Distribution of Anopheles species in malaria endemic areas of Honduras in an elimination setting

Background : Anopheles mosquitoes are the vectors of malaria, one of the most important infectious diseases in the tropics. More than 500 Anopheles species have been described worldwide, and more than 30 are considered a public health problem. In Honduras, information on the distribution of Anopheles spp. and its genetic diversity is scarce. This study aimed to update information on the diversity of Anopheles mosquitoes in Honduras with a morphological and molecular approach. Methods : Mosquitoes were captured in 8 locations in 5 malaria endemic departments during 2019. Two collections methods were used. Adult anophelines were captured outdoors using CDC light traps and by aspiration of mosquitoes at rest. The morphological identification was performed using taxonomic keys. Genetic analyses included the sequencing of a partial region of the cytochrome oxidase I gene (COI) and the ribosomal internal transcribed spacer 2 (ITS2). Results : A total of 1320 anophelines were collected and identified through morphological keys. Seven Anopheles species were identified. An. albimanus was the most widespread and abundant species (74.02%). To confirm the morphological identification of the specimens, 175 and 122 sequences were obtained for COI and ITS2 respectively. Both markers confirmed the morphological identification. COI showed a greater nucleotide diversity than ITS2 in all species. High genetic diversity was observed within the populations of An. albimanus while An. darlingi proved to be a highly homogeneous population. Phylogenetic analyses revealed clustering patterns in An. darlingi and An. neivai in relation to specimens from South America. New sequences for An. crucians , An. vestitipennis , and An. neivai are reported in this study. Conclusions : Here we report the distribution of Anopheles species in endemic areas of malaria in Honduras. According to our results, both taxonomic and molecular approaches are useful tools in the identification of anopheline mosquitoes. However, both molecular markers differ in their ability to detect intraspecific genetic diversity. These results provide supporting data for a better understanding of the distribution of malaria vectors in Honduras.

4 Neotropics [13], or have made notable efforts to predict the distribution of the DVS of malaria in the Americas through intensive literature searches and an evidence-based approach [9,14]. Despite these efforts, there are still important information gaps about Anopheles species in Honduras, and the only verifiable data on their distribution in the country are internal reports by the Ministry of Health, which publishes them as part of routine entomological surveillance since 2013. According to those reports, 12 species of anophelines have been identified through morphometric keys: Anopheles albimanus, An. albitarsis, An. apimacula, An. argyritarsis, An. crucians, An. darlingi, An. gabaldoni, An.
grabhami, An. neomaculipalpus, An. pseudopunctipennis, and An. punctimacula. Another information gap in Honduras is the lack of molecular markers data that support the classification of mosquitoes based on morphometric keys. Molecular markers are critical to distinguish between evolutionarily close or cryptic species, even using immature specimens [15,16].
To optimize the limited resources available for vector control strategies in Honduras, it is necessary to know in depth the distribution and relevant bionomic aspects of DVS and other anophelines considered secondary vectors of malaria. This study aims to provide an update on the diversity of the Anopheles mosquitoes in Honduras, supporting its distribution in morphological data, as well as in two molecular markers.

Study sites
Entomological captures were carried out in 8 sites in 5 departments of the country (Atlántida, Colón, Comayagua, El Paraíso, and Gracias a Dios) from February to October 2019 ( Table 1). The departments of Atlántida, Colón and Gracias a Dios are classified as very humid tropical ecosystems, while Comayagua and El Paraíso are considered as subtropical dry. The average temperature varies between 25ºC and 33ºC, and the relative humidity ranges from 40-91% in all sites depending on the season of the year. The population's livelihood in the selected areas is mainly based on agricultural and livestock activities. The study sites are those monitored by the Ministry of Health of Honduras to undertake routine entomological surveillance as they remain endemic to malaria by Plasmodium vivax. Malaria due to P. falciparum malaria is reported almost exclusively in Gracias a Dios.
Geographical coordinates and altitude of the collection sites are shown in Table 1.

Nucleotide Diversity (π) And Number Of Haplotypes
In order to calculate the nucleotide diversity (π), the sequences of both molecular markers were analysed separately and by species. The sequences were aligned using the MUSCLE algorithm. MEGA v10.0 software with 1000 Bootstrap replicas was used to calculate the pairwise distance using the Maximum Composite Likelihood substitution method, and 95% as the site coverage cut-off. The percentage of identical bases for each species and between species (intra-and inter-specific similarities) was calculated in order to demonstrate the reported "barcoding gap".
The haplotype diversity was calculated with R through the function hap.div of pegas (v0.12 package) and using the Nei and Tajima´s method [24]. Haplotype frequencies were calculated using the Haplotype function with default parameters, and the haplotype network was computed with the haploNet function using an infinite site model, pairwise deletion missing data, and probability of parsimonious link [25].
Phylogenetic Analysis 8 Nucleotide sequences were trimmed and manually corrected using the Geneious ® 9.1.2 software. The ClustalW tool was used to align sequences. Phylogenetic trees were constructed using the Tamura-Nei distance model, the Neighbor-Joining method and a bootstrap of 1,000 replicas with no outgroup.
Length, identical sites and pairwise % identity were calculated for each molecular marker and each species.
To calculate the phylogenetic relationships between specimens collected in Honduras with those collected in other countries of the Americas, analogous COI and ITS2 sequences for all available Anopheles species were downloaded from the GenBank database. Sequences were aligned and phylogenetic trees constructed under the same parameters described above.

