-
essary to provide coverage against both subfamilies. A bivalent
vaccine given to mice, rabbits, monkeys, or humans is capable
of generating serum bactericidal antibody responses against a
variety of heterologous strains [5–7, 14, 31].
In assessing the likelihood that a surface protein–based vac-
cine will recognize a majority of clinical isolates, it is imperative
to monitor the variation and expression of each individual
antigen rather than relying on epidemiological surrogates. In a
few instances, a particular fHBP variant may predict MLST (eg,
97% of B24 strains belong to the ST-32 clonal complex), but
the opposite is not necessarily true (figure 5). Earlier, less com-
prehensive studies [32, 33] showed that neitherPorAnorMLST
is predictive of fHBP variant. Similarly, this study showed that
even those complexes, such as ST-32, that contain a dominant
variant also contain examples of many other fHBP variants of
A number of N. meningitidis surface protein–based vaccines
have been considered for development, but the hurdles have
been considerable. These include a failure to generate bacte-
ricidal antibodies in humans (Neisserial surface protein A
[NspA]) , a high degree of sequence divergence in critical
epitopes (PorA and Tbp) [35–37], and an absence of the gene
in many isolates (N. meningitidis Adhesin A [NadA]) [32, 38].
However, fHBP is ubiquitous in N. meningitidis, is expressed
on the bacterial surface, generates bactericidal antibodies in
humans, and sequence divergence within each of the 2 sub-
families does not limit the development of a broad, subfamily-
specific bactericidal response [5–7, 14, 30, 39]. All of these
attributes make a bivalent recombinant fHBP vaccine a prom-
ising candidate for prevention of MnB disease.
The strains described in this study have also been evaluated
for surface expression of fHBP and serogroup B capsule .
They show a range of expression for both, with no correlation
between amount of capsule and level of surface expression of
fHBP. The best predictor for killing by anti-fHBP antibodies
appears to be the level of fHBP surface expression . All
1837 strains evaluated in this study contained the fHBP gene,
consistent with the important role that fHBP is believed to play
in survival of the organism in vivo. An interesting observation
was the detection of 3 strains with truncated or hybrid variants
of fHBP; additional studies are necessary to determine whether
these proteins are functional. Ongoing studies are aimed at
388 • JID 2009:200 (1 August) • Murphy et al
further refining the correlates of fHBP surface expression and
protection in relation to the selection of appropriate assay
strains for the evaluation of Phase III sera. The epidemiolog-
ically and genetically characterized collection of invasive MnB
strains described here forms the framework for this ongoing
We thank the Active Bacterial Core Surveillance sites for their efforts in
collecting the isolates from the United States, the Group for Enteric, Re-
spiratory, and Meningeal Disease Surveillance in South Africa (GERMS-
SA) for their efforts in collecting the isolates from South Africa, and all
the other laboratories responsible for isolating and documenting the pri-
mary case isolates. The technical assistance of Scott Sigethy and Diane
Pawlyk (Wyeth); Eva Hong, Corinne Ruckly, Dario Giorgini, and Jean-
Michel Alonso (Institut Pasteur); Vlasta Pavlikova, Renata Pospisilova,
Monika Heroldova and Romana Richterova (NIPH Prague); and Linda de
Gouveia and Kedibone Mothibeli (South Africa) is greatly appreciated.
1. Rosenstein NE, Perkins BA, Stephens DS, Popovic T, Hughes JM. Me-
ningococcal disease. N Engl J Med 2001;344:1378–88.
2. Trotter CL, Andrews NJ, Kaczmarski EB, Miller E, Ramsay ME. Ef-
fectiveness of meningococcal serogroup C conjugate vaccine 4 years
after introduction. Lancet 2004;364:365–7.
3. Finne J, Bitter-Suermann D, Goridis C, Finne U. An IgG monoclonal
antibody to group B meningococci cross-reacts with developmentally
regulated polysialic acid units of glycoproteins in neural and extra-
neural tissues. J Immunol 1987;138:4402–7.
4. Finne J, Leinonen M, Makela PH. Antigenic similarities between brain
components and bacteria causing meningitis. Implications for vaccine
development and pathogenesis. Lancet 1983;2:355–7.
