Journcd of General Virology (1997), 78, 2853-2858....
Journcd of General Virology (1997), 78, 2853-2858. Printed in Great Britaln
Variability of the NSs protein among Rift Valley fever virus
isolates
A. A. Sali,’ P. M. de A. Zanotto,’ H. G. Zeller,3 J. P. Digoutte,’ Y. Thiongane4 and M. Bouloy5
‘)
’ Institut Pasteur, 36 Avenue Pasteur, BP 220 Dakar, Senegal
2 DIPA-UNIFESP,
Sao Paulo, Brazil
3 Institut Pasteur de Madagascar, BP 1274, Tananarive, Madagascar
e
4 Institut Sénégalais de Recherche Agronomique, BP 2057 Dakar, Senegal
5 Laboratoire des Bunyavirides, Institut Pasteur, 28 Rue du Docteur Roux, 75724 Paris Cedex 1 5, France
Eighteen strains of Rift Valley fever (RVF) virus
vergence among isolates ranged from 0 to 9.6% at
collected over a period of 38 years and isolated
the nucleotide level and from 0 to 9.5% at the
from diverse localities in Africa and from various
amino acid level. Phylogenetic analysis based on the
hosts (human, animal and arthropod) were investi-
NS, gene revealed two major lineages: Egyptian
gated by RT-PCR followed by sequencing of the NS,
and sub-Saharan. This led to the establishment of
protein coding region. This region was chosen to
the relatedness between strains and insights into
analyse variability because, in contrast to the N
the NS, protein, the function of which is still
protein, the NS, protein differs in various phlebo-
undetermined. Alignment of the deduced amino
viruses and there exists an RVF virus (clone 13) in
acid sequences indicated that the cysteine residues
which 70% of the NS, ORF is deleted, suggesting
are conserved, as are several motifs representing
that this sequence is under a weak selective press-
potential phosphorylation sites.
ure. Sequence data indicated that percentage di-
depend on transovarial transmission of virus in floodwater
Rift Valley fever (RVF) is an arthropod-borne viral disease
Aedes mosquitoes (Linthicum ef al., 1985).
in Africa, primarily affecting ruminants, which results in high
RVF virus belongs to the genus Phlebovirtts in the family
mortality in Young and adult animals as well as abortions. Since
Bunyaviridae, a genus which contains two serogroups: the
the epidemic in Egypt in 1977, RVF virus has been shown to
sandfly fever viruses and the tick-transmitted uukuviruses
infect humans extensively, leading to several disease manifest-
(Murphy et al., 1995). According to serological data based on
ations including acute febrile illness, hepatitis, encephalitis,
complement fixation, neutralization and haemagglutination
haemorrhagic fever and ocular sequelae (Laughlin et al., 1979).
inhibition tests, the sandfly fever group is composed of at least
Recent RVF epidemics in Mauritania (Digoutte & Peters,
eight antigenic complexes: Bajaru, Candiru, Chilibre, Frijoles,
1989), Madagascar (Morvan et a/., 1992) and Egypt (Arthur et
Punta Toro, Rift Valley fever, Salehabad and an unassigned
al., 1993), as well as the enzootic maintenance of the virus in
complex containing 16 viruses.
Senegal (Zeller et al., 1997), have emphasized the emergence of
The genome of phleboviruses consists of three negative-
this virus as a major public health threat in Africa. With the
sense single-stranded RNA segments designated L (large), M
exception of Egypt, RVF virus is confined to sub-Saharan
(medium) and S (small). The L segment codes for the L protein,
Africa, where it is widespread and maintained in endemic/
the viral RNA polymerase. The M segment codes for a
enzootic and epizootic forms (Meegan & Bailey, 1989). The
precursor to the envelope glycoproteins GI and G2 and two
survival of RVF virus during inter-epizootics is believed to
non-structural proteins of 14 and 78 kDa. The S segment codes
r-.-~.~
for nucleocapsid protein N and the non-structural protein NS,
by an ambisense strategy (for reviews see Bouloy, 1991; Elliott
Author for correspondence: Michèle Bouloy.
et al., 1991; Giorgi, 1996; Schmaljohn, 1996).
Fax j-33 1 40 61 31 51. e-mail mbouloy@pasteur.fr
Molecular and biological variability among phleboviruses
The GenBank/EMBL accession numbers of the sequences reported are
has been studied for RVF virus (Battles & Dalrymple, 1988;
Y1 2739-Y12756.
~-~--
--..-~_~-~
Anderson & Peters, 1988; Saluzzo et al., 1989; Besselar et al.,
0001-4905 0 1997 SGM
1991) and to a lesser extent for Toscana virus (Schwarz et al.,
Table 1. Characteristics of the RVF virus isolates used in
19%). These reports indicated that among natural isolates of
this study
RVF virus, the antigenic properties of the glycoproteins and
H, human; Ar, arthropod; An, animal: B, buffalo.
the nucleoprotein appeared to be stable, and sequences of the
G2 epitopes in the M segment are relatively conserved. On the
other hand, the Egyptian strains differed from the sub-Saharan
Isolation
strains by their pathogenicity in Wistar-Furth rats. Since the
NS, protein is the most variable protein among phleboviruses
Code
Strain
Year
Source
Origin
(Giorgi et al., 1991), its use as a marker of variability was
S N S
Smithburn
1944 Strain
Uganda
investigated. The extent of conservation of the NS, gene
Entebbe
among RVF virus isolates is not known. This question was
Ar UG 55
Lunyo
1955 Mosquito
Uganda
raised by the existence of the naturally occurring avirulent
Ar CAR 69
Ar B 1976
1 9 6 9 Mosquito
CAR
isolate ‘clone 13’ (Muller ef al., 1995), which harbours a
H EGY 77
ZH 548
1977 Human
Egypt
prominent deletion in the NS, region. This suggests that this
MPIZ”
MP12
1977 Strain ZH 548
Egypt
domain is under weakly constraining evolutionary forces,
B EGY 93
B EGY 93
1993 Buffalo
Ewpt
H EGY 93
H EGY 93
1993 Human
Eupt
resulting in the potential for accumulation of substitutions in
Ar MAD 79
Ar Mg 811
1979 Mosquito
Madagascar
this region. A pane1 of 18 strains of RVF virus, collected over
An MAD 91
An Mg 990
1991 Bovine
Madagascar
a period of 38 years and isolated from various hosts in different
HI MAU 87
H D 47502
1987 Human
Rosso
countries, was selected and the NS, gene was amplified by
(October 28)
(Mauritania)
RT-PCR and sequenced. This analysis was carried out to
HZ MAU 87
H D 47311
1987 Human
Yeur
(October 24)
Massène
establish the relatedness between strains as well as to gain
(Mauritania)
insights into the conserved motifs of this protein, the function
H3 MAU 87
H D 47408
1987 Human
Tegmaline
of which is still undetermined.
(October 27)
(Mauritania)
H4 MAU 87
H D 48255
1987 Human
Rosso
(November 11) (Mauritania)
An GUI 84
An K 6087
1984 Bat
Kindia
Methods
(Guinea)
W Virus propagation and RNA extraction. The origins of RVF
Ar BUR 84
Ar D 38457
1984 Mosquito
Burkina
virus isolates used in this study are shown in Table I. Viruses were
Faso
propagated in Vero cells cultivated in Leibovitz 15 growth medium
Ar SEN 84
Ar D 38661
1984 Mosquito
Kedougou
supplemented with 5 % foetal calf serum (Gibco BRL). At 48 to 72 h post-
(Senegal)
infection, infected cultures showed cytopathic effects and the presence of
Ar SEN 93
Ar D 104769 1993 Mosquito
Ferlo
(Senegal)
viral antigens could be detected by an indirect immunofluorescence
An D 106417 1993 Bovine
Kolda
assay. Cytoplasmic extracts were obtained by the NP40 lysis method and
An SEN 93
(Senegal)
total RNA was extracted with phenol-chloroform (1: 1) and precipitated
with ethanol as previously described (Bouloy et al., 1984).
n RT-PCR and sequencing procedures. Synthesis of first strand
’ Laboratory attenuated strain derived from a wild strain.
complementary DNA was primed with oligodeoxynucleotide NS3a
SNS, Smithburn neurotropic strain.
