Molecular Identification of Mosquitoes (Diptera: Culicidae) Using COI Barcode and D2 Expansion of 28S Gene

Abstract

The purpose of this study is to improve the identification of Culicidae species from the Vale Ribeira region, São Paulo state, Brazil. Adults were collected in the municipalities of Cananeia and Pariquera-Açu and morphologically identified. Molecular analyses were performed on sequences of COI barcode and a fragment of the D2 expansion of the 28S ribosomal RNA gene generated from field collected mosquitoes. The analyses included species delimitation, phylogeny, and interspecific genetic distances using the Kimura 2-parameter model. Species included in the analyses were Aedes perventor, Aedes scapularis, Aedes serratus/Aedes nubilus, Aedes serratus s.s., Aedes terrens, Haemagogus capricornii, Haemagogus leucocelaenus, Haemagogus janthinomys, Kerteszia bellatrix, Kerteszia cruzii, Psorophora ferox, Psorophora forceps, Sabethes conditus, and Wyeomyia confusa. COI sequences from specimens collected at other localities were included in the analysis for comparison. Results of barcode RESL analysis showed that specimens of Ps. ferox and Hg. janthinomys split into three clusters for each species. Similarly, sequences of Ke. bellatrix and Ke. cruzii were recovered in two groups for each species. Distinct from other species included in analyses, Ps. ferox and Ps. forceps shared 100% similarity in the D2 fragment sequenced. Overall, the analysis of COI barcode sequences revealed the following key findings: (1) the presence of subclades within Hg. janthinomys, with its division into three groups suggests that this species may represent a species complex; (2) Ke. bellatrix from the Atlantic tropical rainforest shares 95.59% sequence similarity with a specimen from the type locality, indicating that specimens from Southeastern Brazil may belong to an unidentified species within the Ke. bellatrix complex; (3) Ke. cruzii also represents a species complex; and (4) D2 sequences successfully identified most species studied, apart from Ps. forceps and Ps. ferox.

1. Introduction

The Culicidae family comprises 113 genera and 3726 species [1], with approximately 150 species implicated in the transmission of pathogens that significantly contribute to human morbidity and mortality [2]. Notably, around 70 species of the genus Anopheles are capable of transmitting Plasmodium species that are causative agents of malaria in humans [3]. Aedes aegypti and Aedes albopictus are key vectors for several arboviruses, including dengue, Zika, chikungunya, and yellow fever [4]. Species within the genus Haemagogus are critical vectors of the yellow fever virus in zoonotic cycles, with Haemagogus leucocelaenus identified as the primary vector in regions of Southeastern Brazil [5].

Various methods are employed to reduce the incidence of mosquito-borne diseases, with vector control being a key strategy [6,7]. A thorough understanding of the ecology, biology, and behavior of mosquito vector species is essential for designing effective and sustainable control programs [8]. Accurate species identification is critical for managing and developing targeted control measures, ensuring the success of vector control efforts. Identifying field-collected mosquitoes can be challenging due to potential damage to key characteristics during collection and storage, or because the vector species may belong to complexes that are morphologically similar. DNA sequencing is a valuable tool for distinguishing between morphologically similar species or specimens that have been damaged, providing precise and reliable identification.

Among the available DNA-based methods, mitochondrial and ribosomal genes are commonly used to identify or define species boundaries [9,10,11,12]. Although the COI barcode alone may not always be sufficient for identifying all species within certain subgenera of the Culex mosquitoes [13], it remains a useful tool for identifying other species [14,15,16]. Another valuable DNA tool is the D2 expansion of the 28S ribosomal RNA gene, which has been used to uncover phylogenetic relationships in various organisms, including triatomines [17], fungi [18], and mosquitoes [14,19], and for species identification through metabarcoding [11].

The Atlantic tropical rainforest biome harbors a high diversity of mosquitoes, including species that are potential vectors of arboviruses and other species that are key for the transmission of Plasmodium species [20]. The Vale do Ribeira region in Southeastern São Paulo state, Brazil, hosts the largest continuous expanse of preserved Atlantic tropical rainforest [21]. This region supports a rich mosquito community, including species that are vectors of several pathogens capable of infecting and causing diseases in humans [22,23,24]. Because of its diverse mosquito fauna, numerous vertebrate reservoirs of various pathogens, and a population frequently exposed to mosquito and other blood-feeding insect bites, the risk of emergence and spread of vector-borne diseases, such as bromeliad malaria, is high [25]. Additionally, there is a significant potential for the emergence of yellow fever [26], Rocio virus [27,28,29], and other viruses [30]. Given the high mosquito diversity and the potential for emerging vector-borne diseases in the Southeastern Atlantic tropical rainforest, this study aimed to identify the mosquito species collected in the region. To achieve this, we utilized the COI barcode and a fragment of the D2 domain of the 28S ribosomal RNA gene (large subunit RNA) to characterize specimens from various locations within the coastal Atlantic rainforest in Vale do Ribeira, Southeastern São Paulo State, Brazil. To ensure accurate identification, specimens from other geographical areas were also included in the analysis.

