Access the full text.
Sign up today, get DeepDyve free for 14 days.
Loading next page...
/lp/de-gruyter/the-applicability-of-dna-barcoding-for-dietary-analysis-of-sika-deer-4XG7xkB7uI
References
References for this paper are not available at this time. We will be adding them shortly, thank you for your patience.
- Publisher
- de Gruyter
- Copyright
- Copyright © 2015 by the
- ISSN
- 2299-1077
- eISSN
- 2299-1077
- DOI
- 10.1515/dna-2015-0021
- Publisher site
- See Article on Publisher Site
Abstract
In Japan, overgrazing by sika deer (Cervus nippon) has been suggested to cause a decline in forest understory vegetation. DNA barcoding has become an accepted method for analyzing the diets of animals and may be useful for evaluating the impact of sika deer on vegetation. However, the applicability of DNA barcoding in the dietary analysis of sika deer, particularly whether all of the food plants can be detected with sufficient taxonomic resolution and whether the results can be evaluated quantitatively, has not been investigated. We conducted a feeding trial by feeding five plant species to a captive sika deer and sequenced the chloroplast trnL P6 loop region from the sika deer's fecal DNA using the Ion PGM sequencer. We detected the sequences of all of the food plants at the species level using the local (selfproduced) database and at the genus or family level with the global database. Although the sequences of some major food plants were detected with high frequency, the proportion of consumed food plants did not match the proportion of sequences obtained from fecal DNA. With further technical advances and the further completeness of the sequence database for vegetation, DNA barcoding will be a useful tool for the dietary study of sika deer. Keywords: Cervus nippon, feeding trial, high-throughput sequencing, trnL P6 loop *Corresponding author: Haruko Ando, Center for Environmental Biology and Ecosystem Studies, National Institute for Environmental Studies, Tsukuba 305-8506, Japan, E-mail: ando.haruko@nies.go.jp Fumiko Nakahara, Hideyuki Ito, Asako Murakami, Naoki Morimoto, Kyoto City Zoo, Kyoto 606-8333, Japan Michimasa Yamasaki, Atsushi Takayanagi, Yuji Isagi, Laboratory of Forest Biology, Division of Forest an Biomaterials Science, Graduate School of Agriculture, Kyoto University, Kyoto 606-8502, Japan © 2015 Fumiko Nakahara et al. licensee De Gruyter Open. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivs 3.0 License. species. However, the usefulness of DNA barcoding in studying the diets of sika deer has not been researched. The objectives of this study were to evaluate the applicability of DNA barcoding for sika deer dietary analysis in two aspects: 1) taxonomic identification using a local (self-produced) and global database and 2) quantitative evaluation of diet composition by the sequence proportion. We conducted next-generation sequencing of the chloroplast trnL P6 loop region for the fecal samples obtained after feeding trials to sika deer. 2 Materials and methods The feeding trial and fecal sampling were performed in the Kyoto City Zoo Wildlife Rescue Center. In this center, three fawns were reared in spring 2013. One of the three was caged separately from the other fawns. Although the age of the fawn was not certain, it had just started to eat leaves in August 2013. The feeding trial was performed in the morning for 7 days between August 17 and August 23, 2013. We fed five food plants to the fawn in different proportions: Quercus glauca, Prunus yedoensis, Acer palmatum, Pueraria lobata, and Microstegium vimineum. These food species were selected and fed according to the feeding plan of the zoo other than M. vimineum, which is one of the favorite grass species of sika deer. The amount of plants consumed by the fawn was estimated on a dry weight basis. On the first day of the feeding trial, 5 to 24 g of each food plant species was dried at 80 °C for 48 hours to obtain their dry weights. The water content of each plant species was then calculated by the difference between their fresh weight and dry weight. On each day of the feeding trial, the fresh weight of food plants given to the fawn was measured before feeding, and the dry weight of food plants remaining after the feeding (not consumed by the sika deer) was measured after 48 hours of desiccation. The dry weight of food plants consumed by the sika deer was calculated as the difference between the dry weight of food before consumption, which was estimated by the fresh weight of food and its water content, and the dry weight of food after consumption. Before