1. Introduction
Arboviruses (arthropod-borne viruses) are the agents of some of the emerging and re-emerging diseases with the greatest impact on human health. Yellow fever is a mosquito-borne disease caused by the yellow fever virus (YFV), which occurs in Africa and the Americas in two main transmission cycles: urban and sylvatic. These are distinguished by the ecological nature of the transmission, or, more specifically, the mosquito vector species: the anthropic
Aedes aegypti in urban areas and several sylvatic Aedini and Sabethini species in forested environments [
1,
2]. In South America, YFV has been maintained only in the sylvatic transmission cycle involving non-human primates (NHPs) and arboreal mosquitoes of the genera
Haemagogus and
Sabethes, where humans can acquire the infection during epizootics [
2,
3,
4,
5].
In 2017–2019, Brazil recorded its most serious outbreak when YFV spread in the Atlantic Forest in the southeast region, where it had not circulated for almost 80 years [
2]. During this outbreak, a high number of epizootics were reported, hundreds of monkeys were killed, and some NHP species were eradicated from some forest fragments and environmental conservation areas. Moreover, 2170 confirmed human infections with 932 deaths were recorded [
5].
Haemagogus spp. play a key role in the transmission of sylvatic yellow fever in South America, mainly due to its primatophilic habit, which facilitates its contact with viremic NHPs [
6,
7,
8]. Two species were identified as the primary YFV vectors during the 2017–2019 outbreak:
Haemagogus janthinomys/capricornii and
Haemagogus leucocelaenus [
8]. Among the species of the genus
Sabethes,
Sabethes chloropterus (Humboldt, 1819) and
Sabethes albiprivus (Theobald, 1903) are commonly found in enzootic YFV foci and were found to be naturally infected with the virus. However,
Sa. chloropterus has greater local epidemiological prominence and was considered a secondary vector of YFV in the Atlantic Forest bioregion [
8,
9,
10].
Brazil has the greatest diversity of species in the world, with more than 20% of the total species on the planet [
11]. The Atlantic Forest is classified as one of the five areas with the highest rates of endemicity and diversity of fauna and flora worldwide [
12].
The occurrence of species depends on factors such as the structural heterogeneity of the biome, with more diverse habitat resources allowing for more species to coexist while also minimizing the effect of competition and, consequently, increasing local biodiversity. Variations in the number of species between communities can be represented and quantified in several ways, with the most common being through diversity indices [
13,
14].
The spatial distribution of mosquito species is mainly related to the choice of oviposition site, which is strongly influenced by climatic and environmental factors. In addition, some Culicidae species’ eggs can withstand a period out of water without hatching, which is a natural ability of mosquitoes that lay their eggs out of water. This ability, which can be influenced by extrinsic factors, allows these species to withstand periods of drought [
7].
The triggering, spread, and severity of an outbreak of a zoonotic arbovirus, such as YFV, are multifactorial. In the case of the 2017–2019 YFV outbreak in southeastern Brazil, the main drivers identified were poor vaccination coverage; increased areas with conditions favorable for NHPs; mosquito population growth in some parts of the Atlantic Forest combined with the loss of natural habitats in others; human contact with the forest; and the ecological plasticity of vectors, such as the ability of sylvatic mosquitoes to bite beyond the forest limits [
2]. This last factor may have an important role, as a high proportion of those infected during the outbreak did not report having entered the forest, though they lived near or approached forest fragments for leisure, resting, and working (planting, weeding, harvesting). The zones with intermediate levels of forest cover and high exposure to the forest edge were identified as being more prone to the occurrence of human infections by YFV [
15,
16,
17].
In this study, we identified and described mosquito fauna and spatial distribution of mosquito species in a fragment of the Atlantic Forest in Rio de Janeiro (RJ) state impacted by the 2017–2019 outbreak to evaluate the risk of YFV infections in distinct environments.
3. Results
A total of 9349 mosquitoes were collected (
Table 1). Just over half of this total (51.1%) was collected with ovitraps, with 5616 specimens reared from eggs (85.0%), and the remainder comprised the larvae and pupae found in the water (15.0%). Most of the 3733 adults (87.7%) were captured with the BG-Sentinel trap and the rest via PHA (12.3%).
