Extracting the amphibian chytrid fungus from formalin-fixed specimens


Correspondence author. E-mail: kathryn.richards@yale.edu.


1.Batrachochytrium dendrobatidis is a chytrid fungal pathogen driving many amphibian species extinct. For some species, the only place they can be found is in natural history museum collections. Harnessing molecular tools to uncover this pathogen’s date of arrival and movement throughout a region and through time will prove useful to broader B. dendrobatidis research. However, it has remained difficult to access B. dendrobatidis DNA from within preserved host tissues, especially specimens that were originally fixed in formalin.

2. Herein, I describe two methods to extract and detect, via qPCR, B. dendrobatidis DNA from herpetological natural history collections. Both extraction methods use a DNA-binding matrix, either a magnetic bead resin or a silica membrane, that retains and separates the extracted DNA from potentially qPCR-inhibiting contaminates. Nine positive control specimens enabled initial method optimizations and coincidently limited empirical comparisons between each extraction method. A museum study involving 164 formalin-fixed amphibians from Connecticut, either having been originally fixed in the 1960s or 2000s, were swabbed and samples extracted by one of the two presented methods.

3. Each method successfully extracted B. dendrobatidis DNA; of the 164 specimens tested, 37 were found B. dendrobatidis-positive, including six specimens (representing three caudate species) from 1968. In addition, results show that an infected individual stored in a jar of multiple specimens does not always contaminate the other specimens with B. dendrobatidis.

4. These results suggest that qPCR methods are well suited to assess B. dendrobatidis presence but to not assign zoospore loads in preserved specimens.


Amphibian populations are declining worldwide (Stuart et al. 2004). Of the six hypothesized drivers for their decline, tremendous attention is given to an emerging infectious disease (Collins & Storfer 2003; Lips et al. 2006; Skerratt et al. 2007). This disease, known as chytridiomycosis, is caused by the fungal pathogen Batrachochytrium dendrobatidis (hereafter, Bd) (Berger et al. 1998; Longcore, Pessier & Nichols 1999). Bd infects the epidermis of amphibians, upsetting the animal’s osmo-regulatory system and ultimately leading to cardiac arrest (Piotrowski, Annis & Longcore 2004; Voyles et al. 2007, 2009). Bd can be found on every continent inhabited by amphibians and in more than 200 species (Skerratt et al. 2007). Strong evidence suggests that Bd has been the proximate, if not the ultimate, cause of multiple population crashes and even extinctions (Stuart et al. 2004; Lips et al. 2006; Skerratt et al. 2007); chytridiomycosis may be one of the most devastating infectious diseases to wild vertebrate life (Skerratt et al. 2007).

Since its identification in the late 1990s (Berger et al. 1998; Longcore, Pessier & Nichols 1999), there has been an intercontinental effort to document amphibian populations and species infected with Bd. Initially, Bd was detected in amphibian tissues through histological examination on either archived specimens (e.g. Berger, Speare & Kent 2000; Ouellet et al. 2005) or on toe clips from living animals or recently deceased individuals (e.g. Bosch, Martínez-Solano & García-París 2001). However, using histology is not very sensitive for Bd detection (at most, its diagnostic sensitivity is 63% when compared to qPCR’s 97%) and it is very time-consuming (Kriger et al. 2006; Hyatt et al. 2007). A highly sensitive quantitative polymerase chain reaction (qPCR) specific to Bd’s ribosomal DNA (Boyle et al. 2004) and a noninvasive epidermal swabbing protocol for Bd DNA collection (Briggs & Vredenberg 2004; Livo 2004; Kriger et al. 2006; Retallick et al. 2006; Hyatt et al. 2007) now dominate as the preferred method to screen for Bd. Scientists can sample living animals noninvasively with increased confidence in Bd detection limits. Because of the relative ease of DNA collection and the quick turnaround time for results, most of today’s sampling efforts for Bd are from fresh epidermal swabs from living animals followed by the Boyle et al. qPCR (e.g. Bielby et al. 2009; Rovito et al. 2009).

