Regional Fluctuation in the Functional Consequence of LINE-1 Insertion in the Mitf Gene: The Black Spotting Phenotype Arisen from the Mitfmi-bw Mouse Lacking Melanocytes

Microphthalmia-associated transcription factor (Mitf) is a key regulator for differentiation of melanoblasts, precursors to melanocytes. The mouse homozygous for the black-eyed white (Mitfmi-bw) allele is characterized by the white-coat color and deafness with black eyes due to the lack of melanocytes. The Mitfmi-bw allele carries LINE-1, a retrotransposable element, which results in the Mitf deficiency. Here, we have established the black spotting mouse that was spontaneously arisen from the homozygous Mitfmi-bw mouse lacking melanocytes. The black spotting mouse shows multiple black patches on the white coat, with age-related graying. Importantly, each black patch also contains hair follicles lacking melanocytes, whereas the white-coat area completely lacks melanocytes. RT-PCR analyses of the pigmented patches confirmed that the LINE-1 insertion is retained in the Mitf gene of the black spotting mouse, thereby excluding the possibility of the somatic reversion of the Mitfmi-bw allele. The immunohistochemical analysis revealed that the staining intensity for beta-catenin was noticeably lower in hair follicles lacking melanocytes of the homozygous Mitfmi-bw mouse and the black spotting mouse, compared to the control mouse. In contrast, the staining intensity for beta-catenin and cyclin D1 was higher in keratinocytes of the black spotting mouse, compared to keratinocytes of the control mouse and the Mitfmi-bw mouse. Moreover, the keratinocyte layer appears thicker in the Mitfmi-bw mouse, with the overexpression of Ki-67, a marker for cell proliferation. We also show that the presumptive black spots are formed by embryonic day 15.5. Thus, the black spotting mouse provides the unique model to explore the molecular basis for the survival and death of developing melanoblasts and melanocyte stem cells in the epidermis. These results indicate that follicular melanocytes are responsible for maintaining the epidermal homeostasis; namely, the present study has provided evidence for the link between melanocyte development and the epidermal microenvironment.

Introduction risk of extinction, the Mitf mi-bw allele was rescued by crossing with C3H/He mice [2]. The Mitf mi-bw mice with pigmented spots were spontaneously born in the late 1990s, during which time we had maintained the Mitf mi-bw allele on the C3;B6 mixed background. Subsequently, the Dct-lacZ transgene was introduced into the C3;B6-Mitf mi-bw /Mitf mi-bw mice with pigmented spots (C3;B6-Mitf mibw /Mitf mi-bw , Tg(Dct-lacZ)). The Mitf mi-bw mouse line with black spots carrying the Dct-lacZ transgene on the C57BL/6J background (C57BL/6J-Mitf mi-bw /Mitf mi-bw , Tg (Dct-lacZ)) was established by crossing the C3;B6-Mitf mi-bw /Mitf mi-bw , Tg(Dct-lacZ) mice with pigmented spots with the C57BL/6J-Mitf mi-bw /Mitf mi-bw , Tg(Dct-lacZ mice (more than 10 generations). In the subsequent text, we use a term, the black spotting mouse, to indicate the homozygous Mitf mi-bw mouse carrying the Dct-lacZ transgene with the black spotting phenotype, unless otherwise specified. Mice were maintained under 12 h light/12 h dark cycle at 23-25°C and were allowed free access to standard mice food and water. This study was carried out in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The animal experiments were performed based on the protocol approved by the Committee on the Ethics of Animal Experiments of Tohoku University Graduate School of Medicine (Permit Number: 2013MdA-251). Animals were euthanized by cervical dislocation. All efforts were made to minimize suffering.

