Journal of Wildlife Diseases, 43(3), 2007, pp. 521–524 # Wildlife Disease Association 2007

Chytrid Fungus in Frogs from an Equatorial African Montane Forest in Western Uganda

Tony L. Goldberg,1,2,3 Anne M. Readel,2 and Mary H. Lee11Department of Pathobiology, University of Illinois, 2001 South Lincoln Avenue, Urbana, Illinois 61802, USA; 2 Program in Ecology and Evolutionary Biology, University of Illinois, 235 NRSA, 607 East Peabody Drive, Champaign, Illinois 61820, USA; 3 Corresponding author (email: [email protected])

ABSTRACT: Batrachochytrium dendrobatidis, grassland, woodland, lakes and wetlands, the causative agent of chytridiomycosis, was colonizing forest, and plantations of exotic found in 24 of 109 (22%) frogs from Kibale trees (Chapman et al., 1997; Chapman National Park, western Uganda, in January and June 2006, representing the first account of the and Lambert, 2000). Mean daily minimum fungus in six species and in Uganda. The and maximum temperatures in Kibale presence of B. dendrobatidis in an equatorial were recorded as 14.9 C and 20.2 C, African montane forest raises conservation respectively, from 1990 to 2001, with concerns, considering the high amphibian mean annual rainfall during the same diversity and endemism characteristic of such areas and their ecological similarity to other period of 1749 mm, distributed across regions of the world experiencing anuran distinct, bimodal wet and dry seasons declines linked to chytridiomycosis. (Chapman et al., 1999, 2005). Kibale has Key words: Africa, amphibians, Anura, experienced marked climate change over Batrachochytrium dendrobatidis,Chytridio- the last approximately 30 yr, with increas- mycota, Uganda. ing annual rainfall, increasing maximum mean monthly temperatures, and decreas- Chytridiomycosis, an emerging infec- ing minimum mean monthly temperatures tious disease caused by the fungus Ba- trachochytrium dendrobatidis, is a major (Chapman et al., 2005). cause of amphibian population declines Medium- and high-altitude equatorial (Daszak et al., 1999). Its global distribu- African forests such as Kibale hold a great tion, responsiveness to microenvironmen- diversity of herpetofauna, including many tal conditions affected by climate change, endemic anuran species (Channing and and association with anuran extinctions in Howell, 2006). Surveys by Vonesh be- Central America and Australia make it tween 1995 and 1997 identified 28 anuran, a disease of special concern for conserva- 15 lizard, and 32 snake species in Kibale, tion (Daszak et al., 1999; Pounds et al., an assemblage that shows an affinity with 2006). Despite growing information about the herpetofauna of the Guinea–Congo- B. dendrobatidis elsewhere in the world, lean rain forests to the west (Vonesh, comparatively little is known about the 2001). The herpetofauna of Kibale is fungus in Africa, where it is thought to markedly different from that of the coastal have originated (Weldon et al., 2004). mountains of Tanzania, which contain This report presents the results of many local endemic anuran species, and a survey for B. dendrobatidis in Kibale where B. dendrobatidis notably contribut- National Park, Uganda. Kibale is a 795- ed to the recent decline of the Kihansi km2 park located in western Uganda near spray toad, Nectophrynoides asperginis the foothills of the Rwenzori Mountains (Krajick, 2006; Lee et al., 2006). To our (0u139–0u419N, 30u199–30u329E), consist- knowledge, ours is the first multispecies ing primarily of moist semideciduous and survey for B. dendrobatidis in Kibale or in evergreen forest, transitional between any similar equatorial African montane lowland rainforest and montane forest forest site. (elevation ranging from approximately Frogs were collected from locations 1,100 to 1,600 m), and interspersed with near Makerere University Biological Field

521 522 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007

TABLE 1. Prevalence and intensity of infection with B. dendrobatidis in frogs collected from Kibale National Park, Uganda.