Scientists today studying Bd largely test hypotheses in contemporary populations rather than in populations or species of the past. The only way to tap into Bd’s past is to screen preserved specimens held in natural history collections (NHC), but this would require employing histology. NHC hold an untapped wealth of information on Bd that will not only help biologists to understand Bd’s past demographics, but NHC could also provide valuable insights into today’s declines. To date, HerpNet (http://www.herpnet.org/), an online collaborative among numerous American museums to make their herpetological collections accessible through the internet, lists more than 1·1 million lot numbers. Accessing these collections could help resolve the first date of Bd’s introduction into naïve populations and uncover how it swept through a particular region. Matching Bd’s occurrence to noted amphibian declines could support changes in host demographic data because of, known or suspected, chytridiomycosis (i.e. Cheng et al. 2011). For many rare or extinct species, NHC may be the only or best place to find that particular species.

To better sample NHC for Bd, we need improved methods, as histology is too laborious and not sensitive enough (Hyatt et al. 2007). There are a few, recent attempts to apply Bd-specific molecular methods to NHC. Walker et al. (2008) digested formalin-fixed Xenopus tissues in a readily available DNA extraction kit (DNeasy; Qiagen, Valencia, CA, USA) followed by a whole genome amplification kit (REPLI-G; Qiagen). Unfortunately, none of the formalin-fixed tissues quantitatively or qualitatively rendered enough DNA for analyses. Later, Soto-Azat et al. (2009) tested three methods (epidermal swabbing, brushing and scraping) to nondestructively harvest fixed DNA from preserved amphibian tissues. Preserved, Bd-infected amphibians from both preservation mediums, ethanol and formalin, were tested. Harvested DNA was extracted according to the Boyle et al. (2004) protocol, using the PrepMan Ultra extraction reagent (Applied Biosystems, Carlsbad, CA, USA) with mechanical disruption (zirconium/silica beads; Biospec Products, Bartlesville, OK, USA). As with Walker et al. (2008), none of the formalin-fixed tissues yielded Bd DNA, although they were able to detect Bd from specimens preserved in ethanol. Even though both studies were unable to amplify Bd DNA from formalin-fixed tissues, the use of interdental brushes as a nondestructive collection tool was encouraging. Many museums do not welcome or are strongly resistant to destructive sampling, and thus, a tool that can collect DNA without harming the specimen is crucial to employ.

The aims of this study were to optimize molecular-based approaches to identify Bd in preserved herpetological specimens (formalin-fixation, included) and to provide recommendations for Bd qPCR applicable to screening both contemporary and preserved samples. The hypothesis being tested is that if Bd DNA can be extracted and detected from preserved, Bd-positive control specimens, Bd can also be detected from preserved specimens of unknown infection status.

Materials and methods


Positive control samples

Seven dismembered limbs from seven different anurans were supplied by Allan Pessier (San Diego Zoo, San Diego, CA, USA). Limbs came from Anaxyrus baxteri (Wyoming toad, collected in December 2009, n = 3), A. baxteri (collected in June 2001, n = 2), A. baxteri (collected in July 2000, n = 1) and Lithobates pipiens (Northern leopard frog, collected in May 1999, n = 1) (all amphibian nomenclature after Frost et al. 2006). These animals were found Bd-infected by histological analyses (A. Pessier, pers. comm.). They were fixed and maintained in 10% buffered formalin. All limbs were stored individually. Two other ranid specimens came from Arizona, USA: a Lithobates yavapaiensis (Lowland leopard frog) specimen collected in April 1972 (Arizona State University herpetological collection #15089) and a Lithobates chiricahuensis (Chiricahua leopard frog) specimen collected in September 1982 (University of Arizona herpetological collection #44789). These animals were also previously found Bd-infected by histological analyses (M. Sredl, pers. comm.). The L. yavapaiensis was fixed in 10% buffered formalin and maintained in 70% ethanol. The L. chiricahuensis was fixed in 10% buffered formalin and maintained in 33% isopropyl alcohol.

