Ethics Statement

This work was carried out under scientific license provided by the OPDS (Provincial Agency for Sustainable Development) for sampling in protected areas, in some private places where specific permits from the owners were needed, and in some free-access lands where permits were not required. None of the endangered or protected species were included in this study.

Sampling area

Mosquitoes were collected in General Pueyrredon district (Figure 1), located on the coast of the Atlantic Ocean, in the South-East of Buenos Aires province (38° 00´ S, 57°33´ W). Its total surface accounts for 1,453.44 km² of which 79.48 km² are taken up by the Mar del Plata urban area. This geographic region is characterized by a great diversity of biotopes, most of which have favorable conditions for the development of mosquitoes. The area is crossed from west to east by numerous streams which rise in the highlands of the province. There are numerous wetlands like “Reserva Integral Laguna de los Padres” (a provincial reserve) and “Laguna Mar Chiquita” (a Ramsar site) which are permanently visited. The climate is mild with maritime influence and the average temperature is 8°C (46°F) in winter and 20°C (68°F) in summer. Rainfalls are abundant throughout the year, and the weather, according to Köppen climate classification, is similar to Western Europe [9].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Sampling area and freshwater habitats from General Pueyrredon municipality.

1, Urban area with natural and artificial containers; 2, roads towards Mar del Plata with puddles alongside; 3 roads in coastal area with puddles alongside; 4, area of nature reserve with wetlands, freshwater swamps and bogs.

https://doi.org/10.1371/journal.pone.0075516.g001

Insect Collections and Identification

Mosquitoes were collected between September 2009 and April 2011. Adult collections were made with a CDC Mini Light Trap with incandescent light (model 2836BQ, BioQuip Products, Rancho Dominguez, California, USA). Hand-held aspirators were used to collect host-seeking females of species that could not be caught in light traps. Larval specimens were collected from vernal pools, freshwater swamps, bogs, and natural and artificial containers (Figure 2), and reared until fourth instar or adulthood. Voucher specimens, prepared from all localities, were submitted to the local museum, Museo de Ciencias Naturales “Lorenzo Scaglia” (Mar del Plata, Argentina). Mosquito fourth instar larvae and adult females were recognized using identification keys by Darsie and Mitchell [10] and Rossi et al. [11]. Specimens were stored at -20°C until DNA extraction.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Surveyed mosquito habitats.

A, Eryngium horridum; B, detail of leaf axils of E. horridum; C, freshwater pool in nature reserve; D, puddles along roads; E-G, artificial containers. Dr. Maciá has given written informed consent, as outlined in the PLOS consent form, to publication of this photograph.

https://doi.org/10.1371/journal.pone.0075516.g002

Nucleic acid isolation, primer design and PCR conditions

Genomic DNA was isolated from a single mosquito at the fourth larval instar or at the adult stage. After all individuals were sorted by species using taxonomical keys, sampling place and breeding habitat type, one individual of each of these groups was preserved as morphological voucher and another one as a source of genomic DNA. Total DNA was extracted with the PureLink Genomic DNA Mini Kit (Invitrogen, Grand Island, New York, USA) according to the manufacturer’s instructions.

For the amplification of a fragment of the COI the standard primers were useful, and for nuclear 18S rDNA some samples were amplified at a first step by EukA/EukB primers [12], but a new oligonucleotides design from some Culicidae sequences was needed for most of the samples. For that, we designed a set of primers using ClustalW2 Server [13] to compare 18S rDNA sequences of the different Culicidae genera retrieved from Genbank/EMBL databases. Primer sequences and PCR annealing temperature of each are shown in Table 1.

The PCR reactions were carried out by using standard protocols. The amplified products were analyzed by electrophoresis in 1% (w/v) agarose gels in tris-acetate buffer and ethidium bromide staining [14], and the fragments of about 1,200; 1,800 and 600 bp of 18S rDNA and COI sequences of about 600 bp were purified using Purelink PCR purification kit (Invitrogen) according to the manufacturer’s protocol.

Sequencing and database analysis of the PCR fragments

The purified PCR product was submitted for direct nucleotide sequencing (Macrogen, Korea). DNA sequence data sets were analyzed by BLASTn and multiple-sequence alignment (Table 2) [15]. Additional 18S rDNA and COI coding-sequences for mosquito species and outgroup taxa were obtained from the National Center for Biotechnology Information (NCBI), http://www.ncbi.nlm.nih.gov (Tables S1 and S2). Retrieved sequences were aligned using ClustalX [16] and phylogenetic analyses were performed. Firstly, DNA alignments were used to estimate the best evolutionary model using the program ModelTest [17]. These models were selected using AIC score [18] and used to obtain Maximum Likelihood trees through the PhyML program [19]. Also, Maximum Parsimony and Distance trees (using Neighbor-Joining method) were obtained using the programs MEGA [20] and HYPHY [21] respectively. Maximum Likelihood and distance trees were obtained using Tamura Nei model [22] with observed frequencies and gamma distribution for rate heterogeneity estimation. Tree support was obtained using a non-parametric bootstrapping (using 500 replicons for Maximum Likelihood estimations and 5,000 for Parsimony and Distance methods) followed by a majority rule consensus to obtain the final trees. Trees obtained from the different methods were compared using TOPD program [23]. Finally, trees were rooted using mid-point protocol with the corresponding outgroups and displayed by the Archaeopteryx program [24].

