GENETIC CHARACTERIZATION OF H13 AND H16 INFLUENZA A VIRUSES IN GULLS (LARUS SPP.) WITH CLINICALLY SEVERE DISEASE AND CONCURRENT CIRCOVIRUS INFECTION

Influenza A viruses (IAVs) are naturally maintained in wild, aquatic waterfowl, mainly ducks and shorebirds (orders Anseriformes and Charadriiformes), which generally remain apparently healthy during infection (Webster et al. 1992). The IAVs are subtyped based on the antigenically distinct surface glycoproteins: the hemagglutinins (HAs) and the neuraminidases (NAs). To date, 16 HAs (H1 to H16) and 9 neuraminidases (N1 to N9) have been identified in birds (Hinshaw et al. 1982; Webster et al. 1992; Fouchier et al. 2005). Two HA subtypes, H13 and H16, are found almost exclusively in gulls (family Laridae) and terns (family Sternidae; Fouchier et al. 2007). Previous genetic analyses of the genes encoding both the surface glycoproteins and the internal proteins of H13 and H16 subtype viruses have revealed the formation of gull-specific clades and gull-specific genetic signatures that are indicative of evolutionary adaptation to gulls (Fouchier et al. 2005; Tonnessen et al. 2013a). In experimental infections with different H13 strains, no replication or limited infections were observed in chickens (Gallus sp.), ducks (Anatidae), and turkeys (Meleagris sp.), whereas mild disease was observed in ferrets (Mustela sp.; Hinshaw et al. 1982; Brown et al. 2012). Similarly, experimental infection of chickens with a H16 virus caused only limited infections (Tonnessen et al. 2011). The molecular specificity behind host-range preference of H13 IAV derives from fine differences in the host cell α2,3-linked sialic acid receptor composition (Gambaryan et al. 2005; França et al. 2013).

Most ducks appear to shed IAVs primarily from the intestinal tract (Costa et al. 2011; Krauss et al. 2013). Although gulls shed H16 from the intestinal epithelium, H13 shedding has been detected from both the oropharynx and intestinal tract (Brown et al. 2012; Hofle et al. 2012; Lindskog et al. 2013). Virus transmission in gull populations is probably similar to transmission in duck populations. The transmission is favored by mass congregations of gulls of different ages during molting, migrating, and wintering (Olsen et al. 2006).

Many shorebird species encountered in the Northern hemisphere are long-distance migrants and are capable of intercontinental migration. Extensive phylogenetic analyses of IAV genes have revealed a distinction between Eurasian and North American strains. Reassortment between these two gene pools is thought to occur by connecting bird populations at sites at which migratory flyways overlap and is possibly facilitated by the involvement of gulls and shorebirds (Wille et al. 2011; Dusek et al. 2014). In addition to the geographic mosaicism, reassortment of gull-origin strains with duck-origin strains contribute to the formation of multiple reassortants, comparable with the swine influenza triple reassortants (Hall et al. 2013).

In the passive surveillance program for influenza A viruses in Finland, conducted by the Finnish Food Safety Authority (Evira, Helsinki, Finland), IAVs have been regularly isolated from gulls. These isolates have, with the exception of one H3N8 virus, been of the H13 or the H16 subtypes. The atypically severe disease seen in most of our IAV-positive gulls and the occasional findings of circovirus (CV)-like inclusion bodies in the bursa of Fabricius in diseased gulls, including IAV-positive gulls, led us to investigate the possible role of a concurrent CV infection in the affected birds. Circoviruses, members of the family Circoviridae, are small, nonenveloped viruses with circular, single-stranded DNA genomes of 1.7–2.3 kilobases. Although CV infections may appear as subclinical, some strains are immunosuppressive and cause mild to severe disease in birds and swine (Todd 2000, 2004). They exhibit a narrow host range and to date, several avian CV species have been distinguished and named according to their host: psittacine beak and feather disease virus, canary CV, duck CV, finch CV, goose CV, gull CV (GuCV), pigeon CV, raven CV, starling CV, and swan CV (International Committee on Taxonomy of Viruses 2015). The GuCV has previously been reported in European Herring Gull (Larus argentatus), Black-headed Gull (Chroicocephalus ridibundus), and Southern Black-backed Gull (Larus dominicanus), but knowledge of their prevalence and distribution is limited (Twentyman et al. 1999; Smyth et al. 2006; Todd et al. 2007).

