p53 regulates ERK1/2/CREB cascade via a novel SASH1/MAP2K2 crosstalk to induce hyperpigmentation

Abstract We previously reported that three point mutations in SASH1 and mutated SASH1 promote melanocyte migration in dyschromatosis universalis hereditaria (DUH) and a novel p53/POMC/Gαs/SASH1 autoregulatory positive feedback loop is regulated by SASH1 mutations to induce pathological hyperpigmentation phenotype. However, the underlying mechanism of molecular regulation to cause this hyperpigmentation disorder still remains unclear. In this study, we aimed to investigate the molecular mechanism undergirding hyperpigmentation in the dyschromatosis disorder. Our results revealed that SASH1 binds with MAP2K2 and is induced by p53‐POMC‐MC1R signal cascade to enhance the phosphorylation level of ERK1/2 and CREB. Moreover, increase in phosphorylated ERK1/2 and CREB levels and melanogenesis‐specific molecules is induced by mutated SASH1 alleles. Together, our results suggest that a novel SASH1/MAP2K2 crosstalk connects ERK1/2/CREB cascade with p53‐POMC‐MC1R cascade to cause hyperpigmentation phenotype of DUH.


Introduction
Melanocytes are epidermal cells of neurocrest origin that synthesize melanin and are responsible for skin pigmentation and protection from ultraviolet (UV) radiation [1,2]. Melanocortin receptor 1 (MC1R), a G protein-coupled receptor (GPCR) coupled to the a subunit of G protein, induces elevated concentrations of intracellular cyclic adenosine monophosphate (cAMP) upon binding a ligand such as a-melanocyte-stimulating hormone (a-MSH). Adenosine 3 0 ,-5 0 -cyclic monophosphate (cyclic AMP, cAMP), an important second messenger in melanogenesis, contributes to both melanin synthesis and movement of melanosomes to newly formed dendrites at the cell periphery [1,2]. MC1R, preferentially expressed in epidermal melanocytes [3], is activated by its ligand, a-MSH, a pigmentation hormone produced and secreted by keratinocytes and melanocytes in the skin following UV exposure. Upon ligand binding, GPCRs impart a signal to heterotrimeric G proteins composed of a, b and c subunits, leading to the dissociation of the Ga subunit from the Gbc subunit. These activated G proteins of the guanine nucleotide-binding protein subunit alpha isoforms short (Gas) class directly catalyse the conversion of adenosine triphosphate (ATP) to cAMP [4]. SASH1, originally identified as a candidate tumour suppressor gene in breast and colon cancer, regulates the adhesive and migratory behaviours of cancer cells during tumour formation.
Functional domain analysis has demonstrated that SAM and SH3 domain containing 1 (SASH1) protein may belong to the previously described novel family of putative adapter and scaffold proteins that transfers signals from ligand-receptor interactions [5][6][7]. Our previous study indicated that SASH1 is associated with Gas downstream of a-MSH/MC1R signalling cascade [8]. Our recent study also revealed that SASH1 is regulated by a novel p53/ POMC/a-MSH/Gas/SASH1 cascade to mediate melanogenesis. A novel p53/POMC/Gas/SASH1 autoregulatory positive feedback loop is regulated by SASH1 mutations to induce pathological hyperpigmentation phenotype [9]. However, whether ERK1/2 signal pathway which mediates melanogenesis [10,11] and microphthalmia transcription factor (Mitf), a melanocyte master transcription factor which is also responsible for pigment cell-specific transcription of the melanogenesis enzyme genes [12], are mediated by SASH1 mutations remains unknown.
P53, a canonical transcriptional factor, has been suggested to directly activate transcription of numerous genes such as those that control cell cycle, apoptosis and others. P53 directly mediates UV induction of POMC-MSH signal cascade in skin and stimulates the POMC promoter in response to UV and is involved in UV-independent pathologic pigmentation and could mimic the tanning response [13]. Based on our previous report that SASH1 binds with Gas,the downstream molecule of POMC/MSH-MC1R signalling cascade [8], we deduce that SASH1 may be mediated by p53-POMC-MC1R cascade.

Pull-down assay and nanoflow LC-MS/MS and bioinformatic analyses
The protocols for the pull-down assay, nanoflow LC-MS/MS, database searches and bioinformatics analysis for functional classification are referred to our previous report [8].

