Amphibian Fungal Disease and the Cutaneous Microbiome

AMPHIBIAN FUNGAL DISEASE AND THE CUTANEOUS MICROBIOME

A Thesis

Submitted to the Faculty

Drexel University

Patrick Joseph McLaughlin

in partial fulfillment of the

requirements for the degree

Doctor of Philosophy

June 2015

iii

DEDICATIONS

To my mother, who taught me more than all the textbooks and professors combined,

especially about how to be a good person. To my father, who instilled in me his deep

appreciation for nature, which has dictated my path ever since. To my brother, who has

always been there to lean on, look up to, and save my ass. To my sister, who constantly

reminds me what is truly important in life. To my Philadelphia friends for all the love, support, and beer that got me through. Lastly, to the frogs, I am sorry I had to kill some of

you, I did not enjoy it, it was for science… forgive me.

ACKNOWLEDGEMENTS

This work was supported by Drexel University, ExxonMobil Foundation, Mobil

Equatorial Guinea Incorporated, la Universidad Nacional de Guinea Ecuatorial, The Explorers

Club, U.S. Fish & Wildlife Service, Hudson Valley Herpetological, North Carolina Museum of

Natural Sciences, and the Bioko Biodiversity Protection Program. J. Rosado (Museum of

Comparative Zoology, Harvard University) and J. Friel (Cornell University Museum of

Vertebrates) loaned specimens in their care. Research and specimen export permits were provided by the Ministerio de Información, Cultura y Turismo, Republica de Guinea Ecuatorial, and the Facultad de Medio Ambiente. T. Butynski, G. R. Mene, A. Fertig, M. Cuzzocreo, K.

Fugett, A. Byrne, N. Stanek, R. Sicheneder, R. Bell, B. Stuart, J. Owens, D. Cronin, D. B. Mene,

P. Łukasik, N. Wong, S. Woloszynek, and D. Rock assisted with fieldwork, labwork, statistics, and/or general improvements to this thesis. Special thanks to my dissertation committee of G.

Hearn, M. O’Connor, J. Russell, D. Blackburn, and S. Honarvar who each contributed essential guidance throughout all facets of this research.

TABLE OF CONTENTS

LIST OF TABLES ...... viii

LIST OF FIGURES ...... ix

ABSTRACT ...... x

1. BIOKO ISLAND AMPHIBANS ...... 1

1.1 Arthroleptis bioko ...... 6

1.2 Wolterstorffina parvipalmata ...... 7

1.3 Didynamipus sjostedti ...... 8

1.4 Phlyctimantis boulengeri ...... 11

2. A NEW LEPTOPELIS (AMPHIBIA: ARTHROLEPTIDAE) FROM BIOKO ISLAND, EQUATORIAL GUINEA ...... 12

2.1 Abstract ...... 12

2.2 Introduction ...... 12

2.3 Materials and Methods ...... 14

2.4 Species Description ...... 16

2.4.1 Holotype ...... 16

2.4.2 Paratypes ...... 16

2.4.3 Referred Materials ...... 18

2.4.4 Diagnosis ...... 18

2.4.5 Description of Holotype ...... 20

2.4.6 Coloration in Preservative ...... 22 vi

2.4.7 Coloration in Life ...... 23

2.4.8 Variation ...... 23

2.4.9 Molecular Divergence ...... 23

2.4.10 Distribution and Natural History ...... 23

2.4.11 Etymology ...... 24

2.5 Discussion ...... 24

3. PREVALENCE OF THE AMPHIBIAN CHYTRID PATHOGEN ( Batrachochytrium dendrobatidis ) ON BIOKO ISLAND, EQUATORIAL GUINEA ...... 27

3.1 Introduction ...... 27

3.2 Origin and Spread of Chytridiomycosis ...... 27

3.3 Materials and Methods ...... 34

3.4 Results ...... 37

3.5 Discussion ...... 44

4. CHARACTERIZATION OF THE AMPHIBIAN CUTANEOUS MICROBIOME ...... 47

4.1 Introduction ...... 47

4.2 Materials and Methods ...... 49

4.2.1 Processing Swabs ...... 52

4.2.2 Statistical Analysis ...... 54

4.3 Results ...... 56

4.3.1 Metabolites: Adonis and Principle Coordinates Analysis (PCoA) ...... 56

4.3.2 Bacteria (97% OTU Composition): Adonis and Principle Coordinates Analysis (PCoA) ..60 vii

4.4 Discussion ...... 65

4.5 Further Analysis ...... 66

5. LIST OF REFERENCES ...... 67

6. APPENDIX 1: Arthroleptis bioko IUCN Red List Assessment ...... 76

7. APPENDIX 2: Specimens Examined ...... 79

8. VITA ...... 80

viii

LIST OF TABLES

1. Amphibian Species Inventory for Bioko Island, Equatorial Guinea ...... 5

2. Measurements of Adult Types of Leptopelis pictilis sp. nov...... 19

3. Distribution of Bd Across Sampled Amphibian Species ...... 38

4. Distribution of Bd During Wet and Dry Seasons Across Sampling Localities ...... 39

5. Amphibian Taxa Sampled for Cutaneous Bacteria and Metabolites ...... 51

6. Genus:Species Abbreviations for Sample ID ...... 55

LIST OF FIGURES

1. Map of Bioko Island, Equatorial Guinea ...... 4

2. Range Map for Didynamipus sjostedti in Pico Basilé National Park ...... 10

3. Sampling Localities for Leptopelis pictilis ...... 17

4. Leptopelis pictilis Holotype ...... 21

5. Bd Sampling Localities ...... 35

6. Plot from Mixed Effects Logistic Regression Model of Bd Prevalence Data ...... 41

7. Bd Infection Load Across Amphibian Families and Season ...... 42

8. Sampling Localities for Amphibian Cutaneous Bacteria and Metabolites ...... 50

9. PCoA of Cutaneous Metabolites from 14 Amphibian Species ...... 57

10. PCoA of Cutaneous Metabolites from 12 Species:Locality Combinations ...... 59

11. Relative Abundance of 12 Most Common Bacterial Taxa (97% OTU) ...... 61

12. PCoA of Relative Abundance for 12 Most Common Bacterial Taxa (97% OTU) ...... 62

ABSTRACT

Amphibian Fungal Disease and the Cutaneous Microbiome Patrick J. McLaughlin Advisor: Gail W. Hearn, Ph.D.

The International Union for Conservation of Nature (IUCN) estimates that over 40% of all known amphibians are currently at risk of extinction due to the cumulative effects of climate change, pollution, habitat loss, and disease. Tropical ecosystems may experience the greatest decline, likely threatening to push many species to extinction that we have not yet discovered.

Amphibians are not well sampled in many areas of Central Africa, which contributes to the difficulty in resolving historical biogeography and phylogenetic relationships. Species distribution information is also essential in developing conservation priorities, particularly in identifying threats to restricted range species or those that may be affected by habitat destruction and/or climate change. Among the most urgent threats to amphibians is the fungal disease chytridiomycosis, which is already assumed to have caused the extinction of many species.

This research sought to conduct an exhaustive inventory of amphibian fauna across Bioko

Island, Equatorial Guinea. Bioko is adjacent to the Guinean Forests, one of the most amphibian species-rich areas in Africa. Over the course of 5 years (2009-2014) these surveys resulted in both new records and new species for Bioko Island, identified species of concern, and refined our knowledge of amphibian species distributions on the island. During this work, amphibians were also sampled for the presence of Batrachochytrium dendrobatidis (Bd ) to determine the nature of chytridiomycosis infection across a wide range of Central African amphibian species.

Results showing positive Bd yet no mortality led us to investigate the cutaneous microbiome, which has been shown to harbor anti-fungal properties in some species. xi

Metabolites and bacteria were sampled via specific swabbing techniques in an attempt to accurately characterize the cutaneous environment. Recently developed techniques for metabolite extraction and HPLC analysis allowed us to identify over 100 distinct compounds from specially prepared foam tipped metabolite swabs. DNA extractions and qPCR analysis performed on bacterial swabs allowed us to quantify the relative abundance of the most common bacteria species found on the amphibian skin. Together, these results show that amphibians tend to have metabolite and bacterial profiles that match other individuals of their species. This suggests the microbiome may be specifically regulated by amphibians, and calls for continued research into the potentially symbiotic role that cutaneous bacteria may play in disease resistance for amphibians.

CHAPTER 1: BIOKO ISLAND AMPHIBIANS

Bioko Island is located 37-km from the Cameroon coast in the Gulf of Guinea, and can be considered a part of the Guinean Forests hotspot, an area characterized by high species richness, endemicity, and significant threats to habitat and biodiversity (Oates et al. 2004). The average rainfall on Bioko’s Southern coast is among the highest recorded in Africa (10,000 mm) and contributes to the diverse and regionally-unique habitat types found across Bioko’s rugged elevation gradient (Oates et al. 2004). The highest point on Bioko is Pico Basilé (3,011 m) and occurs within 10 km of coast, resulting in a steep topography across much of Bioko’s landmass

(2,017 km2). High rainfall, variable landscape, and equatorial temperatures sustain high species richness, particularly with regard to amphibians. Seasonality on Bioko is defined primarily by differences in rainfall throughout the year, as temperatures on the equator do not fluctuate widely

(Oates et al. 2004). The wet season on Bioko occurs from May to October, while the dry season, which still may experience daily rainfall events, occurs from November through February (Oates et al. 2004). Bioko is regarded as a conservation priority in this region of Africa due to its unique landscape harboring a wide range of habitat types and biodiversity that rivals that of the mainland where more widespread habitat destruction has already occurred (Myers et al. 2000).

Amphibians are among the taxa that are under-sampled in West-Central Africa, which, in addition to the cryptic diversity in many amphibian species-complexes, contributes to the difficulty in resolving historical biogeography and phylogenetic relationships for species in the region (Anthony et al. 2015). Species distribution information is important for setting conservation priorities, particularly in identifying threats to restricted range species or those that may be affected by habitat destruction and/or climate change. Many parts of Equatorial Guinea’s 2

Bioko Island remained unexplored by herpetologists until the mid-1990’s (Schiøtz 1999).

Historically, the Equatoguinean government’s distrust of foreign scientists prevented exhaustive inventories of herpetofauna in many areas of mainland Equatorial Guinea and Bioko Island

(Oates et al. 2004). Much of Bioko’s amphibian species inventory has been pieced together from a combination of recent, short-term, and geographically restricted surveys, along with historic records whose taxonomic and geographic accuracy have often come into question. Among the first attempts to survey amphibian diversity for Bioko Island was by Eisentraut (1973) who reported 32 amphibian species.

Oates et al. (2004) suggests the apparent lack of regional endemics reported to exist on

Bioko is a result of limited sampling. Currently, Arthroleptis bioko (Blackburn 2010) and

Leptopelis brevipes (Boulenger 1906) are the only endemic amphibians reported from Bioko

Island, however the latter is widely considered a likely synonym of Leptopelis brevirostris (Frost

1985; Lötters et al. 2005). While the lack of regional endemics on Bioko may be due to a sampling bias, there are likely few species that are endemic only to Bioko because, unlike oceanic islands with higher levels of endemism, Bioko is a continental island that was connected to the mainland as recently as 10,000 years ago (Jones 1994). Despite its lack of endemism,

Bioko does harbor several genera that are unique to the Lower Guinean Forest Zone of Central

Africa, including Didynamipus sjostedti and Wolterstorffina parvipalmata . These two species are threatened throughout their limited range elsewhere in the Guinean Forest hotspot, and their genera are of relatively ancient evolutionary origin, lending further support for Bioko as a priority for conservation in the region (Oates et al. 2004).

From 1910 to 1960, the rate of new species descriptions per year from Africa and South

America were equal (Schiøtz 1999). That rate almost doubled for South America in the 1970’s 3 due to an increase in the number of trained, resident biologists/naturalists working on the continent (Schiøtz 1999). This highlights the vital importance of continued inventories to document amphibian diversity in less-explored areas of West and Central Africa, and the need for visiting biologists/naturalists to invest in education and training of the next generation of

African field biologists. The Bioko Biodiversity Protection Program (BBPP) has focused on research, conservation, and education on Bioko Island since 1999. A central focus of their work has been helping to train students from the La Universidad Nacional de Guinea Equatorial

(UNGE) in ecological research to aid conservation. Over the past 15 years the BBPP has developed an infrastructure that has allowed scientists to explore and study the Gran Caldera and

Southern Highlands Scientific Reserve, thereby instituting the capacity for widespread and sustained surveys and sampling across the island. For the purposes of this work and in collaboration with BBPP, P. McLaughlin began targeted amphibian surveys (2009 ̶ 2013) across

Bioko to update the island’s amphibian inventory and determine species range information. The ultimate goal of these surveys was twofold: to identify rare, threatened, or endangered amphibian populations on Bioko Island and to document previously unknown amphibian diversity on the island via new records or new species.

These surveys represent the most comprehensive and accurate amphibian species inventory for Bioko Island, though there is undoubtedly additional diversity yet to discover

(Table 1.). Fieldwork was conducted under the direction of BBPP and in collaboration with students from UNGE and various visiting scientists. While the eventual goal is to complete an exhaustive field guide for the amphibians of Bioko, herein we provide brief accounts of species that are of particular interest due to their threatened status and/or rarity. Each assessment aims to provide a brief assessment of the species’ natural history and conservation status on the island. 4

Figure 1. Bioko Island, Equatorial Guinea. Sample site names derived from research camps operated by the Bioko Biodiversity Protection Program, and using Bioko Island map produced by Talleres del Instituto Geograficio Nacional – Deposito Legal: Madrid 7955 – 1983.

Table 1. Amphibian species inventory for Bioko Island, Equatorial Guinea Family Genus Species IUCN Status Arthroleptidae Arthroleptis adelphus Least Concern bioko Data Deficient poecilonotus Least Concern sylvaticus Least Concern taeniatus Least Concern variabilis Least Concern Cardioglossa leucomystax Least Concern nigromaculata Near Threatened Leptopelis brevirostris Least Concern calcaratus Least Concern millsoni Least Concern modestus Least Concern pictilis sp. nov. - rufus Least Concern Bufonidae Amietophrynus camerunensis poensis Least Concern gracilipes Least Concern tuberosus Least Concern Didynamipus sjostedti Endangered Nectophryne afra Least Concern Wolterstorffina parvipalmata Vulnerable Dermophiidae Geotrypetes seraphini Least Concern Herpelidae Herpele squalostoma Least Concern Hyperoliidae Afrixalus laevis Least Concern paradorsalis Least Concern Hyperolius endjami Vulnerable ocellatus Least Concern tuberculatus Least Concern Opisthothylax immaculatus Least Concern Phlyctimantis boulengeri Least Concern Petropedetidae Petropedetes johnstoni Near Threatened Phrynobatrachidae Phrynobatrachus auritus Least Concern calcaratus Least Concern Pipidae Xenopus fraseri Least Concern sp. nov. new tetraploid 2 - Ptychadenidae Ptychadena aequiplicata Least Concern Pyxicephalidae Aubria subsigillata Least Concern Ranidae Hylarana albolabris Least Concern Rhacophoridae Chiromantis rufescens Least Concern 12 Families 20 Genera 38 Species

Arthroleptis bioko

Eight specimens of potential Arthroleptis bioko have been collected, but as of yet only three of these have been absolutely identified as A. bioko through diagnostic comparison with

Arthroleptis variabilis . There are some subtle differences between A. bioko, A. variabilis, and the species complex of A. poecilonotus (which likely represents more than one species), and further inspection and possible genetic analysis will be needed to confirm the true ID of the five remaining specimens. While surveys have specifically targeted this endemic species, these eight specimens are the only individuals found in the past five years. A. bioko does not seem to exist in high densities within our collection areas, far less than the other comparable species, so we would characterize them as “rarely recorded.” It is possible with further morphological examination we will identify some existing vouchers specimens from the southern region of the island (Grand Caldera & Southern Highlands) as A. bioko , which could expand their known range and elevation, but such identifications are currently tentative at best.

