Experimental transmission of Zika virus by Aedes japonicus japonicus from southwestern Germany

The invasive mosquito species Aedes japonicus japonicus (Ae. japonicus) is widely distributed in Central Europe and is a known vector of various arboviruses in the laboratory, including flaviviruses such as Japanese Encephalitis virus or West Nile virus. However, the vector competence of Ae. japonicus for the recently emerging Zika virus (ZIKV) has not been determined. Therefore, field-caught Ae. japonicus from Germany were orally infected with ZIKV and incubated at 21, 24, or 27 °C to evaluate the vector competence under climate conditions representative of the temperate regions (21 °C) in the species’ main distribution area in Europe and of Mediterranean regions (27 °C). Aedes japonicus was susceptible to ZIKV at all temperatures, showing infection rates between 10.0% (21 °C) and 66.7% (27 °C). However, virus transmission was detected exclusively at 27 °C with a transmission rate of 14.3% and a transmission efficiency of 9.5%. Taking into account the present distribution of Ae. japonicus in the temperate regions of Central Europe, the risk of ZIKV transmission by the studied Ae. japonicus population in Central Europe has to be considered as low. Nevertheless, due to the species’ vector competence for ZIKV and other mosquito-borne viruses, in combination with the possibility of further spread to Mediterranean regions, Ae. japonicus must be kept in mind as a potential vector of pathogens inside and outside of Europe.


Introduction
Zika virus (ZIKV) is an emerging mosquito-borne virus within the family Flaviviridae that was first isolated from sentinel rhesus macaques in Uganda in 1947 1 . After decades of silent circulation, unprecedented ZIKV epidemics occurred in Micronesia, Polynesia, and, finally, in the Americas in 2015; the hundreds of thousands of human cases finally resulted in the announcement of a Public Health Emergency of International Concern through the World Health Organization 2 . Clinical courses associated with ZIKV infections can range from mild clinical symptoms to severe diseases, including neonatal microcephaly and neurological disorders such as Guillain-Barré syndrome 3 . The mosquito species Aedes aegypti and Aedes albopictus are considered the primary and secondary vectors of ZIKV; however, a wide variety of other Aedes species have been identified as potentially susceptible to ZIKV infection 4 . Recent experimental studies suggested that only Ae. albopictus might play a role in ZIKV transmission in Central Europe, while common members of the genus Culex are probably not important 5,6 . However, north of the Alps, the Asian tiger mosquito is currently established at only a few sites, with relatively low abundance 7,8 . By contrast, the invasive Asian bush mosquito Aedes japonicus japonicus (Ae. japonicus) is widely distributed in Central Europe and is currently established in at least 10 countries, including large parts of Germany 8 . In 2008, the first invasive spreading of Ae. japonicus in Europe was reported in Switzerland 9 , and Ae. japonicus is now listed as one of the nine most dominant mosquito species in Switzerland. Shortly after its introduction in Switzerland, Ae. japonicus was first reported in Germany, followed by the establishment of populations at several sites. In the Netherlands and Belgium, mosquito control programs have been initiated due to the massive Ae. japonicus populations [9][10][11][12] .
Ae. japonicus is a container-dwelling species, colonizing both natural (e.g., bamboo stubs and tree holes) and manmade (e.g., tires and barrels) breeding sites 13 . Due to its tolerance of rather low temperatures, Ae. japonicus has a relatively long seasonal activity compared to other container-breeders 14 . Ae. japonicus has an opportunistic feeding pattern with a preference for mammals, including humans, although avian host species have also been reported 15,16 . Thus, Ae. japonicus could potentially serve as a bridge vector for zoonotic arboviruses. The species is an experimentally proven vector of several flaviviruses, including Japanese Encephalitis virus (JEV), West Nile virus (WNV), and Saint Louis encephalitis virus [17][18][19] , as well as arboviruses of other families, such as La Crosse virus (LACV, Peribunyaviridae) and Chikungunya virus (CHIKV, Togaviridae) 20,21 . Previous studies with an Ae. japonicus population from southwestern Germany also revealed a vector competence for JEV under laboratory conditions 19 .
In light of the continuing spread of Ae. japonicus in Europe and the ongoing circulation of ZIKV in America, the aim of this study was to evaluate whether Ae. japonicus has vector competence for ZIKV under climate conditions representative of tropical and temperate regions.

