First Identification of “Brachyspira hampsonii” in Wild European Waterfowl

Anseriformes deserve special attention in the epidemiology of Brachyspira spp. because diverse Anseriformes species have been described to act as highly efficient carriers of several Brachyspira spp. that can also infect livestock. The aim of this study was to investigate the prevalence and diversity of Brachyspira spp. in waterfowl that winter in Spain. Brachyspira spp. were isolated from 51 of the 205 faecal samples collected from graylag geese and mallards in the Villafáfila Lagoons Nature Reserve (Northwestern Spain). The Brachyspira species identified through phenotyping, PCR and sequencing of the nox gene were B. pilosicoli (5.9%), B. alvinipulli (11.8%), "B. hampsonii" (19.6%), B. murdochii (23.5%) and B. innocens (39.2%). The most relevant finding of this study is the description of "B. hampsonii" in specimens from birds for the first time. Phylogenetic analysis of the nox gene sequences grouped all of the obtained "B. hampsonii" isolates into a cluster with Brachyspira strains previously identified by others as "B. hampsonii" and separated from other Brachyspira spp. isolates and reference strains. Additionally, this cluster was related to clades that grouped B. murdochii and B. innocens isolates. The identification of “B. hampsonii” was also achieved in 8 of the 10 isolates by sequencing the16S rRNA gene and tlyA gene. Regardless of the species identified, no antimicrobial resistance was observed in any of the enteropathogenic isolates recovered. This is the first description of “B. hampsonii” in European waterfowl, which might represent hosts that serve as natural reservoirs of this Brachyspira species. This finding indicates that this spirochete is not limited to North America, and its presence in wild birds in Europe poses a risk of transmission to livestock.


Introduction
Spirochaetes are frequently present in the gastrointestinal tracts of mammals, including human beings, and some birds. The genus Brachyspira belongs to this phylum and comprises Gram negative, helically coiled, highly motile and anaerobic bacteria. A total of seven species have been assigned to this genus (i.e., B. aalborgi, B. alvinipulli, B. hyodysenteriae, B. innocens, B. intermedia, B. murdochii and B. pilosicoli), and several other species have been officially proposed, although they are not yet fully recognised as members of this genus (i.e., ''B. canis'', "B. suanatina", "B. rattus", "B. pulli", "B. ibaraki", "B. corvi", "B. muri" and "B. muridarum"). Although the pathogenicity of some of these Brachyspira species has not been completely elucidated and their role in the development of intestinal diseases in their hosts remains uncertain, some of these species are unquestionably recognised as livestock enteropathogens. Thus, B. alvinipulli, B. intermedia and B.
pilosicoli are considered to cause diarrhoea, reduced egg production and faecal staining of eggshells in chickens [1]. Additionally, B. pilosicoli causes spirochaetal diarrhoea in swine and has been associated with diarrhoea in a variety of mammalian hosts, such as humans, non-human primates, dogs and horses [2][3][4].
However, the most important disease caused by Brachyspira spp. is swine dysentery (SD), a severe mucohaemorragic colitis found in swine. This condition is caused by the strongly β-haemolytic spirochaete B. hyodysenteriae and is characterised by moderate-to-high mortality rates, poor feed conversion, decreased growth rates and a high cost of antibiotic treatments. Due to its high economic impact, SD is considered one of the most significant diseases of swine worldwide. Recent studies have shown that other proposed strongly β-haemolytic Brachyspira species are able to induce a clinical disease that is indistinguishable from that caused by B. hyodysenteriae following the inoculation of growing pigs [5,6].
For example, pigs inoculated with "B. suanatina" obtained from mallards exhibited diarrhoea and colitis similar to those observed in pigs affected by SD [7]. More recently, strongly βhaemolytic isolates of Brachsypira spp. have been isolated from pigs from North America showing clinical cases of mucohaemorragic diarrhoea [6,8], but these isolates have not been identified as B. hyodysenteriae through the usual PCR diagnostic tests. The experimental inoculation of these isolates into growing pigs caused mucohaemorrhagic diarrhoea and colitis that was indistinguishable from SD [6,9]. Phylogenetic analysis of the 16s rRNA and nox genes revealed that these isolates belong to a new species that is virulent for swine, provisionally designated "B. hampsonii", which is circulating among the pig populations of the USA and Canada. However, the origin and/or non-porcine reservoirs of this novel species are entirely unknown.
Regarding the epidemiology of Brachyspira spp. in livestock, Anseriformes deserves special attention. For example, wildliving mallards have been described as highly efficient carriers of several Brachyspira species that are normally found in livestock [10]. Similarly, waterfowl are thought to play a significant role in the faecal contamination of drinking water sources and agricultural crops and may also come into close contact with livestock in outdoor herds [11], thereby allowing any Brachyspira spp. they carry to come into contact with different potential hosts. Furthermore, migratory Anseriformes, travelling from breeding to feeding and resting areas and vice versa, might carry Brachyspira spp. over long distances and act as a potential source of pathogenic isolates for livestock in different countries. In this regard, in Europe, thousands of graylag geese (Anser anser) and other migrating waterfowl travel from summer breeding areas in Northwestern Europe to lagoons in the south, particularly in Spain, where they spend the winter. Most of these waterfowl populations migrate along the Atlantic flyway, which passes through several pig-producing areas of Western Europe. Additionally, Brachyspira spp. of avian origin might harbour antimicrobial resistance that could be transferred to livestock and make the control of Brachyspira spp. in domestic populations more difficult. Although information on the antimicrobial susceptibility of Brachyspira spp. of avian origin is scarce, decreased susceptibility to βlactams has been described in B. pilosicoli isolates obtained from wild-living mallards [12]. This information points towards the necessity of testing the antimicrobial susceptibility of enteropathogenic Brachyspira spp. isolates from waterfowl to determine the occurrence of antimicrobial resistance in these populations and evaluate the risk of transmission to livestock populations.
The aim of this study was to investigate the prevalence, diversity and antimicrobial susceptibility of Brachyspira spp. isolates from waterfowl that winter in Spain, focusing primarily on enteropathogenic species found in swine, such as B. pilosicoli, B. hyodysenteriae, "B. suanatina" and the recently described "B. hampsonii".

