Sika deer presence affects the host–parasite interface of a Japanese land leech

Abstract Since the 1990s, increasing populations of a blood feeding land leech (Haemadipsa japonica) have become a serious issue in several Japanese prefectures, and it may be caused by the increases in sika deer (Cervus nippon) populations seen over the last quarter of the century. Therefore, this study aimed to reveal the host animal species of H. japonica using iDNA (vertebrate DNA isolated from invertebrates) and to test the hypothesis that the increasingly widespread distribution of sika deer results in increased H. japonica populations through changes to the host–parasite interface. We amplified mitochondrial DNA 16S ribosome RNA fragments from iDNA isolated from the blood clots of H. japonica collected across Japan. We identified 17 host animal species, including four orders of Mammalia (Carnivora, Artiodactyla, Rodentia, and Lagomorpha) and two orders of Amphibia (Caudata and Anura). The sika deer was the dominant host species of H. japonica. Additionally, the host animal species composition of H. japonica differed according to the presence or absence of sika deer. In the sites where sika deer were not found, Anura (frog) species were the most commonly identified hosts of H. japonica. These results suggest that the increases in H. japonica populations might have occurred via a change in host preference to sika deer. This change might be driven by the increases in sika deer populations and subsequent increase in the frequency that H. japonica uses the sika deer as easy prey, as well as by sika deer providing more reproductive energy per blood meal than blood meal from frog species. The present study suggests that a more widespread distribution of sika deer resulted in an increase in H. japonica through a change in the host–parasite interface. Therefore, management that focuses on decreasing sika deer populations would likely be an effective method for the reduction of H. japonica populations.

Thus, leech iDNA has been used to detect biodiversity and to create censuses of wild animals through identifying blood meals from a wide range of mammalian orders and by documenting their occurrence in target areas (Abrams et al., 2019;Schnell et al., 2012Schnell et al., , 2015Tessler et al., 2018), thereby contributing to the ecological and natural historical knowledge of various regions. Moreover, the iDNA from the blood of mosquitoes, flies, midges, and ticks has been used to identify the host species of target sanguivorous species, providing basic knowledge for infection control. However, few studies have been conducted into the applications of iDNA in forest ecology or wildlife management (e.g., Kent, 2009).
Haemadipsa japonica Whitman (Haemadipsidae) is a sanguivorous terrestrial leech species endemic to East Asia. The species has been documented in Honshu, Shikoku, Kyushu, and Yakushima islands in the Japanese archipelago (Borda & Siddall, 2010;Morishima & Aizawa, 2019;Nakano, 2017;Whitman, 1886). H. japonica was once restricted to mountainous regions ; however, since the 1990s, the distribution of H. japonica has expanded to areas influenced by human activities, including residential areas, which has become serious issues in several Japanese prefectures (Asada, Ochiai, & Yamanaka, 1995;Sugiyama & Sakaniwa, 2010). For example, the blood-sucking damage caused by H. japonica causes mental stress to forest workers and increases the cost of administering pest-control agents against this leech species (Morishima, Hayashi, & Aizawa, 2018).
In particular, since around the 1990s, the populations of sika deer have substantially increased; presently, the increased populations of sika deer strongly impact forest ecosystems and result in damages to agriculture and forestry (Noguchi, 2017;Takatsuki, 2009). These nearly concomitant increases in land leech and mammalian populations imply that the increases in H. japonica could be due to the increases in mammal populations, particularly sika deer (Sugiyama & Sakaniwa, 2010).
Previous studies have identified the host species of H. japonica with the aim of understanding the causes of H. japonica population increases using iDNA with polymerase chain reaction single-stand conformation polymorphisms (PCR-SSCP; Sasaki, Saito, & Harada, 2005;Sasaki & Tani, 2008;Kanagawa Prefecture, 2009), hybridization probes (Nakanojo Town, 2004, and immunohistological analyses (Yoshiba & Abe, 1989). However, results using these techniques are biased because, prior to the identification of host species, the length of the DNA fragment (i.e., electrophoresis patterns of DNA fragment bands) or the patterns of antigen-antibody reactions must be known for candidate host animals. Therefore if H. japonica feeds on a host animal that was not already identified as a candidate, the host would not be detected. iDNA identification based on nucleotide sequences such as current iDNA studies (e.g., Schnell et al., 2015;Schnell et al., 2012;Tessler et al., 2018) overcomes this bias because iDNA identification was conducted using vast nucleotide DNA database of multiple vertebrates. Therefore, host animal identification using nucleotide sequences of iDNA allows us to understand the causes of the current increases in H. japonica populations through the host-parasite interface of the land leech.
In this study, we aimed to reveal the host species of H. japonica in each site across Japan. We also aimed to test the hypothesis that the increasingly widespread distribution of sika deer resulted in the increases in H. japonica populations through a change in the host-parasite interface by comparing the host animals between sites where sika deer are present versus not present. go.jp/kiso/fnd_list.html). Land leech collection was performed by walking along the forest path to attract leeches and by gathering the leeches attached to the legs or other body parts of the researchers.