feeding, we removed any remaining food plants that were fed the previous day. We prepared a trnL reference database of five food plants and another 17 plant species that are preferred by sika deer for DNA barcoding (Table 1). These 22 species consist of 21 genera and 13 families, including two closely related species (Quercus crispula and Quercus glauca) that belong to the genus Quercus. We extracted DNA from leaves, and the universal primer pair c-d [16] was used for the PCR amplification of whole chloroplast trnL (UAA) introns. PCR amplification was conducted using a Qiagen Multiplex PCR kit (Qiagen, The Netherlands). The amplification mixture contained 20 ng of extracted DNA, 12 l of 2× Multiplex PCR Master Mix, and 0.2 mol/L of each primer pair. The mixture was denatured at 94 °C for 15 min, followed by 35 cycles of 94 °C for 1 min, 57 °C for 1 min, 72 °C for 1 min, and a final cycle of 72 °C for 4 min. Sequencing was performed with a Big Dye Terminator v1.1 Cycle Sequencing Kit (Applied Biosystems, USA) according to the standard protocol. The sequences were visualized by the ABI PRISM 3100 Genetic Analyzer (Applied Biosystems). These sequences were deposited in the DDBJ nucleotide data bank (Accession no. LC057213 LC057234). Fecal sampling was performed for 9 days between August 18 and August 26, 2013. The digestion time of sika deer was estimated to be 20 to 40 hours [9]; therefore, we decided to conduct the feeding trial for more than 2 days. We collected 70 to 100 feces per day and bagged 10 feces from each fecal mass. Ten feces were considered one sample. The collected feces were stored at -30 °C. Each fecal sample was homogenized after freezing by liquid nitrogen. The food plant DNA was extracted from each of the 20 mg fecal samples in dry weight using the DNeasy Plant Mini Kit (QIAGEN) following the manufacturer's instructions. The DNA extracts were recovered in a total volume of 100 L. The amplification mixture contained 20 ng of extracted DNA, 15 l of Multiplex PCR Master Mix, and 0.2 mol/L of each primer pair. Thermal cycling conditions were 95 °C for 15 min, followed by 40 cycles of 94 °C for 30 s, 57 °C for 90 s, 72 °C for 90 s, and a final cycle of 72 °C for 10 min. We checked for the presence of a PCR product of suitable length by electrophoresis on a 1.5% agarose gel. The universal primer pair g and h [17], a part of the trnL c-d region, was used to amplify the trnL P6 loop. The forward primer was tagged with a multiple identifier (MID, Roche Diagnostics, Switzerland) to identify the resulting sequences from each sample. The PCR products were purified using exo/SAP (exonuclease I and shrimp alkaline phosphatase, Takara, Japan and Promega, USA) and a High Pure PCR Products Purification Kit (Roche Diagnostics) and then quantified using a QubitR 2.0 Fluorometer (Life Technologies, USA). The PCR products were mixed such that approximately the same number of molecules from each fecal sample was included in each mix. We used the Ion OneTouchTM System (Life Technologies) to prepare amplicons for sequencing following the appropriate guide protocol. The run was performed using the Ion OneTouchTM 200 Template Kit v2. The resultant enriched Ion SphereTM particles were F. Nakahara, et al. Table 1. List of plant species contained in the database using a local BLAST search. Family Aceraceae Caprifoliaceae Clethraceae Cupressaceae Genus Acer Weigela Clethra Chamaecyparis Cryptomeria Fabaceae Fagaceae Pueraria Quercus Species Acer palmatum Weigela hortensis Clethra barbinervis Chamaecyparis obtusa Cryptomeria japonica Pueraria lobata Quercus crispula Quercus glauca Garryaceae Lauraceae Aucuba Cinnamomum Lindera Liliaceae Pentaphylacaceae Poaceae Hemerocallis Eurya Eleusine Lolium Microstegium Miscanthus Sasa Setaria Polygonaceae Fallopia Polygonum Rosaceae Prunus Aucuba japonica Cinnamomum camphora Lindera umbellata Hemerocallis longituba Eurya japonica Eleusine indica Lolium multiflorum Microstegium vimineum Miscanthus sinensis Sasa kurilensis Setaria viridis Fallopia japonica Polygonum thunbergii Prunus yedoensis Food plant loaded onto 314 Ion semiconductor sequencing chips, and sequencing was carried out on the Ion PGM sequencer (Life Technologies). The software Claident [18] was used to separate the sequences into each sample by MID tags and to exclude short sequences (< 50 bp) and sequences with low quality (mean quality value < 27). The DNA barcoding was carried out with two patterns. One was a local BLAST in BioEdit [19] using a self-produced database. The other pattern was Blast-2-GO [20] using an NCBI database. We excluded sequences with low e-values (< 1.0e-25), as outlined previously [21]. After DNA barcoding, we calculated the frequency of sequence reads: the number of sequences that were assigned to a certain species, genus, or family per the number of total sequences. For quantitative evaluation, we compared the frequency of sequence reads and the proportion of the plant food dry weight. The detected food plants by DNA barcoding that held fewer than 5 sequences in each sample were not included in the analyses to avoid the misidentification of plants due to sequencing errors. The sff file of the fecal sequences was deposited in the DDBJ nucleotide data bank (Accession no. DRR036723). 