Ethics statement
No animal was disturbed, captured or manipulated in any way in carrying out this study, and faeces were collected from the ground only after the birds had left the area by choice. All sampling procedures were authorised by the Villafáfila Lagoons Nature Reserve Director. No ethical authorisation from an Animal Experimental Committee was needed because no animals were handle during sampling, according to European legislation (Directive 2010/63/EU).

Geographical study area, wild bird population and sampling procedure
Villafáfila Lagoons Nature Reserve is located on the grainproducing plains of Zamora province in the Castilla y León region (Spain). The Reserve covers a protected area of 32,682 ha and includes a complex of temporary saline lagoons that occupy approximately 600 ha (approx. 41°49'51'' N; 5°36'31'' W) in the highest flood years. The three main lagoons cover areas of 70 to 192 ha during the winter and lose most of their flooded area during the summer ( Figure 1). The Nature Reserve was established to protect one of the world's densest populations of great bustards (Otis tarda), with more than 3.000 birds inhabiting the area. The Reserve serves as a wintering area mainly for waterfowl coming from Northern European countries. Its populations of geese, particularly the graylag goose (Anser anser), are especially numerous, reaching sizes of up to 40.000 birds, while much smaller populations of other species, such as the greater white-fronted goose (Anser albifrons), are observed. The Reserve's duck populations are also remarkably high during winter, ranging from 3.000 to 14.000 birds. There are many species of ducks, but the most common are mallards (Anas platyrhynchos) and northern shovelers (Anas clypeata). Other migratory birds that winter in Southern Spain, such as the common crane (Grus grus), use the lagoon complex as a resting area during their migrations.
On 24 November 2010 and 31 January 2011, a total of 205 freshly deposited waterfowl faeces specimens were randomly sampled from different locations around the lagoons. During sampling, areas where only one species usually rested were selected, and species identification was performed by Reserve Rangers. When the birds left the area, the team entered the area, and all faeces were analysed by an ornithologist to determine the species of origin. Individual swabs were taken from fresh faecal samples found on the ground. All of the swabs were immediately immersed in Amies medium (sodium chloride 3.00 g/L; potassium chloride 0.20 g/L; calcium chloride 0.10 g/L; magnesium chloride 0.10 g/L; monopotassium phosphate 0.20 g/L; disodium phosphate 1.15 g/L; sodium thioglycollate 1.00 g/L; charcoal 10.00 g/L; bacteriological Agar 7.50 g/L) and submitted to the laboratory.

