Conserv Genet (2007) 8:577–586 DOI 10.1007/s10592-006-9193-y

ORIGINAL PAPER

Characterization of target nuclear DNA from faeces reduces technical issues associated with the assumptions of low-quality and quantity template Mark C. Ball Æ Richard Pither Æ Micheline Manseau Æ Jeff Clark Æ Stephen D. Petersen Æ Steve Kingston Æ Natasha Morrill Æ Paul Wilson

Received: 31 March 2006 / Accepted: 25 July 2006 / Published online: 28 September 2006
Springer Science+Business Media B.V. 2006

Abstract Faecal material has increasingly become an important non-invasive source of DNA for wildlife population genetics. However, DNA from faecal sources can have issues associated with quantity (lowtemplate and/or low target-to-total DNA ratio) and quality (degradation and/or low DNA-to-inhibitor ratio). A number of studies utilizing faecal material assume and compensate for the above properties with minimal characterization of quantity or quality of target DNA, which can unnecessarily increase the risk of downstream technical problems. Here, we present a protocol which quantifies faecal DNA using a two step approach: (1) estimating total DNA concentration using a PicogreenTM fluorescence assay and (2) estimating target nuclear DNA concentration by comparing amplification products of field samples at suspected concentrations to those of control DNA at known concentrations. We applied this protocol to faecal material collected in the field from two species:

woodland caribou (Rangifer tarandus) and swift fox (Vulpes velox). Total DNA estimates ranged from 6.5 ng/ll to 28.6 ng/ll (X = 16.2 ng/ll) for the caribou extracts and 1.0–26.1 ng/ll (X = 7.5 ng/ll) for the swift fox extracts. Our results showed high concordance between total and target DNA estimates from woodland caribou faecal extracts, with only 10% of the samples showing relatively lower target-to-total DNA ratios. In contrast, DNA extracts from swift fox scat exhibited low target DNA yields, with only 38% (19 of 50) of the samples showing comparative target DNA amplification of at least 0.1 ng. With this information, we were able to estimate the amount of target DNA entered into PCR amplifications, and identify samples having target DNA below a lower threshold of 0.2 ng and requiring modification to genotyping protocols such as multiple tube amplification. Our results here also show that this approach can easily be adapted to other species where faeces are the primary source of DNA template.

M. C. Ball (&) Æ S. D. Petersen Æ P. Wilson Natural Resources DNA Profiling and Forensic Centre, Trent University, 1600 East Bank Drive, Peterborough, ON K9J 7B8, Canada e-mail: [email protected]

Keywords Non-invasive Æ DNA quantification Æ Population genetics Æ Swift fox Æ Woodland caribou

R. Pither Æ M. Manseau Western Canada Service Centre, Parks Canada, 145 McDermot Avenue, Winnipeg, MB R3B 0R9, Canada J. Clark Natural Resources Institute, University of Manitoba, 303-70 Dysart Road, Winnipeg, MB R3T 2N2, Canada S. Kingston Æ N. Morrill Ontario Parks, 435 James Street South, Suite 221d, Thunder Bay, ON P7E 6S8, Canada

Introduction Advances in non-invasive DNA collection methods and molecular techniques have created new possibilities for examining population structure, genetic diversity, and dispersal patterns in wildlife species (Fernando et al. 2003). Faecal collections, for example, offer the potential to survey a greater number of individuals than would be possible using invasive techniques, with the added benefit of avoiding any

123

578

disturbance to the animals (Kohn and Wayne 1997; Wasser et al. 1997; Taberlet et al. 1999; Morin et al. 2001; Fernando et al. 2003; Piggott and Taylor 2003; Maudet et al. 2004). Here, sloughed intestinal mucosal cells covering the outer surface of faeces can be isolated to extract DNA template for genetic analyses (Hoss et al. 1992; Flagstad et al. 1999; Banks et al. 2002; Fernando et al. 2003; Maudet et al. 2004; Wehausen et al. 2004). Despite recent successes in providing genotype profiles from faecal DNA, it is still regarded by many as a source of low quantity and quality DNA when compared to traditional sources such as blood and tissue (Gerloff et al. 1995; Flagstad et al. 1999; Morin et al. 2001; Piggott and Taylor 2003; Wehausen et al. 2004). This concern arises from the fact that DNA from faeces can be subjected to degradation, may contain template from sources other than the species under investigation (e.g., microorganisms, digested food products) and samples can include high inhibitorto-DNA ratios. Consequently, many researchers have adopted protocols that potentially increase processing time, cost (Wilson et al. 2003), and analytical error (Gill et al. 2000; Whitaker et al. 2001; Broquet and Petit 2004). To control for an assumed presence of PCR inhibitors, researchers typically dilute faecal DNA extracts using arbitrary volumes (Flagstad et al. 1999). This lowers the inhibitor concentration as well as the amount of DNA template available for replication. To compensate for the reduced concentrations of DNA template, it is common to increase the amount of PCR cycles to provide successful amplification (Ernest et al. 2000; Farrell et al. 2000; Morin et al. 2001; Lucchini et al. 2002; Fernando et al. 2003; Frantz et al. 2003; Bellemain and Taberlet 2004; Hung et al. 2004; Mausdet et al. 2004; Roon et al. 2005). However, increased PCR cycling coupled with low DNA template has been shown to increase the probability of amplification errors such as allelic dropout and false alleles (Gill et al. 2000; Whitaker et al. 2001). Surprisingly, most of these compensatory actions are adopted without any knowledge of whether or not they are necessary, or of the actual DNA concentrations in diluted samples. Research utilizing faecal DNA would benefit from an assessment of DNA quantities and quality prior to research, to avoid potentially unnecessary costs and errors. Morin et al. (2001) employed a direct target DNA quantification protocol using RT-PCR, but this technique that may not be logistically feasible for most labs. Here we report on a two step strategy to characterize faecal DNA using two species, woodland caribou (Rangifer tarandus) and swift fox (Vulpes velox)