Twenty-one species belonging to 12 genera were identified (
Table 1). Most individuals of genera
Culex (Culex) spp. and
Wyeomyia could not be identified at the species level. Five species were collected using all the methods (ovitrap, BG-Sentinel, and PHA): the main YFV vectors
Hg. leucocelaenus and
Hg. janthinomys/capricornii, as well as
Aedes albopictus,
Aedes terrens, and
Limatus durhamii (
Table 1). On the other hand, some species were collected only with PHA captures:
Anopheles cruzii,
Psorophora ferox,
Runchomyia cerqueirai,
Wyeomyia incaudata,
Wyeomyia theobaldi, and the secondary vectors of YFV
Sa. chloropterus and
Sa. albiprivus.
Ae. aegypti,
Limatus pseudomethysticus, and the predator
Toxorhynchites sp. were only collected with ovitraps.
Table 1.
Number and percentage of mosquito specimens collected across all sampling points and capturing methods at the Centro de Primatologia do Rio de Janeiro from December 2018 to December 2019.
Table 1.
Number and percentage of mosquito specimens collected across all sampling points and capturing methods at the Centro de Primatologia do Rio de Janeiro from December 2018 to December 2019.
Genus/Species | Adults | Ovitrap | Total |
---|
Protected Human Attraction | BG-Sentinel | Total | Paddle | Water | Total |
---|
Aedes aegypti (Linnaeus) | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 14 | 0.3% | 0 | 0.0% | 14 | 0.2% | 14 | 0.1% |
Aedes albopictus (Skuse) | 8 | 1.7% | 2 | 0.1% | 10 | 0.3% | 11 | 0.2% | 2 | 0.2% | 13 | 0.2% | 23 | 0.2% |
Aedes scapularis (Rondani) | 59 | 12.9% | 63 | 1.9% | 122 | 3.3% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 122 | 1.3% |
Aedes terrens (Walker) | 4 | 0.9% | 8 | 0.2% | 12 | 0.3% | 788 | 16.5% | 0 | 0.0% | 788 | 14.0% | 800 | 8.6% |
Anopheles cruzii (Dyar & Knab) | 1 | 0.2% | 0 | 0.0% | 1 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 1 | 0.0% |
Culex urichii (Coquillett) | 0 | 0.0% | 1 | 0.0% | 1 | 0.0% | 0 | 0.0% | 444 | 52.7% | 444 | 7.9% | 445 | 4.8% |
Culex (Mcx.) sp. | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 7 | 0.8% | 7 | 0.1% | 7 | 0.1% |
Culex (Culex) spp. | 41 | 9.0% | 3030 | 92.5% | 3071 | 82.3% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 3071 | 32.8% |
Haemagogus janthinomys Dyar/Hg. capricornii Lutz | 54 | 11.8% | 3 | 0.1% | 57 | 1.5% | 509 | 10.7% | 0 | 0.0% | 509 | 9.1% | 566 | 6.1% |
Haemagogus leucocelaenus (Dyar & Shannon) | 99 | 21.6% | 17 | 0.5% | 116 | 3.1% | 3452 | 72.3% | 13 | 1.5% | 3465 | 61.7% | 3581 | 38.3% |
Limatus durhamii Theobald | 11 | 2.4% | 2 | 0.1% | 13 | 0.3% | 0 | 0.0% | 60 | 7.1% | 60 | 1.1% | 73 | 0.8% |
Limatus pseudomethysticus (Bonne-Wepster & Bonne) | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 311 | 36.9% | 311 | 5.5% | 311 | 3.3% |
Mansonia titillans (Walker) | 0 | 0.0% | 6 | 0.2% | 6 | 0.2% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 6 | 0.1% |
Mansonia sp. | 0 | 0.0% | 1 | 0.0% | 1 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 1 | 0.0% |
Psorophora ferox (Humboldt) | 1 | 0.2% | 0 | 0.0% | 1 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 1 | 0.0% |
Runchomyia cerqueirai (Stone) | 2 | 0.4% | 0 | 0.0% | 2 | 0.1% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 2 | 0.0% |
Runchomyia frontosa (Theobald) | 24 | 5.2% | 27 | 0.8% | 51 | 1.4% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 51 | 0.5% |
