Download Leroy et. Al. Gabon 96 phylogeny

Journal
of General Virology (2002), 83, 67–73. Printed in Great Britain
..........................................................................................................................................................................................................
SHORT COMMUNICATION
Sequence analysis of the GP, NP, VP40 and VP24 genes of
Ebola virus isolated from deceased, surviving and
asymptomatically infected individuals during the 1996
outbreak in Gabon : comparative studies and phylogenetic
characterization
Eric M. Leroy,1 Sylvain Baize,1 Elie Mavoungou2 and Cristian Apetrei1
1
2
Centre International de Recherches Me! dicales de Franceville, BP 769 Franceville, Gabon
Institut fu$ r Tropenmedizin, Tu$ bingen-Universita$ t, Wilhelmstrasse 27, 72074 Tu$ bingen, Germany
The aims of this study were to determine if the
clinical outcome of Ebola virus (EBOV) infection is
associated with virus genetic structure and to document the genetic changes in the Gabon strains of
EBOV by sequencing the GP, NP, VP40 and VP24
genes from deceased and surviving symptomatic
and asymptomatic individuals. GP and NP sequences were identical in the three groups of patients
and only one silent substitution occurred in the
VP40 and VP24 genes in asymptomatic individuals.
A strain from an asymptomatic individual had a
reverse substitution to the Gabon-94 sequence,
indicating that minor virus variants may cocirculate
during an outbreak. These results suggest that the
different clinical outcomes of EBOV infection do not
result from virus mutations. Phylogenetic analysis
confirmed that Gabon-96 belonged to the Zaire
subtype of EBOV and revealed that synonymous
substitution rates were higher than nonsynonymous substitution rates in the GP, VP40 and VP24
genes. In contrast, nonsynonymous substitutions
predominated over synonymous substitutions in
the NP gene of the two Gabon strains, pointing to
divergent evolution of these strains and to selective
pressures on this gene.
Ebola virus (EBOV) is one of the most virulent human
pathogens, killing up to 70–80 % of patients within 5–7 days
(Khan et al., 1999). EBOV causes rare but fulminating outbreaks
of haemorrhagic fever in equatorial Africa, being transmitted
Author for correspondence : Eric Leroy.
Fax j241 67 72 95. e-mail leroy!cirmf.sci.ga
0001-7958 # 2002 SGM
from person to person via infected body fluids, such as faeces,
vomit and blood (Dowell et al., 1999). EBOV is subdivided into
four subtypes, namely Zaire, Sudan, Co# te dhIvoire and Reston.
EBOV and Marburg virus (MBGV) are members of the family
Filoviridae, order Mononegavirales, which groups together
nonsegmented, negative-stranded RNA viruses, namely filoviruses, paramyxoviruses, rhabdoviruses and bornaviruses
(Mayo & Pringle, 1998 ; Sanchez et al., 1993). The EBOV
genome displays genes arranged linearly on a single negativestranded RNA molecule that encodes the following seven
structural proteins : nucleoprotein (NP), virion structural
proteins (VP) VP35, VP40, glycoprotein (GP), VP30, VP24
and RNA-dependent RNA polymerase (L) (Feldmann et al.,
1993). Viral proteins are thought to play a key role in
host–virus interactions in EBOV infection and the corresponding genes may thus be subject to selective pressure. In
humans, NP and VP40 elicit strong IgG responses (Baize et al.,
1999 ; Leroy et al., 2000). EBOV GP is thought to induce
endothelial cell disruption and cytotoxicity in blood vessels
(Yang et al., 1998) and to mediate entry into the host cell
(Watanabe et al., 2000). EBOV pathogenicity in guinea pigs is
associated with mutations in VP24 (Volchkov et al., 2000).
