Diversity of West Nile and Usutu virus strains in mosquitoes at an international airport in Austria

Abstract Increased globalization and international transportation have resulted in the inadvertent introduction of exotic mosquitoes and new mosquito‐borne diseases. International airports are among the possible points of entry for mosquitoes and their pathogens. We established a mosquito and mosquito‐borne diseases monitoring programme at the largest international airport in Austria and report the results for the first two years, 2018 and 2019. This included weekly monitoring and sampling of adult mosquitoes, and screening them for the presence of viral nucleic acids by standard molecular diagnostic techniques. Additionally, we surveyed the avian community at the airport, as birds are potentially amplifying hosts. In 2018, West Nile virus (WNV) was detected in 14 pools and Usutu virus (USUV) was detected in another 14 pools of mosquitoes (minimum infection rate [MIR] of 6.8 for each virus). Of these 28 pools, 26 consisted of female Culex pipiens/torrentium, and two contained male Culex sp. mosquitoes. Cx. pipiens/torrentium mosquitoes were the most frequently captured mosquito species at the airport. The detected WNV strains belonged to five sub‐clusters within the sub‐lineage 2d‐1, and all detected USUV strains were grouped to at least seven sub‐clusters among the cluster Europe 2; all strains were previously shown to be endemic in Austria. In 2019, all mosquito pools were negative for any viral nucleic acids tested. Our study suggests that airports may serve as foci of arbovirus activity, particularly during epidemic years, and should be considered when designing mosquito control and arbovirus monitoring programmes.

. Viruses originating from tropical regions in sub-Saharan Africa, like West Nile virus (WNV; Marcantonio et al., 2015), Usutu virus (USUV; Vilibic-Cavlek et al., 2020), or Chikungunya virus (CHIKV; Amraoui & Failloux, 2016), are meanwhile increasingly occurring in Europe. Some of those pathogens can only spread because the appropriate mosquito vector was introduced previously. For example, the spread of invasive Asian tiger mosquitoes (Aedes albopictus) in Europe has been associated with outbreaks of Dengue virus (DENV) in France and Croatia in 2010, and CHIKV in Italy in 2007and France in 2010(Medlock et al., 2012. Other newly emerging pathogens can be transmitted by native mosquitoes. WNV and USUV, which are now well established and widely spread across Europe, were likely introduced by migratory birds and are readily transmitted by competent native mosquito vectors of the Culex pipiens complex (Brugman et al., 2018;Hubálek, 2008).
Possible points of entry for mosquitoes and their pathogens are airports. Although the importation of exotic mosquitoes via air travel seems to be a rare event, importations have been repeatedly reported in recent years (Ibañez-Justicia et al., 2017;Ibáñez-Justicia et al., 2020;Scholte et al., 2014). Apparently, an increasing volume of air transportations elevates the risk of accidental introduction of exotic mosquito species. It is therefore recommended by the World Health Organization (WHO) and the European Centre for Disease Prevention and Control (ECDC) to carry out mosquito monitoring programmes at airports (ECDC, 2012;WHO, 2016).
There are two main ways in which mosquito-borne diseases can be introduced via air travel. First, infected mosquitoes could be transported via the aircraft. For example, malaria cases have been documented in and near international airports among persons who have not recently travelled to areas where the disease was endemic (Gratz et al., 2000). These 'airport malaria' cases are caused by malaria-infected mosquitoes travelling by aircraft from a country where malaria is endemic to a country in which malaria is usually not found, where the mosquito then bites a person in or nearby the airport and transmits the malaria parasite. Those are, however, rare events occurring mainly at airports with high connection frequencies to the main endemic areas in sub-Saharan Africa (Guillet et al., 1998). This transmission route is of course not limited to malaria, as other mosquito-borne diseases seem to spread accordingly (e.g. 'airport Dengue'; Whelan et al., 2012).
Airplane disinsection is an important method to prevent or at least reduce the risk of this transmission route (Gratz et al., 2000). The second major pathway of introducing novel mosquito-borne diseases via aircraft is the transportation of already infected persons (Feng et al., 2019). Upon arriving at their destination, infected passengers may be bitten by native mosquitoes that further spread the pathogens, provided these mosquitoes are competent vectors (Quam et al., 2015).
Autochthonous transmission of arboviruses originating from infected travellers appears to be more frequent, with several examples of DENV infection (Franco et al., 2015;Kutsuna et al., 2015;Marchand et al., 2013) and CHIKV infection (Calba et al., 2017;Delisle et al., 2015) arising across Europe in patients with no travel history.
In this study, we present the results of a mosquito-monitoring programme at a large international airport in Austria. Our monitoring scheme was designed to detect the importation of non-native mosquito species and non-native arboviruses. As some zoonotic arboviruses use birds as amplifying hosts, we additionally surveyed the avian community at the airport and tested birds killed by plane strikes for the presence of arbovirus nucleic acids. The goal was to implement a monitoring programme, in addition to the ongoing national and regional monitoring programmes, and to assess the utility of such a programme to monitor the mosquito population and arbovirus activity with a specific focus on detecting exotic mosquitoes and/or viruses. When the courtyard was no longer used by the employees, the fish were removed; however, the water remained in the pond until the end of July 2019, when the pond was filled up with soil. Moreover, the vegetation at this site was no longer maintained on a regular basis in 2019.

