In Memanbetsu town, Hokkaido island, Japan, a high frequency of natural triploid loaches Misgurnus anguillicaudatus (7.4% on average) was detected by flow cytometry for relative DNA content. Among sympatric diploid females (n=6) from a single population, we found two unique females that laid unreduced diploid eggs. They gave normal diploid progeny even after induction of gynogenesis with genetically inert UV-irradiated sperm. When fertilized with normal loach sperm, some unreduced eggs developed into triploids, but the rest into diploids. Hybridization using goldfish Carassius auratus sperm gave both normal diploid loaches and inviable allotriploid hybrids possessing the diploid loach genome and the haploid goldfish genome. Microsatellite genotyping and DNA fingerprinting demonstrated that the diploid progeny developing from the unreduced eggs were genetically identical to the mother, while the triploids had some of the paternal DNA. These results indicate that the diploid eggs reproduced unisexually as a diploid clone and in other cases developed into triploids after accidental incorporation of sperm nucleus. The presence of at least one clonal line in this area was shown by the identical DNA fingerprint detected in five out of 17 diploid loaches examined.
INTRODUCTION
Unisexual biotypes have been recorded in fishes (Vrijenhoek et al., 1989) and they are often associated with interspecific hybridization, all-female sexuality, aberrant gametogenesis and natural polyploidy (Dawley, 1989; Vrijenhoek, 1989). The occurrence of unisexual, clonal and/or polyploid biotypes caused by apparent hybridization has been reported in fish of the genera Phoxinus (Goddard and Dawley, 1990), Fundulus (Dawley, 1992), Rutilus (Alves et al., 1998, 1999), Cobitis (Kim and Lee, 2000) and Oryzias (Shimizu et al., 2000). Since the unisexual reproduction is generally a rare phenomenon in vertebrates, it is difficult to identify it in wild populations without careful surveillance of unusual occurrences such as aberrant sex ratios or the apparent appearance of hybrids and/or polyploids. However, the cryptic nature of unisexual biotypes suggests that there may be many more examples awaiting discovery (Dawley, 1989).
In the loach Misgurnus anguillicaudatus Cantor (Cypriniformes: Cobitidae), bisexual diploid individuals (2n=50) are most common among Japanese populations, but tetraploid and triploid individuals have been found infrequently among specimens obtained from fish dealers in Japan (Ojima and Takai 1979; Arai et al., 1991). The actual origin of these polyploids is still unknown. Recently, Zhang and Arai (1999) examined loaches in 35 populations in Japan using flow cytometry for relative DNA content. They reported no polyploids in 26 places, a few triploids (1.2–3.2%) in 6 places and relatively high frequencies of triploids in Hirokami Village, Niigata Prefecture (average 7.6%) and Ichinomiya town, Aichi Prefecture (7.7%), Honshu island. In the Hirokami population, they also found a unique diploid female that produced unreduced diploid eggs. When eggs of this female were fertilized with genetically inert UV-irradiated sperm, the gynogens from normal haploid eggs died, whereas those from unreduced diploid eggs survived. Fertilization of such diploid eggs with normal haploid sperm could explain the occurrence of natural triploid in this area. However, unisexual reproduction was not triggered in any unreduced diploid eggs of the loach from the Hirokami population. No clear genetic evidence of unisexual reproduction has been obtained in diploid individuals of this species.
When we examined the distribution of natural triploid loaches among seven localities in Hokkaido island, Japan, samples collected from the northern area (Memanbetsu and Furen town) included many triploids. We also identified diploid females which were capable of producing unreduced eggs. When these eggs were fertilized with normal loach sperm or with normal sperm of the goldfish Carassius auratus (Cypriniformes:Cyprinidae), some progeny were triploid or hybrid but the remainders were diploid and genetically identical to the mother. Therefore, the diploid progeny were actually unisexual clones. In the present paper, we report the presence of a cryptic clonal line in the diploid loach by providing genetic evidence from induced gynogenesis, inter-specific hybridization, microsatellite genotyping and multilocus DNA fingerprinting.