(Muller ef nl., 1995) after denaturation of the RNA with 0.0 L M methyl
mercuric hydroxide for 10 min at room temperature. Methyl mercuric
hydroxide was then complexed with 2 ~1 of 0.7 M p-mercaptoethanol,
were done with the primers used for PCR amplification. Most of the
and reverse transcription was carried out with 100 ng of primer in a final
sequences were read on the two strands. The sequences were deposited
volume of 20 ~1 containing 25 mM Tris-HC1 pH 8.3,25 mM KCI, 5 mM
in the EMBL database (accession numbers Y12739-Y12756).
MgCl,, 0.2 mM of each dNTP (Pharmacia) and 4 units of AMV reverse
transcriptase (Promega).
n Sequence alignments and phylogenetic analysis. The arnino
A 5 ~1 aliquot of the reverse transcription reaction mix was used for
acid sequences of the NS, protein were aligned using Clustal V (Higgins
enzymatic amplification. PCR was carried out with 2.5 units of Tn9 DNA
et ul., 1992) in order to determine the positions of gaps in the ahgnment
polymerase (Amersham) in a 100 ~1 final volume containing 10 mM
of the corresponding nucleotide sequence data. Phylogenetic trees for the
Tris-HC1 pH 8.3, 50 mM KCl, 1.5 mM MgCI,, 0.2 mM of each dNTP
NS, nucleotide sequences were estimated using the FastDNAml program
and 05 pg of each primer (NS3a and NS2g) (Muller ef ai., ~995).
(Olsen ef nl., 1994). The transition/transversion
substitution rate
Amplification was performed after a denaturation step at 95 “C for 3 min
parameter (ts: tv) was calculated by searching for the ts: tv value which
followed by 35 amplification cycles as follows: denaturation at 95 “C for
gives the highest likelihood. In order to establish the robustness of the
1 min, annealing at 53 “C for 1 min and extension at 72 “C for 1 min.
topologies, bootstrap resampling was done (Swofford, 1993). In addition,
Amplified DNA fragments were purified on an agarose gel with the
given an optimal ts: tv parameter, a11 the trees with log likelihoods (1nL)
Geneclean kit (Bio 101) according to manufacturer’s instructions and
not significantly different from the best maximum likelihood tree were
directly sequenced using the thermal cycle sequencing method with
collected. The level of resolution of the 50% majority rule consensus of
oligodeoxynucleotides labelled at the 5’ end with [g-““P]ATI’
and
this tree pool is then indicative of the topology stability at the tree space
polynucleotide kinase (Adams & Blaskesley, 1991). Sequencing reactions
around the InL maxima. These results were also compared to recon-
structions obtained with distance methods in PHYLIP (Felsenstein, 1993)
host species (arthropods, cattle and humans). The character-
and maximum parsimony with PALJP 3.1 (Swofford, 1993).
istics of the isolates are described in Table 1. The attenuated
The S segment sequences of RVF and sandfly fever Sicilian (SFS)
strains MP12 (Caplen et al., 1985) and Smithburn neurotropic
viruses were retrieved from the GenBank database (accession numbers
strain (SNS; Smithburn, 1949), which were obtained by ce11
X53771 and JO4418, respectively).
passages of the parental virulent strains ZH 548 and Entebbe,
n Rates of evolution and selection of different genes of RVF
respectively, were included in this study.
virus. In order to investigate selective regimes, the number of
Two oligodeoxynucleotides, NS2g and NS3ag (Muller et
synonymous (silent) substitutions per synonymous site (ds) and the
al., 1995), were designed based on the published sequence of
number of non-synonymous substitutions (,or amino acid altering
the MP12 strain and utilized as primers for RT-PCR
changes) per non-synonymous site (dn) were calculated for the nucleotide
data sets using the Jukes-Cantor correction method for distances less
amplification. They allowed amplification of a 601-nucleotide-
than 0.4 substitutions per site and the proportional distance method (p-
long DNA fragment which was obtained for a11 the strains
distance) using the MEGA program (Kumar ef a/., 1993) for distances
analysed and subsequently sequenced. Using these primers,
greater than 04.
clone 13 yielded a 12l-nucleotide-long DNA fragment (Muller
et al., 1995), which was too small to be included in this analysis.
Nucleotide substitutions were scattered throughout the entire
Results and Discussion .-....-_-._ ~
region and, with the exception of clone 13, no base insertion or
Phylogenetic analysis of the RVF virus NS, genes
deletion was observed. Pairwise comparison among these
The S genomic sequences of five phleboviruses, Punta
isolates showed percentage divergence ranging from 0 to
Toro, RVF (MPl2 strain), SFS, Toscana and Uukuniemi viruses,
9.6 % at the nucleotide level which corresponded
to 0 to 9.5 %
have already been published (Ihara et al., 1984; Marriott et ai.,
at the amino acid level.
1989; Simons et al., 1990; Giorgi et al., 1991) and were shown
The phylogenetic tree of the nucleotide sequences was
to utilize an ambisense strategy to code for the NS, protein in
generated using the optimal ts: tv value of 2, which was
the genomic sense and for the N protein in the antigenomic
determined empirically. The SFS virus NS, sequence was used
sense. The N and NS, ORFs represent two potential distinct
as an outgroup to determine the coalescence node for RVF
targets for analysis of the S segment. The deduced N protein
virus sequences and therefore indicate a root for the tree. Using
sequences exhibit homologies ranging from 30 to 54 % (Giorgi
ultrametric methods and excluding the SFS virus branch, a
et al., 1991); however, the NS, gene is much less conserved and
congruent root was obtained for Ar MAD 79 (Fig. 1). Similar
cannot be aligned to analyse phylogenetic relationships among
trees were obtained with the maximum likelihood distance
different phleboviruses. Noteworthy is that the function of the
methods and maximum parsimony (not shown). Apart from Ar
protein in the virus life cycle has not yet been determined and
MAD 79, the isolates are distributed in at least two major
that its localization in infected cells, as well as its presence in
lineages, Egyptian and sub-Saharan, with the latter lineage
viral particles, varies depending on the virus. With the
divided into two clusters, Ia and Ib. Al1 the branchings of the
exception of the NS, protein from RVF virus, which has been
tree are highly supported by bootstrap values. Group Ib strains
reported to form filaments in the nuclei of infected cells
were isolated only in West Africa (Senegal, Mauritania, Guinea
(Struthers & Swanepoel, 1982; Struthers et ai., 1984), the NS,
and Burkina Faso), whereas group Ia appears more hetero-
protein of the phleboviruses studied was detected in the
geneous, including strains from Uganda, Madagascar, Central
cytoplasm. In the case of Uukuniemi virus, Simons ef al. (1992)
African Republic as well as strains from Senegal and
demonstrated that the protein was associated with the 40s
Mauritania.
ribosomal subunit and not detected in mature particles. In
The existence of the Egyptian lineage is in good agreement
contrast, for Punta Toro virus, the term ‘non-structural’ does
with a set of data indicating that the strains from the 1977
not seem appropriate since the protein was found associated
outbreak could be distinguished from the sub-Saharan strains
with the ribonucleoproteins in mature particles (Overton et al.,
by several criteria: (i) the N protein of the Egyptian strains
1987).
possesses a specific epitope recognized by the RlP2E7
The diversity of the NS, gene among phleboviruses raised
monoclonal antibody (Saluzzo et al., 1989); (ii) the Egyptian
the question of the extent of conservation of the NS, gene
strains were also highly pathogenic for Wistar-Furth rats; (iii)
among strains of a particular virus. This was underlined by the
these strains were resistant to rat interferon; and (iv) they did
existence of clone 13, a natural, attenuated isolate of RVF virus
not form plaques in primary rat hepatocytes (Anderson &
which harbours a large deletion in the NS, region and replicates
Peters, 1988). As expected, the MP12 strain is closely related
in various hosts including mosquitoes, suggesting that the
to its parent, H EGY 77, but surprisingly, is placed in a more
deleted region represents a potentially variable domain (M&er
ancestral position. The use of a mutagen to obtain MP12 may
et al., 1995). Therefore, we decided to analyse the NS, gene of
have altered the mutation regime in a very biased way for the
various strains of RVF virus and selected a pane1 of 18 RVF
NS, sequence. Moreover, this analysis shows that the strains
virus strains collected over 38 years in eight countries under
isolated in 1977 and 1993 in Egypt are very close to each
epidemic and endemic conditions and isolated from a variety of
other, suggesting that the same virus remained endemic for a
--..