123

Conserv Genet (2007) 8:577–586

by: (1) estimating total DNA concentration using a PicogreenTM fluorescence assay; and (2) estimating target DNA concentration by comparing amplification products of field samples to those of control DNA in a dilution series. In the second step, target DNA is amplified using a robust, taxa-specific DNA marker, in the size range of the largest microsatellite used in genotyping. This comparison will provide information regarding the target-to-total DNA ratio and inhibition of a particular sample. Applying an assessment of quantity and quality provides the categorization of samples with high or low probability of amplification success based upon template thresholds. This allows for the appropriate implementation of either standard genotyping protocols, or the use of modified protocols such as the multiple tube amplification approach (Navidi et al. 1992; Taberlet et al. 1996, 1999; Miller et al. 2002; Frantz et al. 2003).

Methods Samples Woodland caribou (Rangifer tarandus) A total of 20 woodland caribou faecal samples, comprising approximately 15 pellets were collected in February of 2005 from Manitoba, Canada. All samples were shipped from the field frozen and stored at – 20C prior to analysis. In the laboratory, each faecal sample was divided into three sub-samples consisting of three pellets each (approximately 5.0 g). This resulted in 60 sub-samples, which were extracted, quantified, and genotyped as described in the following sections. Control woodland caribou DNA used in the following experiments was extracted from muscle tissue and quantified using PicogreenTM fluorescence assay. Swift fox (Vulpes velox) A total of 50 swift fox scats were collected in August and September of 2005 from Saskatchewan, Canada. The survey routes were cleared of all carnivore faeces one day prior to actual faecal collections, to ensure that the faeces collected were <1-day-old. Collected samples were discriminated by size, and morphology to help ensure that all samples were likely swift fox and not coyote (Canis latrans) or badger (Taxus taxidea), both of which are present in the survey area. However, it was possible that some of the collected samples could have been produced by red fox (Vulpes vulpes), also inhabiting the area. All samples were shipped from the

Conserv Genet (2007) 8:577–586

field frozen and stored at – 20C prior to analysis. Control swift fox DNA used in the following experiments were extracted from muscle tissue, and quantified using PicogreenTM fluorescence assay. DNA extraction A total of 60 (20 · 3 replicates) caribou samples and 50 swift fox scats were prepared for extraction. Samples were thawed in 2 ml of 0.1 M Phosphate Buffered Saline (PBS) for 5 min, then the surface mucosa was wiped using a sterile cotton applicator. Thawing the scat in PBS solution seems to re-hydrate the surface mucous, causing it to thicken. Wiping the scat involved lightly rubbing the outer surface to remove the mucosal coat. Care was taken to minimize the amount of faecal material that was removed with the mucosal coat (Fig. 1). After wiping, the applicator tip was immersed in 300ll of 1 · lysis buffer (Applied Biosystems, Foster City, California) contained in a 1.5 ml epindorff tube. Samples were digested using 20 units of Proteinase K (Qiagen Inc., Mississauga Ontario) and incubated at 65C for 2 h, followed by an additional 20 units of Proteinase K and incubation at 37C for 12 h. DNA from each sample was then extracted using the Qiagen DNAeasy kit (Qiagen Inc., Mississauga Ontario) following the recommended protocol for DNA extraction of tissue using the full volume of lysed material. Each sample was eluted in 65.0ll of 0.1 M TE buffer heated to 70C. PicogreenTM quantification PicogreenTM (Molecular Probes) quantifies the amount of DNA extracted by binding preferentially to doublestranded DNA and fluorescing when excited by laser.

Fig. 1 Removal of outer mucosal layer from faeces using the wiping protocol

579

Five microlitre of each sample was quantified using PicogreenTM. From the RFU (relative fluorescence units) reading, individual sample concentrations were calculated and each sample was diluted to a total DNA concentration of 2.5 ng/ll when possible. Samples were stored at – 20C when not in use.