Moreover, several studies have shown that filovirus subtypes
have different pathogenic potential, causing different mortality
rates and disease severity and variable patterns of haemorrhagic effects. Because of a marked genetic divergence
between subtypes, the lower pathogenic potential of the Sudan
and Reston strains, relative to the Zaire strains, is thought to be
due to mutations in the viral proteins (Bowen et al., 1980 ;
Fisher-Hoch et al., 1992). Taken together, these data suggest
that any mutation of the virus, particularly in GP, NP, VP40 or
VP24, may account for the different clinical outcomes during
natural human EBOV infection and that mutations in any of
these proteins can influence virulence. We determined the
complete open reading frame (ORF) sequences of these genes
in EBOV isolates from deceased patients, surviving symptomatic patients and infected asymptomatic individuals during
GH
E. M. Leroy and others
Table 1. Synthetic oligonucleotides used for PCR and sequencing
Oligonucleotides are shown in the 3h 5h direction and the length (bp) of the fragment amplified from each
gene is also shown. To amplify the GP (GPa and GPb) and NP (NPa and NPb) genes, two overlapping
segments representing almost the entire ORF were used.
Gene
GPa (first-round)
GPb (first-round)
GPa (nested)
GPb (nested)
NPa (first-round)
NPb (first-round)
NPa (nested)
NPb (nested)
VP40 (first-round)
VP40 (nested)
VP24 (first-round)
VP24 (nested)
Sequences 3h
5h
5
GATGAAGATTAAGCCGACAGTGA
GTTGTCTGTTCTGCGGTGATGTTG
TGCAATGGTTCAAGTGCACAGTC
AAGAGATAACTAGATTGATGTCAAA
GTGAGCGTAATCTTCATCTCTCTTA
TTGTTCAACTTGAGTTGCCTCAGAG
CAAGGAAGGGAAGCTGCAGTGTCG
GAATCACATTGGCTATGTTTAAAGC
GACAGTTTCCTTCTCATGC
CCGTTTTCCGAGTAACTCT
CATCCCATTGTTCCATGC
ATTCTTATGAACTTCCATC
TTACAAACTTTTCTTGGAAA
GGCAGTGACGCAGCAGTGA
CATCCCATTGTTCCATGC
ATTCTTATGAACTTCCATC
TACCTCGGCTGAGAGAGTG
ATGCAGGCAATTTGAGGATAAG
6
2174
7
8
5
6
1767
7
8
ATTAACCTTCATCTTGTAAACGTTG
ATTGATCAATTAAAAGTGTCTCCTC
ACATCACTTTGAGCGCCCTCA
AGCTGGCTTACAGTGAGGATT
GCGCAAGGTTTCAAGGTTGAA
TTGAGTCAGCATATATGAGTT
the 1996 Ebola outbreaks in Gabon. The aims of this study
were to investigate if genetic modifications in any of these four
proteins could explain the different clinical outcomes in humans
and to examine the phylogenetic features and profiles of the
Gabon-94 and -96 strains.
Blood samples were obtained from nine individuals selected
randomly during the EBOV outbreak in Booue! , Gabon. These
nine individuals comprised three symptomatic patients who
died, three symptomatic patients who survived (Baize et al.,
1999) and three asymptomatic infected individuals (Leroy
et al., 2000). In each group, one viral gene sequence was
determined in a sample obtained at the beginning of the
epidemic, while the other two were determined from samples
obtained later during the epidemic (Georges-Courbot et al.,
1997). Samples were obtained with the patientsh verbal,
informed consent and all specimens were drawn and manipulated according to the WHO guidelines on virus haemorrhagic
fever agents in Africa (WHO, 1985). Blood samples, collected
on EDTA, were transported on ice to CIRMF (Centre
International de Recherches Me! dicales de Franceville, Gabon)
for analysis. Peripheral blood mononuclear cells (PBMCs) were
separated from whole blood by Ficoll–diatrizoate densityGI
Length (bp)
6
5
1069
7
8
gradient centrifugation. Total RNA was extracted from
PBMCs and first-strand cDNA was synthesized as described
previously (Leroy et al., 2000). Amplifications of cDNA were
carried out as described previously (Leroy et al., 2000) using
the primers listed in Table 1. PCR products were then excised
from the gel and extracted using the QIAquick Gel Extraction
kit (Qiagen). VP24 fragments were sequenced on an ALF
Express DNA sequencer (Pharmacia Biotech) using the Autocycle 200 Sequencing kit (Pharmacia Biotech). GP, NP and
VP40 fragments were sequenced by ACTgene Laboratories
(France) using the ABI 373A or 377 Automated Sequencer
(Applied Biosystems). The CLUSTAL W (Thompson et al.,
1994) profile alignment option was used to align the newly
sequenced GP and NP genes of EBOV Gabon-96 to the
filovirus nucleotide sequences of GP and NP that are available
in GenBank. Gap-containing sites were removed prior to all
analysis. Pairwise distances were calculated using the
DNADIST program of the PHYLIP package (Felsenstein,
1993) using the Kimura two-parameter model of nucleotide
substitution. Phylogenetic trees were constructed using both
the neighbour-joining (NJ) and the maximum-likelihood (ML)
methods. All phylogenetic trees were inferred using PAUP
Evolution of Ebola virus in Gabon
(Swofford, 1998). The sequences of the GP, NP, VP40 and
VP24 genes of the EBOV strain isolated during the 1996
epidemic in Booue! , Gabon are available from GenBank under
accession numbers AY058898, AY058895, AY058896 and
AY058897, respectively.
