Experimental Infection and Repeat Survey Data Indicate the Amphibian Chytrid Batrachochytrium dendrobatidis May Not Occur on Freshwater Crustaceans in Northern Queensland, Australia
- First Online:
- Cite this article as:
- Rowley, J.J.L., Hemingway, V.A., Alford, R.A. et al. EcoHealth (2007) 4: 31. doi:10.1007/s10393-006-0075-z
- 157 Views
Chytridiomycosis is a fatal disease of amphibians, caused by the amphibian chytrid Batrachochytrium dendrobatidis. The disease is unusual in that it may drive many amphibian species to local extinction during outbreaks. These dramatic declines in host population numbers could be facilitated if the pathogen can grow as a saprobe or on alternative hosts, a feature common to other chytrid species. This is also supported by in vitro work that demonstrates B. dendrobatidis can grow and reproduce in the absence of amphibian cells. In a previous study, B. dendrobatidis was detected on freshwater shrimp from rain forest streams in northern Queensland, Australia, using diagnostic PCR. We set out to confirm and further investigate the presence of B. dendrobatidis on crustaceans by carrying out more extensive sampling of shrimp in the field, experimental B. dendrobatidis infection trials using shrimp and crayfish, and PCR verification of the presence of B. dendrobatidis from shrimp samples that previously tested positive. We could not confirm the presence of B. dendrobatidis on shrimp, and report that original positive tests in shrimp reported by Rowley et al. (2006) were likely false. Thus, we suggest that shrimp may not be an important reservoir host for B. dendrobatidis.
KeywordsBatrachochytrium dendrobatidisamphibian declineschytridiomycosisalternative hostfreshwater shrimpcrayfish
Chytridiomycosis, a fatal disease of amphibians, is caused by the chytrid Batrachochytrium dendrobatidis (Berger et al., 1998). B. dendrobatidis is the only chytrid known to parasitize vertebrate hosts (Berger et al., 1998), occurring in the epidermis of adult amphibians and the mouthparts of larval anurans (Berger et al., 1998). Chytridiomycosis is capable of causing rapid mortality (Nichols et al., 2001), with frogs of susceptible species often dying within 3 weeks from infection in the laboratory (Berger et al. 1998). The disease is spread via motile, waterborne zoospores (Longcore et al., 1999). B. dendrobatidis is unusual in that it appears to drive many host species to local extinction during outbreaks (Lips et al., 2006; Schloegel et al., 2006). Mechanisms that could facilitate these extreme epidemics include the ability to grow as a saprobe or on alternative hosts (McCallum and Dobson, 1995, 2002; Daszak et al., 1999; Godfray et al., 1999), a feature common in other chytrid species (Powell, 1993). This would allow the disease to persist in the environment in the absence of amphibian hosts, and could explain its ability to drive host species to extinction.
In the laboratory, B. dendrobatidis zoosporangia can attach to and grow on dead algae and insect exoskeletons (Johnson and Speare, 2003) and survive for at least 3 months in sterile sand or bird feathers (Johnson and Speare, 2005). Recently, Rowley et al. (2006) reported the detection of B. dendrobatidis on freshwater shrimp from rain forest streams in Australia, using diagnostic PCR. Freshwater shrimp may comprise as much as 95% of the standing stock biomass of total stream organisms in some areas (Covich et al., 1991, 1996), potentially making them an important reservoir of the disease. Due to the widespread implications of this finding, we set out to further test and investigate the presence of B. dendrobatidis on shrimp and crayfish. Here we report the results of more extensive sampling of shrimp in the field, infection trials using both shrimp and crayfish in the laboratory, PCR verification of the presence of B. dendrobatidis from shrimp samples that previously tested positive in Rowley et al. (2006), and a reexamination of the raw data from the PCR tests that produced the results reported by Rowley et al. (2006).