Distribution of Anopheles species
Eight locations with active foci of malaria [8] were visited to collect anopheline mosquitoes. A total of 1320 adult individuals of seven Anopheles species were collected and identified by a taxonomic key   databases. This is also the first report of ITS2 sequences for An. vestitipennis. In summary, the morphological identification coincided with the molecular identification of both markers for the species with sequences previously reported in the databases.

Nucleotide diversity and haplotypes
Intraspecific variation was calculated for both markers. COI showed a higher level of polymorphism than ITS2. According to COI, the species with the highest nucleotide diversity was An. crucians (π = 0.05), followed by An. vestitipennis (π = 0.03) ( Table 3). The species with the lower diversity was An.
darlingi. An. albimanus revealed a high number of haplotypes (n = 55). An. pseudopunctipennis showed the highest proportion of COI haplotypes with respect to the number of sequences analysed (11/11) and An. darlingi revealed the lowest haplotype index (3/16). ITS2 showed a low number of haplotypes (1)(2)(3)(4) in all species (Table 3) (Fig. 4). Three analyses were performed to infer phylogenetic relationships between sequences. The first analysis included all the sequences of each marker for six Anopheles species. Both dendrograms (COI and ITS2) showed that the species clearly separated into clades (Fig. 5).
The second analysis included sequences of An. albimanus classified according to geographic region.
Phylogenetic relationships based on COI sequences showed only one separate cluster that included 11 out of 14 sequences of mosquitoes collected in Gracias a Dios. The other sequences were not clustered (Fig.6). ITS2 sequences did not reveal any clustering according to geographical origin. This analysis was not performed for other Anopheles species due to the low intraspecific variation.
The third phylogenetic analysis included the COI sequences of five species obtained in this study (An.
albimanus, An. darlingi, An. pseudopunctipennis, An. punctimacula, and An. neivai) together with analogous sequences available in GenBank in order to understand the relationships between individuals from Honduras with mosquitoes from other countries in the Neotropical region. The same analysis was performed separately with the ITS2 sequences of five species from Honduras (An. albimanus, An. darlingi, An. pseudopunctipennis, An. punctimacula, and An. neivai) and sequences from specimens of other countries. The phylogenetic tree of An. albimanus included 12 COI sequences of mosquitoes from Colombia and 103 sequences of mosquitoes from Honduras, however the sequences of Colombia clustered together with the majority of sequences from Honduras. Eleven sequences of mosquitoes captured in Gracias a Dios formed a well-supported clade (Fig.7a). For An.
darlingi 16 sequences from Honduras, 6 sequences from Colombia, 5 sequences from Brazil, and 4 sequences from Peru were analysed. According to this analysis the population was divided into two clusters, one including all the sequences of Honduras, and another with the sequences of South America (Fig.7b). In addition, 12 sequences of An. pseudopunctipennis from Honduras and nine sequences from Colombia were analysed. For the analysis of An. punctimacula, seven sequences from Brazil, 14 sequences from Colombia and one sequence from Honduras were included. No clusters were detected for both species (Fig.7c, 7d). The analysis for An. neivai included three sequences from French Guiana, six sequences from Colombia, and five sequences from Honduras. The specimens of the three countries showed a defined separation according to geographical origin (Fig.7e).
The phylogenetic analysis of the ITS2 sequences included a total of eight countries of the Americas, including Honduras, Colombia, Brazil, French Guiana, Panama, Nicaragua, Ecuador and Belize. None of the trees could demonstrate separation of populations based on geographic origin (Fig.8).

