Dengue Neurovirulence in Mice:
Identification of Molecular Signatures in the E and NS3 Helicase Domains
Juliano Bordignon,1Daisy M. Strottmann,1Ana Luiza P. Mosimann,1Christian M. Probst,1 Vanessa Stella,1Lucia Noronha,2Sı´lvio M. Zanata,3and Claudia N. Duarte dos Santos1*
1Instituto de Biologia Molecular do Parana´/FIOCRUZ, Rua Prof Algacyr Munhoz Ma´der, Curitiba, Parana´, Brazil
2Laborato´rio de Patologia Experimental/Pontifı´cia Universidade Cato´lica do Parana´, PUC-PR, Rua Imaculada Conceic¸a˜o, Prado Velho, Curitiba, Parana´, Brazil
3Departamento de Patologia/Universidade Federal do Parana´, Rua Cel. Francisco H. dos Santos, Jardim das Ame´ricas, Curitiba-PR, Brazil
Recent observations indicate that the clinical profile of dengue virus (DENV) infection is changing, and that neurological manifestations are becoming frequent. The neuro pathogenesis of dengue, and the contribution of viral and host factors to the disease are not well understood.
To define the amino acid substitutions in DENV potentially implicated in the acquisition of a neurovirulent phenotype we used a murine model to characterize two neuroadapted strains of DENV-1, FGA/NA a5c (previously obtained), and FGA/NA P6 (recently obtained). Only three amino acid substitutions were identified in the neurovirulent strains, mapping to the E and NS3 helicase domains. These mutations enhanced the ability of neuroadapted viral strains to replicate in the CNS of infected mice, causing extensive damage with leptomeningitis and encephalitis. J. Med. Virol. 79:1506–1517, 2007. ß2007 Wiley-Liss, Inc.
KEY WORDS: envelope protein; RNA heli- case; molecular markers; pri- mary neurons; murine model
INTRODUCTION
Dengue virus (DENV) is an arthropod-borne fla- vivirus that belongs to the Flaviviridae family. The four serotypes of DENV (1–4) are transmitted to humans by the mosquito vector Aedes aegypti. The disease is endemic in more than 100 countries and 2.5 billion people are presently at risk of infection [Gubler, 1998; Guirakhoo et al., 2004]. Epidemics with a high frequency of severe dengue cases continue to expand geographically in Asia and South America. Despite its increased impact, the pathogenesis of dengue disease is poorly understood.
The DENV genome consists of a 11 kb single- stranded positive sense RNA that contains a single open reading frame flanked by two untranslated regions (50and 30UTR). The genomic RNA encodes a polyprotein precursor that is co- and post-translationally processed by host-cell and virus-specific proteases to yield the structural proteins C, prM (precursor of M protein) and E, and the non-structural (NS) proteins NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5, which are essential for viral replication [Chambers et al., 1990; Heinz et al., 1994; Lindenbach and Rice, 2001]. The E protein mediates primary attachment of the virus to its target cell and is thus involved in host-cell tropism, and potentially in pathogenesis. After virus endocytosis, a conformational change in the E protein is triggered by an acidification that allows fusion of the viral and cellular membranes, enabling the nucleocapsid to be released into the cytoplasm [Mandl et al., 2000; Modis et al., 2004]. Antibodies against the E protein can neutralize virus infectivity [He et al., 1995].
DENV can cause a self-limiting fever (DF) or a severe hemorrhagic fever (DHF) and shock-syndrome (DSS) [Halstead, 1988; Rothman and Ennis, 1999]. A major pathophysiological symptom that determines the severity of disease in DHF is increased vascular permeability and abnormal homeostasis [Halstead, 1980]. The increase in the more severe forms of DEN disease is particularly apparent in South America.
Following dengue re-introduction in Brazil in 1986
Juliano Bordignon and Daisy M. Strottmann contributed equally to this study.
*Correspondence to: Claudia N. Duarte dos Santos, Instituto de Biologia Molecular do Paran /FIOCRUZ, Rua Prof Algacyr Munhoz M der 3775, 81350-010 Curitiba, Paran, Brazil.
E-mail: [email protected] Accepted 12 June 2007 DOI 10.1002/jmv.20958
Published online in Wiley InterScience (www.interscience.wiley.com)
ß2007 WILEY-LISS, INC.
[Schatzmayr et al., 1986], more than four million cases were reported. Recent observations indicate that the clinical profile of DENV infection is changing, and that neurological manifestations are becoming frequent in patients with either DF or DHF [Lum et al., 1996; Ramos et al., 1998; Solomon et al., 2000; Cam et al., 2001].
Dengue antigens and viral RNA were detected in human brain tissues in fatal cases of DENV infection [Bhoopat et al., 1996; Miagostovich et al., 1997; Miagostovich et al., 2006]. Ramos et al. [1998] demonstrated the presence of DENV-4 by immunohistochemistry in a fatal case of DHF in neurons, astrocytes, microglia, and endothelial cells, and Miagostovich et al. [2006] detected positive RT/PCR signals for DENV-3 in brain sections of a fatal case, emphasizing the role of dengue virus in causing viral encephalitis. These findings are in agree- ment with the hypothesis that the virus may cross the blood brain barrier and directly infect the brain to cause encephalitis [Lum et al., 1996].
The most frequent neurological manifestations observed in patients during dengue infection are encephalitis [Solomon et al., 2000], acute disseminated encephalomyelitis [Yamamoto et al., 2002], post- infectious disseminated acute encephalitis, acute poly- radiculoneuritis (Guillain-Barre´ Syndrome) [Palma- da Cunha-Matta et al., 2004], mononeuropathies [Patey et al., 1993], and also encephalopathy without encephalitis [Chotmongkol and Sawanyawisuth, 2004].
Clinical signs of neurological disease caused by dengue virus are reduced levels of consciousness, severe heada- che, neck stiffness, focal neurological signs, tense fonta- nelle, and convulsions [Solomon et al., 2000]. The pathophysiology of neurological involvement in dengue infection is attributed to several indirect factors, such as cerebral edema, cerebral hemorrhage, hyponatremia, fulminate hepatic failure with post-systemic encephal- opathy, cerebral anoxia, microcapillary hemorrhage, and the release of toxic products [Sumarmo et al., 1983;
Nimmannitya et al., 1987; Lum et al., 1996], or even by a direct effect of viral infection in the central nervous system (CNS) [Bhoopat et al., 1996; Miagostovich et al., 1997; Ramos et al., 1998].
The mechanism(s) by which DENV causes severe disease is still not understood, but it is assumed that the intrinsic biological properties of the infecting viral strains could contribute to the pathogenicity [Rosen, 1977; Leitmeyer et al., 1999; Messer et al., 2002]. The nature of host cell injury may be a key element in disease pathophysiology. Previous results demonstrated that DENV-1 infection of neurons and liver cells induced apoptotic cell death [Despre`s et al., 1998; Duarte dos Santos et al., 2000].
The lack of an animal model that reflects the severe forms of the disease is an obstacle to studying some aspects of dengue pathogenesis. Adaptation of human viral agents to laboratory animals is a useful method for studying the virus–host relationship, although the resulting disease entities are often unlike those naturally occurring in man. Judiciously employed, these laboratory-created systems can nevertheless serve as
models suitable for the study of some basic mechanisms in viral pathogenesis [Cole and Wisseman, 1969].
To better understand dengue neuropathology and map putative molecular markers of neurovirulence, Despre`s et al. [1998] established a murine model of dengue encephalitis by serial passage of DENV-1 FGA/89 in mouse brains and in the AP61 cell line (Aedes pseudocutellaris). After six serial passages, two neuro- virulent strains were generated, FGA/NA d1d and FGA/NA a5c, that caused encephalitis in newborn mice.