5. Farley J, Fletcher L, Bernfield L, et al. Characterization, cloning and
expression of different subfamilies of the ORF 2086 gene fromNeisseria
meningitidis. In: Program and abstracts of the 13th International Path-
ogenic Neisseria Conference (Oslo, Norway). 2002.
6. Bernfield L, Fletcher L, Howell A, et al. Identification of a novel vaccine
candidate for group B Neisseria meningitidis. In: Program and abstracts
of the 13th International Pathogenic Neisseria Conference (Oslo, Nor-
meningitidis 2086 lipoprotein. Infect Immun 2004;72:2088–100.
8. Masignani V, Comanducci M, Giuliani MM, et al. Vaccination against
Neisseria meningitidis using three variants of the lipoprotein GNA1870.
J Exp Med 2003;197:789–99.
9. Madico G, Welsch JA, Lewis LA, et al. The meningococcal vaccine
candidate GNA1870 binds the complement regulatory protein factor
H and enhances serum resistance. J Immunol 2006;177:501–10.
10. Granoff DM, Welsch JA, Ram S. Binding of complement Factor H
(fH) to Neisseria meningitidis is specific for human fH and inhibits
complement activation by rat and rabbit sera. Infect Immun 2009;77:
11. Schneider MC, Exley RM, Chan H, et al. Functional significanceoffactor
H binding to Neisseria meningitidis. J Immunol 2006;176:7566–75.
12. Haralambous E, Dolly SO, Hibberd ML, et al. Factor H, a regulator
of complement activity, is a major determinant of meningococcal dis-
ease susceptibility in UK Caucasian patients. ScandJInfectDis2006;38:
13. Welsch JA, Ram S, Koeberling O, Granoff DM. Complement-depen-
dent synergistic bactericidal activity of antibodies against factor
H–binding protein, a sparsely distributed meningococcal vaccine an-
tigen. J Infect Dis 2008;197:1053–61.
14. Nissen MD, Marshall HS, Richmond P, et al. A randomized, placebo-
controlled, double-blind, phase 1 trial of ascending doses of menin-
gococcal group B rLP2086 vaccine. In: Program and abstracts of the
26th Annual Meeting of the European Society for Paediatric Infec-
tious Diseases (Graz, Austria). 2008.
15. von Gottberg A, du Plessis M, Cohen C, et al. Emergence of endemic
serogroup W135 meningococcal disease associated with a high mor-
tality rate in South Africa. Clin Infect Dis 2008;46:377–86.
16. Centers for Disease Control and Prevention. Active Bacterial Core
Surveillance. Available at: http://www.cdc.gov/ncidod/dbmd/abcs/
meth-surv-pop.htm. Accessed 8 June 2009.
17. Rouaud P, Perrocheau A, Taha MK, et al. Prolonged outbreak of B
meningococcal disease in the Seine-Maritime department, France, Jan-
uary 2003 to June 2005. Euro Surveillance 2006;11:178–81.
18. Jolley KA, Chan MS, Maiden MC. mlstdbNet—distributed multi-locus
sequence typing (MLST) databases. BMC Bioinformatics 2004;5:86.
19. Chenna R, Sugawara H, Koike T, et al. Multiple sequence alignmentwith
the Clustal series of programs. Nucleic Acids Res 2003;31:3497–500.
20. Saitou N, Nei M. The neighbor-joining method: a new method for
reconstructing phylogenetic trees. Mol Biol Evol 1987;4:406–25.
21. Kimura M. A simple method for estimating evolutionary rates of base
substitutions through comparative studies of nucleotide sequences. J
Mol Evol 1980;16:111–20.
22. Rice P, Longden I, Bleasby A. EMBOSS: the European Molecular Bi-
ology Open Software Suite. Trends Genet 2000;16:276–7.
23. Tamura K, Dudley J, NeiM,KumarS.MEGA4:MolecularEvolutionary
Genetics Analysis (MEGA) software version 4.0. Mol Biol Evol 2007;
24. Huson DH, Bryant D. Application of phylogenetic networks in evo-
lutionary studies. Mol Biol Evol 2006;23:254–67.