Holotype specimens were collected in a region and at an elevation that has limited sources of flowing water. The few A. bioko specimens collected from Moka were within 100 m of streams. More focused research is needed to provide accurate information on their association with streams. They are likely to have similar life history characteristics and habitat preferences to those of A. variabilis , which is a leaf-litter species that is reported to occur in proximity to 7 flowing water. In light of the increasing habitat destruction that threatens known populations of

A. bioko , and the possibility that they are likely a restricted range montane species susceptible to potential climate change, we would fully support their categorization by the IUCN as Near

Threatened (Appendix 1). Per this fieldwork, an updated conservation assessment for A. bioko has been prepared and submitted to the International Union for Conservation of Nature (IUCN)

Red List of threatened species (Appendix A: Arthroleptis bioko Red List Assessment).

Wolterstorffina parvipalmata

The BBPP recently collected some additional specimens of Wolterstorffina parvipalmata.

They were all closely associated with a single streamside habitat, perched on leaves about 1.5 m high above a small stream (1-2 m wide). These were collected in the vicinity of Moka, near previous collections in 2012, however those previous encounters were much closer to heavily disturbed agriculture areas/fields. The IUCN Red List entry does not list W. parvipalmata as occurring on Bioko, and this should be updated per our surveys over the past two years that reconfirm past accounts from . It is possible future genetic analysis currently being undertaken by our collaborator Michael Barej (Museum für Naturkunde, Berlin, Germany) may reveal the W. parvipalmata from Bioko to be a unique species. If this holds true, it will represent the second unequivocal endemic Bioko amphibian. Furthermore, it would mean their known range on Bioko 8 is extremely limited, and under heavy threat from development, though additional surveys specifically targeting this species in the future may help to better define and expand their extent of occurrence. Given our limited encounters in recent years despite relatively high sampling intensity around Moka, along with a complete absence of records from other areas on the island before 2012, we would recommend it be listed as Endangered or Vulnerable if it is later found to represent a new species.

Didynamipus sjostedti

Recent surveys (September 2012) carried out by P. McLaughlin and UNGE students re- confirmed the presence of D. sjostedti near the town of Moeri. Our sampling locality is very likely the same region from which previous observations were made, as it represents the only area accessible to researchers on Pico Basilé that contains the necessary habitat requirements

(IUCN 2000). We found clusters of individuals (approximately 3-4 per m²) along moist lowland forest, amongst dead wet leaves and vegetation. Much of this follows the IUCN habitat description, however we were in fairly dense forest, not forest edge or clearings as is stated in that description. We encountered these clusters of individuals abruptly along our trail, and their presence ended just as abruptly, and it seemed during repeated census they were restricted to this narrow range along the trail as it climbed the southern slopes of Pico Basilé. This would support 9 the idea that they are patchily distributed, and thus likely increases threats of fragmentation (ie. - road construction on Bioko). This work was done as part of fieldwork funded by a US Fish &

Wildlife Service amphibian grant.

We have established an estimated range based on locality and elevation in PBNP for D. sjostedti (Fig. 2). Since the habitat type between the low and mid elevation camps did not change dramatically, we are unable to comment on the possible microhabitat characteristics that dictate their patchy distribution within this lowland forest. Owing to the fact their presence seemed to abruptly start and end between specific high and low points, we were able to estimate a high and low elevation range and project it onto Bioko Island. The map shows the documented and potential range of occurrence for this species within the survey region.

For the same reasons cited for A. bioko, I believe Bioko populations are at high risk from development. Additionally, their range on Bioko is extremely small, far smaller than what we know for A. bioko . The current status of populations in Cameroon or Nigeria is unclear. Based on future surveys of the Cameroon population, and given the threats to its range on Bioko, it is highly likely that the classification for D. sjostedti should qualify for one of the higher risk designations above its current Threatened status.

Figure 2. Potential range map for Didynamipus sjoestedti in Pico Basilé National Park (PBNP). Surveys were conducted on the Southern slopes of Pico Basilé. Survey trail is shown with three established camps. Yellow points show the highest (1016 m) and lowest (865 m) elevations for which D. sjoestedti was recorded. Potential range was extrapolated to the East and West of the survey trail based on these high and low points of presence.

Phlyctimantis boulengeri

Only one specimen of Phlyctimantis boulengeri has been collected from Bioko since

2009. This encounter could prove a useful update to the IUCN report of this species from Bioko.

Based on intense sampling effort over the past 5 years, and this being our first encounter, we would categorize this species as rarely encountered on Bioko. This could suggest Bioko populations are severely reduced and at risk of local extirpation, given reports that in other areas where it occurs, it is at high densities. It is unclear what might be preventing this species from a wider distribution on Bioko, but the intense sampling pressure in the region where it was found

(Moeri) suggests this species may be limited in range and/or possibly declining on Bioko.

CHAPTER 2: A NEW LEPTOPELIS (AMPHIBIA: ARTHROLEPTIDAE) FROM BIOKO ISLAND, EQUATORIAL GUINEA

ABSTRACT

A new species of treefrog (genus Leptopelis ) is described from the Gran Caldera and

Southern Highlands Scientific Reserve on Bioko Island, Equatorial Guinea. This new species represents the second endemic amphibian described from this land-bridge island located off the coast of Cameroon in the Gulf of Guinea, and the 54th species in the genus Leptopelis . The new species is distinguishable from all other Leptopelis by having a green dorsum with large maroon spots, and from all species of Leptopelis except L. aubryioides and L. calcaratus by having a dermal spur on the tibiotarsal joint. It is also the only Leptopelis species known to produce distinctive orange eggs. The discovery of a new, large-bodied and conspicuously colored species underscores the extent to which Bioko Island’s herpetofauna and capacity for endemism have been overlooked as an ecologically and evolutionarily important part of the Biafran Forests and

Highlands and the Guinean Forests Biodiversity Hotspot.

INTRODUCTION

THE CAMEROON Volcanic Line encompasses a geological area spanning the border region between Cameroon and Nigeria to the North, and extending south to include the four Gulf of

Guinea Islands: Bioko, Príncipe, São Tomé, and Annobón. This geologic formation, known collectively as the Cameroon Highlands or Biafran Forests and Highlands (BFH), is located within the Guinean Forests Biodiversity Hotspot, and harbors high levels of endemism, particularly with respect to amphibians (Bergl et al. 2007; Blackburn et al. 2010; Cronin et al.

2014). Elevational gradients support a variety of habitat types across the greater BFH, including montane forest, montane scrub, Schefflera forest, and subalpine grasslands that collectively 13 support high levels of biodiversity (Jones 1994; Cronin et al. 2014). Bioko Island is the largest of the Gulf of Guinea Islands and the only land-bridge island within this archipelago (Fig. 1). The island contains three peaks above 2000 m, and harbors a distinctive lowland monsoon forest-type on the southern coast (Guinea 1968) that receives an average annual rainfall of 10,000 mm

(Terán 1962), affording Bioko a wide variety of habitat types within a relatively small

(2017 km²) geographic area. Unlike the oceanic islands in the archipelago, Bioko shares much of its biodiversity with mainland Cameroon because global glacial cycles and retreating sea levels resulted in several periods of connectivity between Bioko and the continent (Jones 1994). This unique biogeographic history, coupled with its location within the BFH and Guinean Forests

Biodiversity Hotspot, make Bioko an extremely unique and important site for evolutionary biologists and conservationists alike (Jones 1994; Cronin et al. 2014).

Due to historical periods of connectivity with the mainland, Bioko hosts over 34 species of amphibians and less than 5% of this diversity is endemic (Jones 1994; Blackburn 2010).

Although overseas dispersal by amphibians is possible (Vences et al. 2003; Measey et al. 2007;

Maddock et al. 2014; Bell et al. 2015), marine incursions have likely restricted amphibian dispersal between Bioko and the continent periodically over the island’s history. Therefore, although there has been little focus on the potential endemism on Bioko due to its proximity to the mainland (37 km) and recent separation (the most recent incursions date to ~12,000 years), the diversity and endemism of Bioko’s herpetofauna is likely underrepresented in the scientific literature (Blackburn 2010). In general, sub-Saharan African anuran species have been difficult to accurately distinguish due to cryptic species complexes, overlapping species descriptions, museum specimen identification errors, and a historic lack of sampling from this region of Africa, 14 the latter being particularly true of Bioko (Barej et al. 2010; Blackburn & Rödel 2011; Portillo &

Greenbaum 2014).

The genus Leptopelis (Arthroleptidae) contains 53 species of medium to large bodied treefrogs distributed across sub-Saharan Africa (Schiøtz 1999; Portillo & Greenbaum 2014).

Members of this genus are readily distinguished by having large eyes, larger body size relative to other arboreal genera of African anurans, and distinctive calls (Schiøtz 1999; Amiet 2012;

Portillo & Greenbaum 2014). The genus contains many cryptic species due to morphological similarity between sympatric species (Schiøtz 1999; Portillo & Greenbaum 2014). Some species of Leptopelis can also be difficult to diagnose owing to insufficient numbers of specimens in museum collections, as many of the species are arboreal and do not form breeding aggregations

(Schiøtz 1999).

Herein, we describe a new species of Leptopelis from high elevations on Bioko Island

(1000–1500 m asl) that, unlike many of its congeners, is morphologically distinctive. The new species represents the second amphibian that is endemic to Bioko Island, and the 54th described species of the genus.

MATERIALS AND METHODS

PJM, RCB, and colleagues collected specimens during March 2008, January ̶ February

2010, August 2011, February 2012, and January 2013. All amphibians were collected following

Drexel IACUC Protocol #18748. Specimens were fixed in 10% buffered formalin after preserving liver in 95% ethanol or RNAlater (Ambion), and later transferred to 70% ethanol.

Specimens and tissue samples were deposited at the North Carolina Museum of Natural Sciences

(NCSM) and Cornell University Museum of Vertebrates (CUMV). Comparative material was examined in the holdings of these institutions and Museum of Comparative Zoology, Harvard 15

University (MCZ; Appendix 2). Additional data on L. aubryioides and L. calcaratus were taken from Köhler (2009) and Amiet (2012).

Measurements were taken to the nearest 0.1 mm with dial calipers: snout-vent length

(SVL); head length from tip of snout to rear of jaws (HDL); maximum head width (HDW); snout length from tip of snout to anterior corner of eye (SNT); eye diameter (EYE); interorbital distance (IOD); horizontal diameter of tympanum (TMP); internarial distance (IND); shank length (SHK); thigh length (TGH); forearm length, from elbow to base of thenar tubercle (LAL); manus length from tip of third digit to base of thenar tubercle (HND); pes length from tip of fourth toe to base of inner metatarsal tubercle (FTL).

We extracted total genomic DNA using a DNeasy Blood & Tissue Kit (Qiagen Inc.,

Valencia, CA, USA) and polymerase chain reaction (PCR)-amplified and sequenced 495 bases of the mitochondrial fragment 16s from a paratype of the new species described below (CUMV

15907) using published primers (16s A-L 5’ CGC CTG TTT ATC AAA AAC AT 3’ and 16s B-

H 5’ CCC GTC TGA ACT CAG ATC ACG T 3’; Palumbi et al. 1991). The PCR was carried out in a final volume of 20µL containing: 20 ng template DNA, 1× Buffer, 0.2 µ M of each primer,

0.4 m M dNTP mix, and 0.125 units of Taq DNA polymerase (Roche Diagnostics, Indianapolis,

IN, USA). Amplification was carried out with an initial denaturation for 5 min at 94 °C, followed by 35 cycles (60 s denaturation at 94 °C, 60 s annealing at 55°C, 60 s extension at 72 °C), and a final extension at 72 °C for 5 min. The PCR product was purified using ExoSAP-IT (USB Corp.,

Cleveland, OH, USA), and sequenced using a BigDye Terminator Cycle Sequencing Kit v.3.1

(Applied Biosystems, Foster City, CA, USA) on an ABI Automated 3730xl Genetic Analyzer

(Applied Biosystems). The DNA sequence was edited using SEQUENCHER 5.0.1 (Gene Codes

Corp., Ann Arbor, MI, USA) and is accessioned in GenBank. A homologous sequence from a 16 specimen of L. calcaratus collected in Littoral Province, Cameroon (MCZ A-138028) was downloaded from GenBank (JQ715684.1) and we calculated uncorrected pairwise distance between the sequences using PAUP* 4.0a134 (Swofford 2002). No sequence data for L. aubryioides are currently available.

SPECIES DESCRIPTION

Leptopelis pictilis sp. nov.

Bioko Painted Treefrog, Rana Arborícola Pintada de Bioko

(Figs. 1, 2, Table 1)

Holotype —NCSM 82300 (field no. RCB 0005), adult female, Republic of Equatorial

Guinea, Bioko Sur Province, Gran Caldera & Southern Highlands Scientific Reserve, Pico Biao,

3.36098ºN, 8.66223ºE, 1376 m elev., coll. Rayna C. Bell, Patrick J. McLaughlin, and Andrew

Fertig, 19 August 2011.

Paratypes —NCSM 82298–99 (two adult males), same data as holotype except

3.36098ºN, 8.66223ºE, 1376 m elev., coll. Rayna C. Bell, Patrick J. McLaughlin, and Andrew

Fertig, 19 August 2011. NCSM 82301 (one adult male), same data as holotype except 3.36393ºN,

8.65797ºE, 1451 m elev., coll. Rayna C. Bell and Patrick J. McLaughlin, 22 August 2011.

NCSM 82310–11 (two adult males), same data as holotype except 3.32183ºN, 8.67125ºE, 1105 m elev., coll. Rayna C. Bell and Patrick J. McLaughlin, 31 August 2011. NCSM 82312 (one immature female), same data as holotype except 3.32438ºN, 8.67223ºE, 1140 m elev., coll.

Rayna C. Bell and Patrick J. McLaughlin, 31 August 2011. NCSM 82302–04 (three adult females), same data as holotype except 3.33270ºN, 8.66792ºE, 1272 m elev., coll. Rayna C. Bell and Patrick J. McLaughlin, 23 August 2011. NCSM 82309 (one adult female), NCSM 82305–08

(four adult males), same data as holotype except 3.32873ºN, 8.67377ºE, 1105 m elev., coll. 17

Figure 1. Distribution of sampling localities for Leptopelis pictilis on Bioko Island, Equatorial Guinea .

Rayna C. Bell and Patrick J. McLaughlin, 23 August 2011. NCSM 82313–14, CUMV 15907

(three adult females), Republic of Equatorial Guinea, Bioko Sur Province, Gran Caldera &

Southern Highlands Scientific Reserve, Pico Gran Caldera, 3.36533ºN, 8.50005ºE, 1073 m elev., coll. Patrick J. McLaughlin and Rachel Sicheneder, 1–2 February 2012 (NCSM 82313–14) and coll. Rayna C. Bell, Amancio Motobe, Anita Yantz, and Romeo Riaba, 16 January 2013 (CUMV

15907).

Referred Material —NCSM 82315 (one adult male), same data as holotype except

3.53562ºN, 8.67133ºE, 1130 m elev., coll. Jake Owens, Jessica Weinberg, and Tom Butynski, 1

March 2008. NCSM 82316 (one adult male), same data as holotype except 3.36133ºN,

8.66217ºE, 1341 m elev., coll. Kerry Fugett, Tom Butynski, and Cayetano Ebano, 30 January

2010. NCSM 82317 (one adult female), NCSM 82318 (one adult male), Republic of Equatorial

Guinea, Bioko Sur Province, Gran Caldera & Southern Highlands Scientific Reserve, Pico Gran

Caldera, 3.36173ºN, 8.49867ºE, 1043 m elev., coll. Patrick J. McLaughlin and Shaya Honarvar,

18 February 2010.

Diagnosis —A medium-sized, arboreal Leptopelis having a sharp canthus rostralis; tubercle (dermal spur) on tibiotarsal joint; no dermal fringe on outer arm; no ulnar tubercle; males with round pectoral glands; bifid distal subarticular tubercle only on Finger IV; green dorsum with large maroon spots; and orange eggs.