3 Results The fawn ate the food plants in variable proportions (Figure 1). The estimated dry weight of the consumed plants per day ranged from 0.19 g to 301.97 g (Figure 1a). Quercus glauca was the most consumed food plant on 6 of the 7 days in the feeding trial because this plant was given in the largest amount and because sika deer have a preference for Quercus. This species and two others (Prunus yedoensis and Acer palmatum) comprised 90% of the consumed food plants (Figure 1b). The sequencing of 81 fecal samples yielded 115,996 reads. In the local BLAST, most of the reads were assigned to food plant species (Figure 2). The sequences of plants that were not fed to the fawn were detected with low frequency. In contrast, in Blast-2-GO, most of the reads, with the exception of Acer palmatum, could not be assigned to food plants at the species level when we used only the one best-hit result (Figure 3a). At higher levels of taxonomy, Microstegium vimineum Pueraria lobata Acer palmatum Prunus yedoensis Quercus glauca (a) (b) Dry weight (g) Figure 1: (a) The dry weight of each of the five food plant species that a fawn consumed on each day of the feeding trial. (b) The of each of the five food plant species that the fawn consumed on each day of the feeding trial. Eurya japonica Aucuba japonica Weigela horetensis Quercus crispula Pueraria lobata Acer palmatum Prunus yedoensis Quercus glauca 100 Figure 2: The of sequences allocated to each of the eight plant species using a local BLAST search with a self-produced sequence database. Colored species indicate food plants that the fawn actually consumed. Species in gray show food plants that were not fed to the fawn. The data are shown separately for 9 days, from 2 to 10 days after the start of the feeding trial. F. Nakahara, et al. approximately 80% to 90% of the reads could be assigned to the genus and the family of food plants (Figure 3b, 3c). When the data from all days of the feeding trial were Unknown Ulmus pumila Hyperbaena domingensis Acer wuyuanense Acer sieboldianum Acer saccharum Acer pseudosieboldianum Acer opalus Acer japonicum Acer erianthum Acer circinatum Acer campestre Acer palmatum Prunus undulata Prunus spinosa Prunus ovalis Prunus mahaleb Trigonobalanus doichangensis Lithocarpus fenestratus Castanopsis sieboldii Castanopsis longzhouica Castanopsis echinocarpa Castanopsis chinensis Castanopsis carlesii Fossil Castanea Castanea sativa Fossil Quercus Quercus suber Quercus pubescens Quercus phillyraeoides Quercus ilex Quercus gilva Quercus cerris pooled, the proportion of consumed food plants estimated by their dry weights did not match the proportion of reads at each level of classification (Figure 4). Unknown Ulmaceae Menispermaceae Aceraceae Rosaceae Fagaceae (a) (b) Unknown Ulmus Hyperbaena Acer Prunus Trigonobalanus Lithocarpus Castanopsis Castanea Quercus 100 (c) Figure 3: The of sequences identified at the (a) species level, (b) genus level, and (c) family level. The data are shown separately for 9 days, from 2 to 10 days after the start of the feeding trial. These sequences were analyzed by Blast-2-GO using the NCBI database. Species Unknown Ulmus pumila Hyperbaena domingensis Acer wuyuanense Acer sieboldianum Acer saccharum Acer pseudosieboldianum Acer opalus Acer japonicum Acer erianthum Acer circinatum Acer campestre Acer palmatum Prunus undulata Prunus spinosa Prunus ovalis Prunus mahaleb Trigonobalanus doichangensis Lithocarpus fenestratus Castanopsis sieboldii Castanopsis longzhouica Castanopsis echinocarpa Castanopsis chinensis Castanopsis carlesii Fossil Castanea Castanea sativa Fossil Quercus Quercus suber Quercus pubescens Quercus phillyraeoides Quercus ilex Quercus gilva Quercus cerris Food plant Microstegium vimineum Pueraria lobata Acer palmatum Prunus yedoensis Quercus glauca Genus Unkno wn Ulmus Hyperbaena Acer Prunus Trigonobalanus Lithocarpus Castanopsis Castanea Quercus Family Unkno wn Ulmaceae Menispermaceae Aceraceae Rosaceae Fagaceae Food plant Species Genus Family Figure 4: The of food plants consumed by the fawn during the 7 days of the feeding trial and the of sequences identified at the species, genus, and family levels using the Blast-2-Go program. The sequence data were pooled for all feces collected during 9 days of sampling. 