Silent circulation of dengue virus in Aedes albopictus (Diptera: Culicidae) resulting from natural vertical transmission

Vertical transmission in Aedes aegypti and Aedes albopictus is considered a maintenance mechanism for dengue virus (DENV) during unfavorable conditions and may be implicated in dengue outbreaks. Since DENV infection dynamics vary among wild-type viruses and vector populations, vertical transmission rates can also vary between regions. However, even though São Paulo is the most populous city in the Americas and has experienced major dengue epidemics, natural vertical transmission had never been detected in this area before. Here we confirm and describe for the first time natural vertical transmission of DENV-3 in two pools of male Ae. albopictus from the city of São Paulo. The detection of DENV-3 in years when no human autochthonous cases of this serotype were recorded suggests that silent circulation of DENV-3 is occurring and indicates that green areas may be maintaining serotypes that are not circulating in the human population, possibly by a vertical transmission mechanism.

A recent review 15 showed that Asia and South America are the main continents where vertical transmission of DENV in Ae. aegypti and Ae. albopictus under natural conditions has been reported. While Brazil is the leading country in the Americas on studies of natural vertical transmission of DENV, few studies have investigated this phenomenon in the state of São Paulo 18,19 .
The state of São Paulo is the most populous in Brazil, with approximately 45 million inhabitants, representing about 21.5% of the entire Brazilian population (209 million inhabitants) 20,21 . Despite the fact that São Paulo, the capital of the state, is the most populous city in Brazil and Latin America 22 and in recent years has experienced major dengue epidemics, such as those in 2014 and 2015 23 , natural vertical transmission has never been detected in the city 15 .
The green areas in municipal parks in the city of São Paulo are remnants of Atlantic Forest vegetation. These parks are used for recreation and provide habitats for several species of mosquitoes of medical importance, such as those of the Culex, Aedes and Anopheles genera 24 . Previous studies performed in Piqueri Municipal Park, for example, reported a great abundance of Ae. albopictus 24,25 , followed by Culex quinquefasciatus, Aedes fluviatilis, Aedes scapularis, Culex nigripalpus and Ae. aegypti 25 , corroborating the findings of studies that showed that Ae. albopictus is usually present in areas with vegetation cover 7,26 . Despite the body of knowledge on mosquito fauna in green areas in São Paulo, little is known about the pathogens circulating in these locations. Investigation into the occurrence of natural vertical transmission of DENV in Ae. aegypti and Ae. albopictus populations is therefore of critical importance for an understanding of the transmission dynamics of this virus and can provide a new outlook on the maintenance of this pathogen in urban green areas and its possible implications for public health.
This study therefore sought to investigate the occurrence of natural vertical transmission of DENV in Ae. aegypti and Ae. albopictus mosquito populations in the city of São Paulo.

Mosquitoes.
To detect natural vertical transmission of DENV in Ae. aegypti and Ae. albopictus during two seasons (spring and autumn), a total of 5,730 mosquito specimens were analyzed. Of these, 3,270 were Ae. albopictus (1,790 males and 1,480 females) and 2,460 Ae. aegypti (940 males and 1,520 females) ( Table 1) The two positive samples were sequenced, edited, aligned with each other and compared with GenBank sequences using BLAST. They had high degree of identity with DENV-3 serotype sequences from GenBank, as shown in the Neighbor Joining tree (Fig. 3), and only three differences in nucleotide sequence were found between the two samples. Both showed high identity with strains found in the city of Guarujá, SP, Brazil, and Belo Horizonte, MG, Brazil. The positive control sample, which was also sequenced and aligned with GenBank sequences, showed high identity with strains of completely different origin. The DENV-3 sequences found in this study and the positive control sequence were deposited in GenBank, respectively, under accession numbers