Antibodies
Anti-MITF polyclonal antibody was produced in rabbits using His-Tag-MITF-M as an antigen that contains a His-tag at the amino-terminus of full-length MITF-M [16]. This antibody recognizes each of MITF-M, MITF-A, MITF-C and MITF-H [16] as well as mouse Mitf [13] and chick Mitf [17,18]. The antibody against Dct as a marker for melanocytes was purchased from Abcam (Cambridge, UK). The antibodies against beta-catenin, phosphorylated ERK1/2 (p-ERK1/2), cyclin D1, and Ki-67 were purchased from Cell Signaling Technology (Danvers, MA, USA). Non-specific normal rabbit IgG as a negative control for immunostain was purchased from DAKO (Produktionsvej, Glostrup, Denmark).

Immunohistochemical analysis of the mouse skin
The skin samples were prepared from male control C57BL/6 mice with the Dct-lacZ transgene, the bw mice carrying the Dct-lacZ transgene, and the black spotting mice carrying the Dct-lacZ transgene. C57BL/6 mice were used as the wild type mouse. The mice used were 8 weeks old, unless otherwise stated. The isolated tissues were fixed with SUPER FIX (KURABO, Chuo-ku, Tokyo, Japan) at room temperature for 24 h-48 h. The tissues were paraffin-embedded and were cut into 2-μm sections for immunostaining. The tissue sections were deparaffinized, and hydrated. Unmasking of antigens was performed with 98°C water bath in Immunosaver (Wako Pure Chemical Industries) for 25 min. For blocking the endogenous peroxidase activity, the tissue sections were treated with Peroxidazed 1 (Biocare Medical, Concord, CA, USA) at room temperature for 5 min. Then, to block non-specific antibody binding, the sections were treated with Background Punisher (Biocare Medical) for 10 min at room temperature. Dilution of primary antibodies in this study was as follows: MITF, 1:400, Dct, 1:1000, beta-catenin, 1:400, cyclin D1, 1:400, p-ERK1/2, 1:400, and Ki-67, 1:400. All antibodies were diluted in Can Get Signal immunostain Solution A. The tissue sections were incubated with primary antibodies at 4°C overnight. Immunohistochemical analyses were performed with MACH4 Universal HRP-polymer (Biocare Medical) according to the manufacturer's instructions. The tissue sections were visualized by staining with 3,3'-diaminobenzidine tetrahydrochloride (DAB) (MITF, beta-catenin, cyclin D1, p-ERK1/2 and Ki-67) or Bajoran Purple (Dct). Hematoxylin was used as counterstaining. As negative control, the tissue sections were incubated with normal rabbit IgG (DAKO), instead of a primary antibody. In case of the immunofluorescence analysis of Dct expression, the frozen section of the black patches of the black spotting mouse was used, as described previously [5].

PCR analysis of genomic DNA
PCR was performed on 100-500 ng of genomic DNA. Cycling conditions of the PCR were as follows: initial denaturation for 2 min at 94°C, followed by a three-step profile: denaturation for 10 sec at 98°C, annealing for 30 sec at 65°C and extension for 5 min for 40 cycles. PCR amplification was performed using the following primers for LINE-1 insertion region (intron 3 -exon 4 of Mitf gene): Fp0, 5 0 -GGAAAAGGGAAGTGGTAGCTTTGTG-3 0 and Rp4, 5 0 -TTCCA GGCTGATGATGTCATCAATTACATC-3 0 .

X-gal staining of embryos from the black spotting mouse
This series of experiments was performed in parallel with the embryos from control transgenic mouse and the bw mouse [5]. Mouse embryos were washed with PBS containing 12.5 mM EDTA to remove blood components. The day of plug detection was taken to be E0.5. Timed embryos (E12.5-E17.5) were fixed with 2% paraformaldehyde by using microwave. Fixed samples were rinsed twice in PBS, and then washed in X-gal detergent (2 mM MgCl 2 , 0.05% BSA, 0.1% sodium deoxycholate, 0.02% NP-40 in 0.1 M sodium phosphate buffer pH7.3). The samples were subsequently incubated in chromogenic solution [5 mM K 3 Fe(CN) 6 , 5 mM K 4 Fe (CN) 6 , 0.085% NaCl, 0.1% X-gal in X-gal detergent] overnight at 20°C. After coloration, embryos were stored in 4% paraformaldehyde in PBS.