2. Materials and Methods

2.1. Mosquito Collections and Species Identification

Adult mosquitoes were collected at the following sites in preserved forest areas in Vale do Ribeira, Southeastern São Paulo State, Brazil: (1) locality 1 (24°47′28.8″ S, 47°54′42.6″ W), Pariquera-Açu municipality; and (2) locality 2 (24°55′09.05″ S, 47°58′12.6″ W), Cananeia municipality. The collections were performed using an entomological net and conducted daily from 8:00 to 14:30 h from 14 December to 17 December 2021 (Locality 2), and on 30 March 2022 (Locality 1). In addition, specimens of some species were collected from other geographical locations in Brazil (Table S1). Field-collected mosquitoes were euthanized with ethyl acetate (C₄H₈O₂) and stored dry in individual vials with silica gel. Specimens were maintained under these conditions until species identification and DNA extraction. Forattini’s [31] key was used for morphological identification. The nomenclature for species of the Anophelinae subfamily followed the system proposed by Foster et al. [10].

All mosquito collections were carried out under permit number 12583–1 from the Instituto Chico Mendes de Conservação da Biodiversidade (ICMBio, SISBIO) granted to MAMS.

2.2. DNA Extraction and Amplification

Genomic DNA was extracted individually from whole mosquitoes using a commercial kit (Table S1) according to the manufacturer’s instructions. The barcode region of the COI gene was amplified with primers LCO1490 5′-GGTCAACAAATCATAAAGATATTGG-3′ and HCO2198 5′-TAAACTTCAGGGTGACCAAAAAATCA-3′ [32] by following the same protocol employed by Torres-Gutierrez et al. [33]. PCR amplification of the D2 region of 28S rDNA was performed with a final volume of 20 µL containing 1 × GoTaq Master Mix (Promega) and 0.2 μM each primer (Ill+Mozzie.D2.Uni.F: 5′ TCGTCGGCAGCGTCAGATGTGTATAAGAGACAGAAGCACTCTGAATAGAGAGTC 3′; Ill+Mozzie.D2.Uni.R: 5′ GTCTCGTGGGCTCGGAGATGTGTATAAGAGACAGTGGTCCGTGTTTCAAGAC 3′) [11], 2–5 ng of DNA, and ultrapure water to fill the remaining volume. The thermocycler conditions consisted of 95 °C for 2 min, 30 cycles of 95 °C for 10 s, 65 °C for 30 s, 72 °C for 30 s, and a final extension at 72 °C for 5 min. PCR products were purified by PEG precipitation (20% polyethylene glycol 8000/2.5 M NaCl).

2.3. Sequencing and Alignment

The COI barcode of the mitogenome and the D2 domain of the 28S rRNA gene were sequenced in both directions using the Big Dye Terminator Cycle Sequencing Kit v3.1 (Applied Biosystems, Foster City, CA, USA) and the same set of PCR primers. Sequencing products were purified with Sephadex G50 columns (GE Healthcare, Chicago, IL, USA), and the amplified DNA fragments were separated by capillary electrophoresis in an Applied Biosystems 3130 DNA Analyzer (PE Applied Biosystems). Sequences were edited using BioEdit v7.2.5 software [34], and the primer regions were removed. Sequences of each region were aligned separately by nucleotide using the Muscle algorithm [35], implemented in MEGA v11.0.13 [36]. COI sequences downloaded from GenBank (accession numbers MF172270, MW339748, MF172347, MN997518, OQ272309, OQ272319, MT985984, KY859893, OQ272332, KU551284-KU551286, KU551289, MF172267, NC_044663, PP915652, PP915660, PP915683, and PP915685) and from specimens collected in other geographical localities in Brazil (Table S1) were included in the COI analyses.