Target DNA quantification PicogreenTM, or any other quantification method that uses fluorescence to determine the amount DNA, provides estimates of cumulative DNA found in each sample extract. As faecal DNA extractions may include DNA from sources other than the target species, PicogreenTM alone will not provide an assessment of the quantity of the target DNA. The goal of this study was to determine quantity of target DNA. Therefore, we compared the amplification products of our faecal DNA samples to those from control DNA. This involved amplifying 4 ll of control DNA of both woodland caribou and swift fox at quantities of 5.0 ng (1.25 ng/ll), 1.0 ng (0.25 ng/ll), 0.5 ng (0.125 ng/ll), 0.2 ng (0.05 ng/ll), and 0.1 ng (0.025 ng/ll) and comparing product amplification of 4 ll of faecal DNA at amounts of 5.0 ng (1.25 ng/ll), 1.0 ng (0.25 ng/ll), 0.5 ng (0.125 ng/ll), 0.2 ng (0.05 ng/ll), and 0.1 ng (0.025 ng/ll) for woodland caribou. Swift fox faecal samples were amplified under a smaller dilution series using only 5.0 ng (1.25 ng/ll), 1.0 ng (0.25 ng/ll) and 0.1 ng (0.025 ng/ll) for comparison to the larger swift fox control DNA series. By amplifying a dilution series for faecal DNA, we are able to assess the presence of high inhibitor-to-DNA ratios. Furthermore, the amplification phenotype of the control DNA dilution series is characterized by an obvious reduction in band intensity at each DNA quantity. This provides an accurate, repeatable measure to accurately characterize the target-to-total DNA ratio in field samples. If PicogreenTM DNA quantity estimates accurately reflect the amount of target DNA in each sample, then the serial dilution amplification should be comparable to their counterparts in the control DNA dilution series. If the target DNA quantity is less than that of the PicogreenTM estimate, then the sample dilutions produce weaker amplified product in the series. The PicogreenTM fluorescence assay has reduced sensitivity when quantifying DNA below 1.0 ng/ll. As such, target DNA characterization of field samples having total DNA PicogreenTM estimates below this threshold were assessed using a serial dilution consisting of an amplification of the stock DNA elute and subsequent 1/10th and 1/100th dilutions.

123

580

Characterization of woodland caribou target DNA PCR amplification of woodland caribou DNA was performed using both high and low molecular weight gender markers (HMW and LMW, respectively) that amplify the zinc finger gene intron on both the X-chromosome and Y-chromosome (Shaw et al. 2003). Our choice of gender specific markers for target DNA characterization was due to our requirement to determine the gender of each faecal sample for downstream analyses. In woodland caribou, the HMW marker set amplifies a region of approximately 900 base pairs (bp) (LGL331 and LGL335; Shaw et al. 2003). An internal primer was developed specifically for caribou SDP730; 5¢-GGAAATCATTCATGAATATCAC-3¢ when coupled with LGL 331; 5¢-CAAATCATGCAAGGATAGAC-3¢ (Shaw et al. 2003) yielded two fragments in males (220 bp and 200 bp) and one in females (220 bp) for the purpose of this study both markers were amplified to more accurately assess the quality of the target DNA template, since DNA degradation would expect to result in reduced amplification of the HMW marker relative to the LMW marker. Typically, amplification of the LMW marker alone will provide information pertaining to the quality of target DNA in the size range of our microsatellite loci. A 10 ll amplification volume was used containing: 1 · PCR buffer; 1.5 mM MgCl; 0.2 lg/ml of bovine serum albumin (BSA); 0.2 lM of each primer; 0.2 mM of each dinucleotide triphosphate; 1 unit of Taq polymerase (Invitrogen Life Technologies) and 4.0 ll of DNA template. The thermocycling protocol consisted of 94C for 5 min, then 29 cycles of 94C for 30 s, 56C for 30 s and 72C for 30 s, followed by a final extension time of 72C for 2 min. Amplified products were electrophoresed and visualized in a 2.0% agarose gel stained with SYBRgreenTM (Invitrogen). Validation of the consistency of genotypes generated by quantified faecal DNA To validate our target DNA quantification protocol all woodland caribou faecal DNA samples were amplified using six polymorphic, fluorescent-labelled microsatellite markers (Rt6, Rt9, Rt24 (Wilson et al. 1997); Map2C, BL42 and BM848 (Moore et al. 1992). Two, 3 primer multiplex reactions were conducted in a 10.0 ll volume containing: 1 · PCR buffer; 2.0 mM MgCl; 0.2 lg/ml of BSA; 0.4 lM of both the forward and reverse primers (more than one set in multiplex reactions); 0.2 mM of each dinucleotide triphosphate; 0.5 units of Taq polymerase (Invitrogen Life Technologies). The amount of target DNA template added

