Identification of Southeastern Bat Species Using Noninvasive Genetic Sampling of Individual Guano Pellets

Since its discovery in Howe's Cave, New York, United States in February 2006, white-nose syndrome (WNS), a disease caused by the fungal pathogen Pseudogymnoascus destructans, has killed in excess of 6 million cave-hibernating bats (U.S. Fish and Wildlife Service [USFWS] 2012) and resulted in devastating declines in bat populations across eastern North America. The disease is currently known to affect nine bat species (USFWS 2017, U.S. Geological Survey [USGS] 2017). Of these, the tri-colored bat Perimyotis subflavus, Indiana myotis Myotis sodalis, little brown myotis M. lucifugus, and northern long-eared myotis M. septentrionalis have experienced the greatest rates of mortality, the latter three being at risk of regional extinction (Frick et al. 2010; Langwig et al. 2012). As a result of population declines, it is anticipated there will be dramatic shifts in temporal and spatial niche partitioning and significant restructuring of bat communities (Jachowski et al. 2014).

In order to monitor changes in bat population status and community structure as a result of WNS and implement appropriate management and conservation strategies, wildlife managers have an increasing need to determine bat species occurrence. Mist netting and roost surveys have traditionally been used to determine occurrence, but the effectiveness of these methods may be limited by WNS-related population declines. White-nose syndrome has resulted in lower bat densities and, in turn, lower capture rates, such that mist netting may no longer be a feasible method for monitoring several WNS-imperiled species (Ford et al. 2011; Coleman et al. 2014; Niver et al. 2014). Roost surveys have traditionally been used to monitor 14 of the 47 bat species known to occur in the United States (U.S) because these species roost in large numbers in relatively accessible locations (Harvey et al. 2011). However, WNS-related declines have made such surveys more difficult. Other species that roost in small numbers or cryptic locations in caves, trees, or buildings are more difficult to observe and monitor using this method (Caceres and Barclay 2000; Tuttle 2003; Kunz et al. 2009; Meretsky et al. 2010; Harvey et al. 2011; Turner et al. 2011).

Ultrasonic acoustic detectors have become a common bat-monitoring tool, and the bat calls recorded by these detectors can be used to determine species occurrence (Weller 2007; Dzal et al. 2011; Ford et al. 2011; Bernard and McCracken 2017). However, call classification and species determination is often limited by the quality of recorded calls (Betts 1998; Loeb and Waldrop 2008). Further, a number of species in the southeastern United States, including members of the WNS-imperiled genus Myotis (Frick et al. 2010; Langwig et al. 2012), share similar call characteristics (Britzke et al. 2011; Kaiser and O'Keefe 2015). These similarities can lead to species misclassification and, as a result, difficulty determining species occurrence (Cox et al. 2016).

When bat species are rare, elusive, or share similar morphological or call characteristics, species identification using genetic markers from fecal material (i.e., guano) found in or around roost locations (i.e., caves, trees, and buildings) may provide an alternative indicator of occurrence (Patrick et al. 2016; Walker et al. 2016; Zinck et al. 2004) identified two genetic markers from 16S ribosomal subunit, mitochondrial deoxyribonucleic acid (mtDNA), for bat species identification using fecal DNA and applied this marker to DNA extracted from guano of bats found in the northwestern United States. Only 5 of the 14 species sequenced by Zinck et al. (2004) are found in the southeastern United States: Rafinesque's big-eared bat Corynorhinus townsendii, big brown bat Eptesicus fuscus, silver-haired bat Lasionycteris noctivagans, hoary bat Lasiurus cinereus, and little brown myotis.

In Great Smoky Mountains National Park (GRSM), located along the border of eastern Tennessee and western North Carolina, WNS was confirmed in 2010 (Carr et al. 2014) and recent studies have since documented declines in numerous bat species (Carpenter et al. 2016; O'Keefe et al. 2016). Following these declines, managers wanted to know what bat species were present in GRSM buildings in order to permit appropriate management and conservation efforts. In surveys of buildings conducted during the summer of 2015 (Fagan et al. 2017), researchers observed roosting big brown bat, eastern small-footed myotis M. leibii, and Rafinesque's big-eared bat. In other buildings, however, guano was the only indication of roosting bats.