Leptopelis pictilis differs from all other Leptopelis by having a green dorsum with large maroon spots and orange eggs, and from all species of Leptopelis except L. aubryioides and L. calcaratus by having a dermal spur on the tibiotarsal joint. Leptopelis pictilis further differs from

L. aubryioides by having a sharp canthus rostralis (rounded in L. aubryioides ), lacking a distinct

Table 1. — Measurements (mm) of adult types of Leptopelis pictilis sp. nov. Abbreviations defined in the text.

Measurement Holotype female All females ( n = 8) Males ( n = 9)

NCSM 82300 Range; Mean ± SD Range; Mean ± SD

SVL 48.7 43.8–53.8; 49.0 ± 3.6 28.5–35.9; 33.2 ± 2.2

HDL 18.1 17.0–21.8; 18.8 ± 1.9 11.9–15.1; 13.6 ± 1.0

HDW 17.8 16.4–21.3; 18.3 ± 1.8 11.5–13.9; 13.0 ± 0.7

SNT 7.4 7.3–9.4; 8.0 ± 0.8 5.0–5.8; 5.4 ± 0.3

EYE 7.8 7.0–7.9; 7.5 ± 0.3 5.3–7.0; 6.0 ± 0.4

IOD 5.7 4.8–6.6; 5.8 ± 0.6 3.4–4.3; 3.9 ± 0.3

TMP 2.8 2.8–3.9; 3.3 ± 0.4 2.4–3.2; 2.7 ± 0.2

IND 4.5 4.5–5.6; 5.0 ± 0.4 3.3–3.9; 3.6 ± 0.2

SHK 26.7 25.2–31.5; 27.4 ± 2.1 16.9–20.3; 18.6 ± 1.0

TGH 25.1 21.0–30.3; 25.9 ± 3.0 15.7–18.7; 17.5 ± 0.8

LAL 9.6 9.0–11.5; 9.9 ± 0.9 6.0–6.9; 6.5 ± 0.3

HND 14.2 13.7–17.0; 15.1 ± 1.4 9.5–11.8; 10.3 ± 0.7

FTL 20.8 19.7–25.2; 21.9 ± 2.3 13.4–16.4; 14.8 ± 1.0 20 dermal fold on the outer arm (present in L. aubryioides ), lacking an ulnar tubercle (present in L. aubryioides ), and having a bifid distal subarticular tubercle only on Finger IV (bifid distal subarticular tubercles on all fingers in L. aubryioides ). Leptopelis pictilis also differs from L. calcaratus by usually lacking a distinct, contrasting, white blotch below the eye (always present in L. calcaratus ) and in the mitochondrial 16s gene (below).

Description of Holotype —Habitus moderately slender; head length subequal to width; snout short, obtusely pointed in dorsal view, round in profile; nostrils lateral, closer to tip of snout than eye, visible in dorsal view; canthis rostralis distinct, slightly constricted behind nostrils; lores concave, oblique; eye diameter subequal to snout length; interorbital distance greater than width of upper eyelid; pineal body not visible; tympanum distinct, round, 40% eye diameter; tympanic annulus raised relative to tympanum, visible on anterior and ventral margin; vomerine teeth on two slightly oblique ridges, medial to choanae, closer to each other than to choanae; tongue heart-shaped, notched posteriorly.

Tips of all four fingers expanded with circummarginal grooves; width of Finger III disc about 1.8 times width of phalanx, 50% diameter of tympanum; relative finger lengths I < II < IV

< III; outer metacarpal tubercle absent, inner small, ovoid; palmar surfaces tubercular; thenar tubercle present; well-developed subarticular tubercles on toes, one on Fingers I and II, two on

Fingers III and IV, distal tubercle on Finger IV bifid; Finger I and preaxial side of Finger II with rudimentary webbing, postaxial side of Finger II webbed to base of subarticular tubercle, preaxial side of Finger III webbed to proximal subarticular tubercle, postaxial side of Finger III and preaxial side of Finger IV webbed to base of distal subarticular tubercle; lateral fringe on all fingers; weak dermal fold on outer edge of lower arm; ulnar tubercle absent.

Figure 2. Leptopelis pictilis holotype (NCSM 82300) in dorsal (A) and ventral (B) views, coloration in life (C). Male and female from Pico Gran Caldera showing variation in coloration (D), egg clutch size and coloration (E), specimen NCSM 82317 with orange dorsum (F). Arrows indicate the tarsal spur.

Tips of all five toes expanded with circummarginal grooves; width of Toe IV disc about

1.3 times width of phalanx, 50% diameter of tympanum; relative toe lengths I < II < III < V <

IV; outer metatarsal tubercle absent, inner metatarsal tubercle distinct, ovoid; plantar surfaces tubercular; well-developed subarticular tubercles on toes, one on Toes I and II, two on Toes III and V, three on Toe IV, distal tubercle on Toe IV bifid; Toe I webbed to distal margin of subarticular tubercle, preaxial side of Toe II webbed to proximal margin of subarticular tubercle, postaxial side of Toe II webbed to about midway between subarticular tubercle and base of disc, preaxial side of Toe III webbed to base of distal subarticular tubercle, postaxial side of Toe III webbed to about midway between distal subarticular tubercle and base of disc, preaxial side of

Toe IV webbed to distal margin of penultimate subarticular tubercle, postaxial side of Toe IV webbed to proximal margin of distal subarticular tubercle, preaxial side of Toe V webbed to about midway between distal subarticular tubercle and base of disc; lateral fringe on all toes; weak dermal fold on outer tarsus; enlarged, conical tubercle (dermal spur) projecting from tibiotarsal joint.

Skin on dorsum finely granular, granulation less pronounced posteriorly; skin on flank and ventral surface coarsely granular, granulation rounder and more pronounced ventrally, smoother on throat and chest; skin on dorsal and ventral surface of limbs smooth; weak, glandular supratympanic fold; dorsolateral fold absent; distinct round rictal gland; weak dermal fringe above vent; enlarged tubercles around vent.

Coloration In Preservative —Dorsum blue, large irregular orange and red spots on back and dorsal part of flank, dorsal surface of forelimb and hindlimb with brown crossbars; side of head brown, tan vertical bar from margin of upper lip to lore, white vertical bar from margin of 23 upper lip to eye; supratympanic fold brown; ventral surfaces cream, throat and ventral surfaces of forelimb and tibiotarsus brown, indistinct brown spots on lateral part of belly.

Coloration In Life —Iris bronze, pupil black; dorsum green, large irregular maroon spots on back and dorsal part of flank, dorsal surface of forelimb and hindlimb with maroon crossbars; side of head maroon, narrow beige line from tip of snout along canthus to outer margin of eyelid, beige vertical bar from margin of upper lip to lore, cream vertical bar from margin of upper lip to eye; supratympanic fold maroon; dorsal surface of fingers and toes yellow; ventral surfaces yellow, chest white, spots on throat and lateral part of belly brown. Dermal spur white. Fringe above vent beige.

Variation —SVL of males 50–80% SVL of females (Table 1). Males with round pectoral glands. Dorsum orange rather than green in specimen NCSM 82317 (Fig. 2F). Size and shape of dorsal maroon spots variable, sometimes merging into irregular bands. Measurements are summarized in Table 1. Supratympanic fold, dermal fold on outer edge of lower arm and fringe above vent weaker or absent in some specimens. Venter in preservative variable in extent of brown coloration, with some specimens having the ventral surface mostly brown with white spots.

Molecular Divergence —The paratype of L. pictilis (CUMV 15907) has an uncorrected pairwise distance of 2.83% in the 16s fragment to L. calcaratus from Cameroon.

Distribution and Natural History —The new species was collected within two geographically distinct localities, separated by approximately 17 km, within the Gran Caldera &

Southern Highlands Scientific Reserve, Bioko. The two localities have similar montane forest habitat, but with differing levels of human disturbance. Specimens at Pico Gran Caldera were taken in a remote field site (elevation 1043–1073 m) in primary montane forest along streams 24 with open canopy areas dominated by bracken ferns. Those found at a field station on Pico Biao, near the village of Moka (elevation 1105–1451 m), were taken in a landscape dominated by subsistence farmland. Specimens at both sites were found during visual surveys at night approximately 1–2 m above the ground on leaves and ferns overhanging streams (Pico Gran

Caldera) or clearings (Pico Biao). The call of the new species is unknown. An amplexing pair collected in January 2012 at Pico Gran Caldera produced a clutch of approximately 70 orange eggs while being held overnight in a plastic bag (Fig. 2 D–E).

The new species was not found by PJM at similar elevations and habitats on Pico Basilé, although fewer surveys have been undertaken there than at the known localities. Pico Basilé is separated from other peaks on Bioko by a deep valley that reaches a maximum elevation of <800 m, while Pico Gran Caldera and Pico Biao are connected by a high plateau that may allow dispersal between the two localities. Additional field surveys on Bioko and the adjacent mainland are needed to determine if the species’ geographic range extends beyond the two known localities.

Etymology —The specific epithet pictilis (L.) means “painted,” in reference to the striking dorsal pattern of the new species.

DISCUSSION

The Guinean Forests Biodiversity Hotspot contains one of most diverse amphibian assemblages in Africa (Penner et al. 2011; Barej et al. 2014). Within this region, the unique biogeographic history of the BFH has generated high levels of endemism across gradients in elevation and within isolated high elevation habitats (Blackburn & Rödel 2011; Barej et al. 2014).

Discovery of the large-bodied and conspicuously colored, endemic L. pictilis underscores the evolutionary significance of the BFH, and the extent to which Bioko Island’s herpetofauna and 25 capacity for endemism have been overlooked as an ecologically and evolutionarily important part of the BFH (Blackburn 2010).

Leptopelis pictilis is morphologically distinct in having bright dorsal coloration (green with maroon or orange pattern) and a dermal spur on the tibiotarsal joint, a trait only found in two other Leptopelis species (L. aubryioides and L. calcaratus ). It is also the only Leptopelis species known to produce distinctive orange eggs, though the reproductive biology of most species in this genus remains undescribed (Schiøtz 1999; Amiet 2012). Leptopelis calcaratus occurrs in sympatry with L. pictilis at the upper reaches of its elevational range on Bioko and exhibits moderate divergence at 16s (2.83%), indicating that the species are closely related.

Future genetic analysis with increased geographic sampling of Bioko and Cameroon will provide further insight into the evolutionary history of this island endemic species.

The known distribution of L. pictilis is restricted to the montane forests of Bioko Island’s southern highlands between 1000–1500 m, encompassing 152 km² within the Gran Caldera and

Southern Highlands Scientific Reserve, and excludes similar habitats on the island’s northern peak, Pico Basilé. Given the lack of conservation management and enforcement within Bioko’s protected areas (Gran Caldera and Southern Highlands Scientific Reserve; Cronin et al. 2014), continued habitat fragmentation within the Pico Biao region, and the limited extent of occurrence and known localities, we recommend L. pictilis is categorized as Vulnerable (VU B1ab(iii); D2) per the definitions outlined in the IUCN Red List Categories and Criteria for extinction risk

(IUCN 2012).

The most recent amphibian inventories for Bioko Island report a combined 34 species

(Eisentraut 1973; Jones 1994), which is likely a considerable underestimate based on surveys conducted by PJM and RCB from 2009–2013. Through these surveys we have extended the 26 ranges for a number of continental species to include Bioko Island, including several Central

African species of Leptopelis (P. McLaughlin & R. Bell, personal observation). Recent surveys also confirmed the existence of the endangered bufonid frog Didynamipus sjostedti on Bioko

(McLaughlin, personal observation; field tags PJM393–397). These findings underscore the importance for continued exploration on Bioko, as it may represent a substantial range expansion for some mainland species, with important conservation implications for species that are threatened. 27

CHAPTER 3: PREVALENCE OF THE AMPHIBIAN CHYTRID PATHOGEN (Batrachochytrium dendrobatidis ) ON BIOKO ISLAND, EQUATORIAL GUINEA

INTRODUCTION

Enigmatic amphibian declines over the past three decades have impacted nearly 30% of all species on Earth, and necessitated a rapid investigation into their causes and possible mitigation strategies (Stuart et al. 2004; Roelants et al. 2007). Among the leading causes for the observed decline is the fungal pathogen Batrachochytrium dendrobatidis (Bd ), a type of chytrid fungus that causes the infectious skin disease chytridiomycosis in amphibians (Berger et al. 2005,

Lips et al. 2006, Longcore et al. 1999). The disease has been reported from all continents where amphibians exist, and is the likely cause for reported extirpation and/or extinction of at least 200 species of the 500+ species known to have been affected (Skerratt et al. 2007; Fisher et al. 2009;

Olson et al. 2013). Amphibians are currently the most threatened class of vertebrates, and despite an intense focus of research on Bd ecology and infection dynamics over the past decade, there remains a significant gap in knowledge for mitigating or preventing further spread of this deadly disease (Stuart et al. 2004; Venesky et al. 2014).

ORIGIN AND SPREAD OF CHYTRIDIOMYCOSIS

Understanding the origins and history of global outbreaks has been a focus of recent Bd research, as it may help to advise future conservation efforts. Much of this work has sought to determine whether Bd is a relatively new and emerging pathogen, or a widespread endemic whose virulence increased due to altered host-pathogen dynamics as a result of environmental perturbations like pollution or climate change (Rachowicz et al. 2005; Fisher et al. 2009). The question of origin is further confounded by the existence of at least four major Bd lineages that have thus far been identified through whole-genome sequencing: the global panzootic lineage 28

(Bd -GPL), the Cape lineage ( Bd -Cape), the Swiss lineage ( Bd -CH), and the Brazilian lineage

(Bd -Brazil) (Rosenblum et al. 2013; Farrer et al. 2011; Rodriguez et al. 2014). Of particular interest is the origin of Bd -GPL, which is found worldwide and implicated as the causative agent in most of the observed amphibian panzootic events (Rodriguez et al. 2014; Lips 2014). Bd -GPL could be an endemic strain that was spread to naïve populations, or a hyper-virulent recombinant lineage resulting from two parental strains that were brought into contact and subsequently spread (James et al. 2009; Farrer et al. 2011; Bataille et al. 2013). Past and ongoing intercontinental spread and possible hybridization of different lineages, including those that may not yet be identified, will only make identification of a source increasingly difficult in the future.

Using sequence data from 17 loci across 59 different Bd strains, James et al. (2009) showed extremely low genetic variation and closely related genotypes of samples taken from across the globe, supporting the assertion that Bd may be a novel pathogen with recent proliferation. This hypothesis is further supported by studies showing waves of infection rapidly advancing through regions such as Australia and Central America (Berger et al. 1998; Lips et al.

2006) and patchy distributions in affected regions reminiscent of the spread of a new pathogen

(http://www. Bd -maps.net/). Under this scenario, a source population will have higher strain heterozygosity than all others, as the pathogen spreads rapidly and subsequently undergoes a genetic bottleneck. Among the 59 strains sampled by James et al. (2009), one sample from a

North American bullfrog ( Rana catesbeiana ) was found to contain as much strain heterozygosity as was represented in all the other strains sampled. The widespread export of R. catesbeianus globally as a food source lends credibility to this scenario, and suggests that North America may be a source for Bd -GPL. In contrast to these findings, Rodriguez (2014) sampled 2799 amphibian museum specimens collected from Brazil, which showed presence of both Bd -GPL and Bd - 29

Brazil as early as 1894. The presence of Bd in these museum specimens predates the observed global declines as well as any possible introduction from North America via R. catesbeiana , and the infection prevalence of these two lineages showed no significant variation in prevalence over time, supporting the hypothesis that they are longstanding endemic pathogens in Brazil

(Rodriguez et al. 2014). These studies highlight the complex and conflicting history of research into Bd origins, along with the importance of continued global sampling to uncover a better idea of the diversity within Bd .

