Limnology

, Volume 17, Issue 1, pp 23–32 | Cite as

Techniques for the practical collection of environmental DNA: filter selection, preservation, and extraction

  • Toshifumi Minamoto
  • Takafumi Naka
  • Kazuhiko Moji
  • Atsushi Maruyama
Research paper

Abstract

Environmental DNA (eDNA) analysis has recently been used for detection of aquatic macro-organisms; however, the analytical procedures used in previous studies have not been optimized for practical use. Here, we compared several methods for DNA enrichment and extraction from water samples to establish widely applicable techniques for eDNA analysis using common carp as the model species. First, several types of filters were compared to identify the optimal filter type. Second, the eDNA yield was compared after a variety of extraction and isolation steps, including a combination of phenol extraction, ethanol precipitation (phenol treatment), and ultrafiltration. Third, DNA fixation with ethanol was tested for the preservation of eDNA on filters. Ethanol precipitation yielded the largest number of eDNA copies, followed by filtering using a 0.2-μm polycarbonate filter and a 0.7-μm glass fiber filter. Phenol treatment resulted in collection of a higher number of eDNA copies than that collected using ultrafiltration. DNA fixation with 15 ml ethanol enabled eDNA preservation on the filters at ambient temperatures for at least 6 days. Finally, combinations of different filter types and DNA enrichment procedures were compared using field water samples. From these results, we propose that the appropriate selection method for eDNA analysis should be chosen based on context. For example, when a high concentration of the target DNA is expected, such as in an aquarium experiment, ethanol precipitation is advantageous. However, when the target DNA is rare, which is the case in most field studies, filtration followed by freezing or DNA fixation by ethanol and phenol treatment are recommended. The filter type should be decided prior to the survey based on the characteristics of the water of interest. Thus, eDNA analysis could be applied to various situations using adaptive combinations of these techniques.

Keywords

DNA extraction Environmental DNA Fish Quantitative PCR 

Introduction

The distribution of species is fundamental information needed for ecological studies, which are necessary for evaluating the risk of extinction and for planning efficient conservation approaches for focal organisms (Krebs 1978). In addition to presence/absence data, information regarding the abundance and biomass of a species would enable quantitative estimation of the extinction risk of endangered species or the expansion risk of introduced species (Brook et al. 2000), facilitating the practical management of wildlife populations. However, the precise evaluation of distribution and biomass is generally time consuming, costly, and labor-intensive. This is particularly true for aquatic ecosystems, where most organisms are not visible or catchable owing to poor visibility, accessibility, and complex vegetation.

In the last two decades, environmental DNA (eDNA), which includes whole organisms and free-drifting DNA fragments in the water, has become a powerful tool for determining the distribution and diversity of aquatic microbes (Lydolph et al. 2005). Specifically, detection of short species-specific eDNA fragments in water has facilitated detection of aquatic microbial biodiversity in aquatic ecosystems (Sogin et al. 2006; Taberlet et al. 2012). Recently, analysis of eDNA has been used to detect aquatic macro-organisms, including amphibians (Ficetola et al. 2008; Fukumoto et al. 2015; Goldberg et al. 2011; Olson et al. 2012), freshwater fishes (Jerde et al. 2011; Minamoto et al. 2012; Takahara et al. 2013; Maruyama et al. 2014; Takahara et al. 2012; Doi et al. 2015b; Janosik and Johnston 2015), mammals (Foote et al. 2012; Thomsen et al. 2012b), and invertebrates (Deiner and Altermatt 2014; Thomsen et al. 2012b; Treguier et al. 2014). Environmental DNA techniques have also been applied to various habitats, including lakes (Takahara et al. 2012; Thomsen et al. 2012b; Eichmiller et al. 2014), ponds (Dejean et al. 2012; Ficetola et al. 2008; Takahara et al. 2013, 2015), rivers (Goldberg et al. 2011; Jerde et al. 2011; Minamoto et al. 2012; Olson et al. 2012; Thomsen et al. 2012b; Fukumoto et al. 2015), and sea (Foote et al. 2012; Thomsen et al. 2012a). Biomass estimation was also shown to be possible by counting the number of eDNA copies in water samples by quantitative polymerase chain reaction (qPCR) (Pilliod et al. 2013; Takahara et al. 2012; Thomsen et al. 2012b; Eichmiller et al. 2014) or droplet digital PCR (ddPCR) (Doi et al. 2015a, b; Nathan et al. 2014). Therefore, the development of eDNA techniques may dramatically reduce the labor requirements for studies of the distribution and biomass of aquatic organisms from many taxa inhabiting various bodies of water. In particular, the development of eDNA techniques could provide rapid and accurate distribution data for alien species (Jerde et al. 2011; Takahara et al. 2013; Fukumoto et al. 2015), which cannot be easily obtained using existing methods, such as direct catching by angling or casting nets, or visual observation by SCUBA divers.

