The occurrence of the pathogenic chytrid fungus Batrachochytrium dendrobatidis (Bd) in the feet of live waterfowl has been documented, but the potential role of birds as dispersers has not been studied. We report the presence of Bd in the feet of preserved aquatic birds in the Bolivian high Andes during the time of drastic amphibian declines in the country. We sampled 48 aquatic birds from the Bolivian Andes that were preserved in museum collections. Birds were sampled for the presence of Bd DNA by swabbing, taking small pieces of tissue from toe webbing, or both. We detected Bd by DNA using quantitative PCR in 42% of the birds sampled via toe tissue pieces. This method was significantly better than swabbing at detecting Bd from bird feet. We confirmed Bd presence by sequencing Bd–positive samples and found 91–98% homology with Bd sequences from GenBank. Our study confirms that aquatic birds can carry Bd and thus may serve as potential vectors of this pathogen across large distances and complex landscapes. In addition, we recommend using DNA from preserved birds as a novel source of data to test hypotheses on the spread of chytridiomycosis in amphibians.
Infection by the chytrid fungus Batrachochytrium dendrobatidis (Bd) is one of the main causes for species extinctions and population declines of amphibians worldwide (Stuart et al. 2004). To understand how the infection front advances, it is important to study potential pathogen dispersers (Lips et al. 2006; Cheng et al. 2011). Previous studies revealed potential vectors and reservoirs of Bd, including amphibians (Kolby et al. 2015), water (Johnson and Speare 2003), crustaceans (McMahon et al. 2013), and waterfowl (Johnson and Speare 2005; Garmyn et al. 2012). In the Andes, this pathogen has been implicated in amphibian declines since the 1980s (La Marca et al. 2005). Spatiotemporal models of disease emergence predicted a southward dispersion at a rate of approximately 200 km/yr for that region (Lips et al. 2008). At such a rate, the fungus should have reached Bolivia by the year 2000. However, drastic amphibian declines in central Bolivia were documented in the early 1990s (De la Riva and Lavilla 2008). Arrival of Bd from Chile or Argentina is unlikely because the first records of Bd in these countries were after the turn of the century (Barrionuevo and Mangione 2006; Solís et al. 2009). Thus, alternative means of spread for this pathogen are worth considering. The possibility that Andean aquatic birds could play a role as dispersers is a plausible explanation for the gaps observed in the way Bd has reached some regions in Bolivia (De la Riva and Burrowes 2011).
Garmyn et al. (2012) detected Bd zoospores in 15% of wild-caught geese and ducks in Belgium. Using sterile in vitro techniques, they showed that Bd zoospores moved toward the keratinous scales of bird's feet, adhered to skin, and formed actively reproducing sporangia that released motile zoospores. Thus, Bd can successfully infect waterfowl, suggesting that these birds could disperse Bd by shedding zoospores into aquatic habitats within their range.
Here we examine the potential role of aquatic Andean birds as Bd dispersers, and we consider their capacity to introduce virulent strains into naïve amphibian populations in Bolivia. To do this, we tested for Bd DNA in 48 birds from museum collections and chose species collected near amphibian population decline sites, before (1980s) and during (1990s) the onset of the amphibian crisis (Table 1). Histology to detect Bd structures (sporangia or zoospores) was discarded because in preservation, bird feet are left to dry untreated, and eventually become mummified, being dark and oily from bone grease. Initially, we sampled 17 bird specimens from the Colección Boliviana de Fauna (CBF) in La Paz, Bolivia, and the next year we sampled 31 additional birds from the Estación Biológica de Doñana (EBD) in Seville, Spain. The locations of the collection sites are given in Table 1. Due to permit restrictions in Bolivia, birds at the CBF collection were sampled only by swabbing the interdigital webbing of the right foot by using sterile dry swabs. Before swabbing, bird feet were cleaned and then wrapped in a wet paper towel for 30–45 min to soften the skin; this procedure facilitated tissue collection. We extracted DNA from swabs with 50 μL of Prepman Ultra (Applied Biosystems, Waltham, Massachusetts, USA) following Hyatt et al. (2007), and Bd was detected via quantitative PCR (Boyle et al. 2004). All 17 birds from CBF that were sampled in this manner tested negative for Bd (Table 1). To determine whether these negative results were associated to sampling methodology, in addition to foot swabs, we took small pieces (10×5 mm) of toe webbing from the birds sampled at the EBD collection. To prevent cross-contamination when handling different bird specimens, we changed gloves and disinfected instruments by dipping in nuclease decontamination solution followed by flaming. We extracted DNA from pieces of toe webbing by using a DNeasy Tissue Kit (QIAGEN, Valencia, California, USA) following modifications for ancient bird DNA (Fulton et al. 2012). These modifications consisted in cutting the tissue into smaller pieces (2×2 mm), extending the initial lysis step by incubating at 50 C and at room temperature for 3 d, and adding an extra 20 μL of proteinase K during incubation. We used the same qPCR method for diagnosis of Bd in all tissue types (Boyle et al. 2004). No frog samples were processed in the laboratory, or shared in the PCR plate, when testing for Bd DNA in bird tissues.