Results
To assess the suitability of the collected Ae. japonicus for vector competence studies, a small number of specimens were challenged with JEV in a preliminary study. In agreement with previous findings 19 , the Ae. japonicus specimens from southwestern Germany were susceptible to JEV. The infection rate (IR) was 51.9%, with an average amount of viral RNA of 5.6 × 10 8 copies/specimen (n = 27). In addition to the previous experiment, we also investigated the transmission of infectious virus particles by analyzing mosquito saliva following incubation of infected mosquitoes at 27°C for 14 days. The results indicated a transmission rate (TR) of 78.6%.
Subsequently, mosquitoes were analyzed for ZIKV infection. Fourteen days post infection, ZIKV RNA was present in the bodies of challenged Ae. japonicus at all of the tested temperatures (Fig. 1). The relative numbers of ZIKV-positive mosquitoes (Fig. 1) and the amount of viral RNA (Table 1) increased with increasing incubation temperatures. The IR increased from 10% (3/30) at 21°C to 24.1% (7/29) at 24°C and to 66.7% (14/21) at 27°C. This pattern is also reflected in the amount of virus RNA within the mosquito bodies, which increased from 1.2 × 10 4 RNA copies/specimen at 21°C to 2.6 × 10 6 RNA copies/specimen at 24°C to 6.4 × 10 8 RNA copies/specimen at 27°C ( Table 1). Dissemination of the virus was found in mosquitoes kept at 24 and 27°C but not in  Table 1 Calculation of the virus titers from bodies or legs for Ae. japonicus specimens from southwestern Germany following challenge of mosquitoes with infectious blood meals and incubation at 21 , 24, or 27°C for 2 weeks  mosquitoes incubated at 21°C. However, the averaged leg titers were substantially higher at 27°C (6.4 × 10 8 RNA copies/specimen) than at 24°C (8.4 × 10 2 RNA copies/ specimen) ( Table 1). This is also reflected by the detection of infectious virus particles in the saliva of two mosquitoes kept at 27°C, resulting in a TR of 14.3% (2/14) and a transmission efficiency of 9.5% (2/21).