Genetic determinants that were potentially responsible for mouse neurovirulence were studied by sequencing the entire genome of neurovirulent strain FGA/NA d1d, the partial genome of neurovirulent strain FGA/NA a5c, and the parental strain FGA/89. Three amino acid differences in the E protein and one in the NS3 helicase domain were found in FGA/NA d1d. Changes in virus replication and in virus assembly may account, in part, for differences in the induction of apoptosis, which seems to be one of the mechanisms responsible for the pathogenesis of these strains [Duarte dos Santos et al., 2000].
To confirm and define the involvement of the amino acid substitutions in the acquisition of DENV neuro- virulent phenotype for mice we used Swiss mice, murine primary neuronal cells (PN), and murine peritoneal exudate cells (PEC) to characterize the DENV-1 FGA/89, FGA/NA a5c, and BR/01-MR (a recent clinical isolate) viruses. A new neuroadapted variant was generated and characterized in vivo using the murine model, and the mutations acquired during the process were mapped and compared with other neurovirulent DENV-1 strains. These mutations may be useful as potential biomarkers, and may be used to elucidate novel mechanisms behind the neuropathology of infection.
MATERIALS AND METHODS Cell Cultures
Murine primary neuron (PN) culture was established as described by Graner et al. [2000]. Swiss mice (16 days old embryos) were euthanized, the hippocampi were aseptically dissected, trypsinized and plated onto plaque in previous matrix gel treated Lab-TekTM Chamber SlidesTM (NuncTM, Rochester, NY) in NeurobasalTM medium (Gibco Laboratories, Grand Island, NY) sup- plemented with B-27 supplement, 10% fetal calf serum (FCS) and antibiotics. Thioglycolate-induced peritoneal exudate cells (PEC) were prepared from the peritoneal cavity of adult Swiss mice 3 days after intraperitoneal injection of 3.0 ml of 3% thioglycolate medium (Sigma Chemical, St. Louis, MO). PEC were obtained by flushing the peritoneal cavity twice with 10 ml ice-cold saline plus heparin (20 UI/ml) and centrifuged at 300g for 10 min. Cells were resuspended in Dulbecco´s Modified Essential Medium (D-MEM/F-12) supple- mented with 25 mM HEPES (N-2-hydroxyethylpiper- azine-N0-2-ethanesulfonic acid, Sigma Chemical) pH 7.4, 2 mM glutamine, 25 mg/ml of gentamicin, 2.0 g/l sodium bicarbonate plus 10% FCS (Gibco
Laboratories, Grand Island, NY). PEC cells were incu- bated at 378C in 5% CO2humidified air in Lab-TekTM Chamber SlidesTM(NuncTM). After 4 hr, non-adherent cells were washed three times with 1PBS and fresh medium was added. C6/36 cells of Aedes albopictus salivary glands were grown in Leibovitz´s L-15 Medium (L-15), supplemented with 5% of Fetal Calf Serum (FCS), 25mg/ml gentamicin and 2% Tryptose solution (13%, Difco Laboratories, Detroit, MI) and incubated at 278C.
Viruses
Virus stocks were amplified in C6/36 cells and purified in a sucrose gradient as described by Gould and Clegg [1985]. The virus titer was determined by the focus forming unit technique in C6/36 cells (ffu) as described by Despre`s et al. [1993]. FGA/89 and the neurovirulent FGA/NA a5c virus strains were kindly supplied by Dr. Philippe Despre`s, from Unite´ des Interactions Mole´culares Flavivirus-Hoˆtes from Pasteur Institut, Paris, France. FGA/89 (GenBank AF226687) was isolated in South America in 1989 from a patient with DF. BR/01-MR (GenBank AF513110) is a recent clinical isolate from a DF patient in Brazil, and was included as an external control.
In Vivo Neuropathogenicity Experiments In vivo characterization of viral strains (FGA/89, FGA/NA a5c, and BR/01-MR) was performed by intra- cerebral inoculation (i.c.) in newborn Swiss mice (48 hr after birth) with 8,000 ffuC6/36of each virus strain. Signs of encephalitis and mortality were recorded until 21 days after the infection (dpi). To evaluate the kinetics of infection, groups of 30 animals were either infected with a viral strain or were mock infected. At each time point (3, 6, 9, 12, 15, 30, and 60 dpi) three animals were euthanized and their central nervous systems (CNS—
brainþcerebellum) were collected, mechanically dis- rupted and homogenized, and 30 mg were used for RNA extraction with the RNeasy1 Mini Kit (QIAGEN, Hilden, GmbH, Germany). Animal experiments were approved by the ethical committee on animal experimentation (CEP/FIOCRUZ 264/05, CEP/UFPR 23075.002348/2007-14).
A 10% suspension of infected brains (in L-15 medium plus 50mg/ml gentamicin, filtered through a 22mm pore size membrane) was prepared from each time point and used to titrate the viral progeny in C6/36 cells as described.
Quantitative Real Time-Reverse Transcriptase-Polymerase Chain
Reaction (qRT/PCR)
Quantitative PCR (qRT/PCR) for DENV-1 was per- formed to determine the viral load in the CNS of infected mice, as described by Poersch et al. [2005]. To normalize viral RNA quantification, qRT/PCR was performed using theMus musculushousekeeping gene glyceralde- hyde-3-phosphate dehydrogenase (murGAPD). The RNA was reverse transcribed using ImPron II RT
(Promega, Madison, WI) and oligo-dT primers (20mM) following the manufacturer’s protocol. The qRT/PCR was performed with the specific primers murGAPD.F 50- CGACTTCAACAGCAACTCCCACTC-30 and murGAP- D.R 50-CACCCTGTTGCTGTAGCCGTATTC-30, using the SYBR Green Master Mix (Applied Biosystems, Inc., Warrington, UK) in an ABI PRISM 7000 Sequence Detection System (Applied Biosystems, Inc.). The following cycle was used for amplification: 508C for 2 min, 968C for 10 min, followed by 40 cycles of 968C for 15 sec, and 628C for 1 min. Melting curves were used to verify product specificity.
In Vitro Characterization of DENV Strains in Primary Cell Cultures
Viral strains FGA/89, FGA/NA a5c, and BR/01-MR were characterized in vitro by infecting PN and PEC cells in 8 well Lab-TekTMslides (1.0105cells/well) at a multiplicity of infection (MOI) of 10, 40, or 100, and mock infection was the control. After 90 min of adsorption, the inoculums were removed and fresh medium was added.
Twenty-four hours after infection, the Lab-TekTMslides were processed in duplicate for indirect immunofluor- escence (IIF) with mouse anti-dengue serotype-specific polyclonal antibodies (anti-DENV1) and fluorescein isothiocyanate-conjugated goat anti-mouse antibodies (Promega), and for RNA extraction from cell pellets (QIAgen Viral RNA Mini Kit) as described by the manufacture. qRT/PCR to quantify viral RNA and murGAPD was performed as described above. Viral progeny in culture supernatants was determined by titration in C6/36 cells as described previously.
Statistical Analysis
Data from viral infections in cell culture (three independent experiments) were compared by Fisher’s test andP<0.05 was considered statistically significant.