While a central focus has been on Bd -GPL and in areas with reported amphibian deaths, histological examination shows evidence of many Bd -type infections from museum specimens from long before global declines were observed, including Africa from as early as 1933 (Weldon et al. 2004; Soto-Azat et al. 2010) and North America as early as 1888 (Ouellet et al. 2005;

Talley et al. 2015). These histological studies often focus only upon confirming presence or absence of Bd infection within a specimen due to difficulties in identifying Bd lineage from museum specimens using qPCR, however there is still substantial evidence showing that Bd is composed of many different endemic and potentially genetically isolated variants that persisted in amphibians long before contemporary declines were observed, with a common ancestor at least 750 years ago (Fisher et al. 2009; Farrer et al. 2011; Schloegel et al. 2012). These records underscore the difficulty in determining the timing, geographic origin, and current distribution of

Bd -GPL. Despite the complex nature of global origins and spread of Bd -GPL, the recent rapid declines seen in many parts of the world were obvious epizootic events. These rapid die-offs seem more likely to be caused by the introduction of new lineages into naïve amphibian populations as opposed to gradual changes in host-pathogen dynamics causing endemic lineages to increase in virulence, though neither scenario can be ruled out. 30

Given the apparent co-occurrence of both non-lethal endemic strains and possibly novel disease-causing strains in many areas (Bataille et al. 2013; Rosenblum et al. 2013), some efforts to disentangle the geographic origins have focused in-part on the international trade in amphibians. Fisher & Garner (2007) documented 28 species known to harbor Bd infections that are traded internationally, and of these the most common are the South American Cane Toad

Rhinella marina , North American Bullfrog Rana catesbeiana , and various species of the genus

Xenopus from Africa. Each of these three were subsequently introduced to local environments in their destination countries, either intentionally or accidentally, and represent the most repetitive and intense potential vectors for the initial introduction of a Bd parent strain or Bd -GPL, and subsequent global proliferation (Fisher & Garner 2007; Schloegel et al. 2012).

Historic trade in R. marina for biocontrol of agricultural insect pests introduced the species extensively throughout the Caribbean and Pacific regions (Easteal 1981;

Shanmuganathan et al. 2010). The introduction of R. marina to Australia in 1935 predates the emergence of chytridiomycosis there by forty years, casting doubt upon their role in the initial dispersal of Bd (Lever 2001; Rachowicz et al. 2005). While contemporary trade in R. marina for biocontrol likely subsided after it became an infamous example of the ecological dangers associated with introducing exotic species, recent evidence suggests the ability of this species to act as a vector for Bd in the wild to be extremely limited. (Lettoof et al. 2013; http://www. Bd - maps.net/surveillance/default.asp).

In addition to harboring Bd with allelic diversity that is emblematic of a possible source strain, R. catesbeiana has been shown to carry a wide variety of different Bd genotypes, and this species is native to North America where some of the earliest known records of Bd have been confirmed (James et al. 2009; Schloegel et al. 2012). However, despite being exported to over 40 31 countries and 4 continents (Lever 2003), the timing and current distribution are inconsistent with the emergence of Bd and amphibian declines in places like Australia and South Africa (Ouellet et al. 2005; James et al. 2009; Talley et al. 2015). While R. catesbeiana is almost certainly the most common vector for intercontinental spread owing to their extensive and ongoing trade and farming, there remains no clear evidence that they harbored the source strain that initiated the global panzootic events (Schloegel et al. 2012).

Among the first hypotheses for a source of pathogenic Bd was from Africa via the dissemination of Xenopus. Trade in Xenopus spp. represents the earliest and most extensive intercontinental trade in potentially Bd -harboring amphibians, having been shipped around the world for use in pregnancy tests as early as the mid-1930’s and lasting into the 1970’s, and thereafter for scientific use as a model organism in the study of immunity, embryology, and molecular ecology (Weldon 2004). During this time, commercial suppliers and recipients did not take into account possible disease transmission when handling any of the three different species of Xenopus that were commonly exported from South and West-Central Africa (Nace et al.

1971). The earliest records of Bd from Africa come from separate studies which independently confirmed it in museum specimens of Xenopus fraseri from Cameroon (1933) and Xenopus laevis from Southern Africa (1938) (Weldon 2004; Soto-Azat et al. 2010). There have been multiple instances recorded of Xenopus spreading and becoming established in their new destinations, and captive colonies have been shown to harbor Bd (Parker et al. 2002;

Cunningham et al. 2005).

While the history of trade in Xenopus confirms it as a top candidate alongside R. catesbeiana as potential early vectors of Bd -GPL, research has focused upon the latter species and in areas such as North, Central, and South America where declines have been well 32 documented and research capacity is much greater. Relatively few studies have focused on documenting its distribution in Africa (Kielgast et al. 2010). Among those, the majority have been conducted in the more arid regions of East and South Africa, with comparatively few occurring within the species-rich tropical regions of sub-Saharan West and Central Africa (Bell et al. 2011; Bd maps). In East and South Africa, widespread Bd has been reported from Ethiopia

(Gower & Loader 2012), Kenya (Kielgast et al. 2010), Tanzania (Moyer & Weldon 2006),

Uganda (Goldberg et al. 2007), Botswana, Swaziland, Lesotho, and South Africa (Weldon 2005).

In West-Central Africa Bd has only been confirmed from Democratic Republic of the Congo

(Greenbaum et al. 2015), Gabon (Bell et al. 2011), Cameroon (Balaz et al. 2012), and Nigeria

(Imasuen et al. 2011; Reeder et al. 2011). Among African islands, Bd has been confirmed on São

Tomé (Hydeman et al. 2013) and Madagascar (Bletz et al. 2015). A recent survey targeting seven

West African countries failed to confirm the presence of Bd across 793 samples (Penner et al.

2013). To date, there are no credible records of Bd West of the Dahomey Gap, suggesting that it may serve as a barrier to West-African Bd expansion. Another possible explanation is that Bd does exist in West Africa but at such low infection loads that the common methodology of qPCR from swab-extracts was unable to detect it. This detection error has been shown to happen in experimental lab assays (Shin et al. 2014), and may also explain previous surveys in Gabon

(Gratwicke et al. 2011) and Madagascar (Vredenburg et al. 2012) that failed to detect Bd despite surveys conducted shortly thereafter confirming its existence. While a non-virulent Bd could have spread rapidly within these populations between these surveys, it seems an unlikely explanation, and highlights the issues associated with detecting Bd in areas where it may be endemic and persisting at low infection loads. 33

The most striking and potentially significant characteristic of African Bd is the complete absence of any reported mass mortality events, and even among the many amphibians that have tested positive for Bd , none have yet shown any clinical symptoms of chytridiomycosis.

However, Africa may not be unique in this aspect, as there are many regions that are not experiencing amphibian declines despite testing positive for Bd (Fisher et al. 2009 ). Gahl et al.

(2012) showed that among seven North American amphibian species exposed to both a local northeastern strain (JEL404) and Panamanian strain (JEL423), mortality was not ubiquitous.

While two species died from both strains, many became infected without succumbing, and one species only experienced full mortality from the Panamanian strain (Gahl et al. 2012). This study highlights the possibility that species may develop resistance to local strains through coevolution

(McCallum 2005) while being susceptible to novel varieties, and may help explain the apparent presence of chytrid in parts of Africa despite an absence of observed symptoms or mass mortality events (Goldberg et al. 2007; Greenbaum et al. 2008; Kielgast et al. 2009; Bell et al.

2011). Variable mortality among species also underscores the important role that some species may play as vectors for the proliferation and persistence of the disease in the wild. Conservation efforts could be aided through future efforts aimed at identifying species that may act primarily as vectors, as opposed to those that are most susceptible to rapid mortality (Bell et al. 2011).

Given the lack of clarity regarding African Bd , specifically in West-Central Africa, and the high plausibility for an Out-of-Africa hypothesis, we sought to sample amphibians for Bd on

Equatorial Guinea’s Bioko Island, located 32 km off the coast of West Africa. Our study aimed to provide information on the distribution and infection dynamics of Bd in Central Africa by conducting community-level sampling of amphibian populations across a wide elevation gradient 34 and during different seasons. Herein we report the first records of Bd from Bioko Island using quantitative polymerase chain reaction (qPCR) methods.

MATERIALS AND METHODS

Two separate amphibian surveys were conducted to provide seasonal variation in chytrid infection prevalence and load within multiple habitats and across as wide an elevation as possible.

Wet season surveys were conducted in 9 localities from August 19 ̶ September 12, 2011. Dry season surveys were conducted in 12 localities from January 12 ̶ March 8, 2012. Among the 16 localities surveyed, five were sampled in both the wet and dry season, as wet season conditions prohibit safe travel to some of the dry season localities. Habitats surveyed, as defined by Zafra-

Calvo et al. (2010), included pristine lowland/coastal, monsoon, montane, and mossy forests, coastal wetlands, disturbed forest/farmland at various elevations, and montane grassland/fern- dominated landscapes on Pico Biao. Sample sites ranged from the coast up to approximately

1936 m during both wet and dry season surveys.

Amphibians were collected during nighttime visual surveys by hand and placed into separate plastic bags that were sealed until swabbing. Live specimens were stored at ambient room temperature overnight, and processed in order of collection the following morning. No live specimens were kept for a period of greater than 24 hours before healthy release to their collection site. We sampled a total of 440 individuals, including 186 adults and two tadpoles in the wet season, and 252 adults in the dry season. Following the methodology of Hyatt et al.

(2007), samples were taken with individually wrapped sterile fine-tip swabs (MW113; Medical

Wire & Equipment Co., Wiltshire, England). All swabbing took place in a controlled environment (field station laboratory) to reduce potential contaminants. During swabbing, individuals were secured by hand through the plastic bags without removing them to reduce 35

Figure 1. Sampling localities on Bioko Island, Equatorial Guinea (top left). Localities sampled by season ( below ) show ratio of Bd positive vs Bd negative specimens. Sample site names derived from research camps operated by the Bioko Biodiversity Protection Program, and using Bioko Island map produced by Talleres del Instituto Geograficio Nacional – Deposito Legal: Madrid 7955 – 1983.

36 possible contamination. Medical-grade nitrile gloves were used during swabbing and were changed between individuals in the rare event when contact was made between the gloved hand and the inside of the bag or the individual being swabbed. Each individual was swabbed continuously for 20-30 seconds across all ventral surfaces, including drink patch, thighs and feet.

Swabs were immediately broken off into sterile 1.5 ml vials containing 95% EtOH. Vials were refrigerated in the field and kept as cool as possible during transit from the field to lab facilities in the U.S., where they were stored permanently at -80°C until processing. A portion of the sampled individuals were euthanized after swabbing and prepared as voucher specimens that are deposited at Cornell University Museum of Vertebrates (CUMV) and the North Carolina

Museum of Natural Sciences (NCSM). All amphibians were collected following Drexel IACUC

Protocol #18748.

DNA extraction and qPCR detection of Bd were conducted following the methods of

Boyle et al. (2004) on a real-time PCR machine (7900 HT; Applied Biosystems, Carlsbad, CA) in 96-well plates. DNA dilutions (1:10) were analyzed for each sample and compared against standards of known zoospore concentrations of Bd strain JEL427 (Puerto Rico). Dilutions were necessary to reduce potential inhibition of target Bd regions from large amounts of contaminant

DNA extracted from swabs, and to allow for duplicate qPCR runs per sample. Flourescence levels were evaluated for each sample using SDS v 2.1 (Applied Biosystems), and all samples were evaluated in duplicate plates with two negative controls (dH2O) per plate. Samples were considered positive for Bd when a sigmoidal amplification occurred in both PCR reactions, and if the average infection intensity (inferred zoospore genome equivalent – GE) from both was greater than one.

RESULTS

Our total sampling included 16 genera and 25 species of frogs from seven families

(Recent phylogenetic work suggests the genera Petropedetes , Phrynobatrachus , and Ptychadena be separated into their own distinct families, but for the purposes of this study they were included in the family Ranidae per previous taxonomic designation) (Table 1). We detected positive Bd samples from all sampled sites, habitats, and elevations in both the wet and dry season (Table 2). Results from all negative controls in each qPCR run showed no amplification, ruling out potential contamination accounting for Bd -positive samples. These samples indicate that Bd is widely distributed on Bioko, and represent the first known records of Bd from the island. Bd was detected in all of the seven families, absent only in two species ( Afrixalus laevis and Wolterstorffina parvipalmata ), however sampling of these species consisted of a single sample each, and thus we cannot reject the possibility that they may harbor Bd infections.

Among the nine species with over 20 individuals sampled, Petropedetes johnstoni had

the highest prevalence (89%) and Chiromantis rufescens had the lowest (48%). The highest

single GE value of 735.9 was recorded during the wet season from Risule ( Hyperolius endjami ),

and the highest dry season value of 165.2 was recorded from Moka ( Arthroleptis variabilis ).

Among the five sites sampled in both seasons, prevalence at each site was consistently higher

during the dry season, while average GE at each site was consistently higher during the wet

season. All dry season sites had at least 90% prevalence except for Hormigas Camp (83%).

Results suggest that seasonality may drive infection prevalence in amphibians on Bioko Island,

as a higher overall proportion of samples tested positive for chytrid in the dry season (94.8%)

than in the wet season (46.3%) when calculated using raw data proportions. Despite the

widespread occurrence of Bd on Bioko, no signs of mortality or individuals suffering from 38

aiis1 eea2 pce 4 2/4 30.6 ± 734.8 326/440 440 25 Species 16 Genera 7 Families Hyperoliidae Rhacophoridae Bufonidae Ranidae Pyxicephalidae Sampled Species Genus Arthroleptidae Family of Distribution 1. Table Bd Batrachochytrium dendrobatidis Pipidae Bd meohyu tuberosus camerunensis poensis Amietophrynus Amietophrynus rhoets variabilis sylvaticus Arthroleptis Arthroleptis Arthroleptis urasubsigillata Aubria fiau paradorsalis laevis Afrixalus Afrixalus - - - rufescens Chiromantis - - parvipalmata afra Wolterstorffina Nectophryne - tcaeaaequiplicata - johnstoni albolabris auritus Ptychadena Phrynobatrachus Petropedetes Hylarana - etpls calcaratus rufus pictilis modestus brevirostris Leptopelis leucomystax - Leptopelis Leptopelis Leptopelis Leptopelis Cardioglossa - - - - sp. nov. new tetraploid 2 Silurana - Hyperolius pshtya immaculatus tuberculatus ocellatus endjami Opisthothylax Hyperolius Hyperolius Hyperolius across Anuran Families and Species Sampled on Biok on Sampled Species and Families Anuran across , GE genome equivalentsGE , ukonjvnl)101-- - 0/1 1 juvenile) (unknown tdoe)1119. - 96.6 1/1 1 (tadpoles) 11/11. 179.0 11.7 ± 31.7 10/21 21 / - 0.8 8.6 7.9 - 1.9 8.3± 1.1 ± 29.4 9.2 ± 15.6 0/1 10/11 2/2 3/5 1 11 2 5 21/21. 0719.5 9.8 6.8 25.5 16.9 ± 50.7 15.7 ± 38.3 9.6 ± 49.1 17.4 ± 93.1 10/12 17/31 99/111 12/19 12 31 111 19 / 26- 52.5 7.5 42.1 7.0 3.9 58.7 22.6 ± 138.3 6.7 ± 13.7 51.8 51.4 ± 80.1 54.8 7.4 ± 13.0 7.8 ± 8.0 24.7 ± 162.4 14/22 33.7 ± 237.1 1/2 3/3 3/5 3/5 4/4 17/25 20/27 22 2 3 5 5 4 25 27 / 73±1. 5.5 17.3 ± 10.6 3/3 3 / . 655.4 8.1 ± 16.5 7/7 7 / - 180.2 - 75.0 ± 713.2 21/25 0/1 25 1 / 68±9. 63.8 53.0 39.4 137.4 76.8 ± 90.2 70.7 ± 163.1 26.8 ± 157.8 78.8 ± 731.6 2/5 10/12 16/31 38/48 5 12 31 48 o Island, Equatorial Guinea Equatorial Island, o 2 8136. 3. 128.4 67.6 ± 734.2 88/123 123 7 3/7 16±9. 11.5 11.6 ± 95.6 138/173 173 11/11. 179.0 11.7 ± 31.7 10/21 21 91/976±2. 7.9 7.6 ± 29.8 15/19 19 46/43. 3. 48.3 33.3 ± 239.1 65/94 94 / 73±1. 5.5 17.3 ± 10.6 3/3 3 / . 655.4 8.1 ± 16.5 7/7 7 Bd -positive/sampled Bd intensity (GE) intensity 74 .3 ST DEV 39

