Sample collection

We collected samples (tissues or feces) and extracted genomic DNA from 20 Japanese giant flying squirrels obtained from four breeding facilities and five field sites (Fig. 1). Tissue samples of four individuals were collected at Tama Forest Science Garden Hachioji and of one individual at Saitama children's Zoo Park (Fig. 1). Among them, three individuals collected at Tama Forest Science Garden were found dead at this site. The cause of death was presumed to be either predation or accidental collision. The entire specimens of two individuals and partial body (only skin) of one individual were stored at –80°C. Since one of these individuals was a partial body (K1), sex of this individual was not identifiable based on genital morphology. Sex of an adult male (K4) and an infant female (K3) were identified based on genital morphology. We note that the female sex of K3 should be interpreted with caution due to difficulty with morphological sex identification of infants. A fourth individual found at this location (K2) was also an infant and had fallen from a tree canopy but was kept alive. A tissue sample was obtained from this individual by swabbing the inside of its cheek with a cotton swab. The cotton swab was immediately soaked in Lysis buffer (0.1 M Tris-HCL, 0.1 M EDTA, 0.01 M NaCl, 0.5% SDS) and was stored at –30°C. Sex of this infant individual (K2) was not identifiable based on genital morphology. One female individual (S1) was collected at Saitama children's Zoo Park. A tissue sample was obtained from this individual by swabbing the inside of its cheek with a cotton swab.

Fig. 1.

Sample information. Samples were collected from four breeding facilities (1–4) and five field sites (5–9). Arabic numbers on the map correspond to the location numbers in the table.

Fecal samples of 16 individuals were collected from four breeding facilities and five field sites (Fig. 1). Fecal samples of five individuals (two males and three females) were collected from four breeding facilities (Chuo University Junior and Senior High School, Tama Zoological Park, Ueno Zoological Gardens, and Saitama Children's Zoo Park) (Fig. 1). All individuals were protected in a forest close to each facility when they were young. Fecal pellets defecated before night were collected from the floor of the breeding facility with disposable toothpicks or chop sticks. Samples were stored independently in 1.5 mL tubes with Lysis buffer (described above) and stored at –30°C. Because two individuals (one male and one female, T1 and T2 in Fig. 1) at Tama Zoological Park were kept in the same cage, it was unclear which feces belonged to which individual. Therefore, we identified defecating individuals by identifying the sex of the DNA samples extracted from each fecal pellet using sex identification markers. Feces collected from Tama Zoological Park were only used for testing of individual identification markers.

Fecal samples of 11 individuals were collected at five field sites (Tama Forest Science Garden, Mt. Takao, Mt. Tsukuba, Lake Tsukui Shiroyama Park and private forest site) from several different trees (Fig. 1). Each fecal sample was stored in separate 1.5 mL tubes containing Lysis buffer and stored at – 30°C. To separate unique individuals from unidentified individuals, we performed sex identification for all samples after development of sex identification markers. Then, we selected two fecal samples with different sex from each site and treated these selected samples as unique individuals. In addition, two to three feces collected from a sufficient distance (1.2 km) apart, which could be safely assumed to have been defecated by different individuals, were also treated as unique individuals. These feces collected from presumably separate individuals were only used for testing individual identification markers. The guidelines for experimental animal management of SOKENDAI were followed throughout the study. The Institutional Animal Care and Use Committee of SOKENDAI approved the animal protocols and procedures (permission #2017004ar and #2018005ar).

DNA extraction

From the stored body specimens, a piece of muscle tissue (∼5 mm square) was cut out with a scalpel and used for DNA extraction. Genomic DNA was extracted from tissue samples using a DNeasy Blood & Tissue Kit (QIAGEN, Hilden, Germany). Fecal DNA was extracted from a single fecal pellet sample using a Cica Geneus DNA Prep Kit for Stool (KANTO KAGAKU) following the manufacturer's instructions.

