Hyperosmotic tolerance of adult fish and early embryos are determined by discrete, single loci in the genus Oryzias

The acquisition of environmental osmolality tolerance traits in individuals and gametes is an important event in the evolution and diversification of organisms. Although teleost fish exhibit considerable intra- and interspecific variation in salinity tolerance, the genetic mechanisms underlying this trait remain unclear. Oryzias celebensis survives in sea and fresh water during both the embryonic and adult stages, whereas its close relative Oryzias woworae cannot survive in sea water at either stage. A linkage analysis using backcross progeny identified a single locus responsible for adult hyperosmotic tolerance on a fused chromosome that corresponds to O. latipes linkage groups (LGs) 6 and 23. Conversely, O. woworae eggs fertilised with O. celebensis sperm died in sea water at the cleavage stages, whereas O. celebensis eggs fertilised with O. woworae sperm developed normally, demonstrating that maternal factor(s) from O. celebensis are responsible for hyperosmotic tolerance during early development. A further linkage analysis using backcrossed females revealed a discrete single locus relating to the maternal hyperosmotic tolerance factor in a fused chromosomal region homologous to O. latipes LGs 17 and 19. These results indicate that a maternal factor governs embryonic hyperosmotic tolerance and maps to a locus distinct from that associated with adult hyperosmotic tolerance.

Adjusting to osmotically different environments is a major factor limiting the ability of freshwater organisms to colonize sea water, or for marine organisms to expand their distribution into freshwater systems. Elucidation of the genetic mechanisms underlying the physiological ability to tolerate osmotic challenges would provide insight into the evolutionary processes related to the alteration/expansion of a species' habitat. Although several factors responsible for osmotic tolerance have been identified in plants and fungi [1][2][3][4] , there has been no significant progress in understanding how it is genetically determined in animals. It is well known that teleost species exhibit considerable intra-and interspecific variation in hyperosmotic/salinity tolerance, and it has been suggested that the ability of fish to adjust to sea water is modulated by genetic factors 5,6 . Furthermore, the transition from freshwater to seawater habitats requires that a species acquires hyperosmotic tolerance during certain life cycle stages, as is known in anadromous salmonids and catadromous eels. The gills (and the yolk sac) are the primary osmoregulatory organs where active extrusion of ions occur. In addition to the gills, major osmoregulatory sites are the gastrointestinal tract, which absorbs ions and water following drinking, particularly for acquisition of water, and the kidney, where divalent ions are excreted into sea water. However, the critical determinant of hyperosmotic tolerance and the genes underlying it have not been comprehensively studied. Previous studies have focused on these osmoregulatory organs and their specific genes, such as those for transporters 7 and those observed in transcriptome changes during seawater challenges 8,9 . Medaka fishes in the genus Oryzias are frequently used as experimental animals for a variety of studies, including those in genetics and physiology. This genus consists of approximately 30 species, most of which inhabit freshwater environments, such as lakes, ponds, and paddy fields, and show diverse tolerance to salinity (e.g.