Runchomyia humboldti (Lane & Cerqueira) | 17 | 3.7% | 18 | 0.5% | 35 | 0.9% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 35 | 0.4% |
Runchomyia reversa (Lane & Cerqueira) | 17 | 3.7% | 15 | 0.5% | 32 | 0.9% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 32 | 0.3% |
Runchomyia sp. | 20 | 4.4% | 19 | 0.6% | 39 | 1.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 39 | 0.4% |
Sabethes albiprivus Theobald | 3 | 0.7% | 0 | 0.0% | 3 | 0.1% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 3 | 0.0% |
Sabethes chloropterus (Humboldt) | 21 | 4.6% | 0 | 0.0% | 21 | 0.6% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 21 | 0.2% |
Toxorhynchites sp. | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 5 | 0.6% | 5 | 0.1% | 5 | 0.1% |
Tricoprosopon sp. | 0 | 0.0% | 1 | 0.0% | 1 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 1 | 0.0% |
Wyeomyia incaudata (Root) | 2 | 0.4% | 0 | 0.0% | 2 | 0.1% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 2 | 0.0% |
Wyeomyia theobaldi) (Lane & Cerqueira) | 26 | 5.7% | 0 | 0.0% | 26 | 0.7% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 26 | 0.3% |
Wyeomyia confusa (Lutz) | 0 | 0.0% | 1 | 0.0% | 1 | 0.0% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 1 | 0.0% |
Wyeomyia spp. | 48 | 10.5% | 61 | 1.9% | 109 | 2.9% | 0 | 0.0% | 0 | 0.0% | 0 | 0.0% | 109 | 1.2% |
Total | 458 | 100.0% | 3275 | 100.0% | 3733 | 100.0% | 4774 | 100.0% | 842 | 100.0% | 5616 | 100.0% | 9349 | 100.0% |
The capture time differed between the three methods, with the ovitrap and the PHA operating for the longest and shortest time, respectively. Even so, the collection method with the greatest diversity and richness was PHA (H’ = 2.26 and 17 species), followed by BG-Sentinel (H’ = 2.02). The paddles of the ovitraps had the lowest diversity (H’ = 0.80) and richness (S = 5) and, consequently, a high level of dominance (D = 0.56), with the species
Hg. leucocelaenus (n = 3452) and
Hg. janthinomys/capricornii (n = 509) representing 72% and 11% of the total number of individuals collected, respectively (
Table 2). However, the paddles had by far the highest abundance (n = 4774), or 78% of the total number of Culicidae collected and identified at the species level.
A total of 4774 adult mosquitoes were obtained from hatched eggs gathered from 2808 recovered paddles collected fortnightly over the 13 months of the study (26 collections), with 842 adults obtained from larvae and pupae found in the water of the ovitraps. The most abundant species among the immatures collected with ovitraps were
Hg. leucocelaenus (61.7%),
Ae. terrens (14.0%), and
Hg. janthinomys/capricornii (9.1%) (
Table 1). These species were the most abundant among the adults that emerged from the eggs found on the paddles (72.3%, 16.5%, and 10.7%, respectively). In contrast, of the larvae and pupae found in the water of the ovitraps, the most abundant species were
Culex urichi (52.7%) and
Li. pseudomethysticus (36.9%) (
Table 1).
Culex (Culex) spp. were the most abundant (82.3%) across all collections of adults. However, the composition of the fauna and abundance changed when analyzing the results of adult collections according to the method used.
Culex (Culex) spp. accounted for 92.5% of the total collected with BG-Sentinel with CO
2 as an attractant, but only 9.0% of the mosquitoes captured with PHA, where other species were much more abundant, such as
Hg. leucocelaenus (21.6%) and
Ae. scapularis (12.9%) (
Table 1).
