Occurrence of Mycoplasma spp. in wild birds: phylogenetic analysis and potential factors affecting distribution

Different Mycoplasma species have been reported in avian hosts. However, the majority of studies focus on one particular species of Mycoplasma or one host. In our research, we screened a total of 1141 wild birds representing 55 species, 26 families, and 15 orders for the presence of mycoplasmas by conventional PCR based on the 16S rRNA gene. Selected PCR products were sequenced to perform the phylogenetic analysis. All mycoplasma-positive samples were tested for M. gallisepticum and M. synoviae, which are considered the major pathogens of commercial poultry. We also verified the influence of ecological characteristics of the tested bird species including feeding habits, habitat types, and movement patterns. The presence of Mycoplasma spp. was confirmed in 498 birds of 29 species, but none of the tested birds were positive for M. gallisepticum or M. synoviae. We found possible associations between the presence of Mycoplasma spp. and all investigated ecological factors. The phylogenetic analysis showed a high variability of Mycoplasma spp.; however, some clustering of sequences was observed regarding particular bird species. We found that wild migratory waterfowl, particularly the white-fronted goose (Anser albifrons) and mallard (Anas platyrhynchos) could be reservoirs and vectors of mycoplasmas pathogenic to commercial waterfowl.

Mycoplasmas are the smallest bacteria widespread in nature. The Mycoplasma genus contains more than 100 species. Some of them appear to be more pathogenic than others but most of them occur as commensals or opportunistic pathogens. Various clinical signs and lesions may be caused by mycoplasmal infections, and these infections may lead to respiratory and reproductive disorders, conjunctivitis, arthritis, and skeletal abnormalities 1 .
The two well-known poultry mycoplasmas that are considered the most pathogenic, Mycoplasma gallisepticum (MG) and Mycoplasma synoviae (MS), have been detected many times in different wild bird species representing different orders. It is well known that many mycoplasmas are host-specific, therefore infections caused by MG and MS mainly manifest clinical signs in wild species of Galliformes [2][3][4][5][6][7][8][9] . The outbreaks of MG infection in wild passerines in the USA provide an example of its ability to adapt to a new host and to develop typical clinical signs of infection. Although MG was found in certain raptors 10 and a broad range of species of the Passeriformes order 11 , the disease symptoms were found only in the Fringillidae [12][13][14] and Paridae 15 families. Conjunctivitis possibly due to MG infection has also been described in other avian species, but these data were based on observation of symptoms in birds and the presence of MG was not confirmed by diagnostic tests. Usually, asymptomatic hosts were considered vectors or reservoirs of MG 16 . A recent study proves that asymptomatic MG infection in the Eastern bluebird (Sialia sialis) can decrease the level of hemoglobin and body weight 17,18 . Another Mycoplasma species that can cause conjunctivitis is Mycoplasma sturni, but its host range is limited to only wild birds. Mycoplasma sturni was identified for the first time in a European starling (Sturnus vulgaris) with severe bilateral conjunctivitis 19 and also isolated from other passerine species such as the northern mockingbird (Mimus polyglottos), blue jay (Cyanocitta cristata) 20 , house finch (Haemorhous mexicanus) 21 , cliff swallow (Petrochelidon pyrrhonota) 22 , American crow (Corvus brachyrhynchos), and American robin (Turdus migratorius) 23 . All these species of birds showed clinical signs comparable to those present in the European starling. However, the presence of M. sturni was also confirmed in birds that did not show any clinical signs 23 and in those birds whose symptoms were due to MG infection 21

Strigiformes / Strigidae
Tawny owl (Strix aluco) AB; TR; S 0 / 2 -  www.nature.com/scientificreports/ closely related (98.4-99.3% similarity) to strains of Mycoplasma hafezii isolated from a peregrine falcon (Falco peregrinus) from Germany (GenBank accession no. EU660528) and a lesser kestrel (Falco naumanni) from Spain (GenBank accession no. EU684063). The sequence incorporated in the subgroup 2.2 was obtained from a mallard (GenBank accession no. MT374123) and showed similarity to a German isolate from a Eurasian coot (Fulica atra) (GenBank accession no. MK615053). The third subgroup (2.3) comprised Polish strains isolated from feral pigeons (GenBank accession nos MT678840 and MT678841), common gulls (GenBank accession nos MT367885 and MT367890), and a white stork (GenBank accession no. MT358569). The nucleotide sequence homology within this cluster ranged from 95.3 to 100%. The Polish sequences isolated from the pigeons had 100% similarity to M. columborale (GenBank accession no. AF412975). Strong sequence identity was also observed between isolates of M. sturni (GenBank accessions nos KU174147 and NR025968) and isolates from common gulls and white storks.