Discussion
Since the first reports of vertical transmission under natural conditions, the role of this phenomenon in the persistence and maintenance of arboviruses in nature has been the subject of much discussion 27,28 . The first report of natural vertical transmission of DENV in the Americas was described in specimens from Trinidad and Tobago 29 and was subsequently reported in South American countries, including Bolivia 30 , Argentina 31 and Peru 32 . In Brazil, the first report of natural vertical transmission was of DENV-1 in Ae. albopictus from the state of Minas Gerais 33 . Subsequently, other occurrences in this species involving other dengue serotypes were also reported in   Under laboratory conditions, females of Ae. albopictus were able to transmit all four DENV serotypes vertically to the offspring. The transmission rates varied according to the virus serotype and strain, and the lowest rate found was for DENV-3 45 . Although the maintenance of DENV-1 was observed in three consecutive generations of Ae. albopictus by vertical transmission 17 , there is no data about the DENV-3 maintenance over successive generations in this species under laboratory conditions. However, the persistence of DENV-3 was demonstrated in Ae. aegypti during seven successive generations 16 .
Vertical transmission was also observed in Ae. aegypti although at a low rate, even for strains known to be more susceptible to oral infection 45 . Similar results were found for a Brazilian DENV strain and populations of Ae. aegypti and Ae. albopictus 46 , indicating that despite the importance of Ae. aegypti for horizontal transmission this species may be less relevant in the maintenance of DENV during periods of low vector density 45,47 . Our results corroborate these findings as we found positive Ae. albopictus pools but no positive Ae. aegypti pools. Other hypothesis for the absence of positive pools of Ae. aegypti in our study is related to the smaller number of specimens analyzed compared to Ae. albopictus. In addition, green areas favor the presence and abundance of Ae. albopictus when compared to Ae. aegypti 24,25,48 which maintains DENV circulation in areas with high concentrations of human beings and houses 49,50 . This may have increased the chances of finding viral circulation in Ae. albopictus and not in Ae. aegypti. The presence of DENV-3 in Ae. albopictus from Piqueri Municipal Park suggests that this mosquito may be acting in the transmission dynamics of DENV, thereby maintaining circulation of the arbovirus in this environment.
In a study conducted with the DENV-2 serotype and strains of Ae. aegypti mosquitoes with high and low susceptibility to infection, Mourya et al. 51 argue that vertical transmission is more frequent in eggs allowed to hatch after longer periods, probably because of an increase in viral copy numbers in the eggs. It seems, therefore, that unfavorable environmental conditions, such as low availability of water, and the consequent increase  48 . Another possible explanation for our findings is that during these two periods more specimens were tested.
Interestingly, no autochthonous human cases of DENV-3 have been reported in the city of São Paulo since 2010 23 (Supplementary Fig. S3). In fact, DENV-3 circulation was detected in the city of São Paulo in 2010 52 , but this serotype was not identified in this area during the epidemic of 2014 53 . This strengthens the hypothesis that vertical transmission could help maintain viral circulation during interepidemic periods and suggests that there is silent circulation of DENV-3 in green areas of São Paulo, a city with no detected autochthonous cases of infection by this serotype.
It is important to highlight that for natural vertical transmission to be detected under the conditions established in this study, an infected female had to have laid eggs in the ovitrap and successfully transmitted the virus vertically and the larvae had to have survived until adulthood to finally be analyzed. When mosquito infection rates are low, more specimens are required for a higher chance of virus detection. For example, for an infection rate of 1, a minimum of 1,609 and 2,301 mosquitoes are needed for an 80% and 90% chance of virus detection, respectively 54 . Although in our study we analyzed 2,460 Ae. aegypti adults and 3,270 Ae. albopictus adults that were raised in the laboratory from eggs collected in the park, if we consider that both species lay their eggs in more than one container in the same gonotrophic cycle 55,56 , it is possible that most of the mosquitoes analyzed originated from the same mother. Nevertheless, considering the difficulty in detecting this phenomenon in natural environments, the successful detection of natural vertical transmission of a serotype with no recent autochthonous cases in São Paulo 23 in Ae. albopictus males in two different seasons indicates that the virus is circulating at significant rates in this vector population and being transmitted vertically.
The MIR values found in the present study are similar to those previously described in the literature, which are mainly low and vary considerably according to the mosquito population 57 . In addition, the fact that the MIR typically exhibits considerable temporal variability in vector populations 58 explains the different results for the various seasons analyzed in this study.
As urban parks are open to the public, infected visitors can contribute to the expansion of several arboviruses. Furthermore, the continuous contact between humans, competent vectors and viruses make these parks high-risk areas for infection by DENV and other arboviruses, raising the question of the relevance of urban green areas in the transmission and maintenance dynamics of arboviruses [59][60][61] . Bearing in mind the biological characteristics of Ae. albopictus, such as its vector competence for dengue, yellow fever, Zika and chikungunya viruses in laboratory conditions 46,62-64 , our findings reinforce the importance of surveillance and control of this species as it may be playing a role in maintaining circulation of DENV, favoring infection of those who live close to parks or visitors to these parks and expanding the distribution of the virus to other areas.