Discussion
To the best of our knowledge, this is the first study on the vector competence of Ae. japonicus for ZIKV (see review by Epelboin et al. 4 ), and our results show a low transmission efficiency at high temperatures. The species is a known competent vector of a variety of arboviruses, including flaviviruses (e.g., WNV, JEV, or Dengue virus), as well as members of other virus families, including Peribunyaviridae (LACV) and Togaviridae (CHIKV) 17,[19][20][21] .
Over the last two decades, Ae. japonicus has successfully invaded Central European countries as well as large parts of North America, and it is found primarily in areas with predominantly temperate climate conditions 8,13,14,22,23 .
Ae. japonicus eggs are resistant to frost and desiccation. Furthermore, the seasonal activity of the species is longer than that of other container-breeding species. Due to these attributes, Ae. japonicus has some developmental advantages over native species that could affect mosquito population patterns as well as pathogen transmission in newly colonized regions. Likewise, it is of considerable interest to collect information on the vector competence of Ae. japonicus for newly emerging viruses such as ZIKV, with a special emphasis on the temperate climate conditions of the species' current distribution range.
Previous studies described relatively low feeding rates of field-caught Ae. japonicus using saturated cotton sticks or feeding systems with chicken skin 19,21 . The experiments presented here demonstrate that artificial feeding via blood drops seems to be an efficient alternative for fieldcaught Ae. japonicus mosquitoes, resulting in a feeding rate of 75%. The experimental results clearly indicate temperature-dependent variations in the susceptibility of Ae. japonicus to ZIKV. Following the experimental challenge with ZIKV-containing blood meals, the number of infected specimens as well as the amount of ZIKV RNA copies per mosquito increased with increasing incubation temperatures (21°C < 24°C < 27°C). This result is consistent with our previous studies performed with ZIKV, where the IRs of various mosquito species from Central Europe were also temperature-dependent 6 . Dissemination was only observed at 24 and 27°C. Infectious virus particles were exclusively detected in two mosquito specimens incubated at an elevated temperature of 27°C, resulting in a TR of 14.3%. There is a high variability in the range of TRs at tropical incubation temperatures (26-28°C) for both Ae. aegypti (21-87%) and Ae. albopictus (18-77%) 6,24-28 . Nevertheless, both species showed considerably higher TRs than Ae. japonicus; whereas Ae. vexans also has a low vector competence for ZIKV, with a TR between 2 and 7% 29 . Likewise, the transmission efficiency of 9.5% for Ae. japonicus at the tropical temperature is lower than the known transmission efficiency for Ae. aegypti (26%) under similar conditions 27 . Previous vector competence studies with Ae. japonicus and various arboviruses were only performed under tropical temperature conditions (i.e., 25-28°C) [17][18][19][20][21] . Therefore, it is unknown if the lack of virus transmission at lower temperatures is a general feature for viral transmission by Ae. japonicus or if this observation is specific for ZIKV. By contrast, the primary and secondary vectors Ae. aegypti and Ae. albopictus can transmit ZIKV below 27°C. Viable ZIKV virus particles were detected in the saliva of Ae. aegypti even at 20°C 30 . Therefore, the lack of transmission by Ae. japonicus is a species-specific observation and not a general pattern for ZIKV.
One explanation for the lack of ZIKV transmission by Ae. japonicus at temperatures below 27°C might be a combination of higher virus replication rates and speciesspecific, temperature-dependent effects on the mosquito microbiome or immune regulatory pathways 31 . However, as shown before, vector competence is influenced by a three-way interaction between the vector population, the virus strain, and temperature 32,33 . Only a suitable combination of these factors allows the virus to replicate and disseminate to the salivary glands to enable transmission through the next bite. However, studies with a combination of one specific mosquito population with one specific virus strain must be interpreted with caution. Even studies with the same combination of vector species and virus can come to varying TRs or even to inconsistent results regarding the species' susceptibility to a virus. Studies performed with either field-caught Ae. aegypti or Ae. albopictus populations from different sites within one country revealed clearly varying TRs for the same ZIKV strains 34,35 . In addition, the vector competence of the same mosquito population can be highly strain-specific. Ae. aegypti from Mexico showed an increased vector competence for African strains of ZIKV compared with an American ZIKV strain 36,37 . Studies with the WNV strain NY99 and Ae. japonicus mosquitoes from northern Switzerland revealed transmission of WNV 38 . By contrast, Ae. japonicus mosquitoes from southwestern Germany were shown to be refractory to the same WNV strain 19 . Similar contradictions have been discussed regarding the vector competence of Culex quinquefasciatus for ZIKV, where some studies detected transmission while others did not [39][40][41] . These differences might be explained by variations in the experimental setup, e.g., the origin of the mosquito population, the virus strain, or vector maintenance protocols in the laboratory 42 . Thus, standardized studies to investigate the vector competence for different local mosquito population/virus strain combinations must be considered to allow a thorough risk assessment. In particular, further analysis of the Ae. japonicus populations from northern America/Asia should be performed to assess the risk of ZIKV transmission in these regions.
Currently, the distribution of Ae. japonicus in Europe is primarily restricted to regions with temperate climates 13 . The lack of vector competence at temperatures below 27°C suggests a limited risk for ZIKV transmission by Ae. japonicus in Europe. However, the rapid spread of Ae. japonicus in Northeast America (as far as 30°N′ latitude in Florida) and the native distribution in Asia at the same latitude illustrate the risk for the species to spread to the Mediterranean region 14,43 . Ae. japonicus may adapt to new environmental conditions and might also have the potential to invade areas of higher temperatures in the Mediterranean region, as has already happened in North America 13 . Nevertheless, for a comprehensive risk assessment of ZIKV transmission in Central Europe, Aedes species such as the native mosquito species Aedes vexans and the invasive species Aedes koreicus should also be considered and investigated as potential ZIKV vectors under temperate climate conditions. Ae. vexans from North America was proven to have a transmission potential for ZIKV of approximately 1-5% at incubation temperatures of 28 or 27°C 29,44 . These low TRs must be considered in light of the locally very high mosquito abundance and aggressive human-biting behavior along rivers in Central Europe 45 . Therefore, further studies should investigate whether this species can transmit ZIKV at lower temperatures. Another candidate of interest is the invasive mosquito species Ae. koreicus, which is closely related to Ae. japonicus. Ae. koreicus is also a vector for arboviruses such as JEV 46 or CHIKV 23 and was quite recently introduced into Central Europe, including in Germany [47][48][49] .
In conclusion, transmission of ZIKV by Ae. japonicus appears to be limited to elevated temperatures. Nevertheless, due to the demonstrated species' vector competence for ZIKV and for other mosquito-borne viruses, in combination with a possible further spread to southern Europe, Ae. japonicus must be considered a potential vector of pathogens, including ZIKV.