Our objectives were to 1) apply the genetic marker from the 16S ribosomal subunit, mtDNA (Zinck et al. 2004), to DNA extracted from tissue and guano samples from the 16 species of bats known to occur in Tennessee, United States (Harvey et al. 2011); 2) use this marker to create a reference database containing DNA sequences of these species; and 3) use the reference database to identify bat species from DNA extracted from guano samples collected in buildings of GRSM.

We used DNA extracted from tissue and guano samples to create a DNA sequence database for the 16 species of bats known to occur in Tennessee (Table 1). All samples were collected from bats captured or found dead as part of other studies conducted from 2010 to 2016. For all samples, species determination was confirmed by at least two researchers experienced in bat identification. Guano samples consisted of 1–5 pellets collected from live-captured bats. Tissue samples varied from 2- to 3-mm wing biopsies from live individuals, to larger pieces of wing or tail membrane from recently dead individuals. We obtained tissue samples from other researchers to use as DNA controls for Townsend's big eared bat Corynorhinus townsendii, Seminole bat Lasiurus seminolus, and southeastern myotis M. austroriparius, which are rarely captured in Tennessee (Table 1). We were unable to obtain guano samples from these species. We stored tissue samples in either silica gel desiccant (4–10 mesh; Fisher Scientific, Pittsburgh, PA; Wasser et al. 1997) or 20% dimethyl sulfoxide (Worthington and Barratt 1996) and guano samples in silica gel desiccant (Brown et al. 2015). We froze all samples at −20°C within 48 h of collection.

All capture and handling protocols were approved by the University of Tennessee Institutional Animal Care and Use Committee (IACUC 2026-0515) and followed the guidelines issued by the American Society of Mammalogists (Sikes et al. 2016). Capture, handling, and sample collection were authorized under scientific collection permits from the USFWS (TE35313B-2), National Park Service (GRSM-2015-SCI-1228), and Tennessee Wildlife Resources Agency (TWRA-3716 and TWRA-3742). Samples provided by other researchers were also collected as part of appropriately approved and permitted activities.

We extracted DNA from tissue samples using DNeasy Blood and Tissue Kits (Qiagen, Germantown, MD), with overnight incubation in proteinase K, and from guano samples using PowerSoil DNA Isolation Kits (Mo Bio Laboratories, Carlsbad, CA). We amplified DNA from both tissue and guano by polymerase chain reaction (PCR) using the Mysp 1 and 2 primers from the 16S ribosomal subunit (Zinck et al. 2004). The 25-μL reactions consisted of 1× PCR gold buffer, 2.5 mM MgCl2, 0.8 mM dNTP blend, 0.125 U AmpliTaq Gold (Applied Biosystems, Foster City, CA), 5 μg BSA (Sigma-Aldrich, St. Louis, MO), 5 μM of each primer (i.e., Mysp 1 and Mysp 2; Integrated DNA Technologies, Coralville, IA), and 5 μL of fecal DNA or 2.5 μL of genomic DNA. Cycling parameters for PCR involved denaturation at 95°C for 10 min, followed by 35 cycles at 95°C for 30 s, annealing for 30 s, and 72°C for 30 s, followed by a final elongation step of 10 min at 72°C. We first amplified samples with an annealing temperature of 55°C. We reamplified at 48°C any samples that did not amplify or amplified weakly. We confirmed positive PCR results on a 2% agarose gel and purified with ExoSap-It (USB Corporation, Cleveland, OH), before sequencing bidirectionally via Sanger sequencing on an Applied Biosystems 3730 (Applied Biosystems) at the University of Tennessee Genomics Core. We aligned and compared sequences using Sequencher v5.0.1 software (GeneCodes, Ann Arbor, MI) to create a reference database with which to compare unknown sequences. We analyzed sequences in DnaSP (v5; Librado and Rozas 2009) to determine intraspecific nucleotide diversity (P_I), as well as nucleotide diversity within genera, when applicable.

