Co- infections: Simultaneous detections of West Nile virus and Usutu virus in birds from Germany

The emergence of West Nile virus (WNV) and Usutu virus (USUV) in Europe resulted in significant outbreaks leading to avifauna mortality and human infections. Both viruses have overlapping geographical, host and vector ranges, and are often co- circulating in Europe. In Germany, a nationwide bird surveillance network was established to monitor these zoonotic arthropod- borne viruses in migratory and resident birds. In this frame-work, co- infections with WNV and USUV were detected in six dead birds collected in 2018 and 2019. Genomic sequencing and phylogenetic analyses classified the detected WNV strains as lineage 2 and the USUV strains as lineages Africa 2 ( n = 2), Africa 3 ( n = 3) and Europe 2 ( n = 1). Preliminary attempts to co- propagate both viruses in vitro failed. However, we successfully cultivated WNV from two animals. Further evidence for WNV-USUV co- infection was obtained by sampling live birds in four zoological gardens with confirmed WNV cases. Three snowy owls had high neutralizing antibody titres against both WNV and USUV, of which two were also positive for USUV- RNA. In conclusion, further reports of co- infections in animals as well as in humans are expected in the future, particularly in areas where both viruses are present in the vector population.


| INTRODUC TI ON
Mosquito-borne viruses represent a growing threat to Europe and its avifauna . In this context, West Nile virus (WNV) and Usutu virus (USUV) have to be considered as two of the more than seventy relevant members of the family Flaviviridae (Calisher & Gould, 2003). These two flaviviruses were historically regarded as viruses of purely African significance, with no evidence of associated bird or human mortality. It was only when these viruses entered Europe and America (for WNV-only) that waves of infection became visible in both birds and mammals (e.g. humans and horses) (Zeller & Schuffenecker, 2004). Although USUV infections in humans are commonly asymptomatic, recent outbreaks in Europe had reported neuroinvasive cases with encephalitis and meningoencephalitis, in patients from Italy and Croatia as summarized by Clé et al., (2019).
USUV and WNV are classified as arthropod-borne viruses (arboviruses) that are primarily transmitted by ornithophilic mosquitoes to birds and mammals. Both viruses are preserved in the environment through a vertebrate host-mosquito life cycle, where different bird species act as amplifying hosts, Culex mosquitoes as primary vectors and mammals, such as humans and horses, as dead-end hosts.
Viral transmission occurs when an infected mosquito takes a blood meal from a non-immune susceptible host. During this process, the virus present in the salivary glands of the mosquito is transferred to the avian host where it is replicated to high titre and appears in the bloodstream as a viraemia, thereby enabling further mosquito transmission cycles of the virus to other hosts (Chancey et al., 2015;Clé et al., 2019).
WNV and USUV mono-infections were reported in a wide range of avian species including owls, birds of prey, passerines, storks, flamingos and others (as summarized in Nikolay, 2015). Typically, infection in most bird species remains inapparent. However, bird species that are highly susceptible for WNV, such as owls, birds of prey and several passerines (e.g. jays, crows and sparrows), can develop a neurological disease which can be fatal (Komar, 2003;Pérez-Ramírez et al., 2014;Troupin & Colpitts, 2016). The same holds true for USUV, with owls and passerines like blackbirds being more susceptible to disease (Becker et al., 2012;Chvala et al., 2004).
The circulation of WNV and USUV is generally influenced by environmental factors that affect the population dynamics of mosquito vectors, the extrinsic incubation periods (the time needed for a mosquito to become infectious after ingestion of a virus) and the population densities of amplifying hosts (Durand et al., 2010;Rubel et al., 2008). Since both WNV and USUV have very similar prerequisites in these regards, it is no surprise that these two viruses cocirculate in Europe in at least 10 countries and that both viruses have been found to infect 34 common bird species belonging to 11 different orders (Nikolay, 2015). Hence, WNV and USUV transmission cycles overlap substantially in many countries (Nikolay, 2015).