Mouse Neuroadaptation of FGA/89 Virus To confirm the involvement of mutations in the E and NS3 proteins in the acquisition of neurovirulence, we neuroadapted the parental strain FGA/89 strain by modifying the protocol of Despre`s et al. [1998]. Six groups (passages 1–6) of at least nine newborn mice (48 hr after birth) were infected by the i.c. route with FGA/89 virus. The first group (P1) was infected with a dose of 1.0105 ffuC6/36, and groups P2–P6 were infected with a suspension of infected brains from the preceding group. Therefore, suspension from P1 was used to directly infect P2, and P2 to infect P3, etc.
Eight days post-infection (dpi), three animals from each group were euthanized, and the remaining animals were observed for 21 days for signs of encephalitis and mortality. Brain tissues from euthanized animals from each group (pool of three animals) were mechanical disrupted, homogenized, weighed and used for either genetic characterization or to prepare a 10% suspension in phosphate-buffered saline, as described above. The suspensions were used for both viral titration in C6/36
and to infect and generate the subsequent passage (40ml/animal i.c.) in another group of newborn mice, and this procedure was repeated until the sixth passage.
qRT/PCR was performed using RNA extracted from the CNS of three animals at 8 dpi to quantify viral RNA after each passage (data normalized by qPCR for murGAPD).
Genomic Characterization of Neuroadapted Viruses
FGA/NA a5c or FGA/NA P1–P6 viral RNA was extracted from C6/36 infected cells with the QIAgen Viral RNA Mini Kit (QIAGEN), or from brain tissue using the RNeasy1 Mini Kit (QIAGEN). Specific primers were used to synthesize overlapping cDNAs using ImPron II RT (Promega, Inc.), following manu- facturer’s protocols. Complete genomes were amplified by overlapping PCR products using the High Fidelity TripleMaster1 PCR System (Eppendorf, Westbury, NY). Nucleotide sequence was determined using a Thermo Sequenase kit (USB, Inc., Warrington, UK) on an ABI 3100 device with the BigDye7 Terminator method (Applied Biosystems, Inc.). Nucleotide sequen- ces were analyzed by a Phred/Phrap/Consed package of programs (www.phrap.org). Sequences from the 50and 30UTR of the FGA/NA P6 genome were obtained after uncapping and RNA ligation as described elsewhere [Duarte dos Santos et al., 2000].
We amplified and sequenced envelope protein (E) and non-structural protein-3 (NS3) from passages one, three, and five, and we sequenced the whole genome of FGA/NA P2, FGA/NA P4 and FGA/NA P6 to determine the dynamics of acquisition of the neurovirulent pheno- type at a molecular level, and to map the amino acid mutations involved. The complete genome of FGA/NA a5c was also determined. For comparative studies, complete nucleotide sequences of FGA/89 (GenBank AF226687), FGA/NA d1d (GenBank AF226686) and BR/
01-MR (GenBank AF513110) were retrieved (http://
www.ncbi.nih.gov). The CLUSTALw software [Thomp- son et al., 1994, 1997] was used to align sequences, and MegaAlign software (DNASTAR, Inc., Madison, WI) was used to quantify nucleotide and amino acid differ- ences between viral strains.
Histopathology
Mice were euthanized and brain tissues were har- vested and fixed in 10% buffered formalin. Fixed tissues were paraffin embedded, sectioned, and stained with hematoxylin and eosin.
Protein Structure Graphics
The prototype FGA/89 sequence was modeled to known three-dimensional structures using the Swiss-model server (http://swissmodel.expasy.org/).
The DENV-2 RNA helicase structure 2BMF [Xu et al., 2005] was selected as the most similar structure and all three-dimensional analysis was done on these structures by using the PyMol software [DeLano and Bromberg, 2004].
RESULTS
The genetic mechanisms of viral adaptation to new hosts remain poorly understood. To confirm the neuro- virulent phenotype of the FGA/NA a5c virus strain and to determine the genetic substitutions that could be involved, newborn Swiss mice were challenged i.c.
with FGA/89 (parental), FGA/NA a5c (neuroadapted) and BR/01-MR (recent clinical isolate) virus strains.
Animals infected with the neurovirulent variant exhibited signs of encephalitis around 10 dpi and died at 13.1 dpi2.2 (meanstandard deviation) (Fig. 1A).
Mock infected animals, or those infected with FGA/89 or BR/01-MR viruses showed normal behavior (Fig. 1A). To evaluate the ability of the DENV strains to replicate and produce infectious particles in infected mouse brains,
Fig. 1. Kinetics of DEN virus infection in mouse brain. Mouse survival after intracerebral challenge of newborn Swiss mice with FGA/89 (}), FGA/NAa5c (&) and BR-01/MR (~) (A). Viral replication determined by qPCR (B) and viral progeny by titration in C636 cells (C), in CNS of Swiss mice infected with FGA FGA/89 (}), FGA/NAa5c (&), and BR-01/MR (~) 3, 6, 9, 12, 15, 30, and 60 days after infection.
qRT/PCR, and viral titration were done during the infection. The highest levels of viral RNA synthesis and viral progeny appeared at 9 dpi, preceding signs of encephalitis in FGA/NA a5c infected animals, and both levels were higher in FGA/NA a5c than in FGA/89 infected animals (Fig. 1B,C). BR/01-MR did not show detectable viral replication or virus titers in the brain tissue of infected animals (Fig. 1B,C). It was believed that only neuron cells could be infected by dengue virus, but our results demonstrated that the DENV-1 used in this study efficiently infected glial cells as well as neurons (data not shown) in contrast to results from Imbert et al. [1994].
Some mouse neuroadapted variants of DENV-1 repli- cate less efficiently in vitro than the parental virus [Bray et al., 1998]. We infected PN and PEC cells with FGA/89, FGA/NA a5c, and BR/01-MR to characterize replication rates and the ability to produce viable viral progeny in different primary cell substrates (Fig. 2). The choice of primary cultures was an attempt to have a more reliable system. In vitro infections of PEC cells with FGA/89, FGA/NA a5c, and BR/01-MR did not show statistical differences in either viral RNA levels or in viral progeny.
The same was observed for infections in PN cells (data not shown). But the results of the two experiments (infections of PN and PEC cells) could not be compared since the murGAPD gene (gene for normalization) was more highly expressed in PN cells than in PEC.
To identify the amino acid changes potentially responsible for the acquisition of the mouse neuro- virulent phenotype, we determined the complete nucleo- tide sequence of the FGA/NA a5c virus. A comparison between FGA/NA a5c with the already published sequence of the mouse-neuroadapted variant FGA/NA d1d [Duarte dos Santos et al., 2000] showed that only the substitutions Met196/Val in the E protein and Leu435/Ser in the NS3 protein were shared by the two strains.
The mutations at positions E365and E405that had been described in FGA/NA d1d were not present in the FGA/
NA a5c genome. Instead, the mutation Thr276/Pro was specifically detected in the E protein of the FGA/NA a5c virus (Table I).
To confirm the involvement of mutations in E and NS3 proteins on the acquisition of the neurovirulent pheno- type, we reproduced the neuroadaptation of FGA/89 by direct serial passage in the CNS of newborn mice, and we monitored the appearance of mutations that may be involved in the mouse-neurovirulent phenotype by partial nucleotide sequencing (E and NS3 genes) of FGA/NA P1, FGA/NA P3 and FGA/NA P5 or complete nucleotide sequencing of FGA/NA P2, FGA/NA P4, and FGA/NA P6 viruses.