The novel sequences reported in this paper have been deposited in the EMBL database (Table 2).

18S rDNA Analysis

Maximum Likelihood, Parsimony and Neighbor-Joining trees based on the information given by mosquito 18S rDNA sequences resulted in very similar trees, with slight differences in the position of species within each genus (Figure 3A and S1). This global topological similarity was also confirmed using the algorithm TOPD showing the robustness of the evolutionary estimation made with different methods. Trees sustained subfamilies (Anophelinae and Culicinae) traditionally accepted for the family Culicidae. Anopheles species produced a basal group to the rest of the family members. Toxorhynchites stemmed in a basal clade with Culicinae, upholding Toxorhynchitini as a separate tribe but not as a subfamily, as proposed by Harbach and Kitching [25]. As expected, Dixella cornuta resulted in an external position with respect to Culicidae.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Phylogenetic estimation by Maximum Likelihood.

A, Using 18S rDNA; B, COI coding-sequence. 500 replicates have been used in the estimation. Arrows indicates local mosquito species.

https://doi.org/10.1371/journal.pone.0075516.g003

Monophyly was supported for the genus Culex with well bootstrapping value and in all the three phylogenies obtained for 18S rDNA (Figure 3A and S1). Members of the other tribes were not classified according to this hierarchical level. Our trees do not support the status of Ochlerotatus as separate genera from Aedes, as stated by Reinert [26], as seen by the internal arrangement of species from both taxa. Psorophora species were grouped together with Culiseta and Coquillettidia, and in a different clade from Ochlerotatus and Aedes. Yet Aedeomyia and Uranotaenia were more closely related to the Ochlerotatus+Aedes group than to the Psorophora, Culiseta and Coquillettidia group (Figure 3A).

COI Analysis

Species from Argentina were placed within the correspondent genus along foreign mosquito species. Although these trees were broadly similar, some differences can be ascertained (Figure 3B and S2), being the tree drawn on the Maximum Likelihood method more fitted to classic taxonomy at genera level (Figure 3B). In this particular tree, genera members from the tribe Aedini were interspersed in several branches. Culex species were clustered and show some inconsistence of some particular sequences analyzed by Parsimony and Distance (Figure S2). All Anopheles species were grouped in a single set, thus reflecting a monophyletic origin in all the COI trees obtained. Orthopodomyia and Toxorhynchites were a sister group of Uranotaenia. Species of Psorophora were clustered together.

Sequences analysis permitted the interpretation of some relationships between native and foreign mosquitoes, which can be summarized as follows:

Ochlerotatus albifasciatus and O. crinifer: These species are floodwater mosquitoes with a diapause in the egg stage. O. albifasciatus is a pest mosquito because females are aggressive biters and can act as vectors of Western equine encephalitis virus [27]; and it develops mainly in shallow puddles of grasslands and adults emerge in large numbers about 7 to 10 days in spring and 11 to 22 days in autumn after flooding of breeding places [28]. O. crinifer shares pools with the former species, but its larval density is usually lower and is more frequent in wooden areas [29], team personal observations]. The congeneric species more closely grouped with both mosquitoes, as seen in the 18S rDNA-based cladogram, were O. sollicitans, O. triseriatus and O. canadensis. COI-based topologies resulted in a close relationship between O. albifasciatus and O. ventrovittis, and O. crinifer and O. atlanticus. According to Crans’ life cycles classification [30], O. atlanticus is the species whose bionomics mostly resembles those of O. albifasciatus and O. crinifer, because they are multivoltine, floodwater mosquitoes. At least some populations of O. albifasciatus are resistant to salt-water [31], as O. sollicitans.

Psorophora cyanescens: This is also a floodwater species which larvae occurs in pools filled with spring and summer rainfall, associated with O. albifasciatus [32]. Females persistently seek hosts for blood-feeding between November and March in other areas of Buenos Aires province and their presence can coincide with P. ferox and P. ciliata [33]. In our 18S rDNA analysis, P. cyanescens was related to P. ferox and P. ciliata, therefore, Maximum Likelihood tree reflects the bionomics of these species. In the COI trees, P. cyanescens was located close to P. longipalpus, independently of the estimating method, while the other species of the genus (P. columbiae, P. signipennis and P. insularia) were placed in a separate cluster. This assemblage within Psorophora agrees with the classification at the subgeneric level, since P. columbiae, P. insularia and P. signipennis belong to Psorophora (Grabhamia), while P. cyanescens and P. longipalpus belong to Psorophora (Janthinosoma).