MATERIALS AND METHODS
Specimens examined and ploidy determination
To survey the frequency of natural polyploidy in loach, a total of 967 individuals were collected from Memanbetsu town, Furen town, Shibetsu city, Asahikawa city, Akkeshi town, Ebetsu city and Nanae town in Hokkaido island, Japan from 1998 and 2000 (Fig. 1). Due to the small sizes of most samples, sex was not determined in individual specimens except for several loaches used for breeding experiments. In addition, some mature females and males were collected from other two localities, Iwamizawa city and Ohno town, for breeding experiments (Fig. 1). All fishes collected were subjected to flow cytometry for relative DNA content of erythrocytic and caudal fin-epithelium cells in order to determine their ploidy status. When the normal diploid loaches exhibit DNA values of 2C in somatic cells, triploid and tetraploid loaches exhibit values of approximately 3C and 4C, respectively. Aneuploidies exhibit values in the range between two euploid DNA contents. Relative DNA content of fishes collected in 1998 was analyzed by excitation at 488 nm with a laser of the flow cytometer FACS Calibur (Becton Dickinson Biosciences, USA) after sample preparation including PI (propidium iodide) staining according to Zhang and Arai (1996). Samples collected in 1999 and 2000 were measured for DNA content by excitation in the near-ultraviolet (UV) range from a mercury lamp of the flow cytometer PA (Partec GmbH, Germany), after DAPI (4′,6-deamidino-2-phenylin-dole) staining with a preparation kit for animal cells (#06-5-4003) provided by the manufacturer of the flow cytometer (Partec GmbH, Germany).
Induced gynogenesis and interspecific hybridization
In June 2, 2000, six mature diploid females (Nos. 1–6) were selected from the samples (n=136) collected from Memanbetsu town, Hokkaido, and injected with human chorionic gonadotrophic hormone (hCG, Gonatropin R, Teikoku Zouki, Japan) to induce ovulation according to the procedure described by Suzuki and Yamaguchi (1975). One male (A) from Shibetsu city, one male (B) from Iwamizawa city (Fig. 1) where no triploid has been recorded (Zhang and Arai, 1999) and two males (C, D) from Memanbetsu town were used for normal fertilization. Sperm was obtained with a capillary tube or syringe by squeezing the abdomen of each diploid male and then diluted with physiological saline as described in Suzuki et al. (1985). Sperm of goldfish (Carassius auratus, Cyprinidae) held in the Aquarium center of the Graduate School of Fish-eries Sciences, Hokkaido University was also obtained in the same manner. Some eggs from each female were fertilized with sperm from one of the four male loaches. The remainder of each female eggs was fertilized with ultraviolet ray (UV) irradiated sperm of the goldfish in order to induce gynogenetic development, following the procedure described by Suzuki et al. (1985). Female No. 5 was not used for normal fertilization due to a shortage of spawned eggs.
Fertilized eggs of each cross were counted and incubated in a plastic pan half-filled with freshwater at room temperature. Hatched fry were counted at one day after fertilization at 28°C. After absorption of the yolk sac, all the resultant progeny were fed brine shrimp Artemia sp. Survival rates at 10 days after fertilization were recorded. Ploidy status of the resultant progeny was assayed by measuring relative DNA content of the whole body of hatched fry using DAPI staining and the PA flow cytometer described above.
In August 24, 2000, female No. 3 reached maturation again and a second artificial ovulation was made. Some eggs of this female were fertilized with UV-irradiated goldfish sperm so as to trigger gynogenesis. The remainder was fertilized with functional goldfish sperm. According to Suzuki (1968) and Kijima et al. (1996), this latter cross is expected to result in inviable interfamilial hybrids. In addition, diploid female No. 7 collected from Ohno town (Fig. 1) and diploid female No. 8, which was the progeny of the cross between an orange variant and an individual from Saitama Prefecture, Honshu island, Japan, were used for induced gynogenesis using UV-irradiated goldfish sperm as well as interfamilial hybridization using normal goldfish sperm. Normal fertilization of female No. 7 with sperm of diploid loach (male E) collected in Akkeshi town, Hokkaido was performed as a control.