We used the reference database to identify bat species from DNA extracted from guano samples collected as part of a study examining roost selection by bats in GRSM (Fagan et al. 2017). To assess bat occurrence in buildings, researchers at the University of Tennessee surveyed 140 buildings in Tennessee and North Carolina three times between 15 May and 14 August 2015, with approximately 1 mo between sampling periods. Researchers cleared buildings of pre-existing guano prior to the beginning of the first sampling period, except when large accumulations made complete removal infeasible. In the case of accumulations that were infeasible to remove entirely, researchers cleared an area of approximately 500 cm². Therefore, all accumulations encountered during the study period were no more than 1 mo old. Researchers encountered 90 guano accumulations in 37 buildings throughout the study period, and collected samples of 1–5 guano pellets from each accumulation. Researchers stored all samples in silica gel desiccant, placed them in a cooler immediately following collection, and froze them at −20°C within 48 h. We extracted and sequenced DNA from one individual pellet from each sample using the methods described above. We compared the resulting DNA sequences to our reference database in Sequencher v5.0.1 in order to infer bat species occurrence in GRSM buildings.

To create the reference database, we sequenced DNA from 62 tissue samples and guano samples from 37 captured bats, representing all 16 species of bat known to occur in Tennessee. We amplified DNA from all 16 species of bats included in this study using one set of primers with one of two PCR cycling conditions (Table 1). We distinguished DNA sequences to species, including all six Myotis species known to occur in the state (Data S1, Supplemental Material). The DNA sequences ranged in length from approximately 175 to 185 base pairs, with gray myotis M. grisescens producing a sequence approximately 10 base pairs longer than all other species (Table 1). Intraspecific nucleotide diversity ranged from 0.00535 to 0.01724, while diversity between species within each genus ranged from 0.05172 to 0.20455 (Table 2). Gray myotis had the highest minimum intrageneric nucleotide diversity, owing to a unique insert region.

Of the 90 DNA samples extracted from individual guano pellets from accumulations in GRSM buildings, 68 (75.6% of samples) amplified successfully. These DNA samples represented 34 of the 37 buildings where guano samples were collected. Guano accumulations were the only indicator of bat occurrence at 19 of these buildings (55.9%). Some of the fecal samples from which DNA was extracted could have been from rodents, potentially explaining the unamplified samples. We identified the DNA extracted from the 68 pellets that were successfully amplified and sequenced to five bat species: big brown bat (31 pellets), eastern small-footed myotis (20 pellets), little brown myotis (11 pellets), Rafinesque's big-eared bat (5 pellets), and northern long-eared myotis (1 pellet). Detection of little brown myotis and northern long-eared myotis in GRSM buildings was only possible through genetic analysis of DNA extracted from guano because these species were not observed in buildings. Through this guano analysis, we also found three buildings were used by more than one species. At two buildings where Rafinesque's big-eared bat roost sites were confirmed visually, we identified additional big brown bat roosts using DNA from collected guano. One building, where no bats were observed, contained roosts used by both eastern small-footed myotis and northern long-eared myotis.

We applied the genetic marker from the 16S ribosomal subunit, mtDNA (Zinck et al. 2004), to DNA extracted from bat tissue and guano to create a reference database for the 16 species of bats known to occur in Tennessee (Harvey et al. 2011). Zinck et al. (2004) required two markers spanning different parts of the 16S gene to distinguish between two sets of species pairs in their study area (i.e., Myotis californicus from M. ciliolabrum, and M. thysanodes from M. evotis). These species are not found in the southeastern United States. On account of the sequence variation of the species in our study area, we required only one marker to distinguish between species.

In light of population declines as a result of WNS (Frick et al. 2010), reliance upon visual and acoustic detection of bats may limit the effectiveness of monitoring and research efforts. Using our 16S DNA sequence database for the 16 species of bats found in Tennessee, we inferred species occurrence at 34 of 39 roost buildings identified by researchers in GRSM. Bats were not observed at 19 (55.9%) of these buildings, and therefore analysis of DNA extracted from guano was the only option for species identification. We not only identified single-species roosts from guano, but also multiple-species roosts. Had bats been visually detected, identification of multiple species may have been difficult because of slight morphological differences between species observed at a distance (Harvey et al. 2011). Therefore, genetic identification using DNA extracted from guano may still have been needed.

The little brown myotis and northern long-eared myotis, two species identified through our case study, are of special concern as a result of declines from WNS (Frick et al. 2010; Langwig et al. 2012). These species were not observed roosting in buildings during surveys in 2015 or 2016 (Fagan et al. 2017). The little brown myotis, identified in this study through the analysis of 11 guano pellets, was commonly captured and documented roosting in buildings before WNS was documented in GRSM (Harvey 1987; E.R. Britzke, U.S. Army Engineer Research and Development Center, unpublished data). In contrast, there were no records of the northern long-eared myotis, identified in this study through analysis of a single guano pellet, roosting in GRSM buildings prior to 2015. By confirming the occurrence of these species in certain buildings, GRSM managers are able to make informed management decisions that balance building maintenance and accessibility with bat conservation and management. These results have important applications for studies monitoring species that are cryptic or infrequently encountered, as with many species of conservation need.

