Myosin 10 is involved in murine pigmentation

Abstract Myosins are molecular motors that are well known for their role in cell movement and contractile functions. Although extensively studied in muscle physiology, little is known about the function of myosins in mammalian skin. As part of the Sanger Institute Mouse Genetics Project, we have identified a role for Myo10 in pigmentation, with a phenotype unlike those of Myo5a or Myo7a. Adult mice homozygous for a disrupted Myo10 allele on a C57BL/6N background displayed a high degree of penetrance for white patches on their abdomen and dorsal surface. Forepaw syndactyly and hind paw syndactyly were also observed in these mice. Tail epidermal wholemounts showed a complete lack of melanocytes in the hair follicles and interfollicular epidermis. Myo10 has previously been implicated in human pigmentation. Our current study reveals involvement of Myo10 in murine skin pigmentation.


| RE SULTS
To assess the efficiency of knockout of the Myo10 allele, we performed real time-qPCR analysis, which showed that Myo10 tm2/tm2 mice have ~15% of wild-type Myo10 transcript levels ( Figure S1A).
Similar to the reported Myo10 m1J allele [10] (http://www.informatics.jax.org/allele/MGI:5578506), Myo10 tm2/tm2 mice typically displayed abnormal dorsoventral coat patterning with white belly spots and extensive depigmented areas in the tail. Depigmentation was limited to the tail tip in control mice ( Figure 1B-C). Myo10 tm2/ tm2 mice showed ocular abnormalities and also typically presented with interdigital webbing-a condition in which the skin between paw digits is not lost during development ( Figure 1D). X-ray analysis confirmed that this abnormality is not due to fusion of bones (http://www.mousephenotype.org/data/genes/MGI:107716).
Unlike Xenopus Myo10 mutants, Myo10 tm2/tm2 mice did not show any cranial or skeletal abnormalities. [12] Abnormal coat colour pattern was observed in all male mutants, whereas females showed slightly reduced penetrance of the phenotype ( Figure 1E). Skin histopathology revealed no obvious structural abnormalities in the epidermis or dermis ( Figure S1B).
The apparent lack of pigment in the tail of Myo10 tm2/tm2 mice prompted us to investigate whether the skin had any melanocyte abnormalities. Confocal imaging of tail epidermal wholemounts revealed a complete lack of pigmented melanocytes in the hair follicles and interfollicular epidermis (Figure 2A). Similar results were observed by imaging the wholemounts under bright field combined with fluorescent detectors ( Figure 2B). We confirmed the lack of melanocyte stem cells and differentiated melanocytes by labelling with antibodies to c-Kit and tyrosinase-related protein 1 (Trp1), respectively. Immunostaining for c-Kit and Trp1 showed no staining in mutant hair follicles and interfollicular epidermis, whereas distinct accumulation of melanocytes with characteristic dendritic projections was observed in WT epidermis ( Figure 2C; Figure S1C). This  Figure S1D). Myo5a is predicted to interact with Myo7a. The other interaction partners are mostly exocyst complex component (Exoc) genes, which are involved in the docking of exocytic vesicles. [13] Surprisingly, the interaction network of Myo10 was separate from that of Myo5a. Top scoring predicted interaction partners included suggests that Myo10 may play a different role from Myo5a in pigmentation. Myo10 is highly expressed in mouse skin, [14] so we analysed expression of Myo10 in different skin compartments using a publicly available skin gene expression (Hair-GEL) database. [15] Interestingly, while Myo10 is predominantly expressed in melanocytes in embryonic and neonatal mice, it is also expressed in other skin compartments such as epidermis and dermal fibroblasts ( Figure S1E).

| CON CLUS IONS
The role of myosins in melanoblast migration is known through studies on Rac1 mouse mutants, [16,17] although direct evidence implicating the involvement of Myo10 has not been reported. An actinbundling protein, Fascin1 (Fscn1), which is a key interacting partner of Myo10, is known to promote migration and proliferation of melanoblasts. [18] In the network analysis of Myo10 interaction partners, we found Fscn1 as a key protein that is known to physically interact with Myo10. This suggests that Myo10 may have a similar role to Fscn1 in actin bundling.
Like most cells, melanoblasts migrate with the help of cytoplasmic projections called filopodia. [19,20] HeLa cells that lack Myo10 fail to form filopodia, [21] indicating that Myo10 is an essential component for filopodia formation. In addition to its role in melanoblast migration, Myo10 has been shown to function in melanosome transportation.
Exposure to UV light increases the number of filopodia and Myo10 expression in differentiated melanocytes. Knockdown of Myo10 leads to decreased numbers of filopodia and reduced melanosome transfer. [9] The digit webbing phenotype may also indicate a role for Myo10 in apoptosis, for example as a dependence receptor required for apoptosis induction. Recent data suggest that Myo10 determines tumour invasiveness and cell cycle regulation (reviewed in), [22] leading us to suggest that it could exert these functions both in melanocytes and in other skin cell populations.
While Myo10 mutant tail epidermis lacked melanocytes, dorsal and ventral skin had a chimeric pattern of normal and abnormal pigmentation. It is intriguing that the pigmentation process in the tail is distinct from dorsoventral skin and reduced penetrance of the phenotype among females, although the exact mechanism of this for sexual dimorphism is unclear. [23] Developmental profiling of melanoblast migration using Myo10 mutants during embryogenesis would be helpful to address this phenomenon. It is also possible that Myo10 has other cellular roles in the skin, for example during wound healing or tumour formation.

ACK N OWLED G EM ENTS
We thank Dr. Ed Ryder and Diane Gleeson (WTSI) for expression data and discussion of the allele. We thank Dr. Clare Garcin for helpful suggestions. Financial support from Medical Research Council UK (G1100073) and Wellcome Trust (096540/Z/11/Z) is gratefully acknowledged.

CO N FLI C T O F I NTE R E S T
The authors have declared no conflicting interest.