Strains

The lf locus was found and isolated by Tomita in the offspring of the wild medaka population collected at Toyokawa, Aichi prefecture, Japan in 1970 and first demonstrated as an autosomal locus (Tomita, 1992). Homozygotes of the lf (lf/lf) have no leucophores throughout the life. In wild-type homozygotes (+/+) and heterozygotes (+/lf), the yellow colored leucophores appear beneath the brain at stage 25 and then develop on the body surface at stage 26 (Tomita, 1992; Iwamatsu, 1994). Table 1 shows the strains used in this experiment. The strains HNI (genotype for the lf locus: +/+; Hyodo-Taguchi, 1980) and Sabae (+/+; Shima et al., 1985) were derived from the northern Japanese population of the medaka fish while the others, Hd-rR (+/+; Hyodo-Taguchi, 1980) and AA2 (lf/lf; Shimada and Shima, submitted) were from the southern population. The two populations were well demonstrated to be genetically much different (Sakaizumi et al., 1983) though the hybrids and backcross progeny among the populations can be obtained (Naruse and Shima, 1989). The Qurt strain was newly established from a domesticated stock of the southern population for this study. The derived stock was originally maintained to obtain a tester strain for the specific locus test system in the medaka fish (Shima and Shimada, 1991). One male with genotype b/b, lf/+, gu/+, and r/r and a female with b/b, lf/lf, gu/gu, and r/r were selected from the massmated stock and mated (b, colorless melanophore and gu, guanineless are autosomal recessive body color mutations; Shima and Shimada, 1991). A male progeny with b/b, lf/+, gu/gu, and r/r was selected and backcrossed with the female-parent. Preliminary results showed that the males had leucophores but the females did not have leucophores in the population (genotypes were considered to be b/b, lf/lf, gu/gu, and r/r in females and b/b, lf/+, gu/gu, and r/r in males). The stock thus obtained was named Qurt and used in this study.

Table 1

The medaka strains used in the present study

Genetic crosses

Following crosses were carried out in this experiment; Cross 1: two AA2 females were mated with two (AA2 female X HNI male) F₁ males, Cross 2: an AA2 female was mated with a (AA2 female X Sabae male) F₁ male, Cross 3: an AA2 female was mated with a (AA2 female X Hd-rR male) F₁ male, Cross 4: two Qurt females were mated with two Qurt males, Cross 5: an AA2 female was mated with a (HNI female X AA2 male) F₁ male. The fish used for the five crosses were kept in plastic cages (34 × 17 × 13cm) in aged tap water at 27 ± 2°C under 14-hr light/10-hr dark photo period cycle. The fish were fed with TetraMin (Tetra Werke, FRG).

Photomicroscopy

The phenotypes of embryos were photographed by transmitting light and epi-illuminating fluorescent light using an Olympus microscope, BX50 with a BX-FLA fluorescence system (using a cube set, U-MWU).

Closely sex-linked inheritance of the lf locus in the five independent crosses

Table 2 summarizes the results of the crosses. In the five independent crosses, a total of 238 males and 226 females was obtained. The sex ratio was reflected the Mendelian 1:1 ratio (χ² value was 0.31, 0.5<P< 0.6). The phenotypes for the lf locus of the progeny were 250 wild types (supposed to be heterozygotes, lf/+) and 214 mutant types (supposed to be homozygotes, lf/lf). Thus the lf locus was also reflected the Mendelian 1:1 segregation ratio (χ² value was 2.79, 0.10 < P < 0.15). In all of the five crosses, the lf locus was closely sexlinked and there was not significant difference in the recombination frequency among the crosses (the recombination frequencies were ranging from 0 to 3.8% and the maximum χ² value was 3.08, 0.3<P< 0.4). Altogether, 10 recombinants (five males and five females) were obtained in 464 backcross progeny. These recombinants were fertile and produced progeny normally. The average recombination frequency between the lf locus and the male determining factor (y) was 2.2% (90% upper and lower confidence limits were 3.6 and 1.3%, respectively).