Gulls were screened for IAV RNA for both passive and active surveillance purposes during the years 2005–10 in Finland. We studied 202 dead or diseased and subsequently euthanized gulls (passive surveillance samples) and 207 clinically healthy birds sampled during hunting and banding (active surveillance samples; Table 1). Birds were tested throughout the year, but most passive surveillance samples were received in late summer and autumn (August–November), after the breeding season. Active surveillance samples were collected from April to October from coastal areas in southern Finland. The samples from apparently healthy birds were collected when they were trapped for banding. The gulls were trapped in cages, and oropharyngeal and cloacal swabs were collected by experienced banders. The swab samples were collected with nylon-flocked swabs, immersed in universal transport medium (both from Copan Diagnostics, Brescia, Italy), and stored at room temperature or chilled until they reached the laboratory, where they were stored at −80 C. Dead birds were sampled during routine necropsy at Evira. Tissue samples included brain, trachea, lung, liver, heart, spleen, pancreas, and intestine. Samples were sent to Evira, where they were processed according to the Diagnostic Manual for Avian Influenza (OIE 2014). All IAV testing was performed with a real-time reverse transcription (RT)-PCR targeting the M gene.

The tested species included European Herring Gulls (58%), Black-headed Gulls (22%), Common Gulls (Larus canus; 8%), Great Black-backed Gulls (Larus marinus; 5%), and Lesser Black-backed Gulls (Larus fuscus; 3%; Table 1). All procedures involving wild birds were approved by the Finnish Animal Experiment Board (permit STH355A).

When initial pathologic findings were suggestive of CV infection in several of the birds, all the IAV-positive (with the exception of one that was no longer available for testing) and 56 IAV-negative control samples (in total, 69 samples) were additionally screened for CV DNA by PCR (Table 2).

Viral RNA was extracted from 140 μL of the sample supernatant using the QIAmp Viral RNA Mini Kit (QIAGEN, Hilden, Germany). The RNA was eluted in 30 μL H₂O, and 2 μL was used as template in real-time RT-PCR assays. Screening of the samples was performed with a real-time RT-PCR assay targeting the IAV M gene (Spackman et al. 2002). In addition, all positive samples were studied with a real-time RT-PCR assay targeting the H5 and H7 genes to identify possible highly pathogenic strains (Slomka et al. 2007). Cycle threshold values of ≤37 were regarded as positive, and all samples yielding values greater than that were regarded as unresolved until confirmation by virus isolation and/or sequencing.

Virus isolation was performed to confirm real-time RT-PCR results and to produce more viral RNA for sequencing. The 8–10-d-old, embryonated chicken eggs were inoculated with 200 μL of sample supernatant. Eggs were checked daily, and the allantoic fluids were harvested when embryos appeared dead or dying. The remaining eggs were harvested on the sixth day. The allantoic fluids were finally tested for hemagglutinating activity (HA). If no HA was detected, the allantoic fluids were subjected to a second round of passaging. Detailed procedures for virus isolation and HA tests are described in the Diagnostic Manual for Avian Influenza (OIE 2014).