Spirochete isolation
The swabs were refrigerated and immediately transported to the laboratory. Samples were selectively cultured anaerobically on tryptose soy agar medium supplemented with 5% ovine blood and antibiotics, as previously described by Jenkinson and Wigar [13]. The cultures showing haemolytic areas within 7 days of incubation were checked for the presence of spirochaetes via phase-contrast microscopy and subsequently subcultured on fastidious anaerobe agar (FAA) plates. To remove any possible contaminating non-spirochaetal bacteria, one 5 µl loop filled with the bacteria was suspended in brain heart infusion broth (BHI), and 200 µL of the inoculum was deposited on sterile 0.45-µm pore-size filter paper, which was applied to the surface of an FAA plate. Following 15-20 minutes of incubation, the filters were removed, and the plates were incubated in an anaerobic atmosphere at 42°C for 72 h. These filtered isolates were subculture twice on FAA plates and examined for purity via phase-contrast microscopy after 72 h of incubation. Finally, all of the isolates were cloned through limiting dilution. For this purpose, 10-fold dilutions of the isolates were performed in BHI and seeded onto new FAA plates to obtain presumably pure spirochetal genotypes. The spirochaetes were stored as described elsewhere [13].

Phenotypic characterisation
Phenotypic characterisation was performed according to a previously described panel system [13]. Under this system, 3day-old cultures were tested for β-haemolysis on sheep blood agar and classified as strongly or weakly haemolytic. Additionally, indole production was investigated through the spot indole test, and finally, hippurate hydrolysis, αgalactosidase and β-glucosidase activity were investigated using commercial tablets (Rosco Diagnostica, Taastrup, Denmark) following the manufacturer´s instructions.

DNA preparation and PCR
DNA was obtained from each isolate by boiling bacterial lysates. For this purpose, a loopfull of a 3-day-old culture was picked from an FAA plate and washed in phosphate-buffered saline, then boiled in nuclease-free-water and centrifuged. Next, the supernatant was transferred to a sterile microtube, and the DNA concentration was adjusted to 20 ng/µL before being used as template for PCR detection of the putative B. hyodysenteriae haemolysin regulatory gene tlyA and the B. pilosicoli 16S rRNA gene [15]. Furthermore, partial sequences of the nox and 16S rRNA genes were amplified by PCR using previously published techniques [16,17]. The B. hyodysenteriae B204 R (ATCC 31212) and B. pilosicoli P43/6/78 T (ATCC 51139) strains and nuclease-free water were employed as positive and negative controls, respectively.

Purification and nucleotide sequencing of the PCR products
The PCR products of the partial nox, 16S rRNA and tlyA gene sequences were purified using a commercial kit (Ilustra TM GFX TM PCR DNA and Gel Band Purification Kit, GE Healthcare, USA), following the manufacturer's instructions. The individual sequences of both strands of the selected PCR products were analysed via capillary electrophoresis in an automated sequencer (MegaBAce 500, GE Healthcare, USA). At least two different PCR products from each isolate were sequenced to verify that no errors had occurred during DNA amplification.

Phylogenetic tree construction and sequence analysis
The partial nox gene nucleotide sequences were determined for all isolates. The sequences were aligned using Clustal W software [18] and manually edited, purged of errors and corrected. For phylogenetic comparison, distance estimation was performed for the partial nox gene nucleotide sequences according to the Kimura 2-parameter method [19]. A phylogenetic tree was constructed using the neighbour-joining method and including "B. corvi" as an outgroup. To assess the statistical reliability of the generated dendrograms, bootstrapping values were calculated (1,000 replicates). The analyses were performed using the MEGA 5.1 software [20]. PCR products for the 16S rRNA gene were only obtained for those isolates that clustered with a Brachyspira spp. classified by others as "B. hampsonii" in the phylogenetic tree based on the partial nox gene sequences. These 16S rRNA sequences were analysed as described above.
The positive amplicons of the partial tlyA gene obtained through diagnostic PCR specific for B. hyodysenteriae were purified and sequenced as described above. To determine whether the Brachyspira spp. isolates positive for the tlyA gene could be classified as B. hyodysenteriae, amino acid sequences were predicted using BioEdit 7.0 [21] and compared to sequence data deposited in the GenBank database. For comparative purposes, Brachyspira spp. sequences were retrieved from the GenBank database and included in the phylogenetic analysis, based on the previous nBLAST results generated in our laboratory (Table 1).

