INTRODUCTION

The genus Oryzias consists of more than 14 species including medaka (O. latipes), a species well known as a laboratory animal (Naruse et al., 1993; Naruse, 1996; Roberts, 1998). Each species of this genus reveals a specific pattern of geographic distribution; some species are endemic to freshwater (FW) lakes in an Indonesian Island, while others are distributed widely in FW, brackish water (BW) and seawater (SW) area of Asia (Naruse et al., 1993; Naruse, 1996; Roberts, 1998). Since such distribution patterns have hardly been altered by human activities because of low commercial value, Oryzias species offer interesting examples for studies of species diversity, evolution and ecology.

The genus Oryzias is also a potential model for studies of adaptation to hyper- and hypoosmotic environments because diverse osmotic adaptability among species is expected. In this paper, we compared the SW adaptability of four Oryzias species, O. javanicus, O. dancena, O. latipes and O. marmoratus inhabiting different osmotic envronments. O. javanicus is found in SW and BW of Thailand, the Malaysian Peninsula and Indonesian Islands (Iwamatsu et al., 1982; Naruse, 1996; Roberts, 1998). O. dancena (also known as O. melastigma) is widely distributed in BW and FW ranging from Pakistan and India to Burma and Thailand (Naruse, 1996; Roberts, 1998). O. latipes is distributed widely in FW of Japan, Korea and China (Naruse et al., 1993; Naruse, 1996) although it is sometimes found in BW (Miyamoto et al., 1986). O. marmoratus is endemic to Lake Towuti, a highland FW lake of Sulawesi Island, where Adrianichthid fish including Oryzias are highly diversified (Naruse et al., 1993; Naruse, 1996). The results showed that adaptability to SW at different life stages correlated with distribution area in their natural habitats.

MATERIALS AND METHODS

RESULTS

Survival of adult fish after direct transfer

All O. javanicus and O. dancena survived the transfer to 50% SW and SW (Fig. 1). In contrast, all O. marmoratus and O. latipes transferred to SW died within 2 hr and 9.5 hr, K. Inoue and Y. Takei respectively. The average time of survival in SW was, 1.6±0.2 hr in O. marmoratus and 4.3±0.8 hr in O. latipes, respectively, which was significantly different from each other (p<0.01). O. marmoratus was not adaptable even to 50% SW; all fish died within 25 hr (19.3±1.7 hr in average) after transfer to 50% SW. On the other hand, O. latipes survived in 50% SW for more than 25 hr (Fig. 1) and no fish died after elongated exposure to 50% SW for up to 48 hr (data not shown). It was also found that prior acclimation of O. latipes in 50% SW for 24 hr allowed them to survive in SW for more than 30 days. SW-acclimated O. javanicus, O. dancena and O. latipes could survive the transfer back to FW.

Fig. 1

Survival of four Oryzias species after direct transfer from freshwater (FW) to seawater (SW), 50% SW and FW (n=8). Circles, triangles and squares indicates survival rates in SW, 50% SW and FW, respectively. Note that all O. dancena and O. javanicus survived in SW, 50% SW and FW and thus the three symbols are piled up at same points.

Spawning and fertilization

O. latipes spawned eggs everyday while other three species spawned several times a week in FW. Most eggs of the four species were fertilized eggs (Fig. 2). O. latipes, O. dancena and O. javanicus also spawned eggs in SW as frequently as in FW. Most eggs of O. dancena and O. javanicus were fertilized as in FW. However, 83.9% of O. latipes eggs spawned in SW were unfertilized (Fig. 2).

Fig. 2

Fertilization rates in eggs of four Oryzias species spawned in seawater and freshwater. n=86 for FW-spawned O. marmoratus eggs, n=242 and 173 for FW- and SW-spawned O. latipes eggs, n=101 and 52 for FW- and SW-spawned O. dancena eggs, and n=75 and 71 for FW- and SW-spawned O. javanicus eggs, respectively.

The frequency of motile sperms of O. javanicus was more than 75% in FW and more than 50% in SW (classes 4 and 3 of Suquet et al., 1992, respectively). The same result was also obtained in O. dancena. While the sperm of O. latipes and O. marmoratus was more than 75% motile in FW (class 4), no apparently motile sperm was found in SW (class 0).

Hatching

Eggs of O. javanicus spawned in FW revealed the highest hatching rate in SW followed by 50%SW and FW (Fig. 3a). In contrast, other three species revealed the highest hatching rates in FW. In all species, hatching tended to be delayed by the increase of salinity. Hatching rates of FW-spawned O. dancena and O. latipes eggs in SW were almost the same, about 60%. O. marmoratus embryos died immediately after the transfer to SW and revealed a low hatching rate even in 50%SW.