The HA and NA segments of positive samples were amplified for genetic subtyping with a universal primer set (Hoffmann et al. 2001). In addition, subtype-specific primers were designed based on the available published sequences of the H13 and H16 genes, and sequences were retrieved from our viruses (primers are available upon request). The amplification reactions were performed with OneStep RT-PCR Kit (QIAGEN). Reaction mixes were prepared as follows: 10 μL 5X PCR buffer, 2.5 μL deoxynucleotides (10 mM), 1 μL of each primer (50 pmol/μL), 2 μL enzyme mix, 0.5 μL RNase inhibitor (20 U/μL), 5–10 μL template RNA filled to 50 μL with water, and thermal amplification cycle conditions of 50 C for 30 min; 95 C for 15 min; 40 × (94 C for 30 s, 58 C for 30 s, 72 C for 4 min); and 72 C for 10 min were applied.

When possible, the original swab specimen was used for sequencing; otherwise, allantoic fluids from virus isolations were used. Sanger sequencing was conducted at the FIMM Technology Center's SeqLab (Helsinki, Finland), applying BigDye 3.1 chemistry (Thermo Fisher Scientific, GmbH, Schwerte, Germany) and run by the Applied Biosystems 3730xl DNA Analyzer (Thermo Fisher Scientific). Sequences of the H13 and H16 genes were edited using MEGA 6.0 software (Tamura et al. 2013). Sequences were subjected to a BLAST search using GenBank (National Center for Biotechnology Information [NCBI], Bethesda, Maryland, USA) to identify viruses that were genetically highly identical (NCBI 2016a). For additional reference sequences, available H13 and H16 nucleotide sequences were retrieved from the NCBI Influenza Virus Sequence Database (NCBI 2016b) and aligned by Clustal W within MEGA software version 7.0 (Kumar et al. 2016). The best-fit substitution models were estimated for the alignments, and maximum-likelihood phylogenies were constructed within MEGA7 (Kumar et al. 2016). Tree support was determined by bootstrap analysis using 1,000 replicates. We deposited H13 and H16 genes with GenBank (Table 3).

We extracted DNA from 200 μL of the original sample suspension by QIAmp DNA Blood Mini Kit (QIAGEN) and eluted it in 30 μL water. Our protocol was based on a previously described method (Todd et al. 2007) for amplifying a 250-nucleotide segment of the CV Rep gene with the primers (CAN-1F: 5′-GGA-GCT-GTT-GCC-GCC-GTG-A-3′; CAN-1R 5′-TAC-CCA-TCC-CAC-CAG-TCA-CC-3′). For the PCR reactions, 5 μL of the DNA template was used in 50 μL reaction mixes, including 5 μL 10X PCR buffer, 10 μL 5xQ solution, 1 μL deoxynucleotides (10 mM), 1 μL of each primer (50 pmol/μL), and 0.25 μL HotStarTaq polymerase enzyme with the HotStarTaq DNA Polymerase Kit (QIAGEN). The following cycling conditions were applied: initial denaturation at 95 C for 15 min; 30 cycles of 94 C for 30 s, 57 C for 45 s, and 72 C for 1 min; and final extension at 72 C for 7 min. Sanger sequencing was conducted using the same primers (Haartman Institute Sequencing Core Facility, Helsinki, Finland), to verify that the amplified PCR products were indeed avian CV.

During years 2005–10, a total of 409 gulls were screened for IAV RNA by PCR: 202 and 207 from the passive and active surveillance programs, respectively (Table 1). The European Herring Gull and Black-headed Gulls were the most numerous species in our sample material, accounting for 58% and 22% of the sample, respectively. The passive surveillance program, which used diseased and dead birds, resulted in a detection of 10 IAVs (5%), and the active surveillance yielded four IAVs (1.9%; Tables 1, 3). Eleven (79%) of the 14 IAV-positive samples were derived from European Herring Gull. All IAVs were detected from samples collected in August, September, and October. Based on sequencing, 11 strains were subtyped as H13, two as H16, and one as H3N8. The NA sequences were not retrieved from all IAV RT-PCR positive samples. When sequences were available, H13 was found in combination with N2 and N6, whereas H16 was combined with N3 (Table 3). We detected IAVs from 11 European Herring Gulls (subtypes H3N8, H13, and H16), two lesser Black-backed Gulls (H13), and one Great Black-headed Gull (H13). Interestingly, a high proportion (10/14; 71%) of the IAV-positive gulls displayed clinical signs of disease or were found dead. Twelve of the IAV-positive birds derived from the Oulu region in northern Finland, and two were from the southern coast of the country.

