Mapping antigenic diversity and strain specificity of mumps virus: a bioinformatics approach.
ABSTRACT Mumps is an acute infectious disease caused by mumps virus, a member of the family Paramyxoviridae. With the implementation of vaccination programs, mumps infection is under control. However, due to resurgence of mumps epidemics, there is a renewed interest in understanding the antigenic diversity of mumps virus. Hemagglutinin-neuraminidase (HN) is the major surface antigen and is known to elicit neutralizing antibodies. Mutational analysis of HN of wild-type and vaccine strains revealed that the hypervariable positions are distributed over the entire length with no detectable pattern. In the absence of experimentally derived 3D structure data, the structure of HN protein of mumps virus was predicted using homology modeling. Mutations mapped on the predicted structures were found to cluster on one of the surfaces. A predicted conformational epitope encompasses experimentally characterized epitopes suggesting that it is a major site for neutralization. These analyses provide rationale for strain specificity, antigenic diversity and varying efficacy of mumps vaccines.
- SourceAvailable from: PubMed Central[show abstract] [hide abstract]
ABSTRACT: In an attempt to account for antibody specificity and complementarity in terms of structure, human kappa-, human lambda-, and mouse kappa-Bence Jones proteins and light chains are considered as a single population and the variable and constant regions are compared using the sequence data available. Statistical criteria are used in evaluating each position in the sequence as to whether it is essentially invariant or group-specific, subgroup-specific, species-specific, etc. Examination of the invariant residues of the variable and constant regions confirms the existence of a large number of invariant glycines, no invariant valine, lysine, and histidine, and only one invariant leucine and alanine in the variable region, as compared with the absence of invariant glycines and presence of three each of invariant alanine, leucine, and valine and two each of invariant lysine and histidine in the constant region. The unique role of glycine in the variable region is emphasized. Hydrophobicity of the invariant residues of the two regions is also evaluated. A parameter termed variability is defined and plotted against the position for the 107 residues of the variable region. Three stretches of unusually high variability are noted at residues 24-34, 50-56, and 89-97; variations in length have been found in the first and third of these. It is hypothesized that positions 24-34 and 89-97 contain the complementarity-determining residues of the light chain-those which make contact with the antigenic determinant. The heavy chain also has been reported to have a similar region of very high variability which would also participate in forming the antibody-combining site. It is postulated that the information for site complementarity is contained in some extrachromosomal DNA such as an episome and is incorporated by insertion into the DNA of the structural genes for the variable region of short linear sequences of nucleotides. The advantages and disadvantages of this hypothesis are discussed.Journal of Experimental Medicine 08/1970; 132(2):211-50. · 13.21 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: Experimental protein structures often provide extensive insight into the mode and specificity of small molecule binding, and this information is useful for understanding protein function and for the design of drugs. We have performed an analysis of the reliability with which ligand-binding information can be deduced from computer model structures, as opposed to experimentally derived ones. Models produced as part of the CASP experiments are used. The accuracy of contacts between protein model atoms and experimentally determined ligand atom positions is the main criterion. Only comparative models are included (i.e., models based on a sequence relationship between the protein of interest and a known structure). We find that, as expected, contact errors increase with decreasing sequence identity used as a basis for modeling. Analysis of the causes of errors shows that sequence alignment errors between model and experimental template have the most deleterious effect. In general, good, but not perfect, insight into ligand binding can be obtained from models based on a sequence relationship, providing there are no alignment errors in the model. The results support a structural genomics strategy based on experimental sampling of structure space so that all protein domains can be modeled on the basis of 30% or higher sequence identity.Proteins Structure Function and Bioinformatics 07/2004; 55(4):942-61. · 3.34 Impact Factor
- [show abstract] [hide abstract]
ABSTRACT: The hemagglutinin-neuraminidase genes of the Kilham strain of mumps virus and three neutralization escape mutants (M11, M12 and M13) of this strain (Löve et al., 1985a) were sequenced using their genomes as template. The predicted amino acid sequences were compared. While one mutant had only one amino acid substitution the other two mutants had four and five respectively. A putative region for the epitope of the selected neutralizing monoclonal antibody was identified in a hydrophilic region encompassing amino acids 352-360, since the single amino acid substitution of one mutant occurred in this region and the other two mutants showed non-conserved amino acid changes in this part of the protein. The previously sequenced prototype strain RW, which lacks capacity to react with the selected neutralizing monoclonal antibody also has one non-conserved amino acid change in the region of the proposed neutralizing epitope. The three mutants showed different biological characteristics. These particular characteristics were therefore interpreted to be primarily associated with strain-specific amino acid changes outside the region of the presumed neutralizing epitope. The decrease in molecular weight in one mutant (M11) was shown to be due to a substitution in position 329 of an asparagine for an aspartic acid, leading to abolishment of a potential N-linked glycosylation site. In the other mutants, one substitution in position 239 of a lysine for a methionine was correlated with an increased neuraminidase activity of strain M12, while a substitution in position 360 of an arginine for a cysteine appeared to represent the most likely explanation for the reduced neurovirulence of strain M13.Virus Research 11/1990; 17(2):119-29. · 2.75 Impact Factor
Mapping antigenic diversity and strain specificity of mumps virus:
A bioinformatics approach
Urmila Kulkarni-Kalea,⁎, Janaki Ojhaa, G. Sunitha Manjaria, Deepti D. Deobagkarb,
Asha D. Mallyac, Rajeev M. Dhere
c, Subhash V. Kapre
aBioinformatics Centre, University of Pune, Pune 411007, India
bDepartment of Zoology, University of Pune, Pune 411007, India
cSerum Institute of India Ltd., 212/2, Hadapsar, Pune 411028, India
Received 13 July 2006; returned to author for revision 18 August 2006; accepted 15 September 2006
Available online 1 November 2006
Mumps is an acute infectious disease caused by mumps virus, a member of the family Paramyxoviridae. With the implementation of
vaccination programs, mumps infection is under control. However, due to resurgence of mumps epidemics, there is a renewed interest in
understanding the antigenic diversity of mumps virus. Hemagglutinin–neuraminidase (HN) is the major surface antigen and is known to elicit
neutralizing antibodies. Mutational analysis of HN of wild-type and vaccine strains revealed that the hypervariable positions are distributed over
the entire length with no detectable pattern. In the absence of experimentally derived 3D structure data, the structure of HN protein of mumps virus
was predicted using homology modeling. Mutations mapped on the predicted structures were found to cluster on one of the surfaces. A predicted
conformational epitope encompasses experimentally characterized epitopes suggesting that it is a major site for neutralization. These analyses
provide rationale for strain specificity, antigenic diversity and varying efficacy of mumps vaccines.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Mumps virus; Hemagglutinin–neuraminidase; HN; Wild and vaccine strain; Homology model; Sequential epitope; Conformational epitope;
Bioinformatics; Molecular immunology
Mumps is an acute infectious viral disease characterized by
enlargement of the parotid and salivary glands. Other
complications of mumps include permanent deafness, orchitis,
pancreatitis and aseptic meningitis (AM). The virus usually
spreads through respiratory droplets and humans are the only
natural host although non-human primates, rodents and other
species can be experimentally infected (Wolinsky, 1996).
Mumps virus is a member of the family Paramyxoviridae,
genus Rubulavirus (Rima et al., 1995). The virus is enveloped
and its 15.3-kb genome is non-segmented, single-stranded,
negative-sense RNA. It codes for seven proteins viz.,
nucleocapsid (N), phosphoprotein (P), matrix (M), fusion (F),
small hydrophobic (SH), hemagglutinin–neuraminidase (HN)
and large (L) proteins. Besides these, the phosphoprotein gene
(P) also codes for two more proteins, V and I (Paterson and
Lamb, 1990). HN is the major antigenic protein, known to elicit
neutralizing antibodies. The SH gene is known to be the most
variable part of the genome and phylogenetic reconstruction
studies of different isolates of mumps virus using SH genes
revealed the existence of 12 genotypes, designated A to L and 3
unnamed potential new genotypes (Jin et al., 2005; Tecle et al.,
Different genotypes have been shown to co-circulate (Afzal
et al., 1997a, 1997b; Wu et al., 1998; Tecle et al., 2001, 2002;
Takahashi et al., 2000) and their distribution may vary among
closely related regions within a country (Takahashi et al., 2000;
Tecle et al., 2001). Strains of mumps virus are also known to
exhibit varying degrees of neurovirulence (Merz and Wolinsky,
1981; Saito et al., 1996; Rubin et al., 1998, 2000; Rafiefard et
al., 2005; Sauder et al., 2006). However, the relative degree of
Virology 359 (2007) 436–446
⁎Corresponding author. Fax: +91 20 2569 0087.