The unusually hot climatic conditions all over Europe in the summer 2018, with an extremely long period of high temperatures, may have provided favourable conditions for the incursion and establishment of WNV and/or USUV into new areas and countries (Camp & Nowotny, 2020). In this context, a large WNV outbreak was observed in southeastern and southern Europe, with a total of 2,083 autochthonous human WNV cases in the European Union (EU) member states and EU neighbouring countries (European Centre for Disease Prevention and Control [ECDC], 2018). In 2018, WNV was detected for the first time in birds and horses in Germany .
While WNV was only recently introduced into Germany, USUV had circulated here for more than ten years. In 2010, USUV was first detected in mosquitoes in southwest Germany (Jöst et al., 2011).
One year later, dead common blackbirds (Turdus merula) were found frequently around the cities of Mannheim and Heidelberg (Baden-Württemberg) followed by a mass mortality among wild birds in southwest Germany (Becker et al., 2012). In the following years, USUV outbreaks stayed geographically restricted to the Upper Rhine Valley in southwest Germany, apart from a few sporadic cases in Berlin and Bonn (Cadar et al., 2015;Ziegler et al., 2015Ziegler et al., , 2016. In 2016, however, USUV case numbers increased dramatically not only in southwestern Germany, but also in the federal states of North Rhine-Westphalia (NRW), Saxony and Saxony-Anhalt Sieg et al., 2017). Since its occurrence in Germany, USUV had been responsible for the mortality of thousands of common blackbirds and many captive and free-living owls in the past . Currently, there are five USUV lineages circulating in the country, namely Africa 2, Africa 3, Europe 2, Europe 3 and Europe 5 .
The initial 2018 WNV outbreak located in the eastern regions of Germany was accompanied by a massive and ongoing USUV epizootic in all federal states of Germany  Naturschutzbund Deutschland e.V. [NABU], 2018) and elsewhere in Europe, such as Austria, Belgium, Croatia and the Netherlands (Benzarti et al., 2020;Oude Munnink et al., 2020;Vilibic-Cavlek et al., 2019;Weidinger et al., 2020). In 2019, an even larger WNV epizootic occurred in Germany with 76 confirmed cases in wild and zoo birds, 36 confirmed cases in horses and five clinical cases in humans (Ziegler et al., 2020). This was also accompanied by numerous USUV outbreaks in wild birds in many different areas in Germany.
With regard to the fact that WNV and USUV often occur in the same geographical regions and use an almost identical transmission cycle between mosquitoes and birds (Nikolay, 2015), it is surprising that co-infections with both viruses have so far only been described once in a human. In that case, both WNV-and USUV-RNA were detected in a blood donor from Austria (Aberle et al., 2018).
Surveillance of emerging arboviruses, such as WNV and USUV, is well suited in zoological gardens since these collection sites contain large diversity of captive species including mammals and birds, and are located within or near urban areas. Captive animals from zoological gardens can serve as important sentinels for newly introduced pathogens in an area. Collection animals within zoological gardens are also routinely monitored by veterinarians and technical staff with expertise in wildlife health; thus, gathering samples and detailed medical records can be readily available (Constant et al., 2020;Cox-Witton et al., 2014). For example, WNV introduction in the western hemisphere and Germany was first recognized in Bronx Zoo/Wildlife Conservation Park, New York City, and Zoo Halle (Saale), Saxony-Anhalt, respectively (Ludwig et al., 2002;.
Our study describes for the first time a co-infection with WNV and USUV in six dead birds in Germany. These co-infected birds were detected in the context of the WNV epizootic in 2018 and 2019 and followed by an extensive molecular and serological investigation of apparently healthy birds held in zoological gardens with confirmed cases of WNV infection.