Mosquito trapping and identification
Weekly sampling of adult mosquitoes was performed using a BG-Sentinel 2 mosquito trap (Biogents AG, Regensburg, Germany) equipped with an additional CO 2 release and a specific lure (BG-Sweetscent Abbreviations: F, forward primer; R, reverse primer; P, probe. a Probes were labelled at the 5′-end with 6-carboxyfluorescein (FAM) and the 3′-end with tetramethyl-6-carboxyrhodamine (TAMRA). b Sane et al. (2012); modified nucleotide in bold.  (Patel et al., 2013). Although this assay was designed as RT-qPCR, it was performed as a conventional RT-PCR (without probe) using

Testing for flaviviruses
OneStep RT-PCR Kit (Qiagen, Hilden, Germany). To determine the identities of the detected viruses and their genetic variants, PCR products (amplicon lengths dependent on the specific virus, app. 260 bp) were subsequently subjected to Sanger sequencing (Eurofins Genomics, Ebersberg, Germany).
To estimate the proportion of infected mosquitoes we calculated the MIR for unequal pool sizes as described previously (Biggerstaff, 2008).
The MIR specifies the ratio of the number of positive pools to the total number of mosquitoes tested per 1000 individuals. Data analysis was conducted in R (R Core Team, 2020).

Phylogenetic analysis
In order to compare WNV and USUV sequences to those previously published from Austria, we used an RT-PCR assay that is universal for the Japanese encephalitis virus complex (Weissenböck et al., 2002), All WNV and USUV sequences generated in this study were uploaded to the NCBI database with the following accession numbers: MW160840-MW160849 and MW160850-MW160861, respectively.

Other molecular investigations
All mosquito extracts were additionally tested by virus-specific RT-qPCRs for two endemic orthobunyaviruses (family Peribunyaviridae): Batai orthobunyavirus (Calovo virus, CVOV) and Tahyna  While CHIKV RT-qPCR (primer set 874F/961R/899P), was performed exactly as described previously (Lanciotti et al., 2007), the published assay for the detection of SINV via RT-qPCR (Sane et al., 2012) was slightly modified. All in-house and published but modified primers and probes, their sequences, positions on the corresponding genes, sizes of the amplicons, as well as reference molecules are listed in Table 1. Each RT-qPCR based on TaqMan® chemistry was optimized for Quanta qScript XLT One-Step RT-qPCR ToughMix Kit (Quantabio, Beverly, MA, USA) with primer and probe concentrations of 0.5 µM each.

Blood meal analysis
In 2019, vertebrate hosts were identified from individual bloodengorged female mosquitoes encountered in the questing traps as described previously (Camp et al., 2019). Briefly, DNA was extracted from homogenized abdomens using a commercial kit (DNeasy, Qiagen, Hilden, Germany) and subjected to PCR to amplify a region of the mitochondrial gene 16S rRNA (Kitano et al., 2007) or cytochrome b (Cupp et al., 2004). The respective primers target vertebrate sequences but do not efficiently amplify invertebrate sequences of these gene regions. The resulting amplicons were sequenced by the Sanger method and compared to in-house voucher sequences, or to publicly available sequences using the online version of BLASTn (Altschul et al., 1990).

Data collection of birds
In 2018, bird monitoring was conducted by driving (walking speed) along a road that encircles the airport (16 km), and counting the number and determining the species of the observed birds. In 2019, the scheme was changed to a 'point stop' method. At 19 locations around the airport's border, the number and the species of the encountered birds within a time period of five minutes were recorded. Bird counts were conducted twice each month, and the numbers averaged for each month.
In 2019, fresh bird carcasses from airplane strikes were collected and analysed for flaviviruses. The carcasses were frozen at −20 • C for up to 1 month and thawed overnight at 4 • C before dissection.
Pooled organ sections (heart, liver, spleen, lung, and leg muscle) were homogenized in Dulbecco's phosphate-buffered saline (DPBS, Gibco, Dublin, Ireland) using metal beads on a bead mill (TissueLyser II, Qiagen, Hilden, Germany), and RNA was extracted from the homogenate using a commercial kit (QIAamp Viral RNA Mini Kit, Qiagen, Hilden, Germany). The RNA was stored at −80 • C until testing for flavivirus nucleic acid as described above.