Suliformes / Phalacrocoracidae
To the third group only one sequence of Mycoplasma spp. was classified. This sequence was obtained from a white-fronted goose (Anser albifrons) and showed identity with previously published M anserisalpingitis sequences from Hungarian (Genbank accession no.

Discussion
Occurrence of Mycoplasma spp. in wild birds. In general, mycoplasmas tend to exhibit host specificity and tissue tropism. Some species show more diverse tissue tropism due to differences between strains (e.g. MS), but the majority of species exhibit a predilection for selected anatomical sites 48 . For mycoplasmas typical for poultry, the sites of swab sampling are well known. In our study, we tested swabs from the oropharynx and/ or cloaca of different species of wild birds. Sampling from both sites gave us an opportunity to determine the site with the higher frequency of Mycoplasma spp.
The occurrence of Mycoplasma spp. other than MG in wild birds is poorly documented. Our findings showed a higher occurrence of other mycoplasma species in some orders of birds such as Charadriiformes, Ciconiiformes, Accipitriformes and Falconiformes than in Passeriformes and others. The high occurrence of Mycoplasma spp. in raptors observed in our study is congruent with results that were reported in lesser kestrels 49 and a western marsh harrier from Germany 31 as well as different raptor species in Illinois 10 . A previous study 50 showed the presence of Mycoplasma spp. in the common buzzard (Buteo buteo) which is in complete agreement with our findings. As far as we know, our study is the first to describe the occurrence of Mycoplasma in two species of raptors: the lesser spotted eagle and white-tailed eagle.
Charadriiformes, especially the Laridae family, are known as reservoirs of numerous viral and bacterial pathogens. However, only a few reports describe the presence of mycoplasmas in gulls 51 . In our study, the prevalence of Mycoplasma spp. in all species of gulls was high (Table 1). This is in good agreement with the results of recent www.nature.com/scientificreports/ studies obtained with the use of next-generation sequencing (NGS), in which Mycoplasma genus bacteria were detected as the most predominant 52,53 . Our results obtained from clinically healthy birds support the view that Mycoplasma species could be a part of the normal microflora of gulls. The majority of feral pigeons were found to be Mycoplasma spp.-positive and this result is in line with previous studies from Japan 54 . Similar results were also obtained in racing and ornamental pigeons 55,56 . The high occurrence of Mycoplasma in white storks observed in our study corroborates the results reported by Möller Palau-Ribes et al. 57 for a white stork population in Germany. We detected mycoplasmas in swabs from both sites in white storks, but their occurrence was lower in cloacal swab samples than in oropharyngeal ones. Samples collected from the Anseriformes order were the most abundant in our study, and 122 out of 462 (26.4%) oropharyngeal and 35 of 93 (37.6%) cloacal swab samples were found to be Mycoplasma spp. positive. Previous studies conducted in the USA by Goldberg et al. 58 showed lower prevalence of Mycoplasma spp. in wild waterfowl (3.5% and 5.5% of tracheal swabs collected from live and dead birds, respectively). The discrepancy between our results and those reported by Goldberg et al. could be related to the method used for Mycoplasma spp. detection and differences in the geographical locations of tested birds. The presence of different mycoplasmas in wild waterfowl has been reported in Europe. Bradbury et al. 59 found M. cloacale in tufted ducks (Aythya fuligula) and common pochards (Aythya ferina) from the United Kingdom. Another study performed in Spain by Poveda et al. 60 describes the presence of Mycoplasma anatis in 3 out of 10 tested shovelers (Anas clypeata). However, the authors did not find any other mycoplasmas in tested mallards or Eurasian teals (Anas crecca), which was contradictory to our findings. However, the lack of mycoplasma-positive results in the Spanish study might be caused by the small number of birds tested and the different detection method. It is worth noting that we also found Mycoplasma spp. in different species of wild geese and in mute swans. Our study, which covers a broad range of wild bird species, may help to deepen the understanding of the role of the tested species as Mycoplasma spp. reservoirs or vectors. Several studies have focused on wild birds as potential vectors of those species of Mycoplasma that are pathogenic to commercial poultry. To the best of our knowledge, our study is the first to reveal the role of wild geese as potential vectors of Mycoplasma strains that may pose a health risk to commercial waterfowl. We wanted to identify Mycoplasma spp. in different species of wild birds. Therefore, we used 16S rRNA gene sequences, which are considered to be an excellent choice for preliminary taxonomic classification and phylogenetic assignment of Mycoplasma isolates to certain groups 61 . It is worth emphasizing that our sequences originated directly from swab samples, which gave the possibility to detect numerous novel sequences that are phylogenetically assigned to the Mollicutes class and potentially represent novel uncharacterized Mycoplasma species. The 16S rRNA gene sequences have become one of the mandatory requirements for the description of new Mollicutes species. However, many additional analyses are required for the discovery and characterization of a novel species. The 16S rRNA genes have lower variability compared to the other genetic markers such as beta subunit (rpoB) genes or the 16S-23S rRNA intergenic transcribed spacer region (ITS) 62 .