Environmental DNA techniques have been used for analysis of aquatic macro-organisms for only a few years (Ficetola et al. 2008). As such, researchers have employed a number of methods to collect eDNA without determination of the optimal methods to use under different contexts, and further studies are needed to optimize eDNA collection methods. Generally, the eDNA density for a target species is low in field water samples, and enrichment of eDNA in water samples is often required. To address this issue, many researchers have adopted the ethanol precipitation method (Dejean et al. 2011, 2012; Ficetola et al. 2008; Foote et al. 2012; Minamoto et al. 2012; Thomsen et al. 2012a). However, for ethanol precipitation, twice the volume of ethanol is typically added to the sample, which restricts the maximum sample volume and detection power of this technique. Therefore, in some studies, filtration is used as the first enrichment step (Goldberg et al. 2011; Jerde et al. 2011; Minamoto et al. 2012; Olson et al. 2012; Takahara et al. 2012, 2013; Thomsen et al. 2012b; Fukumoto et al. 2015). The use of several types of filters has been reported; for example, Turner et al. (2014) showed that eDNA of fish is most abundant in size classes from 1 to 10 μm. Moreover, filters may need to be preserved until analysis for some assays (Renshaw et al. 2015). In some studies, the enrichment yield was further improved after filtration using an ultracentrifugation unit (Minamoto et al. 2012; Takahara et al. 2012, 2013). Finally, a DNA extraction step is generally performed using commercial DNA extraction kits. Thus, the DNA enrichment method for eDNA analysis is not standardized, and a recent study indicated that the choice of protocols for enrichment of eDNA from water can strongly affect detection of eDNA (Deiner et al. 2015).

Therefore, in this study, we compared several enrichment and extraction methods and proposed context-specific selection of methods for eDNA analysis of aquatic vertebrates.

Materials and methods

Study fish and eDNA samples

A freshwater fish, the common carp (Cyprinus carpio, Linnaeus 1758), was used as the model species in this study because it is one of the most widely distributed species worldwide and one of the 100 worst invasive alien species (Lowe et al. 2000). Therefore, efficient and early detection of this fish is required for management of invasion. Moreover, previous studies have examined the application of some eDNA techniques for detection of this fish (Takahara et al. 2012; Doi et al. 2015a, b; Eichmiller et al. 2014; Takahara et al. 2015).

Ten juvenile common carp (20, 20, 20, 22, 28, 28, 28, 30, 38, and 38 g wet weight; sex unknown) were held in a 40-l plastic tank (30 × 45 × 25 cm). The water was continuously aerated and circulated using an internal filter (Roka Boy, GEX corporation, Osaka, Japan). The fish were fed a commercial diet (Saki-Hikari, Kyorin Co. Ltd., Hyogo, Japan), which was mainly comprised of fish meal, flour, and maize (protein >40 %, fat >6.0 %, fiber <3.0 %, and phosphorus >1.0 %). To confirm that the diet did not include carp DNA, real-time PCR analysis was performed as described below. The fish were fed 3 times each week and kept at 19 ± 1 °C with a 12:12-h light–dark cycle. In all experiments performed in this study, water samples were collected from the surface of the water in the fish tank.

Experimental scheme

First, DNA extraction yield using ethanol precipitation and filtration methods was compared (experiment A-a). In the filter comparison, four polycarbonate filters with different pore sizes (0.2, 0.8, 3.0, and 12.0 µm) and a glass fiber filter (mesh size: 0.7 µm) were compared. Second, because the glass fiber filter performed better in the above experiment, a comparison among different glass fiber filters was carried out (experiment A-b). Third, postfiltration eDNA enrichment methods (i.e., phenol treatment versus ultracentrifugation) were compared using eDNA trapped with glass fiber filters (experiment B). Fourth, the validity of ethanol treatment for eDNA preservation on filters was tested (experiment C). After all tank experiments, these methods were tested in field water samples (field survey).

Experiment A-a: identifying the optimal filter type for eDNA collection

To identify the optimal filter type for collecting an adequate amount of eDNA under stable conditions, we compared the numbers of eDNA copies collected by seven different methods, i.e., ethanol precipitation; filtration using polycarbonate filters with mesh sizes of 0.2 µm (GTTP, Isopore membrane filter; Nihon Millipore KK, Tokyo, Japan), 0.8 µm (ATTP), 3.0 µm (TSTP), or 12.0 µm (TKTP); glass fiber filters with a pore size of 0.7 µm (Whatman GF/F; GE Healthcare, Buckinghamshire, UK); and centrifugation (no filtration or precipitation).

Water (4.5 l) was collected from the surface of the fish tank using an autoclaved plastic cup for all methods in order to control possible variations related to sampling. For the filtration treatments, 100 ml of water was collected from the plastic cup and immediately filtered using each of the five filter types. All filtration equipment was washed with tap water and carefully rinsed with distilled water between filtration operations to prevent cross-contamination (Fukumoto et al. 2015). Environmental DNA samples retained on filters were resuspended in 10 ml of ultrapure water twice using vibration for 5 min to obtain approximately 20 ml of eDNA sample solution.

For ethanol precipitation treatment, 100 ml of sample water was poured into an autoclaved plastic bottle, and 10 ml of 3 M NaAc (pH5.2) and 200 ml of ethanol were added. Samples with added ethanol were centrifuged for 1 h at 10,000 × g at 4 °C, and the supernatants were discarded. The pellets were resuspended in 10 ml of ultrapure water twice to obtain approximately 20 ml of eDNA solution. The following procedures were performed using sterilized disposable plastic tubes. Each method (experimental treatment or centrifugation) was performed with four replicates. Environmental DNA sample solutions (20 ml) were reconcentrated by ultrafiltration using an Amicon Ultra-15 (30 kDa; Nihon Millipore KK) to yield a final volume of 200 μl. For the centrifugation sample, 20 ml of water was directly concentrated by ultrafiltration. DNA was further purified using a DNeasy Blood and Tissue Kit (Qiagen GmbH, Hilden, Germany). The final DNA solution (100 µl) was analyzed by real-time qPCR as described below.

Experiment A-b: identifying the optimal filter type from glass fiber filters for eDNA collection

To identify the optimal filter type from glass fiber filters for stable and adequate eDNA collection, the number of eDNA copies collected using glass fiber filters with pore sizes of 0.7 µm (GF/F), 1.6 µm (GF/A: Whatman), and 2.7 µm (GF/D: Whatman) were compared.