| Molecular virus characterization
Viral RNA was extracted from the tissue material from the submitted bird samples and from the frozen (−70°C) coagulated blood of the bird samples (cruor) using RNeasy Kit (Qiagen) according to the manufacturer's instructions. Extracted RNA was analysed with reverse transcription quantitative real-time PCR (RT-qPCR) assays specific for WNV lineages 1 and 2 as described by Eiden et al. (2010) and USUV-specific RT-qPCR described by Jöst et al. (2011). Based on the guidelines of the National WNV Reference Laboratory, quantification cycle (C q ) values <37 were regarded as positive, from 37 to 40 as possible, and >40 as negative. In all RT-qPCR assays, positive controls with 10 3 and 10 4 WNV or USUV genome copies per reaction were included.
Full-genome sequencing was performed with WNV-and USUVpositive bird samples. The organ material from these birds was homogenized in 1 ml TRIzol™ LS Reagent (Invitrogen™) using the TissueLyser II (QIAGEN) with 5 mm steal beads for 2 min at 30 Hz.
After the phase separation step, RNA was extracted from the aqueous phase using the RNAdvance Tissue kit (Beckman Coulter) and the KingFisher Flex System (Thermo Fisher Scientific). The resulting RNA extracts were also tested with the RT-qPCR assays mentioned above.
Selected WNV-positive RNA was subjected to cDNA synthesis and library preparation as described in Wylezich et al. (2018). We performed targeted enrichment of virus sequences in sequencing libraries from four bird samples (Table 1) using custom myBaits ® target capture kit with VirBaits panel (Arbor Biosciences) based on the panel design and protocol described in Wylezich et al. (2020).
The USUV full-genome sequencing was implemented as described in This multiplex PCR was performed in two separate reactions using AccuPrime™ Taq DNA Polymerase, High Fidelity (Invitrogen™).
Amplicons were purified with 1.8 × volumes of Agencourt AMPure XP beads (Beckman Coulter) and quantified using NanoDrop™ spectrophotometer (Thermo Fisher Scientific). These two purified reactions per sample were pooled and adjusted to 500 ng. Fragmentation and library preparation steps were prepared as described in Wylezich et al. (2018). Quantified libraries (GeneRead DNA Library L Core Kit; QIAGEN) were sequenced using an Ion Torrent S5 XL instrument with Ion 530 or Ion 540 chips and the respective reagents (Thermo Fisher Scientific) in 400 bp mode or 200 bp mode, respectively.
After the analysis of USUV sequences, additional library preparation and sequencing were performed to increase the coverage of specific genome positions. Amplification and preparation steps were similar to procedures mentioned above except for primer pairs included in each multiplex PCR mix (library preparation #2 in Table S1). In order to close the gaps within three USUV genome sequences, selected cDNA was amplified using necessary primer pairs for single-plex PCRs (Table S1) and sequenced with a BigDye Terminator v1.1 Cycle Sequencing kit (Applied Biosystems™, Thermo Fisher Scientific) on a 3500 Genetic Analyzer instrument (Applied Biosystems™, Thermo Fisher Scientific).

| Genome characterization of WNV and USUV, and phylogenetic analyses
Sequencing adapters and primers were trimmed using Newbler assembler of the Genome Sequencer Software Suite v. 3.0 (Roche), and sequencing reads were quality controlled using FastQC. Initial  were submitted to European Nucleotide Archive under the BioProject accession number PRJEB41417 (Table 1). All whole-genome sequences of European WNV lineage 2 and USUV were retrieved from GenBank on 01 September 2020 (Tables S2 and S3, respectively). The retrieved USUV and WNV sequences together with the newly obtained sequences were aligned separately per species using MUSCLE (Edgar, 2004). These alignments were visually inspected with Geneious Prime ® 2019.2.3 Software (Biomatters), and genomes with >10% 'N' sequences or gaps were not included in the phylogenetic analyses.
Best fit nucleotide substitution model identified as GTR+I+G4 for both WNV and USUV complete genome sequence data sets was cal- bio.ed.ac.uk/softw are/figtr ee/). Partial USUV envelope sequences were also assessed using phylogenetic analyses mentioned above.
These USUV sequences were retrieved in GenBank on 09 September 2020 (accession nos. indicated in Figures S1 and S2).

| Serological investigation
We analysed all serum samples collected from birds held in four zoological gardens described in Section 2.1 using two specific virus neutralization tests (VNTs), as described by Seidowski et al. (2010) with minor modifications. Briefly, we used a German USUV strain   (Mayr et al., 1977). Serum samples with ND 50 values above 10 were considered as positive; otherwise, samples were regarded as negative. Birds were only regarded positive for WNV if they had a negative (ND 50 < 10) or significantly lower (fourfold lower) USUV titre. The same criteria were also implemented for interpreting USUV VNT results. If a bird had similar antibody titres against both viruses, it is difficult to discriminate between WNVand USUV-specific neutralizing antibodies due to cross-reactivities, and the result must therefore often be interpreted as inconclusive.
However, in cases where the neutralizing titres are very high against both viruses and a high infection pressure of both viruses in a very dense spatial area with many susceptible birds at the same time is obvious, the results can be interpreted with higher certainty as the consequence of infections with both viruses in those birds.