Fertilized eggs were incubated according to the procedure mentioned above. Hatching rate was recorded one day after fertilization and their normal rate of hatched fry was also examined. Flow cytometry with the DAPI staining for DNA content was made to determine the ploidy status as well as the hybrid nature of the resultant progeny.
Microsatellite analysis
The offspring of female No. 1, No. 3 and No. 6 were genotyped at six microsatellite loci. DNA samples were extracted and purified from the fin or muscle of each parental fish used for the cross and from whole body or muscle of surviving fry and juveniles by the procedure previously reported in Arai and Mukaino (1997). Six primer sets for amplification of Mac2, Mac3, Mac24, Mac37, Mac45 and Mac49 microsatellite loci developed by Morishima et al., (2001) were used in the present study. Molecular cloning for microsatellite regions and designing of primer sets based on sequence data deposited in DDBJ/EMBL/GENBANK (accession number AB060172– 060186) were already reported in Morishima et al. (2001). The PCR was carried out in a reaction mixture (20 μl) containing 50–100 ng template DNA and 0.1 U Taq polymerase (TaKaRa, Japan) in the conditions for each primer set reported by Moroshima et al. (2001). PCR product was mixed with loading buffer and electrophoresed in 15% nondenaturing polyacrylamide gel. Amplified fragments were visualized by ethidium bromide staining and UV transillumination. Microsatellite alleles were designated according to their molecular size (base pair).
DNA fingerprinting analysis
For multilocus DNA fingerprinting, purified DNA was digested with Hae III (Takara, Japan), underwent electrophoresis at 60V for 37hr or more in 0.8% agarose gel with the size of 200 mm width and 250 mm length, and then transferred to a nylon membrane by capillary action according to the procedure reported in Arai and Mukaino (1997) with some modifications. The membrane was then baked to fix DNA. Tetranucleotide repeats (GGAT)4 were commercially synthesized and then prepared by a labeling kit Light Smith II TM (Promega, USA). The probe was hybridized and fragments were visualized as described by Spruell et al., (1994). A commercially available NICETM 33.15 (Jeffreys et al., 1985) probe kit (Cellmark Diagnostic, UK) was also used. Hybridization and visualization were made as described in Arai and Mukaino (1997).
DNA fragments were visually compared and scored in the range between 9 and 6 kilobase pairs to calculate band sharing indices (BSIs) according to the following formula proposed by Wetton et al. (1987), 2Nab / (Na + Nb), where Na and Nb are the total number of DNA fragments scored in individual a and b and Nab is the number of fragments shared by both individuals. BSIs were calculated among 9 diploid and 8 triploid individuals from the Memanbetsu population, analyzed on the same gel and the same membrane.
RESULTS
Frequency of natural triploid loaches
Among 967 loach samples collected from seven localities in three years, the occurrence of normal diploids and spontaneous triploids including diplo-triploid mosaics was confirmed by relative DNA content measured by the flow cytometry (Fig. 2). As shown in Table 1, we detected 913 diploid, 51 triploid and 3 diplo-triploid mosaic individuals. No tetraploid or other polyploidies were observed. Of seven localities, Memanbetsu and Furen town gave remarkably high incidences of triploidy including mosaicism, with averages 7.4% (range 4.4–9.2%) and 13.5% (4.3–21.3%), respectively.
Fig. 2
Flow cytometric histograms for relative DNA content of somatic cells when erythrocytes of normal diploid loach are used as standard of 2C. Top to bottom; diploid, triploid and diplo-triploid mosaic showing 2C, 3C and both 2C and 3C peaks, respectively, in the loach collected in wild populations in Hokkaido, Japan.