Results and Discussion
Hyperosmotic tolerance of adult fish determined by a single major locus. All (Fig. 1). When the fish maintained in half-strength sea water for 7 days were transferred to full-strength sea water, no fish died even after prolonged exposure to sea water for 1 month. In contrast, all O. woworae and O. profundicola individuals died within 1 week after transfer to half-strength sea water (N = 20, Figs 1 and 2). Since these two species may not be able to adjust to rapid transfer to hyperosmotic water with handling stress, we performed an additional survival test of O. woworae without the handling in hyperosmotic water (Supplemental Fig. 1). All survived individuals died in full-strength sea water, suggesting that O. woworae cannot adjust to full-strength sea water at least.
To investigate whether the difference in hyperosmotic tolerance is a genetic trait, we initially examined the F 1 hybrids of O. celebensis and O. woworae. All F 1 adults survived after the transfer to half-strength sea water, and  no fish died even after further transfer to sea water, as observed in O. celebensis, suggesting that the hyperosmotic tolerance trait is dominant (Fig. 2). Similar results have been reported in hybrids of Oreochromis niloticus and Sarotherodon galilaeus 12 . Furthermore, we backcrossed the F 1 O. celebensis females to O. woworae males, and 22% (22/102) of the adult backcross (BC) progeny died in half-strength sea water. When the remaining 80 BC progeny were further transferred to full-strength sea water, 48 individuals survived. Thus, the total survival ratio was 47% (48/102), and the hyperosmotic tolerant and non-tolerant phenotypes segregated essentially in a 1:1 ratio. These results suggest that the O. celebensis allele at a single major locus is responsible for the hyperosmotic tolerance but the involvement of handling stress cannot be completely excluded.
The locus for adult hyperosmotic tolerance maps to a fused chromosome corresponding to O. latipes linkage groups 6 and 23. To search for the locus pertaining to adult hyperosmotic tolerance, we conducted a genome-wide linkage analysis on surviving and dead BC progeny (32 individuals each) using 211 polymorphic DNA markers. Figure 3A shows the marker positions for 24 linkage groups (LGs) and the correlation between hyperosmotic-tolerant phenotypes and DNA marker genotypes. This analysis identified eight markers significantly associated (LOD score > 3.0) with hyperosmotic tolerance phenotypes (Fig. 3B). All markers were mapped at the ends of These results suggest that a single locus located approximately in the 4-Mb region of the fused chromosome is responsible for the hyperosmotic tolerance of adult fish.
The corresponding regions on O. latipes LGs 6 and 23 contain 132 annotated genes in the Ensemble genome browser (http://asia.ensembl.org/index.html). Some of these candidates are members of the solute carrier and ATP-binding cassette transporter families that are functionally related to osmotic regulation. For teleosts to maintain their hydromineral balance in sea water, the flux of ions relies primarily on these epithelial transporters, including Na + /K + -ATPase, Na + -K + -2Cl − , and the cystic fibrosis transmembrane conductance regulator (Cl − channel), in osmoregulatory organs, as has been discussed in some reviews 3,7,13 . Quantitative trait analyses for hyperosmotic tolerance (i.e. Na+/K+ -ATPase activity and blood plasma osmolality) in Arctic charr and rainbow trout have suggested that allelic variation at the Na + /K + -ATPase a1b and insulin-like growth factor 2 loci is associated with a difference in tolerance 6 , although the causal gene has not been mapped to the above-mentioned region. Nevertheless, polymorphisms of these transporters or their regulatory factors, including cis-regulatory elements, appear to be involved in the hyperosmotic tolerance of adult fish.
A maternal factor essential for embryonic hyperosmotic tolerance and its genetic basis. After transfer to sea water at the 1 to 2 cell stage, all fertilised eggs/embryos derived from five pairs of O. celebensis developed normally (Fig. 4A), similarly to some euryhaline Oryzias species 10 . These embryos that developed to the somite stage in fresh water (24 h after fertilisation) were also able to survive in sea water, as in the case of adults. In contrast, all embryos derived from five pairs of O. woworae died (cleavage ceased) within 10 h of fertilisation when reared in sea water, but survived and developed normally in fresh water. To elucidate the difference in hyperosmotic tolerance, we conducted an additional survival test on the hybrid embryos. All embryos derived from O. celebensis females and O. woworae males developed in sea water similarly to the embryos from O. celebensis pairs (Fig. 4B). Interestingly, all embryos derived from O. woworae females and O. celebensis males died in sea water similarly to the embryos from the O. woworae pair. These results suggest that a maternal factor(s) derived from O. celebensis eggs is responsible for hyperosmotic tolerance during the cleavage stage, since zygotic gene expression has been reported to begin at the early blastula stage in Oryzias species 14 . In contrast, the hybrid late embryos were able to survive in sea water, as in the case of adults, partially by counterbalancing the undeveloped osmoregulatory gills with the numerous ionocytes in the body and/or yolk sac 7,15,16 . To understand the genetic inheritance of the maternal factor(s) related to embryonic hyperosmotic tolerance, we examined the embryos derived from F 1 hybrid females and O. woworae males (Fig. 4C). These fertilised eggs developed in sea water similarly to those in fresh water and to those from the O. celebensis pairs, suggesting that the trait controlled by the maternal hyperosmotic tolerance factor(s) is dominant. Furthermore, the embryos from approximately 60% (46/72) of BC females developed in sea water similarly to those from the O. celebensis pair, whereas the embryos from the remaining 40% (26/72) of BC females died in sea water (Fig. 4D). These results suggest that one or a few major loci are responsible for the maternal hyperosmotic tolerance factor.
A maternal factor linked to embryonic hyperosmotic tolerance maps to O. latipes LG 17. To search for a locus linked to the maternal factor(s) determining embryonic hyperosmotic tolerance, 42 BC females (26 females spawning hyperosmotic-tolerant eggs and 16 females spawning non-tolerant eggs) were screened for 107 polymorphic DNA markers. Figure 5A indicates a correlation between a maternal hyperosmotic tolerance factor(s) (i.e. hyperosmotic tolerance of eggs) and genotypes of the BC females. One marker on LG 17 was most tightly linked to the causal locus. Further linkage analyses revealed that four markers, AU168434, OLb0909b, NOL7, and PEX5L, were significantly linked to the egg hyperosmotic tolerance phenotypes (LOD score > 3.0). All markers mapped to the middle region of a fused chromosome in O. celebensis and O. woworae corresponding to O. latipes LGs 17 and 19 (Fig. 5B). The marker most tightly linked to the locus responsible for the maternal hyperosmotic tolerance factor(s) (i.e. hyperosmotic-tolerant eggs) was OLb0909b (LOD score = 4.42), which was located between MF01SSA017B03-3 and AU168434, spanning approximately 7.3 Mb of the O. latipes genome sequence (http://asia.ensembl.org/index.html). These results suggest that a single locus on a distinct chromosome is responsible for the hyperosmotic tolerance of embryos during early developmental stages, namely the maternal hyperosmotic tolerance factor(s). The corresponding region on O. latipes LG 17 and 19 includes 187 annotated genes. Among these candidates, phosphofructokinase and acyl-CoA dehydrogenase are implicated in basic metabolism (http://asia.ensembl. org/index.html). After exposure of fertilised eggs and early embryos to a hyperosmotic environment before the organisation of osmoregulatory epithelia, the synthesis/activation of these possible maternal enzymes might be involved in the prevention of dehydration by producing free amino acids as intracellular osmolytes, such as has been observed in osmoconforming marine invertebrates 17 and in developing seabass larvae [18][19][20] .
Several BC individuals, which had no candidate regions from O. celebensis mentioned above, survived in sea water or spawned hyperosmotic-tolerant eggs, suggesting the presence of minor modifiers that affect tolerance at the adult or early embryonic stages. To remove these effects, we are currently establishing congeneric strains with