-
zebu in southern Senegal (Kolda, Casamance). Interestingly,
this strain clustered with those from the neighbouring countries
of Guinea and Burkina Faso. This suggests that Senegal is
a
divided in two areas as far as RVF virus is concemed: the
Sahelian zone where southem Mauritanian and northem
Senegalese strains are circulating, probably through livestock
and mosquitoes, and the Sudano-Guinean zone where south-
em Senegalese strains are in contact with those from bordering
I ArBUR84
countries.
Regarding its geographical origin (i.e. from mosquitoes
collected in Kedougou, a place in southem Senegal close to
H4 MAU 87
Guinea and Mali), Ar SEN 84 would be expected to cluster
Ib
with An GUI 84 which was isolated from a bat in Kindia
- H3MAU 87
(Guinea), a bioclimatic area similar to Kedougou and in the
An SEN 93
vicinity of the Senegalese border. Its presence within the
An GUI 84
Egyptian lineage,
although puzzling, may be due to a
reassortment event. Indeed, the circulation of phylogenetically
distinct strains in the same region and at the same period
provides ideal conditions for a host to be co-infected and
II
generate reassortants as occurs naturally for other members of
the bunyavims family like hantaviruses (Li et al., 19%;
L ArSEN84
l1 Henderson et al., 1995) and in other families of viruses with
/ AI MAD 79
segmented genomes (Murphy & Webster, 1990). Further work
Will be necessary to test whether some of these viruses result
from reassortment.
Fig. 1. Phylogenetic tree for the NS, gene of RVF virus isolates. Values
Finally, the strains isolated in Madagascar in 1991 and 1979
above branches indicate the level (%) of bootstrap support using
belong to different lineages; strain Ar MAD 79 is closely
maximum parsimony after 500 iterations. Values below branches indicate
the number of times a given node was observed on a major@ rule
related to the Egyptian group. This relatedness was noted by
consensus of 50 trees with equivalent likelihood (InL). Branch lengths
Morvan et al. (1992) who analysed the antigenic properties of
shown are proportional to the number of substitutions per 100 residues.
the N protein. The topology of the tree reported in Fig. 1
The rooting shown here was determined by the inclusion of the SFS virus
NS, sequence as the outgroup.
indicates that the virus isolated in the island in 1979 occupies
the most ancestral position and shares a common ancestor with
the Egyptian strains. For these reasons, Ar MAD 79 was
long period in a large area or that the viruses circulating in
included in the Egyptian lineage. On the other hand, strain An
1977 and 1993 were introduced from the same source.
MAD 91 appears more closely related to the eastem or central
Grouping based on geographical origin applies also to the two
African strains. As speculated by Morvan et al. (~EU), it is
strains from Uganda, Ar UG 55 and SNS, a laboratory-
possible that during this latter outbreak, RVF virus had been
attenuated strain derived from the Entebbe strain isolated in
introduced to the island from the southem or eastem toast of
Uganda in 1944. Nevertheless, some of the groupings,
Africa through livestock commerce.
especially those of some Mauritanian and Senegalese strains,
are unexpected. The phylogenetic tree indicates that two
phylogenetically different viruses corresponding to clusters Ia
Insights concerning the conserved amino acids in the
and Ib were circulating during the Mauritanian outbreak in
NS, protein
1987. However, the strains analysed originated from three
Six potential phosphorylation sites for casein kinase 1 and
places less than 50 km apart and from cases occurring over less
2, corresponding to the motifs S/T-X-R/K and S/T-X-X/-D/E,
than I month (see Table 1). It should be noted that strain HI
were identified in the NS, protein of the MP12 strain. Al1 of
MAU 87 was isolated from a febrile case, and the others were
them are conserved in the strains SNS, H EGY 77, B and H
isolated from fatal cases. Moreover, HI MAU 87 is closely
EGY 93, and Ar SEN 84, the complete NS, ORF of which was
related to the Senegalese strain Ar SEN 93 which was isolated
sequenced. Among the other strains studied and within the
from mosquitoes trapped in 1993 in the Sahelian Ferlo, a
region analysed, there are three potential sites at positions 98,
northern region close to the Mauritanian border where the
128 and 149 which are a11 conserved.
This suggests that the
1987 outbreak occurred (Zeller et al., 1997). On the other hand,
protein could be phosphorylated at multiple sites.
the position of strain An SEN 93 within a different cluster
The amino acid sequence alignment also indicates that the
could be explained by its origin; it was isolated from a healthy
five cysteine residues at positions 39,40, 150, 179 and 1% are
conserved. It is possible that these residues form either intra- or
Caplen, H., Peters, C. 1. h Bishop,
D. H. L. (1985). Mutagen-directed
intermolecular disulfide bonds which are important for con-
attenuation of Rift Valley fever virus as a method for vaccine
servation of the three-dimensional structure of the protein and
development. ]ourna~ of General Virology 66, 2271-2277.
the formation of filamentous structures observed in cells
Digoutte, 1. P. & Peters, C. 1. (1989). General aspects of the 1987 Rift
Valley fever epidemic in Mauritania. Research in Virology 140, 27-30.
infected with a11 the strains analysed SO far, except for strain
C13. Also noteworthy is the presence of the RGD motif at
Elliott, R. M., Schmaijohn, C. S. & Colle$ M. S. (1991). Bunyaviridae
genome structure and gene expression. In Current Topics in Microbiology
position 24, conserved in a11 isolates except for An MAD 91 in
and Immuno~ogy,
pp. 91-142. Berlin: Springer-Verlag.
which arginine is replaced by lysine. The existence of such a
Felsenstein, 1. (1993). PHYLIP (phylogeny inference package) version
motif, commonly present in extracellular adhesive proteins, in
3%. Distributed by the Author, Department of Genetics, University of
this intracellular protein is surprising.
Washington, Seattle, USA.
Altogether these data indicate that the structure of the RVF
Giorgi, C. (1996). Molecular biology of phleboviruses. In The Bunyn-
virus NS, gene is conserved and that clone 13 is a very distinct
viridae, pp. 105-128. Edited by R. M. Elliott. New York: Plenum Press.
isolate. The existence of this virus raises the question of
Giorgi, C., Accardi, L., Nicoletti, L., Gro, M. C., Takehara, K., Hilditch,
whether the NS, protein is essential. In an attempt to
C., Morikawa, S. & Bishop,
D. H. L. (1991). Sequences and coding
investigate selective regimes on the NS, gene, ds and dn were
strategies of the S RNAs of Toscana and Rift Valley fever viruses
calculated. The ds:dn ratio was found to be greater than one
compared to those of Punta Toro, Sicilian sandfly fever, and Uukuniemi
viruses. Virology 180, 738-753.
(8.36; P < 0.001) indicating that during evolution more
Henderson, W. W., Monroe, M. C., St Jeor, S. C., Thayer, W. P., Rowe,
synonymous than non-synonymous changes accumulated.
1. E., Peters, C. J. h Nichol, S. T. (1995). Naturally occurring Sin
This suggests the existence of constraints against amino acid
Nombre virus genetic reassortants. Virology 214, 602-610.
changes in the protein.