Immunoprecipitation (IP) and Western blot (WB)
HEK-293T cells were gently washed in PBS for three times and then lysed for 20 min. using IP-Western blot lysis buffer (P0013; Beyotime Biotechnology, Inc.) in the presence of a complete protease inhibitor cocktail, 1 lM sodium orthovanadate and 1 mM sodium fluoride per 6cm dish on ice. Cell lysates were transferred into 1.5-ml microcentrifuge tubes. Extracts were centrifuged for 10 min. at 14007g at 4°C. Then, 300 ll of supernatants were pre-cleaned with 20 ll of Protein A/G PLUS-Agarose (sc-2003; Santa Cruz Biotechnology, Inc.) for 1 hr and immunoprecipitated using 2 ll of DYKDDDDK Flag-tag mouse mAb (M20008; Shanghai Abmart, Shanghai, China) or GFP-Tag mouse Ab (E022030-01; EarthOx, LLC, San Francisco, CA, USA) at 4°C for 2 hrs and mixed with 20 ll of Protein A/G PLUS-Agarose (sc-2003, Santa Cruz Biotechnology, Inc.) at 4°C for 8 hrs to perform immunoprecipitation assay. The immunoprecipitates were washed with PBS three times and subjected to SDS-PAGE and Western blotting. The primary antibodies used in IP-WB analysis were DYKDDDDK Flag-tag mouse mAb (M20008; Shanghai Genomics, Shanghai, China), HA-tag mouse mAb (SG4110-25, Shanghai Genomics), SASH1 rabbit pAb (A302-265A-1; Bethyl Laboratories, Inc. Montgomery, Texas, USA) MAP2K2(MEK1/2) (D1A5) rabbit mAb(#8727; Cell Signaling Technology, Inc, Danvers, MA, USA), b-tubulin mouse Ab (E022030-01; EarthOx, LLC) and GAPDH mouse mAb (M20005; Shanghai Abmart, Inc.). HEK293T transfected cells and NHEMs were gently washed in PBS for three times and then lysed for 20 min. using IP-Western blot lysis buffer. Cell lysates were transferred into 1.5-ml microcentrifuge tubes. Extracts were centrifuged for 10 min. at 12,000 r.p.m. at 4°C. The primary antibodies used in the Western blot experiments were showed in Table S3. Immunoblotting was performed as previously described [16].

Immunohistochemical and melanin staining
Immunohistochemical staining Written informed consent regarding tissues and data used for scientific purposes was obtained from all participating patients. The study was approved by the Research Ethics Committees of Institutes of Biomedical Sciences, Fudan University, Shanghai, China, and Yongchuan Hospital, Chongqing Medical University, Chongqing, China. The fresh human foreskin tissues from a 14-years-old boy were fixed, pulled expansion and spread on a foam board and irradiated for different time length to reach 0.01, 0.05, 0.10, 0.50, 1.00 and 2.00 J/cm 2 , six doses of UVB (wavelength: 311 nm). After irradiation, these irradiated foreskin tissues and normal control (NC) and affected skin epithelial tissues from Y551D-SASH1-mutation individuals were fixed in 10% formalin at 4°C for 24 hrs and embedded in paraffin for immunohistochemistry analyses. The sections covered with foreskin tissues and skin epithelial tissues were deparaffinized and rehydrated using xylene and an ethanol gradient. Sections were incubated with p53 monoclonal antibody(mAb), ACTH polyclonal antibody (pAb), SASH1 pAb, phospho-ERK1/2 mAb, phosphor-CREB(Ser133) mAb, Mitf pAb, SILV mAb, tyrosinase pAb and TYRP1 mAb as well as horseradish peroxidase-linked anti-rabbit and anti-mouse universal secondary antibodies. The detailed usage information of these antibodies is shown in Table S3. Finally, sections were counterstained with haematoxylin and photographed under the positive position microscope BX51 (Olympus, Beijing, China). The staining intensity and percentage of positive cells were calculated and scored in sections of each UV irradiation doses. The staining intensity of p53-, ACTH-, SASH1-, p-ERK1/2-, p-CREB-, Mitf-, Pmel17-, TYRP1-, SILV-, tyrosinase-positive cells was divided into four grades as 'negative (À)', '+', '++' and '+++'and calculated as 0, 1, 2 and 3 score, respectively. The positive cells' percentage was divided into six categories as 'negative (À)', '1-20%', '21-40%', '41-60%', '61-80%' and '81-100%' and calculated as 0, 1, 2, 3, 4, 5 and 6 score,respectively. Total scores of every visual field were determined by the formula  3 min. successively. Then, the GENMED cleaning liquid was removed and replaced with GENMED dyeing working fluid for about 30 min. at 45°C until dark brown of tissue slices appeared. GENMED dyeing working fluid was then removed and replaced with GENMED cleaning liquid for 2-min. incubation,and GENMED cleaning liquid was removed and replaced with GENMED equilibrium liquid for 3-min. incubation. GENMED equilibrium liquid was removed and replaced with GENMED cleaning liquid for 3-min. incubation. Finally, the cleaning liquid was removed and incubated with GENMED stabilizing buffer for 2 min. GENMED stabilizing buffer was removed from the tissue slices, and tissue slices were washed with GENMED cleaning liquid at room temperature for 2 min. GENMED cleaning liquid was removed, and transparent processing of tissue slices was implemented, and tissue slices were mounted with neutral resin and photographed under the positive position microscope BX51 (Olympus). The detailed information of melanin staining reagents was as described as the manufacturer's protocol suggested.

