Introduction

Within the Culicidae family, mosquitoes comprise over 3600 species, categorized into two subfamilies and 113 genera1,2. Among these genera, the Aedes genus, specifically the Stegomyia subgenus, hosts numerous vector species3,4,5,6. As an example, Aedes (Stegomyia) aegypti, belonging to the Aegypti Group, is a major vector of dengue virus (DENV) and can also transmit several other arboviruses, including chikungunya (CHIKV), Yellow fever (YFV) and Zika viruses (ZIKV)5. Another significant species within the Stegomyia subgenus is the invasive Asian tiger mosquito, Ae. (Stegomyia) albopictus. This species can transmit these four arboviruses and has also been associated with the transmission of more than 30 other arboviruses6,7,8.

Aedes albopictus belongs to the Scutellaris Group, which is subdivided in two Subgroups: Albopictus and Scutellaris. This group currently comprises a total of 46 species2 with at least 12 confirmed or potential vectors of arboviruses. For example, Ae. (Stegomyia) malayensis, found in Southeast Asia9, is a vector of DENV10 as well as Ae. (Stegomyia) scutellaris11, which is present in Southeast Asia and in the Pacific region9,12. Aedes (Stegomyia) polynesiensis, present in the eastern Pacific region12, can transmit CHIKV and ZIKV3,4 and is suspected to transmit DENV13. Aedes (Stegomyia) hensilli, an endemic species of Yap island, has the potential to transmit CHIKV and ZIKV14,15. Several other species exclusively found in the Pacific islands, including Ae. (Stegomyia) cooki, Ae. (Stegomyia) hebrideus, Ae. (Stegomyia) kesseli, Ae. (Stegomyia) marshallensis, Ae. (Stegomyia) pseudoscutellaris, Ae. (Stegomyia) rotumae, and Ae. (Stegomyia) tabu, are also suspected to be potential vectors of DENV16.

Understanding mosquito phylogeny holds significant importance for different reasons. It can significantly contribute to define their taxonomy and classification. The classification within the Scutellaris Group species remains a topic of debate, due to their morphological resemblances and potential hybridization of some species observed in laboratory settings without a decrease in fertility17,18. Consequently, exploring their relationships through molecular investigations can be a crucial tool in solving these taxonomical issues. Dating phylogeny could allow a better understanding of the evolutionary history of these species, helping to identify the period of their speciation events.

Few studies have investigated the phylogenetic relationship among the species within the Scutellaris Group, primarily emphasizing slight genetic distinctions between some species even with the use of ribosomal (the internal transcribed spacer 2, the 16S and 28S ribosomal sequences) or mitochondrial (the cytochrome c oxidase subunit 1) genes19,20,21.

The understanding of the Scutellaris Group evolution is also limited due to the limited number of species used in previous investigations. When using mitochondrial data, these works have particularly emphasized the divergence among Ae. polynesiensis and Ae. riversi and between Ae. albopictus, Ae. flavopictus and Ae. subalbopictus during the Paleogene period (66–23 million year ago), along with the divergence between Ae. dybasi and Ae. palauensis during the Neogene period (23–2.58 million years ago)22,23.

In contrast to previous studies, our current research includes broader range of species within the Scutellaris Group from Asia and Pacific regions. Utilizing new sequence data, we assessed the genetic diversity among species of the Scutellaris Group. Furthermore, our investigation aims to estimate speciation time of these species, to better understand their evolutionary history.

Methods

Biological material

Adults and larvae mosquitoes (n = 42) belonging to the Scutellaris Group, were collected in five countries: Cambodia, Fiji, Laos, New Caledonia and Wallis & Futuna (Table 1). Field collections took place between March 2016 and June 2022. Various sampling methods, including the BG-Sentinel® trap, BG-Gravid® Aedes trap (Biogents, Germany), and CDC light trap with CO2 (BioQuip, USA), were employed for adult collection. Larvae were collected directly from their breeding sites using water pipettes and plastic cups and reared in the laboratory until adults at 28 °C ± 2 °C and a relative humidity of 80 ± 10%. They were fed with Innovafeed ad libitum throughout this rearing period. Upon emergence, adults were collected using mouth aspirator and killed by freezing at – 80 °C. Morphological identification of all collected mosquitoes was performed under a stereo microscope using dichotomous identification keys12,17,24, followed by preservation at − 80 °C for subsequent analyses. Among these specimens, Ae. unalom, a newly discovered species within the Albopictus Subgroup, was included25. This species is morphologically close to Ae. albopictus but is genetically distinct. We also included Ae. scutellaris, a species that was recently introduced in New Caledonia26.