| Sample collection and DNA isolation
The collected land leeches were stored in 99.9% ethanol until iDNA isolation. The mid-bodies of the 826 collected leeches were dissected from the anterior to the posterior, and undigested blood clots in the digestive system were visually inspected (i.e., mainly crop and crop ceca). For the 216 leech samples with undigested blood clots, iDNA was isolated from the blood clots extracted from each sample (Appendix S1). Total genomic iDNA was isolated using a DNeasy Blood & Tissue Kit (Qiagen).

| mtDNA sequencing
Mitochondrial DNA 16S ribosome RNA (rRNA) fragments were amplified from the iDNA isolated from 216 H. japonica individuals using the SCPH02500, 5ʹ-TTACCAAAACATCACCTCT-3ʹ and SCPL02981, 5ʹ-ATCCAACATCGAGGTCGTAA-3ʹ primers (Matsui, Rakotondraparany, Hasegawa, & Horai, 2007). Polymerase chain reactions (PCR) were performed in 15 µl reaction volumes, containing 10 ng genomic DNA, 1× PCR buffer, 0.2 mM of each dNTPs, 1.5 mM MgCl 2 , 0.2 µM of each primer, and 0.5 U GoTaq polymerase (Promega). PCR amplifications were performed as follows: an initial denaturation at 94°C for 10 min, followed by 45 cycles of 10 s at 94°C, 30 s at the annealing temperature of 51°C, 1 min at 72°C, and a final extension at 72°C for 10 min using a GeneAmp2720 PCR System (Applied Biosystems, PE Corp.). The PCR products were electrophoretically separated on a 2.0% agarose gel and visualized with ethidium bromide in 1× TAE; all products exhibiting a single DNA fragment were sequenced. Negative controls were used in PCR in our preliminary experiment, confirming no amplification.
The products were purified using an ExoSAP-IT (Affymetrix).
Direct sequencing in both directions was conducted using the ABI PRISM BigDye Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems) on an ABI 3500 Genetic Analyzer. DNA sequences were subjected to visual inspection and were aligned using BIOEDIT 7.2.5.0 (Hall, 1999). The sequences obtained from leech iDNA were identified using the DNA Data Bank of Japan (DDBJ), Nucleotide BLAST Database, and blastn (http://blast.ddbj. nig.ac.jp/). A threshold of more than 98% identity was used. The leech iDNA that was identified as human was excluded from all analyses to remove the possibility of contamination through the experiment.
It was impossible to identify three salamander samples from site 15 to species level using the aforementioned primers. Therefore, we further sequenced mitochondrial cytochrome b (cyt b) fragments for the these three salamander samples using the following primers: L14010,  of 55°C, 1 min at 72°C, and a final extension at 72°C for 5 min. We also sequenced PCR products by direct sequencing.

| Data analyses
In the identification of host animal species, we treated Rana tagoi and R. sakuraii as a single species (called Rana tagoi/sakuraii hereafter) because they are not distinguished from each other by mitochondrial DNA because they showed nonmonophyly due to incomplete lineage sorting (Eto & Matsui, 2014). The differences in H. japonica host animal species composition among sites was examined using nonmetric multidimensional scaling (NMDS) and a Jaccard distance matrix based on the presence and absence of species between sites. A nonhierarchical clustering method, k-means clustering, was performed to determine groups of sites. Indicator species analysis (Dufrêne & Legendre, 1997) was carried out to identify indicative species within the groups of sites determined by k-means clustering. Multi-response permutation procedures (MRPP) and analyses of similarities (ANOSIM) were used to test whether the host animal species composition differed according to sika deer presence. These analyses were performed using the "vegan" (Oksanen et al., 2017) and "labdsv" (Roberts, 2016) R packages in R 3. 6. 0 (R Core Team, 2019) for 14 sites of the 26 sites; 12 (sites 2, 8, 11, 13, 18-21, and 23-26) were excluded from these analyses due to a small sample size (N < 5; Appendix S1). MRPP and ANOSIM were performed between sites with sika deer (sites 5-7, 10, 12, 14-17, and 22) and sites without sika deer (sites 1, 3, 4, and 9).
This study examined the host-parasite interface and whether the host animal species of H. japonica differed according to sika deer distribution using generalized linear mixed models (GLMMs). In the GLMMs, the numbers of H. japonica individuals that fed on wild boar (Nss), Japanese serow (Ncc), Carnivora (carnivore; Nca), and Anura (frog; Nan) were used as response variables. The presence/absence of sika deer (0 = absence; 1 = presence; Dcn) and the altitude of each site (Alt) were used as explanatory variables and fixed effects.
Sites identity was set as a random effect. The correlation between the two explanatory variables (Dcn and Alt) was assessed using a Spearman rank test in R. The details of the GLMMs model are shown in Table 1. The GLMMs were performed for 24 sites using the "glm-mTMB" (Magnusson et al., 2020) package in R. Sites 17 and 18 were excluded from the GLMMs due to a lack of altitude information (Appendix S1). Predicted values and Wald 95% confidence interval (CI) were obtained using R.

