Mayaro virus and dengue virus 1 and 4 natural infection in culicids from Cuiabá, state of Mato Grosso, Brazil

This study aimed to verify the diversity of Culicidae species and their frequency of infection with flaviviruses and alphaviruses in Cuiabá, state of Mato Grosso, Brazil. Mosquitoes were captured with Nasci aspirators and hand net in 200 census tracts, identified alive at species level and pooled in one-20 (11,090 mosquitoes, 14 species). Female pools (n = 610) were subjected to multiplex seminested-reverse transcription-polymerase chain reaction (RT-PCR) for 11 flavivirus and five alphavirus. Positive pools were tested by single RT-PCR followed by nucleotide sequencing, by RT-PCR for E1 gene [Mayaro virus (MAYV)] and by inoculation in Vero cells (MAYV) or C6/36 cells (flaviviruses). One/171 Aedes aegypti was positive for dengue virus (DENV)-1, 12/403 Culex quinquefasciatus, and four/171Ae. aegypti for MAYV, which was isolated from two pools containing two nonengorged females of Ae. aegypti and two ofCx. quinquefasciatus. DENV-4 was detected in 58/171 pools of Ae. aegytpi, 105/403 Cx. quinquefasciatus, two/five Psorophora sp., two/11 Psorophora varipes/Psorophora albigenu, one/one Sabethes chloropterus, two/five Culex bidens/Culex interfor, and one/one Aedes sp. DENV-4 was isolated from two pools containing three and 16 nonengorged Cx. quinquefasciatus females. Phylogenetic analysis revealed MAYV belongs to genotype L, clustering with human samples of the virus previously identified in the city. Cuiabá has biodiversity and ecosystem favourable for vector proliferation, representing a risk for arbovirus outbreaks.

Arboviruses (arthropod-borne viruses) are widely distributed, predominating in tropical areas. These viruses are maintained in nature by epidemiological cycles involving vertebrate hosts and haematophagous arthropod vectors, mainly mosquitoes, ticks, sandflies, and biting midges (Bichaud et al. 2014).
Female mosquitoes are infected by arboviruses during haematophagy in amplification hosts, such as birds, primates, and humans. Vector competence is a component of vector capacity and comprises a variety of biological and genetics factors that combined determine the ability of the invertebrate to transmit infectious agents (Beerntsen et al. 2000). After an extrinsic period of incubation of eight-14 days, a life-long persistent infection is established in the salivary glands of competent vectors (Forrester et al. 2014). Although less frequent, transovarial and venereal transmission also occur in mosquitoes (Coffey et al. 2013).
The state of Mato Grosso (MT), located in Central-West Brazil, presents ecological and climatic conditions favourable to arbovirus circulation. The state is composed of the Amazon, Pantanal, and Cerrado biomes. Cuiabá is the capital of MT and a significant number of confirmed or suspected dengue cases are reported annually in the city (Acendino 2013). Studies involving patients with acute febrile illness during a large dengue outbreak demonstrated the hyperendemicity of DENV and the circulation of SLEV and MAYV in Cuiabá during 2012 (Zuchi et al. 2014, Heinen et al. 2015a. However, until the present, studies to evaluate the genetic diversity and frequency of these viruses are lacking. Toward these ends, the present study was aimed at verifying the diversity of Culicidae species and their frequency of infection by flaviviruses and alphaviruses in Cuiabá. Mosquitoes present a life-long persistent infection and are a common host for several arboviruses. Moreover, these invertebrates represent an important tool for determining the epidemiological status of arboviruses in a given area.

MATERIALS AND METHODS
Study area and Culicidae capture -Cuiabá is located in the south-central region of the state, between the coordinates 15º35'56''S 56º06'01''W, comprising 3,495,424 Km 2 of surface area and 575,480 inhabitants. The climate is semihumid tropical with mean temperature ranging from 21.4-32.8ºC, characterised by a dry season from April-September and a rainy season between October-March (IBGE 2011).