Table 1 Mosquitoes collected from each country.

DNA extraction, PCR and sequencing procedures

Following the method outlined by Rakotonirina et al.21, DNA was individually extracted from each mosquito's body using a DNA Blood and Tissue Kit (Qiagen, Hilden, Germany) in accordance with the manufacturer's guidelines. The extracted DNA underwent amplification via polymerase chain reaction (PCR) utilizing specific primers targeting mitochondrial genes which are widely used markers in evolutionary and taxonomic studies due to the lack of DNA recombination, simplifying inheritance patterns, in contrary to the nuclear gene27. We focused on cytochrome c oxidase subunits 1, 2, and 3 (cox1, cox2, and cox3) and cytochrome b, all known to be under strong purifying selection in mosquitoes27,28. However, during our study, technical limitations resulted in interpretable data only for cox1 and cox3. Consequently, we present our findings based on these two genes.

Cytochrome c oxidase subunit 1 gene is 1537 basepairs (bp). For technical reasons, three sets of cox1 primers were used during this study. The first set of primers (LCO1490: 5′-GGT CAA CAA ATC ATA AAG ATA TTG G-3′ and HC02198: 5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′)21 allow to amplify the 709 bp in the 5′ part of cox1, ranging from nucleotide 17 to nucleotide 725. The second cox1 primer set allowed amplifying 680 bp in the 3′ part of cox1, starting from nucleotide 688 to nucleotide 1367, with COS2183N (5′-CAR CAY YTA TTY TGR TTY TTY GG-3′) and COA3107N (5′-TCY ATT AAA GGA GAA GYW CTA TYT TG-3′) primers20. The last cox1 primer set starting at nucleotide 220 and ending at nucleotide 1508, generated a 1288 bp fragment using C1-J-1718 (5′-GGA GGA TTT GGA AAT TGA TTA GTT CC-3′) and C1R3 (5′-TAA GTA TGT TCT GCA GGA GG-3′) primers25,29. Moreover, cox3 gene is of 789 bp. The primers designed for this study (COIII-F: 5′-ACA GGG GCT ATT GGA GCT AT-3′ and COIII-R: 5′-GCT GCA GCT TCA AAT CCA AAA T-3′) enabled the amplification of 662 bp fragment, initiating from nucleotide 58 and extending to nucleotide 719.

PCR amplifications were conducted in 25 μl reactions, employing isolated DNA at concentrations between 10 ng/μl and 100 ng/μl as the template. Each PCR reaction comprised 0.2 μM of both forward and reverse primers, 1 × of HotStarTaq® DNA Polymerase (Qiagen), and distilled water. The thermal cycling protocol involved several steps: an initial denaturation at 94 °C for 3 min, followed by 35 cycles. During each cycle, there was denaturation at 94 °C for 1 min and annealing at specific temperatures: 50 °C for 1 min (for LCO1490/HC02198), 52 °C for 1 min (for COS2183N/COA3107N), 51 °C for 40 s (for C1-J-1718/C1R3), and 61 °C for 1 min (for COIII-F/COIII-R). After annealing, extension step per cycle occurred at 72 °C for 1 min (for LCO1490/HC02198, COS2183N/COA3107N, and COIII-F/COIII-R), while for C1-J-1718/C1R3, extension was at 72 °C for 1 min 30 s. Finally, a concluding extension step was performed at 72 °C for 10 min. Verification of double-stranded PCR products was done through 1.5% agarose gel electrophoresis, and subsequently, PCR products underwent sequencing using the Sanger technique (conducted by Genoscreen, France, or Macrogen, Korea).

Data analyses

The sequences obtained were initially assembled using PREGAP and GAP software (version 4.10.2)30. Following sequence quality control using the Bioedit software (version 7.2.5)31, the sequences were aligned, and the number of polymorphic sites for each gene was determined using the DNAsp (version 6.12) program32. A haplotype network for the concatenated gene was constructed using the Templeton Crandall and Sing (TCS) algorithm33 within popART (version 1.7)34.