E-mail addresses: firstname.lastname@example.org,
email@example.com (U. Kulkarni-Kale).
0042-6822/$ - see front matter © 2006 Elsevier Inc. All rights reserved.
their neurovirulence could not be determined mainly due to lack
of animal models. Preliminary results obtained using neonatal
rat (Rubin et al., 2000, 2005) and Marmoset monkey models
(Saika et al., 2004) for vaccine strains, though promising,
require further investigations.
Live attenuated mumps vaccines are available as monovalent
mumps, bivalent measles–mumps (MM) and trivalent measles–
mumps–rubella (MMR) vaccines. The strains viz., Jeryl Lynn
(JL), Leningrad-3 (L-3), L-Zagreb, Rubini and Urabe have been
in use all over the world since early 1980s and have been
reviewed extensively (Galazka et al., 1999; Furesz, 2002; Folb
et al., 2004; Ivancic et al., 2005). It has been reported that the JL
vaccine is a mixture of two strains, JL-2 (also called minor) and
JL-5 (also called major) (Afzal et al., 1993; Amexis et al.,
2002). Similarly,the Urabe vaccine strain has also been reported
as a mixture of strains (Brown et al., 1996). The Rubini strain
does not appear to provide long-term protection (Galazka et al.,
1999; Utz et al., 2004; Ong et al., 2005). While offering varying
degrees of protection (Ong et al., 2005), most of the vaccine
strains are known to cause adverse reactions such as AM (Flynn
and Mahon, 2003; Folb et al., 2004; Nagai et al., 2006). Another
challenge is co-circulation of wild-type virus SBL-1 while
mumps vaccine is in use (Orvell et al., 1997; Tecle et al., 1998).
Mumps infection is under control due to the implementation
of vaccination programs. However, there is a renewed interest to
understand the antigenic diversity of mumps virus because of
recent outbreaks of mumps epidemics (CDC, 2006). The
knowledge thus gained will play a decisive role in mumps
vaccine development (Amexis et al., 2001; Furesz, 2002). The
contribution of specific humoral response to mumps virus as a
defense factor has not been definitively explained (Pipkin et al.,
1999; Kacprzak-Bergman et al., 2001). However, HN protein is
a major target for humoral immune response in mumps virus
infection as it elicits neutralizing antibodies (Tanaka et al.,
1992; Cusi et al., 2001). The regions viz., 265–288, 329–340
and 352–360 of HN have been reported to evoke immune
response and are also responsible for virulence (Kovamees et
al., 1990; Orvell et al., 1997; Cusi et al., 2001). Furthermore, the
region 329–340 is shown to have the ability to induce
neutralizing antibodies not only to the attenuated virus strains
but also to wild-type strains (Cusi et al., 2001).
Nucleotide and protein sequences of HN from a number of
mumps virus strains/isolates are available in the public domain
repositories. Extensive analyses of these sequences helped in
identification of strain-specific variations and characterization
of mutants (Yates et al., 1996). Attempts have also been made to
correlate mutations with properties such as antigenicity and
neurovirulence (Sauder et al., 2006). It has been reported that
mutations in HN are distributed over the entire length and no
pattern could be detected using sequence data alone (Ivancic et
It is known that the patterns, which are apparently hidden at
sequence level, become evident when mapped onto structure.
Experimentally derived 3D structure data of HN protein of
mumps virus are not available. Therefore, an attempt has been
made to predict its 3D structure using knowledge-based
homology modeling approach. The structures of HN protein
of one wild-type (SBL-1) and three vaccine strains, namely JL2,
JL5 and L-Zagreb were predicted. The observed variations of
amino acids among the groups of vaccine and wild-type strains
were then mapped on the predicted structure. Sequential and
conformational epitopes were predicted and analyzed in the
context of observed mutations (Kulkarni-Kale et al., 2005;
Kolaskar and Kulkarni-Kale, 1999; Kolaskar and Tongaonkar,
Compilation of HN sequences
A total of sixty-four nucleotide sequences of HN from
various strains/isolates of mumps virus were retrieved from
GenBank. It must be mentioned that curation of data is an
important step in any Bioinformatics analyses. Although
information is available, it is scattered either in different entries
of the same database or among various databases. For example,
only a few sequence entries for HN in GenBank are annotated
with respect to genotype data. Since the genotyping of mumps
virus is carried out using SH gene, the sequence entries of SH
for a given strain/isolate were searched to retrieve the genotype
information. However, only 49 out of 64 strains could be
annotated with respect to genotype information using this
Multiple sequence alignment
The nucleotide and protein sequences of HN were aligned
using ClustalW (Chenna et al.,2003).As expected, though there
exists sequence similarity among the strains of mumps virus, a
few strain-specific variations are observed. Multiple sequence
the MSA) and 96% similarity (identity+favorable substitutions
that are marked with : and . in the MSA) among vaccine strains
(see Annexure V). These values varied up to 74% and 91%
respectively in wild-type strains (MSA data not shown). The
known motifs viz., leucine-zipper, neuraminidase (240-
NRKSCS-245) and receptor-binding site of hemagglutinin
(405-GAEGRV-410) are conserved among all 64 entries
(Jorgensen et al., 1987; Mirza et al., 1994; Lim et al., 2003).
Similarly, there are 9 potential N-linked glycosylation sites with
signature sequence N-X-S or N-X-T (Apweiler et al., 1999) at
positions: 12, 127, 284, 329, 400, 448, 464, 507 and 514. Of
these, glycosylation sites at positions 127, 284, 448, 507, and
514 are conserved in both vaccine and wild-type strains.
Mutation of N to D/S at position 12 results in loss of a potential
cytosolic glycosylation site in a few strains. This site was found
to be missing in the isolates of mumps virus from a vaccinated
population in Singapore (Lim et al., 2003). The 3rd amino acid
position in the glycosylation site 329 contains either, T or S.
However, at similar position in glycosylation site 400, I is found
in entries with GenBank accession numbers: AF448531,
AF448530, AF448527, AF448534 and AF448528, an unfavor-
able mutation as far as glycosylation is concerned. Such strains
may not get glycosylated and this may account for further
437U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
strain-specific antigenic variation, brought in due to post-
translational modifications. Similarly, the N at position 464 is
substituted by K/H, which would result in loss of this
glycosylation site. The mutation N464K is observed in wild-
type strains RW (genotype D: GenBank accession no.
M19933) and 871005 (genotype B: GenBank accession no.
AF314562), as well as in vaccine strains Smith-Kline Beecham
live-attenuated strain (genotype B: GenBank accession no.
AF314559) and JL2 (genotype A: GenBank accession no.
AY584604, AF345290). The mutation N464H is observed in
JL5 strains (GenBank accession no. AY584603 and X93179).
In addition to mutations in the glycosylation site, a few escape
mutants have been characterized by several workers world-
wide. These include positions 239, 266, 269, 329, 352, 354,
356 and 360, which are identified using SBL-1 and Kilham
strains (Orvell et al., 1997; Rafiefard et al., 2005). The MSA
data and the literature-curated data on known escape mutants
were subsequently used to map molecular determinants of
In the next phase, the sequences of all known wild-type and
vaccine strains were aligned separately. These MSA revealed
that there are 44 and 23 mutations respectively in wild-type and
vaccine strains. Of these, 15 sites were found to be common
between both the groups (Table 1). The variability index (VI)
calculated using the method of Wu and Kabat (1970) for wild-
type and vaccine strains are shown as Annexures I and II
respectively. VI gives an idea about the extent of variation at a
given position among a group of aligned sequences and is the
ratio of the number of different amino acids at a given position
and frequency of the most common amino acid at that position.