The neurovirulent phenotype appeared after the third passage of FGA/89 in mouse brains, when 100% of the animals developed encephalitis and died. In the fourth passage, 66.6% of infected animals, and in the fifth passage, 42.9% of infected animals showed signs of encephalitis and died. In the sixth passage, all challenged animals developed encephalitis and no animals survived for more than 15 dpi (Fig. 3A). During the neuroadaptation process, three animals from each group were euthanized at 8 dpi to determine viral titers and viral RNA in infected mouse brains. Viral progeny were 2.3104 ffu/mlC6/36 in passage P2, 5.0104ffu/mlC6/36 in P3, 4.2104 ffu/mlC6/36 in P4, 1.2104ffu/mlC6/36in P5, and 6.7105ffu/mlC6/36in P6 (Fig. 3A). Levels of viral RNA and viral progeny in CNS tissues of infected mice follow the mortality rate fluctuations observed during the viral neuroadaptation process (Fig. 3A).
Nucleotide sequences of viruses generated from passages 1 to 6 during neuroadaptation showed two silent mutations in the E protein (T/C2317and C/T2318), these mutations were also present in the FGA/NA a5c virus. In FGA/NA P4 and FGA/NA P6 genomes, which have been completed sequenced, two other silent mutations were identified A/G325 in C protein (also identified in FGA/NA P2), and G/A3694in NS2A protein (Table II). Two conservative mutations, one in domain III of E protein (Phe402/Leu), and the other in the helicase domain of NS3 protein (Val209/Ile), and a non-conservative mutation in the helicase domain of NS3 protein (Leu480/Ser), changing a hydrophobic to a polar non-charged amino acid were first identified in
Fig. 2. Immunofluorescence assays of PEC (A–D) and PN (E–H) cells. Mock infected (A,E), FGA/89 infected (B,F), FGA/NA a5c infected (C,G) and BR-01/MR infected (D,H). All cells were infected at 100 MOI for 24 hr. Images are from a NIKON ECLIPSE E600 (Nikon, Inc., Melville, NY) with Image-Pro1Plus software version 4.5 (MediaCybernetics, Bethesda, MD) with 400magnification.
FGA/NA P3 and fixed in the subsequent passages (Fig. 3B, Table II). To map the amino acid substitutions of the C-terminal region of NS3 protein, a three- dimensional model was obtained by homology to the high-resolution structure of the helicase domain of NS3 protein from DENV-2 virus (Xu et al., 2005, Fig. 4). The non-conservative amino acid substitutions at NS3435 detected in FGA/NA a5c virus, and at NS3480detected in FGA/NAP3-P6 (both Leu/Ser), map to theb-hairpin of domain II (residues 327–481), flanking the motif VI of superfamily II RNA helicases (Fig. 4). The conservative amino acid substitution at NS3209(Val/Ile) in FGA/NA P3-P6 maps to theb-hairpin of domain I (residues 181–
326) between conserved helicase motif I (Walker A) and the Ia motif. These amino acid mutations arose at the same passage (P3), when animals first showed signs of disease and death, and the mutations persisted during the whole neuroadaptation process, until the sixth passage. The complete nucleotide sequence of FGA/NA P4 and FGA/NA P6 viruses were determined, and it revealed only these three amino acid differences compared to FGA/NA P2 and the parental FGA/89 (Table I). A mutation from Leu435/Ser in the helicase domain was also present in the FGA/NA d1d virus, suggesting a potential role for this non-conservative mutation in the DENV mice-neurovirulent phenotype.
None of the aforementioned mutations was detected in the BR/01-MR viral isolate. Complete nucleotide sequences of FGA/NA a5c (GeneBank EF122232) and FGA/NA P6 (GeneBank EF122231) are available. The 50 and 30UTR were identical in FGA/NA P6, FGA/89, and FGA/NA d1d viruses [Duarte dos Santos et al., 2000].
To evaluate CNS damage, we examined hematoxylin- eosin stained tissue sections from the brains of infected mice (Fig. 5). Mock-infected animals showed thin leptomeninges formed by only one layer of meningothe- lial cells, and showed neuronal cortex with well- preserved structure. These animals also showed rare dead neuronal cells, without signs of inflammation.
FGA/89 and BR/01-MR infected animals showed mild chronic lymphoystiocitic leptomeningitis, but the neuro- nal cortex remained well-preserved, without inflam- mation, similar to mock infected animals. Animals infected with FGA/NA a5c and FGA/NA P6 viruses presented intense chronic lymphoystiocitic leptomenin- gitis and moderate encephalitis, with neuronal necrosis (red neurons), satellitosis of lymphocytes and microglia around dying neurons, and the presence of microglial nodules and perivascular lymphocytes. We did not observe histological differences between FGA/NA a5c and FGA/NA P6 samples (Fig. 5).
DISCUSSION
Dengue virus was first recognized in 1978 to cause episodes of encephalitis [Sumarmo et al., 1978].
Recently, clinical cases of encephalitis associated with dengue virus have increased. Two main mechanisms are involved in neurological pathogenesis during dengue infection, one causing encephalopathy by indirect TABLEI.AminoAcidDifferencesBetweenCompleteGenomeSequencesofFGA/89,FGA/NAd1d,FGA/NAa5candFGA/NAP6AminoacidMutationsinNeuroadapeted VirusesComparedWithParentalStrainFGA/89 Virusstrain
EproteinNS3protein DomainIIDomainIIIHelicaseDomain FGA/89Met196ATGThr276ACAVal365GTCPhe402TTCThr405ACCVal209GTCLeu435TTALeu480TTA FGA/NAd1dVal196GTGThr276ACAIle365ATCPhe402TTCIle405ATCVal209GTCSer435TCALeu480TTA FGA/NAa5cVal196GTGPro276CCAVal365GTCPhe402TTCThr405ACCVal209GTCSer435TCALeu480TTA FGA/NAP6Met196ATGThr276ACAVal365GTCLeu402CTCThr405ACCIle209ATCLeu435TTASer480TCA AminoacididentitysharedbetweenFGA/89,FGA/NAa5c,andFGA/NAP6variesbetween99.9%and99.8%.BR-01/MRisthemostdivergentvirusstrainofthosestudied(identityof99.0–98.9%). NoneofthemutationsdescribedaspotentiallyresponsibleforneurovirulentphenotypesarepresentinBR-01/MR,whichcouldexplainitsattenuatedphenotypelikeFGA/89.Themutationfrom leucinetoserineindomainIIofthehelicaseproteinwastheonlymutationsharedbyallneuroadaptedDENV-1.ByMegaAlignsoftware(DNASTAR,Inc.).
systemic damage, and the other causing encephalitis by the direct action of viral replication in brain tissues [Ramos et al., 1998]. The virus probably crosses the blood–brain barrier to directly infect brain cells, caus- ing symptoms of encephalitis. Chaturvedi et al. [1991]
used a murine model to demonstrate the ability of dengue virus to cross the blood–brain barrier, possibly by releasing a cytotoxic factor (CF) produced during infection.
In a previous report [Despre`s et al., 1996], neuro- virulent variants FGA/NA d1d and FGA/NA a5c were generated during the adaptation of strain FGA/89, a human isolate of DENV-1, to grow in newborn mouse brains. Genetic determinants potentially responsible for the neurovirulent phenotype mapped to the E (three substitutions) and NS3 (one substitution) proteins of the FGA/NA d1d genome [Duarte dos Santos et al., 2000].
To confirm the association of the observed amino acid mutations in the acquisition of the mouse neurovirulent phenotype and to identify putative mechanisms involved in DENV pathogenesis using a mouse model, we studied the kinetics of infection of the parental (FGA/89) and neuroadapted (FGA/NA a5c) viruses in vitro and in vivo, and we generated new neurovirulent variants by serial passage in mouse brains.