Higgins, D. G., Bleasby, A. 1. & Fuchs, R. (1992). Clustal V: improved
Finally, the oligonucleotides used for RT-PCR in this
software for multiple sequence alignment. Cumpufer
Applications in the
analysis appear to be good primers for amplifying the NS,
Biosciences 8, 189-191.
specific region of the 18 isolates presented herein as well as
Ihara, T., Matsuura, Y. & Bishop,
D. H. L. (1984). Novel coding
other isolates tested in our laboratory (data not shown).
strategy (ambisense genomic RNA) revealed by sequence analyses of
Therefore, this work defines the basis for a rapid diagnostic test
Punta Toro phlebovirus S RNA. Virology 136, 293-306.
for RVF virus based on RT-PCR.
Kumar, S., Tamura, K. 81 Nei, M. (1993). MEGA: Molecular Evol-
utionary Genetic Analysis, version 1.0. Pennsylvania State University,
University Park, PA 16802, USA.
The authors thank P. Vialat for excellent technical expertise. We are
Laughlin, L. W., Meegan, 1. M., Strausbaugh, L. W., Morens, D. M. &
grateful to Dr H. Bourhy for providing computer facilities and critically
Watten, R. H. (1979). Epidemic Rift Valley fever in Egypt: observations
reading the manuscript. We thank Dr C. Prehaud, A. Billecocq and B. Le
of the spectrum of human illness. Trunsactions
of fhe Royal Sociefy of
Guenno for fruitful discussions.
Tropical Medicine and Hygiene 73, 630-633.
Li, D., Schmaljohn, A. L., Anderson, K. h Schmaljohn, C. S. (1995).
Complete nucleotide sequences of the M and S segments of two
References ~
hantavirus isolates from Califomia: evidence for reassortment in nature
Adams, S. & Blaskesley, R. (1991). Linear amplification DNA
among viruses related to hantavirus pulmonary syndrome. Virology 206,
sequencing. Facrts 13, 56-57.
973-983.
Anderson, G. W., Jr & Peters, C. 1. (1988). Virai determinants of
Linthicum, K. J., Davies, F. G. & Kairo, A. (1985). Rift Valley fever virus
virulence for Rift Valley fever (RVF) in rats, Microbinl Puthogenesis
5,
(family Bunyaviridae,
genus Phlebovirw). Isolations from Diptera collected
247-250.
during an inter-epizootic period in Kenya. Ja~rnal ofHygiene 95, 197-209.
Arthur, R., El-Sharkawy, M. S., Cope, S. E., Botros, B. A., Oun, S.,
Marriott, A. C., Ward, V. K. & Nuttall, P. A. (1989). The S RNA segment
Morrill, J. C., Shope, R. E., Hibbs, R. G., Darwish, M. A. h Imam, 1. 2. E.
of sandfly fever Sicilian virus. Evidence for an ambisense genome.
( 1 9 9 3 ) . Recurrence of Rift Valley fever in Egypt. Lnncef 342,
Virology 169, 341-345.
1149-1150.
Meegan, 1. M. & Bailey, C. 1. (1989). Rift Valley fever. In The
Battles, J. K. & Dalrymple, J. M. (1988). Genetic variation among
Arboviruses: Epidemiology and Ecology, pp. 51-76. Edited by T. P. Monath.
geographic isolates of Rift Valley fever virus. American Journal of Tropical
Boca Raton, Fla.: CRC Press.
Medicine and Hygiene 39, 617-631.
Morvan, J., Rollin, P. E., Laventure, S., Rakotoarivony, 1. h Roux, J.
Besselar, T. G., Blackburn, N. K. h Meenehan, G. M. (1991). Antigenic
(1992). Rift Valley fever epizootic in the central highlands of
analysis of Rift Valley fever virus isolates: monoclonal antibodies
Madagascar. Research in Virology 143, 407-415.
distinguish between wild-type and neurotropic virus strains. Research in
Muller, R., Saluzzo, J. F., Lopez, N., Dreier, T., Turell, M., Smith, J. &
Virology 142, 469-474.
Bouloy, M. (1995). Characterization of clone 13, a naturally attenuated
Bouloy, M. (1991). Bunyaviridue: genome organization and replication
avirulent isolate of Rift Valley fever virus, which is altered in the small
strategies. Advances in Virus Research 40, 235-266.
segment. American jotcrnnl ej Tropical Medicine and Hygiene 53, 405-411.
Bouloy, M., Vialat, P., Girard, M. h Pardigon, N. (1984). A transcript
Murphy, B. R. & Webster, R. G. (1990). Orthomyxoviruses. In Virology,
from the S segment of the Germiston bunyavirus is uncaped and codes
;;es~1-l152.
Edited by B. N. Fields & D. M. Knipe. New York: Raven
for the nucleoprotein and a nonstructural protein. journal of Virology 49,
717-723.
Murphy, F. A., Fauquet, C. M., Bishop,
D. H. L., Ghabrial, S. A., larvis,
A. W., Martelli, G. P., Mayo, M. A. h Summers, M. D. (editors) (1995).
Simons, J. F., Hellman, U. h Pettersson, R. F. (1990). Uukuniemi virus
Family Bmnyuoiridue. In Virtrs Taxonorny. Sirfh Report of fhe Wernrzfional
S RNA segment: ambisense coding strategy, packaging of comp-
Carnmiffee on Taxonomy
nf Vivuses, pp. 300-315. Vienna % New York:
lementary strands into virions and homology to members of the genus
Springer-Verlag.
Ph&rz~irns. ]ourna~ of Viro/ogy 64, 247-255.
Olsen, G. J., Matsuda, H., Hagstrom, R. & Overbeek, R. (1994).
Simons, J. F., Persson, R. & Pettersson, R. F. (1992). Association of
FastDNAml: a tool for construction of phylogenetic trees of DNA
the nonstructural protein NS, of Uukuniemi virus with the 40s ribosomal
sequences using maximum likelihood. Compzzfer Applications in the
subunit. Jortrnn! of I/irdogy 66, 4233-4241.
Biosciences 10, 41-48.
Smithburn, K. C. (1949). Rift Valley fever: the neurotropic adaptation of
Overton, H. A., Ihara, T. & Bishop,
D. H. L. (1987). Identification of the
the virus and the experimental use of this modified virus as a vaccine.
N and NS, proteins coded by the ambisense S RNA of f’unta Toro
Brifish journal of Eq~erimenfal P~~fhology 30, l-16.
phlebovirus using monospecific antisera raised to baculovirus expressed
Struthers, J. K. h Swanepoel, R. (1982). Identification of a major non-
N and NS, proteins. Virolegy 15 7, 338-350.
structural protein in the nuclei of Rift Valley fever virus-infected cell
Saluzzo, 1. F., Anderson, G. W., Hodgson, L. A., Jr, Digoutte, 1. P. &
]ournu~ of Geaernl Virabgy 60, 381-384.
Smith, 1. F. (1989). Antigenic and biological properties of Rift Valley
Struthers, 1. K., Swanepoel, R. & Shepard, S. P. (1984). Protein
fever virus isolated during the 1987 Mauritanian epidemic. Resecrrch in
synthesis in Rift Valley fever virus-infected cell. Virobgy 134, 118-124.
Vkhgy 140, 155-164.
Swofford, D. L. (1993). PAUP: phylogenetic analysis using parsimony.
Schmaljohn, C. S. (1996). Bnnyaviridae: the viruses and their replication.
version 3.X.1. Computer program distributed by the Illinois Natural
In Fields Virology, 3rd edn, pp. 1447-1471. Edited by B. N. Fields, D. M.
History Survey, Champaign, Illinois, USA.
Knipe & P. M. Howley. Philadelphia: Lippincott-Raven.
Zeller, H. G., Fontenille, D., Traore-Lamizana, M., Thiongane, Y. 6
Schwarz, T. F., Jager, G., Gilch, S. & Nithchko, H. (1995). Nested
Digoutte, J. P. (1997). Enzootic activity of Rift Valley fever virus in
RT-I’CR for detection of sandfly fever virus, serotype Toscana, in clinical
Senegal. American Jmrnd of Tropica/ A4edicine and Hygiene 56, 265-272.
specimens, with confirmation by nucleotide sequence analysis. Reseizrch in
Viralogy 146, 355-362.