PCR amplifications yielded DNA fragments of 2174 (GP),
1767 (NP), 1069 (VP40) and 853 (VP24) bp in size. Under the
same conditions, no EBOV DNA was amplified from cells of
uninfected humans. No differences in the nucleotide sequences
of the GP and NP genes (2174 and 1767 nucleotides,
respectively) were found among the three categories of
patients. The VP40 gene showed only one synonymous
substitution (C T) in strains from the three asymptomatic
individuals compared to those from the symptomatic individuals and the Zaire-76 and -95 strains. Analysis of the VP24
gene only revealed a G A reverse mutation to the Gabon94 strain in one asymptomatic individual. We then sequenced
VP24 sequences from two more asymptomatic subjects (five in
total) and found that these two individuals had identical
sequences to those of the six symptomatic patients and the
other two asymptomatic individuals.
The deduced GP amino acid sequences were compared to
those of all known EBOV sequences. The mean genetic
diversity between Booue! -96 and the other Zaire subtype
strains was much lower (1–2 %) than that between Booue! -96
and the other three subtypes (42–71 %), at both the nucleotide
and amino acid levels, suggesting that Booue! -96 belongs to
the Zaire subtype (Fig. 2). Analysis of multialignment
sequences in the sequenced region of the GP gene (2174 bp)
showed only 36 nucleotide substitutions and 15 amino acid
changes between Booue! -96 and Mayinga-76 (Fig. 1a). Most of
these mutations were located in the middle of the GP gene,
which encodes a variable region that seems to be subtypespecific. In this hypervariable region of 180 nucleotides (60
amino acids), we have observed 14 nucleotide substitutions
(including 10 nonsynonymous) between Booue! -96 and
Mayinga-76 and three nucleotide substitutions (all nonsynonymous) between the two Gabon strains. Almost 40 % of
substitutions are located in less than 9 % of the GP gene
sequence. The two Gabon strains form a subcluster within the
Zaire subtype (Fig. 1). The nucleotide distance in the GP gene
is only 0n39 % (eight substitutions) between Booue! -96 and
Gabon-94, while the nucleotide distance with Mayinga-76 and
Kikwit-95 is 1n79 % (36 substitutions) and 1n04 % (21 substitutions), respectively. At the amino acid level, the genetic
distance between Booue! -96 and Gabon-94 is only 0n99 %,
while the genetic distance with Mayinga-76 and Kikwit-95 is
2n14 % and 0n99 %, respectively. The sequenced region of the
GP gene presented in Fig. 1(a) contains 18 N-glycosylation
sites. Only one site, located in the hypervariable region, had
changed from N–D–S to N–A–S between the Zaire and
Gabon strains.
Amino acid alignment (Fig. 1b), phylogenetic tree construction (Fig. 2) and genetic distance calculations for the NP
gene and its product were consistent with those of GP,
showing that the Booue! -96 strain belongs to the Zaire subtype.
However, the genetic diversity observed between EBOV NP
subtypes was lower than that found in the analysis of GP. The
mean nucleotide distances in the NP gene between the Zaire
and Sudan subtypes are around 30 %, compared to 70 % in the
GP gene. The mean amino acid distances in NP between the
same subtypes are around 45 %, compared to 65–66 % in GP.