The divergence dating analysis relied on a Bayesian Markov Chain Monte Carlo (MCMC) approach, conducted via BEAST (2.6.0 package)35. This aimed to infer the Scutellaris Group's topology and estimate the speciation time of the common ancestor of clades in millions of years (MYA). To achieve this, we used the cox1 gene due to its high representation in the GenBank database for mosquito species. Available cox1 sequences from various species of the Scutellaris Group (Ae. albopictus, Ae. daitensis, Ae. downsi, Ae. dybasi, Ae. flavopictus, Ae. futunae, Ae. galloisi, Ae. hensilli, Ae. malayensis, Ae. miyarai, Ae. palauensis, Ae. polynesiensis, Ae. pseudoscutellaris, Ae. riversi, Ae. scutellaris and Ae. unalom) originating from Asia and the Pacific region were retrieved from the GenBank database and were added to our newly sequenced data. We have removed sequences containing extensive ambiguous “N” symbols based on alignment analysis using the Bioedit software (version 7.2.5)31. A total of 1849 cox1 sequences were included in the analysis (Supplementary Table 1). Aedes aegypti, Aedes notoscriptus and Ae. vigilax were used as outgroups. The full list of cox1 sequences used in this study is available as Excel file (Supplementary Table 1). Two calibration points, representing the speciation event between Ae. polynesiensis and Ae. riversi, and between Ae. notoscriptus and Ae. vigilax were adopted based on previous research within the Culicinae subfamily, using mitochondrial full genome data22.

The best evolutionary model was selected by the ModelFinder36, followed by the tree reconstruction using the IQ-TREE version 1.6.1237 performing the ultrafast bootstrapping with 1,000 replicates. The Bayesian analysis comprised runs of 10 million generations, with tree sampling at every 1000 iterations for each alignment dataset. It used a Birth–Death process of speciation as the Tree Prior. Furthermore, the GTR + G + I evolutionary model was employed. The analysis of these two parts of the cox1 gene was conducted separately because the GenBank database only had sequences for most species in either the 5' or 3′ position. After discarding the initial 25% burn-in, the tree was obtained through the TreeAnnotator program35. For each speciation time, we calculated the highest posterior density (HPD), which represents a set containing 95% of the sampled values (95% HPD). Finally, the consensus trees were visualized and edited using FigTree (version 1.4.4)38.

Results

Sequence variation among the Aedes of Scutellaris Group

All the sequences newly generated during our study (40 cox1 and 42 cox3 sequences) were deposited in the GenBank database (accession numbers PP355451 to PP355490 and PP357893 to PP357934). This study provided the first cox3 sequences of six Aedes species of the Scutellaris Group: Ae. futunae, Ae. malayensis, Ae. polynesiensis, Ae. pseudoscutellaris, Ae. scutellaris and Ae. unalom.

After aligning the sequences and trimming the ends to ensure uniform sequence length across all individuals, the lengths of both the cox1 and cox3 regions decreased. Within the 5′ part of cox1 sequence (420 bp), a total of 102 nucleotide substitutions and 87 polymorphic sites were identified. Few nucleotide mutations were observed among different Ae. albopictus populations from Asia and Pacific countries (two–three mutations), as well as between Ae. polynesiensis from Wallis & Futuna and Ae. pseudoscutellaris from Fiji (five mutations), and between Ae. malayensis from Laos and Ae. scutellaris detected in New Caledonia in 2016 (two mutations) (Supplementary Table 2).

The 3′ part of cox1 gene (493 bp) presented 90 nucleotide substitutions and 79 polymorphic sites. Similar to the 5′ part of cox1 sequence, few nucleotide mutations were identified across the different Ae. albopictus populations (two mutations), between Ae. polynesiensis and Ae. pseudoscutellaris (two mutations), and between Ae. malayensis and Ae. scutellaris (five mutations) (Supplementary Table 3).

Within the cox3 gene sequence (553 bp), 110 mutations occurred with 101 polymorphic sites. Only a few mutations (three–four mutations) were detected among various Ae. albopictus populations. Only three mutations were identified between Ae. polynesiensis and Ae. pseudoscutellaris (Supplementary Table 4).