25. Holmes EC, Urwin R and Maiden MC. The influence ofrecombination
on the population structure and evolution of the human pathogen
Neisseria meningitidis. Mol Biol Evol 1999;16:741–9.
26. Mascioni A, Bentley BE, Camarda R, et al. Structural basis for the
immunogenic properties of the meningococcal vaccine candidate
LP2086. J Biol Chem 2009;284:8738–46.
27. Diermayer M, Hedberg K, Hoesly F, et al. Epidemic serogroup B me-
ningococcal disease in Oregon: the evolving epidemiology of the ET-
5 strain. JAMA 1999;281:1493–7.
28. Yazdankhah SP, Kriz P, Tzanakaki G, et al. Distribution of serogroups
and genotypes among disease-associated and carried isolates of Neis-
seria meningitidis from the Czech Republic, Greece, and Norway. J Clin
29. Pillai S, Howell A, Alexander K, et al. Outer membrane protein (OMP)
based vaccine for Neisseria meningitidis serogroup B. Vaccine 2005;23:
30. Jiang H-Q, Harris SL, Tan C, et al. Prediction of broadvaccinecoverage
for a bivalent rLP2086 based vaccine which elicits serum bactericidal
activity agains a diverse collection of serogroup B meningococci. In:
Program and abstract of the 16th International Pathogenic Neisseria
Conference (Rotterdam, The Netherlands). 2008.
31. Anderson AS, Xu R, Tan C, et al. Functional cross-reactive antibodies
are elicited by a Group B Neisseria meningitidis bivalent recombinant
lipidated LP2086 vaccine in Cynomolgus macaques [abstract P100].
In: Program and abstract of the 16th InternationalPathogenicNeisseria
Conference (Rotterdam, The Netherlands). 2008.
32. Beernink PT, Welsch JA, Harrison LH, Leipus A, Kaplan SL, Granoff
DM. Prevalence of factor H–binding protein variants andNadAamong
meningococcal group B isolates from the United States: implications
for the development of a multicomponent group B vaccine. J Infect
33. Jacobsson S, Thulin S, Molling P, et al. Sequence constancies and
variations in genes encoding three new meningococcal vaccine can-
didate antigens. Vaccine 2006;24:2161–8.
Sequence Diversity of N. meningitidis fHBP • JID 2009:200 (1 August) • 389
34. Halperin SA, Langley JM, Smith B, et al. Phase 1 first-in-human studies
of the reactogenicity and immunogenicity of a recombinant menin-
gococcal NspA vaccine in healthy adults. Vaccine 2007;25:450–7.
35. Rokbi B, Mignon M, Caugant DA , Quentin-Millet MJ. Heterogeneity
of tbpB, the transferrin-binding protein B gene, among serogroup B
Neisseria meningitidis strains of the ET-5 complex. Clin Diagn Lab
36. Tondella ML, Popovic T, Rosenstein NE, et al. Distribution of Neisseria
meningitidis serogroup B serosubtypes and serotypes circulating in the
United States. J Clin Microbiol 2000;38:3323–8.
37. Sacchi CT, Whitney AM, Popovic T, et al. Diversity and prevalence of
PorA types in Neisseria meningitidis serogroup B in the United States,
1992–1998. J Infect Dis 2000;182:1169–76.
38. Comanducci M, Bambini S, Brunelli B, et al. NadA, a novel vaccine
candidate ofNeisseria meningitidis. J Exp Med 2002;195:1445–54.
39. Koeberling O, Giuntini S, Seubert A, Granoff DM. Meningococcal
outer membrane vesicle vaccines derived from mutant strains engi-
neered to express Factor H binding proteins from antigenic variant
groups 1 and 2. Clin Vaccine Immunol 2009;16:156–62.
40. McNeil LK, Murphy E, Zhao X-J, et al. Detection of LP2086 on the
cell surface of Neisseria meningitidis and its accessibility in the presence
of serogroup B capsular polysaccharide. Vaccine 2009;27:3417–21.