4 Discussion In the taxonomic identification of food plants, this study indicated the usefulness of a local database. With the local BLAST, we identified almost 100% of food plant species using a database that included five food plant species and 17 other plant sequences. With Blast-2-GO, we could not identify the food plants at the species level (Figure 3a) but could only do so at the genus or family level (Figure 3b, 3c). The low resolution of Blast-2-GO might be caused by the incompleteness of the NCBI database or the sharing of a sequence between relative species. To identify food plants at lower taxonomic levels, we recommend making a local database of plants that might occur at a research site. However, a local database of the P6 loop may be limited in its taxonomic discrimination when a larger number of plant species is included. Although the self-produced database in this study discriminated between closely related species of the genus Quercus, the discrimination power of the P6 loop may vary among families. For example, the species discrimination rate of the P6 loop database, which consists of 222 plant species that are present on Ogasawara Island in Japan, ranged from 14% to 100% [21]. For a detailed identification of food plants of sika deer at research sites, a combination of genus- or family-specific primers may complement the limitation of the P6 loop. Quantitative evaluation of food composition by sequence proportion is one of the major issues to be considered in diet analysis based on DNA barcoding. In this study, the sequences of several plants that the fawn ate in a large amount were detected at a high frequency from fecal DNA (Figure 4). This result may indicate that the sequence proportion of the P6 loop may be used to detect major food resources for sika deer. However, we obtained the plant sequences that we did not feed to the fawn with low frequency, though we conducted the feeding trial in a cage without any plants. The incorrect sequences came from sequencing errors; thus, the sequences obtained with low frequency may be of low reliability for food detection. Although the number of obtained sequences may reflect the amount of food eaten by sika deer to some extent, the absolute quantification of the diet seems to be inappropriate. Assuming that the digestion time of sika deer is 20 to 40 hours [9], the proportion of plant tissues in feces should correspond to the proportion of the food plant given to a sika deer on a previous day. However, the proportion of sequences obtained from the feces from each day (Figure 2) did not correspond to the proportion of food plants given to the sika deer on the previous day (Figure 1). One of the major possible causes for the differences between the proportion of sequences and food plants is the bias of amplification efficiency in PCR towards different species and loci [22], which may be caused by, e.g., digestive processes [23], the amount of DNA per gram of tissue [24], GC content [25], DNA secondary structure [26], or previous sample treatment [27]. To increase the reliability of the quantitative evaluation of food composition by DNA barcoding, further technical advances in molecular experiments and data analyses may be required to correct the above biases. A feeding trial conducted in this study showed that DNA barcoding could detect the sequences of plants that sika deer eat, thus indicating the applicability of this method to diet analysis of sika deer in the wild. This study also indicated the effectiveness of making a local database for the detailed identification of food plants and to validate and limit the quantitative evaluation of food composition. Although some technical problems exist, diet analysis based on DNA barcoding may contribute to a better understanding of the impact of sika deer on forest vegetation. Acknowledgments: We would like to thank Shingo Kaneko and Ayako Izuno who supported the development of methods and assisted in experiments. We thank the members of Laboratory of Forest Biology, Graduate School of Agriculture, Kyoto University for valuable comments. Conflict of interest: Authors declare nothing to disclose.
Journal
DNA Barcodes – de Gruyter
Published: Jan 1, 2015