Cuiabá is administratively divided into four regions (north, south, central-east and midwest) and 173 neighbourhoods. The neighbourhoods are further subdivided into 804 census tracts (IBGE 2011), based on demographic density (each census tract represents 250-350 residencies) among which 200 census tracts were randomly selected and located using GPS for adult Culicidae collection between January-April, 2013. Three sites were sampled for at least 30 min within each selected unit between 01:00 p.m.-05:00 p.m., using Nasci aspirators (Horst Armadilhas, Brazil) and hand net. The adult specimens were transferred immediately to entomological recipients using manual suction tube (Castro catcher) and transported with adequate humidity and temperature conditions provided with access to 20% glucose solution.
At the laboratory, mosquitoes were identified alive, in a dormant state (4 min at 4ºC), using a dichotomous key (Forattini 2002) within 12-24 h. Alternatively, a nested polymerase chain reaction (PCR) for Culex quinquefasciatus identification (Smith & Fonseca 2004) was performed in 193 pools of Culex sp. after DNA extraction, since morphological identification of the species was not possible in these pools. They were grouped according to date, collection site, species, and gender in pools of one-20 specimens, and stored at -80ºC immediately.
Total RNA and DNA extraction -Specimens were macerated in 800 µL of phosphate-buffered saline and centrifuged at 5,500 g for 4 min at 4ºC. Total RNA was extracted from 400 µL of the supernatant with Trizol reagent (Invitrogen, USA). The pellet obtained during RNA extraction of Culex sp. pools (n = 193) was subjected to total DNA extraction. Procedures followed the manufacturer's instructions.
RT-PCR for a region of the envelope gene 1 (E1) of alphaviruses -A region of the E1 of MAYV-positive pools was amplified as described by Powers et al. (2001) with few modifications. Briefly, total RNA was reverse transcribed with the primer T 25 V-Mlu (50 µM), 2.5 mM MgCL 2 , 100 U of reverse transcriptase (GoScript; Promega, USA), 20 U of RNAse inhibitor (RNAse OUT, Invitrogen) and 1x buffer according to cycling conditions described by the manufacturer. PCR amplification was performed with 6 µL of cDNA, 50 µM of the primer α10247A, 10 mM of dNTP Mix, 2 mM MgCL 2 , 1x buffer, 1 U of DNA polymerase (GoTaq Hot Start; Promega), and ultrapure water to 50 µL reaction volume. Cycling conditions were used as described by the authors, except for the annealing temperature (45ºC instead of 49ºC). The PCR bands (1.3 kb) were excised from the agarose gel, purified with nucleospin kit (Macherey-Nagel, Germany), and sequenced as previously described.
Inoculation of positive pools in cell culture -For virus isolation, the supernatants of MAYV-positive pools were inoculated at 1:10 dilution in Vero cells (ATCC CCL-81) and the pool positive for DENV-1, 10 for DENV-4, and one for SLEV in C6/36 cells (ATCC CLR-1660) cultivated in 24-well polystyrene plates. After 2h of incubation, the inoculum was removed and the monolayers were washed with RPMI-1640 medium (Vero cells) or L-15 medium (C6/36 cells) containing antibiotics. Culture medium (RPMI-1640 or L-15 with 5% foetal bovine serum) was replaced, the plates maintained at 37ºC (Vero cells) or 28ºC (C6/36 cells) in 5% CO 2 incubators, and the monolayers monitored for seven days by examination through inverted microscope. The supernatant was stored at -80ºC and the monolayers subjected to total RNA extraction for single-RT-PCR as described above. Three passages in cell culture were performed for each pool.
Data analysis -Data according to census tract, date of capture, species, gender, and number of specimens were compiled using Microsoft Excel 2013. Minimum infection rate (MIR) was calculated using the formula [(number of positive pools/total specimens of the species tested) x 1,000] (Chow et al. 1998) (Table) and presented with confidence interval of 95% (Epidata Analysis, 2006. Geospatial data were plotted in maps (ArcMap, ESri ArcGIS).