123

Conserv Genet (2007) 8:577–586

to the reaction estimated using our comparative assay. The thermocycling protocol for individual and multiplexed loci consisted of 94C for 5 min, then 30 cycles of 94C for 30 s, 56C for multiplex 1 (Map2C, Rt6 and RT9) and 60C for multiplex 2 (BL42, RT24 and BM848) for 1 min, then 72C for 1 min, followed by a final extension time of 60C for 45 min. As mentioned, each of the 20 samples collected from individual woodland caribou were divided into three sub-samples, (60 sub-samples total) each of which were amplified in triplicate to verify consistency in genotypes among separate amplifications and well as among identical sub-samples. Following amplification, each sample was desalted and then genotyped using a MegaBace 1000 workstation (Amersham Pharmecia) and the scoring software GeneticProfiler v2.3 (Amersham Pharmecia). Alleles were defined based upon size (base pairs) using a Rox 550 (Amersham Pharmecia) size standard. Characterization of swift fox target DNA Since swift fox are carnivorous their scat may contain prey DNA that is likely to amplify gender primers (Murphy et al. 2003). Therefore the PCR assessment of target DNA characterization used a swift fox tetranucleotide microsatellite marker, Vve2-64 (Cullingham et al. 2006) which amplified a region of nuclear DNA approximately 200 bp. PCR amplification involved a 10 ll amplification volume containing: 1 · PCR buffer; 1.25 mM MgCl; 0.2 lg/ml of bovine serum albumin (BSA); 0.3 lM of each primer; 0.2 mM of each dinucleotide triphosphate; 1 unit of Taq polymerase (Invitrogen Life Technologies) and 4.0 ll of DNA template. The thermocycling protocol consisted of 94C for 5 min, next 30 cycles of 94C for 30 s, 56C for 1 min and 72C for 1 min, followed by a final extension time of 65C for 45 min. Amplified products were electrophoresed and visualized in a 2.0% agarose gel stained with SYBRgreenTM (Invitrogen).

Results DNA extractions from woodland caribou faecal material, yielded PicogreenTM DNA concentration estimates ranging from 6.5 ng/ll to 28.6 ng/ll (mean total yield per 65 ll elution volume = 1053 ng). All of the caribou samples used in this report had PicogreenTM estimates well above the 1.0 ng/ll threshold and could be diluted to the defined dilution series. For 52 of the 60 caribou faecal samples, amplification of the sample dilution series, using both the HMW and LMW

Conserv Genet (2007) 8:577–586

gender markers, produced bands of comparable intensity to the corresponding products of the dilution series of control DNA. Six sub-samples had LMW amplification products that were much weaker than the control DNA products at the corresponding concentration. For these sub-samples, by comparing the intensity of the LMW sample product with the control DNA products throughout the entire dilution series, we were able to ascertain their relative amounts. For example in sample Rta-01 (Fig. 2), the product intensity of the 5.0 ng estimate did not correspond to its counterpart in the female control DNA amplification. However, it was in a similar intensity to the 0.5 ng amount in the female control. This represents a targetto-total DNA ratio of approximately 10%. Similarly, amplification of the 1.0 ng total DNA estimate was comparable to the 0.1 ng product in the control series, confirming the 10% target-to-total DNA ratio for this particular sample. We also identified two samples where the actual target DNA amplified was only 4% of the estimated 5.0 ng of total DNA, or 0.2 ng of target caribou DNA in our PCR amplification (i.e., sample Rta-06 in Fig. 2). Two sub-samples exhibited a pattern characteristic of DNA degradation where in the HMW marker had reduced amplification in the dilution series compared to the LMW target-to-total DNA ratios. Subsamples with low target-to-total ratios occurred at random and were not confined to a particular sample of faeces. In addition, the total DNA dilution of each sample to a concentration of 1.25 ng/ll permitted, on

581

average, a 12-fold dilution, alleviating the potential effects of high inhibitor-to-DNA ratios. All caribou faecal DNA samples amplified at all microsatellite loci. Genotype profiles were consistent among all replicate extractions and re-amplifications (3 samples per individual · 3 amplifications = 9 genotype profiles per individual), there was no evidence of allele dropout or the generation of non-specific amplification products that resembled alleles. We did observe a reduction in allele peak heights (measured in RFU) in relation to the estimated amount of target DNA entered into microsatellite amplification. Allele peak heights of samples containing 0.5 ng and 0.2 ng of target DNA were approximately 12.5% (5,000 RFU) and 1.25% (500 RFU), respectively of the allele peak heights generated using 5.0 ng (40,000 RFU) of target DNA template under identical PCR amplification conditions (Fig. 3). Comparatively, swift fox faecal extractions on average provided only half the amount of total DNA ranging from 1.0 ng/ll to 26.1 ng/ll (mean total DNA yield per 65 ll elution volume = 488 ng). Our PCR amplification assay showed that only 19 of the 50 samples extracted had quantities of target DNA above 0.1 ng. Of these 19 samples exhibiting amplification, comparison with the control DNA amplification assay showed that in 10 of these samples, the amplification product of the estimated total DNA of 5.0 ng was comparable to the 0.2 ng amplification of the control, indicating a target-to-total DNA ratio of only 4%. The remaining nine samples exhibited target DNA amplification above 0.5 ng with four samples having amplification product comparable of the 5.0 ng total DNA estimate to 5.0 ng of target DNA (100% target-to-total DNA ratio) and five having amplification product comparable to 1.0 ng of target DNA (20% target-tototal DNA ratio) (Table 1). An example of the swift fox comparative assay is shown in Fig. 4.