Water (1.2 l) was collected from the surface of the fish tank using an autoclaved plastic cup, and 100 ml of water was filtered using each of the three filter types with four replicates. The following procedures were performed using sterilized disposable plastic tubes. Environmental DNA samples on filters were resuspended in 20 ml of DNA extraction buffer (10 mM Tris–HCl, pH 8.0; 10 mM EDTA; 0.5 % SDS), and 50 µg ml−1 of proteinase K was added. The samples were incubated at 56 °C for 15 min. DNA was then extracted by the phenol–chloroform method (Fukumoto et al. 2015) and concentrated by ethanol precipitation. The pellets were resuspended in 250 µl of ultrapure water and stored in 1.5-ml microtubes. DNA was further purified using a DNeasy Blood and Tissue Kit and was analyzed by real-time qPCR as described below.

Experiment B: phenol treatment or ultrafiltration for eDNA enrichment

To identify better methods for increasing the concentration of eDNA collected using filters, postfiltration eDNA concentrations were compared after collection of eDNA using a combination of phenol extraction, ethanol precipitation (phenol treatment), and ultrafiltration, which has been used for quantitative determination of environmental microbes (Honjo et al. 2010; Minamoto et al. 2009).

Sample water (0.8 l) was collected from the surface of the fish tank with a plastic cup, and 100 ml of water was filtered using a GF/F filter with eight replicates (four for phenol treatment and four for ultrafiltration). The procedures used for phenol treatment and ultrafiltration were the same as those used in experiments A-a and A-b, respectively. The final DNA solution was analyzed by real-time qPCR as described below.

Experiment C: efficacy of the ethanol treatment for preserving eDNA

To determine the efficacy of ethanol treatment for preservation of eDNA on filters, we examined changes in the number of eDNA copies on filters treated with or without ethanol.

Water (3.2 l) was collected from the surface of the fish tank with an autoclaved plastic cup, and 32 replicates (100 ml water each) were filtered using a GF/F filter (16 for ethanol treatment and 16 for no treatment). For the ethanol treatment, 15 ml ethanol was added immediately after filtration. The untreated sample received no treatment prior to ultrafiltration. The filters, which contained a small amount of water, were kept at 25 °C and were moved to a freezer at −25 °C on days 0 (immediately after filtration), 1, 3, or 6. Ethanol-treated and untreated samples were replicated using four filters each day. The procedures used for enrichment and extraction were the same as those described in experiment A-a. The final DNA solution was analyzed by real-time qPCR as described below.

Field survey: comparison between DNA extraction methods using field water samples

To confirm the applicability of the methods established by experiments A and B, water samples were collected at a pond and natural lake (Iba-naiko Lagoon), both of which are known habitats of the common carp. Specifically, on July 4, 2013, surface water (25 l) was collected from a pond where carp are reared in the inner yard of the Research Institute for Humanity and Nature, Kyoto, Japan. On July 18, 2013, the same volume of surface water was collected at Iba-naiko Lagoon, which is a known reproductive habitat of the common carp (Takahara et al. 2012; Uchii et al. 2011).

Collected water samples were immediately transported to the laboratory. Briefly, 500 ml of each water sample was filtered using one of the six types of filters (glass fiber filters with pore sizes of 0.7, 1.6, and 2.7 µm or polycarbonate filters with mesh sizes of 0.2, 0.8, and 3.0 µm). Because the finer filters clogged and could not filter the entire 500 ml, filtration was aborted after 10 min. For the pond water sample, 350–380 ml was filtered before clogging with 0.2-µm mesh polycarbonate filters. For the lake water, 125–140, 160–180, and 260–300 ml of water were filtered before clogging with polycarbonate filters with mesh sizes of 0.2, 0.8, and 3.0 µm, respectively. Six replicate filtrations were carried out for each filter type, and the total number of the filters was 72 (two sampling sites × six filter types × six replications). Of the six replications, three were used for phenol treatment and three for ultrafiltration, respectively.

The methods used for DNA enrichment and extraction were the same as those described above. From each water sample, 15 ml was ethanol-precipitated in triplicate. The pellet was then resuspended in 200 ml water, and DNA was extracted with a DNeasy Blood and Tissue Kit. For each treatment, a blank control (500 ml pure water filtered with each filter and 15 ml pure water that was ethanol precipitated) was used, and each blank control was treated the same as the corresponding samples. In total, 26 blank controls were employed. The final DNA solution was analyzed by real-time qPCR, as described below.

Real-time qPCR

Quantification of eDNA copies was performed using real-time TaqMan PCR with a StepOnePlus Real-Time PCR system (Life Technologies, Carlsbad, CA, USA) according to the method of Takahara et al. (2012). Mitochondrial cytochrome b gene fragments were amplified and quantified using the primers CpCyB_496F and CpCyB_573R and the probe CpCyB_550p (Takahara et al. 2012). These primers are specific to the common carp and amplify a 78-bp fragment of the cytochrome b gene. The specificity of the primers was tested with sequences of 56 species (all species for which the sequence data of the target region were available) inhabiting Lake Biwa, Japan (Takahara et al. 2012).

Each TaqMan reaction contained 900 nM of each primer, 125 nM TaqMan probe in 1 × PCR master mix (TaqMan Gene Expression Master Mix; Life Technologies), and 2 µl of the DNA solution. The total volume of each reaction mixture was 20 µl. The PCR conditions were as follows: 2 min at 50 °C, 10 min at 95 °C, and 40 cycles of 15 s at 95 °C and 60 s at 60 °C. Real-time qPCR was performed for each sample in triplicate, and the mean value was used in the analyses. The PCR products of the target sequences were cloned into the pGEM-T Easy Vector (Promega, Tokyo, Japan), and a dilution series of the plasmid containing 3 × 101 − 3 × 104 copies per PCR tube was amplified as standards in triplicate for all qPCR assays.