PN and PEC cells, infected with FGA/89, BR/01-MR or FGA/NA a5c strains, did not show statistically significant differences in viral RNA synthesis and production of infective viral progeny. However, observations of IIF
assays revealed a stronger signal in PN cells infected with FGA/NA a5c than in FGA/89 and BR/01-MR infected cells. These results might suggest more protein produc- tion in neurons infected with the neuroadapted virus, in agreement with results obtained in Neuro2A cells infected with FGA/NA d1d versus the parental strain [Despre`s et al., 1998]. DENV has been shown to infect several types of human cells, such as dendritic cells (DCs), monocytes/macrophages, B and T cells, hepato- cytes, endothelial cells and neuronal cells [Anderson, 2003]. However, more conclusive data are needed to demonstrate whether hepatocytes, endothelial cells, lymphocytes and neurons are targets for DENV replica- tion in vivo [Clyde et al., 2006]. Our results showed that PEC (macrophage like cells) and PN cells were efficiently infected by DENV-1, confirming that DENV can replicate in murine neuronal cells.
In the experiments on infection kinetics in newborn Swiss mice infected i.c. with FGA/NA a5c, animals showed signs of encephalitis around 10 dpi, and died at 13.1 dpi2.2 (meanstandard deviation). These results match those of Despre`s et al. [1998]. The higher levels of viral RNA and viral progeny in the CNS of infected animals were detected at 9 dpi, preceding signs of disease. In contrast with previous observations [Despre`s et al., 1998], we observed both viral replication (qPCR) and the production of infective viral particles (titration in C6/36) in the CNS of animals infected with the parental strain FGA/89, probably due to differences
Fig. 3. Evaluation of the neuroadaptation process of FGA/89. Mouse survival (&), viral RNA in CNS (}) and viral progeny (~) of infected mice during serial passages of FGA/89 in newborn mice (A). Amino acid substitutions in E and NS3 protein during six passages of FGA/89 in newborn mice, compared to nucleotide sequence from FGA/89 (AF226687) (B).
in threshold detection between the assays employed.
Despite viral activity in the CNS of animals infected with FGA/89, no signs of encephalitis were observed until the end of the experiment (21 days). Animals inoculated with FGA/89 showed an important reduction in viral replication and viral progeny at 15 dpi, and complete viral clearance at 30 dpi, suggesting a role for the innate immune response in controlling dengue infection in the CNS [Navarro-Sa´nchez et al., 2005].
The ability of the neuroadapted virus to produce elevated levels of viral RNA and virus particles in the CNS of animals at 9 dpi could largely determine its virulence for irreversible damage in CNS cells, especially neurons, causing encephalitis and death. To evaluate this hypothesis, we did a histopathological analysis. The histological findings showed that FGA/NA a5c and FGA/NA P6 neuroadapted viruses caused much pathological damage in mouse CNS, with encephalitis and leptomeningitis (Fig. 5). Solomon et al. [2000] had demonstrated that encephalitis is the most frequent neurological manifestation in patients during dengue infection. In contrast, animals infected with FGA/89 and BR/01-MR did not show any clinical and histological signs of encephalitis despite the detection of leptome- ningitis. Chotmongkol and Sawanyawisuth [2004] had demonstrated this clinical presentation in a patient TABLE II. Identified Mutations in the Viral Genomes During the Neuroadaptation Process
Virus
Genome sequence
Nucleotide position (FGA89/FGA/NA)
Amino acid position (FGA89/FGA/NA)
FGA/NA P1 E/NS3 E2317ATT/ATC E461Ilea
E2318CTG/TTG E462Leua
FGA/NA P2 Complete C325GCA/GCG C77Alaa
E2317ATT/ATC E461Ilea E2318CTG/TTG E462Leua
FGA/NA P3 E/NS3 E2317ATT/ATC E461Ilea
E2318CTG/TTG E462Leua E2138TTC/CTC E402Phe/Leu NS35144GTC/ATC NS3209Val/Ile NS35958TTA/TCA NS3480Leu/Ser
FGA/NA P4 Complete C325GCA/GCG C77Alaa
E2317ATT/ATC E461Ilea E2318CTG/TTG E462Leua E2138TTC/CTC E402Phe/Leu NS2A3694ACG/ACA NS2A73Thra NS35144GTC/ATC NS3209Val/Ile NS35958TTA/TCA NS3480Leu/Ser
FGA/NA P5 E/NS3 E2317ATT/ATC E461Ilea
E2318CTG/TTG E462Leua E2138TTC/CTC E402Phe/Leu NS35144GTC/ATC NS3209Val/Ile NS35958TTA/TCA NS3480Leu/Ser
FGA/NA P6 Complete C325GCA/GCG C77Alaa
E2317ATT/ATC E461Ilea E2318CTG/TTG E462Leua E2138TTC/CTC E402Phe/Leu NS2A3694ACG/ACA NS2A73Thra NS35144GTC/ATC NS3209Val/Ile NS35958TTA/TCA NS3480Leu/Ser FGA/NA a5c Complete E1520ATG/GTG E196Met/Val
E1760ACA/CCA E276Thr/Pro NS2A3694ACG/ACA NS2A73Thra NS32238TTA/TCA NS3435Leu/Ser
aSilent mutation.
Fig. 4. Diagrammatic representation of the structure of the helicase domain of dengue virus type-2 (strain TSV01 accession no. AY037116) showing mutations found in neuroadapted viruses. Mutations found in FGA/NA P6 (Val209/Ile, dashed-line arrow; Leu480/Ser, large arrow- head) are indicated. The mutation found in FGA/NA a5c (Leu435/Ser) is indicated by a solid arrow. * Indicates the carboxy terminus of the helicase domain.
infected with DENV. Our data suggested that neuro- adapted virus had a greater ability to replicate in CNS, causing extensive inflammatory processes in the neuronal cortex. Conversely, infection with FGA/89 and BR/01-MR viruses caused leptomeningitis without any clinical signs of disease, despite the detection of low (FGA/89) or no (BR/01-MR) viral replication in CNS.
Several authors demonstrated that single amino acid mutations in the flavivirus genome could result in different phenotypes. Liu et al. [2006] demonstrated that a unique amino acid mutation in the NS2A protein of West Nile Virus (WNV) disabled its ability to inhibit interferon alpha/beta (IFN-a/b), reducing virulence in a mouse model. Mandl et al. [2000] showed that site- directed mutagenesis in the E protein of Tick Borne Encephalitis (TBE) virus could attenuate the viral phenotype, and a spontaneous mutation in a neighbor- ing position during viral infection caused reversion to a virulent phenotype. Another example is the increased
mosquito vector infectivity observed in Venezuelan equine encephalitis (VEE) virus, which is due to a single amino acid mutation in the E protein [Brault et al., 2004]. Taken together, these data indicate that single amino acid mutations in both structural and non- structural proteins of Flavivirus are responsible for differences in viral phenotype by affecting viral repli- cation, host response, and vector competence.
Complete genome sequencing of the neurovirulent FGA/NA a5c revealed a conservative mutation at E Met196/Val, also found in FGA/NA d1d, which maps at interface domain I and II in a region that probably acts as a hinge during the fusogenic conformational change of the protein [Rey et al., 1995; Duarte dos Santos et al., 2000]. In addition, a non-conservative mutation (Thr276/ Pro), detected only in FGA/NAa5c E protein, is located at the klb-hairpin (ligand-binding pocket) in the context of the DENV-2 E glycoprotein model, and could be implicated in the enhancement of hydrophobicity.