Discussion

Fig. 2 Comparative amplification assay to determine relative amounts of woodland caribou target DNA. Each sample (Rta-01 to Rta-06) is amplified with LMW gender markers, using estimated total DNA template of 5.0, 1.0, 0.5, 0.2 and 0.1 ng. The two controls male (top row) and female (bottom row) were amplified using known amounts of target DNA at 5.0, 1.0, 0.5, 0.2 and 0.1 ng. These samples were electrophoresed in a 2.0% agarose stained with SYBRgreenTM

Non-invasively collected faeces can provide scientists with a valuable source of DNA for studies on wildlife genetics, but there remains a perception that it provides DNA of only low quantities and quality (Gerloff et al. 1995; Flagstad et al. 1999; Morin et al. 2001; Piggott and Taylor 2003; Wehausen et al. 2004). Even labs that have had success in obtaining genotypes from faecal DNA tend to operate under this perception and adopt protocols, which can increase costs (Wilson et al. 2003), errors (Gill et al. 2000; Whitaker et al. 2001) and/or the unnecessary removal of samples (Woods

123

582

et al. 1999). Surprisingly, little effort has been made to address this issue (but see Morin et al. 2001). In this paper, we have demonstrated a feasible protocol to characterize the quantity and quality of target DNA from faecal sources. PicogreenTM estimates of total DNA concentration indicated that extractions using surface swabbing of three woodland caribou faecal pellets from the same individual, provided on average, over 1,000 ng of total DNA. In the majority of the caribou faecal samples used for this report, concordance between these estimates and the comparative HMW and LMW target DNA amplification assay provided strong evidence that most of the DNA was in fact target DNA, with little DNA from other sources. In some samples there was evidence of DNA degradation, as there were more successful amplifications of the LMW marker than for the HMW marker. The low prevalence of DNA degradation is not surprising given that these samples were collected frozen, reducing mucosa degradation and that they are from herbivores having no cross amplification from mammalian prey items. The amplification of HMW product in the majority of winter-collected woodland caribou samples indicates high yields of quality DNA comparable to tissue samples such as muscle or blood. We did encounter some sub-samples (10%) where there was a discrepancy between the PicogreenTM

Fig. 3 Allelic peak heights of the fluorescently labeled microsatellite BL42 generated using a MEGAbace 1,000 workstation for (a) target DNA at 5.0 ng (1.25 ng/ll); (b) 0.5 ng (0.125 ng/ll) and (c) 0.2 ng (0.05 ng/ll)

123

Conserv Genet (2007) 8:577–586

estimates and the LMW target DNA amplification assay (Fig. 2). In these samples, the estimated total DNA template used in the amplification assay was greater than the actual target DNA present. This resulted in a reduction in the strength of the amplification product of the estimated 5.0 ng, compared to its counterpart in the control DNA series suggesting the presence of DNA from other sources such as microorganisms or undigested items, which are not uncommon for faecal material. The characterization of low target-to-total DNA ratios in a small percentage of our samples is likely due to variation in the amount of mucous on the surface of the pellet. When the mucous layer is thin, wiping the pellet may have removed more solid faecal material into the extraction process. Although we used an extensive dilution series of each faecal sample for the caribou analyses for illustrative purposes, our smaller dilution series of each swift fox faecal sample was also shown to be practical and informative. Our assessment of target DNA extracted from potential swift fox faeces showed less success than that observed from caribou faeces. In these samples, PicogreenTM estimates provided on average 488 ng of total DNA, approximately 50% of that extracted from caribou faeces. Furthermore, only 40% of the samples exhibited amplification of the swift fox microsatellite marker illustrating high discordance between target and total DNA estimates. The lower success in

Conserv Genet (2007) 8:577–586

583

Table 1 A survey of target DNA quantity in faecal extracts from woodland caribou and swift fox, estimated from the amplification of 5.0 ng of total DNA then compared to a control target DNA amplification assaya Species

Amplification marker

5.0 ng ‡ · > 0.5 ng

0.5 ng > · ‡ 0.2 ng

· £ 0.1 ng

Total

Woodland caribou (R. tarandus)

HMW gender (860–920 bp) LMW gender (200–220 bp) Vve2-64 (194–212 bp)

52 (86)b 54 (90) 9 (18)

8 (14) 6 (10) 10 (20)

0 (0) 0 (0) 31 (62)

60 (100) 60 (100) 50 (100)

Swift fox (V. velox) a

Based upon comparison of control DNA amplification dilution series of 5.0 ng (1.25 ng/ll), 1.0 ng (0.25 ng/ll), 0.5 ng (0.125 ng/ll), 0.2 ng (0.05 ng/ll) and 0.1 ng (0.025 ng/ll)

