Evolutionary analysis and immunoinformatic-based epitope prediction of dengue virus serotype 2 strains from Sri Lanka

Methods

This Bioinformatics Study was undertaken by the Genetics and Molecular Biology Unit of the University of Sri Jayewardenepura, Sri Lanka.

Results

Phylogenetic analysis

Both phylogenetic trees (Figs. 1 and 2) displayed strain segregation into two major lineages of the cosmopolitan genotype: major lineage F (2II_F, denoted in red), containing Sri Lankan sequences from 2016-2022, and major lineage A (2II_A, denoted in green), containing Sri Lankan sequences from 2003 and 2004. A clear geographical clustering of sequences from the same country or neighbouring countries was observed in both phylogenetic trees, and both trees followed similar topology with moderate to strong support values.

Close

Fig.1.

Calibrated maximum-clade-credibility tree for DENV-2 using both partial and complete envelope gene sequences. The GTR model was used for this 1500 base pair dataset of the envelope gene fragment. The numbers above each branch represent the posterior probability (PP) values obtained in the Bayesian Index (1.00). Yellow fever virus was used as the outgroup.

Export to PPT

Close

Fig. 2.

Calibrated maximum-clade-credibility tree for DENV-2 using complete envelope gene sequences. The GTR model was used for this 1500 base pair dataset of the envelope gene fragment. The numbers above each branch represent the posterior probability (PP) values obtained in the Bayesian Index (1.00). Yellow fever virus was used as the outgroup.

Export to PPT

A clear segregation of all 2II_A DENV-2 strains into two distinct clades were observed in phylogenetic tree I, illustrating strong and moderate support (PP 0.78 and 0.62) for each of the clades respectively (Fig. 1), while the segregation of all 2II_A DENV-2 strains into a single clade (PP 0.63), illustrating moderate support for the clade, was visible in phylogenetic tree II (Fig. 2).

Clustering of the Sri Lankan strains from 2003 and 2004 (FJ502920 and FJ502914) with a Chinese strain from 2023 (OR418375) and two Indian strains from 2001 and 2013 (DQ448237 and KP940439) was visible in phylogenetic tree I (Fig. 1). Further, sub clustering of a Sri Lankan strain (KX778743) with isolates from Pakistan from the years 2009 and 2011 (KF041230, KP757138, JX042510) was observed in phylogenetic tree I (PP 0.78), illustrating a strong support for the sub clade. Since the Sri Lankan strains of the major lineage A were partial sequences coding for the envelope gene region, none of the above Sri Lankan strains were included in phylogenetic tree II (Fig. 2).

The distinct clustering of 2II_F strains into a single clade was clearly observed in both phylogenetic trees with a moderate support (PP 0.63) and a strong support (PP 0.7), respectively. The clade included the reported strains of DENV-2 from countries of the Southeast Asian Region, the Western Pacific Region, the Eastern Mediterranean Region, and the Reunion Island, including the Sri Lankan strains from 2016-2022. Further sub-clustering of DENV-2 strains with a strong to moderate support was also visible within this major clade in both trees.

The sub clustering of Sri Lankan strains reported in 2016 and 2017 (KY495803 and MZ520913, respectively) with the strains reported from China in 2015 (MG737964), 2016 (ON911798) and 2017 (MG840627, MW365262 and MH827551), Myanmar in 2016 (MW332627), and Taiwan in 2016 (MG895144) were observed in both trees with a strong support value (PP 0.78). Further sub-clustering of Sri Lankan strains (KY495803 and MZ520913) reported in 2016 and 2017, with two Chinese isolates from 2017 (MG840627 and MW365262) and the Myanmar isolate from 2016 (MW332627), were clearly visible in both trees with strong support values of PP 0.97 and PP 0.98, respectively.

The 2017 Sri Lankan isolates displayed distinct sub-clustering patterns in both phylogenetic trees. Specifically, isolate MT006167 clustered with a 2017 Reunion Island isolate (OR236093; PP 0.74 in Fig. 1, PP 0.75 in Fig. 2), while LC312197 clustered with a 2017 Chinese isolate (MH827548; PP 0.64 in Fig. 1, PP 0.70 in Fig. 2). Isolate MN577549 grouped with a 2018 Malaysian strain (PP124252; PP 0.94 in both figures), and MN602605 clustered with sequences from the East Asian and Western Pacific regions (PP 0.69 in Fig. 1, PP 0.64 in Fig. 2).

Furthermore, sub-clustering of Sri Lankan strains reported in 2018 and 2022 (MT006186 and OQ102952, respectively) with a 2023 Chinese isolate (PP564550) was observed in both trees (PP 0.65 and 0.85, respectively). Sub clustering of Sri Lankan strains reported in 2019 (PP434968, PP234965) and 2020 (OR394053) with a Chinese strain reported in 2019 (OP684213) was also evident, with a PP value of 0.98.

Discussion

This study characterised the phylogenetic and evolutionary dynamics of DENV-2 strains circulating in Sri Lanka from 2003 to 2022. Phylogenetic analyses revealed that all Sri Lankan isolates clustered within the cosmopolitan genotype, segregating into two major lineages: lineage A (2II_A) for strains from 2003-2004 and lineage F (2II_F) for strains from 2016 onward, suggesting a lineage shift around late 2016. The 2017 outbreak strains showed close phylogenetic relationships with Southeast Asian isolates from 2014-2016 and significant clustering with Chinese strains, indicating sustained regional viral transmission. Notable phylogenetic associations with isolates from Reunion Island (2017) and Malaysia (2018) indicate active viral gene flow across geographically diverse regions. Selection pressure analysis identified predominantly purifying selection across the envelope gene, with two novel positively selected codon sites (159 and 182). Epitope prediction revealed that several positively selected sites, including the novel site 159, encoded B-cell and T-cell epitopes with high probable antigenicity, suggesting critical roles in immune evasion and viral adaptation.