The two markers identified by Zinck et al. (2004) were recently applied to DNA from mixed guano samples collected in Canada and the northern United States using microarray technology, allowing for the genetic identification of 34 bat species (Patrick et al. 2016). The sequences we present, along with those determined by Patrick et al. (2016) and Zinck et al. (2004), create a large 16S mtDNA sequence database that has broad application for bat-related research. This database and the genetic analysis techniques we employed will be useful to identify bat species in areas where, as a result of WNS-related declines or the cryptic nature of many bat species, roost surveys, mist netting, or acoustic monitoring are ineffective. In addition, these methods may be applied when species identification is limited by biologist inexperience or similarities in bat morphology and call characteristics.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. The mitochondrial 16S ribosomal subunit sequences produced in the development of a DNA database for 16 species of bats known to occur in Tennessee, in fasta format. Samples are labeled with their Genbank Accession numbers, as well as scientific name followed by a letter indicating different sequences within each species. All samples were collected and processed between 2010 and 2016.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S1 (146 KB DOCX).

Reference S1. Harvey MJ. 1987. Bat utilization of historic structures in the Cataloochee Historic Zone and of mines in the Sugar Fork and Eagle Creek Watersheds, Great Smoky Mountains National Park. Report to the National Park Service, Great Smoky Mountains National Park.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S2 (66 KB PDF).

Reference S2. Niver RA, King RA, Armstrong MP, Ford WM. 2014. Methods to evaluate and develop minimum recommended summer survey effort for Indiana bats: white paper. U.S. Fish and Wildlife Service White Paper.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S3 (347 KB PDF); also available at https://www.fws.gov/Midwest/endangered/mammals/inba/surveys/pdf/WhitePaperIBatSurveyEffort13Jan2014.pdf (347 KB pdf).

Reference S3. Tuttle MD. 2003. Estimating population size of hibernating bats in caves and mines. Pages 31–39 in O'Shea TJ, Bogan MA, editors. Monitoring trends in bat populations of the United States and Territories: problems and prospects. Fort Collins, Colorado: U.S. Geological Survey, Biological Resources Discipline. Information and Technology Report USGS/BRD/ITR–2003–0003.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S4; also available at https://www.fort.usgs.gov/publication/21329 (9,544 KB PDF).

Reference S4. U.S. Fish and Wildlife Service [USFWS]. 2012. North America bat death toll exceeds 5.5 million from white-nose syndrome. Arlington, Virginia: U.S. Fish and Wildlife Service. News Release, January 17, 2012.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S5; also available at https://www.whitenosesyndrome.org/sites/default/files/files/wns_mortality_2012_nr_final_0.pdf (109 KB PDF).

Reference S5. U.S. Fish and Wildlife Service [USFWS]. 2017. Bats affected by WNS. Arlington, Virginia: U.S. Fish and Wildlife Service. Whitenosesyndrome.org, July 21, 2017.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S6; also available at https://www.whitenosesyndrome.org/about/bats-affected-wns (143 KB PDF).

Reference S6. U.S. Geological Survey [USGS]. 2017. Alabama survey finds first southeastern bat with white-nose syndrome. Fort Collins, Colorado: U.S. Geological Survey. News Release, June 1, 2017.

Found at DOI: http://dx.doi.org/10.3996/012017-JFWM-007.S7; also available at https://www.usgs.gov/news/alabama-survey-finds-first-southeastern-bat-white-nose-syndrome (335 KB PDF).

We thank everyone who provided DNA samples, including Tim Divoll, Maarten Vonhof, Amy Russell, Jennifer Krauel, and Grace Carpenter. We also thank Gary McCracken, Department of Ecology and Evolutionary Biology at the University of Tennessee, for the use of his genetics lab and the Journal reviewers and the Associate Editor who handled reviews of our paper. Funding for the project was provided by a National Park Service White-nose syndrome Grant.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.