Antimicrobial susceptibility
The antibiotic susceptibility of the isolates classified as "B. hampsonii" in this study was determined via the minimal inhibitory concentration (MIC) method using a commercially available system (VetMIC Brachy antibiotic panels; SVA, Sweden), following the manufacturer's recommendations. VetMIC Brachy antibiotic panels consist of 48-well plates, in which two-fold serial dilutions of tiamulin (0.063 to 8 μg/mL), valnemulin (0.031 to 4 μg/mL), doxycycline (0.125 to 16 μg/ mL), tylvalosin (0.25 to 32 μg/mL), lincomycin (0.5 to 64 μg/mL) and tylosin (2 to 128 μg/mL) are tested. The MIC was determined to be the lowest concentration of the antimicrobial agent that prevented visible growth of the bacteria. The absence of contamination was confirmed by phase-contrast microscopy in wells where visible growth was observed.

Isolation and phenotypic characterisation
The number of faecal samples obtained, the bird species affiliations and the prevalence of spirochaetal faecal shedding are provided in Table 2. A total of 205 faecal swabs from four bird species were collected in this study. Most of these specimens came from graylag goose (71.7%) or wild mallards (16.5%), and only a few came from great bustard and common crane. Growth of spirochaetes exhibiting weak or strong βhaemolysis was observed in 51 samples (24.9%) from waterfowl, whereas no haemolysis and/or spirochaetal growth was observed in any of the samples from other species. Following primary isolation, almost 50% of the isolates were found to be contaminated with coccoid bacteria, as determined via phase-contrast microscopy, but the contaminants were completely removed after the filtration process. The spirochaete isolates varied microscopically in shape and motility, although no correlations were observed between the morphology of the isolates and the Brachyspira species or host of origin. Additionally, they displayed different growth patterns on FAA plates after 72 h of incubation, although the differences could not be ascribed to any particular Brachyspira species. The phenotypic profiles of the pure isolates are provided in

