Skip to main content

Trace of outbreeding between Biwa salmon (Oncorhynchus masou subsp.) and amago (O. m. ishikawae) detected from the upper reaches of inlet streams within Lake Biwa water system, Japan

Abstract

The establishment of fluvial fish populations from anadromous populations by natural or artificial barriers obstructing migration is a good research subject to study life history plasticity. Biwa salmon, Oncorhynchus masou subsp., a salmonid fish endemic to the Lake Biwa water system, exhibit life history variation (e.g., mature stream-resident males) in addition to a typical lacustrine life history type, indicating potential adaptations of life histories in response to emergence of barriers. Currently, fluvial populations that are morphologically similar to both stream-resident Biwa salmon and amago, the fluvial red-spotted masu salmon, Oncorhynchus masou ishikawae, are found upstream of dams which were constructed in the inflowing rivers of Lake Biwa. However, it is unknown whether they are Biwa salmon or amago. To explore that, the genetic characteristics of nine fluvial populations were investigated through AFLP and mtDNA analyses. Bayesian admixture analysis based on the AFLP data revealed that three fluvial populations were admixed populations between Biwa salmon and amago. In addition, a Biwa salmon mtDNA haplotype was detected in some individuals from three populations. However, no genetically pure fluvial populations of Biwa salmon were found, indicating no life history plasticity in this subspecies, and thus hybridization with amago boosted the ability of this subspecies to establish fluvial populations. Nevertheless, other scenarios, such as hybridization after establishment of fluvial populations of Biwa salmon, are also possible. The latter hypothesis could be supported by the fact that amago did not inhabit the river before emergence of barriers. However, a significant genetic population structure was found only in amago, suggesting that this subspecies is native to the Lake Biwa water system. But the possibility that multiple sources of amago have been released into rivers cannot be excluded. Therefore, further studies on the relationships between amago populations in the upper reaches of the Lake Biwa water system and other populations in the surrounding areas of the Lake Biwa water system are needed to clarify the origins of the admixed populations.

Introduction

The establishment of fluvial populations of fish from anadromous populations due to natural or artificial barriers is a good subject to study the plasticity of life histories. In particular, it is known that salmonid fishes that have both an anadromous type and a fluvial type change their life histories relatively flexibly depending on the environment (Sugiwaka and Kojima 1984; Thorpe 1989; Thorpe 1994; Kaeriyama 1996; Morita and Yamamoto 2004). It has been shown that due to the construction of artificial barriers (e.g., check dams), anadromous types of the white-spotted char (Salvelinus leucomaenis) and the masu salmon (Oncorhynchus masou masou) disappeared in some populations, and the population densities upstream of barriers decreased. Some studies have reported that fluvial populations have most probably been established upstream of barriers by increases in initial growth rates (Nakano and Maekawa 1994; Yamamoto and Nakano 1996; Morita et al. 2000; Shimoda et al. 2002; Morita and Nagasawa 2010). In fact, in salmonid species that have both an anadromous type and a stream-resident type, it seems that individuals reaching a critical size at a certain time of the juvenile growth period will mature in the river (Utoh 1976; Thorpe 1986; Nakano and Maekawa 1994; Morita et al. 2000).

In the Lake Biwa water system, there are two subspecies of Oncorhynchus masou. One is Biwa salmon (Oncorhynchus masou subsp.), which is endemic to Lake Biwa and mainly inhabits the lake basin. The other is a fluvial form of red-spotted masu salmon (amago) (Oncorhynchus masou ishikawae), which inhabits the upper reaches of rivers around the lake (Oshima 1957; Yamamoto 1973; Furukawa 1989; Nakano et al. 1989). In the past, it appeared that all individuals of Biwa salmon run to the lake in early summer after hatching and that there were no stream-resident (fluvial) types (Furukawa 1989; Fujioka 1990). In addition, because Biwa salmon ascended rivers from the lake to spawn in the autumn, but did not go up as far as the upper reaches where amago inhabit, it appeared that Biwa salmon and amago inhabited different areas of the Lake Biwa watershed: Biwa salmon in the lower courses of the inflowing rivers and the lake, and amago in the upper courses (Kato 1978; Furukawa 1989). However, Fujioka and Fushiki (1988) suggested that mature stream-resident males of Biwa salmon exist, a result of a stocking experiment, and Kuwahara and Iguchi (1994) identified most parrs collected at spawning sites during the spawning season as Biwa salmon, based on morphological traits, and reported that there were immature males and females as well as mature males at these sites. Moreover, Kuwahara and Iguchi (2007) reported that some Biwa salmon, as well as red-spotted masu salmon, swam up rivers in early summer. These early migrating groups probably swam up rivers to spawn in the autumn and so had to spend the high temperature summer season in the river. Therefore, it was thought that these groups swam up to the cooler upstream areas that amago inhabit (Kuwahara and Iguchi 2007). In fact, based on the oral survey of members of river fishery cooperatives of the Yasu and Ado Rivers, Kuwahara et al. (1992, 1993) reported that until about the 1970s Biwa salmon reached the habitat area of amago and spawned there, and that individuals similar to amago gathered around spawning pairs. In addition, parr males that spawn with lake-run type Biwa salmon, or which were sneakers, have been observed (Kuwahara, unpublished data). From these facts, there is a high possibility that some Biwa salmon originally spawned in the upper reaches of inflowing rivers, in the same areas that amago inhabit, and the population of Biwa salmon includes stream residents. However, it is considered that Biwa salmon cannot reach the habitat of amago at the present time because many rivers around the lake have since been obstructed by artificial barriers, such as dams, and are affected by drought caused by water withdrawal for agriculture (Kuwahara and Iguchi 2007; Fujioka 2009; Kuwahara 2013). Therefore, there is the possibility that fluvial populations of Biwa salmon have been established in areas upstream of the obstructions (Nakano and Maekawa 1994; Yamamoto and Nakano 1996; Morita et al. 2000; Shimoda et al. 2002; Morita and Nagasawa 2010).