Results
Evidence for the fluctuation in the functional consequences of the LINE-1 insertion Using the C57BL/6J mouse carrying the Dct-lacZ transgene, we have established the homozygous Mitf mi-bw mouse carrying the Dct-lacZ transgene on the C57BL/6J background [5], termed bw mouse for simplicity in the present study ( Fig 1A). We have also established the Mitf mi-bw mouse line manifesting black spots on the C57BL/6J background carrying the Dct-lacZ transgene, named the black spotting mouse (Fig 1B). In contrast to the bw mouse that shows the completely white coat (Fig 1A), the black spotting mouse has black spots on the dorsal portion ( Fig 1B). We then confirmed that melanocytes are present in pigmented hair follicles at a pigmented patch of the black spotting mouse, as judged by the presence of melanin pigments and Dct expression (Fig 1C). The distribution of melanocytes in a pigmented hair follicle is similar to that seen in the wild-type C57BL/6J mouse [5], although the number of pigmented hair follicles is limited in a given black patch of the black spotting mouse, as detailed below.

Features of black patches of the black spotting mouse
Representative images of the mice that carry the Dct-lacZ transgene are shown (Fig 2A-2C): the control C57BL/6J mouse (A), the bw mouse (B), and the black spotting mouse (C). In contrast to the bw mouse that shows the completely white coat (Fig 2B), the black spotting mouse has black spots mainly on the dorsal portion and the neck region ( Fig 2C). Importantly, the black spotting mouse retains the black-eyed phenotype, but the pigmented patches of the black spotting mouse show a grayish tone (Fig 2C), compared with the control C57BL/6J transgenic mouse (Fig 2A). The difference in the tone of pigmented coat color is apparent in the paraffin block of each skin tissue (insets of bottom panels in Fig 2). To explore the reason for the grayish tone of black patches of the black spotting mouse, we compared the appearance of hair shafts present in the paraffin blocks of the skin tissues (bottom panels). The hair shafts of the control mouse skin contain melanin pigments with the clearly visible frames (left), whereas the hair shafts of the bw mouse lack melanin pigments (middle). Importantly, there are two types of hairs in a black patch of the black spotting mouse: pigmented hairs and hairs lacking melanin pigments (right). Moreover, the degree of melanin deposits appeared to be lower in the pigmented hair shafts of the black spotting mouse (right), compared to the pigmented hair shafts of the control mouse (left). As expected, non-pigmented hair shafts of the black spotting mouse are indistinguishable from those of the bw mouse. Thus, the lower number of pigmented hairs may account for the grayish tone of a pigmented patch of the black spotting mouse, and a black patch of the black spotting mouse contains hair follicles lacking melanin pigments.
Lack of follicular melanocytes in the white coat area of the black spotting mouse We next immunohistochemically analyzed the Mitf expression in hair follicles at the anagen stage using the skin samples prepared from the bw mouse and the white coat area of the black spotting mouse (Fig 3). The hair follicles of the control transgenic mouse contain melanin pigments ( Fig 3A and 3D), and Mitf expression, stained as brown, was detected in follicular melanocytes (arrowheads in Fig 3A). Because the brown coloration representing Mitf expression may be unclear due to the presence of melanin pigments, the transverse section of a hair follicle is also presented in inset. In contrast, melanin pigments and Mitf-positive cells were absent in the hair follicles of the bw mouse (Fig 3B and 3E). Likewise, melanin pigments and Mitf-positive cells were absent in the hair follicles at the white coat area of the black spotting mouse ( Fig  3C and 3F).
To determine the identity of Mitf-positive cells, we next analyzed Dct expression in the hair follicles at the anagen stage by the immunohistochemical analysis (Fig 4). The Dct expression was detected in the hair follicles of the control mouse ( Fig 4A), but not in the hair follicles of the bw mouse ( Fig 4B) and in the hair follicles at the white coat area of the black spotting mouse ( Fig 4C). Taken together, these results indicate that melanocytes are absent in the hair follicles of the bw mouse and in the hair follicles at the white coat area of the black spotting mouse. In contrast, Dct expression was detected in the hair follicles at a pigmented patch of the black spotting mouse ( Fig 4D). Importantly, besides the lack of follicular melanocytes, there is no noticeable change in the structural organization of the hair follicles of the bw mouse and that of the hair follicles lacking melanocytes of the black spotting mouse.