To avoid contamination, we applied the same precautions adopted in a previous study (Fukumoto et al. 2015). Briefly, we performed filtration, DNA extraction and real-time PCR set-up, and real-time PCR cycling in three separate rooms. To prevent carry-over contamination, after the real-time PCR experiment was started, no other experiments were performed on the same day. To monitor the contamination during real-time PCR set-up, three wells of no-template negative controls were used for all qPCR assays; all negative controls exhibited no amplification. For the field samples, no amplification was observed for blank controls.

Data analyses

The number of common carp eDNA copies per 1-l water sample was calculated on the basis of qPCR results. In experiment A, the number of eDNA copies collected by different types of filters was compared with those using the traditional treatments (experiment A-a, the centrifugation sample; experiment A-b, GF/F) using Dunnett’s test (Dunnett 1964). Coefficients of variations (CVs; standard deviation divided by mean) were calculated for both trials to evaluate the relative stability of the eDNA copies collected. In experiment B, the number of eDNA copies collected after enrichment by phenol treatment and ultrafiltration was compared using Welch’s t-test. In experiment C, the number of eDNA copies for ethanol treated and untreated samples on days 0, 1, 3, and 6 was compared using two types of Dunnett’s tests. In each of the field surveys (pond and lake), the effects of the DNA extraction method (phenol–chloroform or ultracentrifugation), filter type (six types of filters), and interaction conditions (extraction method × type of filter) on the log-transformed number of eDNA copies were examined using two-way analysis of variance (ANOVA). In addition, the numbers of eDNA copies collected by ethanol precipitation and other methods were compared using Dunnett’s test.

All statistical analyses were conducted using R ver. 2.13.0 software (R Development Core Team 2011).

Results

Experiment A-a: optimal filter type for collecting eDNA

The number of eDNA copies collected using ethanol precipitation was significantly higher than that collected using centrifugation (Dunnett’s test, |t| = 5.36, P < 0.01; Fig. 1a). The numbers of eDNA copies collected using filtration with polycarbonate filters with mesh sizes of 0.8, 3.0, and 12.0 μm were significantly lower than those collected using centrifugation (Dunnett’s test, |t| > 6.31, P < 0.01). Filtration using polycarbonate filters with a mesh size of 0.2 µm and glass fiber filters with a pore size of 0.7 μm (GF/F) did not result in significantly different eDNA copy numbers compared with those obtained in the centrifugation samples (Dunnett’s test, |t| < 3.83, P > 0.05). The CV values tended to be higher following ethanol precipitation or filtration by polycarbonate filters with mesh sizes of 0.8 μm or more (Fig. 1b).
Fig. 1

Comparison of the number of environmental DNA (eDNA) copies recovered (a) and each coefficient of variation (CV) (b) using seven methods: ethanol precipitation (ethanol); filtration with polycarbonate filters with mesh sizes of 0.2, 0.8, 3.0, and 12.0 µm; glass fiber filter with a pore size of 0.7 µm (GF/F); and centrifugation (no filtration or preservation). The number of DNA copies is normalized according to the liters of tank water. Each bar represents the mean value of four replicates, and error bars represent one standard deviation (SD). Double asterisks (**) indicate significant differences from the centrifugation sample by Dunnett’s test at P < 0.01

Experiment A-b: the optimal type of glass fiber filter for collecting eDNA

There tended to be fewer eDNA copies collected using glass fiber filters with pore sizes of 1.6 (GF/A) or 2.7 µm (GF/D) compared with a pore size of 0.7 µm (GF/F); however, the differences were not statistically significant (Dunnett’s test, |t| < 3.18, P > 0.05; Fig. 2a). The CV values tended to be lower when 0.7-µm filters were used (Fig. 2b).
Fig. 2

Comparison of the number of environmental DNA (eDNA) copies recovered (a) and each coefficient of variation (CV) (b) using glass fiber filters with pore sizes of 0.7 (GF/F), 1.6 (GF/A), or 2.7 µm (GF/D). Each bar represents the mean value of three replicates, and error bars represent one standard deviation (SD)

Experiment B: phenol treatment and ultrafiltration for concentrating eDNA

More eDNA copies were collected using phenol treatment than by using ultrafiltration (Welch’s t-test, |t| = 3.49, df = 3.833, P < 0.05; Fig. 3a). The CV values tended to be lower when phenol treatment was used (Fig. 3b).
Fig. 3

Comparison of the number of environmental DNA (eDNA) copies recovered (a) and each coefficient of variation (CV) (b) using two methods: phenol treatment and ultrafiltration. The phenol treatment results are the same as those shown in Fig. 2 (GF/F). Each bar represents the mean value of three replicates, and error bars represent one standard deviation (SD). The asterisk (*) indicates a significant difference by Welch’s t-test at P < 0.05

Experiment C: validity of ethanol treatment for preserving eDNA

The number of eDNA copies was not significantly decreased after 6 days for samples treated with ethanol (Dunnett’s test, |t| < 1.20, P > 0.1). However, significant decreases in the numbers of eDNA copies were observed in samples without ethanol treatment (Dunnett’s test, |t| > 7.91, P < 0.01; Fig. 4); without ethanol treatment, the number of eDNA copies decreased by 73 % within 1 day and 99 % within 6 days. The CV values tended to be lower in ethanol-treated samples compared with untreated samples (data not shown).
Fig. 4