Sequence analysis and species affiliations
All of the partial nox gene sequences analysed were 893 nt long. The phylogenetic relationships among the Brachyspira spp. isolates and reference strains are depicted in an  The isolates obtained in this study clustered with 4 of the 7 groups described. Most of the isolates belonged to cluster I, although they were allocated into two separate subclusters, harbouring isolates that are biochemically similar to B. innocens and to B. murdochii reference strains, respectively. Additionally, some isolates that were phenotypically similar to B. alvinipulli reference strain C1 T grouped in cluster IV. Finally, three isolates clustered with B. pilosicoli reference strains in cluster VI. Interestingly, ten isolates grouped together with Brachyspira strains that were recently identified as "B. hampsonii" in cluster II. This cluster was independent of the other Brachyspira spp. isolates and spirochete reference strains, although it was related to cluster I, which included B. murdochii and B. innocens isolates. Phylogenetic analysis of the nucleotide sequences of these isolates revealed that they were divided into two monophyletic clades. The first clade included 8 of the isolates obtained in this study, grouped with the previously described "B. hampsonii" (i.e., Brachyspira sp. KC35 and Brachyspira sp. EB106). Six of these isolates were closely related to previously described isolates, while 2 of them (isolates AIS72 and AIS85) constituted independent branches of the above-mentioned clade. Finally, the remaining two isolates (i.e., AIS 50 and AIS159) clustered with Serpulina sp. P280/1 in a clearly defined and separated clade of cluster II. The species affiliations of the 51 Brachyspira spp. isolates obtained in this study, based on the combination of their phenotypic profiles and phylogenetic analysis of nox gene sequences, as well as the GenBank accession numbers of the nox sequences (i.e., KF386036 to KF386086) are listed in Table 4. As can be observed in the   alvinipulli, while the other 2 isolates lacked indole production and displayed hippurate cleavage and α-galactosidase and βglucosidase activities. Surprisingly, all of the strongly β-haemolytic isolates found in this study were phenotypically and genetically identified as "B. hampsonii". The frequency of the detection of this species differed slightly depending on the origin of the samples, with only 4.9% of the goose samples being positive, while up to 9.0% of the mallard samples contained "B. hampsonii". However, the proportions of "B. hampsonii" isolates among all of the obtained spirochaetes were very similar for the two species (20.6% for geese and 17.6% for mallards). These results are consistent with the general frequency of spirochaete isolation from the faeces samples, which was higher in mallards than in graylag goose (50.0% vs. 23.1%). Nevertheless, it is remarkable that all of the Brachyspira species identified in this study were found in samples from both hosts at least once, indicating that both hosts can carry various Brachyspira species.
To confirm the presence of "B. hampsonii" in the bird fecal samples, additional gene targets were amplified and sequenced from the 10 isolates identified as "B. hampsonii". For the 16S rRNA gene, 1376 nt were determined, and the sequences were used to draw a dendrogram ( Figure 3). These sequences were submitted to GenBank, and accession numbers were assigned (KF386087 to KF386096). According to the phylogenetic tree defined by nox gene sequences, isolates AIS8, AIS64, AIS166 and AIS198 clustered with "B. hampsonii" clade I strains, while isolates AIS50 and AIS159 clustered with "B. hampsonii" clade II strains. Additionally, isolates AIS72 and AIS85, which formed two independent branches of clade I in the nox gene evolutionary tree, were allocated to groups formed by clade I and clade II strains, respectively, in the dendrogram derived from the sequences of the 16S rRNA gene. However, the affiliation of isolates AIS168 and AIS173 with "B. hampsonii" was not confirmed by the phylogenetic dendrogram based on 16S rRNA gene sequences. These isolates were rearranged and clustered with B. murdochii and B. innocens reference strains, respectively.  alvinipulli (2/14), based on phenotyping and phylogenetic analysis of the nox gene. The tlyA gene sequences obtained in this study were submitted to GenBank, and accession numbers were assigned (KF386097 to KF386110). Pairwise comparisons between the isolates indicated that AIS11, AIS18, AIS78 and AIS152, proposed as B. murdochii, show higher sequence similarity to B. murdochii reference strain 56-150T than to any other Brachyspira spp. strain, confirming their affiliation with this species. Similarly, the predicted amino acid sequences of isolates AIS63 and AIS154, which clustered with B. alvinipulli strain C1T in the phylogenetic tree, exhibited 100% similarity to a partial sequence from isolate AN3382/2/03 obtained from mallards in Sweden and classified as B. alvinipulli [10].

Sequence analysis of amplicons from diagnostic duplex PCR
However, isolate AIS63, which was phenotypically and genetically consistent with B. innocens C336, showed a predicted amino acid sequence for the tlyA gene that was 99% similar to B. murdochii 56-150T and isolate AIS11.
Comparisons of the obtained partial tlyA gene sequences between "B. hampsonii" isolates AIS8, AIS64, AIS72, AIS85, AIS166 and AIS198 and reference strains revealed similarities ranging from 87.0% to 96.1% at the nucleotide level and from 92.9% to 97.6% at the amino acid level, depending on the species considered. Specifically, the sequences from "B.

Antimicrobial susceptibility of "B. hampsonii" isolates
The MICs of the six antimicrobial agents tested against the "B. hampsonii" isolates are shown in Table 5. A clear unimodal distribution of the susceptible populations was obtained for all of the antimicrobials tested. Thus, all of the isolates were susceptible to the lowest antibiotic concentration tested, and the MICs were therefore below the lowest concentrations of the antimicrobials included in the panel.