Specimens were handled with new pairs of gloves and/or forceps. Five of the seven specimen limbs and both ranids were swabbed and immediately extracted by a Promega kit. Interdental brushes (Easy Brush Cleaners; DenTek, Maryville, TN, USA) were thoroughly rubbed over the entire specimen, aiming for a minimum of five strokes per ‘surface area’ (c. 1 cm2). Two technical replicates were made from each specimen. The brush portion of the swab was cut and placed into its own 1·5-mL snap cap tube. As the tissue and DNA had been fixed in formalin, no additional preservative was used. The remaining two, unswabbed A. baxteri specimens (two of the three specimens that had remained in formalin since 2009) were transferred to 70% ethanol over a week’s time to mimic traditional museum preparations for herpetological specimens (similar to suggestions made by Simmons 2002). This consisted of 1 day in ddH2O, 2 days in 70% ethanol and a final wash of 70% ethanol for 4 days. On the fourth day, the two specimens were swabbed like the other controls and immediately extracted by the Promega kit.

Specimens for the museum study

The Yale Peabody Museum (New Haven, CT, USA) has more than 13 000 amphibian specimens from around the world, including 4332 from CT. Of those, 164 (3·79% of all CT specimens) were swabbed and analysed for the presence of Bd (following procedures below). All swabbed specimens were nonlarval life stages and came from towns where infection had been found in contemporary field swabs (K.L. Richards-Hrdlicka, J.L. Richardson & L. Mohabir, unpublished data). The date of collection and curation ranged from 1963–1969 and 2002–2007 and represent anurans and caudates (Tables 2 and 3): Anaxyrus americanus (American toad, n = 5), Lithobates clamitans (Green frog, n = 9), Lithobates palustris (Pickerel frog, n = 4), Lithobates sylaticus (Wood frog, n = 11), Pseudacris crucifer (Spring peeper, n = 21), Desmognathus fuscus fuscus (Dusky salamander, n = 52), Eurycea bislineata (Northern two-lined salamander, n = 28), Plethodon cinereus (Red-backed salamander, n = 14), Notophthalmus viridescens (Red spotted newt, n = 6) and Ambystoma maculatum (Spotted salamander, n = 14). These specimens were originally fixed in 10% buffered formalin and maintained in 70% ethanol. Like the positive control specimens, these CT specimens were handled only in the museum facility with new gloves and/or forceps and were swabbed twice. All specimens collected from the 2000’s (n = 60) were extracted using the Promega kit. Specimens from the 1960s (n = 104) were extracted by a Macherey-Nagel kit. In addition, notes were made on how many specimens per jar were stored together and how many of those swabbed individuals were found Bd-positive. It is important to know whether a single positive specimen could contaminate an entire jar filled with other specimens.

DNA Extraction and qPCR

It is currently impossible to distinguish between historic and contemporary Bd DNA; therefore, on the day of any preserved specimen work, no materials, personnel, or clothing came into contact with any known or possible contemporary Bd DNA. All work on preserved specimens was performed in an ancient DNA (aDNA) laboratory with dedicated equipment completely removed from areas where contemporary field samples of Bd are processed.

DNA extraction

Two extraction kits specificity designed for formalin-fixed samples were used: Promega’s (Madison, WI, USA) DNA IQ system (hereafter, DNA IQ) and Machery-Nagel’s (Bethlehem, PA, USA) DNA FFPE (formalin-fixed, paraffin-embedded; hereafter, MN). These two kits were initially selected for testing based on cost comparisons, ease of use and available resources (i.e. the required magnetic stand for DNA IQ was already present in the laboratory). DNA IQ takes advantage of DNA’s electrical charge by capturing and separating DNA on a 2- or 12-position magnetic stand (a magnetic tube rack). Here, a 12-position stand was used. The magnetic bead resin binds to free nucleic acids and when the sample is placed in the magnetic stand, magnetism pulls the resin (along with the bound DNA) to the bottom of the tube, leaving unwanted materials and compounds in the supernatant. MN uses silica membrane spin columns to isolate and purify DNA. MN’s sample throughput is limited by the capacity and number of microcentrifuges available to the researcher. Here, a 30-position microcentrifuge was used. DNA extractions were deemed successful if the target Bd rDNA region was amplified in qPCR.


