Genetic variation in the 3’ untranslated region of dengue virus serotype 3 strains isolated from mosquitoes and humans in Brazil

Summary Background Dengue, a mosquito-borne viral infection caused by one of the four dengue virus (DENV) serotypes (DENV-1 to 4), replicate alternately on the mosquito vector and human host and are responsible for infections throughout tropical and subtropical regions of the world. In Brazil, the disease has become a major public health problem and the introduction of DENV-3 in 2000 in Rio de Janeiro (RJ) was associated with severe dengue epidemics. The potential emergence of strains associated with severe disease highlights the need for the surveillance of DENV in human host and vectors. Methods Aiming to contribute for DENV phylogenetic and vector-virus-human host studies, we sequenced the entire genome of one DENV-3 isolated from naturally infected Aedes aegypti from RJ in 2001 and characterized the 3’ UTR from strains isolated from mosquitoes and humans. Mosquitoes were pooled and submitted to virus isolation in Ae. albopictus C6/36 cells and the infecting serotype was identified by immunofluorescence using type-specific monoclonal antibody. Sequence analysis was performed using BioEdit software, the multiple alignments were performed using CLUSTAL W and the phylogenetic analysis by MEGA 5, using the Neighbor-joining method. Secondary structure prediction was performed by using the MFOLD program. Results Exclusive substitutions and a substitution leading to a stop codon on the NS5 gene were observed in the DENV-3 isolated from a naturally infected Ae. aegypti and fully sequenced. As an 8- nucleotides deletion was observed within the 11- nucleotides (nts) insertion on the variable region (VR) from the 3′UTR in this isolate, we further sequenced other DENV-3 from both mosquitoes and humans. The majority of DENV-3 from RJ analyzed were characterized by the 11-nts insertion in the VR of the 3′UTR, despite the observation of strains carrying the 8-nts deletion. The latter presented similar secondary structures, however not all strains presenting the 11-nts insertion were similar in the predicted secondary structure. Conclusions The phylogeny based on the analysis of the complete genome and 3′UTR characterized the DENV-3 isolated from both vector and human host as belonging to Genotype III (GIII), despite the differences observed on the 3’ UTR. Further studies are needed to address the role of those mutations in the transmission of the different viral populations and vector competence.


Background
Dengue is a mosquito-borne viral infection caused by one of the four dengue virus serotypes (DENV-1 to 4), belonging to genus Flavivirus, family Flaviviridae. The viruses replicate alternately on the mosquito vector, mainly (Ae. aegypti) and human host and are responsible for infections throughout tropical and subtropical regions of the world [1,2].
The rapid global spread of the four DENV serotypes in the last 50 years resulted in the dispersal of genotypes associated with increased disease severity [3]. In Brazil, dengue has been a major public health problem since DENV-1 introduction and spread in 1986 [4], however the introduction of the genotype III of DENV-3, in December 2000, in Nova Iguaçu, State of Rio de Janeiro (RJ), caused one of the most severe epidemics reported in the country in 2002 [5][6][7]. Despite the co-circulation of DENV-1, DENV-2 and DENV-3 in that area, DENV-3 was the only serotype detected in pools of Ae. aegypti during an entomological surveillance performed [8].
Sequencing of distinct DENV genomic regions has identified five genotypes for DENV-3: Genotypes I to III (GI to GIII) which are responsible for most DENV-3 human infections and have been associated with both dengue fever (DF) and dengue haemorrhagic fever (DHF) epidemics in Southeast Asia, Indian Subcontinent, South Pacific and East Africa and Americas, and Genotypes IV and V (GIV and GV) which were not associated with DHF epidemics and are only represented by few early sequences from Americas, South Pacific and Asia [9][10][11][12][13].
The DENV genome is composed by a positive singlestranded RNA of approximately 11 kb in length with an open reading frame encoding for the viral polyprotein, which is cleaved into three structural proteins (C, prM and E) and seven non-structural proteins (NS1, NS2A, NS2B, NS3, NS4 and NS5) flanked by 5′ and 3′ untranslated regions (UTRs) of about 100 and 400 nucleotides, respectively [1]. The flaviviruses UTRs are predicted to form secondary stem-loop (SL) structures, which are highly conserved and play a role in viral replication [14][15][16][17][18].
According to predicted secondary structures, the DENV 3′UTR can be divided into three domains [18]. The domain I, which is located immediately after the NS5 stop codon, is considered the most variable region (VR) within the viral 3′UTR, as it shows large heterogeneity in both length and nucleotide sequences [19][20][21]. Mutations and deletions within these regions may alter infectivity and reduce efficiency of viral replication [22,23] and differences between strains in these regions may correlate with DENV virulence and pathogenicity [24][25][26][27]. Furthermore, deletions and nucleotide variations were also described in the VR within the same serotype [28][29][30]. Domain II is of moderate conservation, comprising several hairpins motifs and where conserved sequence (CS2) and repeated CS2 (RCS2) are present. Domain III is the most conserved region of the 3′UTR with CS1 followed by a terminal stem-loop (3′SL) [18].
Here, aiming to contribute for the studies on human host-virus-vector interactions, we fully sequenced the genome of one DENV-3 isolated from naturally infected fieldcaught mosquitoes in RJ and characterized the viral 3′UTR in comparison to other sequenced DENV-3 isolated from naturally infected mosquitoes and human hosts.

