QTL analysis of modifiers for pigmentary disorder in rats carrying Ednrbsl mutations

Pigmentary variation in animals has been studied because of its application in genetics, evolution, and developmental biology. The large number of known color loci provides rich resource to elucidate the functional pigmentary system. Nonetheless, more color loci remain to be identified. In our previous study, we revealed that two different strains, namely, AGH rats and LEH rats, but which had the same null mutation of the Ednrb gene (Ednrbsl) showed markedly different pigmented coat ratio. This result strongly suggested that the severity of pigment abnormality was modified by genetic factor(s) in each strain. To elucidate the modifier locus of pigment disorder, we carried out whole-genome scanning for quantitative trait loci (QTLs) on 149 F2 (AGH-Ednrbsl × LEH-Ednrbsl) rats. A highly significant QTL, constituting 26% of the total pigmentation phenotype variance, was identified in a region around D7Got23 on chromosome (Chr) 7. In addition, investigation on epistatic interaction revealed significant interactions between D7Got23 and D3Rat78 and between D7Got23 and D14Mit4. Results suggested that a modified locus on Chr 7 was mainly responsible for the variance of pigmentary disorder between AGH-Ednrbsl rats and LEH-Ednrbsl rats, and two modifier loci showing epistatic interaction may, in part, influence pigment phenotype.

Scientific RepoRts | 6:19697 | DOI: 10.1038/srep19697 In this study, we analyzed the difference in pigmentation between the Ednrb sl -mutated rats; we performed quantitative trait locus (QTL) analysis using the intercross descendants with varying severity of pigmentation disorder, and we tried to search the modifier gene(s) affecting the phenotype.

Results
Evaluation of pigmentation in F 2 Ednrb sl/sl pups. Homozygous Ednrb sl/sl rats showed a pigmentary disorder. We reported previously that variations of the pigmentary disorder was observed in Ednrb sl/sl rats with different genetic backgrounds 14 . AGH-Ednrb sl/sl rats almost had no pigmentation on their heads, whereas a large pigmented spot on the head was observed in LEH-Ednrb sl/sl (Fig. 1). We concluded that a modifier locus in the LEH background rescued the pigmentary disturbance to some extent 14 . We calculated the unpigmented coat ratio of AGH, LEH, F 1 , and F 2 Ednrb sl/sl rats by using a camera and Photoshop and revealed the degree of variation of the pigmentary disorder. The severity of pigmentary disorder was calculated as a ratio of the unpigmented area in the head (pigment area/total area), which could be used as quantitative trait of the individual. The range of the severity of pigmentary disorder for each homozygous Ednrb sl/sl rat was presented in a scatter plot. Figure 2a shows that spots for F 2 -Ednrb sl/sl rats were fairly scattered in the range of two extreme values compared with that of both AGH-Ednrb sl/sl and LGH-Ednrb sl/sl rats. Moreover, we calculated the mean value of the pigment disorder ratio for each of the AGH, LEH, F 1 , and F 2 Ednrb sl/sl rats, which were 0.997, 0.755, 0.898, and 0.846, respectively (Fig. 2b). The values obtained in AGH-Ednrb sl/sl rats were highly different from that of LEH-Ednrb sl/sl rats with significance of P = 0.000, suggesting that the role of modifier(s) in the variance of pigment disorder observed in the LEH strain. QTL analysis of the modifiers of pigment disorder in the F 2 Ednrb sl/sl rats. MapManager QTXb20 software was used for QTL scan of the genome to determine the severity of pigmentary disorder in 149 F 2 -Ednrb sl/ sl rats with 91 microsatellite markers (Table 1), which showed polymorphism between AGH and LEH rats. As many as 5,000 random permutations in 1-centiMorgan (cM) steps were performed for each chromosome to calculate the likelihood ratio statistic (LRS). This LRS can be interpreted as a χ 2 statistic or as a LOD score. In addition, the LRS can be converted to the conventional base-10 LOD score by dividing it by 4.61 (twice the natural logarithm of 10) 15 . Results of interval mapping were suggestive, significant, and highly significant linkages, that is, the LRS values were 9.6, 16.5, and 25.6, respectively. A highly significant QTL was detected in the region around D7Got23 on chromosome (Chr) 7 (Figs 3 and 4), which explained 26% of the total phenotypic variance ( Table 2). Epistatic interaction analysis was performed using the interaction function of MapManager QTXb20. The LRS values of suggestive, significant, and highly significant interactions were 29.2, 38.2, and 49.9, respectively. Three two-locus interactions were identified for pigmentary disorder, namely D7Got23 (Chr 7) and D14Mit4 (Chr 14), D7Got36 (Chr 7) and D3Rat78 (Chr 3), D7Rat143 (Chr 7) and D3Rat78 (Chr 3) (Table 3). We performed ANOVA to confirm the significant epistatic interactions in these microsatellite loci and the vicinal loci. ANOVA results revealed that the D7Got23 locus showed highly significant interaction with D3Rat78 locus. Thus, we chose the D7Got23 locus located at the peak position of the QTL on Chr 7 to represent the loci showing significant epistatic interactions.
Allelic effects of Ednrb sl/sl modifier loci. Modifier loci influence the phenotype but cannot revert the effects of a predisposed mutation such as Ednrb sl/sl 16 . To estimate the effects of a modifier locus on the pigmentary     disorder in Ednrb sl/sl individuals, a complete evaluation of the genotypic information in all F 2 -Ednrb sl/sl progenies was performed. Figure 5 shows the extent of pigment disorder that was evidently modified by the LEH alleles at the modifier locus on Chr 7. The ratio of pigmentary disorder was higher in homozygous AGH alleles compared with that in heterozygotes or homozygotes for LEH alleles. Significant difference was found between the ratios of homozygous AGH alleles and heterozygotes (P = 0.000), whereas no significant difference (P = 0.481) was found between heterozyotes and homozygous LEH alleles. Thus, we considered that the effect of LEH allele was approximately dominant.