Conclusions
In this study, we discovered two major genomic loci discretely responsible for hyperosmotic tolerance during the adult and early embryonic stages in teleosts, suggesting that distinct mechanisms play critical roles in hyperosmotic tolerance during their life history. Our findings may provide a basis for further clarifying the essential regulatory mechanisms of osmotic tolerance at the molecular level not only in adult but also early embryonic teleosts. The osmoregulatory mechanisms suggested here for adults are observed in some evolutionarily advanced taxa like vertebrates, whereas most marine invertebrates are osmoconformers requiring intracellular regulation, similar to that suggested here for early embryos 21 . Thus, the present genetic study of the determinants of osmotic tolerance may also provide insight into the ontogeny and phylogeny related to adjustment strategies.   artificial sea water (test group) (15 ppt Sealife, Marinetech, Tokyo, Japan). Survival was monitored daily for 7 days. Death was defined as the termination of opercular movement 8 . Fish that survived in half-strength sea water for 7 days were transferred to full-strength (30 ppt) sea water, and the survival of these fish was similarly monitored.

Methods
Survival test of fertilised eggs/early embryos in sea water. Eggs were spawned and fertilised in fresh water and then directly transferred to experimental aquaria containing either fresh or sea water (30 ppt) after 1 h (1 to 2 cell stage). Eggs were incubated separately using 24-well culture plates at 27 °C ± 2 °C. Embryo survival was monitored after 24 h. Survival ratios were calculated from at least 20 eggs per female. sample using a standard method of proteinase K digestion in lysis buffer, followed by a phenol:chloroform extraction 22 . To search for hyperosmotic tolerance-linked markers, we used expressed sequence tags (ESTs) established for O. latipes, which were amplified using previously published primers designed for O. latipes 23 . PCR genotyping for 15 loci (MEST, AMDHD, OcelKCNJ8, KIN, CCT2, CPA4, 2.9B, 2.9c, MPHOSPH6, RFWD3, MPI, CAT, NOL7, PEX5L, and NOP9) was conducted using primers designed based on the EST sequences (http:// mbase.bioweb.ne.jp/~dclust/medaka_top.html) and the draft genome sequence of O. latipes 24 . PCR amplification of each marker was performed as follows: 4 min at 95 °C, followed by 12 cycles of 30 s at 95 °C, 30 s at 65-56 °C, and 2 min at 72 °C, and 22 cycles of 30 s at 95 °C, 30 s at 53 °C, and 2 min at 72 °C, with a final elongation step of 2 min at 72 °C. Polymorphisms between O. celebensis and O. woworae were identified by length differences in the PCR products, restriction-fragment length polymorphisms, or heteroduplex polymorphisms using discontinuous polyacrylamide gels with a 4% stacking gel and a 9% separating gel [25][26][27] . Marker names, assigned LGs, and primer sequences are listed in Supplemental Table 1. A logarithm of the odds (LOD) score of 3.0 was adopted as an indication of potential hyperosmotic tolerance-linked markers.
Statistical analysis. The effect of salinity on the survival rates of early embryos was initially tested with a two-way analysis of variance (ANOVA) to examine whether there was an interaction effect between strain and salinity. If the two-way ANOVA detected a significant interaction, the survival rate of each strain was further compared with separate one-way ANOVAs followed by a Tukey-Kramer post-hoc test for each salinity level. The probability for establishing statistical significance was P < 0.05.
Availability of data and materials. All data generated or analysed during this study are included in this published article (and its Supplementary Information files).