Time-dependent changes in the relative number of environmental DNA (eDNA) copies on GF/F filters with and without ethanol treatment, as denoted by closed diamonds and open rectangles, respectively, at 25 °C. Each plot represents the mean value of four replicates, and error bars represent one standard deviation (SD). Double and triple asterisks (** and ***) indicate significant differences from day 0 by Dunnett’s tests at P < 0.01 and P < 0.001, respectively

Field survey: filter type and DNA extraction method

The results of the field survey are shown in Fig. 5. For the pond survey, the filter-phenol extraction recovered more eDNA copies than the filter-ultrafiltration method (two-way ANOVA, F(1, 24) = 94.4, P < 0.001). The type of filter also affected the number of recovered eDNA copies (two-way ANOVA, F(5, 24) = 7.5, P < 0.001). The interaction effect (extraction method × type of filter) was also significant (two-way ANOVA, F(5, 24) = 5.8, P = 0.001), which indicated that the optimal filter varied for the different extraction methods. For the lake survey, the filter-phenol extraction recovered more eDNA copies than the filter-ultrafiltration method (two-way ANOVA, F(1, 24) = 12.2, P = 0.002). The difference among filter types was not significant (two-way ANOVA, F(5, 24) = 0.9, P = 0.48); however, the interaction effect was significant (two-way ANOVA, F(5, 24) = 5.8, P = 0.001). Thus, the optimal filter was different between the two extraction methods.
Fig. 5

Comparison of the number of environmental DNA (eDNA) copies recovered by a combination of six types of filters and two postfiltration enrichment methods: phenol treatment and ultrafiltration. The results for field water samples collected at a a pond and b a lake are presented. Each bar represents the mean value of three replicates, and error bars represent one standard deviation (SD)

We compared the number of DNA copies recovered by ethanol precipitation and the other methods. For the pond survey, ethanol precipitation recovered significantly less DNA than 0.2-µm filter/phenol (Dunnett’s test, |t| = 5.22, P < 0.001) and 0.8-µm filter/phenol (Dunnett’s test, |t| = 5.46, P < 0.001) extractions. For the lake survey, ethanol precipitation recovered significantly less DNA than 0.2-µm filter/phenol (Dunnett’s test, |t| = 3.80, P < 0.01), 0.8-µm filter/phenol (Dunnett’s test, |t| = 3.64, P < 0.05), 3.0-µm filter/phenol (Dunnett’s test, |t| = 3.12, P < 0.05), and 0.7-µm glass fiber filter/phenol (Dunnett’s test, |t| = 5.25, P < 0.001) extractions.

Discussion

Environmental DNA techniques are promising methods for timely evaluation of aquatic species distributions in the current biodiversity crisis. Application of eDNA techniques to various environmental samples and quantification of eDNA concentrations in the field require an optimized method for concentrating eDNA from environmental samples. In this study, we compared the collection efficiencies of several methods and combinations of methods. Our data provided important information regarding optimal parameters for eDNA collection, isolation, and purification/enrichment.

In experiment A-a, we found that ethanol precipitation yielded the greatest number of eDNA copies, followed by centrifugation, filtration with a 0.2-µm polycarbonate filter, and filtration with a glass fiber filter with a pore size of 0.7 µm. Thus, ethanol precipitation minimized the loss of eDNA during sample collection. The use of 0.8-, 3.0-, and 12.0-µm polycarbonate filters reduced collection efficiency, probably because most of the eDNA was smaller than the size of the meshes, allowing it to pass through the filters. This result contradicted the findings of previous studies. For example, Turner et al. (2014) reported that the major fraction of eDNA measured between 1 and 10 μm, while Deiner et al. (2015) showed that there is no significant difference between eDNA capturing efficiency by filtration with GF/F and that with ethanol precipitation. However, these previous studies used field samples, and the size distribution of eDNA may differ between water from aquariums and field samples. Our results suggested that ethanol precipitation was preferable for aquarium experiments. However, the CV values of the number of eDNA copies were distinctively higher when ethanol precipitation was applied or when polycarbonate filters with mesh sizes of 0.8 µm or more were used. These high CV values should be considered when using these methods for precise quantification.

In contrast, since the concentration of target DNA is generally low in field samples, it is necessary to process large volumes of water samples, requiring additional labor in the form of transportation and preservation. Moreover, large volumes of water samples generally cannot be processed using ethanol precipitation. The mean value of DNA recovery was lowest for ethanol precipitation of field samples in the present study. Therefore, increasing the eDNA concentration using filtration is desirable for field samples. From the results of experiments A-a and A-b, a 0.2-µm polycarbonate filter and a glass fiber filter (GF/F) are recommended for large-volume samples because of their high collection efficiency and low estimation error (CVs). Given the difference in price (polycarbonate filters are usually more expensive than glass fiber filters) and the time required for filtration (a 0.2-µm filter clogs easily when using environmental samples), we recommend use of a glass fiber filter (GF/F). Most eDNA from macro-organisms has been reported to measure between 1 and 10 μm in field samples (Turner et al. 2014); therefore, the use of a 0.7-μm pore-size filter is reasonable for collection of eDNA from field samples.

Use of the finest filters possible is desirable; however, there could be a trade-off between the amount of water that can be filtered and the fineness of the filter. In the experiments using field samples, finer filters could recover more DNA from relatively clear water obtained from a pond. On the other hand, for turbid water samples, such as samples collected from a natural lake, it was impossible to filter a sufficient volume of water using finer filters because of clogging. Consequently, no significant difference in DNA recovery was observed among filter types. Thus, selection of the appropriate filter is dependent on the properties of the water sample of interest.