Table 1
Frequency of natural triploid loaches in wild populations of Hokkaido island, Japan.
Induced gynogenesis
As shown in Table 2, five diploid females (Nos. 1–4, 6) gave 47.6–94.1% (average 77.9%) hatching rates after fertilization with normal sperm of diploid males. Survival of the 10-day-old feeding fry was 86.6–98.8% (average 96.6%) of the numbers hatched, indicating a good quality of eggs used. When eggs of six diploid females (Nos. 1–6) were inseminated with UV-irradiated goldfish sperm, hatching rates of the resultant gynogenetic progeny were very low (0–7.2%, average 4.5%). Most hatching fry of females No. 1, No. 2, No. 4 and No. 5 exhibited characteristic inviable abnormalities referred to as haploid syndrome (Fig. 3a). Gynogens developing from the other two females No. 3 and No. 6 were externally normal and viable (Fig. 3b). At 10 days after fertilization, a high percentage of these fry (83.7–96.2%) survived and reached the feeding stage. The gyno-genetic progeny of females No. 1, No. 2 and No. 5 died before the feeding stage. Only one gynogen was exceptionally survived in the progeny of female No. 4.
Fig. 3
External appearance of loach progeny developing from induced gynogensis (a, b) and interfamilial hybridization (c). a; abnormal loach larvae, gynogenetically induced from eggs of besexually reproducing females (Nos. 1, 2, 4 and 5) by fertilization with UV-irradiated sperm, b; normal loach larvae, gynogenetically induced from unreduced eggs of putative unisexual females (Nos. 3 and 6), c; abnormal hybrid larvae from the hybridization between loach female and goldfish male.
Table 2
Developmental potential and relative DNA content of the progenies from normal fertilization and induced gyno-genesis of the eggs from six females caught in Memanbetsu, Hokkaido.
Flow cytometry for relative DNA content of surviving fry developing from gynogenetically activated eggs of the two females (No. 3 and No. 6) disclosed that these fry were diploids (Table 2). Thus, the eggs giving rise to viable gynogen even after fertilization with genetically inert sperm are unreduced diploids. In the normal fertilization experiments, female No. 1, No. 2 and No. 4 gave only diploid progeny only with 2C DNA content after normal fertilization. Thus, they generated regular haploid eggs. Females which gave viable diploid gynogens after fertilization with UV-irradiated sperm (No. 3 and No. 6) should give triploids when fertilized with normal sperm. However, female No. 3 gave 93% (65/70) diploid offspring and female No. 6 gave 15% (3/20) diploid offspring in the control crosses (Table 2). Female No. 3 gave 6% (4/70) and female No. 6 gave 60% (12/20) triploid progeny. Both females also produced diplo-triploid, haplodiplo-pentaploid and 2.5n-3.5n aneuploid mosaic progeny. The occurrence of diploid progeny suggests possible uni-sexual development in some eggs of the two unique females No. 3 and No. 6.
Interspecific hybridization
To verify unisexual development of the unreduced eggs, ovulation was induced again in female No. 3 and the eggs were fertilized with functional and with UV-irradiated goldfish sperm. At the same time, eggs of the other two females (No. 7 and No. 8) were investigated for comparison.
As shown in Table 3, normal hatching fry occurred in the gynogenetic progeny of female No. 3 and 86% (51/59) of them were diploid. However, one triploid, five tetraploids (Fig. 4) and two about 2.5n heteroploids were also observed. In contrast, other two females, No. 7 and No. 8, gave only haploid (1C DNA content) fry after induction of gyno-genesis. These results reveal that female No. 3 generated unreduced diploid eggs again, but the other spawned haploid eggs.
Table 3
Developmental potential and DNA content of the progenies from normal fertilization, induced gynogenesis, and hybridization of eggs from three loach females.