Retained LINE-1 insertion in the Mitf gene of the black spotting mouse
We next explored whether the LINE-1 is retained in the Mitf gene of the black spotting mouse ( Fig 5A). Genomic DNA and total RNA were isolated from a pigmented patch or a white coat area of the black spotting mouse (Fig 5B). The PCR analysis of genomic DNA revealed the amplified products of 7.6 kb from bw mouse DNA and black spotting mouse DNA ( Fig 5C), but not from wild-type mouse DNA. In the latter context, the amplified product of 460 bp from wild-type mouse DNA was detected in a separate agarose gel used (data not shown). These results exclude the possibility of the somatic reversion that eliminates the LINE-1 insertion from the Mitf mi-bw gene.
We then performed the RT-PCR using skin samples taken from a pigmented patch and a white coat area of the black spotting mouse (Fig 5D), showing that Mitf-M transcripts are expressed at the black patch of the black spotting mouse as well as in the skin of the wild type mouse, which is consistent with the Dct expression in follicular melanocytes at the pigmented patch (see Fig 4D). In contrast, Mitf-M transcripts are undetectable in the white coat area of the black spotting mouse (Fig 5D) as well as in the bw mouse skin, as reported previously [8]. These results, together with the immunohistochemical data (Fig 4C), indicate the lack of follicular melanocytes in the white coat area of the black spotting mouse.
Attenuated expression of beta-catenin and cyclin D1 in the hair follicles lacking melanocytes As a first step to explore the functional consequence of the LINE-1 insertion that may determine the fate of skin melanocytes, we analyzed the expression profiles of beta-catenin and cyclin D1 in the skin tissues, isolated from the control transgenic mouse, the bw mouse, and the black spotting mouse (Fig 6). Beta -catenin is a downstream component of Wnt signaling [19], the signal of which is actively operating in the hair shaft precursor cells [20] and melanocytes [21]. Cyclin D1 is a target of Wnt/beta-catenin signaling. Furthermore, beta-catenin may function as a coactivator for MITF in melanocytes [22,23]. Overall, the appearance of the skin tissues was similar among the control transgenic mouse (Fig 6A and 6D), the bw mouse ( Fig  6B and 6E), and the black spotting mouse (Fig 6C and 6F), except for the thickened epidermal layer of the bw mouse. Compared with the control mouse (Fig 6A and 6D), the signal intensity for the expression of beta-catenin and cyclin D1 was lower in the hair follicles of bw mouse (Fig 6B and 6E) and the black spotting mouse (Fig 6C and 6F). In contrast, the staining for beta-catenin was apparent in the epidermis of the bw mouse and the black spotting mouse, compared to the control mouse epidermis (Fig 6A-6C). Moreover, the staining intensity for the cyclin D1 expression was higher in the keratinocyte layer of the black spotting mouse ( Fig  6F), compared to the control mouse ( Fig 6D) and the bw mouse ( Fig 6E). Consequently, using larger magnification of the skin tissue sections (x 400), we compared the expression profiles