We sampled shrimp at three relatively undisturbed streams within tropical rain forest of northern Queensland, Australia: Python Creek (145°35′ E, 17°46′ S; 200 m asl); an unnamed creek (“Lower Tully Creek,” 145°41′ E, 17°48′ S; 70 m asl) in Tully Falls Forest Reserve; and Frenchman Creek, in Wooroonooran National Park (145°55′ E, 17°20′ S; 20–100 m asl). All streams contain frog species that have suffered declines in association with outbreaks of chytridiomycosis, and currently persist with B. dendrobatidis [Rowley and Alford, unpublished data]. The area experiences distinct wet and dry seasons, with cool temperatures and minimal rainfall in the winter and high temperatures and rainfall in the summer. Both streams contain at least one Macrobrachium (Decapoda: Palaemonidae) species, but Caridina zebra (Decapoda: Atyidae) occurs only at Python Creek.
We conducted shrimp surveys between 2000–2400 hours during August and September 2005. During each survey, we moved upstream for approximately 100 m, catching shrimp using hand-held dip-nets. Immediately after we captured each shrimp, we swabbed it with a sterile cotton swab (Medical Wire & Equipment Co. [Bath] Ltd., Wiltshire, UK), taking care to swab all regions of the animal. After swabbing, we released shrimp several meters downstream from their point of capture and we continued surveying upstream, making the recapture of individuals highly unlikely. After handling each individual, we washed the nets in the stream. We evaluated the samples we collected from each shrimp for the presence of B. dendrobatidis using Taqman diagnostic quantitative PCR (Boyle et al., 2004). DNA was extracted with PrepMan Ultra, and amplified using primers ITSI-3 Chytr and 5.8S Chytr (Boyle et al., 2004; GeneWorks Pty Ltd, Hindmarsh, Australia). Each sample was tested in triplicate, and a sample was only recorded as positive if all three replicates indicated the presence of B. dendrobatidis.
Controlled Infection Experiments
We tested animals for B. dendrobatidis in the same manner as in field surveys. For all experimental infections, we used the laboratory culture strain Gibbo River-Llesueuri-00-LB-1 (after Berger et al., 2005) grown on nutrient agar at 23°C. B. dendrobatidis zoospores were collected from cultures by flushing the plates with 1 mL nutrient broth. We then spun the zoospores in a centrifuge at 2500 rpm at 4°C for 5 minutes, removed the supernatant containing the zoospores, and diluted it with the nutrient broth to attain the desired concentration of zoospores. We used a hemocytometer to estimate the concentration of zoospores, counting only the motile, and hence viable, individuals. Zoospore concentrations reported here are therefore conservative. We conducted all experiments at room temperature, which did not exceed 26°C.
Experimental Infection of Caridina zebra
We collected 115 C. zebra from Python Creek on September 9, 2005. The shrimp measured 1–2 cm in length. Fifteen shrimp were swabbed and tested for B. dendrobatidis and not used in later experiments. We allocated the remaining 100 into groups of 10, and placed each group into a disposable polyethylene container with about 500 mL of water from Python Creek. We then added approximately 3 mL of nutrient broth containing 125,000 B. dendrobatidis zoospores/mL to each container. During the exposure period, 48 of the shrimp died, probably due to insufficient aeration. After 24 hours, we transferred the surviving shrimp to 10 clear plastic tanks (Hagen Mini Small Pal Pens; 18 × 15 × 11 cm, Rolf C. Hagen, Inc., Montreal, Canada) fitted with filters (Smallworld filter, Penn-Plax, Inc., Hauppauge, NY). Tanks contained 750 mL of water (half stream water from the site of collection and half tap water) and a freshwater plant (Hydrilla sp.), which tested negative for B. dendrobatidis. Seventeen days after exposure to B. dendrobatidis, we tested one shrimp from each tank for B. dendrobatidis.
Experimental Infection of Cherax quadricarinatus
We collected 10 captive-raised juvenile crayfish, Cherax quadricarinatus (Decapoda: Parastacidae), measuring 4–7 cm in length, from breeding tanks at James Cook University on September 5, 2005. We placed the crayfish individually into tanks with filters, identical to those used in the shrimp infection experiment. The tanks were filled with 750 mL water from the communal breeding tanks from which the crayfish had been removed. We allowed an acclimation period of 3 days. During this period, one crayfish moulted and the exoskeleton was left inside the tank for it to consume.