In Shiga Prefecture, stocking with amago from Gifu Prefecture started in the upper reaches of rivers around Lake Biwa from 1970 (Kamata 1979). After that, lake-run amago salmon was caught in Lake Biwa since the 1970s (Kato 1981). Despite the ease of distinguishing between Biwa salmon and the lake-run amago salmon morphologically, there were no reports of lake-run amago salmon in the lake before that (Kato 1981; Fujioka 2009). Kuwahara et al. (2012) compared Biwa salmon, lake-run amago salmon caught in the lake, and amago from Samegai Trout Farm, which was the main source of released seedlings around Lake Biwa, based on mtDNA and nuclear genomes. They found that most individuals of Biwa salmon and lake-run amago salmon had mtDNA haplotypes consistent with species identification, with very few individuals with different haplotypes. On the other hand, unilateral gene introgression in the nuclear genome has occurred from Biwa salmon to lake-run amago salmon, and the proportion of the Biwa salmon genome in lake-run amago salmon is variable. Moreover, the haplotype of the mtDNA and the cluster of the nuclear genome of lake-run amago salmon almost matched with that of amago from Samegai Trout Farm (Kuwahara et al. 2012). From the above results, Kuwahara et al. (2012) inferred that lake-run amago salmon from Lake Biwa originated from stocked individuals, and the Biwa salmon genome detected in lake-run amago salmon was derived from mature stream-resident males that emerged from Biwa salmon seedlings stocked into inflowing rivers. However, whether the amago-like fish originally inhabiting the upper reaches of inflowing rivers of the Lake Biwa area were fluvial populations of red-spotted masu salmon or stream-resident type of Biwa salmon has not been revealed.

From these facts, essentially the following three possibilities are assumed in areas upstream of the obstructions at the inflowing rivers of Lake Biwa. First, there are fluvial populations of Biwa salmon. Second, there are native populations of amago. Third, there are admixed populations of Biwa salmon and amago. In addition, there is a possibility that stocked amago has had an influence on these populations. In this study, we verified these possibilities by clarifying the genetic characteristics of the amago-like fish populations inhabiting upstream areas of rivers around Lake Biwa with sequence data of mtDNA and AFLP data.

In this paper, “fluvial” and “stream-resident” were defined as follows. “Fluvial” was resident individuals in the river which propagated themselves without anadromous individuals. “Stream-resident” was resident individuals in the river which reproduced with anadromous individuals.

Materials and methods

Fish specimens and sampling locality. Specimens were collected from nine rivers in eight river systems around Lake Biwa [Wani River (WAN), Otani River (OTA), Ukawa River (UKA), Aratani River (ARA), Kawachidani River (KAW), Chinai River (CHI), Okawa River (OKA), Obanashi River (OBA), Chaya River (CHA)] with an electric shocker (Model 12 Backpack Electrofisher; Smith-Root Inc., Vancouver, WA, USA) and fishing from 2006 to 2009 (Fig. 1). We selected the former eight rivers which were without fishing rights at the sampling site, or without officially stocked amago near the sampling site from oral survey. The Chaya River was selected as compared to the other eight rivers, because originally amago did not inhabit the river and the population of amago in this river was as a result of official stocking after 1970 (Table 2). There was a check dam without fish ladder just downstream of each sampling site. It is unknown when these check dams were constructed. Fishes can move downstream over the check dam. The number of samples per collection day and the number of samples used for mtDNA sequencing and AFLP analyses are shown in Table 1. Sampling was performed more than once at each sampling site, because we could not collect a sufficient number of samples at one time. Information on the migration of Biwa salmon and the stock of Biwa salmon and amago at the sampling sites is shown in Table 2. The sampled individuals were anesthetized with FA 100 (Tamura Pharmaceutical Co., Ltd.), and clips of the adipose fin and caudal fin from all sampled individuals were collected for DNA analysis. These fin clips were preserved in 100 % ethanol. The total genomic DNA was isolated from the fin clips using QIAamp DNA Mini Kit (QIAGEN K.K., Tokyo).

Fig. 1
figure1

Map of sampling sites shown by dashed line circles in each inflowing rivers of the Lake Biwa basin, Shiga Prefecture

Table 1 Sampling dates and number of samples for mtDNA and AFLP analyses at each sampling site
Table 2 The details of Biwa salmon and amago at each sampling site

We used the data of three populations [Biwa salmon in Lake Biwa (BIW), lake-run amago salmon (LAS) and hatchery-reared amago from Samegai Trout Farm (HAS)] reported by Kuwahara et al.(2012), in addition to the 9 populations in this study.

mtDNA sequencing. The sequence of 943 base pairs from ATPase 6 to cytochrome oxidase subunit 3 (CO III) was determined for 413 specimens. This target region of the mtDNA was amplified by polymerase chain reaction (PCR), following the procedure of Oohara and Okazaki (1996) . The PCR products were purified with QIAquick PCR Purification Kit (QIAGEN K.K., Tokyo) and sequenced with ABI3130XL Genetic Analyzer (Applied Biosystems Japan, Ltd., Tokyo) after reacting with BigDye Terminator version 3.1 Cycle Sequencing Kit (Applied Biosystems Japan, Ltd., Tokyo). The sequences newly detected in this study were deposited in DDBJ/EMBL/GenBank (Accession number LC314728-LC314738).