Phylogenetic analysis -A neighbour-joining phylogenetic three was generated with the nucleotide sequences of a region of NsP1 gene amplified from MAYV-positive pools with Tamura-Nei distance model and 1,000 bootstrap (Geneious software 6.0). Reference sequences, previously classified in genotypes L and D, were retrieved from GenBank (PubMed; NCBI) for comparison. Outgroups included Trocara virus, EEEV, and CHIKV.

RESULTS
Mosquitoes collected in Cuiabá -Between January-April 2013, 11,090 adult Culicidae specimens were collected in Cuiabá, including 6,534 (58.9%) males and 4,556 (41.1%) females. This resulted in 1,419 pools, of which 610 were comprised of females, 114 collected in the midwest region, 180 in the central-east, 118 from the north, and 198 from the south of the city (Table).

Frequency of infection by arboviruses in adult fe-
male mosquitoes -Since only females are haematophagous and due to the large number of specimens collected in the experiment, only the 610 female pools were tested for arboviruses. Of these, 28% (n = 171) were positive for flaviviruses and 2.6% (n = 16) for alphaviruses (Table). Pools of adult females positive for more than one arbovirus represented 0.1%, with three of these five pools with dual infection by DENV and MAYV being composed of only one female without signs of engorgement.
One pool obtained at Bela Vista neighbourhood, captured in February 2013 in the central-east region of Cuiabá containing 18 nonengorged specimens of Ae. aegypti, was positive for DENV-1 (MIR = 0.92 ± 1.7) (KP710881), DENV-4, and MAYV (KP710882) (Fig. 1). Its nucleotide sequence presented, respectively, 99% identity with the DENV-1 HNRG27213 (KC692513.1), genotype V, American/African lineage strain, obtained from humans in 2010 in Argentina, and 99% with the MAYV sequence MAY_BR/MT_CBA_230/2012 obtained from humans in 2012, in Cuiabá (KJ879253.1). One pool obtained in other collection site in the same neighbourhood, containing one female of Cx. quinquefasciatus captured on the same day, was positive for SLEV. The nucleotide sequence (KJ801827) presented 99% of identity with the SLEV previously identified in humans with acute febrile illness in MT during a severe dengue outbreak, belonging to genotype VA (Heinen et al. 2015b) (Fig. 1A).
Unspecific amplification of mosquito sequences was identified within the 1.3 kb amplicon obtained through the E1 gene RT-PCR in MAYV positive pools. Phylogenetic analysis of the nsP1 partial sequences obtained from these pools revealed a similarity with sequences of the virus obtained from human samples in MT. These sequences presented a high similarity with other sequences of MAYV belonging to genotype L from Ixodes spp, Haemagogus janthinomys and humans in the state of Pará (PA) (Fig. 2).

DISCUSSION
This is the first study to investigate the diversity of Culicidae species and their frequency of infection by medically important flaviviruses and alphaviruses in Cuiabá. Cx. quinquefasciatus and Ae. aegypti occurrence has been Cx. quinquefasciatus, the most abundant species, was present in 92% of the census tracts and was positive for DENV-4, SLEV, and MAYV (Fig. 1, Table). The prevalence of this species might have resulted from the competitive success for breeding sites and blood meal sources in comparison to other culicids. This species is a competent vector for WNV, SLEV, Oropouche virus, and MAYV (Hoch et al. 1987, Segura & Castro 2007, Abad-Franch et al. 2012. Data concerning SLEV circulation in Brazil are probably underestimated. Human cases were reported in the Amazon Region, in the city of São Paulo (Rocco et al. 2005, Mondini et al. 2007, and in Cuiabá (Heinen et al. 2015b). One pool containing a nonengorged female of Cx. quinquefasciatus from midwest Cuiabá was positive for SLEV subgenotype V-A during this study; the nucleotide sequence of which presented a high similarity to the SLEV detected in humans in this same area in 2012 (Heinen et al. 2015b).
