Recent advances in molecular biology on pigmentation of the medaka, Oryzias latipes*

In medaka, numerous color variants have so far been discovered, segregated and maintained as homozygous strains (Tomita, 1975), providing excellent materials to study cell biology, embryology and genetics on pigmentation of skin and eyes. Using these variants, much effort has been made to characterize the pigment cells, as well as their hereditary traits, which ascribe pigmentation to the fish, mostly using traditional methods in cell biology (See Obika in this issue). Since the early 1990s, studies on the genes relevant to pigmentation of medaka have been conducted in two ways: one is to determine the causes of albinism by cloning the tyrosinase gene from the genome of this species (Inagaki et al., 1994; Koga et al., 1994, 1995a), and the other is to analyze changes in pigmentary traits by introducing well-known pigmentation-associated, heterologous genes such as the mouse tyrosinase gene into the medaka genome (Matsumoto et al., 1992b).

Successful cloning of the tyrosinase gene from medaka and its homology to the mammalian counterpart

Hori and his associates (Inagaki et al., 1994; Koga et al., 1994) first cloned tyrosinase cDNA and then genomic DNA for tyrosinase from the wild type (BIR) medaka, using authentic DNA probes specific to this gene. Inasmuch as cloning and structural analyses of the mammalian tyrosinase gene had previously been completed (Kwon et al., 1987; Yamamoto et al., 1987; Beermann et al., 1990), the availability of such probes facilitated prompt cloning of the gene from medaka. Analyses of the structure of the medaka tyrosinase gene thus cloned disclosed that its homology with mammalian ones is in a magnitude of 60% in terms of amino acid sequences (Inagaki, et al., 1994; Koga et al., 1994). Further cloning of the tyrosinase gene from the orange-colored variant (bIR) of medaka and subsequent determination of its structure also disclosed that the gene obtained is essentially the same in base sequence as that of the wild type (Inagaki et al., 1994; Koga et al., 1994). This clearly indicates that defective melanogenesis in the skin of this color variants is not derived from mutation or defects of the tyrosinase gene itself. This finding would also support classical cytochemical observations that melanin deposition is inducible in amelanotic melanophores of this color variant by adding a sulfhydryl-blocking agent under a sufficient supply of dihydroxyphenylalanine (DOPA) (Tomita and Hishida, 1961). Ubiquitous deposition of melanin into the eyes of this variant can also be reasonably explained by this finding.

Structural alteration of tyrosinase gene in albino medaka by insertion of a transposone-like element

By applying the technique established for studies of the medaka tyrosinase gene, Hori and his associates then disclosed structural alteration of this gene in spontaneous albinos i (BiR) (Inagaki et al., 1994; Koga et al., 1994, 1995a) and i-4 (Koga et al., 1995b). Comparison of the structure of the tyrosinase gene obtained from the albino i with that from the wild type indicates that exon 1 of the former is approximately 1.9 kb larger than that of the latter. Restriction mapping of tyrosinase alleles disclosed that the distance between the ApaI site and the EcoRV site is 0.3 kb in exon exon 1 from the wild-type clone, whereas it is 2.2. kb in the albino i. Thus, the albino i locus carries an extra 1.9 kb DNA fragment inside exon 1 of the tyrosinase gene. Inasmuch as this albino is unable to produce melanin throughout its entire life, it is quite probable that insertion of this DNA fragment into exon 1 incapacitates the synthesis of functional tyrosinase, thus preventing melanin formation.

Similar analyses of the tyrosinase gene from albino i-4 also disclosed that an extra DNA fragment is inserted in exon 5 of the gene. This albino has black eyes throughout its life, thus indicating autonomy for the production of tyrosinase capable of melanogenesis, at least in the eyes. All these findings would imply that exon 1 of the gene is critical for expression of tyrosinase activity, whereas exon 5 plays a supplemental role in synthesis of the functional tyrosinase, possibly yielding a truncate polypeptide which is inert in skin melanophores. In view of the apparent rescue of the albino phenotype in the albino i by introduction of an intact tyrosinase gene cloned from the wild type by means of microinjection into fertilized eggs (Koga et al., personal communication), it seems probable that the inability in melanogenesis in these albinos is derived primarily from structural alteration of the gene by insertion.