Habitat Number Number Prevalence Mean Species Abundancea typeb tested positive (%)c intensityd

Bufo funereus U 1 5 3 60 (23–88) 2.2262.00 Hyperolius kivuensis C 2 22 3 14 (4–34) 1.6261.40 Leptopelis christyi U 1, 3 7 2 29 (8–65) 1.2861.20 Leptopelis kivuensis U 3, 4, 5 36 9 25 (14–41) 2.1761.92 Phrynobatrachus graueri C 2 1 0 0.0 N/A Ptychadena mascareniensis U 3, 5 18 2 11 (2–34) 2.4062.30 Xenopus wittei U 5 20 5 25 (11–47) 0.6360.32 All species 109 24 22 (15–31) 2.0261.58 a Relative abundance descriptors are taken from Vonesh (2001); C5common (one can find many specimens); U5usual (one can find when looking in the proper habitat during the appropriate season). b Numbers indicate habitat types from which frogs were captured: 15forest interior, near stream; 25forest interior, in/ near permanent pond; 35forest interior, in/near ephemeral pool; 45forest interior, far from water; 55forest edge, in/ near permanent pool. All frogs were captured in habitat types typical for each species (Vonesh, 2001). c Numbers in parentheses indicate 95% confidence intervals calculated with the modified Wald method (Agresti and Coull, 1998) and are given only where samples sizes are $5. d Mean intensities are expressed as log10 zoospore equivalents per positive toe clip6standard error of the mean, calculated according to the methods of Boyle et al., 2004.