R e c e i v e d 2 9 A p r i l 1 9 9 7 ; Accepted 1 9 J u n e 1 9 9 7
Variability of the NSs protein among Rift Valley fever virus
isolates
A. A. Sali,’ P. M. de A. Zanotto,’ H. G. Zeller,3 J. P. Digoutte,’ Y. Thiongane4 and M. Bouloy5
‘)
’ Institut Pasteur, 36 Avenue Pasteur, BP 220 Dakar, Senegal
2 DIPA-UNIFESP,
Sao Paulo, Brazil
3 Institut Pasteur de Madagascar, BP 1274, Tananarive, Madagascar
e
4 Institut Sénégalais de Recherche Agronomique, BP 2057 Dakar, Senegal
5 Laboratoire des Bunyavirides, Institut Pasteur, 28 Rue du Docteur Roux, 75724 Paris Cedex 1 5, France
Eighteen strains of Rift Valley fever (RVF) virus
vergence among isolates ranged from 0 to 9.6% at
collected over a period of 38 years and isolated
the nucleotide level and from 0 to 9.5% at the
from diverse localities in Africa and from various
amino acid level. Phylogenetic analysis based on the
hosts (human, animal and arthropod) were investi-
NS, gene revealed two major lineages: Egyptian
gated by RT-PCR followed by sequencing of the NS,
and sub-Saharan. This led to the establishment of
protein coding region. This region was chosen to
the relatedness between strains and insights into
analyse variability because, in contrast to the N
the NS, protein, the function of which is still
protein, the NS, protein differs in various phlebo-
undetermined. Alignment of the deduced amino
viruses and there exists an RVF virus (clone 13) in
acid sequences indicated that the cysteine residues
which 70% of the NS, ORF is deleted, suggesting
are conserved, as are several motifs representing
that this sequence is under a weak selective press-
potential phosphorylation sites.
ure. Sequence data indicated that percentage di-
depend on transovarial transmission of virus in floodwater
Rift Valley fever (RVF) is an arthropod-borne viral disease
Aedes mosquitoes (Linthicum ef al., 1985).
in Africa, primarily affecting ruminants, which results in high
RVF virus belongs to the genus Phlebovirtts in the family
mortality in Young and adult animals as well as abortions. Since
Bunyaviridae, a genus which contains two serogroups: the
the epidemic in Egypt in 1977, RVF virus has been shown to
sandfly fever viruses and the tick-transmitted uukuviruses
infect humans extensively, leading to several disease manifest-
(Murphy et al., 1995). According to serological data based on
ations including acute febrile illness, hepatitis, encephalitis,
complement fixation, neutralization and haemagglutination
haemorrhagic fever and ocular sequelae (Laughlin et al., 1979).
inhibition tests, the sandfly fever group is composed of at least
Recent RVF epidemics in Mauritania (Digoutte & Peters,
eight antigenic complexes: Bajaru, Candiru, Chilibre, Frijoles,
1989), Madagascar (Morvan et a/., 1992) and Egypt (Arthur et
Punta Toro, Rift Valley fever, Salehabad and an unassigned
al., 1993), as well as the enzootic maintenance of the virus in
complex containing 16 viruses.
Senegal (Zeller et al., 1997), have emphasized the emergence of
The genome of phleboviruses consists of three negative-
this virus as a major public health threat in Africa. With the
sense single-stranded RNA segments designated L (large), M
exception of Egypt, RVF virus is confined to sub-Saharan
(medium) and S (small). The L segment codes for the L protein,
Africa, where it is widespread and maintained in endemic/
the viral RNA polymerase. The M segment codes for a
enzootic and epizootic forms (Meegan & Bailey, 1989). The
precursor to the envelope glycoproteins GI and G2 and two
survival of RVF virus during inter-epizootics is believed to
non-structural proteins of 14 and 78 kDa. The S segment codes
r-.-~.~
for nucleocapsid protein N and the non-structural protein NS,
by an ambisense strategy (for reviews see Bouloy, 1991; Elliott
Author for correspondence: Michèle Bouloy.
et al., 1991; Giorgi, 1996; Schmaljohn, 1996).
Fax j-33 1 40 61 31 51. e-mail mbouloy@pasteur.fr
Molecular and biological variability among phleboviruses
The GenBank/EMBL accession numbers of the sequences reported are
has been studied for RVF virus (Battles & Dalrymple, 1988;
Y1 2739-Y12756.
~-~--
--..-~_~-~
Anderson & Peters, 1988; Saluzzo et al., 1989; Besselar et al.,
0001-4905 0 1997 SGM
1991) and to a lesser extent for Toscana virus (Schwarz et al.,
Table 1. Characteristics of the RVF virus isolates used in
19%). These reports indicated that among natural isolates of
this study
RVF virus, the antigenic properties of the glycoproteins and
H, human; Ar, arthropod; An, animal: B, buffalo.
the nucleoprotein appeared to be stable, and sequences of the
G2 epitopes in the M segment are relatively conserved. On the
other hand, the Egyptian strains differed from the sub-Saharan
Isolation
strains by their pathogenicity in Wistar-Furth rats. Since the
NS, protein is the most variable protein among phleboviruses
Code
Strain
Year
Source
Origin
(Giorgi et al., 1991), its use as a marker of variability was
S N S
Smithburn
1944 Strain
Uganda
investigated. The extent of conservation of the NS, gene
Entebbe
among RVF virus isolates is not known. This question was
Ar UG 55
Lunyo
1955 Mosquito
Uganda
raised by the existence of the naturally occurring avirulent
Ar CAR 69
Ar B 1976
1 9 6 9 Mosquito
CAR
isolate ‘clone 13’ (Muller ef al., 1995), which harbours a
H EGY 77
ZH 548
1977 Human
Egypt
prominent deletion in the NS, region. This suggests that this
MPIZ”
MP12
1977 Strain ZH 548
Egypt
domain is under weakly constraining evolutionary forces,
B EGY 93
B EGY 93
1993 Buffalo
Ewpt
H EGY 93
H EGY 93
1993 Human
Eupt
resulting in the potential for accumulation of substitutions in
Ar MAD 79
Ar Mg 811
1979 Mosquito
Madagascar
this region. A pane1 of 18 strains of RVF virus, collected over
An MAD 91
An Mg 990
1991 Bovine
Madagascar
a period of 38 years and isolated from various hosts in different
HI MAU 87
H D 47502
1987 Human
Rosso
countries, was selected and the NS, gene was amplified by
(October 28)
(Mauritania)
RT-PCR and sequenced. This analysis was carried out to
HZ MAU 87
H D 47311
1987 Human
Yeur
(October 24)
Massène
establish the relatedness between strains as well as to gain
(Mauritania)
insights into the conserved motifs of this protein, the function
H3 MAU 87
H D 47408
1987 Human
Tegmaline
of which is still undetermined.
(October 27)
(Mauritania)
H4 MAU 87
H D 48255
1987 Human
Rosso
(November 11) (Mauritania)
An GUI 84
An K 6087
1984 Bat
Kindia
Methods
(Guinea)
W Virus propagation and RNA extraction. The origins of RVF
Ar BUR 84
Ar D 38457
1984 Mosquito
Burkina
virus isolates used in this study are shown in Table I. Viruses were
Faso
propagated in Vero cells cultivated in Leibovitz 15 growth medium
Ar SEN 84
Ar D 38661
1984 Mosquito
Kedougou
supplemented with 5 % foetal calf serum (Gibco BRL). At 48 to 72 h post-
(Senegal)
infection, infected cultures showed cytopathic effects and the presence of
Ar SEN 93
Ar D 104769 1993 Mosquito
Ferlo
(Senegal)
viral antigens could be detected by an indirect immunofluorescence
An D 106417 1993 Bovine
Kolda
assay. Cytoplasmic extracts were obtained by the NP40 lysis method and
An SEN 93
(Senegal)
total RNA was extracted with phenol-chloroform (1: 1) and precipitated
with ethanol as previously described (Bouloy et al., 1984).
n RT-PCR and sequencing procedures. Synthesis of first strand
’ Laboratory attenuated strain derived from a wild strain.
complementary DNA was primed with oligodeoxynucleotide NS3a
SNS, Smithburn neurotropic strain.