In experiment B, phenol treatment led to collection of more eDNA copies than ultrafiltration, and the CV was lower with phenol treatment. Additionally, from the results of field sample analysis, phenol treatment could recover more eDNA copies than ultrafiltration. These results indicated that phenol treatment was advantageous; however, phenol treatment has some disadvantages, including risks associated with phenol use and the longer amount of time required compared with ultrafiltration. Therefore, methods used to improve eDNA enrichment yield should be chosen depending on the specific context.

There are various restrictions associated with the use of eDNA techniques in the field. For example, electricity is often inaccessible in the field; hence, a method for preserving eDNA samples at ambient temperatures without using a freezer would dramatically expand the conditions under which eDNA techniques can be applied. Thus, the efficacy of DNA fixation using ethanol was examined because eDNA can be stably stored for long periods of time in ethanol (Hajibabaei et al. 2012). In experiment C, the number of eDNA copies collected on filters was dramatically decreased without ethanol treatment. The rate of eDNA decomposition appeared to be comparable to that observed in sample water (Dejean et al. 2011; Thomsen et al. 2012b). In contrast, the amount of eDNA retained on the filters with ethanol treatment did not change, even after 6 days. The lower CV associated with ethanol treatment compared to that of an untreated sample also supported this advantage (data not shown). This result indicated that ethanol treatment was useful for preservation of eDNA on filters. Recently, Renshaw et al. (2015) also reported a method for the preservation of filtered eDNA samples at room temperature; in this method, eDNA is preserved in DNA lysis buffer. The present study showed that eDNA could be preserved on dried filters by adding a small volume of ethanol after filtering the field water samples, and the filters could then be packed in a smaller space. Notably, the amount of ethanol required for this method is less than that used by the traditional ethanol precipitation method, in which twice the volume of ethanol is added to the water sample.

In tank experiments (experiments A, B, and C), we did not adopt blank controls for monitoring cross-contamination during filtration and DNA extraction processes; thus, we could not completely exclude the possibility of cross-contamination. However, we performed the experiments at the same facilities and applied the same precautions and procedures as in the previous study in which no cross-contamination was observed (Fukumoto et al. 2015). In addition, we employed blank controls for field surveys and found no contamination. Therefore, we believe that the results of tank experiments may not be substantially affected by cross-contamination.

In this study, we proposed context-specific selection of methods for eDNA analysis of aquatic vertebrates (Fig. 6). In aquarium experiments, the eDNA concentration is relatively high; thus, ethanol precipitation coupled with ultrafiltration or filtration using 0.7-µm glass fiber filters coupled with phenol treatment are optimal. These two methods differ with respect to risks associated with phenol use, time required for processing, and collection efficiency. In the field, samples should be filtered because large amounts of water are required to collect eDNA, which is present at relatively low concentrations. Moreover, the filter type should be chosen on the basis of the characteristics of the water of interest before starting the survey.
Fig. 6

The proposed context-specific method for the case-by-case selection of alternative methods for environmental DNA (eDNA) capture and extraction. Purified eDNA can be applied to various downstream applications, such as real-time PCR or DNA metabarcoding

Environmental DNA retained on filters can be preserved by ethanol or freezing, extending the scope of eDNA analysis. In particular, DNA fixation by ethanol is expected to have practical use in various fields because this method enables enrichment of eDNA without using any electricity and permits preservation and transportation at ambient temperatures. For example, our research team succeeded in detection of fish DNA from eDNA samples that were collected in tropical Asia, concentrated, preserved, and transported at ambient temperatures for about 1 week (Minamoto et al., unpublished). Thus, eDNA analysis could be applied to different situations using adaptive combinations of these techniques.

In conclusion, our results indicated that the eDNA recovery yield depended on the filter used and the postfiltration enrichment process. Additionally, the filter should be selected depending on the characteristics of the target water. Further studies are needed in the near future to identify patterns of eDNA concentrations associated with environmental factors, such as water quality parameters, and to elucidate optimal eDNA recovery methods, including filter selection and postfiltration eDNA enrichment.

Notes

Acknowledgments

The authors sincerely thank Dr. M. Kondoh, Dr. H. Yamanaka, and the students in the Maruyama laboratory at Ryukoku University for discussions and comments on this study. This study was conducted as a part of the MEXT GRENEei Ecohealth Project (Project Leader: Prof. Chiho Watanabe, the University of Tokyo, Japan). This study was partly funded by the “Environmental Change and Infection Diseases in Tropical Asia” project (R-04) of the Research Institute for Humanity and Nature led by KM and by grants from JSPS KAKENHI (grant numbers 24657020 to TM and AM, and 26440238 to TM).