After testing negative for B. dendrobatidis, we removed crayfish from their tanks and placed them individually into disposable polyethylene containers for 24 hours with approximately 500 ml of breeding tank water. We randomly chose two crayfish as negative controls, and exposed four to water containing approximately 10,000 zoospores, and four to water containing approximately 1000 zoospores. After 24 hours, we returned the crayfish to their individual tanks. One crayfish that was exposed to 10,000 zoospores moulted on day 31. No other crayfish moulted during the experiment. Crayfish that died prior to day 32 were swabbed for B. dendrobatidis at time of death, and all others were swabbed at day 32.
We analyzed the supernatant remaining from the DNA extractions identified as the positive “shrimp swabs” collected by Rowley et al. (2006), plus similar supernatant from swabs known to contain B. dendrobatidis (as positive controls). Initial PCR amplification was conducted using the chytrid specific primers BOB5 and BOB6 (Boyle et al., 2004). In addition, chytrid specific primers (Bd1a, Bd2a) from Annis et al. (2004) were used to amplify all samples. Generic primers, ITS5 and ITS2 from White et al. (1990), were used as positive controls for the presence of rDNA on which the chytrid specific primers are derived (Annis et al., 2004; Boyle et al., 2004). PCR amplification conditions followed those initially described in their respective articles (Annis et al., 2004; Boyle et al., 2004). Optimization testing of a wide range of temperature and amplification conditions followed the protocol of Cobb and Clarkson (1994). Whole genome amplification of DNA extracted from swabs was conducted to detect minute quantities of genomic DNA following manufacturers instructions using a GenomiPhi™ DNA amplification kit (GE Healthcare Bio-Sciences Pty. Ltd., Rydalmere, Australia). Subsequent PCR testing to detect the presence of chytrid DNA and/or any eukaryotic DNA followed the procedures undertaken for pre-genome amplified DNA described above. Negative controls were run using all reagents, using water to substitute for DNA samples in all treatments.
We sampled 34 Macrobrachium from Frenchman Creek, 28 Macrobrachium from Lower Tully Creek, and 38 C. zebra from Python Creek, resulting in a total of 100 shrimp samples. No shrimp tested positive for the presence of B. dendrobatidis.
Controlled Infection Experiments
Experimental Infection of Caridina zebra
All quantitative PCR results for swabs taken from the 15 C. zebra prior to the first experimental infection were negative except for one C. zebra, which yielded a very low positive (one zoospore equivalent), in two of the three wells. We then retested the sample for this individual and found a very low positive for one replicate, which we consider a negative result. All 10 C. zebra tested were negative for B. dendrobatidis 17 days after infection.
Experimental Infection of Cherax quadricarinatus
Of the 10 crayfish, 4 died during the experiment. One control died 35 days after the experimental infection began, two from the high exposure group died at days 31 and 35, and one from the low exposure group died at day 28. Swabs taken at 32 days from the surviving crayfish and from those that died were negative for B. dendrobatidis.
There was no amplification of DNA from the “shrimp samples” using chytrid primers, either direct from swab DNA extractions as used in Rowley at al. (2006) or following GenomiPhi™ amplification of DNA present in the sample. Positive amplification of DNA from the “shrimp samples” was achieved using generic primers following GenomiPhi™ amplification of DNA indicating very small amounts of DNA in the original DNA extraction of the “shrimp samples.” Testing of lower annealing temperatures did not yield any amplification product using chytrid primers. All positive controls amplified readily with both sets of chytrid specific primers and the generic primers. Negative controls remained negative in all tests.
Reexamination of Raw Data from Rowley et al. (2006)
Our inability to confirm the results reported in Rowley et al. (2006) using further sampling, infection trials, and genetic evidence led us to examine every possible source of error in the process. We retrieved and examined the archived original output from the diagnostic PCR tests run for Rowley et al. (2006), and found that there had been an error in transcribing the raw output from the machine to the results spreadsheet that was disseminated to the authors of that report. In the raw output, all shrimp samples were negative.
The total number of shrimp from tropical Australian streams that have been tested and found to be negative for infection by B. dendrobatidis to date is 163 (115 in the present paper and 48 by Rowley et al. ). The exact upper 95% binomial confidence interval for prevalence of infection in shrimp, assuming that those sampled are was representative of shrimp populations in those streams, is 2.24% (StatXact 4.0, Cytel, Inc., Cambridge, MA). Our results thus suggest that it is unlikely that freshwater shrimp are acting as environmental reservoirs or alternative hosts for B. dendrobatidis in tropical rain forest streams in Australia.
