Coronaviruses (CoVs) are found in a wide variety of animals in which they can cause respiratory, enteric, hepatic, and neurological disease of varying severity. In humans, CoVs cause a large fraction of common colds [1], as well as more rare, but serious disease such as the outbreak of Severe Acute Respiratory Syndrome (SARS, caused by SARS-CoV) in 2003, and the recent Middle East Respiratory Syndrome (MERS, caused by MERS-CoV) [2]. Both SARS and MERS are zoonotic diseases as they emerged following spill-over from animal reservoirs [2–4]. CoVs belong to the family Coronaviridae and are divided into four genera or groups: Alpha-, Beta-, Gamma- and Deltacoronavirus [5]. SARS and MERS are Betacoronaviruses; current phylogenetic evidence suggest that bats are the genetic source of alpha- and betacoronaviruses which include human and most other mammalian specific viruses. In contrast, gamma- and deltacoronaviruses are dominated by viruses found in birds [6–8].

The most prominent representatives of gammacoronavirus are infectious bronchitis virus (IBV) and Turkey CoV, which cause substantial economic losses to the poultry industry [9, 10]. Avian coronaviruses, most of which are distinct from IBV and Turkey CoV, circulate in wild birds [11], but the taxonomic status of “avian coronavirus” is contentious [12]. Regardless, avian gammacoronavirus has been detected across an array of avian orders including Anseriformes, Ciconiiformes, Charadriiformes, Columbiformes, Galliformes, Passeriformes, Pelecaniformes, and Psittaciformes. Members of the deltacoronvirus group are also composed of viruses detected in wild birds, which suggests that wild birds play an important role in the epidemiology of gamma- and deltacoronaviruses [6, 7, 11]. Despite this assertion, limited studies have been performed to assess host species breadth, seasonal prevalence, geographic distribution, or genetic structuring.

In this study we aimed to sample a range of waterbirds from Scandinavia for the presence of CoV. For this, we utilized samples collected from wild birds at Ottenby Bird Observatory (56°12′N, 16°24′E), which is located on Öland, an island in the Baltic Sea off the south-east coast of Sweden. It is an important stop over site for migrating birds, including Anseriiformes and Charadriiformes migrating between breeding areas in Scandinavia, the Baltic Countries, Finland, Russia and more southern wintering grounds. Furthermore, CoV have been detected and characterized from Mallards (Anas platyrhynchos) at this study site previously, allowing us to compare viruses circulating in both 2007 and 2011. Finally, we assessed interspecies transmission and geographic subdivision of avian CoV in wild birds.

In this study we aimed to further contribute to the natural history of avian coronaviruses by screening an array of waterbirds for these viruses. We found a prevalence of 18.7% CoV, which is higher than the 0–15% reported previously in wild bird studies [11, 14, 15, 21–27]. This may be due to the high proportion of dabbling ducks, particularly Mallard, in our study, and the temporal and spatial features of the dabbling duck sampling. More specifically, sampling of dabbling ducks occurred at a migratory stopover site, and thus high prevalence could be driven by migrants which not only import viruses through infected individuals, but also replenish the pool of susceptible individuals across the migratory period, resulting in high prevalence at these locations [28, 29]. Directly comparing prevalence estimates to previous studies, however, is challenging due to the array of different methods utilized, ranging from conventional PCR to multiplex real-time PCR, in addition to species distributions and seasonal variations. Regardless, this study corroborates previous studies suggesting the importance of dabbling ducks (Anas sp) as hosts of avian coronaviruses [11–15, 21–25], but also highlighting the importance of diving ducks. The high diving duck prevalence in this study was driven by high prevalence in Greater Scaup, which could be due to congregation in very large flocks facilitating transmission, but could also be due to other factors such as a local variation in disease prevalence or reflect outbreak dynamics wherein the virus rapidly expands across the susceptible population. Overall, diving ducks have only been sampled in this study and by Chu et al. (2011), so it is unclear whether this pattern of high prevalence is present globally. We found a low prevalence in waders, which is in contrast to [14] and [21], which found a very high prevalence (>20%) in waders, but corroborates findings in [22, 25, 26]. This may be due to differences in congregation and feeding patterns of waders between the locations, but has to be interpreted with caution, as numbers of waders sampled in our study, and others, were small. Finally, while assessment of gulls is limited, this study corroborates findings with Muradrasoli et al (2010) indicating that this group requires further scrutiny.

In order to better assess CoV dynamics, resampling the same site across time is imperative, and this is the first study to do so. Coronaviruses from Mallards were previously assessed using samples collected at Ottenby in 2011 [15] and we detected high similarity between viruses from 2007 and those from 2011. Indeed, 8/10 CoV were most closely related to sequences from 2011, which strongly suggests maintenance and/or location transmission of these RdRp lineages at this location. This is unlikely to be a coincidence as Swedish Mallard CoV make up less than 1/3 of those available sequences, and all current RdRp clades are comprised of sequences from Asia. The waterfowl host appears important in these samples, however assessing host species biases is challenging due to the overwhelming number of sequences from Anas species. However, coronaviruses from diving ducks were rather different from existing diversity as demonstrated by the similarity between two of three Scaup CoV sequences from this study and Tufted Duck CoV sequence from Hong Kong. The limited number of sequences in this clade could be due to sampling shortcomings, or that currently developed primers do not amplify this clade effectively.

Despite few studies, small samples sizes and differences in prevalence, what is clear, is that in the Northern Hemisphere waterfowl species, especially dabbling and diving ducks are important in the epidemiology of avian CoVs. It is interesting to note that these patterns are very similar to those found in low pathogenic influenza A viruses: high prevalence in waterfowl and gulls in the Northern Hemisphere [30], and little host species and temporal structuring within waterfowl derived viruses in the conserved polymerase genes (such as PB2, PB1) [31]. Further, detection in fecal or cloacal samples suggest these viruses may also replicate in the gastrointestinal tract, which is the true for Turkey Coronavirus [9]. As demonstrated by years of influenza sampling [32], to better understand the eitiology, epidemiology and ecology of these viruses more systematic surveillance needs to be undertaken and more viruses need to be sequenced, particularly full genomes. Given the importance of coronaviruses as human pathogens, as exemplified by SARS and MERS-CoV and the potential for wild birds as reservoirs, spreaders and mixing vessels, further studies of coronaviruses in wild birds are warranted.