Fig. 5. Histopathology of mouse brains at 10 dpi, either mock infect- ed or infected with FGA/89, FGA/NA a5c, FGA/NA P6, or BR-01/MR, stained with hematoxylin and eosin. Meninges and neuronal cortex of:
mock infected animals (1and6), FGA/89 (2and7), BR-01/MR (3and8), FGA/NA a5c (4and9), and FGA/NA P6 (5and10). L: leptomeninges;
C: neuronal cortex; A: apoptotic neurons; M: microglial nodulo.
In Photos 9 and 10, L and V shows exocytosis of lymphocytes from vessels. Images done in a Olympus BX 50 with Image-Pro1 Plus software version 4.5 (Maryland) with 200(Photos 1–5) or 400 (Photos6–10) magnification.
Mutations affecting the ligand pocket (residues 268–
280 of DENV-2 strain S1) change the pH threshold for fusion, a crucial step in the DENV life cycle [Modis et al., 2003]. Additionally, a non-conservative mutation from Leu435/Ser in the NS3 helicase, which had been previously observed in FGA/NA d1d, was also detected in the a5c strain. This mutation maps to the tip of the b-hairpin emerging from the RNA binding domain (helicase domain II) to contact the C-terminala-helical domain [Duarte dos Santos et al., 2000].
To confirm the involvement of the amino acid substitutions in the neurovirulent phenotype, we used a new process of neuroadaptation of the FGA/89 strain in mice, by intra-cerebral serial inoculation. Genomic sequencing during the neuroadaptation process first detected the mutations in E and helicase proteins from passage 3 (FGA/NA P3), concomitant with the appearance of signs of encephalitis in the mice.
Only three amino acid substitutions were detected in the neuroadapted viruses FGA/NA P3-P6 (E Phe402/Leu, NS3 Val209/Ile and NS3 Leu480/Ser) compared to the parental FGA/89 strain. The E Phe402/Leu substitution, although not identical to those observed in FGA/NA d1d, mapped to the same domain. Sequence alignment analysis revealed a conserved Phe on residue E402in dengue type 1 virus and in other flaviviruses [Stiasny et al., 1996]. Previous reports demonstrated the impor- tance of amino acid changes at E Phe402/Ile of DENV-2 [Bray et al., 1998] and at E Phe401/Leu DEN-4 [Kawano et al., 1993] to the process of neurovirulence. In DEN-3 neuroadaptated virus, mutations in positions E Glu401/ Lys and E Thr403/Ile [Lee et al., 1997] were selected by passage in mouse brain, and in DEN-1 the mutation in the residue E Thr405/Ile appeared to be involved in this phenomenon [Duarte dos Santos et al., 2000]. These findings suggest the importance of H1pred a-helical region E400-412[Stiasny et al., 1996] in the acquisition of neurovirulence in mice. The predicted a-helix was essential for the conversion of soluble protein E dimers to homotrimers and also for a transient exposure of the fusion peptide, which facilitates interaction with the target membrane [Modis et al., 2004; Stiasny et al., 2005]. We did not find any mutation in FGA/Na a5c in the H1preda-helical region of E protein, despite several descriptions about its relevance in neurovirulence, which suggests that other substitutions could be involved. Mutations in the E protein could enhance the pathogenicity of dengue infections by changing the binding affinity of the virus to cell receptors, or by affecting the entry of dengue virus into neuronal cells by altering the fusogenic pathway [Hasegawa et al., 1992; Jiang et al., 1993; McMinn et al., 1996].
Since DENV E glycoprotein is an important antigenic determinant and is involved in cellular tropism, knowl- edge about the molecular determinants in E protein that alter viral strain pathogenicity is of the utmost rele- vance in understanding DENV pathogenesis [Rey, 2003].
Non-structural protein 3 (NS3) is a multifunctional protein with helicase, nucleoside 50-triphosphatase
(NTPase) and RNA 50-triphosphatase (RTPase) activi- ties involved in replication and capping [Benarroch et al., 2004].
The conservative NS3 Val209/Ile mutation present in FGA/NA P6 appears in domain I of NS3 protein (using the DEN-2 helicase three-dimensional model, Xu et al., 2005), corresponding to the NTPase domain. Based on a sequence alignment of the NS3 helicase domains, an Ile at the corresponding 209-position is observed in Japanese Encephalitis Virus (JEV), WNV, and TBE [Xu et al., 2005], all viruses belonging to the encephalitis group. The role of this mutation in the acquisition of the DENV neurovirulent phenotype needs to be determined.
The non-conservative NS3 Leu480/Ser substitution detected in the FGA/NA P6 is located in the same helicase domain II of NS3 Leu435/Ser, present in FGA/NA a5c and FGA/NA d1d, stressing the involvement of this protein in the neurovirulence process in mice (Fig. 3A,B). Duarte dos Santos et al. [2000] demon- strated that the NS3435mutation maps to the tip of the b-hairpin in the helicase domain and that it may alter the activity of the replicative complex (RC) in mouse neuroblastoma cells. A connection between the two domains via the b-hairpin may be important for the NTPase/helicase activity of NS3, which is thought to change conformation upon ATP hydrolysis to unwind viral RNA.
Mutations in the NS3 protein C-terminus could either alter RTPase activity, which seems to be coupled to helicase activity, and to be dependent on the conserved Walker B motif [Benarroch et al., 2004]. Mutations in the NS3 helicase domain seemed to enhance the ability of mutated virus to replicate, and this was corroborated by higher levels of viral RNA and progeny in the CNS of mice infected with FGA/NA a5c (Fig. 1B,C). Supporting this idea, during the neuro- adaptation process, the highest levels of viral RNA replication in mouse CNS were observed at passages where 100% of the animals died (passages P3 and P6) (Fig. 3A). Massive viral replication and progeny production could trigger apoptosis in mouse neuronal cells, leading to encephalitis and death [Jelachich and Lipton, 1996; Despre`s et al., 1998].
High error rates in viral RNA polymerase (103–105/ nucleotide/cycle) and potentially high selective pressure for dengue virus replication in a different environment, like the CNS of newborn mice, are probably the main mechanisms responsible for the appearance of neuro- adapted virus populations [Domingo and Holland, 1997;
Crotty et al., 2001]. Nucleotide sequencing of the viral RNA from all passages revealed predominant consensus sequences, but when the eletroferograms of mutations in E and NS3 proteins were carefully analyzed, a mixture between different nucleotides during the neuro- adaptation process was observed at the same position, suggesting a competition between different viral pop- ulations during passages in mouse CNS tissue. Vignuzzi et al. [2006] demonstrated the idea that quasispecies are essential for adapting to and surviving in environments
with new selective pressures. The interactions between different variants may be an explanation for the fluctuation in mortality observed during FGA/89 neuro- adaptation (Fig. 3A). Guirakhoo et al. [2004] had demonstrated that extensive passages of DEN/YF virus chimera renders viruses less neurovirulent for infant mice and that it could not be supported by a change in the virus genome. They reasoned that possible sub- populations of mutants (estimated to be less than 10% of the total virus population) with reduced neurovirulence were present, which cannot be detected by consensus sequencing. Furthermore, Novella and Ebendick-Cor- pus [2004] working on molecular basis of fitness lost/
recovery of vesicular stomatitis virus propose that a significant level of fitness increase was observed in several bottlenecked strains with no obvious changes in the consensus sequence.
Since reports of the association of DENV infection with neurological complications are more frequent, this murine model might be a suitable alternative to study the pathogenesis, treatment, and diagnosis of dengue virus infection in the central nervous system. The identification of molecular signatures in DENV strains responsible for neurological disease could be used to predict the emergence of epidemics [Brault et al., 2004], and could eventually drive therapeutic interventions.