| RE SULTS
Since the first report of USUV circulation within bird populations in Germany, a nationwide bird surveillance network was established to systematically monitor zoonotic arboviruses, such as WNV and USUV, in migratory and resident birds using molecular and serological diagnostics tools . Among 88 WNV-infected birds reported in the 2018-2019 surveillance programme in Germany, our study detected six deceased birds (cases #1-#6) positive with both WNV and USUV genomes (Ziegler et al., , 2020.

| Results of RT-PCR screening, sequencing and virus phylogeny
Five of the deceased co-infected birds (cases #1-#5) were collected in 'Tierpark Berlin' in 2018 and 2019, while the sixth co-infected bird (case #6) was detected in Dresden in 2019 ( Figure 1, Table 1). These birds with co-infection belong to the taxonomic orders Strigiformes, Charadriiformes, Anseriformes and Passeriformes (Table 1). The presence of both WNV-and USUV-RNA within individual organs was observed in four birds (Table 1). On the other hand, co-infection at the organ level was not confirmed in case #6 since the liver and heart tissue samples were pooled, and was not observed in case #3.
For USUV screening, organ samples with C q values in the range of 37-40, which were defined as 'possible' USUV-positive, were confirmed either with another USUV-specific RT-qPCR assay or by sequencing (Table 1). In all co-infected organ samples, the RNA loads were higher for WNV than for USUV. Organs that were only USUVpositive were, for example, the brain from a snowy owl (Bubo scandiacus, case #1) and both the kidney and the liver from a Chinese merganser (Mergus squamatus, case #3), while a few organ samples from cases #2-#5 were WNV-positive only (Table 1).
WNV and USUV genomic sequencing were implemented on selected organ samples of six co-infected birds to confirm the results of the RT-qPCR and determine their respective virus lineages (Table 1). Two complete (cases #4, #5) and two partial (cases #1, #3) WNV genomes were assembled from this study, while two WNV full genomes (cases #2, #6) were sequenced from our previous study (Ziegler et al., 2020) (

| Flavivirus monitoring in live captive birds from different zoological gardens
The investigation of WNV-USUV co-infection was continued in captured live birds from four zoological gardens (areas A to D) with confirmed WNV cases using molecular and serological diagnostic assays. Molecular screening revealed that the whole blood samples from 67 birds from zoological gardens located in Berlin (area A, Table 2), Saxony-Anhalt, Brandenburg and Bavaria (areas B to C, Table 3) were WNV-RNA-negative. Likewise, whole blood samples from birds located in areas B to D were all negative of USUV-RNA (Table 3). In contrast, USUV-RNA was detected in two snowy owls from Berlin sampled in 2019 (Table 2) (Table 2). These parent owls were confined in the same aviary with their two offspring (cases #1, #2) in 2018, which had confirmed WNV and USUV double infection (Table 1) (Table 3). Among these samples, the taxonomic classification of one bird was not possible. Two Eurasian eagle owls from area C ('Wildpark Poing') had USUV-specific neutralizing antibodies, and serological evidence of WNV was not detected at all. Therefore, the evidence of WNV-USUV co-circulation could be especially detected in the birds from areas B and D (Table 3).