Ethical Statement
All human DENV-3 strains belong to a previously gathered collection from the Laboratory of Flavivirus, IOC/ FIOCRUZ, RJ, Brazil obtained from acute phase human serum through the passive surveillance system from an ongoing Project approved by resolution number CSN196/ 96 from the Oswaldo Cruz Foundation Ethical Committee in Research (CEP 274/05), Ministry of Health, Brazil.
Ae. aegypti examined in this study were collected by the staff of the Dengue Control State Program for the determination of house infestation index, virological and entomological surveillance. No special permission or written consent is required for house entrance for mosquito collection and larval site treatment.

Viral strains
The DENV-3 strains isolated from Ae. aegypti adult mosquitoes (n=4) and human hosts (n=10) naturally infected in RJ were collected from epidemics occurred from 2001 to 2008. The first Brazilian DENV-3 strain (BR74886/02) isolated from a human fatal case fully sequenced [31] was used for comparison purposes and detailed information on the strains is provided on Table 1.

DENV-3 human cases
From 2001 to 2008, the Laboratory of Flavivirus, as a Regional Reference Laboratory for the Brazilian Ministry of Health, received a total of 16,185 dengue suspected cases for routine diagnosis. Virus isolation was attempted in 9,405 cases and DENV-3 was the infecting serotype in 52.8% of the positive isolates. The samples analyzed in this study were chosen randomly and representative of each year, during and after the 2002 epidemic in RJ.

Preparation of vectors
Mosquitoes' pools were macerated in 1 ml of Leibovitz L-15 medium (Sigma) plus antibiotics (penicillin-streptomycin, 10,000 units -Invitrogen) and centrifuged (6,000 rpm at 4°C for 30 min). Supernatant was transferred to an Eppendorf tube containing 100 mL of streptomycin / fungizone and penicillin, kept in an ice bath for 1 hour and centrifuged (3,000 rpm at 4°C for 15 min). Supernatant was transferred to an Eppendorf tube containing 0.3ml of fetal calf serum (Invitrogen) and frozen (−70°C).

Virus isolation
Virus isolation was performed by inoculation into monolayers of C6/36 Aedes albopictus cells [33] in Leibovitz L-15 medium (Sigma) supplemented with 2% fetal calf serum (Invitrogen) and 0.2 mM of nonessential amino acids (Invitrogen). Cells were incubated at 28°C for 5 to 7 days and observed for cytopathic effects. Isolates were identified by indirect fluorescent antibody test (IFAT) using serotype-specific monoclonal antibodies [34] and infected supernatant was clarified by centrifugation and virus stocks stored in 1-mL aliquots at −70°C.

RNA extraction
Viral RNA was extracted using QIAamp Viral RNA Mini kit (Qiagen) following the manufacturer's instructions and stored at -70°C for DENV typing and sequencing. For the viral 3′UTR sequencing, the RNA was extracted directly from serum and mosquitoes macerate, previously detected by RT-PCR. For the full genome sequencing of the DENV-3 strain BR73354/01, the RNA was extracted from the first passage in cell culture.

RT -PCR (Reverse transcriptase-polymerase chain reaction)
RT-PCR for detecting and typing DENV was performed as described previously [35].

Sequencing and phylogenetic analysis
PCR products were sequenced in both directions using the BigDye Dideoxy Terminator sequencing kit (Applied Biosystems). The mosquitoes' DENV-3 full-length genome sequence and 3′UTR sequences obtained in this study were deposited in GenBank (www.ncbi.nlm.nih.gov) and are described on Table 1. Sequence and similarity identity were used for the comparison, DENV-1 (GenBank accession number #AF513110), DENV-2 (#AF489932) and DENV-4 (# AF326573) strains were used as outgroup to root the tree.