Source, rearing, and experimental infection of mosquitoes
Ae. japonicus eggs were collected with ovitraps in southwestern Germany (49°31′26.26″N, 8°40′16.88″E) in summer 2017. Approximately 1200 eggs were flooded in the laboratory, and the larvae and adults were maintained at 26°C, with a relative humidity of 80% and a 12:12 light: dark photoperiod. These temperature conditions were selected because the larval development of Ae. japonicus is known to positively correlate with increasing temperatures up to 30°C. As the pupation limit is reached at 28°C, we chose 26°C as the incubation temperature for successful and rapid development of the mosquitoes 14,50 . Species identification was performed using the morphological key in the "Guidelines for the surveillance of invasive mosquitoes in Europe" 51 . To exclude natural flavivirus infections that could potentially interfere with the experimental outcome, 10 randomly selected adult specimens were tested with pan-Flavi-, pan-Bunya-, and pan-Alphavirus PCRs, but these tests were negative [52][53][54] .
Groups of 20 females (4-14 days old) were placed in plastic vials, starved for 24 h, and challenged with infectious blood meals. The feeding, incubation, and analysis of the mosquitoes were performed in the BSL-3 insectary in the Bernhard Nocht Institute for Tropical Medicine, Hamburg, Germany. To support a high feeding rate, we provided infectious blood in 50 µl drops at the bottoms of the vials (two drops per vial). Thus, a feeding rate of 75% (i.e., the percentage of engorged females to total females) was reached.
For validation of the salivation assay for Ae. japonicus, an infection experiment with JEV was first performed. Female mosquitoes were infected via an infectious blood meal using the SA-14 strain of JEV (GenBank accession number EU073992) 55 at a final concentration of 10 7 plaque-forming units/milliliter (PFU/ml) and were kept at 27°C for 14 days.
Subsequently, a total of 381 female mosquitoes were challenged with blood meals containing ZIKV, strain ZIKV_FB-GWUH-2016 (GenBank accession number KU870645, fifth passage) 56 at a final concentration of 10 7 PFU/ml. Two hundred forty-three engorged females were incubated at 80% humidity and temperatures of 21, 24, or 27°C.

Assessment of ZIKV infection, dissemination, and transmission
Fourteen days post infection, mosquitoes were analyzed for JEV (n = 27) or ZIKV (n = 79) infection, dissemination, and transmission. Infection, dissemination, and virus titers were determined by separate analyses of mosquito bodies and heads without legs and wings (infection and body titer) and of legs (dissemination and leg titer) for the presence of viral ZIKV RNA using a quantitative real-time PCR assay (qRT-PCR; Real Star Zika Virus RT-PCR Kit, Altona Diagnostics, Hamburg, Germany). ZIKV transmission was assessed by testing mosquito saliva for the presence of infectious virus particles using the salivation assay as previously described 6 . In short, mosquitoes were immobilized and the probosces were placed into a filter tip containing 10 µl of phosphate-buffered saline (PBS). After 30 min, saliva-containing PBS was pipetted into the media of Vero cells seeded in a 96-well plate to measure the cytopathic effect, i.e., the presence of infectious virus particles, after 7 days. The presence of ZIKV in the supernatant of cytopathic cells was subsequently tested by the abovementioned qRT-PCR assay.

Statistical analysis
Calculations of the IR, dissemination rate (DR), and TR were performed as described by Fortuna et al. 57 . The IR is defined as the number of virus-positive mosquito bodies per number of fed females, the DR is defined as the number of virus-positive legs per number of virus-positive bodies, and the TR is defined as the number of viruspositive saliva samples per number of virus-positive bodies. Calculation of the transmission efficiency was conducted as described by Chouin-Carneiro et al. 35 and is defined as the number of virus-positive saliva samples per total number of fed females. The R program 58 was used for all calculations and visualizations, including the ggplot2 59 , tidyr 60 , and plyr 61 packages.

Data availability
All relevant data are provided within the paper.