When the cox1 and cox3 genes (1466 bp) were concatenated, 26 haplotypes among the 42 genotypes were identified (Fig. 1). The Albopictus Subgroup was composed of a mix of haplotypes of Ae. albopictus from Cambodia, Fiji, and Laos and two haplotypes of Ae. unalom (Fig. 1). The Scutellaris Subgroup was grouped in three distinct clusters. The first cluster contained two haplotypes of Ae. futunae from Wallis and Futuna. The second cluster was consisted of a combination of haplotypes of Ae. polynesiensis from Wallis and Futuna and Ae. pseudoscutellaris from Fiji. In this cluster, the distinction between Ae. polynesiensis from Wallis and Futuna and Ae. pseudoscutellaris from Fiji was ambiguous even when using the concatenated mitochondrial data (Fig. 1). The third cluster was composed of a mix of haplotypes of Ae. malayensis from Laos and Ae. scutellaris from New Caledonia.

Figure 1
figure 1

Haplotype network of species of Scutellaris Group (n = 42) using the concatenated cox1 and cox3 genes (1466 bp). Nodes in the graph represent distinct haplotypes, and their size indicate the number of sequences within each haplotype. The number in brackets represent the number of mutations between different nodes or haplotypes.

Estimating divergence times

When examining the 5′ segment of the cox1 gene (379 bp after aligning our sequences with those from the GenBank database and trimming the extremity to have the same sequence length), the evolutionary timeline revealed that the divergence of the last common ancestor between Ae. aegypti (one of the species used as outgroup) and the Aedes of Scutellaris Group dated back to approximately 83 MYA (95%HPD: 61–108) (Fig. 2). The Albopictus and Scutellaris Subgroups were monophyletic, and speciation process between these two subgroups occurred around 61 MYA (95%HPD: 50–71).

Figure 2
figure 2

Evolutionary timescale of the Scutellaris Group using the cox1 gene. Tree derived from 1710 cox1 nucleotide sequences (379 bp) of the 5’cox1 gene and generated from BEAST program using the GTR + G + I evolutionary model. Aedes aegypti, Ae. notoscriptus and Ae. vigilax were used as outgroups. Two calibration points were used from prior Culicinae subfamily studies using mitochondrial full genomes: Ae. polynesiensis-Ae. riversi, and Ae. notoscriptus-Ae. vigilax speciation events22. For analysis, the percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown for values over 60 (in red).

The last common ancestor of the Albopictus Subgroup dated back approximately 51 MYA (95%HPD: 41–60). In this subgroup, the distinction between Ae. flavopictus and Ae. miyarai was ambiguous: we observed three distinct clusters of Ae. flavopictus mixed with one single cluster of Ae. miyarai.

Within the Scutellaris Subgroup, we identified two monophyletic clusters for which the last common ancestor dated back approximately 35 MYA (95%HPD: 32–35). The first cluster is mainly composed of species from Asia (Ae. daitensis, Ae. malayensis, Ae. riversi), except Ae. scutellaris, which was introduced to New Caledonia. The second cluster is exclusively composed of species from the Melanesia and Polynesia of the Pacific region (Ae. futunae, Ae. polynesiensis and Ae. pseudoscutellaris). The speciation of these two clusters took place nearly at the same period: around 20 MYA (95%HPD: 13–29) for the cluster comprising Ae. daitensis, Ae. malayensis, Ae. riversi and Ae. scutellaris, and approximately 25 MYA (95%HPD: 16–33) for the cluster composed of Ae. futunae, Ae. polynesiensis and Ae. pseudoscutellaris (Fig. 2). The analysis of the 5’segment of the cox1 gene did not allow to distinguish Ae. malayensis and Ae. scutellaris: three clusters (containing a mix of the two species) were observed, with speciation occurring approximately 11 MYA (95%HPD: 7–14). Also, the analysis of this gene did not allow to distinguish Ae. polynesiensis and Ae. pseudoscutellaris, the last common ancestor of these two species dated back approximately 4 MYA (95%HPD: 1–5).