Secondary structure analysis
The predicted secondary structures were generated by MFOLD web server (http://mfold.rna.albany.edu/? q=mfold/RNA-Folding-Form) with default folding parameters and folding predictions at 37 o C on the VR from the 3' UTR of DENV-3.
In Brazil, the first DENV-3 was isolated in December of 2000 in Nova Iguaçu, RJ [5] when the Ae aegypti infestation level was 8.1% and, 58% of those mosquitoes were resistant to temephos at that time [8,48]. Due to the role of the city of Nova Iguaçu in dengue epidemiology, after the DENV-1 introduction in 1986 [4], field studies were conducted for detection of DENV in field-caught vectors [8,49]. The potential emergence of strains associated with severe disease highlights the need for the surveillance of DENV in human host and vectors, as the detection of DENV in infected field-caught vectors is considered a useful tool for the early prediction of epidemics and detection of new serotypes/ genotypes introductions [50,51].
The entomological surveillance performed in the first semester of 2001 in Nova Iguaçu, RJ, resulted in the isolation of three DENV-3 strains from the districts of Santa Efigênia (BR73354/01), California (BR73356/01) and Morro Agudo (BR73636/01) isolated from three pools (9 mosquitoes/ pool) of naturally infected Ae. aegypti females [8]. In January 2006, one DENV-3 strain (BR81200/06) was isolated from one Ae. aegypti pool composed of three females, collected indoors in the Vargem Pequena neighborhood, west region of RJ.
In order to access the differences among Brazilian DENV-3, we sequenced and deposited on Genbank (FJ177308) the entire genome sequence of one virus isolated from naturally infected Ae. aegypti (BR73354/01) and compared to the Brazilian strain 74886/02 (AY679147), isolated from the liver of a fatal case during the DENV-3 epidemic occurred in 2002 [31]. The nucleotide similarity was 99.3% and the phylogeny based on the analysis of the complete coding region characterized the Brazilian strain as belonging to genotype III (Indian Subcontinent), Figure 1.
Amino acid substitutions were observed throughout the entire coding region, when the Brazilian DENV-3 strains were compared to the prototype PHIL/H87/1956 and strains representative of the other genotypes. Some substitutions were exclusive to the Brazilian DENV-3 strains analyzed (Table 2) and some were shared among the Brazilian strains and the strain isolated from Ae. aegypti in Taiwan in 1998 (TAIWAN/TWmosq/1998), Table 2. Exclusive substitutions to the DENV-3 Brazilian strains analyzed in this study were observed on the capsid, prM and envelope genes. Besides those others observed were also shared by other genotype III strains previously analyzed [11].
On the NS1 gene only one residue substitution was observed, and substitution on NS2A and NS2B were also reported. On the NS3 gene, a substitution on only NS3 115 was exclusive to the Brazilian DENV-3. On NS4B, the same substitution was observed on positions NS4B 21 and NS4B 138 . However, on the NS5 gene, besides the ten substitutions exclusive to the Brazilian DENV-3, substitutions exclusive to the Brazilian DENV-3 isolated from Ae. agypti (BR73354/2001) were observed almost consecutively, on NS5 97 and NS5 98 . On NS5 101 , the substitution resulted in a stop codon and other substitutions throughout NS5 were also observed. Despite this, the original macerate was reinoculated in C6/36 cells culture and the DENV −3 infection confirmed by the presence of cytopathic effect and a positive RT-PCR. The presence of genome-defective DENV-3 containing either stop codons or deletions in vivo has been reported previously [52].
Despite the use of the E gene for DENV phylogenetic and evolutionary studies [9,12,28,[53][54][55][56][57][58][59][60][61][62] due to its biological properties and selective pressure imposed by the host immune response, the role of sequences heterogeneity in other genomic regions which includes the non-structural genes and the genome UTRs cannot be excluded [53]. Table 2 Amino acid differences among selected Brazilian DENV-3 and strains representative of the DENV-3 genotypes, compared to the prototype H87 (Continued)   Previous studies have suggested that the sequence and secondary structures of the 5′ and 3′UTR of flaviviruses play an important role in viral replication and differences in these regions may influence viral virulence [24,26,27,63]. Mutations and deletions within these regions may alter infectivity and reduce efficiency of viral replication [22,23]. Furthermore, the domain I from the DENV 3′UTR is considered the most variable region (VR) from the 3′UTR [18] and can serve as a good marker for DENV evolution [19][20][21].
The analysis of the 3′UTR of the strain BR73354/01 genome showed an 8-nucleotides deletion within the 11-nucleotides insertion on the VR, previously observed for the Brazilian DENV-3 strain isolated from humans [31] and common to genotype III DENV-3 strains from the Latin America/Caribbean and Sri Lanka regions [27,64]. Nucleotides substitutions exclusive to the BR73354/01 were observed on positions 10,383 and 10,391 from the 3′UTR. One substitution on the RCS2 and one on the CS2 were shared by all Brazilian DENV-3 when compared to the prototype.
We additionally analyzed and deposited on GenBank the 3′UTR sequences from other three DENV-3 strains isolated from naturally infected Ae. aegypti isolated in 2001 and 2006 in RJ (BR73356/01, BR73636/01 and BR81200/ 06) and from ten DENV-3 isolated from humans from 2001 to 2008. The strain BR73356/01 presented the same 8-nucleotides deletion observed for the strain BR73354/01. However, the other two strains also isolated from mosquitoes in RJ (BR73636/01 and BR81200/06) presented the 11-nucleotides insertion common to the human strains. The analysis of the 3′UTR from DENV-3 isolates from humans showed that nine out of ten strains also presented the 11-nucleotides previously described. However, one of the strains isolated in Rio de Janeiro in 2002 (BR74792/02) showed the same 8-nucleotides deletion observed on the mosquito strains ( Figure 2). Previous studies have shown deletions and nucleotide variations in the VR within a same serotype [28,29]. Despite those observations, it was also shown a high conservation on the 3′UTR RCS2, CS1 and CS2 conserved regions (Figure 2, gray-shadowed areas) among all the Brazilian strains analyzed. This was quite expected as domain II is of moderate conservation and domain III, the most conserved region of all [18].Therefore, we focused on the VR of the 3′UTR, aiming to better characterize those mutations by predicting the secondary structures of that region. Not all strains presenting the 11-nucleotides insertion were similar in structure ( Figure 3A and 3B). In fact, despite the 11-nucleotides insertion, the strain BR72/2008 presented a unique secondary structure ( Figure 3B). The only difference from the latter is a nucleotide substitution (G →A) on the 11-nucleotides insertion region, when compared to all other Brazilian sequences analyzed ( Figure 2). The strain BR80996/2006, also showed a unique secondary structure due to the nucleotides substitution presented in the VR, despite the presence of the insertion shared with the other strains. The slight difference presented by the strain BR83904/2007, was due to a substitution (C→T) exclusive to this sequence. Furthermore, all three sequences with the 8-nucleotides deletion (BR73354 mosq/2001, BR73356mosq/2001 and BR74792/2002) presented the same secondary structure ( Figure 3C).
Phylogenetic studies based on the 5′and 3′UTR have shown to be very useful for molecular epidemiological studies [19,21,27,28,65]. The Maximum-Likelihood phylogenetic tree of Brazilian DENV-3 strains isolates from naturally infected Ae. aegypti mosquitoes and humans based on the 3′UTR sequence analysis places those strains as belonging to genotype III, corroborating the findings of the full-length genome analysis (data not shown).