The haplotype diversity and nucleotide diversity of each population were evaluated using Arlequin ver. 3.5 (Excoffier and Lisher 2010). In addition, the minimum spanning network (MSN) of haplotypes was estimated using Arlequin software.

Amplified fragment length polymorphism analysis. AFLP analysis was conducted basically following the procedures of Vos et al. (1995). The selective reaction was carried out with nine primer pairs (EcoRI-AAC/MseI-CATG; EcoRI-ACG/MseI-CATG; EcoRI-AGT/MseI-CATG; EcoRI-ATA/MseI-CATG; EcoRI-AAG/MseI-CATG; EcoRI-ATC/MseI-CATG; EcoRI-ATA/MseI-CTGC; EcoRI-AAG/MseI-CTGC; EcoRI-ATC/MseI-CTGC), and each EcoRI primer was labeled with a fluorescent label.

After mixing the GeneScan 500 ROX size standard (Applied Biosystems Japan, Ltd., Tokyo) and Hi-Di-Formamide (Applied Biosystems Japan, Ltd., Tokyo) to each selective reaction solution, amplified fragments of AFLP were detected using ABI 3130 XL Genetic Analyzer (Applied Biosystems Japan, Ltd., Tokyo). The experiment of this paper was done using the same machine under the same conditions and at the same time as those of Kuwahara et al. (2012). Loci that contained ambiguous peaks lower than 50 relative fluorescence units in the allele calling procedure were excluded from the analysis to avoid the possibility of classing instrument noise from the sequencer as real peaks. Based on the obtained data, the heterozygosity of each group was calculated using AFLP-SURV ver. 1.0 (Vekemans, 2002).

Genetic population structures were estimated using STRUCTURE ver. 2.2 (Falush et al. 2007). STRUCTURE analysis was performed using an admixture model. We analyzed twelve populations, including nine populations in this study and three populations in Kuwahara et al. (2012). So, assuming that the number of clusters (K) ranges from 1 to 13, five iterations of 100,000 simulations per K were run with random seed number. Estimation of the K value was proposed by Pritchard et al. (2000) and based on the rate of change of Ln P (D). On the other hand, Evanno et al. (2005) proposed estimation of the K value using the change of ΔK based on the rate of change of Ln P (D) between adjacent Ks. Meirmans (2015) suggested that it was necessary to consider all assumed K values. Therefore, the K value was estimated from both the rate of change of Ln P (D) and the change of ΔK, and all estimated K values were examined in this study.

The mean value of the posterior estimates of the proportion of ancestry (q) from each of the nine populations was then calculated for each individual. We applied a threshold q value of 0.1 on hybrid identification, i.e., individuals with 0.9 > q > 0.1 were assigned as hybrids.

We also used NewHybrids ver. 1.1 beta (Anderson and Thompson 2002; Anderson 2008) to identify early generations of hybrids. NewHybrids assigns posterior probabilities of individuals belonging to six classes (pure parental species, F1 or F2 hybrids and first-generation backcrosses in either direction). As prior information, individuals of the hatchery-reared amago and Biwa salmon with q > 0.99 and q < 0.01, respectively, were designated as pure parental species (Table S2). Jeffrey’s priors were used for the mixing proportions and allele frequencies, and three independent runs of 30,000-iteration burn-in followed by 20,000 iterations of data collection were performed.

Results

mtDNA sequencing. A total of 25 haplotypes were detected from the populations used in this study, 11 of which were haplotypes newly detected in this study, while the remaining 14 had been previously reported in Kuwahara et al. (2012) . The gene diversity and nucleotide diversity values (± SE) of each group, including the three populations of Kuwahara et al. (2012) , are shown in Table 3.

Table 3 The gene and nucleotide diversities of mtDNA in populations of Oncorhynchus species in Lake Biwa and its tributaries

From the results of MSN, the Biwa salmon haplotypes (Hap - 4, 5) greatly differ from the amago haplotypes (Fig. 2).

Fig. 2
figure2

Minimum spanning network (MSN) for mitochondrial DNA haplotypes confirmed from all samples in this study and samples in Kuwahara et al. (2012) . Blue numbers are the new haplotypes found in this study, and the others are from previous studies. One bar and dot indicate one mutational step and hypothetical missing haplotypes, respectively. Red bars and red dots show alternative relationships between haplotypes

Biwa salmon haplotype (Hap-4) was detected in 26 of 52 individuals (50 %) from the Wani River, from 17 of 22 individuals (77.3 %) from the Aratani River, and from 15 of 49 individuals (30.6 %) from the Kawachidani River. Major haplotypes detected from amago in the Samegai Trout Farm (Hap-1, 3, 11) were found in some rivers around Lake Biwa. Hap-1 was found in the Wani River, Ukawa River, Kawachidani River and Obanashi River. Hap-3 was found in the Kawachidani River, Okawa River, Obanashi River and Chaya River. Hap-11 was not found in any rivers (Fig. 3).

Fig. 3
figure3

Frequency distributions of haplotypes confirmed from nine sampling groups in this study and three groups from Kuwahara et al. (2012)

AFLP analysis. A total of 631 loci were detected in this study, and 609 loci were polymorphic. The locus with polymorphism/all loci of each primer set were as follows: AAC+CATG (EcoRI+MseI) 80/81, ACG+CATG 80/80, AGT+CATG 95/95, ATA+CATG 50/60, AAG+CATG 43/49, ATC+CATG 43/47, ATA+CTGC 76/76, AAG+CTGC 66/67, ATC+CTGC 76/76. The fragment size of each AFLP marker from all samples including the three groups in Kuwahara et al. (2012) produced for 9 EcoRI/MseI selective primer combinations are shown in Table S1. The estimated heterozygosity value (± SE) of each population, including the three populations of Kuwahara et al. (2012), are shown in Table 4.