of beta-catenin and cyclin D1 in the hair follicles of the control transgenic mouse (Fig 7A and 7D), the bw mouse (Fig 7B and 7E), and the black spotting mouse (Fig 7C and 7F). The expression of beta-catenin was detected in follicular cells, located around the dermal papilla and in the outer root sheath and the inner root sheath ( Fig 7A); namely, beta-catenin was expressed in matrix cells around the dermal papilla, melanocytes, and the precursor cells for the cortex of hair shaft. However, the nuclear localization of beta-catenin was not clear in follicular melanocytes of the control mouse (white arrowheads in Fig 7A), Variability in the Functional Consequence of LINE-1 Insertion suggesting that the hair follicles analyzed were at later anlagen stage. In contrast, the staining intensity for beta-catenin expression was apparently lower in the hair follicles lacking melanocytes of the bw mouse ( Fig 7B) and the black spotting mouse (Fig 7C), compared to the control mouse ( Fig 7A). Moreover, the signal intensity for beta-catenin was similar in hair follicles between the bw mouse ( Fig 7B) and the black spotting mouse (Fig 7C). Importantly, the cyclin D1 expression was detected in the nuclei of follicular melanocytes of the control mouse (yellow Variability in the Functional Consequence of LINE-1 Insertion arrowheads in Fig 7D), suggesting that Wnt signaling may be active in follicular melanocytes. The cyclin D1 expression was similar in hair follicles between the bw mouse ( Fig 7E) and the black spotting mouse (Fig 7F). These results suggest that follicular melanocytes may support the expression of beta-catenin in their surrounding cells, such as matrix cells.
Enhanced expression of beta-catenin and cyclin D1 in the epidermis of the black spotting mouse We next compared the epidermal expression of beta-catenin and cyclin D1 among the control mouse (Fig 8A and 8D), the bw mouse (Fig 8B and 8E), and the black spotting mouse (Fig 8C  and 8F). The images shown represent the epidermal portions of the same tissue sections used in Fig 7. Apparently, the staining intensity for beta-catenin was higher in the epidermis of the black spotting mouse, compared to the bw mouse and the control mouse epidermis (Fig 8A-8C). Moreover, the signals for beta-catenin appeared to be localized in the membrane region of keratinocytes in the bw mouse epidermis (Fig 8B). In this context, the staining intensity for cyclin D1 was lower in the bw mouse epidermis (Fig 8E), compared to that in the control mouse epidermis (Fig 8D) and the black-spotting mouse epidermis ( Fig 8F). Importantly, the staining intensity for beta-catenin and cyclin D1 was highest in the epidermis of the black spotting mouse (Fig 8C and 8F). We also confirmed that the keratinocyte layer was thicker in the bw mouse (Fig 8B and 8E), compared to the control mouse (Fig 8A and 8D) and the black spotting mouse (Fig 8C and 8F). These results suggest that beta-catenin may not properly function in keratinocyte nuclei of the bw mouse, whereas the function of beta-catenin may be enhanced in keratinocytes of the black spotting mouse.