The analysis of the 3′ part of the cox1 gene (486 bp after aligning our sequences with those from the GenBank database and trimming the extremity to have the same sequence length) revealed sequence differences compared to the 5′ fragment, influencing on the resulting phylogenetic tree structure. Contrary to the observation when analyzing the 5′ part of the cox1 gene, the Albopictus Subgroup formed a paraphyletic group, while the Scutellaris Subgroup was monophyletic (Fig. 3). The most recent common ancestor of Ae. albopictusAe. flavopictusAe. unalom of the Albopictus Subgroup and the species of Scutellaris Subgroup dated back around 64 MYA (95% HPD: 46–87). We also identified two monophyletic clusters within the Scutellaris Subgroup for which the last common ancestor dated back approximately 36 MYA (95%HPD: 32–50). The first cluster is mainly composed of a mix of species from Asia and Micronesia (Ae dybasi, Ae. hensilli, Ae. malayensis, Ae. palauensis and Ae. riversi), and Ae. scutellaris introduced in New Caledonia and their last common ancestor dated back approximately 13 MYA (95% HPD: 7–21). The second cluster included only species from Melanesia and Polynesia of the Pacific region (Ae. futunae, Ae. polynesiensis and Ae. pseudoscutellaris) for which the last common ancestor dated back around 27 MYA (95% HPD: 20–32). The distinction between Ae. polynesiensis and Ae. pseudoscutellaris was ambiguous.

Figure 3
figure 3

Evolutionary timescale of the Scutellaris using the cox1 gene. Tree derived from 167 cox1 nucleotide sequences (486 bp) of the 3′cox1 gene and generated from BEAST program using the GTR + G + I evolutionary model. Aedes aegypti, Ae. notoscriptus and Ae. vigilax were used as outgroups. Two calibration points were used from prior Culicinae subfamily studies using mitochondrial full genomes: Ae. polynesiensis-Ae. riversi, and Ae. notoscriptus-Ae. vigilax speciation events22. For analysis, the percentage of replicate trees in which the associated taxa clustered together in the bootstrap test (1000 replicates) is shown for values over 60 (in red).

Discussion

This study explored the speciation patterns of the species within the Scutellaris Group by analyzing their mitochondrial genes. Our results showed that the speciation process of these species happened long before human migration. Human presence in the Asia–Pacific region began only 55–65 thousand years ago, with subsequent dispersal within the region39, whereas the radiation of Scutellaris species appeared from approximately 60 MYA. The significant temporal distinction between the evolutionary timeline of these species and human migration dismisses any correlation between human colonization from Asia to the Pacific region and the speciation of the Scutellaris Group, which was suggested in the 1960s17. Our research revealed a broader pattern of diversification for many other Scutellaris species, previously uninvestigated.

Due to the lack of complete genome data and limited availability of other mitochondrial genes in GenBank for most Scutellaris Group species, we used only cox1 gene fragments. This approach allowed us to incorporate existing sequences into our study. Despite using partial cox1 genes, our estimated dates closely align with the previously established timelines. Indeed, prior studies using various genetic markers, including mitochondrial DNA, nuclear DNA, and whole genome data, estimate the divergence time between Ae. aegypti and Ae. albopictus to be roughly between 81 and 67 MYA22,23,40. In our study, we observed a divergence between these two species at around 84 MYA and 77 MYA using the 5′cox1 and the 3′cox1 genes, respectively. Previous work has also identified the speciation between Ae. albopictus and Ae. flavopictus at between 32 and 28 MYA when using the concatenated mitochondrial and nuclear genes or the mitochondrial genes23. In our study, we observed a separation between these two species at approximately 18 MYA (95%HPD: 11–28) when analyzing the 5′cox1 or 44 MYA (95%HPD: 27–62) when using the 3′cox1 region. The consistency between our results and previous findings can confirm the reliability of our analyses.

We noted a variation in speciation dates, particularly within the Albopictus Subgroup, when comparing the 5′cox1 gene (which has become the standard barcoding marker for insects41) and the 3′cox1 results. These variations could be attributed to the difference in nucleotide substitution patterns between these regions, previously highlighted in other insect families42, and also attributed to the different number of taxa and sequences included in the analyses. However, our results indicated overlapping dating intervals (95% HPD) for the two gene portions, particularly evident within the Scutellaris Subgroup, suggesting no significant difference between the dating results. For example, the speciation between Ae. futunae and the cluster Ae. polynesiensis/Ae. pseudoscutellaris, dated back to approximately 25 MYA (95%HPD: 16–33), when analyzing the 5′ cox1 and around 27 MYA (95%HPD: 20–33), when analyzing the 3′ cox1. Also, the speciation between Ae. riversi and the cluster Ae. malayensis/Ae. scutellaris was identified at approximately 11 MYA (95%HPD: 6–13) with the 5′ cox1 gene while the radiation between the cluster Ae. riversi/Ae. palauensis/Ae. dybasi and the cluster Ae. hensilli/Ae. malayensis/Ae. scutellaris was also estimated at around 13 MYA (95%HPD: 7–21) with the 3′cox1 gene.