Fig. 4
Flow cytometric histograms for relative DNA content of somatic cells and spermatozoa, when erythrocytes of normal diploid loach are used as standard of 2C. Top to bottom; 1C DNA content of loach spermatozoa, 1.4C DNA content of goldfish spermatozoa, 2C DNA content of diploid loach, 2.4C DNA content of the hybrid between normal eggs of loach female No. 8 and spermatozoa of goldfish male, 3.4C DNA content of the allotriploid hybrid between unreduced eggs of loach female No. 3 and spermatozoa of goldfish male, and 4C DNA content of spontaneous tetraploid fry developing from induced gynogenesis of unreduced eggs of loach female No. 3 after fertilization with UV-irradiated spermatozoa.
In hybridization experiments (Table 3), female No. 3 gave 65.2% (30/46) normal loach fry, whereas only abnormal fry appeared from the eggs of females No. 7 and No. 8 (Table 3, Fig. 3c). Flow-cytometry for relative DNA content showed that 48% (13/27) fry from female No. 3 gave just 2C DNA, equivalent to diploid loach, whereas almost all the other fry had 3.4C DNA content (Fig. 4). A small number of aneuploid mosaics were detected. Since the relative DNA content of goldfish sperm was approximately 1.4C in contrast to 1C of the loach sperm (Fig. 4), the fry with 3.4C were allotriploid hybrids possessing two genomes (diploidy, 2C) from the loach and one genome (haploidy, 1.4C) from the goldfish. Abnormal fry developing from the eggs of female No. 8 had 2.4C DNA content in most cases (Fig. 4), thus being diploid hybrids possessing one loach genome (haploidy, 1C) and one goldfish genome (haploidy, 1.4C). Exceptionally, one fry was a mosaic with loach haploid cells and loach-goldfish hybrid cells. These results clearly indicate that some eggs of the female loach No. 3 developed without any genetic contribution from the father fish, but other eggs did incorporate DNA from the sperm nucleus in the process of fertilization.
Microsatellite genotyping
To confirm unisexual development of some unreduced eggs of the females No. 3 and No. 6, genetic transmission of microsatellite markers was analyzed. As shown in Fig. 5a and Table 4, alleles at the microsatellite loci (primer sets) segregated according to Mendelian expectations in the normal diploid progeny of normal loach female No. 1. Mendelian segregation of the six microsatellite loci used in the present study has already been confirmed in some diploid families by Morishima et al. (2001). Thus, in that study, diploid progeny of the normal loach females had both maternally and paternally derived alleles in each genotype after normal fertilization, revealing bisexual reproduction.
Fig. 5
Microsatellite genotypes in the normally fertilized and gyno-genetically induced progeny of the loach. a; Mac3 genotypes among diploid progeny (Nos. 1–8) developing from normally fertilized eggs of female No. 1, b; Mac37 genotypes among UV-induced gynogenetic diploid progeny developing from unreduced eggs of female No. 3 (left, Nos. 1–4) and No. 6 (right, Nos. 1–5), c; Mac37 geno-types among triploid (Nos. 1–4) and diploid (Nos. 5–10) progeny developing from normally fertilized unreduced eggs of female No. 3.
Table 4
Microsatellite genotypes in the progenies of normal fertilization and induced gynogenesis using the eggs of loach females No.1, No.3 and No.6.
In UV-induced diploid gynogens developing from putative unreduced eggs of females No. 3 and No. 6, only microsatellite genotypes identical to the mother fish were detected at each locus (Fig. 5b, Table 4), clearly indicating the unisexual transmission of maternal genes. Among the progeny developing from the unreduced eggs after normal fertilization with loach sperm, diploid individuals gave unipa-rental genotypes identical to the mother, whereas triploid individuals showed genotypes comprising both two alleles from the mother and one allele from the father (Fig. 5c, Table 4). These genetic results indicate that some unreduced eggs of females No. 3 and No. 6 developed unisexually without any DNA contribution from the sperm nucleus, but other eggs incorporated the sperm genome, probably accidentally. Most unreduced eggs are initially diploid as observed in the genotypes of artificially induced gynogenesis.