2.4. Phylogenetic Analysis

Phylogenetic trees were generated using maximum-likelihood (ML) methods with ultrafast bootstrap values set for 1000 repetitions. These reconstructions were performed in IQ-tree 2 v. 2.2.0 [37] using the –m MFP option to choose the optimal selection model. Repeated sequences were removed, and only unique D2 and COI sequences were used in this analysis. The COI dataset was partitioned by codon positions.

2.5. Interspecific Genetic Distances and Species Delimitation Analyses

Calculations of the interspecific genetic distance of COI and D2 sequences were performed using the Kimura 2-parameter (K2P) distance model [38]. Two methods were employed for species delimitation using the COI barcode. The first method was Assemble Species by Automatic Partitioning (ASAP) [39], which used the Kimura (K80) ts/tv 2.0 substitution model. ASAP is based on pairwise genetic distances and does not require prior knowledge of species composition, the number of species, or any biological information. Pairwise genetic distances were used to obtain a list of the best partitions, which were ranked by a score calculated using the barcode gap width and the probability of panmixia. The second method involved the Refined Single Linkage (RESL) algorithm in BOLD systems (https://www.boldsystems.org/ accessed on 24 September 2024). This approach used sp-distances and an intraspecific threshold of 2.2% to cluster sequences into taxonomic units, generating BOLD BINs (Barcode Index Numbers) from the BOLD database. The D2 dataset was analyzed using only the ASAP method.

3. Results

A total of 137 collected mosquitoes were included in the analyses. The following species were identified: Aedes perventor (n = 8), Aedes scapularis (n = 10), Aedes serratus/nubilus (n = 9), Aedes serratus s.s. (n = 6), Aedes terrens (n = 2), Haemagogus leucocelaenus (n = 6), Haemagogus janthinomys (n = 17), Kerteszia bellatrix (n = 1), Kerteszia cruzii (n = 13), Psorophora ferox (n = 8), Psorophora forceps (n = 3), Sabethes conditus (n = 33), and Wyeomyia confusa (n = 21) (Table S1). The COI sequences were 658 bp long after removal of the primers. The length of the D2 sequences varied from 357 to 420 bp among the species. The D2 sequences from Ae. perventor were 357 bp. The D2 sequence lengths of Ae. serratus/nubilus were 370 bp long, Hg. leucocelaenus (372 bp), Ae. scapularis (373 bp), Ae. terrens (373 bp), Hg. janthinomys (378 bp), Ps. forceps (389 bp), Ps. ferox (389 bp), Wy. confusa (405 bp), Sa. conditus (409 bp), Ke. cruzii (419 bp), and Ke. bellatrix (420 bp).

The analyses included 108 sequences of the D2 domain of the 28S rRNA gene and 147 sequences of the COI barcode. A total of 128 COI sequences were newly obtained for this study (Table S1), and 19 were downloaded from GenBank (Table S1). These comprised three sequences of Ps. ferox, two of Ke. bellatrix, four of Hg. janthinomys, five of Ke. cruzii, two of Haemagogus capricornii, and one each of Ae. scapularis, Ae. serratus s.s., and Wy. confusa. The COI and D2 sequences generated in this study were deposited in GenBank with accession numbers PP823984-PP824111 (COI) and PP843945-PP844052 (D2).

3.1. Phylogenetic Analysis

Ninety-five COI sequences and twenty-five D2 unique sequences were used for phylogenetic analysis. For the COI analysis, the bootstrap support values for some lineages were >98% according to ultrafast bootstrapping, while for other splits, the bootstrap support values were <98% (ranging from 53% to 98%) (Figure 1). The results of ultrafast bootstrapping in the phylogenetic trees with D2 sequences (Figure 2) indicated that except for the tribe Sabethini lineage and the genus Kerteszia, all remaining lineages were poorly supported.

3.2. Interspecific Genetic Distances and Species Delimitation Analyses

A total of 147 sequences were analyzed using the COI barcode. Interspecific genetic distances of COI sequences are shown in Table S2 and ranged from 2.78% to 19.84%. Species delimitation using the COI distance-based ASAP method supported 13 potential taxa, while the RESL method delimited 19 clusters (Table S3). Both ASAP and RESL analyses identified a single lineage for Ae. perventor, Hg. leucocelaenus, Sa. conditus, Wy. confusa, Ae. terrens, and Ae. scapularis (Table S3). Aedes serratus s.s and Ae. serratus/nubilus were grouped in a single cluster. The K2P mean distance between Ae. serratus/nubilus from Vale do Ribeira and Ae. serratus s.s. exceeded 2.7% (Table S4).