Discussion
A number of studies have focused on estimation of the prevalence and diversity of intestinal spirochaetes in different bird populations [10,[22][23][24][25][26][27] because they could represent important reservoirs of Brachyspira species, including those that can infect and cause disease in livestock. Together, the results of these studies show that Brachyspira species are commonly found in the intestines of waterfowl species. Biologically, waterfowl species exhibit prominent migratory behaviour [28]. They usually breed in a widely dispersed pattern, over large geographic areas and concentrate during winter in a few key areas [29]. Spain is one of the main wintering areas for Northern European waterfowl populations, which are distributed among the main wetlands in the country. These high concentrations of waterfowl might facilitate the transmission of spirochaetes among different avian species and from birds to livestock. Special attention should be given to the possibility of the transmission of pathogenic Brachyspira spp. to swine because Spain is the second largest pigproducing country in Europe, with an output of more than 41 million pigs per year. Moreover, outdoor production, particularly associated with the semi-intensive or extensive systems where Iberian pigs are reared, might facilitate contact between reservoirs of Brachyspira spp., including several bird species, and pigs. Therefore, the aim of this study was to investigate the prevalence of Brachyspira spp. species, focusing in particular on spirochaetes that are pathogenic for swine, in the bird populations of the Villafáfila Lagoons Nature Reserve, which is one of the main wintering sites for waterfowl on the Iberian Peninsula.
For this purpose, we collected faeces from different avian species and analysed these specimens for the presence of spirochaetes. Although we cannot completely rule out the possibility that faeces were mistakenly collected from other species, the likelihood of misidentification was low because the sampling areas were selected based on specific resting or feeding areas for a single species (i.e., grain plains for great bustards, crop fields for graylag geese and the banks of lagoons for mallards and common cranes). In addition, only fresh faeces were collected, increasing the probability that the faeces belonged to birds that had just left the area. Finally, the sizes of the faeces produced are quite different between the sampled species, facilitating their identification. However, molecular tools, such as DNA barcoding [30], would be necessary to achieve definitive and unambiguous identification of the species of origin for the faecal samples.
Nevertheless, our results are consistent with previous data, further indicating that the different avian species targeted were correctly identified. Thus, although a number of intestinal spirochaetes are able to colonise wild-living non-anseriform birds, such as game birds, jackdaws and rheas [26,27], Anseriformes have been suggested to represent some of the main carriers of Brachyspira spp. In accordance with this notion, spirochetes were not observed in specimens from the common crane or great bustard in this study, while several Brachyspira species were isolated from both graylag geese and mallards. Wild-living mallards have been reported to support a high biodiversity of Brachyspira spp. in their intestines, frequently excreting spirochaetes in their faeces without showing any clinical signs [10]. In contrast, disease associated with spirochaetal infection has been reported in geese. In particular, B. alvinipulli has been isolated from geese suffering from fibrinonecrotic typhlocolitis [31]. The results of the present study confirm that wild-living mallards can carry a variety of Brachyspira species and point to graylag geese as a relevant host for most of the intestinal Brachyspira spp. (including B. murdochii, "B. hampsonii", B. innocens, B.  pilosicoli and B. alvinipullli). This role has been confirmed in mallards, which have been reported to carry Brachyspira spp. isolates whose nox gene sequences are similar to those of isolates recovered from livestock [10].
Nevertheless, most of the Brachyspira spp. isolates obtained from waterfowl in this study belonged to non-pathogenic species or species of unknown virulence, such as B. murdochii and B. innocens, which have been reported to constitute part of the microbiota of the intestinal tract of birds and do not induce clinical disease or lesions following experimental infections [32].