Results from testing toe-webbing tissue pieces in birds from the EBD collection revealed that 42% (13/31) of the specimens were PCR positive for Bd (Table 1). In all cases except one, the corresponding swab samples tested negative for Bd, confirming that the probability of detecting Bd DNA in bird feet is associated to the sampling method (χ2=13.29, P=0.0003). To corroborate Bd identification, we sequenced all Bd–positive samples by using the Bd–specific internal transcribed spacer (ITS)-1 and 5.8S primers (Boyle et al. 2004). The PCR products were treated with Exo-SAP-IT (United States Biochemical, Cleveland, Ohio, USA) and sequenced using the BigDye Terminator Cycle Sequencing Kit (Applied Biosystems). The sequenced reactions were cleaned by precipitating with ethanol and EDTA salt and then analyzed on a 3130xl DNA Analyzer (Applied Biosystems). Individual sequences were assembled and edited in Sequencher 5.1 (Gene Codes Corporation, Ann Arbor, Michigan, USA). We used the basic local alignment search tool (BLAST) from the National Center for Biotechnology Information website (www.ncbi.nlm.nih.gov) and obtained maximum identity matches to Bd (CW34 clone ITS-1, partial sequence; 5.8S ribosomal RNA gene; and ITS-2; GenBank accession JQ582938) for DNA sequences in seven of the 13 Bd–positive samples (Table 1). There were no BLAST matches to any other organism. The percentage of query cover by alignment to the database sequences ranged from 73% to 87%, and maximum identity matches to Bd were between 91% and 98%. Lack of positive matches to sequences in GenBank for the other six Bd–positive birds may be due to poor DNA alignment (Zhang et al. 2000) or to variation in ITS-1 Bd haplotypes (Longo et al. 2013).
The earliest Bd–positive records in our survey were from the northwestern part of the Bolivian altiplano in a Puna Ibis (Plegadis ridgwayi) and a Yellow-billed Teal (Anas flavirostris) collected in 1982 at Ulla Ulla Dam, followed by Yellow-billed Pintail (Anas georgica) and White-tufted Grebe (Rollandia rolland) from Lake Titicaca in 1983 (Table 1). The fact that a duck (A. georgica) collected in 1985 in Cochabamba, approximately 440 km southeast of Ulla Ulla Dam, tested positive for Bd, suggests that aquatic birds already carried this fungus to the Andes of central Bolivia by the mid-1980s. Because birds can wander between bodies of water from different basins, it is possible that high Andean aquatic birds moved Bd to lower montane streams, resulting in the drastic declines of cloud forest anurans in the early 1990s. Although the three species of cloud forest–dwelling birds examined (Merganetta armata, Cinclus leucocephalus, and Cinclus schulzi) tested negative for Bd by both sampling techniques (Table 1), our sample size is too small to discard their roles as direct Bd dispersers. Regardless, high puna lakes from the Amazonian versant of the Andes that are heavily used by aquatic birds potentially infected with Bd, often drain into rivers flowing down to cloud forest valleys. In this manner, carrier birds may contribute to the spread of Bd across landscapes and into habitats with naïve populations of amphibians.
Cloud forests are of special concern, because it is from this habitat that nine species of aquatic frogs in the genus Telmatobius have already been lost, presumably to chytridiomycosis (De la Riva and Reichle 2014), and where we currently find the highest prevalence of Bd (De la Riva and Burrowes 2011).
Here we confirm that aquatic birds carry Bd and thus may serve as vectors of this pathogen, explaining in part the erratic pattern of disease incidence and declines observed in Bolivia (P.A.B. and I.D.l.R., unpubl.). In addition, we show that Bd DNA can be detected in museum specimens of birds, and furthermore that the probability of detection increases significantly if samples are taken from skin pieces instead of just swabbing the toes. The Bd DNA data from museum specimens of birds can provide valuable information to test hypotheses about the spread of an infectious agent across diverse ecological landscapes with barriers that primary hosts, themselves, cannot overcome. This kind of information can help us reveal patterns of chytridiomycosis occurrence across space and time and begin to understand the causes of isolated decimations of entire amphibian communities from pristine habitats, such as the cloud forests of Bolivia, whereas others persist in ecosystems nearby.
This research was supported by projects CGL2011-30393 and CGL2014-56160-P of the Spanish Government (I.D.l.R.). The DNA sequencing was possible with support from the Sequencing and Genomics Facility and the Molecular Science Research Center at the University of Puerto Rico, Río Piedras, funded by the National Institute of Health award P20GM103475. We are grateful to James Aparicio and Isabel Gómez for help at the CBF and to Teresa García Díez and José Cabot at the EBD.