| D ISCUSS I ON
Although co-circulation of WNV and USUV in the same region has previously been described, co-infection with both viruses in birds has not been reported before. So far, only a single case of WNV and USUV co-infection was reported in humans (Aberle et al., 2018). reported WNV-USUV co-circulation.
The first confirmed case of a WNV-USUV double infection was described from an asymptomatic blood donor residing in Austria (Aberle et al., 2018). The corresponding blood sample collected in August 2018 was tested positive for both WNV and USUV using RT-qPCR and VNT, and genomic analyses classified these F I G U R E 3 Phylogenetic analysis of Usutu virus complete genomes. USUV sequences related to WNV-USUV co-infected birds are highlighted in red and with red dots, while other German USUV sequences are highlighted in green. Taxon information includes the nucleotide accession number, collection year and countries of origin of the samples. The scale bar indicates the number of nucleotide substitutions per site. Numbers before the nodes denote bootstrap values ≥80%. The maximum likelihood tree was constructed using the best-selected nucleotide substitution model (GTR+I+G4). The branch of Europe 5, second cluster of Africa 2, clusters Africa 3.1, Africa 3.2 and several Africa 3.3 strains from the Netherlands, several Italian USUV strains from lineage Europe 2, lineages Europe 1, Europe 4, and Europe 3 were collapsed into triangles, and the numbers in parentheses indicate the number of collapsed USUV strains. USUV sequences from cases #1 to #6 were submitted to European Nucleotide Archive under the project accession: PRJEB41417. Usutu virus sequences retrieved from GenBank are described in Table S3 flaviviruses under WNV lineage 2 and USUV lineage Europe 2 (Aberle et al., 2018). In an earlier study, Tamba (Wang et al., 2020). It is possible that the WNV suppressed the productive replication of USUV within cells, which is a process known as superinfection exclusion (Salaman, 1933;Zou et al., 2009). The superinfection exclusion is considered as a protection strategy of the 'primary virus' from the competing related 'secondary virus' within the same host (Wang et al., 2020). This strategy has been observed in different flaviviruses, such as dengue virus, Japanese encephalitis virus, Nhumirim virus and Culex flavivirus, in mosquito cells (Goenaga et al., 2020;Kanthong et al., 2010;Kenney et al., 2014;Pepin et al., 2008), and other members of the family Flaviviridae, such as

Hepatitis C virus and bovine viral diarrhoea virus in vertebrate cell
lines (Lee et al., 2005;Tscherne et al., 2007).
Our study detected USUV belonging to lineages Africa 2, Africa 3 and Europe 2. USUV sequences from two snowy owls (cases #1, #2) collected in 'Tierpark Berlin' in 2018 were identified as lineage TA B L E 2 Detection of WNV-and USUV-RNA using reverse transcription quantitative real-time PCR (RT-qPCR) and specific antibodies by virus neutralization tests (VNTs) in the area A ('Tierpark Berlin', Berlin). Positive serology results are highlighted in red and bold, and crossreacting antibody titres are also displayed in black Note: Possible USUV-positive samples based on C q value range 37-40 were indicated in italics.

Columbiforms Fantail Pigeons
Columba livia forma dom. European Clade (CEC), and the majority of these strains clustered together to form the Eastern German Clade (EGC) as described in our previous study (Ziegler et al., 2020). Four complete and two partial WNV genomes acquired from birds with co-infections were classified as WNV lineage 2 EGC. This finding was not surprising since cases #1 to #5 were confined in Berlin, which was one of the hotspots for cases infected with WNV lineage 2 EGC in 2018-2019 (Ziegler et al., 2020). In 2020, Berlin remained a WNV hotspot in Germany with again numerous bird cases and occasional reports in both horses and humans (Friedrich-Loeffler-Institut [FLI], 2020).
Specific neutralizing antibodies against USUV had been verified in the German avifauna since 2012 Ziegler et al., 2015), while the occurrence of neutralizing antibodies against WNV in German resident birds was not reported before 2018. The detection of WNV-specific antibodies in live birds corresponds to the first WNV-RNA detection in dead birds .