b

Number of samples with percentage subtended by brackets

obtaining high quantities of target DNA is likely due to a combination of three factors. First, these samples were collected in summer and environmental degradation is probably significant even over short time periods. Previous studies have demonstrated that winter-collected faeces provide higher quality DNA than samples collected in other seasons (Lucchini et al. 2002; Piggott and Taylor 2003; Piggott 2004; Maudet et al. 2004; Ha´jkova´ et al. 2006). Cold winter conditions likely help to preserve DNA by reducing cell degradation and maintaining an intact mucous layer on the surface of faeces. Second, many of the scats swabbed contained digested prey items such as hair and bone. Although care was taken not to include these items in our extractions, it is likely that some of this material was co-extracted and would contribute to total DNA estimates. The third possibility is that the faeces originated from a closely related species such as the red fox or potentially other genera such as coyote. Although designed for swift fox, cross homology of canid primers has been observed (Kitchen et al. 2005) and as a result amplification of the microsatellite locus used here may be successful in other canid species other than swift fox. For the amplified product of the dilution series to be informative, knowledge of the species of origin is required. Identifying the species that contributed the faecal material is critical for individual-based genotyping projects. Species-diagnostic sequences such as mitochondrial DNA (mtDNA) sequences such as control region or cytochrome b are often used (e.g., Dale`n et al. 2004). We recommend applying the nuclear assay before mtDNA amplification to determine the species, particularly if individual microsatellite genotypes are critical to the objectives of the project. Using these species diagnostic markers prior to target DNA has a higher probability of successful mtDNA amplification for species identification; however, nuclear DNA amplifications may fail due to lower copy number. In studies requiring nuclear genotypes, the most costeffective strategy is to apply species identification only

on those samples that have quantities of nuclear DNA that will provide genotypes at microsatellite loci. In this study, mtDNA amplification species identification (e.g., sequencing) of potential swift fox scat would be required on only 19 of 50 samples, saving considerable analytical time and cost. From these target DNA estimates, we were able to determine approximately how much target DNA was being entered into our amplification reactions and thus, remove the ambiguity associated with using arbitrary volumes of DNA template. Furthermore, by knowing the concentration of target DNA, we could reduce the concentration of PCR inhibitors by dilution, while maintaining a sufficient concentration of DNA template. Many researchers dilute their samples when using faecal extractions to unknown concentrations and/or increase the number of PCR cycles (Ernest et al. 2000; Morin et al. 2001; Lucchini et al. 2002; Fernando et al. 2003; Frantz et al. 2003; Bellemain and Taberlet 2004; Hung et al. 2004; Maudet et al. 2004; Roon et al. 2005), which increases the probability of amplification errors such as allelic drop-out and false alleles (Gill et al. 2000; Whitaker et al. 2001). We used a standard PCR protocol of 30 cycles and experienced no amplification errors in the characterized amounts of

Fig. 4 Comparative amplification assay to determine relative amounts of swift fox target DNA. Each sample (Vve-01 to Vve10) is amplified with the tetranucleotide marker Vve2-64, using estimated total DNA template of 5.0, 1.0 and 0.1 ng. The two controls, male (top row) and female (bottom row), were amplified using known amounts of target DNA at 5.0, 1.0, 0.5, 0.2 and 0.1 ng. These samples were electrophoresed in a 2.0% agarose stained with SYBRgreenTM

123

584

target DNA from woodland caribou faeces. Although we used lower quantities of BSA than recommended by Kreader (1996) and Wehausen et al. (2004) due to the constraints of using a capillary-based genotyping unit, we still observed no evidence of PCR inhibition. The target DNA characterization process also provides the opportunity to remove or categorize samples that have inadequate DNA quantities for successful genotype amplification. Woods et al. (1999) presented a method to exclude poor quality DNA extracted from hair samples by removing samples having unsuccessful target DNA amplification at a single microsatellite locus. This method does provide a valid screening process for samples such as hair; however, it cannot be confidently used for DNA from faeces. Since DNA extracted from faecal material may have high concentrations of inhibitor and low target-to-total DNA ratios, this screening procedure could result in the unnecessary removal of faecal DNA samples. Applying our protocol, the characterization of target DNA and dilution from total DNA estimates would identify such problems, allowing more samples to be used for analysis. In terms of categorizing samples of high and low probability of genotyping success, recent research has shown that allelic dropout is common when less than 0.2 ng of target DNA is used (Morin et al. 2001; Krenke et al. 2002; Leclair et al. 2003) and forensic laboratories recommend range of 2.0–0.5 ng of DNA for human profiling (Leclair et al. 2003). We were able to maintain genotype consistency among all replicate caribou faecal samples, but we did notice a considerable reduction in RFU signal of microsatellites amplified from DNA template less than 0.5 ng (Fig. 3). At target DNA amounts of approximately 0.2 ng, allele peak heights were considerably lower and at risk to allelic dropout. In profiling woodland caribou samples, we implemented a strategy to remove samples in this range from downstream analyses in order to decrease the risk of genotype error and significant labour in confirming genotypes (e.g., multiple-tubes approach). While this approach is optimal for winter collected woodland caribou scat, due to our high success rate, other projects with less than ideal samples can still utilize this protocol to categorize target DNA template. The swift fox project represents the latter scenario as our quantification protocol resolved many of the samples having target DNA at thresholds £ 0.2 ng, requiring microsatellite amplification using such modifications as a multiple tubes approach (Navidi et al. 1992; Taberlet et al. 1996, 1999; Miller et al. 2002; Frantz et al. 2003). In the laboratory, the use of a surface wiping technique is an effective way to maximize the collection of