References

  1. Brook BW, O’Grady JJ, Chapman AP, Burgman MA, Akçakaya HR, Frankham R (2000) Predictive accuracy of population viability analysis in conservation biology. Nature 404(6776):385–387. doi:10.1038/35006050 CrossRefPubMedGoogle Scholar
  2. Deiner K, Altermatt F (2014) Transport distance of invertebrate environmental DNA in a natural river. PLoS One 9(2):e88786. doi:10.1371/journal.pone.0088786 PubMedCentralCrossRefPubMedGoogle Scholar
  3. Deiner K, Walser J-C, Mächler E, Altermatt F (2015) Choice of capture and extraction methods affect detection of freshwater biodiversity from environmental DNA. Biol Conserv 183:53–63. doi:10.1016/j.biocon.2014.11.018 CrossRefGoogle Scholar
  4. Dejean T, Valentini A, Duparc A, Pellier-Cuit S, Pompanon F, Taberlet P, Miaud C (2011) Persistence of environmental DNA in freshwater ecosystems. PLoS One 6(8):e23398. doi:10.1371/journal.pone.0023398 PubMedCentralCrossRefPubMedGoogle Scholar
  5. Dejean T, Valentini A, Miquel C, Taberlet P, Bellemain E, Miaud C (2012) Improved detection of an alien invasive species through environmental DNA barcoding: the example of the American bullfrog Lithobates catesbeianus. J Appl Ecol 49(4):953–959. doi:10.1111/j.1365-2664.2012.02171.x CrossRefGoogle Scholar
  6. Doi H, Takahara T, Minamoto T, Matsuhashi S, Uchii K, Yamanaka H (2015a) Droplet digital polymerase chain reaction (PCR) outperforms real-time PCR in the detection of environmental DNA from an invasive fish species. Environ Sci Technol 49(9):5601–5608. doi:10.1021/acs.est.5b00253 CrossRefPubMedGoogle Scholar
  7. Doi H, Uchii K, Takahara T, Matsuhashi S, Yamanaka H, Minamoto T (2015b) Use of droplet digital PCR for estimation of fish abundance and biomass in environmental DNA surveys. PLoS One 10(3):e0122763. doi:10.1371/journal.pone.0122763 PubMedCentralCrossRefPubMedGoogle Scholar
  8. Dunnett CW (1964) New tables for multiple comparisons with a control. Biometrics 20(3):482–491CrossRefGoogle Scholar
  9. Eichmiller JJ, Bajer PG, Sorensen PW (2014) The relationship between the distribution of common carp and their environmental DNA in a small lake. PLoS One 9(11):e112611. doi:10.1371/journal.pone.0112611 PubMedCentralCrossRefPubMedGoogle Scholar
  10. Ficetola GF, Miaud C, Pompanon F, Taberlet P (2008) Species detection using environmental DNA from water samples. Biol Lett 4(4):423–425. doi:10.1098/Rsbl.2008.0118 PubMedCentralCrossRefPubMedGoogle Scholar
  11. Foote AD, Thomsen PF, Sveegaard S, Wahlberg M, Kielgast J, Kyhn LA, Salling AB, Galatius A, Orlando L, Gilbert MTP (2012) Investigating the potential use of environmental DNA (eDNA) for genetic monitoring of marine mammals. PLoS One 7(8):e41781. doi:10.1371/journal.pone.0041781 PubMedCentralCrossRefPubMedGoogle Scholar
  12. Fukumoto S, Ushimaru A, Minamoto T (2015) A basin-scale application of environmental DNA assessment for rare endemic species and closely related exotic species in rivers: a case study of giant salamanders in Japan. J Appl Ecol 52(2):358–365. doi:10.1111/1365-2664.12392 CrossRefGoogle Scholar
  13. Goldberg CS, Pilliod DS, Arkle RS, Waits LP (2011) Molecular detection of vertebrates in stream water: a demonstration using Rocky Mountain tailed frogs and Idaho giant salamanders. PLoS One 6(7):e22746. doi:10.1371/journal.pone.0022746 PubMedCentralCrossRefPubMedGoogle Scholar
  14. Hajibabaei M, Spall J, Shokralla S, van Konynenburg S (2012) Assessing biodiversity of a freshwater benthic macroinvertebrate community through non-destructive environmental barcoding of DNA from preservative ethanol. BMC Ecol 12(1):28. doi:10.1186/1472-6785-12-28 PubMedCentralCrossRefPubMedGoogle Scholar
  15. Honjo MN, Minamoto T, Matsui K, Uchii K, Yamanaka H, Suzuki AA, Kohmatsu Y, Iida T, Kawabata Z (2010) Quantification of cyprinid herpesvirus-3 in environmental water by using an external standard virus. Appl Environ Microb 76:161–168. doi:10.1128/AEM.02011-09 CrossRefGoogle Scholar
  16. Janosik AM, Johnston CE (2015) Environmental DNA as an effective tool for detection of imperiled fishes. Environ Biol Fishes. doi:10.1007/s10641-015-0405-5 Google Scholar
  17. Jerde CL, Mahon AR, Chadderton WL, Lodge DM (2011) “Sight-unseen” detection of rare aquatic species using environmental DNA. Conserv Lett 4(2):150–157. doi:10.1111/J.1755-263x.2010.00158.X CrossRefGoogle Scholar
  18. Krebs C (1978) Ecology: The Experimental analysis of distribution and abundance, 2nd Edn. Harper and Row, New YorkGoogle Scholar
  19. Lowe S, Browne M, Boudjelas S, DePoorter M (2000) 100 of the world’s worst invasive alien species: a selection from the global invasive species database. 1–12Google Scholar
  20. Lydolph MC, Jacobsen J, Arctander P, Gilbert MTP, Gilichinsky DA, Hansen AJ, Willerslev E, Lange L (2005) Beringian paleoecology inferred from permafrost-preserved fungal DNA. Appl Environ Microb 71(2):1012–1017. doi:10.1128/Aem.71.2.1012-1017.2005 CrossRefGoogle Scholar