Table 2

The recombination frequencies between the lf and y loci in the 5 independent backcrosses

Expression of leucophores in early developmental stages

Figure 1 shows the photographs of embryos of the Hd-rR strain which carries wild type alleles for the lf locus homozygously (+/+). The leucophores were not visible in a 1-day embryo at the late neural stage by a transmitting illumination (Fig. 1a) nor by an epi-illumination fluorescent condition (Fig. 1b). However, in a 2-day embryo when the heart started beating, the leucophores were observable beneath the brain by a transmitting illumination (Fig. 1c) as well as by an epi-illumination fluorescence condition (Fig. 1d). In a 4-day embryo, when retinal pigmentation and optic vesicle formation were evident, the leucophores were more easily observable (Fig. 1e, f). Figure 2 shows the photographs of embryos of the Qurt strain. The half of the embryos have wild type leucophores (the genotype was supposed to be lf/+), although the rest do not have visible leucophores at all (the genotype was supposed to be lf/lf). Most of the phenotypically wild type embryos developed to males while the mutant type embryos developed to females as shown in the experiment of genetic crosses.

Fig. 1

Photographs of embryos of the Hd-rR strain (wild type +/+). (a,b) 1-day embryo (26 hr after fertilization, 27°C), (c,d) 2-day embryo (50 hr), (e,f) 4-day embryo (74 hr). (a,c,e) photographed by transmitting illumination, (b,d,f) photographed by epi-illuminating fluorescent light. Arrows show leucophores. Scale bar: 500 μm.

Fig. 2

Photographs of 2-day (30 hr) embryos of the Qurt strain; the heterozygous lf/+ fish on the left and the mutant lf/lf fish on the right. (a) photographed by transmitting light, (b) photographed for autofluorescence. Leucophores (arrows) are present in the heterozygous lf/+ fish but absent from the mutant homozygous lf/lf fish. Scale bar: 500 μm.

The sex-linked inheritance of the lf locus

The lf locus was firstly demonstrated as an autosomal locus (Tomita, 1992), however, in the present study, the closely sex-linked inheritance was demonstrated in all of the five crosses. We observed recombinations between sex chromosomes derived from different strains and populations. In Cross 1, X chromosomes from the AA2 strain (abbreviated to AA2X) and Y chromosome from the HNI strain (HNI-Y); Cross 2, AA2-X and Sabae-Y; Cross 3, AA2-X and Hd-rR-Y; Cross 4, Qurt-X and Qurt-Y; and Cross 5, HNI-X and AA2-Y. Although the HNI and Sabae strains were genetically much different from the other stains, the backcross progeny were obtained with normal 1:1 sex ratio. A total of 130 males and 137 females was obtained in Cross 1, 2 and 5 (χ² value was 0.18, 0.6<P< 0.7). In the crosses using southern strains, 108 males and 89 females were obtained (Cross 3 and 4, χ² value was 1.83, 0.15<P<0.17). These results indicated that all the strains used in this study shared the same sex determination system (i.e., the same sex chromosome).

Expression of the leucophores in wild type and mutant type embryos were the same as originally described (Tomita, 1992). In the Tomita's experiment, an exceptional autosomal inheritance might have been detected. Such inheritance could be explained by chromosomal rearrangements (translocation) or high recombinations between sex chromosomes (pseudoautosomal inheritance).

Prediction of sex of each embryo in early developmental stages by using the expression of leucophores

As shown in the photographs, expression of wild type leucophores began on two days after fertilization, while in the mutant lf/lf homozygotes, visible leucophores were not observed at all. Although a question still remained to be addressed whether the pigment cells were colorless or the cells themselves were absent, the lf locus should be a very useful genetic marker for prediction of sex in each embryo because it is very easy to distinguish lf/lf phenotype from lf/+ pheno type. For example, in the Qurt strain where the Y chromosomes carry the wild type lf allele while the X chromosomes carry the mutant type lf alleles, males (X^lf/Y⁺) always have leucophores but females (X^lf/X^lf) do not have leucophores as far as X and Y chromosomes do not recombine. Because the recombination frequency between lf and y was 2.2%, we can predict sex of each embryo with approximately 97% confidence in this strain. In the medaka fish, differentiation of gonads begins in 5-day embryos just before hatching (Satoh and Egami, 1972). Because the phenotypic expression of the lf locus is much earlier than the gonadal differentiation, this genetic marker is a powerful tool for studying sex determination or sex differentiation mechanisms in the medaka fish.

The first sex-linked color locus demonstrated in the medaka was the r locus (Aida, 1921). The recombination frequency between r and y was estimated to be 0.2% (Yamamoto, 1964). Experiments using double recessive mutant strains to determine gene orders for the r, lf and y loci are now under way.