Table 4 Estimated heterozygosity values (±SE) from AFLP data within the 12 sampled groups

For the change in the value of ΔK, peaks were found at K = 2, 4 and 5. The highest peak is at K = 2 and the lowest at K = 5 (Fig. 4). The rate of change of Ln P (D) was large at K=2 and 4, and no significant change was observed at K = 5 (Fig. 4). Meirmans (2015) argued that all possible clusters should be considered, because it is difficult to determine the K value by STRUCTURE software. Therefore, it is necessary to examine K = 2 and 4.

Fig. 4
figure4

Ln P (D) computed using STRUCTURE version 2.2., for estimating the most suitable number of clusters (K). Posterior probability of the data Ln P (D) against the number of K clusters (below) (Pritchard et al., 2000), and estimating the most suitable number of clusters (K) (ΔK = m|L’’(K)|/s[L(K)]) (above) (Evanno et al. 2005). Probability of the most suitable number of clusters (K) is calculated for K = 1-13 in the AFLP data of nine river populations and three groups from Kuwahara et al. (2012)

In the case of K = 2, cluster 2 was detected at more than 10 % in 23 individuals (43.4 %) from the Wani River and 4 individuals (7.8 %) from the Ukawa River. In addition, cluster 2 was detected at more than 40 %, 15 %, and 20 % from all individuals from the Aratani River, Kawachidani River and Chinai River, respectively (Fig. 5). Cluster 1 and 2 correspond to amago and Biwa salmon, respectively (Kuwahara et al. 2012).

Fig. 5
figure5

Admixture analysis of nine river populations and three groups from Kuwahara et al. (2012) based on AFLP data, calculated by STRUCTURE version 2.2 with k = 2 (above) and k = 4 (below). Each individual is represented as a vertical bar partitioned into K segments, with each length proportional to the estimated membership in each cluster. The proportion of mtDNA haplotypes of each sampling group is shown by pie charts (top)

2, 1, 13, 23, 4, and 22 individuals were detected as hybrids (0.9 > q > 0.1) from hatchery-reared amago from Samegai Trout Farm, Biwa salmon in Lake Biwa, lake-run amago salmon, Wani River, Ukawa River, and Aratani River. All individuals from Kawachidani River and Chinai River were identified as hybrids. All individuals from other rivers were not identified as hybrids (Table S2).

The NewHybrids analyses were conducted independently for the Wani River, Ukawa River, Aratani River, Kawachidani River, and Chinai River, where hybrids were detected in STRUCTURE analysis. For each population, 8 individuals of genetically pure amago (q > 0.99) and Biwa salmon (q < 0.01) in the hatchery-reared amago and Biwa salmon, respectively, were designated as pure parental species (Table S2). Except for the Aratani River, consistent results were not obtained in the three independent runs. In the Aratani River, 13 and 5 individuals were assigned with high posterior probability (> 0.99) to F1 hybrid and first-generation backcrosses toward Biwa salmon, respectively, and consistent results were obtained in the three independent runs.

In the case of K = 4, cluster 1, which is the main genome of amago from Samegai Trout Farm, was detected at 60.7 %, 33.7 %, 57.7 %, and 85.6 % from the populations in the Wani River, Aratani River, Kawachidani River, and Okawa River, respectively (Fig. 5). Cluster 3 was detected most frequently at 96.1 %, 85.3 %, 60.2 %, and 80.9 % in the Otani River, Ukawa River, Chinai River, and Okawa River, respectively (Fig. 5). On the other hand, cluster 4 was detected at 97.1 % in the Chaya River where only released individuals exist (Fig. 5).

Discussion

From the results of the AFLP analysis, it is suggested that the fluvial populations of the Aratani River, Kawachidani River, and Chinai River are admixed populations of Biwa salmon and amago (Fig. 5). The Biwa salmon genome was detected from some individuals from the Wani River and Ukawa River (Fig. 5). Originally, Biwa salmon did not migrate to the Chaya River, and there is no information about Biwa salmon in the Otani River and Obanashi River. However, in the six other rivers, it has been confirmed that there is a high possibility that Biwa salmon migrated to the upper reaches of each river in the past (Table 2). The rivers that we sampled in this study are now divided by transverse constructions, such as dams, and we cannot find migrated Biwa salmon from Lake Biwa to the upper reaches of these constructions. Nonetheless, the fact that the Biwa salmon genome was detected at least in five rivers suggested that there was genetic exchange between Biwa salmon and amago when Biwa salmon migrated to the upper reaches of these rivers. However, in this study, genetically pure fluvial populations of Biwa salmon were not found (Fig. 5). These facts suggest the following two possibilities. The first possibility is that hybridization with stocked amago occurred after the establishment of fluvial populations of Biwa salmon. Another possibility is that hybridization with amago boosted the ability of Biwa salmon to establish fluvial populations.