The phylogenetic analysis showed identity between Mycoplasma spp. isolates obtained from two different groups of birds: white storks (MT358571, MT358576, MT358586, MT374253, and MT549663) and birds of prey (EU684058, MK615041, MK615064, and MT367903). This could be explained by those birds possibly sharing species of commensal mycoplasma 13,63 . The other sequences that were found by us in white storks showed similarity with German and Austrian isolates. However, none of the sequences collected in our study were similar to M. ciconiae sp. nov. which is considered to be common in the white stork population 57 . Möller Palau-Ribes et al. made a serological analysis to characterize the M. ciconiae and obtained a faint reaction with M. sturni antiserum. Surprisingly, the sequence found by us in Polish white storks (GenBank accession no. MT358569) was similar to that in M. sturni (96.6%), which could explain the positive reaction described in the aforementioned study. Our phylogenetic analyses also showed high similarity between sequences obtained from white storks and gulls in some clades. This finding revealed a potential role of the host diet in the frequency of detection of commensal Mycoplasma species. Furthermore, the dendrogram demonstrated that the sequences originating from gulls were clustered and those clusters were dispersed among clades. This may suggest that Laridae may be hosts of different mycoplasmas that could be unique to gulls. We believe that future work needs to be done to investigate and characterize the mycoplasmas that could be found in this family of birds.
The phylogenetic analysis of Mycoplasma spp. sequences obtained from birds of prey leads to the conclusion that most of them demonstrate close relatedness and specificity only for these hosts. Similar relationships were observed in wild waterfowl. The sequences of Mycoplasma spp. originating from mallards (MT374250 and MT374251), a mute swan (MT374120), and a graylag goose (MT374249) showed close similarity to mycoplasmas typically associated with Anseriformes. Our attention was drawn to the similarity of the sequence from a white-fronted goose (MT374256) to sequences of M. anserisalpingitis that were isolated from domestic geese from Hungary, China and Poland 34 . The results demonstrate that wild anserids can be reservoirs and vectors of various Mycoplasma species pathogenic to commercial waterfowl.
We have also identified a 100% sequence similarity between sequences obtained from common redpolls and M. orale (GenBank accession nos NR043199 and NR113660). This Mycoplasma species is considered a nonpathogenic or opportunistic part of the flora of the human upper respiratory tract 64 and is well known as a common cell culture contaminant 65 . Microbiome analysis has revolutionized knowledge of the health consequences of transmission of non-pathogenic microbes between humans and animals. A recent review of the literature on this topic found significant consequences of the transfer of non-pathogenic microbes 66 . Furthermore, metagenomic studies have proved that bacteria of the human oral microbiota could possess antibiotic resistance genes [67][68][69] . It is crucial to note that antibiotic resistance genes can be transferred to bacteria of the same or different species. The presence of antibiotic resistance genes in bacteria isolated from wild animals may be associated with their proximity to human populations 70 . Wild birds which spend time close to human settlements may acquire bacteria with these genes and contribute to their widespread dissemination, especially if the birds are migratory 71 www.nature.com/scientificreports/ Further work needs to be done to verify whether the common redpoll, as a synanthropic and migratory bird, could play a role as carrier or vector of antimicrobial-resistant strains of Mycoplasma. Knowledge of the impact of feeding behavior on bacterial microbiota in wild birds is limited. There are only few elucidated examples of the relationship between feeding habits of birds and the occurrence of particular bacteria. A good example is the difference in occurrence of Streptococcus spp. between granivorous and omnivorous birds 73 . Another study documented the impact of diet diversity on richness of the microbiota in 13 different bird species from Slovenia 74 . Stenkat 75 found that the presence of Enterobacteriaceae was significantly associated with an animal-based diet, especially a piscivorous one, while the presence of Pseudomonadaceae correlated positively with a herbivorous diet. Our findings revealed a higher occurrence of Mycoplasma spp. in birds that have animal-based diets than in herbivorous and omnivorous birds.
Habitat was verified as another potential factor which may have an influence on the occurrence of Mycoplasma spp. in birds. Our results show that the occurrence of Mycoplasma spp. was more frequent in birds inhabiting aquatic environments than in those living in terrestrial ones. The aquatic environment provides habitats to a wide range of bird species, for many of which it is the principal one, while for others it is only a feeding place. Some of them are gregarious, feeding side by side in mixed-species flocks, and leave a large amount of feces which also could be a source of Mycoplasma spp. 76 .