Eight sites show VI >4 among wild-type strains as compared
with only three sites in vaccine strains (Annexures I and II). The
eight variable sites of wild-type strains (12, 20, 25, 121, 266,
354, 385, 533), when mapped on the predicted structure,
revealed that three sites (12, 20, 25) are part of cytosolic region;
site 121 is not modeled; site 385 is buried and 3 sites (266, 354
and 533) are present on the surface of HN and are accessible for
interactions with antibodies. The positions 266 and 354 are of
significance since they are part of the experimentally character-
ized neutralizing epitope. Among the vaccine strains, only three
sites: 8, 80 and 121 are found to have variability index of 4 or
higher. Of these sites, 8 and 80 are part of cytosolic region; site
121 is found to be hypervariable in both vaccine and wild-type
strains and is a part of the stalk region. It must be mentioned that
3D structures are predicted only for the extracellular domain of
Prediction of 3D structure
Homology modeling is the method of choice for the
prediction of 3D structure of proteins when structure of at least
one orthologue is available and the sequence similarity
between the query and its orthologue is ≥30% (DeWeese-
Scott and Moult, 2004). The 3D structures of HN of one wild-
type (SBL-1) and three vaccine strains (viz. JL2, Jl5 and L-
Zagreb) were predicted using the said approach. The choice of
SBL-1, a wild-type virus, was based on two reasons: firstly,
SBL-1 is less neurovirulent than other wild-type strains; and
secondly it has been reported to co-circulate among the
populations vaccinated with JL vaccine (Orvell et al., 1997;
Tecle et al., 1998). The 3D structure data of HN of two
paramyxoviruses viz., parainfluenza virus (PIV) and Newcastle
disease virus (NDV) are available in Protein Data Bank (PDB).
The HN sequences of the four mumps virus strains mentioned
above were aligned with their orthologues in PDB to identify
template(s) for homology modeling. The sequence of HN of
mumps virus was found to have 35% and 27% similarity with
NDVand PIV respectively. On the basis of sequence similarity,
stereochemical and geometrical evaluation of 3D structures of
HN of NDV and PIV, the structure of HN of NDV was
identified as a template (PDB ID: 1E8U; Crennell et al., 2000;
Takimoto et al., 2000). The pairwise sequence alignment of
HN of mumps virus (SBL-1) with its template is shown in Fig.
1. The initial structure of HN of SBL-1 was then built by
assigning the coordinates from the aligned residues of
template. Care was taken to avoid assignment of initial
coordinates to Proline from non-Proline residues and to non-
Glycine from Glycine residues. Coordinates for the loop
regions were assigned by searching the database of loops
(Hobohm and Sander, 1994). Energy of the model is calculated
using amber all atom force field (Seibel et al., 1990) and
distance-dependent dielectric constant (4rij). This helps to
simulate the effect of the solvent in an implicit fashion. The
bond lengths, bond angles and omega values were checked
and corrected only if they were outside the acceptable range.
For bond length and bond angle, the acceptable range was
fixed as (standard bond length/bond angle ±3σ) and for
Analysis of mutations in HN protein of mumps virus
SitesMutations: in vaccine
strains, derived using
MSA of vaccine strains
Mutations: in wild-type
strains, derived using
MSA of wild-type strains
V135I, V212I, V218A,
V259I, D266A, T279I,
V287I, T288K, S295T,
S336L, Q354P, E356D,
S372N, S442Y, S462L,
N464K/H, E468K, V470I,
T473I, V474A, S490R,
N131S, I132V, D156E,
V212I, N240I, D251G,
V287I, S295T/A, K335R,
S336L, Y347C, S351L/P,
D356E/S, R363K, F370L,
N372S, V375I, A378V,
P397A, S398L, N399S,
T402I, L403M, L411I,
I435V/L, T438I, Q444L/P,
V447G, N464K, E468K,
T473I, A474V, P483S,
Note. Mutations only in the extracellular domain, i.e., residues 132–582 are
listed for both wild-type and vaccine strains. The mutations are reported in the
format ‘predominant residue–position-mutated residue’. Sites common in
vaccine and wild-type strains are underlined. Sites mapped on predicted
structure (see Fig. 2b) are shown in bold.
438 U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
Models of all four strains were built and optimized independ-
ently to eliminate bias. The structure of the monomer for every
strain was built initially followed by dimer assembly and
optimization. The order for optimization was: loops, SCRs, all
side-chains and all atoms.
Evaluation of predicted structures
The essential accuracy and correctness of the models
were evaluated using PROSTAT (module in Homology),
ProsaII (Sippl, 1993) and PROCHECK (Laskowski et al.,
1993). Evaluation of the models using more than one
method was essential in this particular case because the
sequence identity between the HN protein of mumps virus
and NDV is only 35%. The models were evaluated in terms
of stereochemical and geometrical parameters like bond
lengths, bond angles, torsion angles and packing environ-
ment and were found to satisfy the said criteria (Ramachan-
dran and Sasisekharan, 1968) (Table 2). It must be
mentioned that the accuracy of any model built using
homology modeling is limited not only by the accuracy of
the template but also sequence alignment, choice of initial
structures of loop regions and extent of optimization (Rost
and O'Donoghue, 1997).
Further objective check on the quality of models was
obtained using the program ProsaII (Sippl, 1993). The energy
graphs drawn using ProsaII display the energetic architecture of
protein folds as a function of amino acid sequence. The ProsaII
Z-score for the template and the four models are found to be in
the range typical for globular proteins (Table 2). The
PROCHECK G factor is the log-odds score based on the
observed distributions of various stereochemical parameters
(Laskowski et al., 1993). The over all G factor score for the
template and models lie in the range −0.5 to 0.5 which
indicates that all four structures are stereochemically acceptable
Structural description of HN of mumps virus
Predicted structure of HN of SBL-1 is shown in Fig. 2a.
As can be seen, the predominant secondary structure is β-
sheet. The fold of the protein is β-propeller. Every monomer
is a six-bladed β-propeller consisting of six 4-stranded β-
sheet motifs along with four helices. Superimposition of the
predicted structure of HN of all strains on the template
structure shows very high similarity at 3D structure level,
though sequence similarity is only 35%. The root–mean–
square (rms) deviations (Å) between the template and each of
the models for ∼428 structurally aligned residues are SBL-1:
0.63, Jl2: 0.67; JL5: 0.65 and L-Zagreb: 0.67, respectively.
Furthermore, all the four helices and majority of the strands
are conserved and show similar boundaries. However, some
variations in the length of the strands have been observed
when compared with the template structure. A few short
strands that are present in the template have disappeared in
the model. Similarly, a few strands are split into two short
strands. The rmsd among the models of HN ranged between
0.37 and 0.5 Å indicating that there exists an overall
structural similarity for the backbone atoms. Subsequent to
our homology modeling studies, the coordinates of SV-5, a
member of Rubulavirus genus, were available in PDB (Yuan
et al., 2005). Since SV-5 is evolutionarily closely related to
mumps virus as compared to NDV (chosen template), we
compared the crystal structures of HN of SV-5 (PDB: 1Z4V)
and NDV (PDB: 1E8U). The overall structural rmsd between
1Z4V and 1E8U (for a monomer) was found to be 0.94 Å
Evaluation of predicted structures of HN monomer of mumps virus
et al., 1993)
Sasisekharan, 1968) (%)
−9.37 99.4I127, F156,
Fig. 1. Pairwise sequence alignment of HN (monomer) of mumps virus (query:
SBL-1) and that of NDV (template: 1E8U; chain A). The structurally conserved
regions (SCRs) and loop regions are identified by superimposition of the
structures of HN proteins and boundaries of secondary structural elements of
NDV. SCRs are shown as highlighted text.