Conclusions
Here, we analyzed the coding region and the 3′UTR of DENV-3 from both human host and mosquitoes and described insertions, deletions and a substitution leading to stop codon formation. The majority of DENV-3 in this study was characterized by the 11-nucleotide insertion in the 3′UTR, despite the observation of strains carrying the 8-nucleotide deletion. In spite the presence of distinct viral variants, it is suggested that the major variant is transmitted. However, how those distinct viral populations are maintained or transmitted is not fully understood, therefore the availability of viruses isolated from both hosts are crucial for the better comprehension of the vector-virus-human host interactions and for quasispecies investigations. Furthermore, the analysis of those distinct viral populations in experimentally infected mosquitoes may help to elucidate those observations.

Competing interest
The authors have no conflict of interest.
Authors' contributions FBS, RMRN and RLO designed the study. MGC, FNB performed the experiments, MGC, FBS and RLO wrote the paper. All authors read and approved the final manuscript.
(See figure on previous page.) Figure 3 On (A) the predicted secondary structure of the variable region (VR) from Brazilian DENV-3 strains isolated from humans and mosquitoes Ae. Aegypti (nucleotides 1 to 108) presenting the 11-nucleotides insertion and compared to the prototype PHIL/87/1956 (nucleotides 1-104). On (B) the predicted secondary structures from the Brazilian strains (nucleotides 1 to108) presenting the 11-nucleotides insertion, but with nucleotides substitutions differing from the strains on (A) and on (C) the structures from the sequences (nucleotides 1 to 100) with the 8-nucleotides deletion.
Financial support CNPq, CAPES, FIOCRUZ, FIOCRUZ/PAPES V and FAPERJ. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.