(Muller ef nl., 1995) after denaturation of the RNA with 0.0 L M methyl
mercuric hydroxide for 10 min at room temperature. Methyl mercuric
hydroxide was then complexed with 2 ~1 of 0.7 M p-mercaptoethanol,
were done with the primers used for PCR amplification. Most of the
and reverse transcription was carried out with 100 ng of primer in a final
sequences were read on the two strands. The sequences were deposited
volume of 20 ~1 containing 25 mM Tris-HC1 pH 8.3,25 mM KCI, 5 mM
in the EMBL database (accession numbers Y12739-Y12756).
MgCl,, 0.2 mM of each dNTP (Pharmacia) and 4 units of AMV reverse
transcriptase (Promega).
n Sequence alignments and phylogenetic analysis. The arnino
A 5 ~1 aliquot of the reverse transcription reaction mix was used for
acid sequences of the NS, protein were aligned using Clustal V (Higgins
enzymatic amplification. PCR was carried out with 2.5 units of Tn9 DNA
et ul., 1992) in order to determine the positions of gaps in the ahgnment
polymerase (Amersham) in a 100 ~1 final volume containing 10 mM
of the corresponding nucleotide sequence data. Phylogenetic trees for the
Tris-HC1 pH 8.3, 50 mM KCl, 1.5 mM MgCI,, 0.2 mM of each dNTP
NS, nucleotide sequences were estimated using the FastDNAml program
and 05 pg of each primer (NS3a and NS2g) (Muller ef ai., ~995).
(Olsen ef nl., 1994). The transition/transversion
substitution rate
Amplification was performed after a denaturation step at 95 “C for 3 min
parameter (ts: tv) was calculated by searching for the ts: tv value which
followed by 35 amplification cycles as follows: denaturation at 95 “C for
gives the highest likelihood. In order to establish the robustness of the
1 min, annealing at 53 “C for 1 min and extension at 72 “C for 1 min.
topologies, bootstrap resampling was done (Swofford, 1993). In addition,
Amplified DNA fragments were purified on an agarose gel with the
given an optimal ts: tv parameter, a11 the trees with log likelihoods (1nL)
Geneclean kit (Bio 101) according to manufacturer’s instructions and
not significantly different from the best maximum likelihood tree were
directly sequenced using the thermal cycle sequencing method with
collected. The level of resolution of the 50% majority rule consensus of
oligodeoxynucleotides labelled at the 5’ end with [g-““P]ATI’
and
this tree pool is then indicative of the topology stability at the tree space
polynucleotide kinase (Adams & Blaskesley, 1991). Sequencing reactions
around the InL maxima. These results were also compared to recon-
structions obtained with distance methods in PHYLIP (Felsenstein, 1993)
host species (arthropods, cattle and humans). The character-
and maximum parsimony with PALJP 3.1 (Swofford, 1993).
istics of the isolates are described in Table 1. The attenuated
The S segment sequences of RVF and sandfly fever Sicilian (SFS)
strains MP12 (Caplen et al., 1985) and Smithburn neurotropic
viruses were retrieved from the GenBank database (accession numbers
strain (SNS; Smithburn, 1949), which were obtained by ce11
X53771 and JO4418, respectively).
passages of the parental virulent strains ZH 548 and Entebbe,
n Rates of evolution and selection of different genes of RVF
respectively, were included in this study.
virus. In order to investigate selective regimes, the number of
Two oligodeoxynucleotides, NS2g and NS3ag (Muller et
synonymous (silent) substitutions per synonymous site (ds) and the
al., 1995), were designed based on the published sequence of
number of non-synonymous substitutions (,or amino acid altering
the MP12 strain and utilized as primers for RT-PCR
changes) per non-synonymous site (dn) were calculated for the nucleotide
data sets using the Jukes-Cantor correction method for distances less
amplification. They allowed amplification of a 601-nucleotide-
than 0.4 substitutions per site and the proportional distance method (p-
long DNA fragment which was obtained for a11 the strains
distance) using the MEGA program (Kumar ef a/., 1993) for distances
analysed and subsequently sequenced. Using these primers,
greater than 04.
clone 13 yielded a 12l-nucleotide-long DNA fragment (Muller
et al., 1995), which was too small to be included in this analysis.
Nucleotide substitutions were scattered throughout the entire
Results and Discussion .-....-_-._ ~
region and, with the exception of clone 13, no base insertion or
Phylogenetic analysis of the RVF virus NS, genes
deletion was observed. Pairwise comparison among these
The S genomic sequences of five phleboviruses, Punta
isolates showed percentage divergence ranging from 0 to
Toro, RVF (MPl2 strain), SFS, Toscana and Uukuniemi viruses,
9.6 % at the nucleotide level which corresponded
to 0 to 9.5 %
have already been published (Ihara et al., 1984; Marriott et ai.,
at the amino acid level.
1989; Simons et al., 1990; Giorgi et al., 1991) and were shown
The phylogenetic tree of the nucleotide sequences was
to utilize an ambisense strategy to code for the NS, protein in
generated using the optimal ts: tv value of 2, which was
the genomic sense and for the N protein in the antigenomic
determined empirically. The SFS virus NS, sequence was used
sense. The N and NS, ORFs represent two potential distinct
as an outgroup to determine the coalescence node for RVF
targets for analysis of the S segment. The deduced N protein
virus sequences and therefore indicate a root for the tree. Using
sequences exhibit homologies ranging from 30 to 54 % (Giorgi
ultrametric methods and excluding the SFS virus branch, a
et al., 1991); however, the NS, gene is much less conserved and
congruent root was obtained for Ar MAD 79 (Fig. 1). Similar
cannot be aligned to analyse phylogenetic relationships among
trees were obtained with the maximum likelihood distance
different phleboviruses. Noteworthy is that the function of the
methods and maximum parsimony (not shown). Apart from Ar
protein in the virus life cycle has not yet been determined and
MAD 79, the isolates are distributed in at least two major
that its localization in infected cells, as well as its presence in
lineages, Egyptian and sub-Saharan, with the latter lineage
viral particles, varies depending on the virus. With the
divided into two clusters, Ia and Ib. Al1 the branchings of the
exception of the NS, protein from RVF virus, which has been
tree are highly supported by bootstrap values. Group Ib strains
reported to form filaments in the nuclei of infected cells
were isolated only in West Africa (Senegal, Mauritania, Guinea
(Struthers & Swanepoel, 1982; Struthers et ai., 1984), the NS,
and Burkina Faso), whereas group Ia appears more hetero-
protein of the phleboviruses studied was detected in the
geneous, including strains from Uganda, Madagascar, Central
cytoplasm. In the case of Uukuniemi virus, Simons ef al. (1992)
African Republic as well as strains from Senegal and
demonstrated that the protein was associated with the 40s
Mauritania.
ribosomal subunit and not detected in mature particles. In
The existence of the Egyptian lineage is in good agreement
contrast, for Punta Toro virus, the term ‘non-structural’ does
with a set of data indicating that the strains from the 1977
not seem appropriate since the protein was found associated
outbreak could be distinguished from the sub-Saharan strains
with the ribonucleoproteins in mature particles (Overton et al.,
by several criteria: (i) the N protein of the Egyptian strains
1987).
possesses a specific epitope recognized by the RlP2E7
The diversity of the NS, gene among phleboviruses raised
monoclonal antibody (Saluzzo et al., 1989); (ii) the Egyptian
the question of the extent of conservation of the NS, gene
strains were also highly pathogenic for Wistar-Furth rats; (iii)
among strains of a particular virus. This was underlined by the
these strains were resistant to rat interferon; and (iv) they did
existence of clone 13, a natural, attenuated isolate of RVF virus
not form plaques in primary rat hepatocytes (Anderson &
which harbours a large deletion in the NS, region and replicates
Peters, 1988). As expected, the MP12 strain is closely related
in various hosts including mosquitoes, suggesting that the
to its parent, H EGY 77, but surprisingly, is placed in a more
deleted region represents a potentially variable domain (M&er
ancestral position. The use of a mutagen to obtain MP12 may
et al., 1995). Therefore, we decided to analyse the NS, gene of
have altered the mutation regime in a very biased way for the
various strains of RVF virus and selected a pane1 of 18 RVF
NS, sequence. Moreover, this analysis shows that the strains
virus strains collected over 38 years in eight countries under
isolated in 1977 and 1993 in Egypt are very close to each
epidemic and endemic conditions and isolated from a variety of
other, suggesting that the same virus remained endemic for a
--..