6Sml ie -96--8/8 26±1174. 15-193 41.6 72.6 ± 131.7 87/188 - - 7-1936 16 Sample Sites Moka Cascades Trail North Camp Hormigas Camp Bao Lago Biao ᵇDry Season:January-MarchᵇDry 2012 Season:August-September ᵅWet 2011 Bd Batrachochytrium dendrobatidis # Species (m) Elev. Avg. (m) Range Elev. abaRi Site Sample of 2.Distribution Table Puntas Cabras Risule Rio Bacha Camp Bate te Caldera Trail Punto Santiago L uba Moaba Camp Moraka Playa 7-8 8 /02. - - 5.5 7.1 7.3 10.9± 21.32 ± 39.0 - - 30/36 59/61 - - 5 9 27 - 1047 - 572 7 1003-1071 541-623 - - - 1434 - - 1144-1527 - 94.6 50.4 0 - - 84.8 ± 430.2 58.9 ± 139.6 26.0 - - 24/31 8/33 - 1/10 - 5 7 4 - 1426 - 1221 1376-1555 288 1105-1330 - - 277-387 3913 8631/37. 6. 0913 9622217.5 2/2 2 1936 1936 50.9 77.1 ± 163.1 12/13 3 1876 1379-1936 -1 441/6177±7802821 531/710.2 ± 26. 16/17 3 15 15 238.2 157.7 ± 708.0 10/16 4 84 7-315 07 5451 94±6. 111-01 61 11.1 ± 20.6 16/16 4 19 18-30 31.1 29.4 ± 66.6 5/16 4 45 10-70 2 2 22 19±2. 5.7 3.2 - 11.3 - 11.9 ± 20.9 4.6 4.4 9.3 ± 8.1 9.5 ± 49.2 - - 6.7 ± 13.4 9.2 ± 16.8 22/22 7.3 ± 18/18 7/7 5 7/8 19/21 - - 5 1 16/17 222 2 7 - - 136 94 6 222 27 44 - - 136 443 - 94 - 28-131 - 27 - - 443 - - 92.7 159.3 - 0 - - 97.5 ± 86.7 237.2 ± 734.2 - - 35.1 - - 21/56 5/8 - - - 11 1/5 - - 3 - - - 2 467 - 171 - - 262-780 - 78 - - 171 - - 64-81 - - Bd During Wetᵅ and Dryᵇ Seasons by Sample Site on Bio on Site Sample by Seasons Dryᵇ and Wetᵅ During , GE genomeGEequivalents , WET SEASON Bd -positive/sampled ko Island, Equatorial Guinea Equatorial Island, ko Bd intensity (GE) intensity ST DEV Elev. (m)Range Elev. 3/5 07±1. 3.5 10.7 ± 10.8 239/252 - - 6 Avg. Elev. (m) Elev. Avg. # Species DRY SEASON Bd 2 71±10929.9 17.1 ± 160.9 /27 -positive/sampled Bd intensity (GE) intensity ±020.2 ± 0.2 1. 3.8 14.2 6.7 3 6.1 TS DEV

40 chytridiomycosis were detected during the sampling period. Regions of the world where chytridiomycosis has had significant negative impacts upon amphibians (ex. Central and South

America) have had associated mass mortality events that signaled the apparent outbreak. Despite a sustained research presence in the forests of Bioko Island since the mid 1990’s by researchers from the Bioko Biodiversity Protection Program (BBPP) and the National University of

Equatorial Guinea (UNGE), there have been no accounts of any mass mortality events anywhere on Bioko. Severe chytridiomycosis is believed to occur at infection loads of approximately

10,000 GE (Vredenburg et al. 2010), which is 14-fold greater than the highest infection load found in any of our samples (Tables 1, 2).

Results were evaluated using a mixed effects logistic regression model using a variety of fixed and mixed effects consisting of both a random intercept and slope corresponding to genus:species combinations (hereafter referred to as species or species factor) and elevation, respectively. A “season” predictor was nested at a second level within species factors, thus allowing the species component to vary. This variation is visualized as the normalized random intercepts in Fig. 2 ( top plot ). The species factor was chosen due to the assumption that there may be variance in infection load between different genera and/or species due to skin type and/or associated habitats common within each genus or species represented in the sampling. The elevation factor was chosen as a proxy for temperature, since Bd infection load may vary at temperatures outside its optimal growth range of 17 ̶ 25 °C (Piotrowski et al. 2004); hence, species acted as the random intercept for the model.

Figure 2. (top plot ) Plot from mixed effects logistic regression model. Fixed effects (season, elevation category, habitat usage of species) and mixed effects consisting of a random intercept and random slope corresponding to genus:species combinations and elevation, respectively (genus:species intercept: 0.39509; genus:species elevation Z slope: 0.18384 ; elevation was z-scored; x-axis represents normalized random intercept variance). Plots of Bd presence by elevation using factors of genus ( low plot, left ) and family ( low plot, right ).

Figure 3. Bd infection load (GE) plotted according to family and season. (a) Positive samples by family: Arthroleptidae (Avg: 33.3; SD: 48.3), Bufonidae (Avg: 7.6; SD: 7.9), Pyxicephalidae (Avg: 17.3; SD: 5.5), Hyperoliidae (Avg: 67.7; SD: 128.4), Pipidae (Avg: 8.1; SD: 5.4), Ranidae (Avg: 11.6; SD: 11.5), Rhacophoridae (Avg: 11.7; SD: 9.0). Top six GE values omitted (Arthroleptidae: 1; Hyperoliidae: 5). (b) Positive samples wet season vs. dry season: Wet (Avg: 86.5; SD: 127.1), Dry (Avg: 10.6; SD: 11.9). Top six GE values omitted (Arthroleptidae: 1; Hyperoliidae: 5). (c) Positive wet season samples by family: Arthroleptidae (Avg: 77.4; SD: 56.7), Bufonidae (Avg: 2.7; SD: 0.9), Hyperoliidae (Avg: 106.8; SD: 155.8), Ranidae (Avg: 32.8; SD: 30.3), Rhacophoridae (Avg: 19.2; SD: 22.4). Top six GE values omitted (Arthroleptidae: 1; Hyperoliidae: 5). (d) Positive dry season samples by family: Arthroleptidae (Avg: 12.3; SD: 24.0), Bufonidae (Avg: 8.8; SD: 8.4), Pyxicephalidae (Avg: 17.3; SD: 5.5), Hyperoliidae (Avg: 11.0; SD: 6.0), Pipidae (Avg: 8.1; SD: 5.4), Ranidae (Avg: 10.2; SD: 7.1), Rhacophoridae (Avg: 9.9; SD: 3.5). Top GE value omitted (Arthroleptidae: 1). 43

Utilizing the equation p(infection) , we can calculate the approximate probability of Bd infection by season for each species factor using values from the plot for each species (α species ) along with the season variable (Dry = 0; Wet = 1). The following is an example of this calculation for Hyperolius tuberculatus (α species = 1.25) dry season infection probability:

3.3174+1.25 - 3.6623 ( 0 ) e p() infection = = 0.9897 3.3174+1.25 - 3.6623 () 0 1+ e

Thus, in the dry season the species H. tuberculatus had the highest probability of infection

(99.0%) on average, while Phrynobatrachus auritus had the lowest probability of infection

(88.8%) on average. There is not a great deal of ecological or biological significance that can be attributed to such a slight overall variation in probability of infection. Both values are still extremely high in terms of infection probability, and with all species falling within this 10.2% range, it does not suggest any major species-specific trends in infection probabilities in the dry season. Conversely, when this calculation is done for the wet season species, we find a more compelling amount of variation. In the wet season, H. tuberculatus had the highest probability of infection (71.2%) on average, while P. auritus had the lowest (16.9%) on average. Unlike the dry season results, the 54.3% range of species infection probabilities suggests that species-specific characteristics are a more important driver of variability in the wet season. Furthermore, it shows that there are considerable differences in infection probabilities among most species between the dry season and wet season, possibly owing to environmental conditions within each that favor or discourage infection, respectively.

The variation in infection probability across elevation for genus and family is demonstrated in Fig. 2 ( low plots ). Model validation was performed sequentially with a full model (fixed effects: season, elevation category, habitat usage of species) compared to a null 44

model via a chi squared test. These results suggest there is no apparent taxon-specific variation in

infection probability across elevation, as might be suspected given that Bd infection loads and/or presence may vary with temperature.

Finally, Bd positive samples were plotted in four different combinations using Bd

infection intensity (GE) by family or season (Fig. 3). Even with outliers omitted, no significant

variation in GE can be seen across families (Fig. 3a), across families within the wet season (Fig.

3c), or across families within the dry season (Fig. 3d). The only noteworthy variation is seen in

GE by season (Fig. 3b), supporting the results from raw data proportions of variability between

seasons.

DISCUSSION

Despite testing a variety of factors, our analysis shows that Bd prevalence and infection

load are best explained by seasonality. Our samples show higher prevalence of Bd in the dry

season, but higher infection load in the wet season. However, despite higher infection loads in

the wet season, results from the mixed effects logistic regression model suggest that wet season

conditions may allow many species to avoid Bd in the environment, as the overall probability of

infection for each species was considerably lower in the wet season as compared to the dry

season. This suggests that there may be a species-specific factor that allows some species to

avoid infection during the wet season. Alternatively, the authors hypothesize that wet season

conditions may simply increase the amount of favorable (wet) habitat available to amphibians,

while simultaneously reducing or “flushing” aquatic Bd spores from much of the optimal habitat.

If there is more available habitat (streams, rivers, pools) that is devoid of Bd , or with lower

concentrations of Bd due to this “flushing”, then infection probabilities will likely be reduced for

most species. Concurrently, while dry season infection loads are lower than in the wet season, 45

the model suggests the probability of infection is higher across all species. Again, this may best be explained by habitat/water availability, where amphibians in the dry season are forced into

fewer areas of optimal (wet) habitat where there is a higher likelihood of aquatic Bd spore

retention and concentrations.

Taken altogether, our results are mostly consistent with other studies from tropical

habitats showing significantly higher prevalence in the drier-cooler months in Australia (Kriger

& Hero 2007), Puerto Rico (Longo et al. 2010), and Madagascar (Bletz et al. 2015). Surveys

of Bd in temperate habitats also show a seasonal pattern, with higher infection load and prevalence occurring in colder months in the Northeast (Lenker et al. 2014) and Southwest

(Savage et al. 2011) United States. While a simple presence/absence ratio of Bd prevalence is

comparable between seasons, sites, and species, it is impossible to make the same comparisons

with infection load. A variety of different confounding factors dictate infection load which we

could not account for in our sampling regime. Primary among these is that different amphibian

species tolerate different infection loads, and the breadth of species we sampled prevents us from

comparing infection intensities between seasons.

We did not see an expected correlation between elevation and infection prevalence. Bd persistence is temperature dependent, surviving within a range of 4 ̶ 25 °C, with optimal growth between 17 ̶ 25◦C (Piotrowski et al. 2004). Since temperature varies with elevation, the lack of a

correlation between elevation and prevalence or infection load was surprising, but not

inconsistent with results from other studies. Despite the temperature-dependent nature of Bd

growth in lab settings, many field studies have also failed to find a direct correlation between

elevation and infection dynamics (Kriger & Hero 2008; Sapsford et al. 2013). We hypothesize

that micro-climate characteristics do not vary as much as air temperatures across elevation, 46

especially on Bioko Island where constant rainfall (even in the dry season) may help to further

homogenize micro-climate temperatures at all elevations. Further research should attempt to

measure the variation in air vs. micro-climate temperature across seasons to determine if this

scenario may be plausible.

Future Bd research in Central Africa should continue to assess the geographic and

taxonomic extent of Bd in the region, with particular attention paid to monitoring susceptible populations (high elevation species, endangered species) and areas West of the Dahomey Gap

where Bd is yet to be confirmed (Kielgast et al. 2010; Hydeman et al. 2013). Opportunistic Bd

surveys are the most common within the literature, and while they are useful in providing range

expansion information, future research should prioritize seasonal variation, host susceptibility,

Bd diversity and virulence, and disease dynamics. With the wide diversity of Bd now being

revealed, it is critically important that future sampling, extraction, and analysis standards are

followed appropriately to allow comparison between studies (Shin et al. 2014; Bletz et al. 2015).

More extensive surveys, larger sample sizes, and temporal surveys are needed to give a better picture of chytrid dynamics as they vary between locations, species, and by Bd strain.

Understanding these dynamics will help target future research and mitigation efforts to be more

effective and prioritize the most at-risk regions and species. Surveys and analysis are relatively

simple, and increasingly less expensive to process. These surveys should be promoted amongst

researchers leading expeditions to remote regions of the world, no matter what the central focus

of their fieldwork. Surveys can have additional benefits in bolstering species inventories and

involving undergraduates and local citizens in scientific research, and thereby increasing

awareness and public education of the risks faced by amphibians.

CHAPTER 4: CHARACTERIZATION OF THE AMPHIBIAN CUTANEOUS MICROBIOME

INTRODUCTION

Characterizing the microbiota, and its effect upon organismal physiology, disease

resistance, and diversification has become increasingly important for ecologists. With the

emergence of the hologenome theory (Zilber-Rosenberg et al. 2008), and as molecular

techniques to isolate and analyze host bacterial communities have advanced and become

increasingly more cost-effective, there has been an increased focus upon discovering the roles bacterial symbionts play in an organism’s health and function.

Unnaturally high extinction rates in amphibians across the globe due to the fungal skin pathogen Batrochytridium dendrobatidis (Bd ) has precipitated a rapid effort to explore antifungal properties of the amphibian skin environment among species that are resistant, in hopes of

finding a way to counteract this devastating pandemic (Jani & Briggs 2014; Wake & Vredenburg

2008). Research has shown that the addition of Janthinobecterium to Bd -infected frogs was able

to prevent mortality in the species Rana muscosa (Harris et al. 2009). These bacteria produce the

antifungal metabolite violacein, and this work has helped to pioneer a new focus on developing

an amphibian probiotic treatment against chytridiomycosis. Future mitigation efforts, including

the development of probiotics, will require additional research to be done on host-bacteria

dynamics, synergistic effects of microbial communities, and the identification of additional

antifungal bacteria. Unfortunately, much of the work thus far has focused on laboratory assays

with captive-bred amphibians, as opposed to testing wild-caught individuals (Harris et al. 2009).

There have been very few studies aimed at characterizing the amphibian microbiome, and

fewer still that have explored whether these bacterial community assemblages vary between 48

different amphibian species (McKenzie et al. 2012). While there has been a surge in laboratory

studies examining the ability of amphibian-derived bacterial symbionts to inhibit the growth of

Bd via anti-fungal bacterial metabolites, relatively few have attempted to sample and

characterize bacterial communities or metabolites of wild-caught amphibians. Of these,

McKenzie et al. (2012) showed that host species was a highly significant indicator of cutaneous bacterial community composition, though this study was limited in scope, sampling only larval

individuals of two anuran species and one salamander species. In another such study, Kueneman

et al. (2013) sampled adults of five species ranging from tadpoles to adults, and also found host-

specific trends of cutaneous bacteria communities. Both studies focused on species in temperate

regions of the Western US (Colorado and California, respectively) meaning that tropical

amphibian communities have not yet been sampled as such. To our knowledge, there has been no

comprehensive attempt to sample both bacteria and metabolites from the same individuals across

many wild (or captive) species. Given the limited research in this area, there is need for

continued investigation into the amphibian cutaneous microbiome.