Mosquito species
In total, 4850 mosquitoes were collected (
( Table 3)  In addition to the sequence similarity analyses, the phylogenetic analysis confirmed the high diversity of WNV and USUV strains detected in this study. The WNV strains detected in this study belong to five sub-clusters within the WNV sub-lineage 2d-1 (Figure 2).

Other viral investigations
All mosquito pools collected in 2018 and 2019 tested negative for CVOV/BATV, CHIKV, TAHV and SINV by their corresponding RT-qPCRs. During the bird monitoring at the airport, 11 species were recorded ( Figure 4); the most common were crows (Corvus corone) and kestrels (Falco tinnunculus). Eurasian magpies (Pica pica), the most common host for the identified Cx. pipiens/torrentium blood meals, accounted only for 1.5-8.5% of the bird counts from May to October. Flaviviral nucleic acids were not detected in any of the tested bird tissues taken from carcasses following airplane strikes.

DISCUSSION
During two years of monitoring at the airport, we detected an unusually high number of WNV-and USUV-positive mosquito pools only in the first year, 2018. Although we detected only endemic WNV and USUV strains, the partial genomic sequence of each mosquito poolderived virus was unique. We collected no exotic mosquito species, except a single specimen of Ae. japonicus. The observed mosquito population was comprised largely of Cx. pipiens/torrentium, which is common in heavily built-up areas such as airports (Ibañez-Justicia et al., 2017) or other urban environments in central Europe (Krüger et al., 2014;Lebl et al., 2015) and North America (Pecoraro et al., 2007;Trawinski & MacKay, 2010). A closer examination of sampled individuals showed that they belonged to Cx. pipiens f. pipiens, which seems to be the predominant variant of this species complex in eastern Austria (Zittra et al., 2016). Culex pipiens mosquitoes are known to be competent vectors of several arboviruses, including those endemic to Europe (WNV and USUV), as well as St. Louis encephalitis virus, SINV, and Rift Valley fever virus (Turell, 2012). Otherwise, the mosquito community at the airport was similar to other mosquito surveys in nearby urban environments in Austria (Lebl et al., 2015).
Although both WNV and USUV have been established in Austrian mosquito populations for many years Weissenböck et al., 2003;Wodak et al., 2011), only targeted vector surveys at sites of confirmed human, horse, or bird WNV cases ( The results of the blood meal analysis, with a high proportion of Cx. pipiens/torrentium females feeding on birds, are in line with previous studies showing that Cx. pipiens mosquitoes are ornithophilic, but occasionally feed on mammals, reptiles, or amphibians as well Radrova et al., 2013;Rizzoli et al., 2015;Roiz et al., 2012). Although they were encountered rarely during the bird counting events, Eurasian magpies (Pica pica) were the preferred hosts of Cx. pipiens/torrentium females based on the blood meal analysis. These results match those of Rizzoli et al. (2015), who demonstrated that, in Europe, magpies and blackbirds (Turdus merula; not recorded at the airport) were significantly preferred by Cx. pipiens. Although we did not calculate a preference index, magpies were not the most abundant birds at the airport, as we encountered carrion/hooded crows (Corvus corone ssp.) and common kestrels (Falco tinnunculus) more frequently ( Figure 4). However, only one of 18 blood meals from Cx. pipiens/torrentium was identified as a common kestrel. As magpies have been shown to be highly susceptible to WNV infections, they could thus be an important factor in the transmission cycle of this virus (Jiménez de Oya et al., 2018;Napp et al., 2019). Of note, it is possible that the high proportion of blood from magpies relative to their abundance represents a sampling bias of our study, as we found a magpie nest in a courtyard near where the mosquito sampling was performed.
Although Cx. pipiens/torrentium mosquitoes clearly prefer avian hosts, flaviviruses were not detected in any of the bird carcasses collected in this area during the study period. This could be due to a sampling bias, as no magpies were among the tested carcasses, which also reflects their infrequent encounters during the ornithological survey.
Other bird species observed during the study, from which carcasses were tested for virus, are known to support relatively high viremia following WNV infection, particularly the corvids and birds of prey (Komar et al., 2003;Work et al., 1955). However, the unexpectedly high infection rate of mosquitoes trapped at the international airport in 2018 is in accordance with the reported, relatively high numbers of WNV and/or USUV infections in humans , horses (de Heus et al., 2020), and birds (Weidinger et al., 2020) in Austria in