439U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
indicating a similar fold. The structural superimposition
between four models of HN of mumps virus (one wild-type
and three vaccine strains) built using 1E8U as a template and
1Z4V showed a deviation in the range of 0.85–0.88 Å. Thus,
comparison with orthologue further validated the predicted
The predicted structure of SBL-1 was compared with that of
vaccine strains to study the effect of mutations. The comparison
of SBL-1 and L-Zagreb structures shows that the mutations
P459R, V259I, A577T that are part of SCR regions have effect on
local region rather than on surrounding regions. However, in
case of N161H mutation, more effect is seen in surrounding
regions rather than in local regions as it is a part of splice site
connecting two SCRs. Some of the mutated residues are part of
predicted antigenic determinants, which may have implications
in antibody recognition.
Fig. 2. (a) Predicted structure of HN (dimer) of L-Zagreb. Kabsch and Sander secondary rendering is shown. The β-sheets form the major secondary structure and four
short α-helices are present in each chain. The strands, helices, turns and coils are shown in yellow, red, blue and green respectively. (b) Mutations mapped on the
solvent accessible surface of the predicted structures of HN (monomer) of SBL-1, JL5, JL2 and L-Zagreb. Mutations are colored according to predominance of a given
amino acid. Dark color indicates that the said amino acid is present in 3 of the four strains.
440U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
Mapping of epitopes
The regions, 265–288, 329–340 and 352–360, of HN are
known to be antigenic (Kovamees et al., 1990; Orvell et al.,
1997; Cusi et al., 2001). Mapping of these regions on the
predicted structure of mumps revealed that they are present on
the surface of HN and hence are available for interaction with
antibodies. The B cell epitopes were predicted using both,
sequence data and 3D models of HN. Some of the residues
predicted as antigenic using sequence-based approach are
buried in the core (Kolaskar and Tongaonkar, 1990). This is
due to the fact that sequence-based approaches use only the
amino acid propensity data for calculation of solvent accessi-
bility. The epitopes were also predicted using CEP server, a
structure-based epitope prediction program (Kulkarni-Kale et
al., 2005; Kolaskar and Kulkarni-Kale, 1999) that predicts both,
conformational and sequential epitopes (see Supplementary
material: Annexures III and IV) using accessibility of amino
acids in an explicit manner. The conformational epitopes or
probable antibody-binding sites are composed of multiple
individual sequential epitopes that are brought together due to
protein folding (Van Regenmortel, 1998). The accuracy of the
Conformational Epitope Prediction algorithm was found to be
76% when evaluated using the co-crystal structures of antigen–
antibody complexes deposited in PDB. The algorithm was also
found to correctly predict one of the sequential epitopes of the
hyaluronidase, a major allergen of bee venom that has been
characterized using the Fab fragment of a mouse monoclonal
IgG antibody. Out of 14 residues predicted by CEP server, 9
sequential residues are involved in Fab binding (Zora Housley,
personal communication). The CEP server predicted 20
antigenic determinants (see Supplementary material: Annexures
III and IV) that form 16 conformational epitopes and 3
sequential epitopes. The predicted B cell epitopes, 261–266,
269–272, 284–296, 327–331 and 334–363, overlap with the
experimentally characterized epitopes of mumps virus. Further-
more, one of the predicted epitope (CE5) comprising of regions
261–266 and 269–272, has already been validated to be a
neutralizing epitope (Kovamees et al., 1990; Orvell et al., 1997).
Similarly, CE4 is made up of regions 261–266, 269–272, 199–
207 and 220–240 suggesting that these regions form a major
antibody-binding site, capable of being recognized by multiple
antibodies. Antibodies with overlapping binding sites have been
well characterized for lysozyme (Davies and Cohen, 1996;
Smith-Gill, 1996). The region 213–372 has been reported to
induce hemagglutination-inhibiting and mumps virus neutraliz-
ing antibodies (Cusi et al., 2001). As can be seen from the
supplementary material (Annexures III and IV), CEP predicts
six epitopes in the region 213–372. It has also been reported
that the antibodies raised against the synthetic peptide 329-
NSTLGVKSAREF-340 neutralized the mumps wild-type virus.
However, these antibodies failed to neutralize the Urabe Am-9
vaccine strain (Cusi et al., 2001). It must be mentioned that
Urabe AM-9 was found to have K335E mutation, which would
require charge reversal in the CDR regions of respective
antibodies. The lower efficacy of Urabe AM-9 could be
attributed to such mutations (Ong et al., 2005). Two epitopes
327–331 and 334–359 that encompass the peptide 329-
NSTLGVSAREF-340 are predicted by the CEP server
(Annexure IV). It was observed that the region 334–359
extends up to 363 when predicted for JL2, JL5 and SBL-1 (data
Hemagglutinin–neuraminidase (HN) protein has been iden-
tified as a major target for the humoral immune response in
mumps virus infection (Cusi et al., 2001). Nucleotide sequences
of HN of 64 strains/isolates of mumps virus are available in
GenBank release 146 (Feb. 2005). Most of these were
sequenced to characterize the outbreaks in the vaccinated
populations and to identify the spread and circulation of
genotypes (Tecle et al., 2001; Lim et al., 2003).
The complete genomes of mumps virus have also been
sequenced and are available in VirGen database (Kulkarni-Kale
et al., 2004) of which 9 are vaccine strains and 12 are wild-type
strains. The complete genome of the L-Zagreb mumps
vaccine strain has been sequenced recently (Ivancic et al.,
2005). Comparison of the nucleotide and protein sequences of
L-Zagreb with other strains revealed that although the
functional regions of HN, V and L proteins are conserved,
there are substantial variations observed in the known antigenic
regions (Ivancic et al., 2005). Furthermore, no molecular
pattern was identified which can serve as a distinction marker
between virulent and attenuated strains (Ivancic et al., 2005).
The structures of HN of three viruses, viz., NDV, PIVand SV-5
that belong to the family Paramyxoviridae have been solved
(Crennell et al., 2000; Takimoto et al., 2002; Lawrence et al.,
2004; Yuan et al., 2005). The knowledge gained from these
structures has been used to understand the mechanism of viral
membrane fusion (Zaitsev et al., 2004; Lamb et al., 2006). HN
has three functions, it recognizes the sialic acid containing
receptors on the cell surface; promotes the fusion activity of F
protein, viral penetration through the cell surface and removes
the sialic acid from progeny virus particles to prevent self-
agglutination (Crennell et al., 2000; Takimoto et al., 2002;
Lamb et al., 2006). Due to its multi-functional nature, HN
molecule is a target for development of antiviral drugs
(Crennell et al., 2000). Similarly, due to its ability to elicit
neutralizing antibodies, HN is a preferred candidate for vaccine
Sequence- and structure-based analyses of HN protein of
mumps virus revealed that although there exists very high
sequence similarity both at the nucleotide and protein level,
various strains have acquired certain mutations which confer
strain-specific properties. It is known that phenotypic properties
such as antigenicity, immunogenicity and neurovirulence are
the result of ‘spatio-temporal’ hierarchical processes. Exact
mapping of the same at molecular level is difficult due to the
fact that there exists complexity in terms of intermolecular
interactions which may be discrete or continuous (Flower et al.,
2003). Nevertheless, analyses of sequences and structures can
help in providing the rationale to explain the phenotypic
properties in related organisms.
441U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
The MSA revealed the positions of HN protein that are
under selection pressure. There are about 23 and 44
‘informative and variable’ sites respectively in vaccine and
wild-type strains of mumps virus and both the groups share
15 such sites (see Table 2). Similarly a few ‘singleton
variable’ sites are observed in both the groups. For example,
T265A, S368R and R459P mutations are unique to SBL-1
and could contribute towards its strain specificity. From the
plots of variability index, it is interesting to note that the
wild-type strains have a large number of sites that are
subjected to selection pressure and a wide variety of
substitutions are observed among them. Further strain
specificity is conferred by post-translational modifications.
The attachment of the carbohydrate moiety is known to play
an important role in the antigen–antibody interactions at
molecular level (Petrescu et al., 2004) and variation of the
immune response can be correlated to the extent of
glycosylation (Lisowska, 2002; Palomo et al., 2000; Cole et
al., 2004). The mutation at position 464 from N to K and H
in the JL2 and JL5 can result in loss of potential
glycosylation in these vaccine strains. Absence of carbohy-
drate moiety will alter the structural and chemical surface in
these vaccine strains and contribute to its efficacy (Ong et al.,
2005). Since this site is surrounded by the known neutralizing
epitopes (see Fig. 2b), it can be hypothesized that the
antibodies raised against these vaccines may not be able to
neutralize the genotypes that are glycosylated at this position.