All nine positive control specimens and every swab from specimens collected in the 2000s were extracted with DNA IQ (n = 69). Swabs remained in the tubes throughout the procedure. Swabs were incubated overnight at 37°C in 200 μL of digestion buffer (10 mm Tris pH 8·0, 10 mm NaCl, 50 mm EDTA pH 8·0, 0·5% SDS and 0·1 m DTT) and 20 μL pK (10 mg mL−1). After incubation, 100 μL of lysis buffer and 7 μL of the homogenized magnetic resin mixture were added and incubated under agitation at room temperature (23°C) for 1 h. Resin was collected on the magnetic stand, supernatant removed, and another 100 μL of lysis buffer added, vortexed and supernatant removed. Two wash steps were repeated, each time adding 100 μL of the 2× wash buffer, vortexing and removing the supernatant while tubes were in the magnetic stand. When the second wash step was completed and supernatant removed, tubes remained open to the air for 20 min to allow for residual ethanol evaporation. DNA was eluted with 50 μL of elution buffer warmed to 70°C and samples were incubated for at least 1 min. The samples were then cooled to 65°C for a final 5 min incubation period. Finally, samples were placed in the magnetic stand and the supernatant pipetted into a new tube (the supernatant now contains the DNA). The magnetic stand accompanying the kit holds 12 tubes; the last position always held a blank, negative control swab. Every positive control or every 11 specimens from CT were extracted alongside a negative control swab to test for contamination during extractions.


All swabs from the 1960s specimens were extracted by MN (n = 104). To the tube containing the swab, 200 μL of FL buffer and 20 μL of pK (10 mg mL−1) were added, vortexed briefly and spun down at 11 000 g for 1 min. Tubes were incubated overnight at 37°C. 200 μL of D-Link buffer was added, vortexed briefly and spun down at 11 000 g for 30 s and incubated at 90°C for 30 min. Tubes were cooled to room temperature, then 400 μL of 98% ethanol was added and vortexed. The manufacturer’s protocol was followed, except for the following modifications: (Wash step) Tubes were allowed to sit with the B5 buffer for 5 min, then they were spun down, the flow through discarded, they were again spun down and columns placed into new 1·5 mL tubes. (DNA elution) Tubes were left open for 30 min for ethanol evaporation, and then, 50 μL of BE buffer, warmed to 70°C, was added to the column. Tubes sat for 1 min, then spun down and elution collected. Since a 30-position microcentrifuge was used, every 29 specimens was extracted alongside a blank, negative control swab. The negative control swab was similarly extracted.


Extracted DNA was analysed by qPCR according to Boyle et al. (2004) and Garland et al. (2010) with some minor optimizations. Reaction volumes were 20 μL, consisting of 10 μL qPCR master mix (SensiMix II Low Rox; Bioline, Taunton, MA, USA), 5 μL of DNA template, 1 μL of bovine serum albumin (BSA, 100×, 10 mg mL−1), 1 μL of each primer (900 nm each) and 2 μL of the MGB probe (250 nm). BSA helps to reduce PCR inhibition (Garland et al. 2010). All specimen samples were tested in duplicate and included the negative control swab extractions, a nontemplate control (qPCR negative control, DNase-free water) and pure Bd culture (JAM081) positive control standards (100, 10, and 1 zoospore per well). Protocol for preparing the positive control standards followed ‘Quantitation standards’ from Boyle et al. (2004). The qPCR conditions were as follows: 95°C for 10 min, followed by 45 cycles of 95°C for 15 s and 60°C for 1 min on an ABI 7500 qPCR machine. qPCR plates were assembled and sealed in the aDNA laboratory prior to entrance into the contemporary DNA laboratory where the qPCR machine is located.