Overall, our results demonstrated that most of the speciation occurred during the Paleogene and Neogene periods (66 MYA to 23 MYA). Previous works suggested that mammalian diversification began before the Cretaceous–Tertiary (K/T) boundary (before 66 MYA)43,44. The peak of mammalian diversification rate is estimated to have occurred between 33 MYA, to 30 MYA, and diversification rates remained high and constant until 8.55 MYA, before declining significantly45. A co-radiation between Aedes species of Scutellaris Group and mammals might, therefore, have occurred. However, further data and in-deep studies are necessary to provide enough evidence of this possible co-evolution.

Significant fluctuations in global climate and environmental perturbations due to an asteroid impact have been identified as a cause of mass extinctions, leading to the extinction of dinosaurs 66 MYA46,47 . This asteroid impact was followed by extensive volcanic activity, known as the Deccan Traps, which influenced the atmospheric CO2 levels48,49. Elevated atmospheric CO2 levels have previously been associated with increased speciation in Culicidae50. Interestingly, our analysis identified a separation between the Albopictus and the Scutellaris Subgroup occurring just after this massive extinction event, estimated to have taken place around 64 MYA or 61 MYA.

In Eurasia and the Pacific region, the Paleogene and Neogene periods were characterized by geological dynamics that influenced the landscape51,52,53. One significant geological event was the creation of Palau's volcanic arc system in Micronesia. This happened about 40 MYA when the Pacific Plate moved under the Philippine Plate along the Kyushu-Palau Ridge54. At the same period, changes occurred along the Eurasian coastline, leading to the creation of the Philippine Plate and the Izu Arc, a volcanic chain in the western Pacific Ocean. This later event played a role in shaping Japan and forming the Japan Sea and Okhotsk Sea55. Within the Scutellaris Subgroup, we identified a separation after these periods. Indeed, a divergence between some species from Asia–Micronesia and Melanesia-Polynesia occurred 36 -35 MYA involving two clusters: the first cluster comprised Ae. daitensis and Ae. riversi (only known from Japan), Ae. dybasi Ae. hensilli, Ae. palauensis (present in Palau islands)2,12,20 and the second cluster consisted of Ae. futunae (endemic to Wallis and Futuna), Ae. pseudoscutellaris (only known from Fiji) and Ae. polynesiensis (widely distributed in the South Pacific)12.

Finally, we found low genetic divergence between Ae. polynesiensis and Ae. pseudoscutellaris, which corroborated with previous studies19,21. Even when using concatenated mitochondrial genes, our results demonstrated that the number of mutations observed between these two species was comparable to the mutations found among different Ae. albopictus populations. Researchers have hypothesized that these two species might either constitute a single species18 or are in the process of speciation19. To definitively determine the taxonomic status of these closely related species, which also exhibit morphological similarities12, a comprehensive genome sequencing approach would be essential. We, also, observed a difficulty in distinguishing Ae. malayensis and Ae. scutellaris in 5′cox1 region, when including in our analysis: Ae. malayensis sequences from Singapore10 and Ae. scutellaris from Thailand56. This ambiguity was not observed when using only the sequences generated from our laboratory: the phylogenetic tree generated from the 5′cox1 part was previously published in 202321. Additionally, ambiguity in separating Ae. flavopictus and Ae. miyarai was also highlighted in our study. These findings could be attributed to the specimens being collected from different locations or the potential misidentification of the species due to their morphological similarities.

Conclusion

Overall, we characterized the phylogeny of several Aedes species of the Scutellaris Group in the Asia–Pacific region. This research represents a significant step forward in understanding their evolution. Further studies with more extensive sampling and analyzing complete genome data should be done that will offer a more understanding of the evolutionary history of the Scutellaris Group.