Furthermore, this particular glycosylation site is flanked by
mutations specific to vaccine strains viz., JL, JL-2, JL-5,
Rubini and Sipar-02, indicating that this site may play a
deterministic role in molecular recognition of antigen and
offers a probable explanation for the observation that SBL-1
strain remains endemic in Europe (Orvell et al., 1997; Tecle
et al., 1998; CDC, 2006).
MSA of 15 vaccine strains is shown as Annexure V. The
variations observed have been marked by taking into account
the amino acid predominant at the corresponding position in the
MSA of wild-type strains. All the vaccine strains have a set of
unique variations that may be responsible for attenuation.
However, a few vaccine strains deviate minimally from the wild
type while acquiring attenuation. For instance, out of the forty-
seven mutations only four in L-Zagreb are such that the given
amino acid is not the preferred amino acid in all wild-type
strains. On the contrary, two Rubini vaccine strains show 18 and
19 such variations where the amino acid at the respective
position is not the preferred one among the wild type (see
Annexure V). As seen from Annexure V, the experimentally
characterized epitope 329–340 is conserved among the vaccine
strains. However, variations were observed in a few vaccine
strains in epitope 265–288. Similarly, two sites vary in Rubini,
JL-2 and JL-5 in the epitope 352–360. Apart from the above
mentioned experimentally characterized epitopes, MSA also
revealed the existence of a hypervariable region 462–474.
In addition to MSA of vaccine strains, pairwise
alignments of 15 vaccine strains with 9 representative
strains of genotypes A to E and G to J (as listed by Jin et
al., 2005) were carried out. In the absence of HN sequence
for Urb/Jap67 strain, 871005 (GenBank: AF314562) was
selected to represent genotype B. No HN sequence was
available for the representative strain or any other strain of
genotype F. For genotypes K and L only partial sequence data
for HN were available and hence were not included in the
analysis. Global alignments were carried out by needle
program of EMBOSS package (Rice et al., 2000). Annexure
VI shows %identity, %similarity and %difference. %Differ-
ence is calculated as (100−%similarity). Table 3 shows the
%difference between 15 vaccine strains and 10 representa-
tive sequences. Minimum deviation was observed between
vaccine strain and the representative strain of the genotype
to which it belongs. For example, SkBv, Rub1v, Sipar02
and Urabe deviate only by 0, 0.2, 0.3 and 0.3 respectively.
The range of minimum and maximum deviation varies from
0.9–1.9 for L-Zagreb to 0.2–2.9 for Rub1v. These observa-
tions provide a rationale for the varying specificities of the
vaccine strains for different genotypes.
Comparison of 3D structures of four strains showed
structural similarity with the HN of NDV indicating that the
fold and the function of the mumps HN is conserved despite
65% sequence variation. Overall structural similarity was also
observed among four strains and a few structural deviations
were found only in the regions of mutations, as expected. It
must be mentioned that none of the observed mutations are
part of dimerization interface, although mutations at dimer-
ization interface have been observed among members of the
genus Rubulavirus. There are 24 mutations among the SBL-1
and three vaccine strains viz., L-Zagreb, JL2 and JL5 (Fig.
2b). Of these, 16 (259, 265, 266, 287, 288, 354, 356, 368,
372, 442, 462, 464, 468, 470, 473, 474) are present on one
Sequence variation (100−%Similarity) between vaccine and wild-type strains
computed using pairwise global alignments of HN protein of mumps virus
Note. representative wild-type strains of various genotypes as listed by Jin et al.
(2005) have been used. For more details refer to Annexure V.
442 U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
surface (front view) of HN whereas only 8 (135, 161, 218,
279, 336, 459, 490) are present on the other surface (back
view: not shown). Of the 24 mutations, twenty-three are
accessible (%accessibility >25%) and only one at position
577 is buried in the core. Twelve of the sixteen mutations
present on one surface are part of experimentally character-
ized epitopes. Furthermore, these 12 mutations are localized
and predicted to be part of the conformational epitopes
indicating that this region is a major site for antibody
recognition as well as neutralization. Mutations are colored
according to predominance of a given amino acid (Fig. 2b).
Dark color indicates that the said amino acid is present in
three of the four strains. The mutations, A265T, N266D,
R368S, K468E, are shown in pink (SBL-1) and magenta
(vaccine strains). Residues shown in magenta are identical in
the vaccine strains viz., L-Zagreb, JL5, and JL2. The
mutations, A265T and R368S, are specific only to SBL-1.
All vaccine as well as wild-type strains except SBL-1 have T
and S at the respective positions. Similarly, mutation N266D
is specific to SBL-1 as all other wild-type strains and all other
vaccine strains except Rubini have D at this position. In case
of the mutation K468E, E is a preferred amino acid at this
position in all wild-type and vaccine strains except SBL-1,
Edinburgh-4/3 (wild type) and Rubini (vaccine strain). The
fact that SBL-1 co-circulates in JL vaccinated population may
be attributed to the mutations unique to SBL-1 (Orvell et al.,
1997; Tecle et al., 1998).
Mutations, V259I, I287V, P354Q, E356D, S372N, V474A,
are shown in blue (SBL-1, JL2, JL5) and cyan (L-Zagreb).
Substitutions of V with I/A may not bring about a major change
as they are known to be conformationally similar (Kolaskar and
Kulkarni-Kale, 1992) and hence are favorable mutations. The
mutation V259I is unique to L-Zagreb and may contribute
towards the specificity of this vaccine strain. All wild-type and
vaccine strains have V at this position. Similarly, I287V
mutation is common at this position in all wild-type and vaccine
strains. Majority of wild and vaccine strains except JL-2, JL-5
and Rubini are found to have V at this position. Similarly, the
mutation E356D does not alter the biochemical property of the
respective molecules. All wild-type strains except SBL-1,
Kilham and Edinburgh-4/5 and all vaccine strains except JL-2,
JL-5 and Rubini have E at this position. The mutation, P354Q
is observed in JL-5 and L-Zagreb. All vaccine as well as wild-
type strains have Q where as vaccine strains viz., JL-2 and
Rubini and wild-type strains viz., SBL-1, Edinburgh-4/5 have
P; Kilham strain has H and Zagreb cro69 strain has K at this
position. Similarly S is a preferred residue at position 372 in the
wild-type and vaccine strains including L-Zagreb. However,
the wild-type strains viz., SBL-1, Edinburgh-4/5 and the
vaccine strains viz., JL-2, JL-5, JL minor, major component
have N (S372N). These analyses indicate that SBL-1 is very
different from the vaccine strains that are in use. The vaccine
strains JL-2, JL-5 and Rubini also have unique mutations
The mutations V470I, N464K/H, S462L and T288K are
shown in orange (SBL-1, JL5 and L-Zagreb) and yellow (JL2,
JL5). All the above mutations are found in JL2 whereas JL5
has retained some residues of the wild type at a few positions
but acquired mutations in others. The site N464 in SBL-1 is
replaced by two different positively charged residues viz., K in
JL2 and H in JL5 indicating that this site is under selection
pressure in these vaccine strains. The vaccine strain of
Smith-Kline Beecham live-attenuated Urabe AM9 vaccine
(GenBank: AF314559) also has K at the equivalent posi-
tion. L-Zagreb, like most of other vaccine and wild-type strains,
is found to have N at this position (Annexure V). The wild-type
strain, RW of genotype D (GenBank: M19933) has K at this
position. Mutations T473I, S442Y are shown in dark green
(SBL-1, L-Zagreb, JL2) and light green (JL5). Hence these
mutations are unique to JL5 and may contribute to its strain
Mutations identified using MSA of vaccine strains
(Annexure V) when mapped on the predicted 3D structure
revealed that the residues of hypervariable region 462–474
are found to cluster on one of the surfaces of HN.