To validate the involvement of the observed mutations in the appearance of neurovirulent viral phenotypes, we are neuroadapting other low passage DEN-1 recent clinical isolates, to exclude the possibility of a pheno- menon specific to FGA/89. The model presented in this study, based on neuroadapted viruses and primary cell cultures, is an interesting alternative to study DENV neuropathogenesis.
ACKNOWLEDGMENTS
The authors are indebted to Dr. Philippe Despre`s, from Unite´ des Interactions Mole´culares Flavivirus- Hoˆtesof Pasteur Institut, Paris, France, for providing FGA/89 and FGA/NA a5c. The authors thank Paulo Arauco and Ana Paula Camargo Martins for technical assistance and Anaclete Felini from LACEN/PR for providing DENV-1 patient serum. The authors thank CNPq, Fiocruz, Fundo Parana´, and Fundac¸a˜o Arauca´ria for financial support.
REFERENCES
Anderson R. 2003. Manipulation of cell surface macromolecules by flaviviruses. Adv Virus Res 59:229–274.
Benarroch D, Selisko B, Locatelli GA, Maga G, Romette JL, Canard B.
2004. The RNA helicase, nucleotide 50-triphosphatase, and RNA 50- triphosphatase activities of dengue virus protein NS3 are Mg2þ— Dependent and require a functional Walker B motif in the helicase cataclytic core. Virology 328:208–218.
Bray M, Men R, Tokimatsu I, Lai JC. 1998. Genetic determinants responsible for acquisition of dengue type 2 virus mouse neuro- virulence. J Virol 72:1647–1651.
Brault AC, Powers AM, Ortiz D, Estrada-Franco JG, Navarro-Lopez R, Weaver SC. 2004. Venezuelan equine encephalitis emergence:
Enhanced vector infection from a single amino acid substitution in the envelope glycoprotein. Proc Natl Acad Sci USA 101:11344–
11349.
Bhoopat L, Bhamarapravati N, Attasiri C, Yoksarn S, Chaiwun B, Khunamornpong S, Sirisanthana V. 1996. Immunohistochemical characterization of a new monoclonal antibody reactive with dengue virus-infected cells in frozen tissue using immunoperox- idase technique. Asian Pac J All Immunol 14:107–113.
Cam VB, Fonsmark L, Hue BN, Phuong TN, Poulsen A, Heegaard DE.
2001. Prospective case–control study of encephalopathy in children with dengue hemaorrhagic fever. Am J Trop Med Hyg 65:848–851.
Chambers TJ, Hahn CS, Galler R, Rice CM. 1990. Flavivirus genome organization, expression, and replication. Annu Rev Microbiol 44:
649–688.
Clyde K, Kyle JL, Harris E. 2006. Recent advances in deciphering viral and host determinants of dengue virus replication and pathogen- esis. J Virol 80:11418–11431.
Cole AC, Wisseman CL, Jr. 1969. Pathogenesis of type 1 dengue virus infection in suckling, wealing and adult mice: The relation of virus replication to interferon production and antibody formation. Am J Epid 89:669–680.
Chaturvedi UC, Dhawan R, Khanna M, Mathur A. 1991. Breakdown of the blood–brain barrier during dengue virus infection of mice.
J Gen Virol 72:859–866.
Chotmongkol V, Sawanyawisuth K. 2004. Case report: Dengue hemorrhagic fever with encephalopathy in an adult. Southeast Asian J Trop Med Pub Health 35:160–161.
Crotty S, Cameron CE, Andino R. 2001. RNA virus error catastrophe:
Direct molecular test by using ribavirin. Proc Natl Acad Sci USA 98:6895–6900.
DeLano WL, Bromberg S. 2004. The PyMOL user’s manual. San Carlos, CA: DeLano Scientific.
Despre`s P, Frenkiel MP, Deubel V. 1993. Differences between cell membrane fusion activities of two dengue type-1 isolates reflect modifications of viral structure. Virology 196:209–219.
Despre`s P, Flamand M, Ceccaldi P-E, Deubel V. 1996. Human isolates of dengue virus type-1 induce apoptosis in mouse neuroblastoma cells. J Virol 70:4090–4096.
Despre`s P, Frenkiel MP, Ceccaldi PE, Duarte dos Santos C, Deubel V.
1998. Apoptosis in the mouse central nervous system in response to infection with mouse-neurovirulent dengue viruses. J Virol 72:
823–829.
Domingo E, Holland JJ. 1997. RNA virus mutations and fitness for survival. Annu Rev Microbiol 51:151–178.
Duarte dos Santos CN, Frenkiel MP, Courageot MP, Rocha CF, Vazeille-Falcoz MC, Wien MW, Rey FA, Deubel V, Despre`s P, 2000.
Determinants in the envelope E protein and viral RNA helicase NS3 that influence the induction of apoptosis in response to infection with dengue type 1 virus. Virology 274:292–308.
Gould EA, Clegg JCS. 1985. Growth, titration and purification of togaviruses. In: Mahy BWJ, editor. Virology: A practical approach.
Washington, DC: IRL Press. 43–78.
Graner E, Mercadante AF, Zanata SM, Forlenza OV, Cabral ALB, Veiga SS, Juliano MA, Roesler R, Walz R, Minetti A, Isquierdo I, Martins VR, Brentani RR. 2000. Cellular prion protein binds laminin and mediates neuritogenesis. Mol Brain Res 76:85–92.
Gubler DJ. 1998. Dengue and dengue hemorrhagic fever. Clin Microbiol Rev 11:480–496.
Guirakhoo F, Zhang Z, Myers G, Johnson BW, Pugachev K, Nichols R, Brown N, Levenbook I, Draper K, Cyrek S, Lang J, Fournier C, Barrere B, Delagrave S, Monath TP. 2004. A single amino acid substitution in the envelope protein of chimeric yellow fever-dengue 1 vaccine virus reduces neurovirulence for suckling mice and viremia/viscerotropism for monkeys. J Virol 78:9998–10008.
Hasegawa H, Yoshida M, Shiosaka T, Fujita S, Kobayashi Y. 1992.
Mutations in the envelope protein of Japanese encephalitis virus affect entry into cultured cells and virulence in mice. J Virol 191:
158–165.
Halstead S. 1988. Pathogenesis of dengue: Challenges to molecular biology. Science 239:46–481.
Halstead SB. 1980. Dengue haemorrhagic fever—A public health problem and a field for research. Bull WHO 58:1–21.
He RT, Innis BL, Nisalak A, Usawattanakul W, Wang S, Kalayanarooj S, Anderson R. 1995. Antibodies that block virus attachment to Vero cells are a major component of human neutralizing antibody response against dengue virus type 2. J Med Virol 45:451–461.
Heinz FX, Stiasny K, Pueschner-Auer G, Holzmann H, Allison S, Mandl CW, Kunz C. 1994. Structural change and functional control of the tick-borne encephalitis virus glycoprotein E by the hetero- dimeric association with protein prM. Virology 198:109–117.
Imbert JL, Guevara P, Ramos-Castan˜eda J, Ramos C, Sotelo J. 1994.
Dengue virus infects mouse cultured neurons but not astrocytes.
J Med Virol 42:228–233.
Jelachich ML, Lipton HL. 1996. Theiler’s murine encephalomyelitis virus kills restrictive but not permissive cells by apoptosis. J Virol 70:6856–6861.
Jiang WR, Lowe A, Higgs S, Reid H, Gould EA. 1993. Single amino acid codon changes detected in Louping ill virus antibody-resistant mutants with reduced neurovirulence. J Gen Virol 74:931–935.