Three clusters were detected from amago in the case of K = 4 (Fig. 5). Cluster 1 is a major cluster detected from amago cultivated in the Samegai Trout Farm, which is the main stocked seedling in the Lake Biwa water system. Amago of Samegai Trout Farm is cultivated using individuals from Gifu Prefecture. Therefore, it is highly likely that cluster 1 is not a genome derived from the Lake Biwa water system. Cluster 4 is the main cluster of the population in the Chaya River. It was reported that amago did not originally inhabit the Chaya River and only white-spotted charr (Salvelinus leucomaenis) lived there as the salmonid fishes (Kuwahara et al. 1994). In the Chaya River, after the trial stocking of amago from Samegai Trout Farm in 1970 (Kamata 1979), the stocking of amago has been carried out each year by the Echi River Upstream Fishery Cooperative, with stocked individuals brought from Mie Prefecture and not only from Samegai Trout Farm (Ikeda, personal communication). Therefore, it is considered that cluster 4 is also very unlikely to be derived from the Lake Biwa water system. These facts suggest that there are multiple origins of stocked amago in the Lake Biwa water system. On the other hand, cluster 3 was the main cluster in the Otani River, Ukawa River, Chinai River, and Okawa River, where official amago stocking has not been done, and this cluster was also detected in about 30 % from the Wani River (Fig. 5). Amago from Samegai Trout Farm has been widely used for the seedling of cultivation and stocking in Shiga Prefecture from the beginning. Therefore, it seems that amago from Samegai Trout Farm was easy to use also for non-official stocking. Additionally, white-spotted char inhabiting the Lake Biwa water system is closely related with those of adjacent rivers flowing to the Sea of Japan, suggesting the possibility that dispersal by watershed exchange into the Lake Biwa water system can occur (Yamamoto et al. 2004; Kikko et al. 2008). Furthermore, there are some studies that have reported close genetic relationships of fish species between the Lake Biwa water system and adjacent water systems, e.g., kawayoshinobori (Rhinogobius flumineus), hariyo (Gasterosteus microcephalus), and gigi (Pelteobagrus nudiceps) (Shimizu et al. 1993; Watanabe et al. 2003; Watanabe and Nishida 2003). This suggests that amago that have the cluster 3 genome, of which the origin is unknown, came into the Lake Biwa water system from adjacent water systems by watershed exchange. It is considered that amago originally existed in the upper reaches of rivers around Lake Biwa (Oshima 1957; Yamamoto 1973; Furukawa 1989; Nakano et al. 1989). So it is highly probable that this cluster 3 is a native genome.

Assuming that cluster 3 is a native genome, prior to river fragmentation, the fluvial populations of amago that have this genome and Biwa salmon that migrate from Lake Biwa would be spawning sympatrically (Kuwahara et al. 1992, 1993; Kuwahara and Iguchi 2007). Moreover, because salmonid fishes spawning sympatrically are susceptible to crossbreeding among different species (Kato 1977; Kitano et al. 2009; Ichimura et al. 2011; Kitano et al. 2014), crossbreeding between Biwa salmon and amago would have occurred. Populations including individuals with both genomes have therefore become established in the upper reaches of the rivers flowing into Lake Biwa. This is supported by artificial hybridization experiments, which have shown that crossbred individuals (F1) can occur (Fujioka 1991). Therefore, it is inferred that even after the fragmentation of the rivers, the crossbreeding populations remained in upstream areas and then further mating with artificially stocked amago occurred.

Individuals regarded as lake-run amago salmon have been caught in Lake Biwa from 1970 when the stocking of amago from Samegai Trout Farm began (Kawabata 1931; Kato 1981; Fujioka 2009). Amago from Samegai Trout Farm originated from the eyed eggs received from Gifu Prefectural Fresh Water Fish Research Institute in 1967 (Tazawa and Kamata 1969). Seedlings of amago from Gifu Prefectural Fresh Water Fish Research Institute originated from Kiso River water system (Honjoh 1977). In Kiso River, it is known that many red-spotted masu salmon originally migrated upstream from the Ise Wan Sea (Oshima 1957), and Honjoh (1977) said that silvering individuals appeared from amago reared in the pond of fishery. In fact, in the oral survey some members of Katsuragawa fishery cooperative in Adogawa River said that silvering individuals easily appeared from amago from Samegai Trout Farm (Kuwahara et al. 1993). These facts also seem to suggest that the origin of the lake-run amago salmon caught in Lake Biwa is amago from Samegai Trout Farm. In addition, the genome detected from lake-run amago salmon caught in the lake was consistent with the genome of amago from Samegai Trout Farm (Kuwahara et al. 2012). In the case of K = 4, cluster 1 was also detected at about 80 % in lake-run amago salmon in this reanalysis and was also detected from all individuals (Fig. 5). Moreover, the mtDNA haplotypes detected from lake-run amago salmon were almost consistent with the haplotypes of amago from Samegai Trout Farm (Kuwahara et al. 2012). From these facts, it seems that the lake-run form originated only from individuals having cluster 1 genome, except for pure Biwa salmon. In other words, it is highly probable that the crossbred individuals between Biwa salmon and amago that have the cluster 3 genome, which is presumed to have occurred before the fragmentation of the rivers, did not migrate to the lake.

On the other hand, assuming that cluster 3 is not a native genome, the genomes of native amago were not detected in this study (Fig. 5) and we could not confirm the possibility that native amago existed in the Lake Biwa water system. From this, it seems that native amago did not exist in the Lake Biwa water system. In this case, it is inferred that individuals that inhabited the upper reaches of rivers flowing into Lake Biwa and were regarded as amago (Oshima 1957; Yamamoto 1973; Fururkawa 1989; Nakano et al. 1989) were actually a stream-resident type of Biwa salmon (Fujioka and Fushiki 1988; Kuwahara and Iguchi 1994) and that they mated with Biwa salmon that migrated upstream from Lake Biwa (Kuwahara et al. 1992, 1993; Kuwahara and Iguchi 2007).