Furthermore, the residues 462, 464, 468, 470, 473 and 474
are solvent accessible and are in close proximity to the known
epitopes. This prompts us to propose that the region 462–474
is an additional antigenic site capable of playing a
deterministic role in neutralization (see Fig. 2b).
Hence, both sequence- and structure-based analyses explain
the fact that L-Zagreb has its own set of unique mutations
acquired during the process of attenuation. However, it has been
observed that the residues spanning known and predicted
neutralizing epitopes are conserved in L-Zagreb with reference
to wild-type strains, thereby rendering L-Zagreb the capacity to
elicit strong and avid neutralizing antibodies. Thus, these
studies explain the observed antigenic diversity and strain
specificity of mumps virus apart from providing an insight into
underlying molecular mechanisms responsible for observed
variations in the efficacy of mumps vaccines.
Materials and methods
Compilation of data sets
Nucleotide and protein sequences of HN and SH genes of
mumps virus were retrieved from GenBank (Benson et al.,
2006) and GenPept (Wheeler et al., 2006) databases respec-
tively. The 3D coordinates of orthologous proteins, which were
used as templates in homology modeling, were obtained from
PDB (Berman et al., 2000). The data on known antigenic
regions and escape mutants were compiled from the published
Pairwise global sequence alignments were carried out using
the needle program of EMBOSS package (Rice et al., 2000).
Multiple sequence alignments (MSA) were carried out using
ClustalW (Chenna et al., 2003). Variability plots were obtained
by a program developed in-house, which implements Wu and
Kabat (1970) index. B cell epitopes of HN were predicted with
the program ANTIGEN (Kolaskar and Tongaonkar, 1990) and
443 U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
Conformational Epitope Prediction (CEP) Server (Kulkarni-
Kale et al., 2005; Kolaskar and Kulkarni-Kale, 1999).
Three-dimensional structure of HN protein of mumps virus
was predicted using knowledge-based homology modeling
approach described previously (Kolaskar and Kulkarni-Kale,
1999; Kulkarni-Kale and Kolaskar, 2003). The Homology,
Discover and Biopolymer modules of the InsightII suite of
programs (version 2000) were used for this purpose (Accelrys
The steps specific to the modeling of HN protein of mumps
virus are as follows:
▪ Identification of strains for prediction of structure: four
strains of mumps virus were identified for molecular
modeling studies viz., one wild-type strain SBL-1 (genotype
A; GenBank accession number: M55065) and three vaccine
strains, namely L-Zagreb, JL5 and JL2 (sequenced at the
Serum Institute of India, Pune, India and deposited in
GenBank with accession numbers AY583323, AY584603
and AY584604 respectively).
▪ Identification of templates: the orthologues with known
structures were identified using the program BLAST
(Altschul et al., 1997) against PDB database. Crystal
structures of HN protein of two members of Paramyxoviri-
dae viz., human parainfluenza virus (PIV) PDB-ID: 1V3B,
1V3C, 1V3D, 1V3E and Newcastle Disease virus (NDV)
PDB-ID: 1E8U, 1E8V, 1E8T, 1USR, 1USX were short-
▪ Template evaluation and prioritization: the template struc-
tures were ranked according to resolution. A few structures
are complexes of HN with a natural ligand and/or with small
molecules. Templates were evaluated for multiple criteria
such as geometry, stereochemistry and occupancy of
dihedral angles in the allowed regions of Ramachandran
plot. Structural alignment of templates (1E8U, 1E8V, 1E8T,
1USR, 1USX) was carried out to identify the structurally
conserved and variable regions. The overall rmsd was found
to be 0.3 Å. On the basis of the sequence similarity and the
template evaluation studies, HN of NDV (PDB: 1E8U;
Crennell et al., 2000; Takimoto et al., 2000) was chosen as a
▪ Identification of SCRs and loop regions: the structurally
conserved regions (SCRs) and loops were identified using
structural alignment of templates and multiple sequence
alignment of HN protein of members of Rubulavirus genus.
Care was taken not to allow breaks in the known secondary
▪ Optimization of predicted structures: the amber all atom
force field (Seibel et al., 1990) with distance-dependent
dielectric constant was used to predict the 3D structures of
HN protein. Optimization of structure was carried out using
the Steepest Descents (SD) and Conjugate Gradient (CG)
minimization methods till the average rms derivative criteria
reached 0.01 and 0.001 kcal/mol/Å, respectively.
▪ Evaluation of models: the essential accuracy and correctness
of the models were evaluated using various methods viz.,
PROSTAT (module in Homology), ProsaII (Sippl, 1993) and
PROCHECK (Laskowski et al., 1993).
Perl scripts and useful comments by Shriram Bhosle are
gratefully acknowledged. Dr. S.R.R. Reddy is profusely
thanked for his help in improving the readability of the
manuscript. Thanks to the editor and the anonymous reviewers
for their constructive and enjoyable criticism. This work was
supported by a contract as research grant to UKK by the Serum
Institute of India Research Foundation. Molecular modeling
facility funded by the Department of Biotechnology, Govern-
ment of India at the Bioinformatics Centre, University of Pune,
is acknowledged gratefully.
Appendix A. Supplementary data
Supplementary data associated with this article can be found,
in the online version, at doi:10.1016/j.virol.2006.09.040.
Afzal, M.A., Pickford, A.R., Forsey, T., Heath, A.B., Minor, P.D., 1993. The
Jeryl Lynn vaccine strain of mumps virus is a mixture of two distinct
isolates. J. Gen. Virol. 74, 917–920.
Afzal, M.A., Buchanan, J., Dias, J.A., Cordeiro, M., Bentley, M.L., Shorrock,
C.A., Minor, P.D., 1997a. RT-PCR based diagnosis and molecular
characterisation of mumps viruses derived from clinical specimens collected
during the 1996 mumps outbreak in Portugal. J. Med. Virol. 52, 349–353.
Afzal, M.A., Buchanan, J., Heath, A.B., Minor, P.D., 1997b. Clustering of
mumps virus isolates by SH gene sequence only partially reflects
geographical origin. Arch. Virol. 142, 227–238.
Altschul, S.F., Madden, T.L., Schaffer, A.A., Zhang, J., Zhang, Z., Miller,
W., Lipman, D.J., 1997. Gapped BLAST and PSI-BLAST: a new
generation of protein database search programs. Nucleic Acids Res. 25,
Amexis, G., Fineschi, N., Chumakov, K., 2001. Correlation of genetic
variability with safety of mumps vaccine Urabe AM9 strain. Virology
Amexis, G., Rubin, S., Chizhikov, V., Pelloquin, F., Carbone, K., Chumakov,
K., 2002. Sequence diversity of Jeryl Lynn strain of mumps virus:
quantitative mutant analysis for vaccine quality control. Virology 300,
Apweiler, R., Hermjakob, H., Sharon, N., 1999. On the frequency of protein
glycosylation, as deduced from analysis of the SWISS-PROT database.
Biochim. Biophys. Acta 1473, 4–8.
Benson, D.A., Karsch-Mizrachi, I., Lipman, D.J., Ostell, J., Wheeler, D.L.,
2006. GenBank. Nucleic Acids Res. 34, D16–D20 (Database issue).
Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H.,
Shindyalov, I.N., Bourne, P.E., 2000. The protein data bank. Nucleic Acids
Res. 28, 235–242.
Brown, E.G., Dimock, K., Wright, K.E., 1996. The Urabe AM9 mumps vaccine
is a mixture of viruses differing at amino acid 335 of the hemagglutinin–
neuraminidase gene with one form associated with disease. J. Infect Dis.
CDC, 2006. Update: multistate outbreak of mumps – United States, January 1 –
May 2, 2006. MMWR 55, 559–563.
Chenna, R., Sugawara, H., Koike, T., Lopez, R., Gibson, T.J., Higgins, D.G.,
Thompson, J.D., 2003. Multiple sequence alignment with the Clustal series
of programs. Nucleic Acids Res. 31, 3497–3500.
444U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
Cole, K.S., Steckbeck, J.D., Rowles, J.L., Desrosiers, R.C., Montelaro, R.C.,
2004. Removal of N-linked glycosylation sites in the V1 region of simian
immunodeficiency virus gp120 results in redirection of B-cell responses to
V3. J. Virol. 78, 1525–1539.