Regarding the pathogenic spirochaetes recovered, B. pilosicoli was only identified in 2 of the 34 mallard samples and 1 of the 147 goose samples. The low frequency of isolation of this Brachyspira species was unexpected, especially because its reported prevalences in water bird species in Australia and wild-living mallards in Sweden are 40% and 47%, respectively [10,24]. The ultimate reasons for these differences in the frequency of detection of B. pilosicoli in waterfowl between our study and previous studies remain unclear. However, it is noteworthy that the prevalence reported for B. pilosicoli in pigs in Spain is also very low [33], and the isolation of this Brachyspira species from clinical samples from swine submitted to the Diagnostic Laboratory of the Department of Infectious Diseases of the Veterinary Faculty of the University of Leon, Spain, is an infrequent event. Together, these data indicate that B. pilosicoli might not be a common spirochete in the environment in Spain. It should be noted that none of the isolates described as B. pilosicoli in the present study showed the typical phenotypic profile of biochemical group IV. However, this finding was not surprising because phenotypic diversity is a common finding among B. pilosicoli isolates from birds and pigs [10,34], and classification must be performed via PCR, multilocus enzyme electrophoresis and/or gene sequencing.
A total of 10 strongly β-haemolytic Brachyspira spp. isolates were obtained from mallards and geese. Unexpectedly, none of these spirochaetes produced indole. Furthermore, the genomic analyses indicated that none of them clustered with any of the reference strains of B. hyodysenteriae and "B. suanatina" included in the study. Although "B. suanatina" has been isolated from mallards and pigs in Sweden and Denmark [35], no isolates of this newly proposed species have been identified among Brachyspira spp. collected in Germany [36]. There are also no previous reports concerning the presence of this species in Spain, and to our knowledge, it has never been isolated from pigs or birds. In the case of B. hyodysenteriae, previous studies carried out in Sweden have indicated that up to 20% and 10% of the spirochaetes isolated from farmed and wild-living mallards, respectively, show a phenotypic profile similar to B. hyodysenterie [3]. However, no information related to the distribution of this species in waterfowl in other geographic regions is available. Thus, it is possible that this Brachyspira species is not as frequent in waterfowl as inferred from the available studies from other areas. Nevertheless, this spirochete is commonly found in pigs in various countries, including Spain [33,36].
Surprisingly, all of the strongly β-haemolytic Brachyspira isolates recovered from faecal samples from waterfowl species were classified as "B. hampsonii" through phenotyping and genotyping. The genotypic characterisation was initially based on partial sequences of the nox gene. Based on these sequences, the isolates identified as "B. hampsonii" were clustered into two previously described clades [8]. Clade I was composed of 8 of the 10 isolates analysed. These isolates were more diverse than the two isolates grouped in clade II with Serpulina sp. P/280. It is remarkable that the isolates recovered from waterfowl were divided into the same two clades previously described for swine "B. hampsonii" isolates and that the genetic distance of the isolates included in clade I was greater for both the swine and waterfowl isolates compared to the diversity of the isolates of both origins included in clade II.
However, the analysis of 16S rRNA gene sequences resulted in rearrangement of the phylogenetic tree. Thus, 2 isolates previously considered to be "B. hampsonii" clustered with B. innocens and B. murdochii. The lack of correlation between these two phylogenetic dendrograms could be due to the genomic plasticity of spirochaetes. Thus, genomic rearrangements and recombination events are possible between different species of spirochaetes due to horizontal gene transfer promoted by a chromosomal prophage [37]. These events are facilitated by the fact that most Brachyspira spp. are commensal organisms found in the intestinal tracts of waterfowl species [7,10] and can infect the same individual concurrently [9]. For example, Jansson et al. [10] suggested that recombination had occurred in the nox and 16S rRNA genes of three B. pilosicoli isolates obtained from a single sample from a mallard. The intestinal tracts of waterfowl might represent a highly suitable habitat for recombination between spirochaetes, and these events greatly complicate the classification of isolates belonging to the Brachyspira genus, especially in avian hosts and/or new environmental niches. Nevertheless, based on phenotypic characterisation, all of the Brachyspira spp. isolates obtained in this study that were grouped together with "B. hampsonii" clade I or clade II based on nox gene sequences are proposed as "B. hampsonii" in this paper.
It is remarkable that the present study is the first to report the isolation of the newly proposed Brachyspira species "B. hampsonii" from wild mallards and graylag geese in Europe. Furthermore, although the sample size was limited, this newly proposed Brachyspira species was isolated at a relatively high frequency from waterfowl wintering in Northwestern Spain, showing an estimated prevalence close to 20%. Prior to this work, the presence of "B. hampsonii" was believed to be restricted to North America, where veterinary diagnostic laboratories have reported circulation of the bacteria in pig populations in Canada and the USA [5,7,8]. The presence of "B. hampsonii" in wild birds in Europe observed in this study might be explained by the intercontinental transmission of this species between American and European waterfowl populations because the global migratory pathways of waterfowl overlap between Europe and eastern North America, which facilitates pathogen exchange between the two populations [38,39]. Additionally, "B. hampsonii" has been recently isolated from lesser snow geese (Chen caerulescens caerulescens) in the Canadian arctic [40]. However, the origin of "B. hampsonii" in the European region and its relationship with North American isolates is still unknown.