123

Conserv Genet (2007) 8:577–586

target DNA from the mucous layer while minimizing contamination from non-target DNA and PCR inhibitors. Our wiping technique differs from the surface scraping technique used by Fernando et al. (2003) and Wehausen et al. (2004), which may bring solid faecal material and consequently inhibitors into the extraction procedure (Waits and Paetkau 2005). In summary, we have developed a protocol of characterizing faecal DNA using fluorescence quantification assays and a comparison of nuclear amplification product to predetermine the quantity of total and target DNA prior to downstream analysis, we feel that it can be easily utilized or adapted to other species where faeces is the primary source of DNA template. We speculate that many of the problems associated with faecal DNA are a consequence of the uncertainty involved with the amount of target DNA available and arbitrary compensation for assumed low-template and its myriad of problems. In light of these issues, it is highly recommended that two steps be taken to characterize faecal DNA prior to analysis: (1) Quantification of total DNA content in faecal extracts using a fluorescence assay (or whichever technique is feasible for your laboratory); and (2) Quantification of targetto-total DNA ratio through the comparison of amplified product from a dilution series of faecal sample total DNA estimates, with control DNA of the target species at known quantities. Based on the outcome of these first two steps, decisions can be made regarding which samples to pursue (i.e., those from critical areas) in downstream analyses and the analytical method to employ (i.e., multiple tubes). When developing an amplification assay for individual projects, we recommend using a DNA marker in the size range of the largest microsatellite to be profiled which is taxonspecific. Gender markers optimized for a particular taxon are ideal as they provide important sex identification information in addition to sample DNA quality and quantity. Ultimately, the initial investment of thoroughly characterizing DNA extracted from faeces can significantly reduce downstream genotyping problems or allow for the appropriate implementation of compensatory protocols to account for low template on specific batches of samples. Acknowledgements Financial support for this project was provided in part by NSERC, separately to Paul Wilson and Mark Ball, Parks Canada Species at Risk Recovery Action and Education Fund, a program supported by the National Strategy for the Protection of Species at Risk, Ontario Parks (OMNR), Manitoba Department of Conservation and Manitoba Hydro. We would like to thank Medea Curteanu who provided the swift fox samples used in this report. Finally, we want to thank the reviewers of this manuscript for their valuable comments.

Conserv Genet (2007) 8:577–586

References Banks SC, Piggott MP, Hansen BD, Robinson NA, Taylor AC (2002) Wombat coprogenetics: enumerating a common wombat population by analysis of faecal DNA. Aust J Zool 50:193–204 Bellemain E, Taberlet P (2004) Improved non-invasive genotyping method: application to brown bear (Ursus arctos) feces. Mol Ecol Notes 4:519–522 Broquet T, Petit E (2004) Quantifying genotyping errors in noninvasive population genetics. Mol Ecol 13:3601–3608 Cullingham CI, Smeeton C, White BN (2006) Isolation and characterization of Swift Fox tetranucleotide microsatellite loci. Mol Ecol Notes (in press) Dale`n L, Go¨therstro¨m A (2004) Identifying species from pieces of faeces. Conserv Genet 5:109–111 Ernest HB, Penedo CT, May BP, Syvanen M, Boyce WM (2000) Molecular tracking of mountain lions in the Yosemite Valley region in California: genetic analysis using microsatellites and faecal DNA. Mol Ecol 9: 433–441 Farrell LE, Romant J, Sunquist ME (2000) Dietary separation of sympatric carnivores identified by molecular analysis of scats. Mol Ecol 9:1583–1590 Fernando P, Vidya TN, Rajapakse C, Dangolla A, Melnick DJ (2003) Reliable noninvasive genotyping: fantasy or reality. J Hered 94:115–123 Flagstad O, Roed K, Stacy JE, Jakobsen KS (1999) A reliable non-invasive genotyping based on excremental PCR of nuclear DNA purified with a magnetic bead protocol. Mol Ecol 8:879–883 Frantz AC, Pope LC, Carpenter PJ, Roper TJ, Wilson GL, Delahay RJ, Burke T (2003) Reliable microsatellite genotyping of the Eurasian badger (Meles meles) using faecal DNA. Mol Ecol 12:1649–1661 Gill P, Whitaker JP, Flaxman C, Brown N, Buckleton J (2000) An investigation of the rigor of interpretation sites for STRs derived from less than 0.1 ng of DNA. Forensic Sci Int 112:17–40 Gerloff U, Schlotterer C, Rassmann K, Rambold I, Hohmann G, Fruth B, Tautz D (1995) Amplification of hypervariable simple sequence repeats (microsatellites) from excremental DNA of wild living bonobos (Pan paniscus). Mol Ecol 4:515–518 Ha´jkova´ P, Zemanova´ B, Bryja J, Ha´jek B, Roche K, Tkadlec E, Zima J (2006) Factors affecting success of PCR amplification of microsatellite loci from otter faeces. Mol Ecol Notes 6:559–562 Hoss M, Kohn M, Paabo S, Knauer F, Schroder W (1992) Excremental analysis by PCR. Nature 359:199 Hung CM, Li SH, Lee LL (2004) Faecal DNA typing to determine the abundance and spatial organisation of otters (Lutra lutra) along two stream systems in Kinmen. Anim Conserv 7:301–311 Kitchen AM, Gese EM, Waits LP, Karki SM, Scaustet ER (2005) Genetic and spatial structure of a swift fox population. J Anim Ecol 74:1173–1181 Kohn MH, Wayne RK (1997) Facts from faeces revisited. Trends Ecol Evol 12:223–227 Kreader CA (1996) Relief of amplification inhibition in PCR with bovine serum albumin or T4 gene 32 protein. Appl Environ Microbiol 62:1102–1106 Krenke BE, Tereba A, Anderson SJ, Buel E, Culhane S, Finis CJ, Tomsey CS, Zachetti JM, Masibay A, Rabbach DR, Amiott EA, Sprecher CJ (2002) Validation of a