The findings are consistent with and extend previous research on DENV-2 evolution and transmission dynamics. The co-circulation of all DENV serotypes in Sri Lanka^5,6 has established DENV-2 as a significant contributor to dengue-associated morbidity and mortality. The persistent dominance of the cosmopolitan genotype is consistent with global phylogenetic studies,^17-21 with its widespread distribution across WHO’s Southeast Asian, Western Pacific, and Eastern Mediterranean Region, reflecting the virus’s remarkable adaptive capacity across different environmental contexts.

The observed lineage shift in 2016²² corroborates with independent investigations²³ showing that 2017 Sri Lankan isolates form a distinct clade separate from 1985 – 2004 isolates, reinforcing the hypothesis that this transition contributed to outbreak severity. Following the emergence and dominance of DENV-1 from 2009 to mid-2016,^8,24 the DENV-2 reintroduction in 2016 marked a major serotype shift with profound implications for population-level immunity dynamics. The 2017 outbreak resulted in 186,101 suspected cases and 440 deaths, with DENV-2 as the predominant serotype.^8,23,25 Phylogenetic sub-clustering of 2017 isolates with Southeast Asian strains from 2014-2016 suggests introduction of multiple DENV-2 variants.¹⁹ The quasi-species nature of the virus,²⁶ shaped by diverse evolutionary mechanisms, likely facilitated the emergence of multiple viral strains across different regions. Consequently, the concurrent circulation of multiple DENV-2 strains, combined with the serotype shift and a potential lineage transition, may have created conditions conducive to antibody-dependent enhancement-mediated disease severity,²⁷ contributing significantly to the 2017 outbreak magnitude.

The close phylogenetic clustering of Sri Lankan DENV-2 strains with Chinese isolates, correlates with a documented 72.5% increase in Chinese tourist arrivals to Sri Lanka during 2010-2016,²⁸ positioning China as the third-largest tourist source to the country. The intensification of bilateral tourism likely facilitated viral transmission, providing phylogenetic evidence for the role of international travel in dengue dissemination. The identification of imported DENV cases in Reunion Island in 2017, including sequences from Sri Lanka^29,30 supports an indirect epidemiological link between the 2017 Sri Lankan outbreak and the 2018 Reunion Island outbreak. These transmission patterns represent active viral gene flow facilitated by Sri Lanka’s strategic geographical position at the crossroads of major Indian Ocean sea routes and the expanding international air travel.^19,31 The clustering of 2017 Sri Lankan strains with strains from major aviation hubs (Singapore, Malaysia, India, China, and Taiwan) from 2013-2016 suggests that air travel and human migration have been pivotal in transmitting viral strains from Southeast Asian and Western Pacific regions. Additional transmission routes may include inadvertent transport of infected mosquito eggs and larvae through international trade and the silent transmission via asymptomatic carriers.^32,33

The predominance of purifying selection in the envelope gene aligns with previous evolutionary studies,^19,34,35 reporting the conserved nature of DENV-2 structural proteins.³⁴ However, this study advances the field by identifying two novel positively selected sites (159 and 182) not previously documented. The identification of epitope-bearing codons under positive selection extends previous immunological studies by directly linking evolutionary pressures to functional antigenic regions. The envelope protein plays a critical role in mediating viral attachment and host cell entry and serves as a major target for neutralising antibodies.³⁶ Therefore, the positioning of positively selected codons within this region may facilitate immune evasion, suggesting natural selection to ensure viral survival. Identifying these epitope-bearing codons provides functional insight into the evolutionary patterns in circulating DENV-2 strains, as sites under positive selection that overlap with immunogenic regions are likely shaped by antibody or T-cell-mediated immune responses. This work provides novel insights for rational vaccine design strategies and establishes a foundation for future peptide-based vaccines or therapeutic strategies against DENV infections.

This study provides a comprehensive phylogenetic analysis of Sri Lankan DENV-2 strains spanning two decades, offering valuable insights into viral evolution, lineage dynamics, and international transmission patterns. However, several limitations warrant consideration. The relatively small dataset may have limited the detection of additional positively selected sites, and incomplete temporal representation may not fully capture viral evolution dynamics. Additionally, phylogenetic inferences regarding transmission routes remain correlative and require experimental validation. Future studies incorporating larger sample sizes, whole-genome sequencing, and functional assays would provide more robust insights into evolutionary mechanisms driving DENV-2 circulation in Sri Lanka.

The findings underscore the need for enhanced genomic surveillance systems in Sri Lanka to track viral evolution in real time. Given the strong correlation between international travel and viral transmission, targeted vector control at ports of entry and strengthened surveillance of travellers from endemic regions should be prioritised. The novel positively selected epitopes identified, particularly at codon site 159, provide valuable targets for developing region-specific diagnostics and vaccine candidates. The documented association between serotype shifts and severe outbreaks highlights the importance of maintaining robust serotype-specific surveillance. Regional collaboration for data sharing across affected countries remains essential for early outbreak detection and effective dengue control.