Regardless of its origin, the circulation of "B. hampsonii" in European graylag geese and mallards poses a risk for swine because this pathogenic spirochete could theoretically spread from wild waterfowl to pigs. Thus, the relatively high frequency of the isolation of "B. hampsonii" from goose faecal samples points towards a potential role of graylag geese in the epidemiology of infections by this Brachyspira spp. in domestic populations. However, the isolation of this species from mallards deserves special attention because, unlike the graylag goose, which is found in specific wetlands, mallards are widely distributed across a broad range of habitats, and they can be frequently observed at shallow water depths in the field. This particular behaviour might facilitate contact between contaminated water and/or crops and pigs reared in outdoor units, thereby enabling the transmission of "B. hampsonii" to the swine population.
Consequently, the results of this study highlight the importance of carrying out studies examining the prevalence of "B. hampsonii" in the pig populations of Spain and Europe. Furthermore, our results indicate that cross-reactions can occur between B. hyodysenteriae and some isolates of "B. hampsonii" when a PCR assay designed for the routine diagnosis of swine dysentery is applied [14]. Cross-reactions between B. hyodysenteriae and other spirochaetes in speciesspecific PCR assays have been described in the past, though only for isolates phenotypically similar to B. intermedia, B. innocens and B. alvinipulli [10]. However, "B. hampsonii" can exhibit strong β-haemolysis in culture, similar to that produced by B. hyodysenteriae. Additionally, consistent with their similar culture properties [6], some strains of "B. hampsonii" have been described to be as virulent as a reference strain of B. hyodysenteriae [9] and to cause severe mucohaemorragic diarrhoea that is undistinguishable from swine dysentery. In fact, it has been reported that in North America, strongly βhaemolytic spirochaetes classified as "B. hampsonii" are more frequently isolated in cases of dysentery than B. hyodysenteriae [9]. Hence, the existence of cross-reactions between "B. hampsonii" and B. hyodysenteriae in a theoretically B. hyodysenteriae-specific PCR assay could have important consequences for the appropriate diagnosis of brachyspiral colitis in swine and for evaluation of the real incidence of "B. hampsonii" in the European pig population. In this regard, de Jong et al. [41] recently described a β-strongly haemolytic isolate in Europe that was positive for the tlyA gene but was identified as B. intermedia based on nox gene sequencing, suggesting an affiliation with "B. hampsonii". All of these data indicate that routine diagnosis based on commonly used PCR assays targeting the tlyA gene might be misleading in this complex genomic context and that caution should be employed in the molecular identification of Brachyspira spp. Other PCR assays, such as those based on the nox and 23S rRNA genes, may be more specific for B. hyodysenteriae identification. However, further studies are needed to assess these diagnostic PCR assays and to improve the specificity of all the diagnostic techniques applied for the diagnosis of porcine spirochaetal colitis.
Finally, regarding the antimicrobial susceptibility of the identified "B. hampsonii" isolates, all of the tested isolates showed susceptibilities that were below the previously reported wild type cut-off values for the antimicrobial agents [42]. These results were not unexpected and are in accord with those obtained by Jansson et al. for different Brachyspira spp. isolated from wild-living mallards [12], likely due to the limited environmental exposure to antimicrobials among intestinal spirochaetes originating from wild birds. However, these values could be used as a baseline for monitoring the antimicrobial susceptibility of avian "B. hampsonii" isolates.
In conclusion, this study describes the presence of the newly described Brachyspira spp. "B. hampsonii" for the first time in waterfowl populations wintering in Northern Spain. This finding indicates that this spirochete is not limited to North America, and its presence in wild birds in Europe poses a risk of transmission to livestock. Consequently, additional studies are needed to determine the prevalence of "B. hampsonii" in the European swine population and the existence of other nonavian hosts for this Brachyspira species and the potential pathogenicity of "B. hampsonii" isolates recovered from waterfowl for swine.