DENV-1 Genotype V in Brazil: Spatiotemporal dispersion pattern reveals continuous co-circulation of distinct lineages until 2016

In Brazil, DENV-1 introduced in the 80’s, remained the prevalent serotype from 2012 to 2016. After its re-emergence in the country in 2009, the co-circulation of different viral lineages was identified, however, its transmission dynamics afterwards, was not fully characterized. In this study, we performed the continuous molecular surveillance after the reemergence period (2012 to 2016), covering the 30 years of circulation of DENV-1 in Brazil. Phylogenetic analysis allowed confirmation of the continued presence of genotype V, as well as three distinct co-circulating lineages. The molecular characterization of the E gene presented two new amino acid substitutions previously unidentified in the country. Phylogeographic analysis has shown that a large flow of migrations has occurred between Brazil and Argentina in the last 10 years.

The dengue virus (DENV) belongs to the family Flaviviridae, genus Flavivirus, and presents four serotypes antigenically distinct, DENV-1 to DENV-4 1 . They are classified as arboviruses, since its maintenance in nature occurs through a transmission cycle of vertebrate hosts and hematophagous arthropods 2 , most prevalently in tropical and subtropical areas worldwide 3 .
Dengue is one of the main public health problems in the world, with relevant social and economical impact due to the increased geographic extension, number of cases and disease severity 4 and, in Brazil, DENV activity has grown significantly since the introduction of DENV-1 in Rio de Janeiro (RJ) in 80's. This serotype re-emerged in 2009 after approximately eight years without being related to epidemics, causing in 2010, more than one million probable cases, with the highest hospitalization rate reported in the country, especially in those over 60 years old. In the last 32 years, more than twelve million dengue cases have been reported in Brazil 5,6 .
After reemergence period (2009)(2010), the DENV-1 and DENV-4 were responsible for a high number of cases in the following years, reviewed in dos Santos et al. 7 , being the DENV-1 prevalent in years of 2014 to 2016, representing proximally 90% of isolated cases in country 8,9 . This serotype continue to be identified until today 10 .
Phylogenetic analysis of DENV-1 since its introduction in the country, characterized those viruses as belonging to genotype V, with its origin in the Caribbean and Venezuela and showed that several independent introductions occurred over time with high genotype variability [11][12][13][14][15][16][17] .
Molecular characterization and evolutionary studies can be important tools to monitor the introduction and understanding of the viruses spread in a region, as well as to predict possible epidemiological consequences of such events. In this scenario, the DENV serotypes surveillance is of great relevance, as it may provide information on the spread of potentially virulent strains, as well as, to assess their impact on the population during an outbreak. Here, we performed the molecular characterization and investigated the spatiotemporal patterns of

Results
Phylogenetic analysis of DENV-1 strains based on Envelope gene. For phylogenetic analysis of current DENV-1 lineages isolated from patients from 2012 to 2016 in Brazil, the envelope (E) gene sequencing of viral strains (n = 30), of four distinct geographic region: Southeast (n = 20), Central-Western (n = 4), North (n = 1) and Northeast (n = 5) (Table 1), was performed. The Maximum Likelihood (ML) phylogenetic analysis of 717 E gene sequences of DENV-1 genotype V available in Genbank, additionally to thirty sequences generated in this study, and sequences representing each DENV-1 genotype (I to IV), totalizing 751 sequences, showed that all strains currently belong to genotype V, but grouped in two distinct clades, previously identified by the our group, as Clade I and Clade II into the cosmopolitan clade of genotype V (Fig. 1A).
In order to analyze the branches that actually had recent Brazilian sequences, we separated from Clade I, the subset corresponding to Sub-Clade I, containing 240 sequences and submitted to a new ML analysis to identify the specific branches with Brazilian representatives (Fig. 1B). From this analysis, three subsets were separated for phylogeographic reconstruction: Sub-Clade Ia (n = 106), Sub-Clade Ib (n = 56) and Clade II (n = 116) ( Table S2).  (Table S2, Sub-Clade I). Moreover, the Sub-Clades Ia and Ib identified within this Sub-Clade were submitted separately to a Bayesian phylogeographic analysis.   (Table S2, Subset-Clade II).

ID samples Year
The subset was also subjected to a Bayesian phylogeographic analysis and the results showed that this subset appears to have arisen in the Caribbean Lesser Antilles ( Molecular characterization of DENV-1 strains based on E gene. The sequences alignment demonstrated different patterns of nucleotides responsible for conserving some amino acid in domains I, II, III and stem region of the E gene among the strains of the different clusters. In order to identified those differences, we used a strain from the period of DENV-1 introduction in Brazil, for comparison purposes (Table 2).
In domain I, at position E 297 , the amino acid methionine (M) was observed in all 30 sequences studied, different from the 80's sequence which has a threonine (T) at that same position. In domain II, we identified the Alanine (A) at position E 230 in only two of our sequences studied, belonging to the Sub-Clade Ib. All other sequences have a T at this position. Analyzing all 240 sequences of the Sub-Clade I, only those two sequences and one from Argentina from 2016 (KX768377), presented this amino acid.