  21. Maruyama A, Nakamura K, Ymanaka H, Kondoh M, Minamoto T (2014) The release rate of environmental DNA from juvenile and adult fish. PLoS One 9(12):e114639. doi:10.1371/journal.pone.0114639 PubMedCentralCrossRefPubMedGoogle Scholar
  22. Minamoto T, Honjo MN, Kawabata Z (2009) Seasonal distribution of cyprinid herpesvirus 3 in Lake Biwa. Appl Environ Microb 75:6900–6904. doi:10.1128/AEM.01411-09 CrossRefGoogle Scholar
  23. Minamoto T, Yamanaka H, Takahara T, Honjo MN, Kawabata Z (2012) Surveillance of fish species composition using environmental DNA. Limnology 13(2):193–197. doi:10.1007/S10201-011-0362-4 CrossRefGoogle Scholar
  24. Nathan LM, Simmons M, Wegleitner BJ, Jerde CL, Mahon AR (2014) Quantifying environmental DNA signals for aquatic invasive species across multiple detection platforms. Environ Sci Technol 48(21):12800–12806. doi:10.1021/es5034052 CrossRefPubMedGoogle Scholar
  25. Olson ZH, Briggler JT, Williams RN (2012) An eDNA approach to detect eastern hellbenders (Cryptobranchus a. alleganiensis) using samples of water. Wildlife Research 39(7):629–636. doi:10.1071/WR12114 CrossRefGoogle Scholar
  26. Pilliod DS, Goldberg CS, Arkle RS, Waits LP (2013) Estimating occupancy and abundance of stream amphibians using environmental DNA from filtered water samples. Can J Fish Aquat Sci 70(8):1123–1130. doi:10.1139/cjfas-2013-0047 CrossRefGoogle Scholar
  27. R Development Core Team (2011) R: a language and environment for statistical computing. R Foundation for Statistical Computing, ViennaGoogle Scholar
  28. Renshaw MA, Olds BP, Jerde CL, McVeigh MM, Lodge DM (2015) The room temperature preservation of filtered environmental DNA samples and assimilation into a phenol-chloroform-isoamyl alcohol DNA extraction. Mol Ecol Resour 15(1):168–176. doi:10.1111/1755-0998.12281 PubMedCentralCrossRefPubMedGoogle Scholar
  29. Sogin ML, Morrison HG, Huber JA, Welch DM, Huse SM, Neal PR, Arrieta JM, Herndl GJ (2006) Microbial diversity in the deep sea and the underexplored “rare biosphere”. Proc Natl Acad Sci USA 103(32):12115–12120. doi:10.1073/pnas.0605127103 PubMedCentralCrossRefPubMedGoogle Scholar
  30. Taberlet P, Coissac E, Hajibabaei M, Rieseberg LH (2012) Environmental DNA. Mol Ecol 21(8):1789–1793. doi:10.1111/J.1365-294x.2012.05542.X CrossRefPubMedGoogle Scholar
  31. Takahara T, Minamoto T, Doi H (2013) Using environmental DNA to estimate the distribution of an invasive fish species in ponds. PLoS One 8(2):e56584. doi:10.1371/journal.pone.0056584 PubMedCentralCrossRefPubMedGoogle Scholar
  32. Takahara T, Minamoto T, Yamanaka H, Doi H, Kawabata Z (2012) Estimation of fish biomass using environmental DNA. PLoS One 7(4):e35868. doi:10.1371/journal.pone.0035868 PubMedCentralCrossRefPubMedGoogle Scholar
  33. Takahara T, Minamoto T, Doi H (2015) Effects of sample processing on the detection rate of environmental DNA from the Common Carp (Cyprinus carpio). Biol Conserv 183:64–69. doi:10.1016/j.biocon.2014.11.014 CrossRefGoogle Scholar
  34. Thomsen PF, Kielgast J, Iversen LL, Moller PR, Rasmussen M, Willerslev E (2012a) Detection of a diverse marine fish fauna using environmental DNA from seawater samples. PLoS One 7(8):e41732. doi:10.1371/journal.pone.0041732 PubMedCentralCrossRefPubMedGoogle Scholar
  35. Thomsen PF, Kielgast JOS, Iversen LL, Wiuf C, Rasmussen M, Gilbert MTP, Orlando L, Willerslev E (2012b) Monitoring endangered freshwater biodiversity using environmental DNA. Mol Ecol 21(11):2565–2573CrossRefPubMedGoogle Scholar
  36. Treguier A, Paillisson J-M, Dejean T, Valentini A, Schlaepfer MA, Roussel J-M (2014) Environmental DNA surveillance for invertebrate species: advantages and technical limitations to detect invasive crayfish Procambarus clarkii in freshwater ponds. J Appl Ecol 51:871–879. doi:10.1111/1365-2664.12262 CrossRefGoogle Scholar
  37. Turner CR, Barnes MA, Xu CCY, Jones SE, Jerde CL, Lodge DM (2014) Particle size distribution and optimal capture of aqueous macrobial eDNA. Method Ecol Evol 5:676–684. doi:10.1111/2041-210X.12206 CrossRefGoogle Scholar
  38. Uchii K, Telschow A, Minamoto T, Yamanaka H, Honjo MN, Matsui K, Kawabata Z (2011) Transmission dynamics of an emerging infectious disease in wildlife through host reproductive cycles. The ISME Journal 5:244. doi:10.1038/ismej.2010.123 PubMedCentralCrossRefPubMedGoogle Scholar

Copyright information

© The Japanese Society of Limnology 2015

Authors and Affiliations

  • Toshifumi Minamoto
    • 1
    • 2
  • Takafumi Naka
    • 3
  • Kazuhiko Moji
    • 2
    • 4
  • Atsushi Maruyama
    • 3
  1. 1.Graduate School of Human Development and EnvironmentKobe UniversityKobeJapan
  2. 2.Research Institute for Humanity and NatureKyotoJapan
  3. 3.Faculty of Science and TechnologyRyukoku UniversityOtsuJapan
  4. 4.Graduate School of Tropical Medicine and Global HealthNagasaki UniversityNagasakiJapan

Personalised recommendations