In 1970, a dam was constructed on the Ishida River, about 5 km downstream from where the Kawachidani River, which was one of the five rivers where the Biwa salmon genome was detected, joins it. During construction of the dam, change in the river channel was undertaken about five years before completion of the dam. Therefore, it is considered that the migration of Biwa salmon to the upper reach above the Ishida River dam was stopped from this time. In addition, according to records, stocking of amago in the Ishida River started in 1972, although the detailed locations of releases are unknown (Kamata 1979). So, at the very least, seven years had elapsed since the migration of Biwa salmon stopped and stocking of amago started. Likewise, in the Wani River, although the precise times at which river improvements and dam construction were started are unknown, Biwa salmon migrated to the sampling site in the past. However, Biwa salmon have not been seen since the late 1950s in the upper reaches of the Wani River (Nishimura, personal communication). Public stocking of amago in Shiga Prefecture began in 1970 (Kamata 1979), and the private stocking is also thought to have occurred since 1969 when amago seedlings began to be mass produced (Honjoh 1977). Therefore, it is thought that the stocking of amago was carried out in the Wani River after more than 10 years since the migration of Biwa salmon was stopped. In addition, since the Biwa salmon type mtDNA haplotype was detected from both rivers (Figs. 3 and 5), it is suggested that females of Biwa salmon were involved in spawning after migration of Biwa salmon stopped in both rivers. In other words, at least in the Wani and Ishida rivers, it is inferred that both the lake-run and stream-resident types of Biwa salmon originally existed. It was thought that fluvial populations of Biwa salmon formed after the migration of Biwa salmon was stopped, and when the artificial stocking of amago started, the fluvial Biwa salmon crossbred with amago.

As described above, the following two possibilities were suggested regarding the formation of the genetic structure of the populations inhabiting the upper reaches of inflowing rivers into Lake Biwa. Firstly, there were originally crossbred groups of native amago and lake-run Biwa salmon that migrated upstream from Lake Biwa. Since the artificial division of rivers occurred afterward, it appears that lake-run Biwa salmon was not able to migrate upstream, the crossbred groups remained and further crossbreeding occurred with artificially stocked amago. Another is the possibility that fluvial populations of Biwa salmon were established in upstream areas after fragmentation of rivers occurred, and then hybridized with artificially stocked amago. However, as mentioned earlier, cluster 3 is highly likely to be native, so it seems that the previous possibility is supported.

To verify these possibilities and clarify the details of the relationship between Biwa salmon and amago, which is supposed to inhabit inflowing rivers of Lake Biwa, it is necessary to clarify the origins of the cluster 3 genome shown in Fig. 5. Therefore, it is necessary to clarify the details of the relationship of the cluster 3 genome with amago and masu salmon (yamame) Oncorhynchus masou masou in the surrounding areas of the Lake Biwa water system.

Biwa salmon is currently regarded as one subspecies of the Oncorhynchus masou species complex. However, Nakabo (2009) and Hosoya (2013) suggested that Biwa salmon is an independent species based on biological information (Kato 1978; Kuwahara and Iguchi 2007), genetic data (Tega et al. 2012), etc. The recent phylogenetic analysis based on AFLP data suggested that Biwa salmon is considerably genetically differentiated from the other three subspecies (Takahashi et al. 2016). In this study, it was suggested that there was a possibility that a crossbred group of Biwa salmon and red-spotted masu salmon (amago) existed in the upstream areas of inflowing rivers of Lake Biwa. Nevertheless, it is also shown that a population of pure Biwa salmon existed in Lake Biwa. This seems to support the views of Nakabo (2009) and Hosoya (2013) that Biwa salmon is an independent species. To clarify the details of the relationship between Biwa salmon and red-spotted masu salmon and to promote the conservation of Biwa salmon, which is endemic to Lake Biwa, detailed taxonomic studies on Biwa salmon are necessary.

References

  1. Anderson, EC (2008) Bayesian inference of species hybrids using multilocus dominant genetic markers. Phil Trans R Soc B 363:2841–2850

  2. Anderson EC, Thompson EA (2002) A model-based method for identifying species hybrids using multilocus genetic data. Genetics 160:1217–1229

  3. Evanno G, Regnaut S, Goudet J (2005) Detecting the number of clusters of individuals using the software STRUCTURE : a simulation study. Mol Ecol 14:2611–2620

  4. Excoffier L, Lischer H (2010) Arlequin suite ver. 3.5: a new series of programs to perform population genetics analysis under Linux and Windows. Mol Ecol Res 10:564–567

  5. Falush D, Stephens M, Pritchard JK. 2007. Structure software, Version 2.2. University of Oxford, Oxford

  6. Fujioka Y (1990) Biwa salmon : its ecological and morphological characteristics. Technical Reports of Hokkaido Salmon Hatchery (Fish and Eggs) 159:25–38

  7. Fujioka Y (1991) Morphological, physiological and ecological studies on Biwa salmon. BULL SHIGA PREF SAMEGAI TROU FARM 3:1–112

  8. Fujioka Y (2009) Introduction of Biwa salmon Migrating between River and Lake. Sunrise Shuppann, Hikone, 216pp

  9. Fujioka Y, Fushiki S (1988) Downstream migration and body silvering in underyearling Biwa salmon Oncorhynchus rhodurus. Nippon Suisan Gakkaishi 54:1889–1897

  10. Furukawa T (1989) Biwamasu: Oncorhynchus masou subsp. In:Kawanabe K, Mizuno N (eds) Freshwater fishes of Japan. Yamakei, Tokyo, pp 180–181

  11. Hosoya K (2013) 96. Family Salmonidae. In: Nakabo K (ed) Fishes of Japan with pictorial keys to the species third edition. Tokaidaigaku Shuppannkai, Kanagawa, pp 1833–1835