Herein we attempt to characterize the cutaneous bacterial communities and metabolite profiles of 17 different amphibian species. Using a non-lethal swab-based technique developed by Umile et al. (2014), we sampled amphibian skin for bacteria and metabolite compounds. We extracted and sequenced bacterial 16S rRNA from skin swabs to identify the taxonomic diversity and relative abundance of cutaneous bacterial communities from wild-caught adult amphibians.

Using high performance liquid chromatography (HPLC) we assessed the relative composition of metabolite compounds on these same individuals. Together, the results show that amphibian species may have very specific assemblages of both bacteria and metabolites. In regard to the current global chytridiomycosis pandemic affecting many amphibian populations, our results call 49

for further refinement of these techniques and broader sampling across different amphibian

species and regions to more clearly determine the role bacteria and metabolites play in

amphibian disease resistance.

MATERIALS AND METHODS

All metabolite and bacterial swabs were collected at the same time from individual live

amphibians using two separate, sterile swabs. In some cases individuals were only swabbed for

metabolites. The majority of samples came from amphibians collected by hand from five

different sites on Bioko Island, Equatorial Guinea from September 8–17, 2013. To avoid

contamination during capture events, medical-grade nitrile gloves were used during each capture.

Individuals were placed in separate plastic bags and stored overnight at ambient temperatures.

Swabbing was done the following morning in the same order in which the amphibians were

caught. Metabolite collection was done using polyurethane-tipped foam swabs (Fisher Scientific;

cat no. 14-960-3J), with a pretreatment as described by Umile et al. (2014). Bacteria collection

was done using individually wrapped sterile fine-tip swabs (MW113; Medical Wire &

Equipment Co., Wiltshire, England). To reduce possible contamination, swabbing took place in a

controlled environment (Moka Wildlife Center- Bioko Island, Equatorial Guinea). During

swabbing, individuals were secured by hand through the plastic bags without removing them to

reduce possible contamination. Each individual was swabbed inside their collection bag for 20 ̶

30 seconds across all ventral surfaces, including drink patch, thighs and feet. Medical-grade

nitrile gloves were used while swabbing and changed between each individual. To avoid possible bias caused by the first swab on small bodied frogs, the order in which frogs of a given species

were swabbed (metabolite vs. bacteria) was alternated for each successive individual. Both

metabolite and bacteria swabs were immediately placed into a sterile 1.5 ml vial, with vials for 50

Figure 1. Sampling localities for amphibian cutaneous bacteria and metabolites on Bioko Island, Equatorial Guinea. Sample site names derived from research camps operated by the Bioko Biodiversity Protection Program, and using Bioko Island map produced by Talleres del Instituto Geograficio Nacional – Deposito Legal: Madrid 7955 – 1983.

aiis1 eea17 Species 13 Genera 9 Families Rhacophoridae Ranidae Ptychadenidae Pipidae Hylidae Petropedetidae Bufonidae Bac. Swabs: Met. Swabs Bac. Swabs Abbreviation Species Hyperoliidae Genus Arthroleptidae Family C for Sampled Species Familiesand Anuran 1. Table Number of individuals swabbed bacteria; for swabbed individuals of Number eou laevis fraseri Xenopus Xenopus hrmni rufescens (eggs) rufescens Chiromantis Chiromantis fiau paradorsalis (eggs) paradorsalis Afrixalus Afrixalus variabilis bioko Arthroleptis Arthroleptis aasphenocephala albolabris Rana Hylarana tcaeaaequiplicata Ptychadena iuaasp. nov. new tetraploid 2 Silurana hyoarcu auritus Phrynobatrachus afra Nectophryne yeoistuberculatus sp. (eggs/developing tadpoles) ocellatus endjami Hyperolius Hyperolius Hyperolius Hyperolius brevirostris Leptopelis H yla cinerea utaneous Bacteria and Metabolites and Bacteria utaneous Met. Swabs: Number of individuals swabbed for for metabolites swabbed individuals of Number eg11 7 1 6 Cegg Cruf sh6- 16 6 13 Rsph Halb aq77 7 Paeq le2- 5 3 2 3 - Xlae Xfra Snt2 ar11 1 Paur 1 1 Nafr tb910 3 9 8 9 2 3 19 6 7 Htub 2 Hegg 16 Hoce Hend 1 Aegg 5 Apar 3 1 2 2 Lbre Avar Abio cn3 Hcin 1101 91 -

52 bacteria containing 95% EtOH. All vials were then immediately frozen until processing. After

swabbing, amphibians were released healthy to the collection site within 24 hours of capture,

except when voucher specimens were taken. All amphibians were collected following Drexel

IACUC Protocol #18748. 11 of the amphibians from which swab samples were taken were purchased from Carolina Biological Supply (Burlington, North Carolina) to serve as a

comparison between wild and lab-reared individuals. They were swabbed on November 11, 2013

immediately upon arrival at Drexel University, Philadelphia, PA. The purchased species were:

Hyla cinerea (3), Rana sphenocephala (6), Xenopus laevis (2). For the same comparative purposes, five individuals of the species Hyperolius tuberculatus collected from Lago Biao were

transported live to the United States where they were held in captivity at Cornell University,

Ithaca, NY. These five specimens were swabbed on November 7, 2013 after spending 52 days in

captivity. In total, 101 metabolite swabs and 99 bacteria swabs were taken. However, eight of the bacteria swabs were omitted due to failed DNA extraction/amplification.

Processing Swabs

All metabolite swabs were processed in the lab of Dr. Kevin Minbiole at Villanova

University. Concentrated swab extracts were reconstituted in methanol, and 25ul of this extract

was analyzed by reversed phase, high performance liquid chromatography (HPLC) using a

Shimadzu LC-20 liquid chromatograph equipped with an ACE C18 column (3 um, 150 x 4.6

mm) and SPD-M20A diode array detector. Compounds were characterized by retention time and,

where applicable, characteristic UV-Vis chromophores. The detailed protocol used for extracting

small molecules from swabs and subsequent HPLC analysis can be found in Umile et al. (2014).

DNA from cutaneous bacteria swabs was extracted using the DNeasy Blood & Tissue Kit

(Qiagen) according to the manufacturer's instructions, with some pre-protocol sample 53

modifications. Specifically, we evaporated existing ethanol in each sample vial by carefully

cleaning the outside of the tube with ethanol, then unscrewing vial caps enough to allow the

ethanol to evaporate from inside the tube but prevent any possible airborne contamination. Vials

with unscrewed caps were placed in an incubator (40°C) until ethanol was evaporated. After all

ethanol had evaporated, we added enzymatic lysis buffer (180 ul) to the vials, leaving the swab

in the vial, and vortexed for 30 s, and centrifuged (8000 rpm; 30 s). The swab tip was then

inverted, with tip pointing toward the top of the tube, and the vial was again vortexed for 30 s,

and centrifuged (8000 rpm; 30 s). Two vortexing steps were done to maximize the release of bacteria from the swabs into the lysis buffer, and the swab inversion kept the swab tips from reabsorbing the lysis buffer after the final centrifugation. The swab was removed and discarded from the vial, and the resulting pellet was thoroughly re-suspended in the lysis buffer by pipette.

The vial was capped and placed in an agitating incubator (56°C; 30 min). Extraction then proceeded according to the Qiagen protocol for Gram-positive bacteria, however only 100 ul of buffer AE was used during the elution step of this protocol.

To estimate the concentration of bacterial 16S rRNA genes in individual samples, the

extracted DNA was amplified using universal PCR primers for the eubacterial 16S rRNA gene,

(9Fa and 1513R (Russell et al. 2009)) at a concentration of 500 nM each, MyTaq HS Red Mix

(Bioline), and the following thermocycler protocol: initial denaturation at 95°C for 1 min,

followed by 40 cycles of denaturation at 95°C for 15 s, annealing at 55°C for 15 s and extension

at 72°C for 20 s, this was followed by a final extension at 72°C for 2 min. PCR products were

run on a 1% agarose gel to assess concentration and suitability for sequencing; only eight of 99

samples failed to produce a visible band, and these samples were excluded from later sequencing.

The 91 samples which produced visible bands considerably brighter than those from included 54

negative controls (negatives from blank swabs in the field and during extractions in the lab) were

submitted to Argonne National Laboratory (Lemont, IL) along with blank extractions for 16S

rRNA amplicon library preparation and sequencing. Amplicon libraries were prepared following

the Earth Microbiome Project protocols (http://www.earthmicrobiome.org/emp-standard- protocols/16s/), and sequenced in a multiplexed MiSeq 2 x 150bp run.

Data was analyzed using a customized pipeline in mothur v. v.1.33.3 (Schloss et al. 2009).

After read assembly, strict quality filtering and generation of a table of unique genotypes, we

removed those unique genotypes which were abundant in negative controls (assumed

contaminants amplified from blank extractions) and those which were rare in the whole dataset

(noisy reads) using a custom script (P. Lukasik, in preparation). The purpose of this step was to

reduce the effects of contamination with bacterial DNA derived from reagents (Salter et al. 2014).

We then proceeded to chimera removal using uchime, followed by classification of taxonomy via

the GenBank sequence database. Operational taxonomic units (OTUs) were identified at 97%

sequence similarity using the average neighbor clustering algorithm. Using this data set, we produced a table of non-chimeric OTUs, containing information on abundance and taxonomy of bacteria from all samples in our study.

Statistical Analysis

Steinhaus distances were generated in R v3.2.0 using metabolite data (presence/absence)

data by species, allowing us to compare metabolite profiles between amphibian species. Principal

coordinates analyses (PCoA) were performed on the resulting Steinhaus distance matrices using

the dsvdis() function in the labdsv package in R. We also performed permutational MANOVA

(1000 permutations) with Vegan function Adonis to assess whether observed differences between species metabolite profiles were significant. 55

Bray-Curtis distances were generated in R v3.2.0 using cutaneous bacteria data (97%

OTU composition), allowing us to compare cutaneous bacterial profiles between amphibian

species. Principal coordinates analyses (PCoA) were performed on the resulting Bray-Curtis

distance matrices using the dsvdis() function in the labdsv package in R. We also performed permutational MANOVA (1000 permutations) with Vegan function Adonis to assess whether

observed differences between species bacterial profiles were significant.

For both the metabolite and bacterial analyses, the specific species of a given sample was

abbreviated using the first letter of its genus and first three letters of its species (Table 2). In

some instances, the sample name may refer to an egg mass, a blank control, and/or include

collection locality in parenthesis after the designated species abbreviation.

Table 2. Genus:Species Abbreviations for Sample ID Family Genus:Species Abbreviation Arthroleptidae Arthroleptis (egg mass) Aegg Arthroleptis bioko Abio Arthroleptis variabilis Avar Leptopelis brevirostris Lbre Bufonidae Nectophryne afra Nafr Hylidae Hyla cinerea Hcin Hyperoliidae Afrixalus paradorsalis Apar Hyperolius (egg mass) Hegg Hyperolius endjami Hend Hyperolius ocellatus Hoce Hyperolius tuberculatus Htub Phrynobatrachidae Phrynobatrachus auritus Paur Pipidae Silurana new tetraploid 2 Snt2 Xenopus fraseri Xfra Xenopus laevis Xlae Ptychadenidae Ptychadena aequiplicata Paeq Ranidae Hylarana albolabris Halb Rana sphenocephala Rsph Rhacophoridae Chiromantis (egg mass) Cegg Chiromantis rufescens Cruf control (blank swab) ctrl 56

RESULTS

Metabolites: Adonis and Principal Coordinates Analysis (PCoA)

We analyzed the variation in metabolite presence/absence for all of the 142 distinct

compounds detected by HPLC analysis across all samples. The first aimed to detect any species-

specific trends, while the second used factors of species by locality to determine if location may be a factor influencing metabolite profiles. Each is described in detail below.

The initial analysis was done using all samples to examine major trends in metabolite profiles across sampled species, and included 95 amphibians and six egg masses across 14

different species (Fig. 2). Adonis analyses indicated that a considerable amount of the variation

in metabolite profiles can be explained by species specificity (Steinhaus: F = 2.848, R2 = 0.352,

P = 0.0001, Deg. Freedom = 16, Residual Deg. Freedom = 84). To better visualize the variation between species, we plotted the results of a PCoA on Steinhaus distances. Most notably,

Afrixalus paradorsalis (Apar) samples cluster together and have little overlap with other species.

Hylarana albolabris (Halb) and Ptychadena aequiplicata (Paeq) also showed overlapping clusters that were distinct from other species, except for the single sample of Phrynobatrachus

auritus (Paur) which was also within these overlapping clusters. Additional overlapping clusters

can be seen in samples of Arthroleptis variabilis (Avar), Arthroleptis bioko (Abio), and

Hyperolius tuberculatus (Htub), and separately from those were clusters of Hyperolius endjami

(Hend), Hyperolius ocellatus (Hoce), and Chiromantis rufescens (Cruf). The few additional

samples shown were inconclusive in the relationship as they were either not clustered very

tightly or represented single samples from a given species. Overall, the results suggest that 35%

of the variability we can see in the PCoA is attributable to species, and given the possibility that

many metabolite swabs detected a great deal of environmental (non-amphibian derived) 57

Figure 2. Principal coordinates analysis (PCoA) plot of cutaneous metabolites from 95 amphibians and six egg masses across 14 different species from Bioko Island.

metabolites, these data suggest strong support for species-specific metabolite profiles. Future

analyses may attempt to eliminate the low abundance metabolites from the analysis, which are

likely environmental contaminants. Additional steps to reduce contaminants in the field, such as

rinsing amphibians before swabbing, may also help to reduce non-amphibian derived metabolites

and reveal a stronger species-specific correlation to metabolite profiles.

A secondary analysis was done using a subset of samples to examine the possible contribution of

locality to the metabolite profile within and between species (Fig. 3). We chose only species that

had high sample sizes and which were distributed across many sites, which included 62

amphibians, with 12 Species:Locality combinations created from five different species collected

across five different localities (Localities: Risule, Riaba, Arena Blanca, Lago Loreto, and Lago

Biao; Species: Afrixalus paradorsalis (Apar) , Hylarana albolabris (Halb) , Hyperolis endjami

(Hend) , Hyperolis ocellatus (Hoce), and Hyperolis tuberculatus (Htub)). Adonis analyses indicated that only a slight amount of the variation in metabolite profiles can be explained by the interaction between species and locality. The variation explained by species alone was more than twice that explained by locality, again suggesting species is the most influential factor in determining metabolite profiles (Steinhaus for Species: F = 5.831, R2 = 0.264, P = 0.0001, Deg.

Freedom = 4; Steinhaus for Locality: F = 3.004 , R2 = 0.102, P = 0.0001, Deg. Freedom = 3;

Steinhaus for Species:Locality Interaction: F = 1.520, R2 = 0.069, P = 0.0114, Deg. Freedom =

4; Residual Deg. Freedom = 54). To better visualize the variation between species, we plotted the results of a PCoA performed on Steinhaus distances. A. paradorsalis again showed a distinct

cluster, though samples from Arena Blanca and Riaba seem to be concentrated on one side of the

grouping as compared to those from Risule. This variation seems slight, and likely does not

imply any major difference in metabolite profile due to locality. In all other Species:Locality 59

Figure 3. Principal coordinates analysis (PCoA) of cutaneous metabolites from 62 amphibians. Each abbreviated sample ID includes collection locality in parenthesis. 12 Species:Locality combinations were created across five different species and four localities from Bioko Island.

groupings, there was no discernable within-species differences based on locality, supported by

the Adonis analyses showing only 10.2% of the variation being attributable to locality, likely

from environmental contaminants in swabs as previously discussed. With 26.4% of the variation

explained by species alone, and assuming a fair amount of environmental contaminants during

sampling, this shows further support for species-specific metabolite profiles.

Bacteria (97% OTU Composition): Adonis and Principal Coordinates Analysis (PCoA)

This analysis utilized the relative abundance of the 12 most common bacterial 16S rRNA

OTUs (97% OTU similarity) to determine variation by species and/or locality (Figs. 4, 5).