Inasmuch as most oculo-cutaneous albinism in mammalian species is caused by point mutation in the exons of the tyrosinase gene (King et al., 1995), it is noteworthy that albinism in medaka, as far as the albinos examined so far are concerned, is caused by a mechanism different from that in mammals. Since the fragment inserted in the tyrosinase gene of the albinos resembles transposable elements with regard to its structure, it can be postulated that albinism in medaka is caused by structural changes of the tyrosinase gene by insertion of a transposon-like fragment in the exon 1 rather than a defect in the cellular vehicles for melanin deposition. The possibility still exists however, that point mutation in the tyrosinase gene also causes albinism in medaka as in mammals.

Induction of melanogenesis by introduction of mouse tyrosinase gene into the orange-colored variant of medaka

Another line of study to elucidate the molecular basis of skin pigmentation in medaka was to generate transgenic medaka bearing tyrosinase genes using non-melanized color variants. In 1987, Takeuchi and his associates succeeded in cloning a tyrosinase gene from the mouse (Yamamoto et al., 1987) and then constructed the functional mini-gene, mg-Tyrs-J (4.5 kb) by fusing the 5' -flanking genomic non-coding sequence, which includes the regulatory region, the complete coding region of the cDNA, and its 3' -flanking sequence (Yamamoto et al., 1989). We began to produce transgenic fish using this constructed DNA which is heterologous to medaka. Such transgenic fish were expected to provide a useful, " gain-of-function" type animal model for studies of the integration, expression and transmission of pigmentation-associated genes. It is also expected that the use of the heterologous gene would facilitate discrimination of its expression from that of the genomic counterpart, provided that specific antibodies to distinguish the two are available.

In fish, including medaka, it is well established that skin pigmentation is composed of three to four different sorts of chromatophores, vis., melanophores, xanthophores or erythrophore, leucophores and/or iridophores (Bagnara and Hadley, 1973), and that despite their common ontogenic origin in the neural crest, the pigment composition of these cells after terminal differentiation differs completely with respect to its molecular species, indicating multiple differentiation toward different phenotypes (Bagnara et al., 1979). Developmental studies on pigmentation of transgenic fish carrying mouse tyrosinase genes would shed slight on the role of the tyrosinase gene in the differentiation of chromatoblasts to melanophores.

Successful production of transgenic medaka bearing the gene for mouse tyrosinase

In 1990, production of transgenic fish using three color variants of medaka, albino i (BiR), albino i-3 and orange-colored variant (bIR) was begun. In the initial phase of these studies, a linear form of the transgene was microinjected into the pronucleus of matured oocytes, followed by insemination according to the procedure developed by Ozato et al. (1986). Mainly due to low fertility of the strains used, we failed to yield transgenic fish from the albinos. Eventually, we succeeded in obtaining transgenic fish from the orange-colored variant. Of 64 oocytes microinjected, 13 embryos grew normally beyond hatching and three of them exhibited brownish skin, which was proved to be due to distribution of melanized melanophores (Matsumoto et al., 1992b). A variation in the degree of melanization was observed in the skin among these three founders, one being melanized heavily whereas the remaining two were weakly melanized. Crossing this heavily melanized female founder with the homozygous normal orange-colored variant male yielded progeny with melanized and non-melanized skin at an approximate ratio of 1 : 1 (Matsumoto et al., 1992b). This indicates that the transgene of the murine origin is capable to transmitting in medaka through the germ line, presumably according to the Mendelian law of hereditary.

Due to low fertility and limited availability of transgenic fish produced, we failed to expedite these studies further. Subsequently, we succeeded in producing similar transgenic fish efficiently by use of electroporation (Matsumoto et al., 1992a). This has enabled us to generate 16 founders exhibiting brownish body coloration from the orange-colored variant of medaka (Fig. 1), though no melanogenic founders were obtained from albino i-3.