Crennell, S., Takimoto, T., Portner, A., Taylor, G., 2000. Crystal structure of the
multifunctional paramyxovirus hemagglutinin–neuraminidase. Nat. Struct.
Biol. 7, 1068–1074.
Cusi, M.G., Fischer, S., Sedlmeier, R., Valassina, M., Valensin, P.E., Donati, M.,
Neubert, W.J., 2001. Localization of a new neutralizing epitope on the
mumps virus hemagglutinin–neuraminidase protein. Virus Res. 74,
Davies, D.R., Cohen, G.H., 1996. Interactions of protein antigens with
antibodies. Proc. Natl. Acad. Sci. U.S.A. 93, 7–12.
DeWeese-Scott, C., Moult, J., 2004. Molecular modeling of protein function
regions. Proteins 55, 942–961.
Flower, D.R., McSparron, H., Blythe, M.J., Zygouri, C., Taylor, D., Guan, P.,
Wan, S., Coveney, P.V., Walshe, V., Borrow, P., Doytchinova, I.A., 2003.
Computational vaccinology: quantitative approaches. Novartis Found.
Symp. 254, 102–120.
Flynn, M., Mahon, B.P., 2003. Cell mediated and humoral immune responses to
mumps virus: recent developments. In: Pandalai, S.G. (Ed.), Recent
Research Developments in Virology. Research Signpost, Trivandrum,
India, pp. 97–115.
Folb,P.I., Bernatowska, E., Chen, R.,Clemens,J.,Dodoo,A.N.,Ellenberg,S.S.,
Farrington, C.P., John, T.J., Lambert, P.H., Macdonald, N.E., Miller, E.,
Salisbury, D., Schmitt, H.J., Siegrist, C.A., Wimalaratne, O., 2004. A global
perspective on vaccine safety and public health: the Global Advisory
Committee on Vaccine Safety. Am. J. Public Health 94, 1926–1931.
Furesz, J., 2002. Safety of live mumps virus vaccines. J. Med. Virol. 67,
Galazka, A.M., Robertson, S.E., Kraigher, A., 1999. Mumps and mumps
vaccine: a global review. Bull. World Health Organ 77, 3–14.
Hobohm, U., Sander, C., 1994. Enlarged representation set of protein structures.
Protein Sci. 3, 522–524.
Ivancic, J., Gulija, T.K., Forcic, D., Baricevic, M., Jug, R., Mesko-Prejac, M.,
Mazuran, R., 2005. Genetic characterization of L-Zagreb mumps vaccine
strain. Virus Res. 109, 95–105.
Jin, L., Rima, B., Brown, D., Orvell, C., Tecle, T., Afzal, M., Uchida, K.,
Nakayama, T., Song, J.W., Kang, C., Rota, P.A., Xu, W., Featherstone, D.,
2005. Proposal for genetic characterisation of wild-type mumps strains:
preliminary standardisation of the nomenclature. Arch. Virol. 150,
Jorgensen, E.D., Collins, P.L., Lomedico, P.T., 1987. Cloning and nucleotide
sequence of Newcastle disease virus hemagglutinin–neuraminidase mRNA:
identification of a putative sialic acid binding site. Virology 156, 12–24.
Kacprzak-Bergman, I., Zaleska, I., Jasonek, J., 2001. Specific humoral
immunity in mumps meningitis in children. Med. Sci. Monit. 7, 977–981.
Kolaskar, A.S., Kulkarni-Kale, U., 1992. Sequence alignment approach to pick
up conformationally similar protein fragments. J. Mol. Biol. 223,
Kolaskar, A.S., Kulkarni-Kale, U., 1999. Prediction of three-dimensional
structure and mapping of conformational epitopes of envelope glycoprotein
of Japanese encephalitis virus. Virology 261, 31–42.
Kolaskar, A.S., Tongaonkar, P.C., 1990. A semi-empirical method for prediction
of antigenic determinants on protein antigens. FEBS Lett. 276, 172–174.
Kovamees, J., Rydbeck, R., Orvell, C., Norrby, E., 1990. Hemagglutinin–
neuraminidase (HN) amino acid alterations in neutralization escape mutants
of Kilham mumps virus. Virus Res. 17, 119–129.
Kulkarni-Kale, U., Kolaskar, A.S., 2003. Prediction of 3D structure of envelope
glycoprotein of Sri Lanka strain of Japanese encephalitis virus. In: Chen,
Yi-Ping Phoebe (Ed.), Conferences in Research and Practice in Information
Technology, 19, pp. 87–96.
Kulkarni-Kale, U., Bhosle, S., Manjari, G.S., Kolaskar, A.S., 2004. VirGen: a
comprehensive viral genome resource. Nucleic Acids Res. 32, D289–D292.
Kulkarni-Kale, U., Bhosle, S., Kolaskar, A.S., 2005. CEP: a conformational
epitope prediction server. Nucleic Acids Res. 33, W168–W171.
Lamb, R.A., Paterson, R.G., Jardetzky, T.S., 2006. Paramyxovirus membrane
fusion: lessons from the F and HN atomic structures. Virology 344, 30–37.
Laskowski, R.A., MacArthur, M.W., Moss, D.S., Thornton, J.M., 1993.
Procheck: a program to check the stereochemical quality of protein
structure. J. Appl. Crystallogr. 26, 283–291.
Lawrence, M.C., Borg, N.A., Streltsov, V.A., Pilling, P.A., Epa, V.C., Varghese,
J.N., McKimm-Breschkin, J.L., Colman, P.M., 2004. Structure of the
haemagglutinin–neuraminidase from human parainfluenza virus type III.
J. Mol. Biol. 335, 1343–1357.
Lim, C.S., Chan, K.P., Goh, K.T., Chow, V.T., 2003. Hemagglutinin–
neuraminidase sequence and phylogenetic analyses of mumps virus isolates
from a vaccinated population in Singapore. J. Med. Virol. 70, 287–292.
Lisowska, E., 2002. The role of glycosylation in protein antigenic properties.
Cell. Mol. Life Sci. 59, 445–455.
Merz, D.C., Wolinsky, J.S., 1981. Biochemical features of mumps virus
neuraminidases and their relationship with pathogenicity. Virology 114,
Mirza, A.M., Deng, R., Iorio, R.M., 1994. Site-directed mutagenesis of a
conserved hexapeptide in the paramyxovirus hemagglutinin–neuraminidase
glycoprotein: effects on antigenic structure and function. J. Virol. 68,
Nagai, T., Okafuji, T., Miyazaki, C., Ito, Y., Kamada, M., Kumagai, T., Yuri, K.,
Sakiyama, H., Miyata, A., Ihara, T., Ochiai, H., Shimomura, K., Suzuki, E.,
Torigoe, S., Igarashi, M., Kase, T., Okuno, Y., Nakayama, T., 2006. A
comparative study of the incidence of aseptic meningitis in symptomatic
natural mumps patients and monovalent mumps vaccine recipients in Japan.
Vaccine [Jan 31, Epub ahead of print].
Ong, G., Goh, K.T., Ma, S., Chew, S.K., 2005. Comparative efficacy of Rubini,
Jeryl-Lynn and Urabe mumps vaccine in an Asian population. J. Infect. 51,
Orvell, C., Alsheikhly, A.R., Kalantari, M., Johansson, B., 1997. Characteriza-
tion of genotype-specific epitopes of the HN protein of mumps virus. J. Gen.
Virol. 78, 3187–3193.
Palomo, C., Cane, P.A., Melero, J.A., 2000. Evaluation of the antibody
specificities of human convalescent-phase sera against the attachment (G)
proteinofhuman respiratory syncytial virus:influenceof strainvariationand
carbohydrate side chains. J. Med. Virol. 60, 468–474.
Paterson, R.G., Lamb, R.A., 1990. RNA editing by G-nucleotide insertion in
mumps virus P-gene mRNA transcripts. J. Virol. 64, 4137–4145.
Petrescu, A.J., Milac, A.L., Petrescu, S.M., Dwek, R.A., Wormald, M.R., 2004.