Interestingly, microsatellite genotypes at the six loci examined were completely identical between the diploid sib-progeny and the mother loaches (No. 3 and No. 6) as shown in Table 4. This indicates that the unreduced diploid eggs of such females develop unisexually as well as clonally. Identical microsatellite genotypes between female No.3 and No.6 also suggest that these two females may be a member of a single cryptic diploid clonal line which reproduces uni-sexually.
Multilocus DNA fingerprinting
So as to confirm the clonal nature of the progeny, genetic identity was then examined in a wide range of loach genomes by means of the multilocus DNA fingerprinting technique. DNA fingerprints detected with (GGAT)4 probing after Hae III digestion depicted that the diploid progeny (n=4) were genetically identical and absolutely clonal to the mother fish (female No. 3) without any contribution from the alleged father (Fig. 6). DNA fingerprints in the triploid progeny included some of the fragments from the father together with all the fragments possessed by the mother (Fig. 6). DNA fingerprinting with a different probe 33.15 in the same Hae III digested samples also gave similar results (figure not shown). Thus, the clonal nature of the diploid eggs is proven and the occurrence of triploid progeny can be explained by genomic incorporation of sperm nuclear DNA. DNA fingerprinting of the diploid and triploid progeny of No. 6 gave the same conclusion.
Fig. 6
Multilocus DNA fingerprints hybridized with (GGAT)4 probe in Hae III digested samples from diploid (lanes 5–8) and triploid (lanes 1–4) progeny developing from the fertilization of unreduced eggs of female No. 3 with sperm of male A. Triangles denote sperm DNA.
Identical DNA fingerprints between females No. 3 and No. 6 suggest the existence of a clone in the wild population of Memanbetsu town. Finally, 17 diploid individuals including females No. 3 and No. 6 and 8 triploid individuals from the wild population of the same town were genetically assayed in 2000 to identify clonal individuals by DNA fingerprinting. As shown in Fig. 7, clonal individuals were easily detected by their absolutely identical DNA fingerprints. Five out of 17 diploids examined had an identical pattern and therefore were considered to be a clonal line. At least one clonal line exists in this population.
Fig. 7
Multilocus DNA fingerprints hybridized with (GGAT)4 probe in Hae III digested diploid (lanes 1–9) and triploid (lanes 10–12) loach samples collected from the Memanbetsu population, Hokkaido, Japan. Note the identical fingerprints among loach in lanes 1, 8 (=female No. 3) and 9 (=female No. 6) and close similarities between clonal diploids (lanes 1, 8 and 9) and natural triploid individuals (lanes 10–12).
In the DNA fingerprints probed with (GGAT)4 and probe 33.15, ranging between 6 and 9 kilobase pairs, 6 to 20 (mean 15.3) and 6 to 18 (mean 12.3) fragments were counted, respectively. BSIs analyzed in five clonal diploids, 12 aclonal common diploids and 8 triploids are shown in Table 5. BSIs within clonal line were 1.0 for both probes, while those within common diploid loaches were 0.40 for (GGAT)4 and 0.44 for 33.15 probe. BSIs between clonal and aclonal diploids were 0.28 for (GGAT)4 and 0.25 for 33.15 probe, suggesting a genetically remote relationship between the two.
Table 5
Band sharing indices(BSIs) in DNA fingerprints probed with (GGAT)4 and 33.15 inHae III digested loach samples.
DNA fingerprinting revealed that all the natural triploid individuals examined (n=8) were not genetically identical, because each individual exhibited a specific pattern as shown in Fig. 7. However, all triploids were genetically very similar to each other as indicated by high values of BSIs between triploids, 0.93 for (GGAT)4 and 0.84 for 33.15 probe (Table 5). BSIs between triploids and clonal diploids were high (0.96 for (GGAT)4 and 0.89 for 33.15 probe), but those between triploids and aclonal diploids are very low (0.31 for (GGAT)4 and 0.32 for 33.15 probe). This indicates that natural triploids in this area may be generated by fertilization of clonal diploid eggs with normal haploid sperm.