Samples consisted of 85 individuals and six egg masses (with three blank controls) across 16

species and five localities. Individuals of Afrixalus paradorsalis (Apar), Hylarana albolabris

(Halb), Hyperolius ocellatus (Hoce), and Hyperolius tuberculatus (Htub) were swabbed at two

different collection localities to assess the contribution that locality may have upon their bacterial profile. Of these, five H. tuberculatus specimens collected from Lago Biao were transported

alive to Cornell University, Ithaca, NY, where they were held in captivity before being swabbed

to determine if captivity had a noticeable effect upon their bacterial profile. These individuals

were not swabbed in the field, but four H. tuberculatus from Lago Loreto were swabbed in the

field to use as an approximate comparison.

Adonis analyses indicated that over half of the variation in the relative abundance of bacteria can be explained by species alone, while locality and the interaction of species and

locality both accounted for only a slight amount of this variation (Bray Curtis for Species: F =

4.830, R2 = 0.517, P = 0.0001, Deg. Freedom = 18; Bray Curtis for Locality: F = 4.779 , R2 =

0.057, P = 0.0001, Deg. Freedom = 2; Steinhaus for Species:Locality Interaction: F = 1.838, R2

= 0.022, P = 0.0209, Deg. Freedom = 2; Residual Deg. Freedom = 68). To better visualize the 61

Figure 4. Relative abundance of the 12 most common bacterial taxa (97% OTU similarity at 16S rRNA) occurring in all amphibian samples. Each sample includes collection locality in parenthesis. Samples consisted of 85 individuals, six egg masses, and three negative controls across 16 species and five localities from Bioko Island. 62

Figure 5. PCoA plots of relative abundance for the 12 most common bacterial taxa (97% OTU similarity of 16S rRNA) occurring in all amphibian samples. Samples were plotted in separate but comparable plots within their respective collection locality. Samples consisted of 85 individuals and 6 egg masses across 16 species and 5 localities on Bioko Island. Four species (Apar, Halb, Hoce, Htub) were swabbed at two different localities.

variation between skin communities of different species by location, we constructed graphs

showing the relative abundance of the 12 most common bacterial OTU’s for each sample (Fig. 4).

Visual inspection of these graphs shows that OTUs which are the most abundant across the

samples also tend to occur with the greatest species-specificity. For example, when OTU001

does occur in samples, it typically occurs at a higher relative abundance than other OTUs and

show strict species specificity, regardless of locality. OTU001 is notably absent in all three

Arthroleptidae species, as well as in H. tuberculatus , X. fraseri , and P. aequiplicata . Similarly,

OTU002 and OTU003 seem to show relatively strict species specificity, and in these instances

they are often found at relatively high abundance in those species’ samples. However, there are

some instances where OTU002 and OTU003 do occur at very low abundance in some species’

samples but not others. The single sample of a Chiromantis rufescens egg mass showed

considerable OTU diversity, which may be a significant finding for future study, as their egg

masses are covered in foam that may host a variety of protective bacteria for developing embryos.

The OTU category labeled as “other” represents multiple bacterial taxa that are not defined here by a single OTU category, yet collectively represented a significant portion of the bacterial

abundance found in some samples. The high incidence of this “other” OTU in X. fraseri samples suggest that its microbiota may be considerably more diverse, possibly owing to the type of habitats it lives in (mud, stagnant pools) and/or the wider variety of environmental bacteria that may have been sampled during swabbing of this species.

Using the relative abundances in each sample, we also plotted the results of a PCoA performed on Bray–Curtis distances (Fig. 5). Samples were plotted on five comparable axes according to collection locality. Distinct clusters by species are apparent, though some species clusters overlap with others. This suggests some level of bacterial similarity across many species, 64

which is apparent in the case of OTU001 in the graphs from Fig. 4. Species showing relatively

tight clusters include A. bioko , A. variabilis , A. paradorsalis , C. rufescens , H. albolabris , H. ocellatus , H. tuberculatus , and P. aequiplicata . One notable insight from the PCoA is the

separate clustering in two species purchased from Carolina Biological Supply (Locality plot

USA- CB), H. cinerea and R. sphenocephala . This suggests each species not only has a distinct bacterial community, but that the species-specific variation in these communities is not

homogenized through captivity. Overall, the relative abundance of bacteria visualized in the

graphs and PCoA clustering, along with the Adonis results showing that species account for

51.7% of the variation in bacterial profiles, provide support for a high level of species-specificity

in cutaneous bacterial profiles of amphibians.

DISCUSSION

The results strongly suggest that the amphibian cutaneous environment is not simply a

reflection of environmentally derived compounds and bacteria , and that there is a significant

degree of amphibian species specificity in both the metabolite and bacterial profiles. This also

suggests that the amphibian cutaneous environment is somehow regulated to select for this

specificity. Given the nature of amphibian skin and the typical habitats amphibians are found in, preventing bacterial and fungal infection is presumably an important function of these cutaneous bacteria and metabolites. The specific role of cutaneous bacteria, and their additive effects on the amphibian, needs to be further explored via laboratory assays. This study provides the basis upon which future research can attempt to further characterize of the most common cutaneous bacteria, which may be critical in conservation efforts against amphibian diseases such as chytridiomycosis. This work may also help to reveal unknown symbiotic relationships, and the possible role of these symbiotic relationships in allowing amphibians to exploit new niches as they diversified over the past 400 million years. Finally, this study provides practical application of newly developed techniques for non-lethal sampling of amphibian skin compounds, and may prove to be an effective method for fingerprinting species to reveal species-level diversity in cryptic species complexes (Umile et al. 2014). Results from a 2012 pilot study intended to test this new metabolite swab technique showed two separate metabolite clusters for H. ocellatus .

This led the author to further investigate photos and vouchers specimens of these H. ocellatus ,

which revealed that samples actually consisted of two extremely similar looking species, H. ocellatus and H. endjami . The latter species had not yet been recorded from Bioko, and this

discovery represented a considerable range expansion for a species listed as vulnerable by the

IUCN Red List. 66

FURTHER ANALYSIS

Initial analysis of these data has revealed unique bacteria and metabolite communities

across multiple tropical amphibians for the first time. Neither bacteria nor metabolites appear to be dictated by habitat or site-specific dynamics. However, with such a robust data set there are

additional questions that can be investigated.

One of the most important of these is to investigate additional within-OTU diversity

across species. While OTU’s are useful in characterizing basic bacteria taxonomy, they do not

necessarily represent exact species designations. Some within-OTU diversity may actually

represent species-level variation, and thus functional variation as it relates to amphibians. Further

analysis of the 16S rRNA sequence data should focus on characterizing within-OTU diversity.

Additional taxonomic information for each bacterial OTU, cross referenced against known bacterial 16S rRNA databases, may also provide insight into the specific function of some

species we detected. This may provide useful in the identification those species that may offer

antifungal resistance. A combined analysis of the presence/absence metabolite data and bacterial

OTU sequence data could be used to identify correlations between metabolites and particular bacterial taxa. An additional model that attempts to combine a sample’s total bacterial and

metabolite profile together may also be useful in characterizing taxa-specific trends in skin

environment diversity and ultimately its relative effectiveness in regard to disease resistance.

Finally, future studies should build on this work to investigate whether bacteria in amphibians

are primarily transferred vertically or horizontally, and if the acquisition of new bacterial

symbionts over time may have led to amphibian diversification over evolutionary time.

LIST OF REFERENCES

Amiet, J.-L. (2012). Les rainettes du Cameroun (Amphibiens Anoures) . la Nef des livres, France.

Anthony, N. M., Atteke, C., Bruford, M. W., Dallmeier, F., et al. (2015). Evolution and Conservation of Central African Biodiversity: Priorities for Future Research and Education in the Congo Basin and Gulf of Guinea. Biotropica, 47, 1, 6-17.

Balaz, V., Kopecky, O., & Gvozdik, V. (2012). Presence of the amphibian chytrid pathogen confirmed in Cameroon. Herpetological Journal, 22, 3, 191-194.

Barej, M., Schmitz, A., Günther, R., Loader, S.P., et al. (2014). The first endemic West African vertebrate family - a new anuran family highlighting the uniqueness of the Upper Guinean biodiversity hotspot. Frontiers in Zoology , 11 , 8.

Barej, M. F., Bohme, W., Rodel, M.-O., Gonwouo, L. N., et al. (2010). Review of the genus Petropedetes reichenow , 1874 in Central Africa with the description of three new species (Amphibia: Anura: Petropedetidae). Zootaxa, 2340, 1-49.

Bataille, A., Fong, J. J., Cha, M., Wogan, G. O., et al. (2013). Genetic evidence for a high diversity and wide distribution of endemic strains of the pathogenic chytrid fungus Batrachochytrium dendrobatidis in wild Asian amphibians. Molecular Ecology, 22, 16, 4196-209.

Bd -Maps Species Summaries. http://www.bd-maps.net/surveillance/default.asp

Bd -Maps. http://www.bd-maps.net/

Bell, R. C., Drewes, R. C., & Zamudio, K. R. (2015). Reed frog diversification in the Gulf of Guinea: Overseas dispersal, the progression rule, and in situ speciation. Evolution, 69, 4, 904-915.

Berger, L., Speare, R., Daszak, P., Green, D. E., et al. (1998). Chytridiomycosis causes amphibian mortality associated with population declines in the rain forests of Australia and Central America. Proceedings of the National Academy of Sciences of the United States of America, 95, 15, 9031-6.

Berger, L., Speare, R., & Skerratt, L. F. (2005). Distribution of Batrachochytrium dendrobatidis and pathology in the skin of green tree frogs Litoria caerulea with severe chytridiomycosis. Diseases of Aquatic Organisms, 68, 1, 65-70.

Bergl, R.A., Oates, J.F., and R. Fotso. (2007). Distribution and protected area coverage of endemic taxa in West Africa’s Biafran forests and highlands. Biological Conservation, 134 , 195–208.

Blackburn, D. C. (2010). A New Squeaker Frog (Arthroleptidae: Arthroleptis ) from Bioko Island, Equatorial Guinea. Herpetologica, 66, 3, 320-334.

Blackburn, D. C., & Rödel, M.O. (2011). A New Puddle Frog (Phrynobatrachidae: Phrynobatrachus ) from the Obudu Plateau In Eastern Nigeria. Herpetologica, 67, 3, 271- 287.

Blackburn, D. C., Gvozdík, V., & Leaché, A. D. (2010). A New Squeaker Frog (Arthroleptidae: Arthroleptis ) from the Mountains of Cameroon and Nigeria. Herpetologica, 66, 3, 335- 348.

Bletz, M. C., Rosa, G. M., Andreone, F., Courtois, E. A., et al. (2015). Widespread presence of the pathogenic fungus Batrachochytrium dendrobatidis in wild amphibian communities in Madagascar. Scientific Reports, 5.

Bletz, M. C., Rebollar, E. A., & Harris, R. N. (2015). Differential efficiency among DNA extraction methods influences detection of the amphibian pathogen Batrachochytrium dendrobatidis . Diseases of Aquatic Organisms, 113, 1, 1-8.

Boulenger, G. A. (1906). Report on the batrachians collected by the late L. Fea in West Africa . Annali del Museo Civico di Storia Naturale di Genova. Serie 3, 2, 157–172.

Channing, A. (2001). Amphibians of Central and Southern Africa . Ithaca, N.Y: Comstock Pub. Associates.

Cronin, D. T., Libalah, M. B., Bergl, R. A., & Hearn, G. W. (2014). Biodiversity and Conservation of Tropical Montane Ecosystems in the Gulf of Guinea, West Africa. Arctic, Antarctic, and Alpine Research, 46, 4, 891-904.

Cunningham, A. A., Garner, T. W., Aguilar-Sanchez, V., Banks, B., et al. (2005). Emergence of amphibian chytridiomycosis in Britain. The Veterinary Record, 157, 13, 386-7.

Easteal, S. (1981). The history of introductions of Bufo marinus (Amphibia: Anura); a natural experiment in evolution. Biological Journal of the Linnean Society, 16, 2, 93-113.

Eisentraut, M. (1973). Die Wirbeltierfauna von Fernando Poo und West Kamerun. Bonner Zoologische Monographien, 31–428.

Farrer, R. A., Weinert, L. A., Bielby, J., Garner, T. W., et al. (2011). Multiple emergences of genetically diverse amphibian-infecting chytrids include a globalized hypervirulent recombinant lineage. Proceedings of the National Academy of Sciences of the United States of America, 108, 46, 18732-6.

Fisher MC, Garner TW. (2007). The relationship between the introduction of Batrachochytrium dendrobatidis , the international trade in amphibians and introduced amphibian species. Fungal Biol. Rev . 21 , 2–9.

Fisher, M. C., Garner, T. W., & Walker, S. F. (2009). Global emergence of Batrachochytrium dendrobatidis and amphibian chytridiomycosis in space, time, and host. Annual Review of Microbiology, 63, 291-310.

Frost, D. R., Duellman, W. E., & Association of Systematics Collections. (1985). Amphibian species of the world: A taxonomic and geographical reference . Lawrence, Kan: Allen Press.

Frost, D. R. (2015). Amphibian Species of the World: an Online Reference. Version 6.0 (Accessed: May 2015). American Museum of Natural History, New York, USA. Electronic Database accessible at http://research.amnh.org/herpetology/amphibia/index.html .

Gahl, M.K., Longcore, J.E., & Houlahan, J. E. (2012). Varying Responses of Northeastern North American Amphibians to the Chytrid Pathogen Batrachochytrium dendrobatidis . Conservation Biology, 26, 1, 135-141.

Goldberg, T. L., Readel, A. M., & Lee, M. H. (2007). Chytrid Fungus in Frogs from an Equatorial African Montane Forest in Western Uganda. Journal of Wildlife Diseases, 43, 3, 521-524.

Gower, D. J., Doherty-Bone, T. M., Aberra, R. K., Mengistu, A., et al. (2012). High prevalence of the amphibian chytrid fungus ( Batrachochytrium dendrobatidis ) across multiple taxa and localities in the highlands of Ethiopia. Herpetological Journal, 22, 4, 225-233.

Gratwicke, B., Alonso, A., Elie, T., Kolowski, J., et al. (2011). Batrachochytrium dendrobatidis not detected on amphibians from two lowland sites in Gabon, Africa. Herpetological Review, 42, 1, 69-71.

Greenbaum, E., Kusamba, C., Aristote, M. M., & Reed, K. D. (2008). Amphibian Chytrid Fungus Infections in Hyperolius (Anura: Hyperoliidae) from Eastern Democratic Republic of Congo. Herpetological Review, 39, 1, 70-121.

Greenbaum, E., Meece, J., Reed, K. D., Kusamba, C. (2015). Extensive occurrence of the amphibian chytrid fungus in the Albertine Rift, a Central African amphibian hotspot. The Herpetological Journal, 25 , 2, 91-100(10).

Guinea, E. (1968). Fernando Po. In Conservation of Vegetation in Africa South of the Sahara. (I. Hedberg and O. Hedberg, eds) Acta Phytogeographica Suecica, 54 , 30–132.

Harris, R. N., Brucker, R. M., Walke, J. B., Becker, M. H., et al. (2009). Skin microbes on frogs prevent morbidity and mortality caused by a lethal skin fungus. The Isme Journal, 3, 7, 818-24. Hyatt, A. D., Boyle, D. G., Olsen, V., Boyle, D. B., et al. (2006). Diagnostic assays and sampling protocols for the detection of Batrachochytrium dendrobatidis . Diseases of Aquatic Organisms, 73, 3, 175.

Hydeman, M. E., Bell, R. C., Zamudio, K. R., & Drewes, R. C. (2013). Amphibian chytrid fungus confirmed in endemic frogs and caecilians on the Island of São Tomé, Africa. Herpetological Review, 44, 2, 254-257.

Imasuen, A. A., Aisien, M. S. O., Weldon, C., Du, P. L. H., et al. (2011). Occurrence of Batrachochytrium dendrobatidis in amphibian populations of Okomu national park, Nigeria. Herpetological Review, 42, 3, 379-382.

Imasuen, A.A., Weldon, C., Aisien, M.S.O., Dupreez, L.H. (2009) Amphibian chytridiomycosis: first report in Nigeria from the skin slough of Chiromantis rufescens . Froglog , 90 , 6-8.