We screened a panel of 69 birds, including 13 of the IAV-positive gulls and 56 IAV-negative gulls, for CV DNA (Table 2). Overall, CV DNA was detected in 30% (21/69; 95% confidence interval [CI]: 19.6, 41.3) of these samples. Among clinically healthy birds, CV was detected in 13% (3/24; 95% CI: −0.7, 25.7), whereas it was detected in up to 40% (18/45; 95% CI: 25.7, 54.3) of the dead or diseased birds (P=0.027 by Fisher's exact test). In the group of dead or diseased birds, one third of the CV-positive birds were concurrently infected with IAV (6/18). Only tissue samples tested positive for CV DNA, although both cloacal and oropharyngeal swabs have been shown to be suitable sample types for pigeon CV detection (Duchatel et al. 2006).

We detected CV DNA in 54% (7/13; 95% CI: 26.7, 80.9) in the IAV-positive group, whereas it was found in a significantly smaller proportion, 25% (14/56; 95% CI: 19.2, 30.8), in the IAV-negative group (P=0.016 by Fisher's exact test). One of the birds concurrently infected with IAV and CV appeared clinically healthy, whereas the remaining six were found dead or diseased (Table 2). Five of the coinfected birds were Herring Gulls, one a Great Black-headed Gull and one a Lesser Black-backed Gull.

Using BLAST searches for a short sequence (200 nucleotides) showed that the highest nucleotide identities were shared with GuCVs (98% and 99% nucleotide identity with KT454927 and DQ845074, respectively) previously detected in Europe. Our GuCV sequences shared only up to 75% sequence identity with beak and feather disease viruses known to cause both acutely fatal and chronic disease in psittacine birds.

Partial HA sequence was obtained from eight of the 11 H13 viruses. Because of poor sample quality or low RNA gain, longer reads for phylogeny and GenBank submission were only sequenced from a few of them (Table 3). Three viruses—A/European Herring Gull/Finland/9875/2005, A/European Herring Gull/Finland/9611/2007, and A/European Herring Gull/Finland/13833/2008—were subjected to phylogenetic analyses (Fig. 1). In the maximum-likelihood tree generated from full-length HA gene sequence alignments (Fig. 1), A/European Herring Gull/Finland/9875/2005 and A/European Herring Gull/Finland/9611/2007 showed strong phylogenetic relatedness with Norwegian, Swedish, Dutch, and Central Asian viruses (98% to 99% nucleotide identity). Sequence alignment (positions 445–943) of the remaining Finnish viruses showed 98% to 99% sequence identity with A/European Herring Gull/Finland/9875/2005 and A/European Herring Gull/Finland/9611/2007 and clustered phylogenetically closely in analyses of 300–600 nucleotide long sequence alignments of the HA genes. Interestingly the HA gene of A/European Herring Gull/Finland/13833/2008 diverged from our other isolates by 24–27% and clustered exclusively with North American and South American strains (up to 98% nucleotide identity).

The HA genes of the two H16 strains were both from Herring Gulls and shared 97% nucleotide identity (nucleotides 1–1,006). By BLAST analysis and supported by phylogenetic trees based on the full-length HA gene of A/European Herring Gull/Finland/13022/2005 (Fig. 2), the virus isolated in Finland was most closely related to northern European strains (99% nucleotide identity) and to Central Asian strains (98–99% nucleotide identity). The H3N8 virus is genetically highly similar to duck-origin viruses that circulate readily among wild ducks in Finland and was not analyzed further for this study (Lindh et al. 2008).