Crossing of the transgenic founders (females) thus obtained with the homozygous orange-colored variant fish (males) yielded offspring with melanized and non-melanized skin at a ratio of 1 : 1 (Matsumoto et al., 1992a). Crossing between different transgenic founders yielded offspring with melanized skin only, some of which assumed black body coloration comparable to the wild type fish of this species (Fig. 1). Further crossing between different lines of transgenic F1 or F2 yielded heavily melanized progeny, suggesting increased degrees of melanization in the skin with increased copy numbers of the transgene. Recent PCR assays using a probe discriminating the mouse tyrosinase gene disclosed unequivocally that the transgene is present in transgenic F1 and F2 fish (unpublished data). Thus, transmission of the transgene over three generations according to Mendelian law has been decisively confirmed using a new line of transgenic fish thus generated (Matsumoto et al., 1992a).

Recognition of melanogenesis-competent cells by mouse tyrosinase gene in transgenic medaka

Immunocytochemical assay of skin tissues from fully grown, melanized transgenic F1 fish using the polyclonal antibody against mouse tyrosinase (Yamamoto and Takeuchi, 1981), which was not cross-reactive to medaka tyrosinase, disclosed that melanophores are unequivocally reactive, indicating expression of the transgene (Matsumoto et al., 1992a, b) (Fig. 2). Reactivity of this antibody was specific to melanophores but not to xanthophores and leucophores (Matsumoto et al., 1992a, b). This implies that the transgene in medaka is expressed in a cell-type specific manner which results from recognition of melanogenesis-competent cells by the regulatory sequence of the mouse tyrosinase gene.

Skin melanophores in these transgenic fish varied widely with respect to the degree of melanization, some being melanized heavily and others quite faintly even in the same limited area of the skin (Ono et al., 1994). Such variation was never observed in the skin of the wild-type adult fish. Reactivity in an immunocytochemical assay tends to be much stronger in weakly melanized melanophores than heavily melanized ones. This could be explained by a well-known reciprocal relationship between tyrosinase activity and melanin deposition during maturation of melanophores (Seiji and Fitzpatrick, 1961). Upon careful comparison of their cell shape, population density and distribution pattern, it is considered that melanized melanophores present in the skin of transgenic F1 fish correspond to amelanotic melanophores, which are abundant in the orange-colored variant of this species (Hishida et al., 1961). It may be concluded that, whatever the biochemical mechanism may be, mouse-type tyrosinase allows melanogenesis to proceed under the inhibitory influence presumed to exist in melanophores of the orange-color variant of medaka.

Occurrence of mammalian-type melanosomes in transgenic medaka carrying mouse tyrosinase gene

Electron microscopy of the skin of transgenic F1 showed that the melanophores are laden with numerous melanosomes, most of which are installed with a peculiar internal structure (Matsumoto et al., 1992b) (Fig. 3). The ultrastructural feature of these melanosomes apparently differs from that of skin melanophores found in the wild type of this species, particularly in the adult form. The presence of highly organized lattice-like internal structures strongly suggests that the manner of melanogenesis is much closer to that of mammals. In medaka, it is well established that melanogenesis of the melanosomes is conducted mostly by participation of multivesicular particles (Nakajima and Obika, 1986), and that the occurrence of lattice-like structures at an intermediate stage of maturation is rather rare, even though similar internal structures are often seen in the melanophores appearing at the embryonic to larval stages.

In mammals, the pheomelanosomes are formed from spherical, multivesicular particles, whereas the eumelanosomes are formed from elliptical, lattice structure-based particles (Prota, 1992). In view of recent findings on the switching mechanisms between eumelanin and pheomelanin formation in agouti mice (Barsh, 1996), it is likely that combinations of the genes under operation would determine pigmentary composition and ultrastructural pattern of melanosomes. In medaka, excess production of mammalian-type tyrosinase would force the internal structure of melanosomes to assume the form resembling that of mammalian eumelanosomes.

Perspective

With regard to the distribution pattern of skin melanophores in the transgenic medaka deriving from the orange-colored variant, at least two distinct types are recognized: some transgenic lines present a striped or spotted pattern while others have evenly distributed melanophores, as in the wild type. Such fish lines would provide a novel animal model for molecular studies of pigment pattern formation. In crossing transgenic fish, we have noticed that among the progeny, there appears a unique color variant which is deficient in xanthophores. Since it is likely that a certain portion of color variants in medaka is caused by insertion of a transposable element such as that discovered by Hori and his associates, it is interesting to speculate that the xanthophore-deficient color variant may have resulted from a mechanism like insertion mutation upon introduction of the transgene.