SensiMix II Low Rox (hereafter, SensiMix) was the chosen qPCR master mix as initial trials indicated an advantage over the traditionally used TaqMan Universal No AmpErase UNG master mix (Applied Biosystems), especially in samples with low zoospore loads. A representative comparison between the multicomponent plots of the two master mixes can be found in Fig. 1. Both qPCR master mixes are equally sensitive, reporting the successful detection of one zoospore in four of six wells; however, there is greater visual discrimination between the reference dye, ROX, and the Bd-specific MGB (minor groove binder) probe dye, FAM, when using SensiMix (Fig. 1). Greater visual discrimination at low Bd zoospore loads may be critical in some contexts, such as detection of early or light infections. The amount of ROX in SensiMix matches the requirements of the qPCR machine used here and is the suspected reason for the enhanced visual discrimination.

Figure 1.

 A representative multicomponent plot comparison between two qPCR master mixes, SensiMix (a) and TaqMan Univerisal No AmpErase UNG (b). Each graph displays the internal reference dye, ROX, to Bd-specific MGB probe dye, FAM, for six wells testing one zoospore per 20 μL reaction.


DNA Extraction and qPCR of Positive Controls

Batrachochytrium dendrobatidis DNA was successfully extracted by DNA IQ and amplified via qPCR in nine positive control specimens dating back to 1972, including the two specimens moved from formalin into 70% ethanol (Table 1). Mean number of zoospores per specimen ranged from 9 to 105 740. Of particular note, the L. pipiens collected in 1999 had a reportedly ‘light infection’ as determined by histology (A. Pessier, pers. comm.), whereas by qPCR, it had the highest zoospore load. No Bd was detected in any negative controls.

Table 1.   Mean number of zoospores for each Bd-positive specimen in a 20 μL reaction volume
SpeciesYear collectedStatus/originMean # zoospores
  1. All specimens were originally fixed in 10% buffered formalin and were extracted by DNA IQ.

  2. *After fixation, these specimens were maintained in 70% ethanol or 33% isopropyl alcohol.

  3. These specimens were washed in ethanol as per traditional museum preparations prior to extraction and qPCR (see Positive control samples).

Lithobates yavapaiensis*1972Wild/AZ9
Lithobates chiricahuensis*1982Wild/AZ14
Lithobates pipiens 1999Wild/Maine105 740
Anaxyrus baxteri 2000Wild/Laramie, WY335
A. baxteri 2001Wild/Laramie, WY212
A. baxteri 2001Wild/Laramie, WY464
A. baxteri 2009Captive/Saratoga, WY2696
A. baxteri 2009Captive/Saratoga, WY14 647
A. baxteri 2009Captive/Saratoga, WY1648

Museum Study

In total, 164 formalin-fixed specimens representing 10 species from CT were swabbed and tested for the presence of Bd. Of all 164 specimens, 37 were found Bd-positive. Ten of the specimens from the 1960s were infected (Table 2), whereas 27 of the more recently collected specimens were Bd-positive (Table 3). From the 1960s, no anurans were found infected vs. 10 caudates (eight D. fuscus fuscus, one E. bislineata and one P. cinereus; Table 2). From the 2000s, 14 anurans were found Bd-positive (two A. americanus, six L. clamitans, three L. palustris and three Lithobates sylvaticus) and 13 caudates were positive (seven D. fuscus fuscus, four E. bislineata and two N. viridescens; Table 3). The mean number of zoospores detected from each specimen ranged from 0·05 to 935 zoospores (Table S1). Some zoospore loads were quite low (>0 to 1), but extracts were retested and replicated in additional qPCRs, indicating low zoospore loads are real. The earliest date and location of detection came from six specimens (four D. fuscus fuscus, one E. bislineata and one P. cinereus) collected from Guilford, CT in October 1968. No Bd was detected in any negative controls.