  12. Honjoh T (1977) Studies on culture and transplantation of Amago salmon salmon, Oncorhynchus rhodurus. Report of Gifu Prefectural Fresh Water Fish Research Institute 22:1–103

  13. Ichimura M, Yanagimoto T, Kobayashi T, Masaoka T, Kaeriyama M (2011) Hybrid identification of “Sakemasu” collected in the Nemuro Strait, eastern Hokkaido, using DNA analysis. Nippon Suisan Gakkaishi 77:834–844

  14. Kaeriyama M (1996) Effect of population density and habitat environment on life history strategy and migration of juvenile Sockeye (Oncorhynchus nerka) and Chum Salmon (O. keta). Scientific Reports of the Hokkaido Salmon Hatchery 50:101–111

  15. Kamata T (1979) The native stocks and transplantation of Amago salmon and Iwana trout in Shiga Prefecture. In: Shiga Shizen Kankyo Kenkyukai (ed) Land and life in Shiga: scientific studies of Shiga Prefecture, Japan. Foundation of Nature Conservation in Shiga Prefecture, Otsu, pp 615–622

  16. Kato F (1978) Morphological and ecological studies on two forms of Oncorhynchus rhodurus found in Lake Biwa and adjoining inlets. Jpn J Ichthyol 25:197–204

  17. Kato F (1981) The Amago salmon salmon, Oncorhynchus rhodurus, collected in Lake Biwa. Jpn J Ichthyol 28:184–186

  18. Kato K (1977) Natural hybrids of salmonid fishes from the Nippara River, Tokyo. Jpn J Ichthyol 23:225–232

  19. Kawabata J (1931) Fishes from Lake Biwa. The society that posthumaous published the late Mr. Jugorou Kawabata’s manuscript, Tokyo, 198pp

  20. Kikko T, Kuwahara M, Iguchi K, Kurumi S, Yamamoto S, Kai Y, Nakayama K (2008) Mitochondrial DNA Population Structure of the White-Spotted Charr (Salvelinus leucomaenis) in the Lake Biwa Water System. Zool Sci 25:146–153

  21. Kitano S, Hasegawa K, Maekawa K (2009) Evidence for interspecific hybridization between native white-spotted charr Salvelinus leucomaenis and non-native brown trout Salmo trutta on Hokkaido Island, Japan. J Fish Biol 74: 467–473

  22. Kitano S, Ohdachi S, Koizumi I, Hasegawa K (2014) Hybridization between native white-spotted charr and nonnative brook trout in the upper Sorachi River, Hokkaido, Japan. Ichthyol Res 61:1–8

  23. Kuwahara M (2013) Present status and conservation of the endemic subspecies of Oncorhynchus masou (Biwa salmon) in Lake Biwa. Jpn J Ichthyol 60:63–67

  24. Kuwahara M, Iguchi K (1994) Occurrence of mature stream resident males of Biwa salmon, Oncorhynchus masou subsp. Jpn J Ichthyol 40:495–497

  25. Kuwahara M, Iguchi K (2007) Occurrence of summer upstream migration in Biwa salmon (Oncorhynchus masou subsp.). Jpn J Ichthyol 54:15–20

  26. Kuwahara M, Maehata M, Akiyama H, Matsuda M, Okano T (1992) Fishes living in Yasu River. Ann Rep Biwako Bunkakan 10:47–61

  27. Kuwahara M, Maehata M, Akiyama H, Matsuda M, Okano T (1993) Fishes living in Ado River. Ann Rep Biwako Bunkakan 11:11–21

  28. Kuwahara M, Maehata M, Akiyama H, Matsuda M, Okano T (1994) Fishes living in Echi River. Ann Rep Biwako Bunkakan 12:9–18

  29. Kuwahara M, Takahashi H, Kikko T, Kurumi S, Iguchi K (2012) Introgression of Oncorhynchus masou subsp. (Biwa salmon) genome into lake-run O. m. ishikawae (Amago salmon) introduced into Lake Biwa, Japan. Ichthyol Res 59:195–201

  30. Meirmans PG (2015) Seven common mistakes in population genetics and how to avoid them. Mol Ecol 24:3223–3231

  31. Morita K, Yamamoto S (2004) Consequences of riverine fragmentation by damming: rivers are corridors between forest and the sea. In: Maekawa K (ed) Ecology and evolution of salmon and trout. Bunichi Sougou Shuppan, Tokyo, pp 281–312

  32. Morita K, Yamamoto S, Hoshino N (2000) Extreme life history change of white-spotted char (Salvelinus leucomaenis) after damming. Can J Fish Aquat sci 57:1300–1306

  33. Morita K, Nagasawa T (2010) Latitudinal variation in the growth and maturation of masu salmon (Oncorhynchus masou) parr. Can J Fish Aquat sci 67:955–965

  34. Nakabo T (2009) Zoogeography of Taiwanese fishes. Korean J Ichthyol 21:311–321

  35. Nakano S, Taguchi S, Shibata Y, Furukawa T (1989) Satsukimasu Amago salmon: Oncorhynchus masou ishikawae In: Kawanabe K, Mizuno N (eds) Freshwater fishes of Japan. Yamakei, Tokyo, pp 169–178

  36. Nakano S, Maekawa K (1994) Charr with special reference to the life history change and damming. In: Goto A, Tsukamoto K, Maekawa K (eds) Diadromous freshwater fishes-life history and evolution. Tokaidaigaku Shuppannkai, Kanagawa, pp. 86–100