Culex spp.: Poor intraspecific resolution was observed regarding this mosquito genus as expected using these markers. Neither the COI- nor the 18S rDNA-based topologies lead to unambiguous positions of each species of Culex (Figure 3), so we cannot establish any relationship among them with certainty. Low bootstrapping values indicate that there is not enough information either in COI or 18S rDNA to resolve Culex species. Harbach [34] pointed out that Culex contains a high degree of polymorphism and exceptional forms. All species in our samples were classified within the subgenus Culex (Culex), which is diversified to the highest degree at the Neotropical region [35]. Most of the taxonomic research on Culex was only directed to differentiate species instead of a thorough classification that reflects natural relationships [36]. Demari-Silva et al. [8] established the relationships between Culex species from Brazil and suggested that C. dolosus may comprise a species complex. Intraspecific variation may add further confusion during identification. In Argentina, the distribution and hybridization of the Culex pipiens subgroup (sensu Harbach [36]) members such as C. pipiens and C. quinquefasciatus were studied in previous works. Humeres et al. [37] provided genetic evidence about the subspecific status of C. pipiens and C. quinquefasciatus, and contributed with information about genetic distance and gene flow between these taxa as well. However, other authors considered both taxa as separate species. C. quinquefasciatus is distributed from the provinces of Buenos Aires and Mendoza northwards, while C. pipiens was found from Buenos Aires southwards to Santa Cruz province [38,39]. Also, hybrid forms have been found in the central area of the country between 30° and 33° S [40,41]. Finally, Micieli et al. [42] were able to identify C. pipiens form molestus, C. quinquefasciatus and hybrids between them by morphological criteria and high-resolution molecular markers. The status of the members of the C. pipiens complex is still controversial, since some researchers consider C. pipiens and C. quinquefasciatus different species [36,42], while others ascribed them as subspecies of C. pipiens [43-46]. We consider all individuals in our samples as C. pipiens, because these were obtained within its geographical range [47], and did not attempt to identify subspecies or forms. Culex members probably could have been better discriminated using different gene polymorphisms, like the second intron of the acetylcholinesterase-2 (ace-2) locus [48], but this analysis was not the aim of this work. All the preceding analyses show that Culex species deserve further research using multidisciplinary procedures.

Uranotaenia lowii: Larvae of U. lowii are found in permanent, grassy shallow ponds, usually exposed to sunlight, from Southern USA to Argentina [49]. As females feed on amphibians, they are unimportant from a medical standpoint. This fact may explain in part the lack information on the phylogeny of the genera in the Neotropical region other than Lane [50], Galindo et al. [51] and Belkin et al. [52]. An interesting result of the 18S rDNA trees is the association between Aedeomyia and Uranotaenia, which suggests the relationship between them previously reported by Harbach and Kitching [25] (Figure S1).

In Mar del Plata we detected the presence of C. pipiens, C. apicinus, C maxi, O. albifasciatus, O. scapularis, O. serratus, O. crinifer, P. ciliata and U. lowii, which were reported as potential vectors of pathogens (http://www.cdc.gov/ncidod/dvbid/westnile/mosquitospecies) [53-57]. Several viruses transmitted by these mosquitoes were reported in Argentina, as well as the vectorial capacity of some of these species [58-64], justifying the importance of knowing the species present in this area of study and the increase health risk if we take into account the following issues. First, Mar del Plata increases its population during summer seasons, with a large movement of people around the country including areas of high prevalence of Aedes (=Stegomyia) aegypti (L). Second, there is a relative proximity of this city with Buenos Aires and La Plata (located at 400 and 370 Km away, respectively, to the North), both with a significant abundance of A. aegypti [65,66]. Third, the latter showed a progressive advance southwards [67]. Fourth, the gradual increase of the average temperature in the region has led to the introduction and development of insects such as lepidopterans common in warmer areas [68]. Thus, we might think that Mar del Plata could be a potential source for the introduction and spread of agents of medical interest.

New diseases are constantly being discovered and are becoming more widely distributed with the increase in traveling, to and from tropical and disease-endemic countries. According to the Pan American Health Organization [69], the prevention or reduction of transmission of most diseases being vectorized by mosquitoes depend entirely on the control of mosquito vectors or the interruption of human-vector contact, for this reason, the correct identification of the potential vectors involved and their geographic distribution is a critical step, as well as the active monitoring and surveillance of the natural mosquito populations.