Fig. 3

Comparison of hatching and death of eggs of four Oryzias species cultured in freshwater (FW), 50% seawater (SW) and SW. Results using eggs spawned in FW and SW were shown in panels a and b, respectively. In the left graph of each panel, rates of embryos hatched out (open bars) and those of embryos died before hatching (solid bars) are shown. In the center graph, the time of hatching after fertilization in hatched embryos is shown as means±S.E.M. In the right graph, the time of death after fertilization in embryos died before hatching is shown as means±S.E.M. Significant differences among eggs cultured in FW, 50% SW and SW were indicated by asterisks. Significant differences between the time of hatching and that of died before hatching were indicated by arrowheads at the right end. Single and double asterisks and arrowheads represent p<0.05 and p<0.01, respectively. Numbers of fertilized eggs used for each group were 23–30, 22–40, 14–19, and 14–30 for O. marmoratus, O. latipes, O. dancena, and O. javanicus, respectively.

Similar results were obtained with O. latipes, O. dancena and O. javanicus eggs spawned in SW (Fig. 3b). In this case, more than 90% of O. javanicus embryos hatched out in all salinity conditions. The hatching rate of SW-spawned eggs was higher than that of FW-spawned eggs in O. latipes, but it was lower in O. dancena.

To examine the hatching ability in SW separately from the difference of development, we cultured FW-spawned fertilized eggs of O. javanicus, O. dancena and O. latipes in SW until the hatching of the first embryo and then transferred to newly prepared SW and FW (Fig. 4). After transfer, the hatching of O. javanicus and O. latipes embryos immediately started both in FW and SW possibly by stimulus of the change of water or elevated temperature. The hatching rate of O. latipes 24 hr after the transfer was almost the same in SW and in FW and that of O. javanicus was slightly higher in SW than in FW. On the other hand, the hatching of O. dancena embryos delayed in SW and the hatching rate 24 hr after the transfer was higher in FW than in SW.

Fig. 4

Comparison of hatching rates of embryos of three Oryzias species in freshwater (FW) and seawater (SW). Fertilized eggs, spawned in FW, were cultured in SW until the hatching of the first embryo was observed and then equal number (n=53, 37, and 30 for O. latipes, O. dancena, and O. javanicus, respectively) of pre-hatching embryos were transferred into FW and SW. Hatching and death of embryos were checked every 30–120 min. Squares and circles indicate embryos transferred to FW and SW, respectively.

DISCUSSION

To expand the habitat from FW to the sea, it is indispensable for adult fish to have ability to adapt to hyperosmotic condition in SW. In addition, processes in early development such as the fertilization, embryonic development and hatching are influenced directly by ambient salinity in oviparous fish. Thus, it is necessary to adapt all such processes to complete the whole lifecycle in the sea.

Adult fish of O. javanicus and O. dancena revealed high adaptability to sudden changes of the ambient osmotic pressure in both directions, i.e., from FW to SW or vise versa, whereas O. marmoratus exhibited the lowest ability for SW adaptation among the four species (Fig. 1). The result of O. latipes was intermediate: Although this species failed to survive direct transfer to SW, it became adaptable after 24 hr acclimation in 50% SW. This result suggests that O. latipes is potentially euryhaline as suggested by previous reports (Yanagishima, 1957; Miyamoto et al., 1986; Haruta et al., 1991; Sakamoto et al., 2001) but lacks the adaptability to a sudden increase of the ambient osmotic pressure. This characteristic is especially interesting for studies of osmo-regulatory mechanisms: Using this species, it becomes possible to study functions of fast-acting osmoregulatory factors such as natriuretic peptides and angiotensin II separately from slow-acting hormones such as cortisol and growth hormone (Takei, 1993; Takei and Hirose, 2002).

It is well known that the sperms of SW fish are activated when they are exposed to water with osmotic pressure or concentration of ions such as K⁺ higher than the seminal plasma, while those of FW fish are activated by lower osmotic pressure or ion concentration (Morisawa and Suzuki, 1980; Billard and Cosson, 1992). We found in this study that the sperm of O. dancena and O. javanicus can be activated not only in FW but also in SW. These species offers a new model to study sperm activation mechanisms. In O. latipes, approximately 16% of eggs spawned in SW were fertilized (Fig. 2) although the sperm was immotile in SW. It is possible that a very small number of sperms are motile in SW, or the immotile but mature sperm might be transferred passively to the micropyle.