Enhanced expression of Ki-67 in the basal layer of the bw mouse epidermis
To explore the molecular basis of the thickened keratinocyte layer in the bw mouse, we next analyzed the expression of p-ERK1/2 and Ki-67 in hair follicles (Fig 9) and the epidermis (Fig  10) of the consecutive skin tissue sections. The expression of p-ERK1/2 and Ki-67 was similar in hair follicles among the three mouse groups (Fig 9). Especially, Ki-67 expression was detected in matrix cells surrounding the dermal papilla. In contrast, the staining intensity for Ki-67 was higher in the basal layer of the bw mouse epidermis (Fig 10E), compared to the control mouse and the black spotting mouse (Fig 10D and 10F). Importantly, the epidermal expression of p-ERK1/2 and Ki-67 was similar in the control mouse and the black spotting mouse (Fig 10). The enhanced expression of Ki-67 may reflect the thickening of the keratinocyte layer of the bw mouse. We therefore suggest that the epidermal microenvironment of the black spotting mouse may be suitable for normal growth of keratinocytes, which may contribute to the survival of developing melanoblasts in the selected areas.

Formation of presumptive black spots by E15.5 in the black spotting mouse
To explore the timing of the formation of the black spots, we took the advantage that the black spotting mouse carries the Dct-lacZ transgene. The number and the distribution of Dct-lacZpositive melanoblasts, stained blue, were analyzed during embryonic development of the black spotting mouse (Fig 11A and 11B). In addition to the ectopic expression of the transgene in caudal nerves [24,25], the RPE and the telencephalon, where the endogenous Dct gene is expressed [26], were stained blue. At E12.5, Dct-lacZ-positive melanoblasts were easily detected in the trunk region of the black spotting mouse embryo (Fig 11A and 11B), although the density of the Dct-lacZ-positive cells was apparently lower in the black spotting mouse embryo, compared to the control mouse embryo [5]. In fact, the number of the Dct-lacZ-positive cells per counted area was consistently lower in the black spotting mouse embryo than that in the control mouse embryo (Fig 11C). In contrast, at this stage, the Dct-lacZ-positive cells were hardly detected in the bw mouse [5]. Thus, the migration ability of neural crest cells is better retained in the black spotting mouse, compared to the bw mouse.  Subsequently, at E15.5 (Fig 11A and 11B), the Dct-lacZ-positive cells accumulated in restricted areas, while the Dct-lacZ-positive cells were undetectable in other areas. In other words, certain groups of melanoblasts survived in the presumptive pigmented patches, and other groups of melanoblasts disappeared in the presumptive white coat areas. Thus, the fate of developing melanoblasts is determined by E15.5 in the black spotting mouse. In contrast, the Dct-lacZ-positive cells disappeared by E13.5 in the bw mouse [5]. Such a difference in the fate of developing melanoblasts may reflect the fluctuation in the functional consequence of the LINE-1 insertion. At E17.5, Dct-lacZ-positive melanoblasts were easily detected as a cluster at  Variability in the Functional Consequence of LINE-1 Insertion the presumptive pigmented patches (Fig 11B). These results suggest that melanoblasts can proliferate and differentiate into melanocytes in certain restricted areas of the black spotting mouse, despite the LINE-1 insertion in its Mitf gene.

Aberrantly spliced Mitf-M transcripts expressed in the pigmented patches
To explore the molecular basis for the formation of the black spotting phenotype, we performed the RT-PCR analysis using other primer set: exon-1M forward primer (Fp1) and    (Fig 12). With this primer set, we were able to detect the aberrantly spliced Mitf-M transcripts related to the LINE-1 insertion [8]. In the wild-type mouse skin, the PCR products were detected as a single band that represents authentic Mitf-M mRNA, whereas multiple PCR products were detected in the skin samples taken from the separated pigmented patches of the black spotting mouse (Fig 12). Sequencing analysis revealed that these PCR products represent (a) Mitf-M mRNA containing a LINE-1 segment of 104 bp, (b) authentic Mitf-M mRNA, (c) Mitf-M mRNA lacking exon 3 or exon 4, and (d) Mitf-M mRNA lacking both exon 3 and exon 4. These splicing products were identical to those detected in the bw mouse brain [8]. These data reconfirm that the LINE-1 insertion is retained in the Mitf gene of the black spotting mouse, thereby excluding the possibility of the somatic reversion. Importantly, the Mitf-M mRNA lacking the exon-4 sequence was shown to encode a dominant-negative form of Mitf-M protein [8], because exon 4 encodes the transactivation domain of Mitf [27]. Thus, the LINE-1 insertion may result in the decrease in the level of functional Mitf-M in melanoblasts and melanocytes at the pigmented patches of the black spotting mouse, which may also account for the grayish tone of pigmented patches (see Fig 2). However, the Mitf-M deficiency does not impair the development of melanoblasts in restricted areas of the black spotting mouse.
Age-related graying of pigmented patches As described above, the black spotting mouse may provide a good model to study the phenotypic consequence of the Mitf-M deficiency, prompting us to analyze the age-related changes of black spots. As expected, the black spotting mouse shows apparent graying (Fig 13). By about 6 months after birth, graying was noticeable, and by about 12 months, almost all hairs of the black spots turn to gray or white. These results suggest that melanocyte stem cells or melanocytes in the hair follicles at the black patches may be more vulnerable to certain aging-related events and thus these cells fail to survive. We have therefore proposed that the Mitf-M deficiency may lead to premature death of melanocyte stem cells or melanocytes in the hair follicles of the black spotting mouse. Alternatively, the phenotypic consequence of the LINE-1 insertion may vary with age.