Table 2.   Number of species, organized by year of collection, found positive or negative for Bd
Year curatedSpeciesDNA kitNo. of Bd-positiveNo. of Bd-negativeGrand total
  1. All specimens were collected from CT and fixed in 10% buffered formalin, now maintained in 70% ethanol and extracted by MN.

1960      11
Desmognathus fuscus fuscus MN 11
1963      11
Lithobates clamitans MN 11
1968     61622
Ambystoma maculatum MN 33
D. fuscus fuscus MN4 4
Eurycea bislineata MN189
L. clamitans MN 11
Plethodon cinereus MN145
1969     47680
A. maculatum MN 1111
Anaxyrus americanus MN 11
D. fuscus fuscus MN42226
E. bislineata MN 55
L. clamitans MN 11
Lithobates sylvaticus MN 88
Notophthalmus viridescens MN 33
P. cinereus MN 55
Pseudacris crucifer MN 2020
Grand total    1094104
Table 3.   Number of species, organized by year of collection, found positive or negative for Bd
Year curatedSpeciesDNA kitNo. of Bd-positiveNo. of Bd-negativeGrand total
  1. All specimens were collected from CT and fixed in 10% buffered formalin, now maintained in 70% ethanol and extracted by DNA IQ.

2002   11
Desmognathus fuscus fuscus DNA IQ 11
2003  142034
Anaxyrus americanus DNA IQ213
D. fuscus fuscus DNA IQ2810
Eurycea bislineata DNA IQ 77
Lithobates clamitans DNA IQ3 3
Lithobates palustris DNA IQ213
Lithobates sylvaticus DNA IQ3 3
Notophthalmus viridescens DNA IQ2 2
Plethodon cinereus DNA IQ 22
Pseudacris crucifer DNA IQ 11
2004   99
A. americanus DNA IQ 11
D. fuscus fuscus DNA IQ 44
E. bislineata DNA IQ 11
N. viridescens DNA IQ 11
P. cinereus DNA IQ 22
2005  11 11
D. fuscus fuscus DNA IQ3 3
E. bislineata DNA IQ4 4
L. clamitans DNA IQ3 3
Lithobates palustris DNA IQ1 1
2007  235
D. fuscus fuscus DNA IQ213
E. bislineata DNA IQ 22
Grand total  273360

The 164 swabbed specimens were stored in 43 jars and of those, seven jars had specimens reporting both Bd-positive and Bd-negative qPCR results; the other jars had entirely Bd-positive or Bd -negative specimens.


I present two methods to detect Bd from amphibian specimens preserved in NHC, even specimens originally fixed in formalin back in 1968. Moreover, it does not appear that an infected individual will always contaminate other co-stored specimens. The methods presented here can provide Bd scientists with a temporal dimension for their research. Adding this dimension could generate more powerful predictions about how Bd can impact a host species or population. Currently, our knowledge about Bd comes mostly from studies conducted just once or at most, spanning a few field seasons (but see Rovito et al. 2009b; Cheng et al. 2011). The recent work by Cheng et al. (2011) illustrates the utility of preserved specimens screened in connection to noted amphibian declines in Central America by uncovering Bd’s arrival and movement through the studied region. Their work also shows that the methods presented here are not the only methods available to detect Bd in preserved specimens via qPCR; they successfully adapted the PrepMan Ultra extraction method described in Boyle et al. (2004). As molecular techniques continue to improve and more companies begin to release kits designed for use on formalin-fixed specimens, there will likely be many more ways to extract and detect Bd via PCR or qPCR.