The variability between subtypes is located mainly in the Nterminal part of the NP gene (Fig. 1b). Substitutions were not
located in a particular region but were observed throughout
the NP gene and its product (Fig. 1b). NP amino acid distances
between the Zaire and Sudan subtype strains are much higher
than nucleotide distances (45 versus 30 %), in contrast with
those in the GP gene (66 versus 72 %). In contrast with GP,
genetic distances in the NP gene between Booue! -96 and
Gabon-94 are similar to, or higher than, those between Booue! 96 and the other Zaire subtype strains. This nucleotide distance
is 0n91 % (18 substitutions) between Booue! -96 and Gabon-94,
0n74 % between Booue! -96 and Mayinga-76 and only 0n51 %
between Booue! -96 and Kikwit-95. Genetic variability in NP
between the two Gabon strains is high (1n27 %), relative to that
between Booue! -96 and Mayinga-76 or Kikwit-95 (0n95 and
0n63 %, respectively). This higher variability in NP between the
Gabon strains compared with other viral proteins seems to be
a feature of the ‘Gabon cluster’.
One goal of this study was to determine whether the
clinical outcome of EBOV infection is associated with a
particular virus genetic structure. No difference was found in
the nucleotide sequences of the GP and NP genes among
strains from fatalities, survivors and asymptomatic individuals.
Furthermore, we identified only one synonymous substitution
in the VP40 gene and another in the VP24 gene between
symptomatic patients and asymptomatically infected individuals. Taken together, these results show that GP, NP, VP40
and VP24 from the three categories of patients are identical
and suggest that fatalities, survivors and asymptomatic
individuals were infected with the same virus. Thus, the
different clinical outcomes of EBOV infection in humans does
not seem to be due to virus mutations. These data are
consistent with those obtained during the Kikwit outbreak in
1995, when no substitutions were identified in the most
divergent region of the GP gene (249 nucleotides) between
isolates from fatalities and isolates from survivors. Moreover,
no substitutions were identified in the same GP gene region of
isolates from convalescing patients (Rodriguez et al., 1999).
The strong conservation of virus genetic structure observed in
the three groups of EBOV-infected individuals in our study
suggests that differences in clinical outcome are related to
host immune reactivity. Despite this high genetic stability,
sequencing of the genes encoding L, VP30 and VP35 will be
necessary to confirm that the three groups of individuals were
infected by the same virus. Also, these proteins could also be
involved in EBOV pathogenesis, as VP35 has been described
GJ
E. M. Leroy and others
(a)
Fig. 1. For legend see facing page.
HA
Evolution of Ebola virus in Gabon
(b)
Fig. 1. (a) Predicted amino acid sequences of the GP gene of all known EBOV strains. The reference sequence of EBOV
(Mayinga-76) and that of the Yambuku outbreak strain in the Democratic Republic of the Congo (DRC) in 1976 are shown at
the top. The N-glycosylation sites, mapping as NXT (where X can be any amino acid residue except proline or tryptophan), are
boxed. Amino acid changes between the reference sequence and Booue! -96 are shaded in grey. (b) Predicted amino acid
sequences in the NP gene of all known EBOV strains. The reference sequence of EBOV (Mayinga-76) and that of the Yambuku
outbreak strain in DRC are shown at the top. Amino acid changes between the reference sequence and Booue! -96 are shaded
in grey. Dots indicate residues that are identical to the reference sequence and dashes indicate gaps introduced for optimal
alignment. Amino acids are numbered according to their position in the sequence.
as a type I interferon antagonist (Basler et al., 2000) and VP30
and L are part of the replicase complex (Muhlberger et al.,
1999).
Phylogenetic analysis of the VP24 gene showed only one
and eight substitutions in the Booue! -96 sequences from the
three deceased patients, the three survivors and the four
asymptomatic individuals relative to Gabon-94 and Zaire-76,
respectively, indicating a low genetic variability in this gene.