2018.
In Europe, 2018 has been widely recognized as a year with a recordbreaking incidence of WNV cases in both, humans and horses . In Austria, the number of human WNV cases in 2018 was 4-to 5-fold higher than in previous and following years (2018, n = 20; 2016 and 2017, n = 5 each; 2019 n = 4). All but two cases occurred in the City of Vienna and the directly adjacent regions (ECDC, 2018(ECDC, , 2019. It is therefore not surprising that we also report a relatively high MIR for WNV in 2018 at our study site. In general, a MIR >1 seems to be associated with an increased incidence of human WNV cases, compared to longitudinal surveys with similar sampling methods in the neighbouring country of Italy (Calzolari et al., 2015(Calzolari et al., , 2020Cerutti et al., 2012), although strict comparisons of MIR between surveillance programmes may not be reliable (Chakraborty & Smith, 2019). Moreover, we note that the incidence of WNV in humans in Austria is much lower than in Italy, in both epidemic and non-epidemic years ECDC, 2018ECDC, , 2019. Increased WNV activity in 2018 was likely due to specific environmental conditions that promoted mosquito abundance and increased the likelihood of earlyseason encounters between infected mosquitoes and suitable competent vertebrate reservoirs . It remains to be seen why WNV cases in Europe returned to baseline in 2019, given the similar meteorological conditions. We did not record meteorological variables in this study, but we monitored both the mosquito and avian communities. In our analysis, we averaged bird counts at the airport from 2018 and 2019 (Figure 4), as there was no difference in a given species' abundance between the years. As has been noted by others (reviewed in , the observed decline of infections in 2019 may be due to increased flock immunity of avian amplifying hosts, acquired during the 2018 epizootic/epidemic year. However, we did not test the serological status of the bird carcasses that we investigated. Because USUV utilizes similar competent mosquito vectors as WNV (primarily Cx. pipiens; Fros et al., 2015), and probably similar vertebrate amplifying hosts, with blackbirds being notable for their high mortality rate associated with USUV infection Weidinger et al., 2020;Weissenböck et al., 2002), it is not surprising that we detected the same high MIR for USUV as for WNV in 2018, but found no infections in 2019.
The relatively high prevalence of WNV and USUV in 2018 was similarly experienced throughout Europe, and was likely due to environmental factors which led to increased mosquito abundance and increased transmission efficiency in the early season   .
The mosquito-derived USUV strains identified here were distributed widely over the phylogenetic tree within the cluster Europe 2 ( Figure 3). They are highly related to Austrian human sequences from 2017 (Bakonyi et al., 2017), as well as to Austrian and Hungarian bird sequences from 2017 and 2018 (Weidinger et al., 2020). Sequencing of Austrian human USUV strains detected in 2018  was performed in another genomic region; therefore, a direct phylogenetic comparison with the current mosquito strains was not possible. However, upon closer examination of joint sequences, at least three mosquito strains would most probably also cluster together with the majority of USUV strains generated from Austrian blood donors in 2018. Others have noted the comparatively low genetic variability of USUV within the established genetic lineages circulating across Europe, particularly since the spread of the Europe 2 lineage beginning in 2015/2016 (Bakonyi et al., 2017;Weidinger et al., 2020), and even over broader geographic areas (Nikolay et al., 2013). Thus, our data represent an interesting aspect of the co-circulation of USUV and WNV that has yet to be addressed. Namely, albeit the arthropod vector population and host availability are presumably the same for both viruses, we observed a much higher mutation rate for WNV than for USUV. There are likely many factors that may contribute to this, which should be investigated more thoroughly in future studies.
The uniqueness of the WNV and USUV strains we identified is additionally underlined by deletions of different lengths within the 3′UTRs in two WNV and two USUV strains (Figures 2 and 3

CONCLUSIONS
We report the establishment of a mosquito monitoring programme at a large international airport in Central Europe, to augment the national surveillance programme. The first year of monitoring coincided with the largest outbreak of WNV in Europe on record, and our observations reflect the high transmission activity of both WNV and USUV in 2018.
Combining bird surveys and mosquito blood meal analyses to identify hosts provided additional information about the transmission patterns of WNV and USUV at the airport, both of which likely involve certain avian hosts (especially the Eurasian magpie, Pica pica) and mosquito species (Cx. pipiens/torrentium) for virus maintenance, amplification and spillover. Therefore, mosquito control efforts at airports should consider targeting these species specifically.