DNA Extraction and qPCR of Positive Controls

Most herpetological specimens are fixed or preserved in a manner that is very damaging to DNA. Even with the advent of tissue banks, the required tissue for molecular work may not be available and requests to destructively sample specimens are often met with resistance. The procedures presented here for collecting and extracting DNA are nondestructive, lending themselves for use on rare or delicate specimens. Results from tests performed on nine positive controls confirmed Bd DNA originally fixed in and stored in 10% buffered formalin as far back as 1972 is extractable using DNA IQ and detectable via qPCR. One specimen had a reportedly light infection, but after qPCR the zoospore load was found to be the highest. This inconsistency between histology results and qPCR results may be random (but supported elsewhere, see Cheng et al. 2011), as this is a singular example, but it may also reflect qPCR’s greater degree of sensitivity over histology (Kriger et al. 2006; Hyatt et al. 2007). The two specimens moved from formalin into 70% ethanol were found positive for Bd, showing that a 6-day exposure to ethanol does not impact pathogen detection. Although a 6-day exposure to ethanol is a comparatively short period of time that will likely never be encountered in actual NHC.

Museum Study

Of all 164 tested amphibians, 37 were found Bd-positive, showing it is possible to detect Bd from formalin-fixed amphibians dating back to at least 1968. 1968 is the earliest record for Bd in New England, but it was not the goal of this research to determine the date of Bd’s first arrival to the state or the prevalence in particular years, collection sites or species. Enabled by these extraction techniques, these hypotheses can be tested in the future. From other research, it is known Bd is pervasive in CT (28% of 916 individuals swabbed in 2010; K.L. Richards-Hrdlicka, J.L. Richardson & L. Mohabir, unpublished data). Thus, it was not surprising to find Bd in specimens collected from the recent past in an area with widespread contemporary prevalence and potentially a long-term presence. The earliest published record for Bd in this region of the world is 1961, from Saint-Pierre-de-Wakefield, Quebec (Ouellet et al. 2005). Of the 37 infected specimens, more were found infected when collected recently and extracted with DNA IQ than specimens collected in the 1960s and extracted by MN. To explain this trend, Bd could be increasing in prevalence since its arrival to the state; however, this study cannot address this putative trend. This museum study did not control for species type or sample size. Moreover, there is autocorrelation between decade collected and extraction kit. It is not possible to determine from these tests if Bd has increased in prevalence as there are untested, confounding factors.

Of the two kits, MN potentially has increased throughput capability when compared to DNA IQ and may be preferred by researchers requiring greater sample throughput. Bd can be successfully extracted from specimens fixed in 1968 using MN or 1972 using DNA IQ. A preliminary test between extraction methods (data not shown) on duplicate swabs suggest that the two methods are very comparable, with perhaps increased yield (zoospore load) detected from MN extractions. These data are preliminary and come from only eight replicate swabs, but the observation is suggestive and worthy of additional research into which method yields more DNA or is more sensitive to low zoospore loads or degraded DNA.

Seven of the 43 jars accessed for sample collection had both Bd-positive and Bd-negative co-stored specimens. These seven jars illustrate that a single infected specimen will not always yield Bd-positive results for other co-stored specimens. Future research would benefit from uncovering the degree of contamination between specimens maintained in a single jar housing infected and uninfected individuals.

Recommendations for Detecting Bd in Preserved Specimens


Care should be taken when choosing a method or kit or qPCR master mix to extract and detect Bd DNA with attention paid to budget, sample throughput, sensitivity, downstream use, and ease. Regardless of the particular molecular method, it is crucial all work on preserved specimens be executed with the proper set of negative controls and in facilities with dedicated equipment and clothing that have never been exposed to contemporary Bd. Guidelines and other precautions for working with preserved specimens from NHC are reviewed elsewhere (see Wandeler, Hoeck & Keller 2007).


For most research into NHC, a sample size of at least 30 individuals per species per sampling location per year is recommended (Cheng et al. 2011). Even though there are at least one million amphibian lots currently in American NHC, researchers will quickly find that not every species at particular locations, sufficient sample sizes, or desired collection dates are represented. Reaching the target sample size may prove impossible, but should not be an impediment to experimental design, especially if research goals are to examine rare, extinct, or older specimens. Gaps in collection records may be even more apparent for species known or suspected to be especially sensitive to chytridiomycosis (or to any other disease for that matter), whereas tolerant species may be well represented in collections.