Despite the phylogenetic tree of COI sequences and COI RESL analysis recovering Hg. janthinomys and Hg. capricornii in different two separate groups (Figure 1), the ASAP analysis showed that both species grouped in a single cluster (Table S3). In COI RESL, specimens of Hg. janthinomys were grouped into three clusters: one that included specimens from northern Brazil, Águas da Prata municipality (São Paulo state), Colombia, and Trinidad and Tobago (type locality); the second cluster was formed by specimens from Vale do Ribeira (São Paulo state); and the third cluster with only one specimen from the municipality of Águas da Prata (São Paulo state). The mean K2P distance between these clusters was less than 2% (Table S5).

The results of the COI RESL analysis showed that Ke. bellatrix and Ke. cruzii were separated, with each species composed of two distinct clusters (Table S3). For Ke. bellatrix, one cluster included specimen from Vale do Ribeira (BOLD BIN: AAF0614) (Ke. bellatrix G1), while the second cluster comprised specimen from Trinidad and Tobago (BOLD BIN: AAJ2798) (Ke. bellatrix s.s.). The mean K2P distance between these groups was greater than 4% (Table S6). The BLAST (Basic Local Alignment Search Tool) alignment between the COI sequences of Ke. bellatrix from Vale do Ribeira and from the type locality showed a similarity of 95.59% (e-value 0.0). For Ke. cruzii, the cluster containing specimens from Vale do Ribeira (BOLD BIN: AAG3843) (Ke. cruzii s.s.) differed from the Ke. cruzii sample from Parque Estadual da Cantareira, São Paulo municipality (BOLD BIN: ADV0367) (Ke. cruzii G1), with a mean K2P distance greater than 4.3% (Table S7).

Specimens of Ps. ferox from Vale do Ribeira grouped into two clusters in the COI RESL analysis: one cluster that included sequences from French Guiana and Colombia (BOLD BINs AAO0580) (Ps. ferox s.s.) and a second cluster formed by only one specimen from Vale do Ribeira (BOLD BIN AFE1443) (Ps. ferox G1). The third cluster of this specie includes only one specimen from Mexico (BOLD BIN: AFE7024) (Ps. ferox G2) (Table S3). The K2P mean distance between these groups was 2% (Table S8).

Analyses with the D2 expansion of the 28S rRNA gene were performed using sequences from 12 species (n = 108 specimens). Using ASAP, 11 clusters were recovered from the D2 dataset (Table S9). Most species formed distinct cluster groups, including Hg. leucocelaenus, Sa. conditus, Wy. confusa, Ae. terrens, Ae. scapularis, Hg. janthinomys, Ke. bellatrix, Ae. perventor, Ae. serratus/nubilus, and Ke. cruzii. Psorophora forceps and Ps. ferox grouped together. The pairwise genetic distances between D2 sequences of different species are in Table S10. Psorophora ferox and Ps. forceps shared 100% similarity in the D2 fragment sequenced. The distance between the remaining species ranged from 2.67% to 40.47%, with the smallest distance observed between Ae. serratus/nubilus and Ae. perventor, whereas the largest distance was between Sa. conditus and Ke. cruzii.

4. Discussion

Precise identification of vector mosquito species is essential for the effective control of vector-borne diseases like malaria and arboviruses. DNA tools are particularly useful for distinguishing species when morphological similarities among females complicate precise identification [40]. This study utilized two DNA sequence markers for species identification: D2 and COI [11,12,19,41,42].

Species delimitation methodologies are employed in studies with anophelines [12,42]. Among the existing methods, two were used in this study: ASAP [39] and RESL (https://www.boldsystems.org/, accessed on 29 May 2024). Both techniques are appropriate for species partitioning from single locus sequence alignments. ASAP is an exploratory method, in which hierarchical clustering is based on pairwise genetic distances, and no prior biological knowledge about intraspecific diversity is required. RESL is a network-based method, in which hierarchical clustering is based on connectivity statistics. In this technique, Barcode Index Number (BIN) is generated from Molecular Operational Taxonomic Units (MOTUs). Considering that each technique has its limitations, both methodologies were used to delimit the species collected in the Vale do Ribeira, as well as to verify the diversity of a given species in different localities.