TA B L E 3 (Continued)
We reported also the presence of similarly high titres of neutralizing antibodies against WNV and USUV in three snowy owls located in areas A and B, while USUV-RNA was only detected in two snowy owls from area A. The detection of both WNV-and USUV-specific antibodies exclusively in snowy owls is debatable since the sample set was rather small. Mono-infection of WNV or USUV can cause fatal diseases in different bird species. As an example, the first confirmed case of WNV infection in Germany was reported in area B.
This infection resulted in the death of a great grey owl (Strix nebulosa, FLI sample code: ED-I-33/18, accession no. MH924836 ). Its partner owl also succumbed to acute viral infection caused by USUV (U. Ziegler and M. Keller, unpublished data).
The observed high titres of neutralizing antibodies against both WNV and USUV could be a rare event due to sequential or simultaneous infections as a consequence of a high infection pressure for both flaviviruses. However, as described in previous surveillance studies , some level of serological crossreactions cannot be ruled out since multiple flaviviruses including WNV and USUV share antigenic characteristics with their structural outer proteins (e.g. envelope protein) (Blázquez et al., 2015).
However, the identification of three resident birds with similar and very high titres of neutralizing antibodies against both flaviviruses in zoological gardens with confirmed WNV-USUV co-circulation is more indicative of simultaneous or sequential infections rather than cross-reactions. These data are in line with the observation that two juvenile snowy owls from area A had double infection with WNV and USUV, which was detected by highly specific molecular diagnostics (see Table 1). Therefore, infection with WNV and USUV in the same bird is a possibility that occurs but has so far only been detected in isolated cases either by high titre serological evidence or by specific molecular diagnostics.
USUV circulation in Berlin, Saxony-Anhalt and Saxony was reported at least a year before the introduction of WNV in those regions Sieg et al., 2017). Hence, there was the possibility that the surviving snowy owls acquired a protective USUV infection prior to the WNV infection. This could explain why the parent snowy owls (Table 2)  whereas their offspring died a few months later in 2018 (cases #1, #2, Table 1) and 2019 (accession nos. LR743424 and LR743428) (Ziegler et al., 2020). This hypothesis is also supported by the study of Blázquez et al. (2015), who reported that USUV pre-infected mice challenged with WNV infection were protected against WNV disease and death but not against infection since WNV-RNA was still detected in 50% of WNV-challenged mice 7 days post-infection.
Hence, USUV is a potential low pathogenic flaviviral vaccine candidate against WNV since it elicits cross-protective immunity against the heterologous neurovirulent WNV (Blázquez et al., 2015).
It was hypothesized previously that an increased risk of a higher virulence and mortality rate of a secondary infection by a closely related virus could be, for example, explained by the so-called 'antibody-dependent enhancement' (ADE) (Porterfield, 1986). This, however, was not detected in the studies of Percivalle et al. (2020) and Sinigaglia et al. (2019). Results reported by Sinigaglia et al. (2019) implied that prior USUV infections followed by WNV infections did not lead to ADE in humans, likewise in the cases described by Percivalle et al. (2020). There was no clear evidence that a secondary WNV or USUV infection in humans, which were previously infected with either WNV or USUV, caused more severe disease or even death (Percivalle et al. 2020). Hence, we cannot rule out the possibility that a prior WNV infection generated cross-protective immunity against USUV infection in these snowy owls.
Based on these studies, simultaneous infections as well as se-  (Wang et al., 2020). Moreover, the observed superinfection exclusion in WNV and USUV co-infected cells and mosquitoes lowers the probability of WNV and USUV recombination.

| CON CLUS IONS
WNV and USUV are important arboviruses in the world in the context of public and animal health. Mono-infection as well as co-circulation of WNV and USUV had been previously described in mosquitoes, birds, horses and humans in several countries. We present here the first extensive surveillance of WNV-USUV co-infections in deceased and live birds from 2018 to 2019. The co-infections in five birds from zoological gardens and in one wild bird were detected by RT-qPCR and characterized by genomic sequencing. Flavivirus double infections in three birds from zoological gardens were verified using virus neutralization tests.
Taken together, the use of two virus-specific PCR systems has proven to be successful in the surveillance of wild and captive avian species for WNV and USUV. This study showed that the WNV and USUV RT-qPCR assays used in the surveillance network are highly specific in distinguishing these two flaviviruses. Our study also emphasized the importance of zoological birds, which are resident birds, as a tool for developing an early warning system to detect an introduction and circulation of zoonotic pathogens, especially within cities and urban settings. In the future, further co-infections can be expected in animals as well as humans, in areas with WNV and USUV co-circulation. Therefore, the consequences of co-infections for public health must also be taken into consideration.

E TH I C A L A PPROVA L
Blood samples for this study were taken based on the official veterinarian order to screen birds held in areas where WNV had been detected. The collection of the samples was carried out by the respective zoo veterinarians. Their handling is, therefore, not subject to further approval.

CO N FLI C T S O F I NTE R E S T
The authors declare no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The data that support the findings of this study are available in the main manuscript and in the supplementary material of this article.