Take at least two technical replicate swabs of each specimen and plan to extract the first swab. As Bd encysts in the epidermis, the first and most thorough swab may collect the most Bd DNA of all the replicates. Specimens are not always easy to get a hold of, so when given the chance take multiple swabs.


Keep qPCR volumes at 20 μL, or whatever volume allows for no less than 5 μL of template DNA to be added. Using a smaller reaction volume will save money and is appropriate for zoospore loads ≥19·1 ± 10·6 per sample (Ruthig & DeRidder 2012); however, in initial trials for this study, it reduced chances of detecting low Bd zoospore loads. In two identical qPCR plates, one plate run with 10 μL reaction volumes (2·5 μL template DNA/well) and the other plate with 20 μL reaction volumes (5 μL template DNA/well), with a series of positive controls (made from pure Bd culture) and Bd-infected field samples, only 24% of the wells from the smaller reaction volume reported Bd, whereas 91% of the wells in the larger reaction volume reported Bd. Zoospore loads in these plates ranged from 0·05 to >30 000. It was not surprising that larger zoospores loads were successfully detected, but it remained a consistent pattern among subsequent tests that with 10 μL reaction volumes, ≤1 zoospore was either haphazardly detected or not at all. If reaction volumes must be reduced, it may be possible to maintain or increase detection probability with increased replicate wells (i.e. test extracts in >3 wells). Bd DNA that has been in museum curation may be degraded and/or fragmented and, in most cases, the starting template is a complex genomic sample in which Bd may be far less prevalent than host DNA, for example. Increasing the amount of template DNA increases the likelihood of successfully detecting Bd, especially for low zoospore loads.


New qPCR technologies and chemistries allow researchers to run qPCR at ‘fast’ and ‘standard’ times. Ruthig & DeRidder (2012) found ‘fast’ settings appropriate for contemporary field samples with ≥19·1 ± 10·6 zoospores per sample using Applied Biosystem’s TaqMan Fast Universal PCR MasterMix. In other initial tests for the methods presented here, there was a reduction in Bd detection when ‘fast’ settings were used. Two identical qPCR plates tested six replicates each of 1 and 0·1 zoospores per well in 20 μL reaction volumes using SensiMix; one plate was run at the ‘standard’ settings and the other was run under the 'fast' settings. None of the wells amplified Bd under ‘fast’ settings, whereas 67% of the wells, run with the ‘standard’ reaction time, successfully detected Bd. If SensiMix is used, run reactions at the standard time (c. 1·75 h) as opposed to the fast time (c. 45 min). This finding also highlights the importance of Recommendation #1: it is important to select the best qPCR master mix given research goals (i.e. detection of ≤1 zoospore) and equipment (i.e. particular make and model of qPCR machine). As most starting templates are complex, it is important to give the primers and probe ample time to align to their target sequences. Running complex genomic samples at the standard time also gives the DNA more time to denature, at which formalin-fixed DNA may be slower. The standard reaction time not only gives primers more time to find their target sequence, but also makes the target more accessible as DNA denatures.


This work was supported by the Yale Institute for Biospheric Studies (YIBS) Molecular Systematics Conservation Genetics Laboratory, a YIBS Center for Field Ecology grant and by the US Environmental Protection Agency Science to Achieve Research Fellowship. I would like to thank Allen Pessier, Michael Sredl, George Bradley, Charlotte Johnston, Thomas Dowling, Oliver Hyman, Erica Bree Rosenblum and Suzanne Joneson for the generous supply of and help with positive control materials. Thanks to Samantha Attwood and Greg Watkins-Colwell for their help in the laboratory and collections and to Adalgisa Caccone for helpful manuscript comments. The author has no conflicts of interest to declare.