Development of DNA markers to identify individuals

The concentration of the genomic DNA extracted from tissue samples of the male (K4) and presumed female (K3) giant flying squirrels found at the Tama Forest Science Garden (Fig. 1) was measured using a Qubit Fluorometer 2.0 (Thermo Fisher Scientific). Sex of K3 was ambiguous because this individual was an infant. After development of a sex identification marker, this individual was confirmed to be female. The NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Bio Labs, Ipswich, MA, USA) was used to construct libraries from genomic DNA. Paired-end (2 × 250 bp for a female [K3] and 2 × 150 bp for a male [K4]) sequencing was performed on the Illumina HiSeq X or Hiseq 2500 platform. In total, 5 Gb (2 × 250 bp) and 15 Gb (2 × 150 bp) sequences were determined from each individual. For sequences from a female (K3), the average length of DNA fragments was prepared to be less than 500 bp for Paired-end 2 × 250 bp sequences by Covaris M220 (Covaris. Inc., Woburn, MA, USA). When the paired-reads were overlapped, the two reads were merged as one sequence by the merge reads algorithm implemented in CLC Genomic Workbench 9.0 ( https://www.qiagenbioinformatics.com/). The merged sequences were used as a database for BLASTN searches (Altschul et al., 1990). We searched for the sequences containing trinucleotide repeats using nine types of trinucleotide repeat (AATn, AGCn, AGTn, ATCn, ATTn, CATn, GATn, TAAn, TATn) sequence motifs as query sequences (see  Supplementary Table S2 (zs220045_TableS2.xlsx)). The sequence reads from a male (K4) were mapped to merged sequences (K3) using CLC Genomic Workbench. Reads showing high similarity (> 90% in > 90% of read length) were mapped to merged sequences to avoid mapping the low similarity reads. When the number of repeats was different between merged sequences and the mapped reads, the sequence was selected as a candidate for development of a marker region. We selected 12 candidate sequences (accession numbers: LC726459–LC726470) and designed two pairs of primers on the flanking regions of the repeats in each candidate sequence. The sequences of primers are listed in  Supplementary Table S2 (zs220045_TableS2.xlsx).

Development of DNA markers to identify sex

SRY and XIST regions were searched from the sequences of the male (paired end 150 bp) and the female (paired end 250 bp) Japanese giant flying squirrels, using homologs of SRY (Gene ID 21674) and XIST (Gene ID 213742) from the house mouse (Mus musculus) as queries. To perform nested PCR, two sets of primers were designed on SRY and XIST sequences (accession numbers: LC726471, LC726472), respectively. The sequences of primers are listed in  Supplementary Table S1 (zs220045_TableS1.xlsx).

Test of DNA markers

Since the amounts of Japanese giant flying squirrel DNA extracted from fecal samples were expected to be low, we used nested PCR to obtain a sufficient amount of PCR products. Furthermore, for efficient amplification, nested PCR was performed independently for each locus. Nested PCR reactions were performed with final volume of 6 µl with concentration of KOD Fx Neo (TOYOBO, Osaka, Japan) (0.12 U), 1× KOD Fx Neo buffer, dNTPs (0.4 mM), and 0.2 µM forward and reverse primer, distilled water and template DNA to amplify SRY, XIST, and microsatellite regions. For template DNA, we used 1 µl of DNA (0.05∼376 ng) extracted from tissues or feces for the first round of nested PCR. For the second round of nested PCR, PCR products after the first round were diluted 100 times with water and 1 µl was used as a template. For the first and second round of nested PCR, different sets of primers were used (see  Supplementary Table S1 (zs220045_TableS1.xlsx) and  Supplementary Table S2 (zs220045_TableS2.xlsx): we used primers “F1” and “R1” for the first round of nested PCR and “F2” and “R2” for the second round of PCR. Primers without numbers (SRY_R and m6_R) were used for the first and the second rounds of PCR). The conditions of PCR were as follows: 94°C for 2 min of denaturation step, followed by 35 cycles of 98°C for 10 seconds of denaturation step, 58 – 62°C for 30 seconds of annealing step (listed in  Table S1 (zs220045_TableS1.xlsx) and  S2 (zs220045_TableS2.xlsx) for each locus), 68°C for 30 seconds of extension step, and final extension at 68°C for 1 minute. After the nested PCR, 3 µl of PCR product were electrophoresed on 2% agarose gels, with a size marker (φX174/HaeⅢ, Takara, Shiga, Japan). Samples were stained with DNA Fluorescent Staining Dye (SMOBIO, Hsinchu City, Taiwan).

Sex was identified by electrophoresis of PCR products. When both SRY and XIST regions were amplified, sex was assigned as male. When only XIST region was amplified, sex was assigned as female. To verify the sex identification, amplification of XIST and SRY regions was performed two or more times.

For microsatellite regions, we performed an additional round of PCR using second round primers tagged with a fluorescent label. The labeled PCR products were purified, and the fragment length was analyzed using a 3500 Genetic Analyzer and the Gene Mapper software 6 (Applied Biosystems). For efficiency, when we analyzed the fragment length, one to four PCR products with different color labels or different length ranges were mixed. To verify the results of amplification, PCR for each microsatellite region was performed two to seven times. We only genotyped the sample if a clear peak was observed two or more times by fragment analysis (Taberlet et al., 1996; Sefc et al., 2003). When peak observation was less than two times, the sample-specific locus was treated as failed amplification. Also, when we observed more than two peaks at a locus, the sample-specific locus was treated as a failed allele. We therefore excluded loci with either failed amplifications or failed alleles from the data analyses. Contamination was checked by negative controls for all PCR reactions.