Station (elevation approximately 1500 m) a previously described real-time quantita- in Kibale National Park in January and tive PCR was used that can detect a single June 2006. Sites were chosen that were zoospore in a diagnostic sample (Boyle et known to have high amphibian densities, al., 2004). Samples were run in triplicate, both from anecdotal local reports and with the use of 1 ml of extracted DNA from previous amphibian surveys (Vonesh, template per reaction. 2001). Frogs were captured by hand or To confirm positive results, a nested with hand nets, placed individually into PCR was used that targets the fungal single-use plastic bags until a toe clip was cysteinyl tRNA synthase (ctsyn1)gene obtained from the right rear foot, and (Morehouse et al., 2004; Garner et al., released at the site of capture. Toe clips 2006), although with newly designed in- were placed in approximately 5 volumes of ternal PCR primers (ctsyn1aF: 59- RNAlaterH stabilization solution (Ambion, TCAGCTGCCGTCGTTTGTGAATTG - Inc., Austin, Texas, USA) and stored at 4 39 and ctsyn1aR: 59-CCAGAGCAGTTTG- C in the field and at 220 C after transport CCAGCATCAAA -39). Amplicons were to the laboratory. Equipment and hands electrophoresed in 1.5% agarose gels, were disinfected with 10% bleach and which were stained with ethidium bro- alcohol wipes, respectively, between indi- mide and visualized under ultraviolet viduals. light. Amplicons were purified from gels Total DNA was extracted from toe clips with the use of the ZymoClean Gel DNA with the use of the ZR Genomic DNA II Recovery Kit (Zymo Research) and were KitTM (Zymo Research Corp., Orange, sequenced directly and in both directions California, USA). For the tissue homoge- with primers ctsyn1aF and ctsyn1aR. nization step, a rotor-stator homogenizer Sequencing was performed at the Univer- with single-use individually sterilized sity of Illinois Roy J. Carver Biotechnology probes was used to prevent cross-contam- Center. ination. Negative-control DNA extractions A total of 109 frogs from seven species were also performed with every 10 toe clip were collected (Table 1). Frogs of all extractions. To test for B. dendrobatidis, species were adults, with the exception SHORT COMMUNICATIONS 523 of Leptopelis christyi and L. kivuensis, for and it thereby significantly expands the which approximately 50% of frogs were known geographic range of the fungus in post-metamorphic juveniles. All frogs ap- Africa. Kibale is approximately 800 km peared clinically normal at the time of from the nearest other location where sampling. Batrachochytrium dendrobati- frogs infected with B. dendrobatidis have dis was detected in 24/109 (22.0%) toe been documented (P. anchietae in Nair- clip samples from six species (Table 1). obi, Kenya; Speare and Berger, 2000). Prevalence estimates varied between 0% Kibale is approximately 1,200 km from and 60% among species, although sample Kihansi Gorge, Tanzania, where B. den- sizes within species were too small for drobatidis contributed to the decline of N. meaningful statistical comparisons. Among asperginis (Krajick, 2006; Lee et al., 2006), infected frogs, there was no significant and over 2,000 km from sites in southern association between intensity of infection Africa where infected Xenopus spp. have and species (Kruskal-Wallis rank sum test existed since at least as early as 1938 statistic510.92; P50.09). Batrachochy- (Weldon et al., 2004). trium dendrobatidis was not detected in It has been suggested that B. dendro- any negative control samples, indicating batidis may have emerged through the that laboratory contamination did not export of X. laevis from southern Africa for affect these results. human pregnancy testing beginning in the With the use of nested PCR, a 235 bp 1930s (Weldon et al., 2004). We note that segment of the B. dendrobatidis ctsyn1 the prevalence of B. dendrobatidis in X. gene was successfully amplified and se- wittei in Kibale (25%) is higher than that quenced in three positive field samples reported in archived specimens of X. (two L. kivuensis and one Ptychadena laevis, X. meulleri, and X. gilli collected mascareniensis). All three sequences from southern Africa between 1879 and matched exactly that of the published B. 1999 (2.7%; Weldon et al., 2004), but is dendrobatidis ctsyn1 gene (GenBank ac- within the range of location-specific prev- cession number BH001044), and all con- alence values for Xenopus spp. collected tained adenine at position 400, a single from South Africa more recently (Weldon, nucleotide A/G polymorphism previously 2005). It is unclear whether this pattern documented in a geographically distribut- represents a recent increase in B. den- ed sample of B. dendrobatidis isolates drobatidis prevalence across Africa, or (Morehouse et al., 2003). These three whether it merely reflects typical differ- samples also had among the highest ences in prevalence among spatially and intensities of infection of the field samples temporally separated populations of eco- determined to be positive by real-time logically similar species. quantitative PCR (all had .1.7 log zoo- No accounts exist of mass anuran spore equivalents per sample; attempts to mortality events in Kibale or surrounding amplify field