  37. Oohara I, Okazaki T (1996) Genetic relationship among three subspecies of Oncorhynchus masou determined by mitochondrial DNA sequence analysis. Zool Sci 13:189–198

  38. Oshima M (1957) Sakuramasu and Biwamasu. Nire Shobo, Sapporo

  39. Pritchard JK, Stephens M, Donnelly P (2000) Inference of population structure using multilocus genotype data. Genetics 155:945–959

  40. Shimizu T, Taniguchi N, Mizuno N (1993) An electrophoretic study of genetic differentiation of a Japanese freshwater goby, Rhinogobius flumineus. Jpn J Ichthyol 39:329–343

  41. Shimoda K, Nakano S, Yamamoto S (2002) Landlocking of anadromous white-spotted charr Salvelinus leucomaenis by damming. Jpn J Ichthyol 49:25–32

  42. Sugiwaka K, Kojima H (1984) Influence of individual density on smoltification in wild juvenile masu salmon (Oncorhynchus masou) in the Atsuta River. Scientific reports of the Hokkaido Fish Hatchery 39:19–37

  43. Takahashi H, Kondou T, Takeshita N, Hsu TH, Nishida M (2016) Evolutionary process of iwame, a markless form of the red-spotted masu salmon Oncorhynchus masou ishikawae, in the Ôno River, Kyushu. Ichthyol Res 63: 132–144

  44. Tazawa S, Kamata T (1969) 1. Study to rear the native trout. Showa 42 fiscal year annual report of Shiga Prefectural Samegai Trout Farm 11–20

  45. Tega T, Noguchi D, Sano T, Tanaka H, Kuwada T, Mito T (2012) Evaluation of genetic diversity and genetic relationship between red-spotted masu trout Oncorhynchus masou ishikawae and Biwa saimon O. m. subsp. Fish Genet Breed Sci 41:51–61

  46. Thorpe JE (1986) Age at first maturity in Atlantic salmon, Salmo salar: freshwater period influences and conflicts with smolting. Can Spec Publ Fish Aquat Sci 89:7–14

  47. Thorpe JE (1989) Developmental variation in salmonid populations. J Fish Biol 35:295–303

  48. Thorpe JE (1994) An alternative view of smolting salmonids. Aquaculture 121:105–113

  49. Utoh H (1976) Study of the mechanism of differentiation between the stream resident form and the seaward migratory form in Masu salmon Brevoort. 1. Growth and sexual maturity of precocious masu salmon parr. Bull Fac Fish Hokkaido Univ 26:321–326

  50. Vekemans X (2002) AFLP-SURV 1.0: a program for genetic diversity analysis with AFLP (and RAPD) population data. Université Libre de Bruxelles, Bruxelles. http://www.ulb.ac.be/sciences/lagev/aflp-surv.html. Accessed on 6 January 2018

  51. Vos P, Hogers R, Bleeker M, Reijans M, van de Lee T, Homes M, Frijters A, Pot J, Peleman J, Kuiper M, Zabeau M (1995) AFLP: a new technique for DNA fingerprinting. Nucleic Acids Res 23:4407–4414

  52. Watanabe K, Nishida M (2003) Genetic population structure of Japanese bagrid catfishes. Ichthyol Res 50:140–148

  53. Watanabe K, Mori S, Nishida M (2003) Genetic relationships and origin of two geographic groups of the freshwater threespine stickleback, ‘Hariyo’. Zool Sci 20:265–274

  54. Yamamoto S (1973) Fishing in the mountain streams of west Japan. Tsurinotomosha, Osaka

  55. Yamamoto S, Nakano S (1996) Growth and development of a bimodal length-frequency distribution during smolting in a wild population of white-spotted charr in northern Japan. J Fish Biol 48:101–104

  56. Yamamoto S, Morita K, Kitano S, Watanabe K, Koizumi I, Maekawa K, Takamura K (2004) Phylogeography of white-spotted charr (Salvelinus leucomaenis) inferred from mitochondrial DNA sequences. Zool Sci 21:229–240

Download references

Acknowledgements

We thank Drs. Katsuro Yahiro, Yasufumi Satoguchi, Robin J. Smith, Ryoichi Tabata (Lake Biwa Museum), Prof. Kouichi Kawamura (Faculty of Bioresources, Mie University), and Dr. Kentaro Morita (Hokkaido National Fisheries Research Institute) for encouragement and critical reading of this manuscript. Our special thanks go to Mr. Makoto Kobayashi (Residing in Takashima), Mr. Noriyuki Ikeda (Chief of the Echi River upstream fishery cooperative), and Mr. Hayato Nishimura (Residing in Otsu) for valuable information on stocking Biwa salmon and amago, and upstream migrating Biwa salmon, respectively. We are also grateful to Mr. Akihiko Mori and members of his fishing group for sampling fish for this study. This work was conducted as part of a Collaborative Research Project of the Lake Biwa Museum (Kyo 06-02).

Author information

Affiliations

Authors

Corresponding author

Correspondence to Masayuki Kuwahara.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Supplementary material 1 (XLS 37 kb)

Supplementary material 2 (XLS 128 kb)

About this article

Verify currency and authenticity via CrossMark

Cite this article

Kuwahara, M., Takahashi, H., Kikko, T. et al. Trace of outbreeding between Biwa salmon (Oncorhynchus masou subsp.) and amago (O. m. ishikawae) detected from the upper reaches of inlet streams within Lake Biwa water system, Japan. Ichthyol Res 66, 67–78 (2019). https://doi.org/10.1007/s10228-018-0650-7

Download citation

Keywords

  • Lake Biwa water system
  • Oncorhynchus masou
  • Crossbreeding
  • mtDNA
  • AFLP