Discussion and Conclusion
The Sub-Clade Ia had Argentina as the gateway to Brazil, spreading to the Northeast, Midwest and Southeast regions of the country. The observation of Argentina as an ancestor for the Brazilian strains, corroborate our previous study 15 . With the introduction in the Southeast region, we observed most of the dispersions of the viral strain within Brazil, as well as returning to Argentina, which circulated at least until 2016.
The viral spread from the Midwest region to China, probably refers to an imported case. At Genbank, that Chinese strain does refer to an imported case study, but it has not been published. However, the results obtained here suggest this probable importation from Brazil, more specifically, from the Midwest region.
Sub-Clade Ib presented an introduction route in Northeast Brazil starting directly from Venezuela, as already demonstrated previously 15 . This clade seems to have remained circulating only in the North of the country, with a punctual dispersal event for the French Guianas. However, we observed a more recent movement of this clade The molecular characterization of the 30 strains allowed the identification of different nucleotides patterns responsible for the conservation of some amino acids in the domains I, II, III and stem-region of the E gene among the different strains and the changes in domain II and one in domain III, were not previous identified. In domain I, at position E 297 , the amino acid M was observed in all 30 sequences studied, different from the sequence of the 80s. In domain II, we identified an A at position E 230 in only two of our studied sequences, belonging to Sub-Clade Ib. All other sequences have a T at this position. Analyzing all 240 sequences of Clade I, only those two sequences and one from Argentina 2016 had this amino acid. During a study conducted in Argentina from December 2015 to April 2016, this amino acid was not reported 18 . In fact, in all the dataset analyzed here, the only sequences that present this shared amino acid are two of our study detected in 2014 and 2015 and the one from Argentina, detected afterwards in that country. Preserved amino acids were observed in domain III and in the stem region. The sequences of Sub-Clade Ia presented in E 394 an R instead of K, like the other sequences of Sub-Clade I and Clade II. However, Clade II sequences showed the largest differences among shared amino acids.
Recently, Dutra and collaborators investigated circulating DENV-1 strains in an epidemic in the state of Minas Gerais, and reported the co-circulation of two distinct lineages in 2013 and five independent introductions of genotype V in the country since 1982 19 . The nomenclature of the different DENV-1 lineages circulating in Brazil is yet to be defined. In our study, we designated the clades with roman numerals I and II, and letters for the lineages within a clade. However, Dutra et al. 19 suggest the standardization of BR1 to BR5 to the lineages, according to the chronological order of introduction in country.
In conclusion, our data show that DENV-1 remains circulating in the country through three distinct lineages, introduced by independent pathways in the last two decades. One of the lineages seems to be restricted to the North region of the country, while the other two are more dispersed by Northeast, Southeast and Midwest. Continuous surveillance is necessary, as new viral variants may arise by local evolution or by introductions from other countries. New viral strains may present nucleotide and/or amino acid changes in viral genome important for infection and replication, impacting the dynamics of disease maintenance and transmission.  (Table 1). Virus isolation was performed by inoculation into C6/36 Aedes albopictus cell line 20 and isolates were identified by indirect fluorescent antibody test (IFAT) using serotype-specific monoclonal antibodies 21 .   22 that detects all four serotypes simultaneously in a semi-nested procedure, generating amplification products with specific size in base pairs (bp) of each DENV serotype, was used to confirm the DENV-1 strains positivity. , Southeast (n = 62), South (n = 1), Northeast (n = 38) and North (n = 42) and one with no information (Table S2). GenBank accession numbers, countries of origin and year of isolation of all included sequences are shown in Table S1. Nucleotide sequences were aligned using MAFFT v6.902b program and the alignments may be available from the authors upon request.

Spatiotemporal dispersion of current lineages of DENV-1 genotype V in Brazil.
The rate of nucleotide substitution per site per year (subs./site/year), the time to the most recent common ancestor (T MRCA ) and the spatial diffusion were jointly estimated for the DENV-1 genotype V lineages using the Bayesian Markov Chain Monte Carlo (MCMC) statistical framework implemented in the BEAST v1.8 package 28 with BEAGLE 29 to improve run performance. The DENV-1 genotype V sequences were subdivided into three subsets that contained Brazilian sequences until 2016: Sub-Clade Ia (n = 106), Sub-Clade Ib (n = 56) and Clade II (n = 116) ( Table S2). The spatiotemporal scale of evolutionary process for analysis was directly estimated from the sampling dates of the sequences using the GTR + Г4 (Sub-Clade Ia), TN93 + Г4 (Sub-Clade Ib) and TN93 + I + Г4 (Clade II) nucleotide substitution model as determined by jModelTest program 30 , a relaxed uncorrelated lognormal molecular clock model 31 , a Bayesian Skyline coalescent tree prior 32 and reversible discrete phylogeography model 33 .