In the Atlantic tropical rainforest of Brazil, human and zoonotic malaria transmission is primarily associated with species of the genus Kerteszia in the subfamily Anophelinae [43]. Kerteszia cruzii is recognized as the primary vector, with Ke. bellatrix serving as a secondary vector in the Southeastern Atlantic Forest, Brazil [44]. Currently, the genus Kerteszia includes 12 valid species, though this number could rise to 28 species, with 16 species unknown to science [42]. Kerteszia cruzii has been identified as a species complex based on the banding pattern of the polytene chromosome [45,46], PCR-RAPD and PCR-RFLP of the ITS2 and parts of the 5.8S and 28S genes [47], DNA sequences of nuclear genes [48,49], and the mitochondrial genome [50]. The results of current analyses using the COI barcode and the D2 region of the 28S rRNA gene further support the hypothesis of a species complex. In RESL, the COI barcode sequences revealed two clusters within Ke. cruzii. The first cluster (Ke. cruzii) comprises sequences from specimens collected in Vale do Ribeira and additional sequences from GenBank, which include specimens from Cananeia (São Paulo State), Maquiné (Rio Grande do Sul State), and Itatiaia (Rio de Janeiro State). The second cluster (Ke. cruzii G1) consists of a GenBank sequence from Parque Estadual da Cantareira, São Paulo State, Brazil.

DNA sequences of Ke. bellatrix (formerly An. (Ker.) bellator) grouped separately into two clusters. The COI dataset included one sequence from a specimen from the type locality, Trinidad and Tobago (GenBank: OQ272309). Results of the RESL analysis showed that this COI sequence formed one cluster, while sequences from the Vale do Ribeira and São Paulo municipality grouped together into a distinct cluster, Ke. bellatrix G1. The clusters generated by the Kerteszia sequences are the same as those previously identified by Bourke et al. [41], and the inclusion of Vale do Ribeira sequences did not alter them. It is important to highlight that only one specimen of Ke. bellatrix and two of Ae. terrens from Vale do Ribeira were used in the analyses, restricting the ability to perform reliable inferences regarding the genetic diversity of these species. New studies with a greater number of samples may contribute to a more robust result.

Twenty-eight species of the genus Haemagogus have been identified in Brazil [51]. Haemagogus leucocelaenus, Hg. capricornii, and Hg. janthinomys are vectors in the sylvatic yellow fever cycle, making these species a public health concern [52]. Hg. capricornii is known to inhabit the Atlantic rainforest region and semi-deciduous forests on plateaus [31], while Hg. janthinomys has a wider geographic distribution. In certain areas, the species are sympatric. In Brazil, both species are found both in lowland regions and in areas with altitudes above 800 m of the sea level [53]. Females and larvae of Hg. capricornii and Hg. janthinomys are morphologically similar and difficult to identify when specimens are not well preserved, and no male is available to verify identification using characteristics of the male genitalia. Molecular techniques and wing morphometry have been employed to distinguish these two species [54,55]. Recently, Telles-de-Deus et al. [55] obtained COI barcode sequences from morphologically identified male specimens of Hg. capricornii and verified that this gene fragment is effective in the molecular identification of the Haemagogus species. The present study used as reference four of the sequences of Hg. capricornii and Hg. janthinomys obtained in Telles-de-Deus et al. [55]. In the results of the ML analysis employing COI sequence data, the bootstrap support for the split leading to these both species was 95%. Already, the same analysis supported Hg. leucocelaenus as a single lineage with 98% bootstrap support.

Studies indicate that Hg. janthinomys may be a species complex because, among others, (1) it has better adaptation to different biomes compared to Hg. capricornii [53], (2) it presents significant intraspecific differences in wing centroid size [54], and (3) it presents polymorphisms in analyses with COI barcoding [55]. In this study, the generation of three clusters of Hg. janthinomys in COI RESL and the presence of some subclades of this species in ML analysis, corroborates with another study [55], and supports the hypothesis that this species may be a species complex.