-
zebu in southern Senegal (Kolda, Casamance). Interestingly,
this strain clustered with those from the neighbouring countries
of Guinea and Burkina Faso. This suggests that Senegal is
a
divided in two areas as far as RVF virus is concemed: the
Sahelian zone where southem Mauritanian and northem
Senegalese strains are circulating, probably through livestock
and mosquitoes, and the Sudano-Guinean zone where south-
em Senegalese strains are in contact with those from bordering
I ArBUR84
countries.
Regarding its geographical origin (i.e. from mosquitoes
collected in Kedougou, a place in southem Senegal close to
H4 MAU 87
Guinea and Mali), Ar SEN 84 would be expected to cluster
Ib
with An GUI 84 which was isolated from a bat in Kindia
- H3MAU 87
(Guinea), a bioclimatic area similar to Kedougou and in the
An SEN 93
vicinity of the Senegalese border. Its presence within the
An GUI 84
Egyptian lineage,
although puzzling, may be due to a
reassortment event. Indeed, the circulation of phylogenetically
distinct strains in the same region and at the same period
provides ideal conditions for a host to be co-infected and
II
generate reassortants as occurs naturally for other members of
the bunyavims family like hantaviruses (Li et al., 19%;
L ArSEN84
l1 Henderson et al., 1995) and in other families of viruses with
/ AI MAD 79
segmented genomes (Murphy & Webster, 1990). Further work
Will be necessary to test whether some of these viruses result
from reassortment.
Fig. 1. Phylogenetic tree for the NS, gene of RVF virus isolates. Values
Finally, the strains isolated in Madagascar in 1991 and 1979
above branches indicate the level (%) of bootstrap support using
belong to different lineages; strain Ar MAD 79 is closely
maximum parsimony after 500 iterations. Values below branches indicate
the number of times a given node was observed on a major@ rule
related to the Egyptian group. This relatedness was noted by
consensus of 50 trees with equivalent likelihood (InL). Branch lengths
Morvan et al. (1992) who analysed the antigenic properties of
shown are proportional to the number of substitutions per 100 residues.
the N protein. The topology of the tree reported in Fig. 1
The rooting shown here was determined by the inclusion of the SFS virus
NS, sequence as the outgroup.
indicates that the virus isolated in the island in 1979 occupies
the most ancestral position and shares a common ancestor with
the Egyptian strains. For these reasons, Ar MAD 79 was
long period in a large area or that the viruses circulating in
included in the Egyptian lineage. On the other hand, strain An
1977 and 1993 were introduced from the same source.
MAD 91 appears more closely related to the eastem or central
Grouping based on geographical origin applies also to the two
African strains. As speculated by Morvan et al. (~EU), it is
strains from Uganda, Ar UG 55 and SNS, a laboratory-
possible that during this latter outbreak, RVF virus had been
attenuated strain derived from the Entebbe strain isolated in
introduced to the island from the southem or eastem toast of
Uganda in 1944. Nevertheless, some of the groupings,
Africa through livestock commerce.
especially those of some Mauritanian and Senegalese strains,
are unexpected. The phylogenetic tree indicates that two
phylogenetically different viruses corresponding to clusters Ia
Insights concerning the conserved amino acids in the
and Ib were circulating during the Mauritanian outbreak in
NS, protein
1987. However, the strains analysed originated from three
Six potential phosphorylation sites for casein kinase 1 and
places less than 50 km apart and from cases occurring over less
2, corresponding to the motifs S/T-X-R/K and S/T-X-X/-D/E,
than I month (see Table 1). It should be noted that strain HI
were identified in the NS, protein of the MP12 strain. Al1 of
MAU 87 was isolated from a febrile case, and the others were
them are conserved in the strains SNS, H EGY 77, B and H
isolated from fatal cases. Moreover, HI MAU 87 is closely
EGY 93, and Ar SEN 84, the complete NS, ORF of which was
related to the Senegalese strain Ar SEN 93 which was isolated
sequenced. Among the other strains studied and within the
from mosquitoes trapped in 1993 in the Sahelian Ferlo, a
region analysed, there are three potential sites at positions 98,
northern region close to the Mauritanian border where the
128 and 149 which are a11 conserved.
This suggests that the
1987 outbreak occurred (Zeller et al., 1997). On the other hand,
protein could be phosphorylated at multiple sites.
the position of strain An SEN 93 within a different cluster
The amino acid sequence alignment also indicates that the
could be explained by its origin; it was isolated from a healthy
five cysteine residues at positions 39,40, 150, 179 and 1% are
conserved. It is possible that these residues form either intra- or
Caplen, H., Peters, C. 1. h Bishop,
D. H. L. (1985). Mutagen-directed
intermolecular disulfide bonds which are important for con-
attenuation of Rift Valley fever virus as a method for vaccine
servation of the three-dimensional structure of the protein and
development. ]ourna~ of General Virology 66, 2271-2277.
the formation of filamentous structures observed in cells
Digoutte, 1. P. & Peters, C. 1. (1989). General aspects of the 1987 Rift
Valley fever epidemic in Mauritania. Research in Virology 140, 27-30.
infected with a11 the strains analysed SO far, except for strain
C13. Also noteworthy is the presence of the RGD motif at
Elliott, R. M., Schmaijohn, C. S. & Colle$ M. S. (1991). Bunyaviridae
genome structure and gene expression. In Current Topics in Microbiology
position 24, conserved in a11 isolates except for An MAD 91 in
and Immuno~ogy,
pp. 91-142. Berlin: Springer-Verlag.
which arginine is replaced by lysine. The existence of such a
Felsenstein, 1. (1993). PHYLIP (phylogeny inference package) version
motif, commonly present in extracellular adhesive proteins, in
3%. Distributed by the Author, Department of Genetics, University of
this intracellular protein is surprising.
Washington, Seattle, USA.
Altogether these data indicate that the structure of the RVF
Giorgi, C. (1996). Molecular biology of phleboviruses. In The Bunyn-
virus NS, gene is conserved and that clone 13 is a very distinct
viridae, pp. 105-128. Edited by R. M. Elliott. New York: Plenum Press.
isolate. The existence of this virus raises the question of
Giorgi, C., Accardi, L., Nicoletti, L., Gro, M. C., Takehara, K., Hilditch,
whether the NS, protein is essential. In an attempt to
C., Morikawa, S. & Bishop,
D. H. L. (1991). Sequences and coding
investigate selective regimes on the NS, gene, ds and dn were
strategies of the S RNAs of Toscana and Rift Valley fever viruses
calculated. The ds:dn ratio was found to be greater than one
compared to those of Punta Toro, Sicilian sandfly fever, and Uukuniemi
viruses. Virology 180, 738-753.
(8.36; P < 0.001) indicating that during evolution more
Henderson, W. W., Monroe, M. C., St Jeor, S. C., Thayer, W. P., Rowe,
synonymous than non-synonymous changes accumulated.
1. E., Peters, C. J. h Nichol, S. T. (1995). Naturally occurring Sin
This suggests the existence of constraints against amino acid
Nombre virus genetic reassortants. Virology 214, 602-610.
changes in the protein.