We also found that hatching rates in SW are different among species in this study (Figs. 3, 4). The hatching rate of FW-spawned eggs in SW was the highest in O. javanicus, followed by O. latipes and O. dancena, and the lowest in O. marmoratus. The hatching rate of SW-spawned eggs was the highest in O. javanicus, followed by O. latipes and then O. dancena. The low hatching rate of O. marmoratus in SW is obviously due to a low survival rate of embryos in SW, not due to the failure in hatching itself, because all embryos transferred in SW died within 1 day. The cause of the decease of hatching rate of O. latipes embryos in SW also seems to be the death of embryos before reaching the hatching stage because the average period for hatching is longer than that of death of embryos (Fig. 3). On the other hand, it seems that the decrease of hatching rate of O. dancena embryos in SW is due to the low ability of hatching in SW, i.e., embryos of this species developed successfully to hatching stage but failed to hatch out. This idea is supported by (i) the average time of embryo death longer than the average hatching time (Fig. 3), (ii) our microscopic observation, in which we found that O. dancena embryos that failed to hatch out died after the absorption of most of the yolk without apparent morphological abnormalities (data not shown), and (iii) the higher hatching rate in FW than in SW when embryos that had been cultured in SW until the hatching stage were transferred to FW or SW (Fig. 4). The mechanism of hatching has been extensively studied in O. latipes; it has been reported that two enzymes, low choriolytic enzyme (LCE) and high choriolytic enzyme (HCE), cooperatively function at the time of hatching (Yasumasu et al., 1997 for review). Although a possible cause of interspecific differences in hatching ability in SW is that the range of salt concentration in which hatching enzymes can function is different among species, information on relationship between activities of hatching enzymes and salinity is not available at present except for the report in O. latipes by Yamagami (1973). Molecular cloning of cDNAs encoding hatching enzymes of Oryzias species is now in progress to compare enzyme activities under the different salt concentration using recombinant techniques.

Results obtained in this study are consistent well with the natural habitat of each species (Table 1). O. javanicus, a BW and SW species, fulfilled all the requirements. We also found that the reproduction of this species became less active after a long culture in FW (unpublished observation). Thus, it seems that the lifecycle of this species has adapted to SW than to FW. O. dancena also fulfills most requirements to live in SW. It seems, however, more advantageous for this species to inhabit BW and FW than SW because the hatching system is not fully adapted to SW. O. latipes seems more suitable to live in FW than the above two species, which is consistent with the fact that the major habitat of this species is FW area. The current results showed that it became adaptable to SW after a short acclimation period in 50% SW and spawned normal eggs in SW although the fertilization rate is low. These results are consistent with the fact that this species is sometimes observed in a pond containing high concentration of salt (Yanagishima and Mori, 1957). O. marmoratus, found only in a FW lake, revealed the poorest adaptability to SW at every stage of the lifecycle.

Table 1

Adaptability to seawater in four Oryzias species

Adaptability to SW may be one of the accounts for the geographical distribution of each species. Myers (1938) and Darlington (1957) proposed a classification of FW fish into three categories based on salt tolerance; the primary division consisting of strict FW fish that cannot or do not enter the sea, the secondary division consisting of fishes that occur chiefly in FW but can enter the sea and survive there for a limited time, and the peripheral division consisting of fishes that are found in FW but connected closely with the sea. Darlington (1957) emphasized that SW is the most effective barrier to stop the dispersal of fishes in primary division and the present distribution of fishes in other divisions are often the result of dispersal through the sea. Although he classified Oryzias into the secondary division, results in the present study suggested that SW-tolerance of each species is too variable to put them into the same division. O. marmoratus is apparently a member of the primary division. While Lake Towuti has a river to the sea, the sea is merely a “barrier” to prevent the migration to other regions (Fig. 5). O. latipes is likely to be in the secondary division. While this species is essentially a FW species, results in this study suggested that it has enough adaptability to achieve the transition through the sea, which may be one of the causes of wide geographic distribution of this species. O. dancena seems to be at an intermediate position between the secondary and the peripheral divisions. For this species, the sea is not the best habitat but can be used as “passageway” to the next BW or FW area. O. javanicus can be classified into the peripheral division but is somewhat SW fish-like. The sea is a part of the “habitat” and it seems easy to expand the distribution area along the coast for this species.

Fig. 5

Possible route of expansion of distribution area through the sea in four Oryzias species. White area and dark gray area represent FW and SW, respectively.

Among the four species examined in this study, O. latipes is well known as a laboratory animal. This species has various advantages for experimental purposes: it is small in size, tough enough to be maintained in small aquarium in the laboratory, frequently spawn transparent eggs, and have relatively short generation time (Ozato and Wakamatsu, 1994; Ozato et al., 1989). We found that other three species, O. javanicus, O. dancena and O. marmoratus share such advantages with O. latipes. Thus, it seems possible to apply to these species experimental approaches including foreign gene transfer (Ozato et al., 1986; 1989; Inoue et al., 1990; 1992), mutagenesis (Shima and Shimada, 2001) and nuclear transplantation (Wakamatsu et al., 2001). We expect that the application of such experimental techniques to various Oryzias species would bring us new information to understand how fish has developed diverse mechanisms to adapt to hyper- and hypo-osmotic environments.