Influenza A viruses of the subtypes H13 and H16 circulate in the gull populations of the Northern Hemisphere. Significant differences in IAV prevalence have been reported among sampling times, locations, and sampled species (Munster et al. 2007). In a recent study of Black-headed Gull populations in the Netherlands, high fluctuations in monthly H13 and H16 prevalence was observed, ranging from 0% to 47%, with epidemics limited to first-year birds and during June–July (Verhagen et al. 2014). Among gulls screened in Norway between 2005 and 2010, the prevalence varied between species with 8% and 24% recorded in Herring Gulls and Black-headed Gulls, respectively (Tonnessen et al. 2013b). In this study, in which both active and passive surveillance samples were screened for IAV RNA, 1.9% and 5%, respectively, of the two sample types were found to be positive. Low-pathogenic IAVs generally cause inapparent signs of disease in naturally infected gull hosts (Munster et al. 2007; Höfle et al. 2012; Verhagen et al. 2014). Surprisingly, most (10/14) IAV-positive samples in this study originated in dead or diseased gulls. The atypical severity of the disease observed in these birds led us to test all IAV-positive samples for CV DNA. Interestingly, seven of the 13 IAV samples tested were also positive for CV. Testing of a panel of passive surveillance samples (n=45) revealed that up to 40% of gulls found dead or diseased were CV positive. One third were also IAV-positive, whereas the cause of disease/death for the remaining two thirds was unknown. Coinfections by IAV and CV have not, to our knowledge, been reported, although the question of the involvement of another unrecognized pathogen contributing to the atypical outcome of IAV infections has been raised in a study of North American gulls (Hall et al. 2013). Typically, the target organ of avian CVs is the bursa of Fabricius (cloacal bursa) in which B lymphocytes develop. The resulting loss of lymphocytes exposes the bird to secondary pathogens. Circoviruses are believed to be often more widespread in the affected population than the disease incidence that they cause (Todd 2004). A clear association between IAV and CV in the clinical outcome in the gulls in this study cannot be drawn based solely on concurrent PCR detection on these viruses. Infection studies as well as histopathologic data would be needed for understanding the pathogenesis of such coinfections. However, our findings warrant more studies to assess the prevalence and distribution of CV infections in gull and wild bird populations, as well as their role as agents causing an underlying immunosuppression in birds.

A genetic separation is often seen between North American and Eurasian IAVs, probably because of geographic constraints on migratory bird movement, such as the Atlantic Sea (Olsen et al. 2006). These constraints appear more easily overcome by birds of the order Charadriiformes than of the order Anseriformes. Phylogenies of the H13 genes detected in gulls in Finland showed them to be of Eurasian origin with highly identical viruses detected in northern Europe and Central Asia, with the exception of one H13N2 virus detected in northern Finland with an HA gene of North American origin. The H16 viruses detected in Finland were closely related to viruses from northern Europe and Central Asia. Overall, the H16 gene phylogeny did not display a phylogenetic separation as clean-cut as the one for H13 because North American genes were found clustering closely with northern European and Central Asian viruses.

The phylogenies of gull IAVs are likely to be somewhat biased because they remain relatively little studied, and few sequences are available for analyses. In addition, screening of gulls has mainly been focused on North American and northern European populations. Many aspects of IAV dynamics in gulls are still incompletely understood. Prevalence and molecular epidemiology data from gull-origin strains from different parts of the world are needed for a better understanding of the ecology of these viruses.

We gratefully acknowledge Veli-Matti Väänänen for his expertise in avian ecology, Esa Aalto for his invaluable effort in sample collection in Oulu and ornithological advice, Jukka Tanner for field assistance, and Merja Hautala for her excellent technical assistance in the laboratory. This work was partially supported by the Ministry of Agriculture and Forestry MAKERA foundation (grant 1771/312/2014). The authors declare no conflicts of interest.