Discussion
Coat color generally depends on the amount of melanin produced by melanocytes derived from neural crest cells 17,18 . Endothelin receptor B (EDNRB) is a G-protein-coupled receptor with seven transmembrane domains that is necessary during the development of neural crest and melanocytes 19 . Mouse with null Ednrb gene appeared albino 11 . However, Ednrb sl/sl rats with different genetic background showed variation in their extent of pigmentation 14 . More severe albino phenotype was observed in AGH-Ednrb sl/sl strain, whereas the pigmentation disturbance was reverted to some extent in LEH-Ednrb sl/sl rats (Fig. 1). This observation suggested that the modifier loci in the LEH allele affected the phenotype. We performed a genome-wide scan in F 2 progenies (AGH-Ednrb sl/sl × LEH-Ednrb sl/sl ) to examine the effect of the modifier loci in pigment disorder in rats using QTL analysis. Results revealed a highly significant QTL around D7Got23 on Chr 7 (Fig. 3); it has an extremely high LRS value and has 26% contribution to the total phenotypic variance (Table 2). Therefore, we assumed that one or more modifier genes responsible for pigment disorder may be located in the same region in Chr 7. The LEH allele in the modifier locus on Chr 7 increased the extent of the pigmented area (Fig. 5). Highly significant difference was observed between the ratios of homozygous AGH alleles and heterozygotes indicating that LEH allele was dominant (Fig. 5). While significant difference was detected between the ratios of homozygous LEH alleles and heterozygotes revealing other loci could influence pigment phenotype as well. Furthermore, a synteny analysis in other mammals revealed that the QTL around D7Got23 locus on Chr 7 in rats corresponded comparatively to a region on Chr 10 in mouse, which housed an Ednrb s modifier locus (k10) determining the expressivity of a white forelock and dorsal hypopigmentation 20 . This result suggested that D7Got23 locus is a region of conserved synteny between rat and mouse genomes. In addition, D7Got23 locus showed epistatic interactions with D14Mit4 locus and D3Rat78 locus. When D7Got23 locus had LEH-homozygous genotype, the LEH-homozygous genotype of D3Rat78 locus showed the highest extent of pigmented coat in F 2 progenies, followed by AGH/LEH-heterozygous and then the AGH-homozygous genotypes (Fig. 6). The LEH allele resisted the pigmentary disorder in both D7Got23 and D3Rat78 loci. We noted that the physical position of D14Mit4 (43.8 Mb, RGSC 5.0) is close to that of D14Got40 (35.3 Mb, RGSC 5.0), which was detected as the hooded locus 21 . The hooded phenotype is one of coat color phenotypes in rat with many alleles causing different extents of pigmented coat area [22][23][24] . Therefore, the modifier locus we detected for   (Fig. 6). The D7Got23 locus was the main one responsible for the variation of pigmentary disorder. The D14Mit4 locus showed significant effect on the pigmentary phenotype when the D7Got23 locus was homozygous for LEH allele. The F 2 rats owning the LEH-homozygous genotype at the D7Got23 locus and the AGH-homozygous genotype at the D14Mit4 locus showed the highest extent of pigmented coat. Possible modifier genes responsible for the pigmentary disorder within the identified chromosomal region were identified using some bioinformatics methods, such as genome annotation combined with literature searches to check the confidence interval of the QTL on Chr 7 for potential genes that might be involved in the development of melanocytes 25 . More than 70 genes were identified (Supplementary Table S1). Among these candidates, Lgr5 and Wif1 attracted our attention. Both of them are associated with WNT signal pathway, which is responsible for the development of melanocytes 26,27 . Mass spectrometry demonstrated that Lgr4 and Lgr5 associate with the Frizzled/Lrp Wnt receptor complex 28 . However, the relationship between Lgr5 gene and pigmentation was unknown. The Wif1 gene, a Wnt inhibitory factor 1, was expressed not only in the melanocytes of normal human skin but also in cultured melanocytes and promoted melanogenesis in normal melanocytes 29 . In addition, we sequenced the coding region of these genes; however, we failed to find any difference between AGH and LEH strains. The causal difference might exist within the regulatory region of Lgr5 or Wif1 gene. But it was possible that other causative genes had not been sequenced in our study. The main reason that we failed to identify the modifier gene(s) was the broad confidence interval of the QTL on Chr 7 (between D7Rat31 and D7Rat143, from 28.5 Mb to 79.0 Mb, RGSC 3.4). We tried hard to find more markers showing polymorphism between AGH and LEH strains in this interval to improve the confidence interval, but we found none. The origin of this interval in AGH and LEH strains might be the same. In addition, the low number of the F 2 rats limited the precision of the result. To elucidate the pigment disorder, other biological methods must be employed to identify the modifier gene(s) in this region.
In conclusion, we have identified a highly significant QTL on Chr 7 by using two pigmentary disorder strains. Other two loci, namely, D14Mit4 locus on Chr 14 and D3Rat78 locus on Chr 3, show interaction with the main QTL on Chr 7 and may synergistically affect pigmentation.