Interestingly, another asymptomatic individual had exactly the
same VP24 sequence as the Gabon-94 strain. This indicates
that several virus variants can cocirculate during the same
Ebola outbreak and that the variant corresponding to this
asymptomatic individual is minor. The low genetic variability
of EBOV was shown further by analysis of the NP and GP
genes. The high mutation rate between Booue! -96 and the Co# te
dhIvoire, Sudan and Reston subtypes, and the low mutation
rate between Booue! -96 and the other Zaire subtype strains
demonstrates that Booue! -96 belongs to the Zaire subtype.
One feature of members of the family Filoviridae is the
contrast between the high genetic diversity between subtypes
and the low intrasubtype variability. Indeed, the Booue! strain
diverges from other strains of the Zaire subtype by only 1–2 %,
despite the fact that Booue! -96 and Zaire-76\95 were isolated
20 years apart and more than 3000 km apart. This is consistent
with another study of the most variable 249 nucleotide region
of the GP gene (Rodriguez et al., 1999). In contrast to most
RNA viruses, EBOV is characterized by high genetic stability,
which may be due to four main factors : low error rate of RNA
polymerase, slow replication in the natural host, small number
of natural hosts and weak immunological pressure. In contrast
HB
E. M. Leroy and others
(a)
(b)
Fig. 2. Phylogenetic trees of EBOV strains for the (a) GP and (b) NP gene sequences presented in Fig. 1 (a) (GP) and (b)
(NP). Analysis was based on amino acid sequences and the trees were constructed by the ML method. Distance matrix-based
trees were constructed with the NJ method (Saitou & Nei, 1987) using the HKY85 model of nucleotide substitution.
Confidence estimates were assessed using 1000 bootstrap replicates (implemented using SEQBOOT, NEIGHBOR and
CONSENSE from the PHYLIP package), with JTT distance matrices generated using PUZZLEBOOT (http ://www.tree-puzzle.de).
ML analysis was carried out with the PROTML program of the MOLPHY package using the JTT empirical matrix model of amino
acid substitution (Jones et al., 1992) ; the trees presented are those with the highest likelihood score. ML trees were also
constructed with the NUCML program of the MOLPHY package, using the HKY85 model of nucleotide substitution
incorporating a transition/transversion ratio of four and an α shape parameter (using eight discrete categories of rate variation)
for a gamma distribution of rate heterogeneity among sites, both estimated on the NJ tree. The NJ tree topology was used as
the starting tree in a heuristic search using tree bisection–reconnection branch swapping to search for the ML tree. MBGV
sequences (Musoke and Popp strains) were used as the outgroup strains. Vertical branches are for clarity only ; the lengths of
the horizontal branches are proportional to the single-base changes. Sequences used in tree construction are available from
GenBank. MBGV strains are indicated as MAR. EBOV strains are indicated as EBO-Cl (Co# te dhIvoire), -Z (Zaire), -S (Sudan) and
-R (Reston).
with other genes, the rate of nonsynonymous substitutions
between the two Gabon strains in the NP gene is significantly
higher than that between Booue! -96 and Mayinga-76 and
between Booue! -96 and Kikwit-95 (1n27 versus 0n95 and
0n63, respectively). This comparatively high level of nonsynonymous substitutions between the two Gabon strains in
the NP gene relative to other genes points to selective
constraints on NP. It has been shown that when nonsynonymous substitutions predominate over synonymous
substitutions in a viral gene, this gene is subjected to positive
selection (Hughes & Nei, 1988). This positive Darwinian
selection is driven generally by host immune pressure (Fitch
et al., 1991 ; Hughes & Nei, 1989). This particular NP pattern
observed in the two Gabon strains of EBOV differs from the
evolutionary patterns of most RNA virus genes, in which
synonymous substitutions predominate strongly over nonsynonymous substitutions (the neutral theory of molecular
HC
evolution) (Gojobori et al., 1990). Two main hypotheses may
be proposed to explain this particular NP pattern : (i) the two
Gabon strains may have diverged much earlier than suggested
by GP analysis (Suzuki & Gojobori, 1997) and (ii) the two
Gabon strains diverged recently, but the separation was
accompanied by a change in the natural host (species or genus),
leading to different immunological pressures on the NP gene.