The abundance of immature forms in ovitraps (
Table 3) was significantly heterogeneous among the sampling sites (P1–P9; Kruskal–Wallis,
p = 0.01), with P3 (inside the forest) having the highest abundance of specimens sampled with ovitraps, followed by P8 (situated roughly in the intermediate ecotone between the modified environment and the forest) (
Table 3). The abundance of eggs collected with ovitrap paddles was greater than that of the larvae and pupae found in the water at all sampling sites except P6 (
Table 3). Sampling sites P4 and P5 were similar in species composition and abundance, considering the specimens gathered with the ovitraps, but differed from the other sites. The Kruskal–Wallis analysis indicated significant differences in the abundances of mosquito specimens obtained with BG-Sentinel versus the ovitraps at P2, P4, and P5 (P2:
p = 0.037, P4:
p = 0.000, P5:
p = 0.017). Dunn’s post-test indicated that the abundances in the BG-Sentinel at these three points differed from those of the ovitrap paddles and ovitrap water.
Adult mosquito collections made with BG-Sentinels had the greatest abundance at P5 (51.5%), located in the forest ecotone (
Table 3), with a practically uniform distribution at the remaining sampling sites, ranging from 9.1% (P3) to 13.5% (P2), regardless of location, whether inside the forest or the modified environment. A similar pattern was found with captures made with PHA, with no significant difference (Kruskal–Wallis,
p = 0.50) in the abundances between captures made inside the forest (P10 and P11) and the modified environment (P12). However, when analyzing the species composition by sampling site (
Table 4), P5, which was located in the forest ecotone, had the highest number of mosquito specimens (n = 2047), while the greatest richness was recorded at two sites located inside the forest: P4 and P11. In contrast, the greatest diversities were registered in captures made with PHA at P11 (inside the forest) and P12 (in the modified environment). Furthermore, the highest evenness values were observed at the three collection points where captures with PHA were conducted (P10 to P12), regardless of the site environment. That is, there was a more similar distribution of the number of specimens of each mosquito species when captures were made with PHA, regardless of whether the sampling occurred in the open field or the forest (
Table 4).
Although Hg. leucocelaenus was more abundant than Hg. janthinomys/capricornii across the study area (Mann–Whitney, p = 0.007), there was no significant difference in the abundances of both species between the sampling sites (Kruskal–Wallis, p = 0.204).
When we analyzed the similarity between the collection points (Sørensen’s similarity (IS)), we found that the highest similarity (IS > 0.81) was between the sites where captures were performed only with PHA (
Table 5). This may indicate that the collection method influenced the composition of the captured fauna more than the sampling location. When comparing the points where there were collections of adults with BG and immature ones with ovitraps (P1 to P5), there was a high similarity for all paired comparisons (IS ranging from 0.59 to 0.83). Among these comparisons, it is interesting to highlight that particularly high similarity values were observed in comparisons between the modified environment (P1) and sites P3 and P4, which were located further into the forest (IS = 0.69 and 0.80, respectively).
A total of 287 adult mosquitoes caught with BG-Sentinel or PHA were screened for YFV, all of which were negative (
Table 6). Almost 86% of tested specimens were
Hg. leucocelaenus (n = 116),
Hg. janthinomys/capricornii (57), and
Ae. scapularis (73).
4. Discussion
Greater awareness of the diversity of mosquito species is fundamental for assessing possible changes in their behavior and adaptations according to the different environmental conditions in areas where the environment has been subjected to and/or is undergoing anthropogenic change [
27].
In this study, 21 species of mosquitoes belonging to 12 genera were recorded in sites located on a disturbance gradient from the open and modified environment to within the Atlantic Forest at CPRJ. Intriguingly, this figure was considerably lower than the diversity of mosquitoes previously described in sites on the same slope of Serra do Mar. For instance, studies undertaken at Serra dos Órgãos National Park and Guapiaçu Reserve, located around 9 and 20 km from the CPRJ, recorded 44 and 59 mosquito species, respectively [
28,
29].