Psorophora ferox is associated with wet woodland habitats [56] and holds epidemiological significance [20,57]. This species has been found naturally infected with the yellow fever virus in Ribeirão Preto, São Paulo State, Brazil [57], and with the Rocio virus in the Vale do Ribeira region [20]. Morphological variations in the number and size of the egg tubercles were found among three populations of Ps. ferox [58]. The analysis of the COI sequence data identified three distinct lineages: (1) the first lineage comprises only one sequence from a specimen collected in Vale do Ribeira, (2) the second lineage includes sequences from mosquitoes collected in Vale do Ribeira, Colombia, and French Guiana, and (3) the third lineage consists of a single sequence from a specimen collected in Mexico. Considering the specimens collected in Vale do Ribeira that split into two lineages, it is plausible to hypothesize that the lineage encompassing only Ps. ferox from Vale do Ribeira may represent Ps. pseudomelanota, which were potentially misidentified. Conversely, the second lineage may accurately represent Ps. ferox. The females and males of these species share morphological similarities but can be distinguished based on the characteristics of fourth-instar larvae and pupae [59]. Female differentiation is possible by examining the distribution of dark and golden scales on the scutum. In Ps. ferox, the scutum is predominantly covered with golden scales, with only a small number of intermixed dark scales. In contrast, Ps. pseudomelanota exhibits a predominance of dark scales on the scutum, with golden scales being sparse and fewer in number. Despite these species being morphologically distinct, it is essential to preserve all scutal scales intact in adult specimens for accurate species identification. Already, for the specimen present in the third lineage, it is reasonable to hypothesize that they may correspond to Psorophora posticatus, originally described by Wiedemann (1821), with Mexico as their type locality. This species was later considered a synonym of Ps. ferox by Belkin (1968). Further research is needed to confirm whether Ps. posticatus is a valid species.

Species of the genus Aedes are distributed worldwide. Aedes species are aggressive blood feeders, and several are vectors for various arboviruses across sylvatic, rural, and urban environments. This genus includes species responsible for transmitting Chikungunya, Zika, dengue, and yellow fever viruses [4]. Ae. scapularis has been identified as a potential vector for yellow fever and Venezuelan equine encephalitis viruses [60,61]. Additionally, Ae. serratus has been found naturally infected and acts as a secondary vector for yellow fever virus in Northwestern Rio Grande do Sul State, Brazil [52]. Aedes serratus is morphologically similar to Ae. nubilus, Ae. oligopistus, and Ae. hastatus. However, these species can be distinguished by characteristics of the fourth-instar larva and male genitalia, making identification based on female traits more challenging. Sequences from Ae. serratus/nubilus in Vale do Ribeira and Ae. serratus in Amazonas and Amapá States clustered together in both the RESL and ASAP analyses. Despite this clustering, the Kimura 2-parameter (K2P) distance between these sequences exceeded 2%. The females from Vale do Ribeira, identified morphologically as Ae. serratus/nubilus, are likely Ae. serratus, despite the K2P distance exceeding 2% when compared to the remaining sequences from specimens collected in Amapá and Amazonas states, Brazil, and Colombia. The COI sequence divergence among conspecific individuals of Aedes were found to be higher than 3% in Canadian mosquitoes [62]. Thus, the COI sequence divergence observed in Ae. serratus across geographically distant localities in Brazil was anticipated.

The D2 region of 28S rDNA has been used for identifying species worldwide [63,64]. The D2 fragment used in the present study enabled the recognition of some species, but not all of the species examined, such as Ps. ferox and Ps. forceps, which were grouped together in a single lineage. A possible explanation for these species grouping together could be associated with the small fragment employed in the study, resulting in a deficiency of phylogenetic evidence to differentiate species of the subgenus Janthinosoma from the genus Psorophora. In addition, the genus taxon sampling was limited to two species that clustered together, as expected. It is also plausible that the addition of other species of Psorophora could increase the likelihood of recovering the species separately.

Although the D2 region presented variation in length in the species studied, the analyses performed were not affected by the presence of shorter sequences, such as those from the species Ae. serratus/nubilus (370 bp), Hg. leucocelaenus (372 bp), Ae. scapularis (373 bp), and Ae. terrens (373 bp). The D2 fragment used was sufficient to delimit the species collected from Vale do Ribeira, except to the Psorophora species, as discussed above.

5. Conclusions

This study has contributed to the understanding of the diversity of Culicidae in the Vale do Ribeira region and has examined the effectiveness of two commonly used fragments for the molecular identification of mosquitoes. The D2 sequences were able to distinguish the samples by species except for Ps. forceps and Ps. ferox. The presence of subclades of Hg. janthinomys and the division into three groups suggest that this species is a species complex. The findings reinforce the hypothesis of previous studies and evidence that Ke. cruzii is a species complex. The COI sequences of Ke. bellatrix from Vale do Ribeira share 95.59% similarity with those of specimens from the type locality, and further investigation is needed to determine whether it could be a potential species.