Nevertheless, the coexistence of two different strains, without
apparent genetic recombination, in similar geographical and
temporal conditions may be possible if the viruses have their
own ecological niche. These features of EBOV genetic structure
may have implications for the development of treatments and
vaccines against EBOV infection. The NP gene pattern of
nonsynonymous nucleotide substitutions suggest that NP is
subject to functional constraints that induce relatively rapid
molecular evolution under immunological pressure. For this
reason, the NP gene is a potential candidate target for antiviral
Evolution of Ebola virus in Gabon
drugs. Likewise, the extremely high genetic stability in each
EBOV subtype may have implications for the development
of vaccines.
We thank V. Volchkov for helpful support on technical procedures.
We thank also P. Roques for comments on the manuscript. CIRMF is
supported by the Government of Gabon, Total-Fina-Elf Gabon and le
Ministe' re de la Coope! ration Franiaise.
References
Baize, S., Leroy, E. M., Georges-Courbot, M.-C., Capron, M., LansoudSoukate, J., Debre! , P., Fisher-Hoch, S. P., McCormick, J. B. & Georges,
A. J. (1999). Defective humoral responses and extensive intravascular
Jones, D. T., Taylor, W. R. & Thornton, J. M. (1992). The rapid
generation of mutation data matrices from protein sequences. Computer
Applications in the Biosciences 8, 275–282.
Khan, A. S., Tshioko, F. K., Heymann, D. L., Le Guenno, B., Nabeth, P.,
Kerstie$ ns, B., Fleerackers, Y., Kilmarx, P. H., Rodier, G. R., Nkuku, O.,
Rollin, P. E., Sanchez, A., Zaki, S. R., Swanepoel, R., Tomori, O.,
Nichol, S. T., Peters, C. J., Muyembe-Tamfum, J. J. & Ksiazek, T. G.
(1999). The reemergence of Ebola hemorrhagic fever, Democratic
Republic of the Congo, 1995. Journal of Infectious Diseases 179 (Suppl. 1),
S76–S86.
Leroy, E. M., Baize, S., Volchkov, V. E., Fisher-Hoch, S. P., GeorgesCourbot, M.-C., Lansoud-Soukate, J., Capron, M., Debre! , P.,
McCormick, J. B. & Georges, A. J. (2000). Human asymptomatic Ebola
apoptosis are associated with fatal outcome in Ebola virus-infected
patients. Nature Medicine 5, 423–426.
infection and strong inflammatory response. Lancet 355, 2210–2215.
Mayo, M. A. & Pringle, C. R. (1998). Virus taxonomy – 1997. Journal of
General Virology 79, 649–657.
Basler, C. F., Wang, X., Muhlberger, E., Volchkov, V., Paragas, J.,
Klenk, H. D., Garcia-Sastre, A. & Palese, P. (2000). The Ebola virus
Muhlberger, E., Weik, M., Volchkov, V. E., Klenk, H. D. & Becker, S.
(1999). Comparison of the transcription and replication strategies of
VP35 protein functions as a type I IFN antagonist. Proceedings of the
National Academy of Sciences, USA 97, 12289–12294.
Marburg virus and Ebola virus by using artificial replication systems.
Journal of Virology 73, 2333–2342.
Bowen, E. T., Platt, G. S., Lloyd, G., Raymond, R. T. & Simpson, D. I.
(1980). A comparative study of strains of Ebola virus isolated from
Rodriguez, L. L., De Roo, A., Guimard, Y., Trappier, S. G., Sanchez, A.,
Bressler, D., Williams, A. J., Rowe, A. K., Bertolli, J., Khan, A. S.,
Ksiazek, T. G., Peters, C. J. & Nichol, S. T. (1999). Persistence and
southern Sudan and northern Zaire in 1976. Journal of Medical Virology 6,
129–138.
Dowell, S. F., Mukunu, R., Ksiazek, T. G., Khan, A. S., Rollin, P. E. &
Peters, C. J. (1999). Transmission of Ebola hemorrhagic fever : a study
of risk factors in family members, Kikwit, Democratic Republic of the
Congo, 1995. Journal of Infectious Diseases 179 (Suppl. 1), S87–S91.
Feldmann, H., Klenk, H. D. & Sanchez, A. (1993). Molecular biology
and evolution of filoviruses. Archives of Virology 7, 81–100.