Diluting effect of the Mitf mi-bw allele on the black spotting phenotype
To explore the inheritance of the black spotting phenotype, we crossed a male black spotting mouse with a female bw mouse or a female black spotting mouse with a male bw mouse. Irrespective of the combination, all the offspring of black spotting mice mated with bw mice have pigmented patches with diluted color, compared to those of black spotting mice (Fig 14A and  14B). In addition, the size of pigmented patches tends to be smaller in the heterozygous black representation for the distribution of surviving melanoblasts, with the emphasis on the appearance of presumptive black patches at E17.5. Note that the drawing of each embryo does not reflect the actual image of the embryo at the indicated age. C. Difference in the number of X-gal-stained cells among C57BL/6J transgenic mouse embryos (control), black spotting mouse embryos (spotting), and bw mouse embryos at E12.5. The area used for cell counting is schematically shown with red square. Note that the tissue sections containing many Dct-lacZ-positive cells were selected from the black spotting mouse embryos at E12.5, whereas every tissue section contained a large number of Dct-lacZ-positive cells from control mouse embryos at E12.5 [5]. The Dct-lacZ-positive cell number/counted area is presented as mean ± S.D. of at least three embryos, although the data were obtained with a semi-quantitative measure. The data of control mouse embryos and bw mouse embryos were taken from the published paper [5] and are shown for comparison. At E12.5, the number of developing melanoblasts was consistently lower in the black spotting mouse than that in control mouse. doi:10.1371/journal.pone.0150228.g011 Variability in the Functional Consequence of LINE-1 Insertion spotting mouse with bw mouse, compared to the patches seen in the homozygous offspring of the black spotting mouse (Fig 14, bottom panels). Thus, the black spotting phenotype is diluted in the presence of the Mitf mi-bw allele; namely, the number of pigmented hair follicles is decreased in the offspring of the black spotting mouse mated with the bw mouse. We also crossed the black spotting mouse with the C57BL/6 mouse, yielding the offspring with black coat color that is indistinguishable from the C57BL/6 mouse (photos not shown). In conclusion, the black spotting phenotype is inherited in a semi-dominant manner for the bw phenotype, whereas the black spotting phenotype appears to be inherited as a recessive trait for the white-coat phenotype.

Features of the grayish tone of pigmented hairs
The presence of non-pigmented hairs is responsible in part for the grayish tone of a black patch of the black spotting mouse (Fig 2), which may reflect the lower number of developing melanoblasts in the trunk region of the black-spotting mouse embryo at E12.5 (about 50% reduction), compared to the control mouse embryo (Fig 11C). It is therefore conceivable that follicular melanocytes may also decrease in their number even at a given black patch. In addition, the amount of functional Mitf-M protein may be decreased in follicular melanocytes at a pigmented patch, because of the expression of aberrantly spliced Mitf-M transcripts, including the transcript encoding a dominant negative isoform lacking exon 4 (the activation domain) [8]. Taken together, these findings suggest that melanin production may be decreased in follicular melanocytes at a black patch of the black spotting mouse.
The early graying seen in the black spotting mouse is reminiscent of a recessive vitiligo mouse carrying the Mitf vit allele [28]. The Mitf vit mouse was spontaneously arisen in the C57BL/6J strain and is characterized by the congenital white spotting on its black coat background, associated with the age-related graying of the pigmented hairs [28]. In contrast, the black spotting phenotype was originally arisen in the bw mouse on the mixed C3;B6 background [2]. The Mitf vit gene carries a single-base change that results in the Asp-to-Asn substitution at position 222 in the basic helix-loop-helix region [29], and the Mitf vit mutation is expected to alter the function of all Mitf isoforms. Indeed, the Mitf vit mutation appears to impair the cooperation with Lef-1 or other factors [22]. Thus, the pathogenesis for early graying of the Mitf vit mouse is the Mitf deficiency, which is essentially similar to that of the black spotting mouse.