Study area. To investigate the occurrence of vertical transmission of DENV in female and male Ae. aegypti
and Ae. albopictus mosquitoes, adults were raised under laboratory conditions from eggs collected in urban green areas of Piqueri Municipal Park (23°31′39.98″S and 46°34′24.88″W), in Tatuapé district, in the east of the city of São Paulo, Brazil (Fig. 4). This park was selected as the study area because it is in the region of São Paulo that had one of the highest numbers of autochthonous human cases of dengue during the last major epidemic of the disease, in 2014/2015 23 . The park, which extends over approximately 100,000 m² and has dense vegetation, harbors approximately 90 species of animals 63  www.nature.com/scientificreports www.nature.com/scientificreports/ Thirty-six ovitraps distributed around Piqueri Municipal Park were used to collect Ae. aegypti and Ae. albopictus eggs. All the paddles were collected and replaced weekly and taken to the Public Health Entomology Laboratory at the School of Public Health, University of São Paulo (FSP/USP).
Mosquito rearing. The paddles were allowed to dry at room temperature for 3-4 days, assessed for the presence of eggs under a stereoscope and placed individually in plastic trays containing 500 mL of water and ground fish food (Tretramin ® ) for larval hatching. All larvae in the L4 stage were identified and separated by species (Ae. aegypti or Ae. albopictus). Pupae were placed individually in small glass containers, and after the adults emerged they were killed by freezing and separated by species, season and sex into pools of 10 individuals, which were placed in 1.5 mL polypropylene microtubes. All 573 pools were kept at −80 °C until used to isolate DENV. were used for RNA extraction according to the manufacturer's instructions. To avoid non-specific amplification, we treated the viral RNA with amplification-grade DNase I (Invitrogen, California, USA) following the manufacturer's protocol.

Isolation of viral genetic material.
RT-PCR was performed according to Lanciotti et al. 65 with a second round of PCR amplification (semi-nested PCR) that allows simultaneous detection of the four DENV serotypes by generating DNA products of unique sizes, which can be used to differentiate the four serotypes (  www.nature.com/scientificreports www.nature.com/scientificreports/ confirmation, all samples showing positive bands were reanalyzed from the extracted RNA using only the oligonucleotides specific to the serotype detected. Nucleotide sequencing. The resulting PCR products were purified with Exosap-IT (GE) following the manufacturer's instructions. The purified DNA was then sequenced by the Sanger method using the Big Dye Terminator Sequencing Kit (ABI) and DENV serotype specific primers (Table 2) in the ABI 3100 Automated DNA Sequencer (ABI). The double-stranded DNA fragments were sequenced in both directions to generate a consensus double-stranded sequence for each sample. Editing, analysis and alignment of the nucleotides were conducted using version 6.0.7 of Bioedit Sequence Alignment Editor. All nucleotide sequences were aligned using ClustalW 66 , and sequence polymorphisms within the amplified region were identified by comparing our sequences with a variety of sequences of the four dengue virus serotypes available from GenBank (Supplementary  Table S1). A similarity tree was built (Fig. 3) with Neighbor Joining algorithm, based on a Kimura 2-parameter model, nucleotide distances, determined by MEGA software (Molecular Evolutionary Genetics Analysis, version 6.0), with 1,000 replications in the bootstrap test 67 .
Statistical analysis. The proportion of infected specimens is commonly estimated from the MIR (number of infected pools/total number of specimens x 1000 mosquitoes tested) based on the assumption of only one infected mosquito per positive pool 13,68 . This ratio is reasonable for estimates based on low-occurrence data and small pools, as observed in this study of natural vertical transmission 13,68 .
For the purposes of comparison, we calculated a general MIR for both species considering all the seasons studied (number of infected pools/total number of specimens for all the seasons analyzed x 1000), a MIR for Ae. albopictus only (number of infected pools of Ae. albopictus/total number of specimens analyzed for all seasons x 1000) and a MIR for Ae. albopictus for each season with positive pools (number of infected pools of Ae. albopictus/total number of specimens analyzed for each season with positive pools x 1,000).

Data availability
All data generated or analyzed during this study are included in the manuscript.