We found a high frequency of Mycoplasma spp. in migrating birds. The similarity of the Mycoplasma spp. sequence originating from white fronted geese to isolates of M. anserisalpingitis that were reported from other parts of the world confirmed that wild birds could play a potential role in the transmission and spread of mycoplasmas. The white-fronted goose is a migratory bird of which the breeding grounds cover almost the whole circumpolar Arctic 77 . However, these birds choose various wintering destinations such as eastern Asia 78 or western Europe 79 . They make many stopovers during migration, which increases the risk of contamination of domestic goose water and pastures by their feces. Excluding an example of transmission of M. gallisepticum between house finch populations in the USA, no evidence of pathogenic mycoplasmas in migratory birds has been documented. The second reason that migration could have an influence on the occurrence of Mycoplasma spp. is the similarity between sequences from Polish white storks and those originating from Spanish imperial eagles (Aquila adalberti). The presence of the same Mycoplasma spp. in those two species of birds could be caused by their sharing of a wintering area 80,81 as well as a probable diet.
In our study, we used a conventional PCR based on the 16S rRNA gene which might be considered either its potential or its limitation. We are aware that culture isolation is essential for the identification of the organism. However, the culture of unknown Mycoplasma species can be extremely hard due to the fastidious conditions required for their growth. Moreover, our aim was to detect different Mycoplasma species simultaneously, therefore we chose the PCR based on the 16S rRNA gene. The majority of Mycoplasma sequences in the GenBank database are sequences of that gene, which was helpful in the phylogenetic analysis of our results. As was mentioned above, the use of 16S rRNA helped us to detect sequences that potentially represent novel uncharacterized species. Another possible limitation of our work is the small sample size for some bird species. However, we decided to report these results, because in some cases, they were obtained from rare bird species.
In conclusion, the observed distribution of Mycoplasma spp. revealed that particular bird species have a predisposition to be hosts of these microorganisms. The identification of mycoplasmas detected in gulls and white storks suggested that they could be hosts of more than one species of Mycoplasma. We observed possible relationships between the presence of Mycoplasma spp. and the ecology of each bird species, i.e., its feeding habits, preferred habitats, and migration patterns. Further studies are necessary to verify the role of these mycoplasmas as opportunistic pathogens. Our results support the idea that wild migratory waterfowl could be a reservoir and vector of mycoplasmas pathogenic to commercial geese. Nevertheless, our results based on a wide range of tested bird species did not confirm the presence of MG or MS, thus the risk of transmission of those pathogens from wild birds to commercial poultry is relatively low.

Materials and methods
Birds. Samples of swabs were taken from 1141 wild birds representing 55 species, 26 families, and 15 orders between 2011 and 2019 (Table 1). A total of 990 clinically healthy live birds (86.8% of birds tested) were caught and sampled by ornithologists during regular ringing actions. The sampling was performed under the framework of the Polish national avian influenza monitoring program. The remaining 151 sampled birds (13.2%) had been found dead and were specimens that did not present any clinical signs of infection and died accidentally due to injuries caused by collisions with glass, power lines, and vehicles and were submitted for diagnostic purposes by ornithological stations and bird rehabilitation centers to the laboratory of the Department of Poultry Diseases of the National Veterinary Research Institute. All those birds were categorized into three groups of dietary preference (animal-based, mixed, and plant-based) based on the main food items known to be preferred by particular species of bird. Birds species were also categorized by preferred habitat (aquatic or terrestrial) and movement pattern (migratory or sedentary). All information about dietary preferences, habitats, and movement patterns was taken from the Birds of the World webpage 82 .
Samples. Oropharyngeal swabs were collected from 1053 birds, cloacal swabs from 88 birds, and within this total, both oropharyngeal and cloacal swabs were taken from 60 birds. The samples were collected with the use of two types of swab: dry swabs from dead birds and swabs with a commercial transport system from live ones (ESwab Collection and Transport System, Copan Diagnostic, Murrieta, CA, USA). Each dry swab was placed into a tube with 500 µl of phosphate-buffered saline solution (PBS) and centrifuged at 20,000×g for 60 s before DNA isolation. The DNA of samples collected from live birds by swabs with the commercial transport system was extracted directly from the transport medium that was centrifuged at 20,000×g for 60 s. The DNA DNA sequencing and phylogenetic analysis. The phylogenetic analysis was performed using representative samples selected from different species of wild birds. Selected PCR products were sent for sequencing by the Sanger method to the Genomed laboratory (Warsaw, Poland). Each amplification product was sequenced in both directions with the forward and reverse amplification primers. Raw sequence data was analyzed with the Statistics. The prevalence of Mycoplasma spp. in wild birds was calculated as the proportion of Mycoplasmapositive birds to the total number of birds examined with 95% Wilson's confidence intervals (95% CI). The relationships between results of the PCR and categorical explanatory variables including diet type (animal-based, mixed, or plant-based), type of habitat preferred (aquatic or terrestrial) and movement pattern (migratory or sedentary) were examined using the Fisher exact test in the rcompanion package version 2.3.26 86 . The dplyr package version 1.0.2 was used for data manipulation 87