Kawano H, Rostapshov V, Rosen L, Lai CJ. 1993. Genetic determinants of dengue type 4 virus neuovirulence for mice. J Virol 67:6567–
6575.
Leitmeyer KC, Vaughn DW, Watts DM, Salas R, Villalobos I, de Chacon Ramos C, Rico-Hesse R. 1999. Dengue virus structural differences that correlate with pathogenesis. J Virol 73::4738–4747.
Lee E, Weir RC, Dalgarno L. 1997. Changes in dengue virus major envelope protein on passaging and their localization o three- dimensional structure of the protein. Virology 232:281–290.
Lindenbach BD, Rice MR. 2001. Flaviviridae: The viruses and their replication. In: Knipe MD, Howley MP, editors. Fields virology 4th edition. Philadelphia: Lippincott Williams & Wilkins. 991–1042.
Liu WJ, Wang XJ, Clark DC, Lobigs M, Hall RA, Khromykh AA. 2006. A single amino acid substitution in the West Nile virus nonstructural protein NS2A disables its ability to inhibit alpha/beta interferon induction and attenuated virus virulence in mice. J Virol 80:2396–
2404.
Lum LCS, Lam SK, Choy YS, George R, Harun F. 1996. Dengue encephalitis: A true entity? Am J Trop Med Hyg 54:256–259.
Mandl CW, Allison SL, Holzmann H, Meixner T, Heinz FX. 2000.
Attenuation of tick-borne encephalitis virus by structure-based site-specific mutagenesis of a putative flavivirus receptor binding site. J Virol 74:9601–9609.
Messer WB, Vitarana UT, Sivananthan K, Elvtigala J, Preethimala LD, Ramesh R, Withana N, Gubler DJ, De Silva AM. 2002.
Epidemiology of dengue in Sri Lanka before and after the emergence of epidemic dengue hemorrhagic fever. Am J Trop Med Hyg 66:765–773.
Miagostovich MP, Ramos RG, Nicol AF. 1997. Retrospective study on dengue fatal cases. Clin Neuropathol 16:204–208.
Miagostovich MP, dos Santos FB, Fumian TM, Guimara˜es FR, da Costa EV, Tavares FN, Coelho JO, Nogueira RM. 2006. Complete genetic characterization of a Brazilian dengue virus type 3 strain isolated from a fatal outcome. Mem Inst Osw Cruz 101:307–313.
McMinn PC, Weir RC, Dalgarno L. 1996. A mouse-attenuated protein variant of Murray Valley encephalitis virus with altered fusion activity. J Gen Virol 77:2085–2088.
Modis Y, Ogata S, Clements D, Harrison SC. 2003. A ligand-binding pocket in the dengue virus envelope glycoprotein. Proc Natl Acad Sci USA 100:6986–6991.
Modis Y, Ogata S, Clements D, Harrison SC. 2004. Structure of the dengue virus envelope protein after membrane fusion. Nature 427:
313–319.
Navarro-Sa´nchez E, Despre`s P, Cedillo-Barro´n L. 2005. Innate immune response to dengue virus. Arch Med Res 36:425–435.
Nimmannitya S, Thisyakorn U, Hemsrichart V. 1987. Dengue haemorrhagic fever with unusual manifestations. Southeast Asian J Trop Med Pub Health 18:398–406.
Novella IS, Ebendick-Corpus BE. 2004. Molecular basis of fitness loss and fitness recovery in vesicular stomatitis virus. J Mol Biol 342:
1423–1430.
Palma-da Cunha-Matta A, Soares-Moreno SA, Cardoso-de Almeida A, Aquilera-de-Freitas V, Carod-Artal FJ. 2004. Neurological compli-
cations arising from dengue virus infection. Rev Neuro 39:233–
237.
Patey O, Ollivaud C, Breuil J, Lafaix C. 1993. Unusual neurologic manifestations occuring during dengue fever infection. Am J Trop Med Hyg 48:793–802.
Poersch OPC, Pavoni DP, Queiroz MH, de Borba L, Goldenberg S, dos Santos CN, Krieger MA. 2005. Dengue virus infections: Comparison of methods for diagnosing the acute disease. J Clin Virol 32:272–
277.
Ramos C, Sa´nchez G, Pando HR, Baquera J, Herna´ndez D, Mota J, Ramos J, Flores A, Llausa´s E. 1998. Dengue virus in the brain of fatal case of hemorrhagic dengue fever. J Neuro Virol 4:465–468.
Rey FA, Heinz FX, Mandl C, Kunz C, Harrison SC. 1995. The envelope glycoprotein from tick-borne encephalitis virus at 2A˚ resolution.
Nature 375:291–298.
Rey FA. 2003. Dengue virus envelope glycoprotein structure: New insight into its interactions during viral entry. Proc Natl Acad Sci USA 100:6899–6901.
Rothman AL, Ennis FA. 1999. Immunopathogenesis of dengue hemorrhagis fever. Virology 257:1–6.
Rosen L. 1977. ‘‘The emperor’s new clothes’’ revisited or reflections on the pathogenesis of dengue hemorrhagic fever. Am J Trop Med Hyg 26:337–343.
Schatzmayr HG, Nogueira RM, Travassos da Rosa AP. 1986. An outbreak of dengue vı´rus at Rio de Janeiro—1986. Mem Inst Osw Cruz 81:245–246.
Solomon T, Dung MN, Vaughn WD, Kneen R, Thao TTL, Raengsa- kulrach B, Loan TH, Day JPN, Farrar J, Myint ASK, Warrell JM, James SW, Nisalak A, White JN. 2000. Neurological manifestations of dengue infection. Lancet 355:1053–1059.
Stiasny K, Allison SL, Bauer AM, Kunz C, Heinz FX. 1996. Structural requirements for low-pH-induced rearrangements in the envelope glycoprotein of tick-borne encephalitis virus. J Virol 70:8142–
8147.
Stiasny K, Ko¨ssl C, Heinz FX. 2005. Differences in the postfusion conformations of full-length and truncated class ii fusion protein e of tick-borne encephalitis virus. J Virol 70:6511–6515.
Sumarmo HW, Jahja E, Gubler DJ, Sutomenggolo TS, Saroso SJ. 1978.
Encephalopathy associated with dengue virus. Lancet 449–
450.
Sumarmo HW, Jahja E, Gubler DJ, Suharyono W, Sorensen K. 1983.
Clinical observations on virologically confirmed fatal dengue infections in Jakarta, Indonesia. Bull WHO 61:693–701.
Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: Improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positions-specific gap penalties and weight matrix choice. Nucl Acid Res 22:4673–4680.
Thompson JD, Gibson TJ, Plewniack F, Jeanmougin F, Higgins DG.
1997. The ClustalX windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucl Acid Res 24:4876–4882.
Vignuzzi M, Stone JK, Arnold JJ, Cameron CE, Andino R. 2006.
Quasispecies diversity determines pathogenesis through coopera- tive interactions in a viral population. Nature 439:344–348.
Xu T, Sampath A, Chao A, Wen D, Nanao M, Chene P, Subhash G, Lescar J. 2005. Structure of the dengue virus helicase/nucleoside triphosphatase catalytic domain at a resolution of 2.4A˚ . J Virol 79:
10278–10288.
Yamamoto Y, Takasaki T, Yamada K, Kimura M, Washizaki K, Yoshikawa K, Hitani A, Nakamura T, Iwamoto A. 2002. Acute disseminated encephalomyelitis following dengue fever. J Infect Chemother 8:175–177.