Methods
Animals. The F 1 family was produced by crossing 2 AGH/Hkv (aganglionosis Hokkaido)-Ednrb sl/+14 males and 8 LEH/Hkv (Long-Evans Hokkaido)-Ednrb sl/+14 females. Five heterozygous males and 20 heterozygous females of F 1 rats were bred to generate the F 2 animals (n = 592), from which 149 Ednrb sl/sl pups were selected by genotyping through PCR amplification using a pair of specific primers (F-CCTCCTGGACTAGAGGTTCC and R-ACGACTTAGAAAGCTACACT) and then flanking the site of deletion with 301 bp in Ednrb gene. PCR products were electrophoresed in 1.5% agarose gels to distinguish the wild (511 bp) type and the mutant (210 bp) type. AGH (n = 35), LEH (n = 34), and F 1 rats (n = 32) were raised for the determination of the severity of pigment disorder in each strain. Animals were maintained in specific pathogen-free conditions, fed, and supplied with water ad libitum. All research and experimental protocols were in accordance with the Regulation for the Care and Use of Laboratory Animals, Hokkaido University and approved by the President of Hokkaido University following to the review of the Institutional Animal Care and Use Committee (Approval ID: No. 110226). Measurement of unpigmented coat ratio. Photographs from the dorsal side of the rats were taken with a COOLPIX 4500 digital camera (Nikon, Tokyo, Japan). The total area of the head and of the pigmented area in the head were traced with lasso tools and calculated as pixels using the histogram function of Photoshop Elements 4.0 (Adobe Systems, California, USA). The ratio of unpigmented area of the head (1-pigmented area/total area) was then calculated as the phenotypic value of pigment disorder.

Microsatellite genotyping.
A total of 149 intercross progenies were selected for the genome-wide scan.
Extraction of genomic DNA from the tail clips was performed using the standard methods. A total of 91 polymorphic microsatellite markers covering all 20 autosomes were selected from the National Center for Biotechnology Information < NCBI; http://www.ncbi.nlm.nih.gov/> (RGSC 5.0) for the genome-wide scan at 10-30 Mbp resolution ( Table 1). The X chromosome was not scanned in our study because gender bias of pigment phenotype was not found in F 2 rats. The PCR procedure is described as follows: each 10 μ L reaction contains 0.5 μ L (10 ng) DNA, 0.5 μ L (5 pmol) of each primer, 5.5 μ L 2 × Taq MasterMix (CWBIO, Beijing, China), and 3.0 μ L ddH 2 O. Touchdown-PCR was performed as follows: 2 min at 95 °C, followed by 10 cycles for 30 s at 95 °C, 30 s at 63-54 °C, 30 s at 72 °C, then 25 cycles for 30 s at 95 °C, 30 s at 54 °C, 30 s at 72 °C, and a final extension at 72 °C for 10 min. PCR products were electrophoresed using 10% acrylamide gels at 160 V for 1.5-3 h, stained with ethidium bromide, and photographed under an ultraviolet lamp ( Supplementary Fig. S1).

Linkage analysis.
Genotyping results and phenotypic values were analyzed using MapManager QTXb20 to identify the pigment disorder modifier loci 30 . A maximum likelihood algorithm with interval mapping was performed to examine linkage probability. Permutation tests were done in 1-cM steps for 5,000 permutations to determine the suggestive, significant, or highly significant levels of statistics. Interactions between all pairs of markers were also screened using MapManager QTXb20 program. We used 5,000 permutations in 1-cM steps to detect possible interactions using regression.

Two-locus interaction analysis.
Interaction effects or epistasis of all pairs of marker loci were tested with the interaction function of Map Manager QTXb20 15 . According to the manual, pairs of loci had to pass two tests to claim significant interaction effects. First, the total effect of the two loci had a p-value < 10 −5 . Second, the interaction effects itself had a p-value < 0.01. Statistical analyses. Statistical analyses was perforemed using the SPSS 19 software and R. Significance test for pigment disorder ratio in AGH-Ednrb sl/sl , LEH-Ednrb sl/sl , F 1 , and F 2 , as well as in three types of alleles of D7Got23 located on the peak were analyzed using ANOVA in SPSS 19. Significance test for the corresponding phenotypic mean values of nine types of alleles of two pairs loci showing epistatic interactions were analyzed by ANOVA in R.