MAYV is endemic in the Amazon Region, where is maintained in sylvatic cycles involving Ha. janthinomys as the main vector and birds and primates as amplifying hosts. Human infections are generally incidental, due to occupational exposure (Pinheiro & Leduc 1998). An urban cycle of transmission has been proposed for MAYV in Manaus  and Cuiabá (Zuchi et al. 2014). In this study, viral isolation was achieved from two pools of Cx. quinquefasciatus and Ae. aegypti, which were only positive to MAYV, with no signs of engorgement. Isolation of MAYV from other Culicidae species such as Aedes, Culex, Psorophora, Sabethes, and Haemagogus spp has been reported (Pinheiro & Leduc 1998, Long et al. 2011. Vector competence of Cx. quinquefasciatus and Ae. aegypti for MAYV transmission has been demonstrated experimentally (Long et al. 2011, Abad-Franch et al. 2012. The importance of these species as MAYV vectors is based on their urban and anthropophilic habits and their wide geographic distribution, favouring the contact with humans and, therefore, viral transmission and urbanisation of MAYV. The positivity of these species to MAYV in the present study might have resulted from natural infection in the mosquito, since most of the positive pools were not engorged and therefore the virus detected was not acquired from the blood. Additional studies are necessary to elucidate the involvement of these mosquito species in the epidemiological cycle of MAYV. Nucleotide sequences of MAYV obtained from Cx. quinquefasciatus showed 98-100% identity with sequences of the virus obtained from humans in Cuiabá (KJ879256.1; KJ879257.1) and in the city of Várzea Grande, MT (KJ879258.1). In a similar vein, sequences obtained from Ae. aegypti showed 99-100% of identity with virus sequences obtained from humans in Cuiabá (KJ879253.1; KJ869256.1) during 2012, indicating the same virus was identified in humans and in mosquitoes in MT (Zuchi et al. 2014). The patients positive for MAYV in Cuiabá in 2012 were urban residents and denied recent travel to rural or sylvatic areas. Associations to sex and occupation were not identified (Forattini 2002). Taken together, these findings corroborate the occurrence of urban transmission in Cuiabá. previously documented in Cuiabá (Campos & Andrade 2003, Carvalho-Leandro et al. 2010. In 2012, DENV hyperendemicity and SLEV and MAYV circulation were also reported in the city (Zuchi et al. 2014, Heinen et al. 2015a. In the same year, the frequency of transovarial infection of Ae. aegypti by DENV-4 was 10.5% in a region of Cuiabá with a high incidence of dengue (Cruz et al. 2015). However, studies involving adult Culicidae and other arboviruses were lacking in this area. Generally, mosquitoes infected by arboviruses are detected for up to six weeks prior to the beginning of an outbreak (Chow et al. 1998). Moreover, infected mosquitoes carry the virus for their entire life, presenting as an epidemiological indicator of arboviruses circulation in a given area.
Although E2/E1 structural sequences of alphaviruses are more appropriate for phylogenetic analysis, a similarity in the clustering among E2/E1 and nsP1 sequences has been described (Vieira et al. 2015). This finding was also observed in this study. MAYV nucleotide sequences formed a monophyletic group separated in clusters comprising genotypes L and D. The genotype D is composed of several isolates from South America and genotype L of strains from northern and central regions of Brazil (Powers et al. 2006, Vieira et al. 2015. All mosquitoes and human sequences of nsP1 from MT grouped within the genotype L of MAYV in the same cluster and are in close proximity to sequences obtained from Ixodes spp and humans in PA. These results indicate that the same virus might be circulating in mosquitoes and humans in this region. Vieira et al. (2015) suggested that MAYV isolates consist of a sympatric group in MT.