DISCUSSION
The appearance of viable diploid progeny after UV-induced gynogenesis without any additional treatment to duplicate the chromosome set of the eggs shows that some diploid loaches in the Memanbetsu population, Hokkaido, Japan, generate unreduced diploid eggs. Surprisingly, a considerable number of diploid individuals occurred among the progeny when such unreduced eggs were fertilized with normal haploid loach sperm, although only triploid individuals were predicted. Since viable diploid loach fry were produced in female loach x male goldfish crosses, and only maternal microsatellite genotypes were observed in the diploid progeny of female loach x normal male loach crosses, unisexual development of unreduced eggs is the most likely explanation for the appearance of diploid progeny in these crosses. This is confirmed by DNA fingerprinting which revealed that each diploid fry was genetically identical to its diploid mother. In conclusion, two diploid females (No. 3 and No. 6) and their diploid progeny appear to belong to a clonal line that reproduces unisexually.
All-female inheritance found in these loaches is likely to be gynogenesis because UV-treated sperm always triggered such reproduction in all the unreduced eggs. However, normal sperm was sometimes incorporated into the zygotic genome, probably after its accidental pronucleation as discussed below. In teleosts, unisexual development is exclusively gynogenesis (embryogenesis initiated by sperm stimulation without syngamy) and no parthenogenesis (all-female reproduction in the absence of sperm) has been found (Dawley 1989; Vrijenhoek et al. 1989). In the cases of loaches studies here, it is not yet established whether uni-sexual development is simply gynogenetic or absolute parthenogenetic.
The occurrence of unreduced diploid eggs was previously found in the diploid loach from the Hirokami village, Niigata prefecture, Honshu island, Japan in which a high incidence of natural triploids was recorded (Zhang and Arai, 1999). In that case, the diploid loach laid both normal haploid and unreduced diploid eggs at the same time and both diploid and triploid fry were observed in a normally fertilized batch. Thus, the unreduced eggs of the loach from Niigata are different from those from Hokkaido; the former develop bisexually and the latter reproduce unisexually. All the unreduced eggs examined in the present study were genetically identical to the mother and no genetic variation was found among the samples examined.
The cytological mechanism leading to unreduced oogenesis is unknown in the clonal diploid loach observed in the present study. However, two major systems, apomixis and premeiotic endomitosis, may explain such a phenomenon as reported in other polyploid and unisexual vertebrates (Dawley, 1989). Apomixis is the mechanism in which abortive meiosis I inhibits synapsis of homologous chromosomes that never leads to subsequent recombination and segregation. Thus, mature eggs are formed by somatic division. Apomictic oogenesis has been well documented in the trip-loid silver crucian carp Carassius auratus gibelio (Cherfas, 1966), its related Japanese subspecies Carassius auratus langsdorfii (Yamashita et al., 1993) and the amazon molly Poecilia formosa and its triploid hybrids (Monaco et al., 1984). Premeiotic endomitosis is a process which results in the doubling of chromosomes without cytokinesis before meiosis, followed by two successive divisions in a quasinormal manner. In the loach, unreduced triploid eggs are formed in triploid (diploid female x tetraploid male) individuals by this system (Zhang et al., 1998). A similar mechanism has been found in triploid Poeciliopsis (Cimino, 1972) and artificial medaka hybrids between Oryzias latipes and O. curvinotus (Shimizu et al., 2000). In this system synapses and recombination occur essentially between the sister replicate chromosomes produced by the endomitosis. This gives rise to no genetic variation, and the genetic composition of the resultant unreduced eggs is theoretically identical to that of the mother. As the mechanism of the formation of unreduced diploid eggs has not been identified in this study, further cytological studies are required in near future.