Felsenstein, J. (1993). PHYLIP : Phylogeny Inference Package, version
3.5c. University of Washington, Seattle, WA, USA.
Fisher-Hoch, S. P., Brammer, T. L., Trappier, S. G., Hutwagner, L. C.,
Farrar, B. B., Ruo, S. L., Brown, B. G., Hermann, L. M., Perez-Oronoz,
G. I., Goldsmith, C. S., Hanes, M. A. & McCormick, J. B. (1992).
Pathogenic potential of filoviruses : role of geographic origin of primate
host and virus strain. Journal of Infectious Diseases 166, 753–763.
Fitch, W. M., Leiter, J. M., Li, X. Q. & Palese, P. (1991). Positive
Darwinian evolution in human influenza A viruses. Proceedings of the
National Academy of Sciences, USA 88, 4270–4274.
Georges-Courbot, M.-C., Sanchez, A., Lu, C.-Y., Baize, S., Leroy, E. M.,
Lansoud-Soukate, J., Te! vi-Benissan, C., Georges, A. J., Trappier, S. G.,
Zaki, S. R., Swanepoel, R., Leman, P. A., Rollin, P. E., Peters, C. J.,
Nichol, S. T. & Ksiazek, T. G. (1997). Isolation and phylogenetic
characterization of Ebola viruses causing different outbreaks in Gabon.
Emerging Infectious Diseases 3, 59–62.
Gojobori, T., Moriyama, E. N. & Kimura, M. (1990). Molecular clock of
viral evolution, and the neutral theory. Proceedings of the National Academy
of Sciences, USA 87, 10015–10018.
Hughes, A. L. & Nei, M. (1988). Pattern of nucleotide substitution at
major histocompatibility complex class I loci reveals overdominant
selection. Nature 335, 167–170.
Hughes, A. L. & Nei, M. (1989). Nucleotide substitution at major
histocompatibility complex class II loci : evidence for overdominant
selection. Proceedings of the National Academy of Sciences, USA 86,
958–962.
genetic stability of Ebola virus during the outbreak in Kikwit, Democratic
Republic of the Congo, 1995. Journal of Infectious Diseases 179 (Suppl. 1),
S170–S176.
Saitou, N. & Nei, M. (1987). The neighbor-joining method : a new
method for reconstructing phylogenetic trees. Molecular Biology and
Evolution 4, 406–425.
Sanchez, A., Kiley, M. P., Holloway, B. P. & Auperin, D. D. (1993).
Sequence analysis of the Ebola virus genome : organization, genetic
elements, and comparison with the genome of Marburg virus. Virus
Research 29, 215–240.
Suzuki, Y. & Gojobori, T. (1997). The origin and evolution of Ebola and
Marburg viruses. Molecular Biology and Evolution 14, 800–806.
Swofford, D. L. (1998). PAUP : Phylogenetic Analysis Using Parsimony,
version 4.0. Sunderland : Sinauer Association.
Thompson, J. D., Higgins, D. G. & Gibson, T. J. (1994). CLUSTAL W :
improving the sensitivity of progressive multiple sequence alignment
through sequence weighting, position-specific gap penalties and weight
matrix choice. Nucleic Acids Research 22, 4673–4680.
Volchkov, V. E., Chepurnov, A. A., Volchkova, V. A., Ternovoj, V. A. &
Klenk, H. D. (2000). Molecular characterization of guinea pig-adapted
variants of Ebola virus. Virology 277, 147–155.
Watanabe, S., Takada, A., Watanabe, T., Ito, H., Kida, H. & Kawaoka,
Y. (2000). Functional Importance of the coiled-coil of the Ebola virus
glycoprotein. Journal of Virology 74, 10194–10201.
WHO (1985). Recommendations for management of viral haemorrhagic
fevers in Africa. Meeting of the World Health Organization, Sierra Leone,
1985.
Yang, Z.-Y., Delgado, R., Xu, L., Todd, R. F., Nabel, E. G., Sanchez, A.
& Nabel, G. J. (1998). Distinct cellular interactions of secreted and
transmembrane Ebola virus glycoproteins. Science 279, 1034–1037.
Received 5 July 2001 ; Accepted 31 August 2001
HD