Acknowledgement

The authors wish to thank Drs. H. Yamamoto (Tohoku University), T. Akiyama (Keio University), K. Miyazaki (Kyoto University) for their productive collaboration in the studies reviewed herein. Particular thanks are also due to Dr. T. Takeuchi (Ishinomaki-Senshu University) who passed away during the course of this study. This study was supported in part by a research grant from Keio University.

References

Bagnara, J.T. and M.E. Hadley (1973) In: Chromatophores and Color Changes. Prentice-Hall, Englewood Cliffs, New Jersey.

Bagnara, J.T., J. Matsumoto, W. Ferris, S.K. Frost, W.A. Turner, Jr., T.T. Tchen and J.D. Taylor (1979) Science, 203: 410- 415.

Barsh, G.S. (1996) Trends Genet., 12: 299- 305.

Beermann, F., S. Ruppert, E. Hummler, F.X. Bosch, G. Muller, U. Ruther and G. Schuts (1990) EMBOJ., 9: 2819- 2826.

Hishida, T., H. Tomita and T. Yamamoto (1961) Embryologia, 5: 335- 346.

Inagaki, H., Y. Bessho, A. Koga and H. Hori (1994) Gene, 150: 319- 324.

King, R.A., V.J. Hearing, D.J. Greel and W.S. Ometting (1995) In: The Metabolic Basis of Inherited Disease (Eds., C.R. Scriver, et al.) McGraw-Hill, pp.4353- 4392.

Koga, A., H. Inagaki, Y. Bessho and H. Hori (1994) Fish Biol. J. MEDAKA, 6: 7- 15.

Koga, A., H. Inagaki, Y. Bessho and H. Hori (1995a) Mol. Gen. Genet., 249: 400- 405.

Koga, A., S. Shimada, M. Suzuki, S. Takeda, H. Inagaki, Y. Bessho and H. Hori (1995b) Abstract for the 8th Annual Meeting of Japanese Society for Molecular Biology.

Kwon B.S., A.K. Haq, S.H. Pomerantz and R. Halaban (1987) Proc. Natl. Acad. Sci. USA, 84: 7473- 7477.

Matsumoto, J., E. Hirose, K. Miyazaki, H. Yamamoto and T. Takeuchi (1992a) Zool. Sci., 9: 1213.

Matsumoto, J., T. Akiyama, E. Hirose, M. Nakamura, H. Yamamoto and T. Takeuchi (1992b) Pigment Cell Res., 5: 322- 327.

Nakajima, Y. and M. Obika (1986) Cell Tissue Res., 244: 279- 283.

Ono, H., J. Matsumoto, H. Yamamoto and T. Takeuchi (1994) Pigment Cell Res., 7: 367.

Ozato, K., H. Kondoh, H. Inohara, T. Iwamatsu, Y. Wakamatsu and T.S. Okada (1986) Cell Diff., 19: 237- 244.

Prota, G. (1992) In: Melanins and Melanogenesis. Academic Press. San Diego, California.

Seiji, M. and T.B. Fitzpatrick (1961) J. Biochem., 49: 700- 706.

Tomita, H. (1975) In: Medaka (Killifish), Biology and Strains (Ed., T. Yamamoto), Keigaku Publishing Co. Ltd., Tokyo, pp.251- 272.

Tomita, H. and T. Hishida (1961) Embryologia 5: 347- 356.

Yamamoto, H. and T. Takeuchi (1981) J. Histochem. Cytochem., 29: 953- 958.

Yamamoto, H., S. Takeuchi, T. Kudo, C. Sato and T. Takeuchi (1989) Jpn. J. Genet., 64: 121- 135.

Yamamoto, H., S. Takeuchi, T. Kudo, K. Makino, A. Nakata, T. Shinoda and T. Takeuchi (1987) Jpn. J. Genet., 62: 271- 274.