The collection method considerably influenced the results as they related to mosquito species diversity, richness, and abundance across the sampling sites. Species that do not breed in open containers, such as ovitraps, were not collected at points where only this collection method was used (P6–P9). The distribution of
Ae. scapularis illustrated this finding since it was found only at sites where adult collections were performed, whether with BG-Sentinels or PHA. This behavior was also shared by the species of genera
Runchomyia and
Wyeomyia. The distribution of some species appeared to be affected more by the location of the sampling site, as was the case of the two
Sabethes species, which were collected only at sites where captures were carried out with PHA (P10–P12). The use of ovitraps essentially selected species of tribe Aedini, such as
Hg. leucocelaenus,
Ae. terrens, and
Hg. janthinomys/capricornii on the paddles, and
Culex urichii and
Li. pseudomethysticus in the water held in the ovitraps. Adults of
Culex (
Culex) spp. were abundant where BG-Sentinels were used (P1–P5) but accounted for only 9.0% of mosquitoes captured using PHA where, in contrast, primary and secondary YFV vectors, such as
Ae. scapularis,
Hg. janthinomys/capricornii, and
Hg. leucocelaenus, were much more frequent.
Hg. leucocelaenus represented 21.7% of mosquitoes captured with PHA, followed by
Ae. scapularis (12.9%). The latter species was considered a secondary vector of yellow fever during the recent epidemic in Rio de Janeiro [
8], was already found infected with YFV in other parts of Brazil [
30,
31], and was experimentally demonstrated to be able to transmit the virus between monkeys [
32,
33].
Interestingly, around 70% of immatures collected with ovitraps, regardless of their location, belonged to the two
Haemagogus species considered to have been the primary YFV vector in RJ and other outbreaks in southeastern and southern Brazil [
8,
31,
34,
35,
36,
37,
38,
39,
40].
Altogether, the results of adult and immature collections indicated that Hg. leucocelaenus is more abundant than Hg. janthinomys/capricornii. However, there was no difference in the abundance of both species across the sampling sites in the transects. That is, they occurred in similar abundance from the open field and modified environment to 400 m into the forest. This remarkable aspect of the species’ distribution had a major impact on YFV transmission.
These two vectors can move for several kilometers [
41], covering forest and modified environments, a behavior that facilitates the spatial dissemination of the YFV and viral transmission from viremic NHPs living deep in the forest to humans, not only when they enter the forest or the forest edges but also in open fields and anthropic, modified environments with a certain proximity to the forest. Their ability to move from inside the forest, where they breed in tree holes and usually bite NHPs, to attacking humans in the intermediate ecotone or the modified environment was indicated as an important risk factor during the 2017–2019 YFV outbreak. Many people who became infected or died in this outbreak believed themselves unreachable by epizootic transmission because they lived in peri-urban areas and did not enter forests or their surroundings and consequently did not prioritize getting vaccinated [
2]. However, the ability to fly a considerable distance [
41] and the spatial distribution of
Hg. leucocelaenus and
Hg. janthinomys/
capricornii, as observed in the CPRJ, illustrate how people can become exposed to YFV-infected sylvatic mosquito bites in rural, peri-urban, and even urban areas adjacent to the Atlantic Forest.
The failure to detect natural infections in the tested mosquitoes from the CPRJ area was consistent with the local epidemiological data. Our collections started in December 2018 and lasted until December 2019, a period during which no human cases were diagnosed in the entire transmission season from July 2018 to June 2019 across RJ. Moreover, the last suspected infections in NHPs were reported five months earlier (July 2018) and were considered remnants of the previous season. The only and last record of YFV circulation in the state was of a dead howler monkey found approximately 90 km from the CPRJ [
24].
Our findings on mosquito diversity and spatial distribution at CPRJ, particularly concerning the primary and secondary vectors of YFV, also indicated that the captive NHPs were similarly exposed to their bites and the risk of mosquito-borne agents, including arboviruses, regardless of the location of the outdoor enclosures. The local fauna includes several species indicated as vectors of several arboviruses in Brazil [
42]. Antibodies against some
Flaviviruses and other arboviruses were reported in captive NHPs born at CPRJ [
43]. Altogether, these data triggered the vaccination of NHPs against YFV with the YF 17DD attenuated virus, which was promptly initiated in 2018 at CPRJ to protect captive animals of susceptible endangered species and to aid in reducing transmission [
18,
44]. The protocol we used to test mosquitoes can detect both wild-type and vaccinal YFV [
23]. The negative tests of wild mosquitoes collected in this study during the period of vaccination of the animals in the CPRJ corroborate the conclusions of Miranda et al. (2022) regarding a lack of evidence of uncontrolled transmission of this vaccine virus in nature from viremic New World NHPs [
44].