International Union for Conservation of Nature and Natural Resources. (2000). IUCN red list of threatened species . Cambridge, U.K.: IUCN. http://www.iucnredlist.org/

IUCN. (2012). IUCN Red List Categories and Criteria: Version 3.1. Second Edition. Gland, Switzerland and Cambridge, UK: IUCN. iv + 32pp.

James, T. Y., Litvintseva, A. P., Vilgalys, R., Morgan, J. A., et al. (2009). Rapid global expansion of the fungal disease chytridiomycosis into declining and healthy amphibian populations. Plos Pathogens, 5, 5.

Jani, A. J., & Briggs, C. J. (2014). The pathogen Batrachochytrium dendrobatidis disturbs the frog skin microbiome during a natural epidemic and experimental infection. Proceedings of the National Academy of Sciences of the United States of America, 111, 47, 5049-58.

Jones, P. J. (1994). Biodiversity in the Gulf of Guinea: an overview. Biodiversity & Conservation, 3, 9, 772-784.

Kielgast, J., Rödder, D., Veith, M., & Lötters, S. (2010). Widespread occurrence of the amphibian chytrid fungus in Kenya. Animal Conservation, 13, 36-43.

Köhler, J. (2009). The Identity of Hylambates rufus aubryioides Andersson, 1907 (Anura: Arthroleptidae) from Cameroon. Copeia, 2009, 1, 57-61.

Kriger, K. M., & Hero, J.-M. (2007). Large-scale seasonal variation in the prevalence and severity of chytridiomycosis. Journal of Zoology, 271, 3, 352-359. 71

Kriger, K. M., & Hero, J.-M. (2008). Altitudinal distribution of chytrid ( Batrachochytrium dendrobatidis ) infection in subtropical Australian frogs. Austral Ecology, 33, 8, 1022- 1032.

Kueneman, J. G., Parfrey, L. W., Woodhams, D. C., Archer, H. M., et al. (2014). The amphibian skin-associated microbiome across species, space and life history stages. Molecular Ecology, 23, 6, 1238-50.

Lenker, M. A., Savage, A. E., Becker, C. G., Rodriguez, D., et al. (2014). Batrachochytrium dendrobatidis infection dynamics vary seasonally in upstate New York, USA. Diseases of Aquatic Organisms, 111, 1, 51-60.

Lettoof, D. C., Greenlees, M. J., Stockwell, M., & Shine, R. (2013). Do invasive cane toads affect the parasite burdens of native Australian frogs?. International Journal for Parasitology. Parasites and Wildlife, 2, 155-64.

Lever, C. (2003). Naturalized reptiles and amphibians of the world . Oxford: Oxford University Press.

Lever, C. (2001). The cane toad: The history and ecology of a successful colonist . Otley: Westbury Academic and Scientific Publishing.

Lips, K. R., Brem, F., Brenes, R., Reeve, J. D. (2006). Emerging infectious disease and the loss of biodiversity in a Neotropical amphibian community. Proceedings of the National Academy of Sciences of the United States of America, 103, 9, 3165-70.

Lips, K. (2014). A tale of two lineages: unexpected, long-term persistence of the amphibian- killing fungus in Brazil. Molecular Ecology, 23, 4, 747-749.

Longcore, J.E., Pessier, A.P., Nichols, D.K. (1999) Batrochytrium dendrobatidis gen. et sp. nov., a chytrid pathogenic to amphibians. Mycologia , 91 , 219-227.

Longo, A. V., Burrowes, P. A., & Joglar, R. L. (2010). Seasonality of Batrachochytrium dendrobatidis infection in direct-developing frogs suggests a mechanism for persistence. Diseases of Aquatic Organisms, 92, 253-260.

Lötters, S., Rödel M.-O., and Burger M. (2005). A new large tree frog from north-western Gabon (Hyperoliidae : Leptopelis). Herpetological Journal . London, 15 , 149–152.

Maddock, S. T., Day, J. J., Nussbaum, R. A., Wilkinson, M., et al. (2014). Evolutionary origins and genetic variation of the Seychelles treefrog, Tachycnemis seychellensis (Dumeril and Bibron, 1841) (Amphibia: Anura: Hyperoliidae). Molecular Phylogenetics and Evolution, 75, 194-201.

McCallum, H. (2005). Inconclusiveness of Chytridiomycosis as the Agent in Widespread Frog Declines. Conservation Biology, 19, 5, 1421-1430.

McKenzie, V. J., Bowers, R. M., Fierer, N., Knight, R., et al. (2012). Co-habiting amphibian species harbor unique skin bacterial communities in wild populations. The Isme Journal, 6, 3, 588-96.

Measey, G.J., Vences, M., Drewes, R.C., Chiari, Y., Melo, M., et al. (2007). Freshwater paths across the ocean: molecular phylogeny of the frog Ptychadena newtoni gives insights into amphibian colonization of oceanic islands. Journal of Biogeography , 34 , 7–20.

Moyer, D., and C. Weldon. (2006). Chytrid distribution and pathogenisity among frogs of the Udzungwa Mountains Tanzania. Final report. CEPF Small Grant Final Project Completion. March 2006. Available at: www.cepf.net/Documents/Final.Chytrid.report.pdf .

Myers, N., R. A. Mittermeier, C. G. Mittermeier, do Fonseca, G.A.B., Kent, J. (2000). Biodiversity hotspots for conservation priorities. Nature , 403 , 853-858.

Nace, G.W., Waage, J.K., and Richards, C.M. (1971) Sources of Amphibians for Research. BioScience , 21 , 14, 768-773.

Oates, J. F., Bergl, R. A., Linder, J. M., Wildlife Conservation Society (New York, N.Y.), Center for Applied Biodiversity Science., & BioOne (Organization). (2004). Africa's Gulf of Guinea forests: Biodiversity patterns and conservation priorities . (BioOne.) Washington, D.C: Conservation International.

Olson, D. H., Aanensen, D. M., Ronnenberg, K. L., Powell, C. I., et al., & Bd Mapping Group. (2013). Mapping the global emergence of Batrachochytrium dendrobatidis , the amphibian chytrid fungus. Plos One,8, 2.

Ouellet M., Mikaelian I., Pauli B.D., Rodrigue J., & Green D.M. (2005). Historical Evidence of Widespread Chytrid Infection in North American Amphibian Populations. Conservation Biology, 19, 5, 1431-1440.

Palumbi, S.R., Martin, A., Romano, S., McMillan, W.O., Stice, L., and Grabowski, G. (2002). The Simple Fool's Guide to PCR, Version 2.0 . Privately Published, University of Hawaii.

Parker, J. M., Mikaelian, I., Hahn, N., & Diggs, H. E. (2002). Clinical diagnosis and treatment of epidermal chytridiomycosis in African clawed frogs ( Xenopus tropicalis ). Comparative Medicine, 52, 3, 265-8.

Penner, J.W., Wegmann, M., Hillers, A., Schmidt, M., et al. (2011). A hotspot revisited - a biogeographical analysis of West African amphibians. Diversity and Distributions , 17 , 1077–1088.

Penner, Johannes, Gilbert B. Adum, Matthew T. McElroy, Thomas Doherty-Bone, Mareike Hirschfeld, Laura Sandberger, Ché Weldon, et al. (2013). "West Africa - A Safe Haven for Frogs? A Sub-Continental Assessment of the Chytrid Fungus ( Batrachochytrium dendrobatidis )". PLoS ONE. 8 (2): e56236.

Piotrowski, J.S., Annis, S.L., & Longcore, J.E. (2004). Physiology of Batrachochytrium dendrobatidis , a chytrid pathogen of amphibians. Mycologia , 96 , 9–15.

Portillo, F., & Greenbaum, E. (2014). At the edge of a species boundary: A new and relatively young species of Leptopelis (Anura: Arthroleptidae) from the itombwe plateau, democratic republic of the congo. Herpetologica, 70, 1, 100-119.

Rachowicz, L.J., Hero, J.-M., Alford, R.A., Taylor, J.W., et al. (2005). The Novel and Endemic Pathogen Hypotheses: Competing Explanations for the Origin of Emerging Infectious Diseases of Wildlife. Conservation Biology, 19, 5, 1441-1448.

Reeder, N. M. M., Cheng, T. L., Vredenburg, V. T., & Blackburn, D. C. (2011). Survey of the chytrid fungus Batrachochytrium dendrobatidis from montane and lowland frogs in eastern Nigeria. Herpetology Notes, 4, 1, 83-86.

Rodriguez, D., Becker, C. G., Pupin, N. C., Haddad, C. F. B., et al. (2014). Long-term endemism of two highly divergent lineages of the amphibian-killing fungus in the Atlantic Forest of Brazil. Molecular Ecology, 23, 4, 774-787.

Roelants, K., Gower, D. J., Wilkinson, M., Loader, S. P., et al. (2007). Global patterns of diversification in the history of modern amphibians. Proceedings of the National Academy of Sciences of the United States of America, 104, 3, 887-92.

Rosenblum, E. B., James, T. Y., Zamudio, K. R., Poorten, T. J., et al. (2013). Complex history of the amphibian-killing chytrid fungus revealed with genome resequencing data. Proceedings of the National Academy of Sciences of the United States of America, 110, 23, 9385-90.

Russell, J. A., Moreau, C. S., Goldman-Huertas, B., Fujiwara, M., et al. (2009). Bacterial gut symbionts are tightly linked with the evolution of herbivory in ants. Proceedings of the National Academy of Sciences of the United States of America, 106, 50, 21236-41.

Salter, S. J., Cox, M. J., Turek, E. M., Calus, S. T., et al. (2014). Reagent and laboratory contamination can critically impact sequence-based microbiome analyses. Bmc Biology, 12.

Savage, A. E., Sredl, M. J., & Zamudio, K. R. (2011). Disease dynamics vary spatially and temporally in a North American amphibian. Biological Conservation, 144, 6, 1910-1915.

Schiøtz, Arne. (1999). Treefrogs of Africa . Frankfurt am Main: Edition Chimaira.

Schloegel, L. M., Toledo, L.F., Longcore, J. E., Greenspan, S. E., et al. (2012). Novel, panzootic and hybrid genotypes of amphibian chytridiomycosis associated with the bullfrog trade. Molecular Ecology, 21, 21, 5162-5177.

Schloss, P. D., Westcott, S. L., Ryabin, T., Hall, J. R., et al. (2009). Introducing mothur: open- source, platform-independent, community-supported software for describing and comparing microbial communities. Applied and Environmental Microbiology, 75, 23, 7537-41.

Shanmuganathan, T., Pallister, J., Doody, S., McCallum, H., et al. (2010). Biological control of the cane toad in Australia: a review. Animal Conservation, 13, 16-23.

Shin, J., Bataille, A., Kosch, T. A., & Waldman, B. (2014). Swabbing often fails to detect amphibian Chytridiomycosis under conditions of low infection load. Plos One, 9, 10.

Skerratt, L., Berger, L., Speare, R., Cashins, S., et al. (2007). Spread of Chytridiomycosis Has Caused the Rapid Global Decline and Extinction of Frogs. Ecohealth, 4, 2, 125-134.

Soto-Azat, C., Clarke, B. T., Poynton, J. C., & Cunningham, A. A. (2010). BIODIVERSITY RESEARCH: Widespread historical presence of Batrachochytrium dendrobatidis in African pipid frogs. Diversity & Distributions, 16, 1, 126-131.

Swofford, D.L. (2002). PAUP*: Phylogenetic Analysis Using Parsimony *(and other methods). Version 4.0a134. Sinauer Associates, Sunderland, Massachusetts.

Talley, B. L., Muletz, C. R., Vredenburg, V. T., Fleischer, R. C., et al. (2015). A century of Batrachochytrium dendrobatidis in Illinois amphibians (1888-1989). Biological Conservation, 182, 254-261.

Terán, M.D. (1962). Síntesis geográfica de Fernando Poo. Madrid, Instituto de Estudios Africanos, Consejo Superior de Investigaciones Científicas.

Umile, T. P., McLaughlin, P. J., Johnson, K. R., Honarvar, S., et al. (2014). Nonlethal amphibian skin swabbing of cutaneous natural products for HPLC fingerprinting. Anal. Methods, 6, 10, 3277-3284.

Vences, M., Vieites, D. R., Glaw, F., Brinkmann, H., et al. (2003). Multiple overseas dispersal in amphibians. Proceedings: Biological Sciences, 270, 1532, 2435-2442.

Vredenburg, V. T., Knapp, R. A., Tunstall, T. S., & Briggs, C. J. (2010). Dynamics of an emerging disease drive large-scale amphibian population extinctions. Proceedings of the National Academy of Sciences of the United States of America, 107, 21, 9689-94.

Vredenburg, V.T., Du Preez, L., Raharivololoniaina, L., Vieites, D. R., et al. (2012). A molecular survey across Madagascar does not yield positive records of the amphibian chytrid fungus Batrachochytrium dendrobatidis . Societas Europaea Herpetologica. Herpetology Notes , 5, 507- 517.

Wake, D. B., & Vredenburg, V. T. (2008). Are we in the midst of the sixth mass extinction? A view from the world of amphibians. Proceedings of the National Academy of Sciences of the United States of America, 105, 11466-11473.

Weldon, C., du, P. L. H., Hyatt, A. D., Muller, R., & Spears, R. (2004). Origin of the amphibian chytrid fungus. Emerging Infectious Diseases, 10, 12, 2100-5.

Weldon, C. (2005). Chytridiomycosis in South Africa. Ph.D Thesis. North-West University, Potchefstroom, South Africa.

Zafra-Calvo, N., Cerro, R., Fuller, T., Lobo, J. M., et al. (2010). Prioritizing areas for conservation and vegetation restoration in post-agricultural landscapes: A Biosphere Reserve plan for Bioko, Equatorial Guinea. Biological Conservation, 143, 3, 787-794.

Zilber-Rosenberg, I., & Rosenberg, E. (2008). Role of microorganisms in the evolution of animals and plants: the hologenome theory of evolution. Fems Microbiology Reviews, 32, 5, 723-735.

APPENDIX 1: Arthroleptis bioko IUCN Red List Assessment

APPENDIX 2: SPECIMENS EXAMINED

Leptopelis calcaratus : Equatorial Guinea: Bioko Sur Province: Gran Caldera & Southern Highlands Scientific Reserve, Pico Biao, NCSM 82319–20 (two adult males). Gabon: Ogooué- Ivindo Province: Ipassa, MCZ A-139741, NCSM 77691, NCSM 77697 (three adult males), MCZ A-139740, MCZ A-139756–57, MCZ A-139799, NCSM 77690, NCSM 77695 (six adult females); Kongou, NCSM 77693 (one adult male), NCSM 77694 (one adult female).

Vita

Patrick Joseph McLaughlin was born in Schenectady, NY on December 10 th , 1982. He attended Scotia-Glenville Senior High School (Scotia, NY) and upon graduation was accepted into St. Lawrence University (Canton, NY) where he focused his studies in ecology. He received a university fellowship to conduct research in the Bahamas with his advisor, Dr. Brad Baldwin, which led to a publication on the nursery-effect of mangrove lagoons for juvenile fish. Since graduating in 2005 with a B.S. in Biology and Environmental Studies, he has worked as a marine ecology instructor in the Florida Keys, a fisheries technician with the U.S. Fish and Wildlife Service in Vermont, a naturalist in Yellowstone National Park, and as a guide for National Geographic’s Student Expeditions. In 2009 he began fieldwork for his Ph.D. with Dr. Gail Hearn, conducting fieldwork in West Africa via the Bioko Biodiversity Protection Program. His research is focused on amphibian ecology, conservation, and fungal disease resistance. Over the past five years Patrick has acted as a research mentor to numerous undergraduate and masters students conducting independent projects on Bioko Island. Patrick has taught or co-taught a variety of courses such as General Biology (cell/molecular), Gross Anatomy, Population Ecology, Marine Ecology, Tropical Ecology, and Field Methods in Tropical Ecology. He served as President of the Biology Graduate Student Association from 2013-14.