Statistical analysis of the protein environment of N-glycosylation sites:
Pipkin, P.A., Afzal, M.A., Heath, A.B., Minor, P.D., 1999. Assay of humoral
immunity to mumps virus. J. Virol. Methods 79, 219–225.
Rafiefard, F., Johansson, B., Tecle, T., Orvell, C., 2005. Characterization of
mumps virus strains with varying neurovirulence. Scand. J. Infect. Dis. 37,
Ramachandran, G.N., Sasisekharan, V., 1968. Conformation of polypeptides
and proteins. Adv. Protein Chem. 23, 283–438.
Rice, P., Longden, I., Bleasby, A., 2000. EMBOSS: the European Molecular
Biology Open Software Suite. Trends Genet. 16, 276–277.
Rima, B.K., Alexander, D.J., Billeter, M.A., Collins, P.L., Kingsbury, D.W.,
Lipkind, M.A., Nagai, Y., Orvell, C., Pringle, C.R., ter Meulen, V., 1995.
Paramyxoviridae. In: Murphy, F.A., Fauquet, C.M., Bishop, D.H.L.,
Ghabrial, S.A., Jarvis, A.W., Martelli, G.P., Mayo, M.A., Summers, M.D.
(Eds.), Virus Taxonomy. Sixth Report of the International Committee of
Viruses. Springer-Verlag, New York, pp. 268–274.
Rost, B., O'Donoghue, S., 1997. Sisyphus and prediction of protein structure.
Comput. Appl. Biosci. 13, 345–356.
Rubin, S.A., Pletnikov, M., Carbone, K.M., 1998. Comparison of the
neurovirulence of a vaccine and a wild-type mumps virus strain in the
developing rat brain. J. Virol. 72, 8037–8042.
Rubin, S.A., Pletnikov, M., Taffs, R., Snoy, P.J., Kobasa, D., Brown, E.G.,
Wright, K.E., Carbone, K.M., 2000. Evaluation of a neonatal rat model for
prediction of mumps virus neurovirulence in humans. J. Virol. 74,
Rubin, S.A., Afzal, M.A., Powell, C.L., Bentley, M.L., Auda, G.R., Taffs, R.E.,
Carbone, K.M., 2005. The rat-based neurovirulence safety test for the
assessment of mumps virus neurovirulence in humans: an international
collaborative study. J. Infect. Dis. 191, 1123–1128.
445U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446
Saika, S., Kidokoro, M., Aoki, A., Ohkawa, T., 2004. Neurovirulence of
mumps virus: intraspinal inoculation test in marmosets. Biologicals 32,
Saito, H., Takahashi, Y., Harata, S., Tanaka, K., Sano, T., Suto, T., Yamada, A.,
Yamazaki, S., Morita, M., 1996. Isolation and characterization of mumps
virus strains in a mumps outbreak with a high incidence of aseptic
meningitis. Microbiol. Immunol. 40, 271–275.
Sauder, C.J., Vandenburgh, K.M., Iskow, R.C., Malik, T., Carbone, K.M.,
Rubin, S.A., 2006. Changes in mumps virus neurovirulence phenotype
associated with quasispecies heterogeneity. Virology 350, 48–57.
Seibel, G., Singh, U.C., Weiner, P.K., Caldwell, J., Kollman, P., 1990. AMfBER
3.0 revision A. University of California at San Francisco, San Francisco.
Sippl, M.J., 1993. Recognition of errors in three-dimensional structures of
proteins. Proteins 17, 355–362.
Smith-Gill, S.J., 1996. Molecular recognition of lysozyme by monoclonal
antibodies. EXS 75, 277–300.
Takahashi, M., Nakayama, T., Kashiwagi, Y., Takami, T., Sonoda, S.,
Yamanaka, T., Ochiai, H., Ihara, T., Tajima, T., 2000. Single genotype of
measles virus is dominant whereasseveral genotypes of mumps virus are co-
circulating. J. Med. Virol. 62, 278–285.
Takimoto, T., Taylor, G.L., Crennell, S.J., Scroggs, R.A., Portner, A., 2000.
Crystallization of Newcastle disease virus hemagglutinin–neuraminidase
glycoprotein. Virology 270, 208–214.
Takimoto, T., Taylor, G.L., Connaris, H.C., Crennell, S.J., Portner, A., 2002.
Role of the hemagglutinin–neuraminidase protein in the mechanism of
paramyxovirus-cell membrane fusion. J. Virol. 76, 13028–13033.
Tanaka,K., Baba,K., Okada,S., Okuno,Y., Yamanishi, K., Ueda, S.,Takahashi,
M., Yamada, A., 1992. Nasal antibody response to mumps virus after
vaccination and natural infection. Vaccine 10, 824–827.
Tecle, T., Johansson, B., Jejcic, A., Forsgren, M., Orvell, C., 1998.
Characterization of three co-circulating genotypes of the small hydrophobic
protein gene of mumps virus. J. Gen. Virol. 79, 2929–2937.
Tecle, T., Bottiger, B., Orvell, C., Johansson, B., 2001. Characterization of two
decades of temporal co-circulation of four mumps virus genotypes in
Denmark: identification of a new genotype. J. Gen. Virol. 82, 2675–2680.
Tecle, T., Mickiene, A., Johansson, B., Lindquist, L., Orvell, C., 2002.
Molecular characterisation of two mumps virus genotypes circulating during
an epidemic in Lithuania from 1998 to 2000. Arch. Virol. 147, 243–253.
Utz, S., Richard, J.L., Capaul, S., Matter, H.C., Hrisoho, M.G., Muhlemann, K.,
2004. Phylogenetic analysis of clinical mumps virus isolates from
vaccinated and non-vaccinated patients with mumps during an outbreak,
Switzerland 1998–2000. J. Med. Virol. 73, 91–96.
Van Regenmortel, M.H., 1998. Mimotopes, continuous paratopes and hydro-
pathic complementarity: novel approximations in the description of
immunochemical specificity. J. Dispers. Sci. Technol. 19, 1199–1219.
Wheeler, D.L., Barrett, T., Benson, D.A., Bryant, S.H., Canese, K., Chetvernin,
V., Church, D.M., DiCuccio, M., Edgar, R., Federhen, S., Geer, L.Y.,
Helmberg, W., Kapustin, Y., Kenton, D.L., Khovayko, O., Lipman, D.J.,
Madden, T.L., Maglott, D.R., Ostell, J., Pruitt, K.D., Schuler, G.D., Schriml,
L.M., Sequeira, E., Sherry, S.T., Sirotkin, K., Souvorov, A., Starchenko, G.,
Suzek, T.O., Tatusov, R., Tatusova, T.A., Wagner, L., Yaschenko, E., 2006.
Database resources of the National Center for Biotechnology Information.
Nucleic Acids Res. 34, D173–D180.
Wolinsky, J.S., 1996. Mumps Virus. Lippincott-Raven, Philadelphia.
Wu, T.T., Kabat, E.A., 1970. An analysis of the sequences of the variable
regions of Bence Jones proteins and myeloma light chains and their
implications for antibody complementarity. J. Exp. Med. 132, 211–250.
Wu, L., Bai, Z., Li, Y., Rima, B.K., Afzal, M.A., 1998. Wild type mumps
viruses circulating in China establish a new genotype. Vaccine 16,
Yates, P.J., Afzal, M.A., Minor, P.D., 1996. Antigenic and genetic variation of
the HN protein of mumps virus strains. J. Gen. Virol. 77, 2491–2497.
Yuan, P., Thompson, T.B., Wurzburg, B.A., Paterson, R.G., Lamb, R.A.,
Jardetzky, T.S., 2005. Structural studies of the parainfluenza virus 5
hemagglutinin–neuraminidase tetramer in complex with its receptor,
sialyllactose. Structure 13, 803–815.
Zaitsev, V., von Itzstein, M., Groves, D., Kiefel, M., Takimoto, T., Portner, A.,
Taylor, G., 2004. Second sialic acid binding site in Newcastle disease
virus hemagglutinin–neuraminidase: implications for fusion. J. Virol. 78,
446 U. Kulkarni-Kale et al. / Virology 359 (2007) 436–446