Discussion
This study provides updated information on the distribution of Anopheles species in endemic malaria regions of Honduras. Seven Anopheles species were found. An. albimanus was the most common species and the most widely distributed. This is consistent with the existing literature. An. albimanus has been described as the dominant species in Central America, the Caribbean and some coastal regions of northern South America [9,12,13]. This has been demonstrated through studies conducted in Colombia [26], Panama [27], Belize [28], and Guatemala [29]. The predominance of this species -considered as a generalist species -can be attributed to the wide range of habitats, feeding preferences, and heights in which the larvae can develop [30,31]. In this study, mosquitoes were collected at eight geographical sites. In seven sites, An. albimanus was the most frequently captured 13 species despite the ecological differences between all locations. Five of the eight sites are classified as wet coastal ecoregions (Atlántida, Colón, and Gracias a Dios), with less than 550 m.a.l.s., while two sites (Comayagua and El Paraíso) are classified as Piedmont, with heights above 550 m.a.l.s. and drier ecosystems [13]. La Ceiba (Atlántida) and Kaukira (Gracias a Dios) yielded greater diversity of species (4 to 5), similar to reports from Cordoba, in the coastal region of the Colombian Caribbean [26], and in a study carried out in Papua New Guinea describing the abundance and composition of anophelines [32]. Such biological diversity of anopheline species could be due to the fact that both sites have tropical ecosystems and are regions rich in lagoons, coral reefs, mangroves, beaches, rivers and abundant flora and fauna. In addition, Kaukira is located in the department of Gracias a Dios, also known as La Mosquitia, the region with less anthropogenic intervention in the country.
The second most abundant species collected in Kaukira was An. crucians. This finding is remarkable since this species was not registered anywhere else in this study. Its abundance is probably related to the high density of phanerogam plants that have been previously associated with the development of their larvae [28]. An. crucians has been recognized as one of the five most important malaria vectors in the country [33], and has been reported as one of the most frequent species in Belize, Guatemala, Honduras and Nicaragua [28,34]. Since La Mosquitia is the main region with permanent transmission and the highest number of malaria cases in the country throughout the year, it would be interesting to further explore the importance of this species in the malaria transmission. On the other hand, An.
darlingi was collected only in two coastal departments (Atlántida and Colón), consistent with previous reports [35]. This species is known for its preference to inhabit areas of high rainfall and where the tropical forest is close to the ocean [13,36].
In addition to the morphological identification of specimens, sequences of the COI gene and the ITS2 ribosomal region were obtained for all species. Four and six species of anophelines were properly identified by BLAST of the COI and ITS2 sequences, respectively. Up to the moment of the analysis, there were no analogous sequences available of COI for An. crucians, An. vestitipennis and An. neivai, nor any sequences of ITS2 for An. vestitipennis in the GenBank database. Consequently, these would be the first sequences reported. This findings support the barcode strategy as a useful tool to confirm the correct assignment of misidentified or unidentified species using morphology [26,37,38,39].
When comparing the individual ability of both markers to identify or confirm Anopheles species in Honduras, it seems that both are informative enough and fulfil their purpose [39,40,41]. Some authors report problems to solve and identify species when those markers are used individually [42], and they suggest that a multi-locus approach might have a greater power of discrimination [43,44].
However, our study shows that both molecular markers are useful separately and are a good complement to the identification of Anopheles based on taxonomic keys [45].
Intraspecific variation was calculated for five Anopheles species. A greater nucleotide diversity (π) and number of haplotypes with COI than with ITS2 were observed. According to this result, COI would be more informative to decipher the intraspecific phylogenetic relationships. Some authors reported different findings when analysing the phylogeny of the Anopheles Hyrcanus Group using ITS2 sequences downloaded from GenBank. They concluded that ITS2 would be more reliable than COI as a phylogenetic analysis tool among very close taxa. This discrepancy could be attributed to the fact that the Hyrcanus Group includes at least 25 species widely distributed in a large geographic area [15,16]. Discrepancies between markers are expected since there are different evolutionary processes that act differently on mitochondrial and nuclear genes [46]. Nevertheless, COI could be considered a more useful marker for evidencing intraspecific genetic diversity between Anopheles spp. in Honduras.
The species with the lowest genetic diversity was An. darlingi when 16 COI sequences were analysed.
Although the number of sequences studied is low, it is possible to say that the population is relatively homogeneous. High homogeneity within the population could be attributed to the fact that the geographical area in which the mosquitoes were collected was small. Similar results were reported in a study conducted in Darien, at the border between Panama and Colombia, with 40 individuals who showed low nucleotide diversity (π = 0.0006) [47].
On the other hand, when the phylogenetic relationship of An. darlingi specimens collected in the Central and South America reveal significant differences through the use of morphological and behavioural markers [48], RAPDs [49], COI [50], and microsatellite loci [51]. It has been hypothesized that this geographic isolation could be attributed to the absence or low population densities of An.
An. neivai was the second species that showed well separated clades within the dendrogram. One isolation should be confirmed in the future by more robust and informative molecular markers such as microsatellite loci [53]. Future sampling should also include specimens from other geographical 16 regions, particularly from the Honduran islands in the Caribbean.

Conclusions
In this study, the distribution of Anopheles species in malaria endemic areas of Honduras has been described through a morphological approach and two molecular markers. Conventional taxonomy, COI, and ITS2 proved to be useful tools for the correct identification of anopheline species. However, both molecular markers differ in their ability to detect intraspecific genetic diversity. According to phylogenetic analyses, the only two species that seem to show some level of structuring with respect

Consent for publication
Not applicable.

Availability of data and materials
The datasets used and/or analysed during the current study are available from the corresponding author on reasonable request.

Competing interests
The authors declare that they have no competing interests. GAbstract.pdf