Data analysis

To measure the discrimination efficiency of each individual identification marker, P_(ID) and P_(ID)sib values were calculated. P_(ID) is the probability that two individuals chosen from a random mating population have the same alleles and P_(ID)sib is the probability that two siblings chosen from a random mating population have the same alleles.

p_i and p_j are ith and jth allele frequencies. i is not equal to j. (Paetkau and Strobeck, 1994)

Cumulative P_(ID) and total P_(ID)sib were calculated by multiplying P_(ID) or P_(ID)sib values from the lowest P_(ID) or P_(ID)sib loci to obtain the discrimination efficiency when multiple markers were analyzed. Deviation from Hardy–Weinberg equilibrium was tested for each locus using the software Genepop (ver 4.7.5.) (Raymond and Rousset, 1995).

Availability of data

The nucleotide sequences were deposited in the DDBJ Sequenced Read Archive under accession numbers DRX346687 and DRX346688.

Development of sex identification markers

We isolated the sequences of the SRY and XIST genes and designed two sets of primers based on the sequences of these genes. We used these primers as sex identification markers. To assess the sex identification markers, we used each of two individuals of males and females of the Japanese giant flying squirrel (G1, U1, S1, K4: Fig. 1). DNAs extracted from two samples each of tissue and feces were used to amplify XIST and SRY regions. We predicted that the XIST region on the X chromosome is amplified from both males and females, whereas the SRY region on the Y chromosome is amplified from only males. Indeed, the XIST region (258 bp) was amplified from both males and females, while the SRY region (103 bp) was amplified only from males (Fig. 2). Therefore, sex of Japanese giant flying squirrel individuals could be identified using DNA markers on the X and Y chromosomes. Furthermore, both regions were amplified from DNA extracted from even a single fecal pellet, suggesting that DNA extracted from a fecal sample is sufficient to non-invasively identify sex.

Fig. 2.

Amplification of XIST and SRY genes from males and females of Japanese giant flying squirrels. Arrows indicate the expected length of XIST (258 bp) and SRY (103 bp) genes. Amplification from DNA extracted from tissue and feces are shown as (tissue) and (feces), respectively.

Individual identification markers

We isolated 12 sequences containing microsatellites with different numbers of repeats between two individuals, male (K4) and female (K3), and designed primers to amplify these sequences. To assess the efficiency of individual identification, DNA fragments were amplified from each of 12 microsatellite loci (see  Supplementary Table S2 (zs220045_TableS2.xlsx)) using DNA from 20 individuals (Fig. 1). Two to 10 DNA fragments with different lengths were amplified from each of 12 loci (Table 1). The frequencies of the DNA fragments with different lengths did not deviate from Hardy-Weinberg equilibrium (P > 0.05) for all loci, indicating that the DNA fragments were the alleles of each locus. Values of P_(ID) and P_(ID)sib were lowest on the 12th microsatellite locus (0.05 and 0.34) and highest on the 6th microsatellite locus (0.46 and 0.68) (Table 1). Cumulative P_(ID) and Cumulative P_(ID)sib values were calculated to obtain the probability of individual discrimination when multiple markers were used (Schwartz and Monfort, 2008). When three or more microsatellite markers were selected from the loci with lowest P_(ID) values, Cumulative P_(ID) became lower than 0.001 (Waits et al., 2001). Similarly, when four or more microsatellite markers selected from the loci with lowest P_(ID) value, Cumulative P_(ID)sib became lower than 0.05 (Woods et al., 1999). As shown in Table 2, PCR results showed that six individuals at seven loci were unreliable due to low amplification rate (labeled “-”) and eight individuals at eight loci were unreliable due to genotyping errors (labeled “FA”). We excluded these unreliable PCR results from our analysis.

Table 1.

Alleles, P(ID) and P(IS)sib values of 12 microsatellite loci.

Table 2.

Amplification patterns of 12 microsatellite loci for each sample.

Even after excluding the unreliable PCR results, none of the PCR products amplified from the 12 microsatellite loci of the 20 individuals showed the same amplification pattern as the others. Thus, all 20 individuals from five different field sites and four breeding facilities (Fig. 1) were discriminated by newly developed microsatellite markers (Table 2). These identifications were successfully analyzed from DNA extracted from a single fecal pellet.