Molecular basis of the black spotting phenotype
In this study, we are unable to clarify the molecular basis for the black spotting phenotype; namely, it remains to be investigated why the black spotting mouse could partly rescue melanoblasts during embryonic development. One possibility is that the black spotting mouse carries a mutation probably in a modifier gene that may support Mitf expression or attenuate the influence of L1 insertion on Mitf expression in developing melanoblasts. In this context, we have shown that the forced expression of Mitf-M in cultured neural tube isolated from E9.5 bw embryos increases the number of migrating neural crest cells and enhances the differentiation of migrating melanoblasts [5]. Thus, the Mitf-M deficiency associated with the LINE-1 insertion reduces the migration of neural crest cells from the neural tube and the differentiation of migrating melanoblasts. In fact, the expression level of Mitf-M mRNA is lower in bw mouse embryos than that in control mouse embryos at E11.5, and subsequently, it was undetectable at E13.5 in bw mice [5]. Moreover, we have shown the enhanced apoptosis of bw melanoblasts during embryonic development [5]. These results indicate that the critical level of Mitf-M is required for the survival of melanoblasts around E13.5. Thus, a certain population of melanoblasts can survive at E13.5 probably due to the increase in Mitf-M function of the black spotting mouse. The relevant mutation may assure the activity of Mitf-M that is very close to the threshold level for the melanoblast survival.
Moreover, we have shown the apparent differences in the epidermal phenotype between the bw mouse and the black spotting mouse (Figs 8 and 10). The bw mouse shows the thickened keratinocyte layer with the higher expression of Ki-67 and the lower expression of cyclin D1, compared to the black spotting mouse. In fact, the nuclear expression of beta-catenin was not apparent in the bw keratinocytes (Fig 8). In other words, the epidermal features of the black spotting mouse are similar to those of the control mouse. Thus, a gene mutation responsible for the black spotting phenotype may also influence the epidermal development. Alternatively, a certain environmental factor may cause an epigenetic change, such as histone modification, in the Mitf gene with L1 insertion. Further analyses are required, such as the sequencing the LINE-1 present in the black patches and high-density SNP genotyping. We would like to leave these issues for the future.
On the other hand, we have shown that the immunoreactive Mitf is expressed in projection neurons, mitral cells and tufted cells, of the wild-type mouse olfactory bulb, and the distribution of these Mitf-expressing neurons was similar to that found in the bw mouse and the control transgenic mouse [13]. Moreover, the real-time RT-PCR analysis revealed the expression of Mitf-M mRNA in the wild-type mouse olfactory bulb as well as in the bw mouse olfactory bulb, but not in the bw mouse eyeball [13]. These results suggest the regional difference in the functional consequence of the LINE-1 insertion between certain neurons and melanocytes located in the skin [5], inner ear [14,30] and retinal choroid [2]. Thus, understanding the molecular basis for the neuron-specific expression of Mitf will help us to identify the responsible gene for the black spotting phenotype.

Implications of the black spotting mouse
We have shown that follicular melanocytes are responsible for maintaining the expression of beta-catenin in the hair follicle as well as for the keratinocyte homeostasis. Thus, the present study has provided evidence for the link between the epidermal microenvironment and melanocyte development/survival. Moreover, the presence of the black spotting mouse indicates the regional difference in the epidermal environment that may determine the Mitf-M expression level and the fate of melanocytes. Thus, the black spotting mouse provides a model system to study the molecular basis for dynamic changes in the functional consequence of the LINE-1 insertion. Lastly, the present study suggests the possibility that human diseases associated with LINE-1 insertion may present the phenotypic variability.