The experimental infections of live shrimp and crayfish indicate that B. dendrobatidis is not likely to grow on their exoskeletons in the laboratory. However, in retrospect, the experimental infection attempts would have been a more powerful test of the ability of B. dendrobatidis to survive on shrimp and crayfish if the experiments had included positive controls, in which zoospores recovered from the inoculation media were cultured to ensure that they had not been killed by some aspect of the procedures followed.
The lack of amplification of chytrid specific DNA from the “shrimp samples” collected and identified as positives by Rowley et al. (2006), even under highly relaxed PCR conditions, indicates the lack of any intact chytrid DNA in these samples. It is possible that this DNA may have degraded during the intervening period between the analysis conducted in Rowley et al. (2006) and this study (several months). However, GenomiPhi™ amplification of the limited DNA present in these samples indicated that intact DNA had survived that could not be amplified with the chytrid specific primers but was amplified with the generic primers, suggesting the presence of eukaryotic DNA. We take this evidence in support of our other findings, which indicate that there is a low probability that shrimp act as alternative hosts for B. dendrobatidis in rain forest streams.
These findings demonstrate that it is important to closely scrutinize the results of PCR runs, particularly when positive results for samples are likely to be negative or are of great significance. Such positives may result from gross contamination of samples that can occur very rarely and go undetected, or, as occurred here, manual errors in the process of disseminating results to authors. In such cases, a replicate sample should be taken, the result should be confirmed by another diagnostic technique such as histology, or the original sample should be retested. Additionally, efforts should be made to reduce and double-check any manual tasks that may be sources of error.
This study also highlights the importance of testing samples in triplicate to ensure a high specificity for the PCR test. Sporadic low level contamination of the PCR test via aerosolization appears to occur based on negative controls occasionally having one or two wells positive at very low levels of zoospore equivalents [Campbell and Skerratt, unpublished data]. Therefore, low positives in one or two wells should be regarded as negative or suspicious positives to maximize specificity. When we examined the map of the 96-well plate used for extraction of the sample with the low positive in two wells, we found that it was located next to a positive control/standard. While we cannot be sure that the low positive from the controlled shrimp in the infection experiments was due to contamination, it is certainly possible that aerosolization of DNA during pipetting may be responsible for this result. Recently, Kriger et al. (2006) recommended using the test in singlicate to reduce costs. This would reduce the specificity of the test and may increase its sensitivity. A singlicate test could be used as a screening test or as a diagnostic test if false positives were not an important consideration such as in ecological studies. However, it would be important to intermittently quantify the level of contamination to ensure that the specificity of the test was not changing.
While we have found no evidence for freshwater shrimp or crayfish acting as alternate hosts, it is still possible that environmental reservoirs or alternate hosts may contribute to the extremely high transmission rates seen in chytridiomycosis outbreaks in some systems (Berger et al., 1998; Lips et al., 2006). For example, B. dendrobatidis was recently detected via quantitative PCR on six of seven substrate samples associated with dead frogs and one of nine stream boulders during a mass mortality event in Panama (Lips et al., 2006). Additional work is needed to determine how important alternative hosts may be in the host–pathogen relationships between B. dendrobatidis and amphibians.
This research was supported by funding from the Australian Geographic Society; the Society for the Study of Amphibian and Reptiles; the Peter Rankin Trust Fund for Herpetology; the Australian Government Department of Environment and Heritage; the US National Science Foundation Integrated Research Challenges in Environmental Biology grant DEB-0213851; and the US National Science Foundation and Australian Academy of Science’s East Asia Summer Pacific Institute, and was carried out under a Scientific Purposes Permit issued by the Queensland Parks and Wildlife Service (WISP01715204), as approved by the James Cook University Animal Care and Ethics Committee (A863). J.J.L.R. was supported by an Australian Postgraduate Research Scholarship. Kathy La Fauce and Leigh Owens provided the crayfish and technical support for the experimental infections. Numerous volunteers assisted in the field.