Ae. aegypti females were the second most frequent specimens, captured in 165/200 census tracts of Cuiabá. This species is highly anthropophilic, extremely adaptable to diverse urban environmental conditions, and it is widely distributed in Brazil DENV-1, 4 were identified among culicids in Cuiabá. The increased incidence of DENV-4 in humans in the metropolitan area of Cuiabá during 2011-2012 (Heinen et al. 2015a) was accompanied by a significant frequency of this serotype in mosquitoes during 2013. Among 171 pools positive for DENV-4, 58/171 (9%) were of Ae. aegypti. Culex, Psorophora, and Sabethes spp positivity to DENV-4 may have resulted from the presence of viraemic blood in the abdomen of engorged females, due to haematophagy in humans or from natural infection of these specimens. Ng et al. (2011) also reported the detection of viruses in mosquitoes arising from the alimentary source. Natural infection of different mosquito species with arboviruses is commonly described; although there is no confirmed association with vector competence for DENV serotypes for these species, i.e., participation in the transmission cycle. Moreover, Cx. quinquefasciatus naturally infected by DENV-2 after a blood meal on viraemic host was already described (Luo 1993) and Haemagogus leucocelenus positivity for DENV-1 in Bahia, where a sylvatic cycle of transmission was suggested, represents a possible adaptation of urban viruses in sylvatic mosquitoes and their maintenance in enzootic cycles (Figueiredo et al. 2010). In Brazil, natural infection of Psorophora spp by Maguari virus, ILHV, MAYV, YFV, and ROCV (de Souza-Lopes et al. 1981), and of Sa. chloropterus by SLEV, an important YFV vector, were reported previously (Svoboda et al. 2014).
Most of the collected mosquitoes and positive pools are from the south, followed by central-east, north and midwest regions of the city. These regions also present socioeconomic conditions favouring the proliferation of mosquito populations. The south and central-east regions presented elevated MIR relative to other regions. In 2013 alone, 3,750 dengue cases were reported in Cuiabá (SES-MT 2014). The neighbourhoods with higher dengue incidence are part of the administrative regions with a high number of positive pools. Higher rates of Ae. aegypti infection correlate with increased possibility of DENV transmission and are directly linked to human infections, since most adult females become infected after biting an infected host (Chow et al. 1998, Urdaneta et al. 2005, Zeidler et al. 2008. Dengue hyperendemicity was previously described in Manaus, in the city of Rio de Janeiro, and Cuiabá (Nogueira & Eppinghaus 2011, Heinen et al. 2015a). The co-circulation of more than one DENV serotype associated to susceptibility of human populations to the serotypes and density of Ae. aegypti populations are factors contributing to the occurrence of epidemics (Guedes 2012). For these reasons, understanding vector population characteristics is essential to establish appropriate vector control measures and, consequently, checking arboviruses dissemination.
The high mosquito population density identified in the study may explain dual positivity for MAYV and DENV-4 detected in five pools. Since most of the pools are composed of more than one specimen, probably different females were infected by different arboviruses, although co-infection could also occur and one or both viruses could also originate from haematophagy of a viraemic hosts in the pools with engorged females. Co-infections by DENV and CHIKV in Ae. albopictus pools have been described from Africa, raising the possibility of human dual infection through one single mosquito bite. Experimentally, co-infection and super-infection of this species by both viruses have been demonstrated (Vazeille et al. 2010).
Cuiabá represents a heterogeneous environmental setting. The uncontrolled demographic growth, precarious sanitary conditions, and of infrastructure observed in several neighbourhoods, associated to the presence of growing areas, parks, streams, and rivers inside the urban perimeter, may explain the presence of sylvatic culicid species in the study. Permanent preservation areas (extensive areas with native sylvatic woods in the urban perimeter protected by law) and parks also allow the existence of birds and primates that are important amplifying hosts for several arboviruses (Navarro- Silva et al. 2004).
The knowledge of culicid species diversity and their frequency of infection by arboviruses in regions where epidemics are frequent are important in predicting which arboviruses may circulate in humans or disseminate after their introduction. These observations highlight the necessity to deepen our understanding about the epidemiological cycle and molecular evolution of these viruses in MT, as well as to improve entomological surveillance measures.