samples with infection in- areas. Because long-term historical re- tensities less than this were unsuccessful). cords are lacking for anuran populations These data, combined with information in Kibale and in equatorial Africa in from amplifications of dilutions of positive general, it is currently unclear how B. control samples of known concentrations dendrobatidis might be affecting amphib- (data not shown), indicate a detection limit ian health and conservation in the region. of approximately three zoospore equiva- Considering the remarkable biodiversity lents per reaction for the nested PCR. and endemism of equatorial Africa’s me- This survey represents the first account dium- and high-altitude forests, and con- of B. dendrobatidis in the six positive sidering their general ecological similarity species listed in Table 1. It is also the first to forests in Central America and Australia account of B. dendrobatidis in Uganda, that have been profoundly affected by 524 JOURNAL OF WILDLIFE DISEASES, VOL. 43, NO. 3, JULY 2007 chytridiomycosis, an improved under- ———, ———, T. T. STRUHSAKER, ———, C. J. standing of the disease in equatorial Africa CLARK, AND J. R. POULSEN. 2005. A long-term is exigent. evaluation of fruiting phenology: Importance of climate change. Journal of Tropical Ecology 21: We gratefully acknowledge the assistance 31–45. in the field of I. Rwego, T. Gillespie, C. DASZAK, P., L. BERGER,A.A.CUNNINGHAM,A.D. Chapman, L. Chapman, the personnel of HYATT,D.E.GREEN, AND R. SPEARE. 1999. the Kibale EcoHealth Project, and the staff Emerging infectious diseases and amphibian of Makerere University Biological Field population declines. Emerging Infectious Dis- eases 5: 735–748. Station. We also thank J. Vonesh for helping GARNER, T. W. J., M. W. PERKINS,P.GOVINDARAJULU, with amphibian identifications, V. Beasley, D. SEGLIE,S.WALKER,A.A.CUNNINGHAM, AND J. Levengood, and J. Longcore for providing M. C. FISHER. 2006. The emerging amphibian fungal isolates, M. Fisher and T. Garner for pathogen Batrachochytrium dendrobatidis glob- providing information about PCR primers ally infects introduced populations of the North and protocols, and T. Garner and one American bullfrog, Rana catesbeiana. Biology Letters 2: 455–459. anonymous reviewer for helpful comments KRAJICK, K. 2006. The lost world of the Kihansi toad. on the manuscript. The Uganda Wildlife Science 311: 1230–1232. Authority and the Uganda National Council LEE, S., K. ZIPPEL,L.RAMOS, AND J. SEARLE. 2006. for Science and Technology kindly granted Captive-breeding programme for the Kihansi us permission to conduct this research. This spray toad Nectophrynoides asperginis at the work was supported by the University of Wildlife Conservation Society, Bronx, New York. International Zoo Yearbook 40: 241–253. Illinois, the Geraldine R. Dodge Founda- MOREHOUSE, E. A., T. Y. JAMES,A.R.GANLEY,R. tion Frontiers for Veterinary Medicine VILGALYS,L.BERGER,P.J.MURPHY, AND J. E. Program, and by the Morris Animal Foun- LONGCORE. 2003. Multilocus sequence typing dation under Award D04ZO-67. suggests the chytrid pathogen of amphibians is a recently emerged clone. Molecular Ecology 12: LITERATURE CITED 395–403. POUNDS, J. A., M. R. BUSTAMANTE,L.A.COLOMA,J.A. AGRESTI, A., AND B. A. COULL. 1998. Approximate is CONSUEGRA,M.P.FOGDEN,P.N.FOSTER,E.LA better than ‘‘exact’’ for interval estimation of MARCA,K.L.MASTERS,A.MERINO-VITERI,R. binomial proportions. The American Statistician PUSCHENDORF,S.R.RON,G.A.SANCHEZ- 52: 119–126. AZOFEIFA,C.J.STILL, AND B. E. YOUNG. 2006. BOYLE, D. G., D. B. BOYLE,V.OLSEN,J.A.MORGAN, Widespread amphibian extinctions from epi- AND A. D. HYATT. 2004. Rapid quantitative demic disease driven by global warming. Nature detection of chytridiomycosis (Batrachochy- 439: 161–167. trium dendrobatidis) in amphibian samples SPEARE,R.,AND BERGER, L. 2000. Global distribution of using real-time Taqman PCR assay. Diseases of chytridiomycosis in amphibians. http://www.jcu. Aquatic Organisms 60: 141–148. edu.au/school/phtm/PHTM/frogs/chyglob.htm. CHANNING, A., AND K. M. HOWELL. 2006. Amphibians of East Africa. Cornell University Press, Ithaca, VONESH, J. R. 2001. Natural history and biogeography New York. of the amphibians and reptiles of Kibale National Park, Uganda. Contemporary Herpetology, 4, CHAPMAN, C. A., AND J. E. LAMBERT. 2000. Habitat alteration and the conservation of African http://www.cnah.org/ch/ch/2001/4/index.htm. primates: Case study of Kibale National Park, WELDON, C. 2005. Chytridiomycosis, an emerging Uganda. American Journal of Primatology 50: infectious disease of amphibians in South Africa. 169–185. PhD Thesis, North-West University, Potchefst- ———, L. J. CHAPMAN,R.WRANGHAM,G.ISABIRYE- room, South Africa. BASUTA, AND K. BEN-DAVID. 1997. Spatial and ———, L. H. DU PREEZ,A.D.HYATT,R.MULLER, temporal variability in the structure of a tropical AND R. SPEARS. 2004. Origin of the amphibian forest. African Journal of Ecology 35: 287–302. chytrid fungus. Emerging Infectious Diseases ———, ———, L. KAUFMAN, AND A. E. ZANNE. 1999. 10: 2100–2105. Potential causes of arrested succession in Kibale National Park: Growth and mortality of seed- lings. African Journal of Ecology 37: 81–92. Received for publication 17 July 2006.