| Sequencing of mtDNA from iDNA of H. japonica
Of

| The differences in host animal species composition among sites
The NMDS analysis and k-means clustering produced two separate groups (clusters 1 and 2) of sites ( Figure 3). Cluster 1 included 10 sites (sites 5-7, 10, 12, 14-17, and 22) in which sika deer are found.

| Relationships among host animals
The correlation between the presence of sika deer (Dcn) and altitude (Alt) was not significant, according to the Spearman rank test  Table 1). In the other models, estimates of the explanatory variables were not significant (p > .05). In each model, the predicted values mostly fit the observed values within 95% CI (Figure 4; Appendix S4).

| The host animal species of H. japonica
The present study identified 17 host animal species of H. japonica ( Figure 2; Appendix S2 has also been reported in South-East Asia (Schnell et al., 2018). The result of the present study suggest that H. japonica also feeds on a wide range of large-bodied Carnivora species (e.g., Asian black bear) and small mammals (e.g., Japanese weasel and Siberian chipmunk), but perhaps as infrequently as has been previously reported in other Asian Haemadipsa species.
Notably, the present study revealed that several H. japonica fed on blood meals from frog species (Figure 2; Appendix S2). Though it is known that H. japonica feeds on frog blood meals in the experimental rearing environments (Yoshiba, 1996), only two previous studies have reported that H. japonica feeds on frogs and salamander in the wild in Japan. One of the aforementioned studies discovered that iDNA of H. japonica belonged to Yakushima Tago's brown frog (Rana tagoi yakushimensis) from Yakushima Island (Hanya et al., 2019). The other study found that a haemadipsid leech (identified as H. zeylanica japonica) fed on Sword-tail newts (Cynops ensicauda) in the Ryukyu islands (Miyata, Miyagi, & Tsukamoto, 1978). However, several studies outside of Japan have described frogs as hosts of various land leech species (e.g., Rocha, Borda, Andreone, & Rosa, 2012;Schnell et al., 2018), and salamanders have also been infrequently described as hosts (Lunghi et al., 2018). This lack of studies reporting on amphibians as hosts of H. japonica can be attributed to the analytical bias that is present when using DNA analyses or immunohistological analyses; amphibians were not used as candidates for host species identification in previous studies using fragment size-based DNA analyses or immunohistological analyses. In general, amphibians dwell in moist, mountainous habitats, such as along small streams and damp areas. Therefore, the habitats of amphibians overlap with the habitats of H. japonica, and consequently, the leech has opportunities to feed on blood meals from amphibians.

| The host-parasite interface between H. japonica and C. nippon
The result of NMDS and k-means clustering in this study showed that the host animal species compositions of the sites were divided into two groups, which were mainly separated according to sika deer presence or absence (Figure 3). In addition, the results of the MRPP and ANOSIM analyses suggested that host species composition differed significantly between sites with sika deer and sites without sika deer. These results strongly indicate that sika deer distribution affects the host-parasite interface of H. japonica. Since around the 1990s, the populations of sika deer have substantially increased in Japan (Takatsuki, 2009 (Yoshiba, 1996). In addition, Davies and McLoughlin (1996) found that mammalian blood provides more energy than amphibian blood, resulting in more available energy for somatic and reproductive growth in an aquatic leech, Hirudo medicinalis.

CO N FLI C T O F I NTE R E S T S
None declared.