585 16-locus fluorescent multiplex system. J Forensic Sci 47: 773–785 Leclair B, Sgueglia JB, Wojtowicz PC, Juston AC, Fre´geau CJ, Fourney RM (2003) STR DNA typing: increased sensitivity and efficient sample consumption using reduced PCR reaction volumes. J Forensic Sci 48:1001–1013 Lucchini V, Fabbri E, Marucco F, Ricci S, Boitani L, Randi E (2002) Non-invasive molecular tracking of colonizing wolf (Canis lupus) packs in the western Italian Alps. Mol Ecol 11:857–868 Maudet C, Luikart G, Dubray D, Von Hardenberg A, Taberlet P (2004) Low genotyping error rates in wild ungulate faeces sampled in winter. Mol Ecol Notes 4:772–775 Miller CR, Joyce P, Waits LP (2002) Assessing allelic drop-out and genotype reliability using maximum likelihood. Genetics 160:357–366 Moore SS, Barendse W, Berger KT, Armitage SM, Hetzel DJ (1992) Bovine and ovine DNA microsatellites from the EMBL and Genbank databases. Anim Genet 23: 463–467 Morin PA, Chambers KE, Boesch C, Vigilant L (2001) Quantitative PCR analysis of DNA from noninvasive samples for accurate microsatellite genotyping of wild chimpanzees (Pan troglodytes verus). Mol Ecol 10: 1835–1844 Murphy MA, Waits LP, Kendall KC (2003) The influence of diet on faecal DNA amplification and sex identification in brown bears (Ursus arctos). Mol Ecol 12:2261–2265 Navidi W, Arnheim N, Waterman MS (1992) A multiple tubes approach for accurate genotyping of very small DNA samples by using PCR: statistical considerations. Am J Hum Genet 50:347–359 Piggott MP (2004) Effect of sample age and season of collection on the reliability of microsatellite genotyping of faecal DNA. Wildl Res 31:485–493 Piggott MP, Taylor AC (2003) Extensive evaluation of faecal preservation and DNA extraction methods in Australian native and introduced species. Aust J Zool 51:341– 355 Roon DA, Thomas ME, Kendall C, Waits LP (2005) Evaluating mixed samples as a source of error in non-invasive genetic studies using microsatellites. Mol Ecol 14:195–201 Shaw CN, Wilson PJ, White BN (2003) A reliable molecular method of gender determination for mammals. J Mammal 84:123–128 Taberlet P, Griffin S, Goossens B, Questiau S, Manceau V, Escaravage N, Waits LP, Bousvet J (1996) Reliable genotyping of samples with very low DNA quantities using PCR. Nucleic Acids Res 24:3189–3194 Taberlet P, Waits LP, Luikart G (1999) Non-invasive genetic sampling: look before you leap. Trends Ecol Evol 14: 323–327 Waits LP, Paetkau D (2005) Noninvasive genetic sampling tools for wildlife biologists: a review of applications and recommendations for accurate data collection. J Wildl Manage 69:1419–1433 Wasser SK, Houston CS, Koehler GM, Cadd GG, Fain SR (1997) Techniques for application of faecal DNA methods to field studies in Ursids. Mol Ecol 6:1091–1097 Wehausen JD, Ramey RR II, Epps CW (2004) Experiments in DNA extraction and PCR amplification from bighorn sheep feces: the importance of DNA extraction method. J Hered 95:503–509 Whitaker JP, Cotton EA, Gill P (2001) A comparison of the characteristics of profiles produced with the AMPFISTR

123

586 SGM Plus multiplex system for both standard and low copy number (LCN) STR DNA analysis. Forensic Sci Int 123:215–223 Wilson GA, Strobeck C, Wu L, Coffin JW (1997) Characterization of microsatellite loci in woodland caribou Rangifer tarandus, and their use in other artiodactyls. Mol Ecol 6:697–699

123

Conserv Genet (2007) 8:577–586 Wilson GJ, Frantz AC, Pope LC et al (2003) Estimation of badger abundance using faecal DNA typing. J Appl Ecol 40:658–666 Woods JG, Paetkau D, Lewis D, McLellan BN (1999) Genetic tagging of free-ranging black and brown bears. Wildl Soc Bull 27:616–627