Higgins, D. G., Bleasby, A. 1. & Fuchs, R. (1992). Clustal V: improved
Finally, the oligonucleotides used for RT-PCR in this
software for multiple sequence alignment. Cumpufer
Applications in the
analysis appear to be good primers for amplifying the NS,
Biosciences 8, 189-191.
specific region of the 18 isolates presented herein as well as
Ihara, T., Matsuura, Y. & Bishop,
D. H. L. (1984). Novel coding
other isolates tested in our laboratory (data not shown).
strategy (ambisense genomic RNA) revealed by sequence analyses of
Therefore, this work defines the basis for a rapid diagnostic test
Punta Toro phlebovirus S RNA. Virology 136, 293-306.
for RVF virus based on RT-PCR.
Kumar, S., Tamura, K. 81 Nei, M. (1993). MEGA: Molecular Evol-
utionary Genetic Analysis, version 1.0. Pennsylvania State University,
University Park, PA 16802, USA.
The authors thank P. Vialat for excellent technical expertise. We are
Laughlin, L. W., Meegan, 1. M., Strausbaugh, L. W., Morens, D. M. &
grateful to Dr H. Bourhy for providing computer facilities and critically
Watten, R. H. (1979). Epidemic Rift Valley fever in Egypt: observations
reading the manuscript. We thank Dr C. Prehaud, A. Billecocq and B. Le
of the spectrum of human illness. Trunsactions
of fhe Royal Sociefy of
Guenno for fruitful discussions.
Tropical Medicine and Hygiene 73, 630-633.
Li, D., Schmaljohn, A. L., Anderson, K. h Schmaljohn, C. S. (1995).
Complete nucleotide sequences of the M and S segments of two
References ~
hantavirus isolates from Califomia: evidence for reassortment in nature
Adams, S. & Blaskesley, R. (1991). Linear amplification DNA
among viruses related to hantavirus pulmonary syndrome. Virology 206,
sequencing. Facrts 13, 56-57.
973-983.
Anderson, G. W., Jr & Peters, C. 1. (1988). Virai determinants of
Linthicum, K. J., Davies, F. G. & Kairo, A. (1985). Rift Valley fever virus
virulence for Rift Valley fever (RVF) in rats, Microbinl Puthogenesis
5,
(family Bunyaviridae,
genus Phlebovirw). Isolations from Diptera collected
247-250.
during an inter-epizootic period in Kenya. Ja~rnal ofHygiene 95, 197-209.
Arthur, R., El-Sharkawy, M. S., Cope, S. E., Botros, B. A., Oun, S.,
Marriott, A. C., Ward, V. K. & Nuttall, P. A. (1989). The S RNA segment
Morrill, J. C., Shope, R. E., Hibbs, R. G., Darwish, M. A. h Imam, 1. 2. E.
of sandfly fever Sicilian virus. Evidence for an ambisense genome.
( 1 9 9 3 ) . Recurrence of Rift Valley fever in Egypt. Lnncef 342,
Virology 169, 341-345.
1149-1150.
Meegan, 1. M. & Bailey, C. 1. (1989). Rift Valley fever. In The
Battles, J. K. & Dalrymple, J. M. (1988). Genetic variation among
Arboviruses: Epidemiology and Ecology, pp. 51-76. Edited by T. P. Monath.
geographic isolates of Rift Valley fever virus. American Journal of Tropical
Boca Raton, Fla.: CRC Press.
Medicine and Hygiene 39, 617-631.
Morvan, J., Rollin, P. E., Laventure, S., Rakotoarivony, 1. h Roux, J.
Besselar, T. G., Blackburn, N. K. h Meenehan, G. M. (1991). Antigenic
(1992). Rift Valley fever epizootic in the central highlands of
analysis of Rift Valley fever virus isolates: monoclonal antibodies
Madagascar. Research in Virology 143, 407-415.
distinguish between wild-type and neurotropic virus strains. Research in
Muller, R., Saluzzo, J. F., Lopez, N., Dreier, T., Turell, M., Smith, J. &
Virology 142, 469-474.
Bouloy, M. (1995). Characterization of clone 13, a naturally attenuated
Bouloy, M. (1991). Bunyaviridue: genome organization and replication
avirulent isolate of Rift Valley fever virus, which is altered in the small
strategies. Advances in Virus Research 40, 235-266.
segment. American jotcrnnl ej Tropical Medicine and Hygiene 53, 405-411.
Bouloy, M., Vialat, P., Girard, M. h Pardigon, N. (1984). A transcript
Murphy, B. R. & Webster, R. G. (1990). Orthomyxoviruses. In Virology,
from the S segment of the Germiston bunyavirus is uncaped and codes
;;es~1-l152.
Edited by B. N. Fields & D. M. Knipe. New York: Raven
for the nucleoprotein and a nonstructural protein. journal of Virology 49,
717-723.
Murphy, F. A., Fauquet, C. M., Bishop,
D. H. L., Ghabrial, S. A., larvis,
A. W., Martelli, G. P., Mayo, M. A. h Summers, M. D. (editors) (1995).
Simons, J. F., Hellman, U. h Pettersson, R. F. (1990). Uukuniemi virus
Family Bmnyuoiridue. In Virtrs Taxonorny. Sirfh Report of fhe Wernrzfional
S RNA segment: ambisense coding strategy, packaging of comp-
Carnmiffee on Taxonomy
nf Vivuses, pp. 300-315. Vienna % New York:
lementary strands into virions and homology to members of the genus
Springer-Verlag.
Ph&rz~irns. ]ourna~ of Viro/ogy 64, 247-255.
Olsen, G. J., Matsuda, H., Hagstrom, R. & Overbeek, R. (1994).
Simons, J. F., Persson, R. & Pettersson, R. F. (1992). Association of
FastDNAml: a tool for construction of phylogenetic trees of DNA
the nonstructural protein NS, of Uukuniemi virus with the 40s ribosomal
sequences using maximum likelihood. Compzzfer Applications in the
subunit. Jortrnn! of I/irdogy 66, 4233-4241.
Biosciences 10, 41-48.
Smithburn, K. C. (1949). Rift Valley fever: the neurotropic adaptation of
Overton, H. A., Ihara, T. & Bishop,
D. H. L. (1987). Identification of the
the virus and the experimental use of this modified virus as a vaccine.
N and NS, proteins coded by the ambisense S RNA of f’unta Toro
Brifish journal of Eq~erimenfal P~~fhology 30, l-16.
phlebovirus using monospecific antisera raised to baculovirus expressed
Struthers, J. K. h Swanepoel, R. (1982). Identification of a major non-
N and NS, proteins. Virolegy 15 7, 338-350.
structural protein in the nuclei of Rift Valley fever virus-infected cell
Saluzzo, 1. F., Anderson, G. W., Hodgson, L. A., Jr, Digoutte, 1. P. &
]ournu~ of Geaernl Virabgy 60, 381-384.
Smith, 1. F. (1989). Antigenic and biological properties of Rift Valley
Struthers, 1. K., Swanepoel, R. & Shepard, S. P. (1984). Protein
fever virus isolated during the 1987 Mauritanian epidemic. Resecrrch in
synthesis in Rift Valley fever virus-infected cell. Virobgy 134, 118-124.
Vkhgy 140, 155-164.
Swofford, D. L. (1993). PAUP: phylogenetic analysis using parsimony.
Schmaljohn, C. S. (1996). Bnnyaviridae: the viruses and their replication.
version 3.X.1. Computer program distributed by the Illinois Natural
In Fields Virology, 3rd edn, pp. 1447-1471. Edited by B. N. Fields, D. M.
History Survey, Champaign, Illinois, USA.
Knipe & P. M. Howley. Philadelphia: Lippincott-Raven.
Zeller, H. G., Fontenille, D., Traore-Lamizana, M., Thiongane, Y. 6
Schwarz, T. F., Jager, G., Gilch, S. & Nithchko, H. (1995). Nested
Digoutte, J. P. (1997). Enzootic activity of Rift Valley fever virus in
RT-I’CR for detection of sandfly fever virus, serotype Toscana, in clinical
Senegal. American Jmrnd of Tropica/ A4edicine and Hygiene 56, 265-272.
specimens, with confirmation by nucleotide sequence analysis. Reseizrch in
Viralogy 146, 355-362.
R e c e i v e d 2 9 A p r i l 1 9 9 7 ; Accepted 1 9 J u n e 1 9 9 7