When the unreduced diploid eggs laid by the Memanbetsu loaches were fertilized with normal sperm, some eggs unisexually (probably gynogenetically) develop into clonal diploids, while the others develop into triploids incorporating nuclear DNA of the sperm. These observations imply that these loaches are not exclusively gynogenetic. Accidental incorporation of sperm nuclear DNA has also been reported in some other unisexual species. In gynogenetically reproducing triploid crucian carp C. a. gibelio in China, a few multiple tetraploid individuals possessing the triploid genome of crucian carp and the haploid genome of common carp occurred (Gui et al., 1993). In Japanese silver crucian carp C. a. langsdorfii, artificial tetraploids were produced by the incorporation of a haploid sperm genome into triploid eggs at high or low temperature (Takai and Ojima, 1983; Dong et al., 1997; Mada et al., 2001). Similar sperm nuclear DNA incorporation at high temperature was also reported in polyploid hybrid salamanders of the genus Ambyostoma which produced unreduced eggs and reproduced gynogenetically at low temperature (Bogart et al., 1989; Elinson et al., 1992). The fact that the two clonal females No. 3 and No. 6 gave different frequencies of triploids after sperm nuclear incorporation suggests that the occurrence of gynogenetic or bisexual reproduction depends on environmental factors at the time of fertilization and subsequent early embryogenesis.
Microsatellite genotyping and DNA fingerprinting reveal the presence of at least one clonal line among the Memanbetsu loaches: a relatively large proportion (29.4%, 5 /17) of diploid loaches appear to be members of the same clone. The distribution and frequency of such clonal lines have not been surveyed in other parts of Hokkaido, but their existence is suggested by the high incidence of triploids. Most unisexual fish species have been shown to comprise several clonal lines as evidenced in natural populations of silver crucian carp (Dong and Taniguchi, 1996; Umino et al., 1997; Ohara et al., 1998, 1999, 2000), Poeciliopsis (Quattro et al. 1992ab), Rutilus (Alves et al. 1998) and other species reviewed by Dawley (1989). Thus, further genetic analyses are necessary to disclose the number, distribution and diversity of clonal lineage(s) among the loaches in Hokkaido. Apparently low BSIs between clonal and aclonal diploid loach lineage indicate that they are dissimilar, suggesting a remotely distant genetic relationship between them.
Sympatric triploids may be the consequence of sperm nuclear incorporation into the clonal diploid eggs. Thus, although all the triploids are genetically similar, they are not identical due to the genetic variability of each sperm nucleus incorporated, as indicated by the various DNA fingerprints observed. High BSI values within triploid groups and those between clonal diploids and triploids also support this conclusion.
It is believed that unisexuality and polyploidy are closely associated with hybridization between bisexual species and subspecies (Dawley, 1989; Vrijenhoek, 1989). A new uni-sexual lineage can be established if genetic differences between two hybridizing species are sufficient to disrupt the meiotic mechanism (Vrijjenhoek, 1989). The loach M. anguillicaudatus is considered as a single species entity in Japan, but recent population genetic studies using allozymes suggest the presence of at least three genetically different groups with subspeces level differentiation (Khan and Arai, 2000). Two biologically distinct types were also recognized based on genetic and morphological data in the loach specimens from Ibaraki prefecture (Dong et al., 1999). These results suggest the possible existence of some cryptic bisexual species in the loach. An unusually high incidence of triploids was found in Memanbetsu population in the present study. Zhang and Arai (1999) found high frequencies of triploids in Hirokami village, Niigata and Ichinomiya town, Aichi prefecture. These three localities seem to be in a zone between genetically distinct groups shown by Khan and Arai (2000). Possible hybridizations between them may cause ameiotic reproduction in the hybrids with the unbalanced genomes. Further investigations are required before the origin and genetic mechanisms incolved in the occurrence of unisexual reproduction and natural trip-loidy in the loach can be conclusively determined.
Acknowledgments
We gratefully acknowledge Prof. N. P. Wilkins, National University of Ireland, galway for his critical reading of our manuscript.