Quantification of intracellular cAMP levels
To identify the NHEMs (purchased from Promocell) and the HEK-293T cells that were transfected HA-MC1R-pcdna3.0 vector beforehand have MC1R responsiveness, NHEMs and HEK-293T transfected cells on 6well plates with four groups of repetition were starved for 8 hrs and treated with 10 À5 M of NDP-MSH for 30 min., respectively. HEK293T transfected cells and NHEMs were diluted with 0.1 M HCL and incubated at room temperature for 20 min. and scraped with a cell scraper. After homogeneous bleeding by pipetting up and down, cell lysates were transferred into a 1.5-ml centrifuge tube and centrifuged for 10 min. Protein concentrations of HEK293T transfected cells and NHEMs were determined by BCA kit (Beyotime Biotechnology Inc.) with bovine serum albumin (BSA) as a standard. Cell samples and standard cAMP stock were treated with acetylating reagent. The detailed subsequent procedures for detection of cAMP levels and producing a standard curve were as per the manufacturer's protocol (cAMP Direct Immunoassay Kit; BioVision Shanghai, China). Finally, absorbance readings were determined at OD 450 , and relative cAMP levels in HEK-293T cells and NHEMs were calculated.

Tyrosinase activity and melanin content assays
Tyrosinase activity analysis.
NHEMs transfected or infected with WT and mutant SASH1 or transfected or infected NHEMs silenced by MAP2K2 and CREB siRNAs were washed with GenMed cleaning liquid provided by the manufacturer (GenMed Scientifics Inc.), lysed with GenMed lysis buffer and centrifuged at 16,000 9 g for 5 min. to obtain the supernatant for tyrosinase activity and protein concentration analyses. The protein concentration was determined with Bradford protein concentration quantification kit (GenMed Scientifics), and the remaining supernatant was placed on the ice for tyrosinase activity analyses. Based on the principle that dopachrome can be generated by tyrosine under the catalysis of tyrosinase, activity of tyrosinase can be determined by the changes in dopachrome absorbance at 475 nm wavelength. GENMED buffer and GENMED substrate solution were mixed and filled with oxygen for 2 min. Negative control liquid and 50 ll of remaining supernatant were added into the mixed solution and mixed with up-down blending for background control measurement and sample tyrosinase activity determination, and the mixture was incubated at room temperature for 60 min. Melanin content assays.
NHEMs infected with WT and mutant SASH1 were washed with GenMed cleaning liquid provided by the manufacturer (GenMed Scientifics Inc.), lysed with GenMed lysis buffer at 4°C for 30 min. and proceeded with 16,000 9 g centrifugation for 5 min. at 4°C. The cell lysates of NHEMs were centrifuged at 16,000 9 g for 5 min. at 4°C, and 10 ll of supernatant was implemented to protein concentration determination. 500 ll of GENMED treating fluid was added to the remaining supernatant and centrifuged at 16,000 9 g for 5 min. The supernatant was removed, and 500 ll of GENMED dissolving solution was added into cell deposits and mixed. 500 ll of GENMED buffer was added and stored at 60°C for 30 min. avoiding exposure. The melanin content was determined by measuring the absorbance at 360 nm wavelength. Meanwhile, 500 ll of GENMED buffer was added to five numbered Eppendorf centrifuge tubes and 500 ll of GENMED melanin standard solution was added to No.1 Eppendorf centrifuge tube and mixed. 500 ll of mixture transferred from No.1 Eppendorf centrifuge tube was added into No.2 centrifuge tube. And the remaining tubes were added in turn. Finally, the melanin standard solution was also added with 500 ll of GENMED dissolving solution and GENMED buffer and incubated at 60°C for 30 min. The absorbance of standard melanin content at 360 nm wavelength was determined, and the standard curve was drawn to analyse the melanin content of samples.

Statistical analyses
The data are presented as mean AE standard derivation (S.D.). These data were first analysed using the homogeneity of variance test, followed by the change of variable test. Statistical significance was determined by a one-factor analysis of variance (ANOVA) using LSD correction on (IBM (International Business Machine), Shanghai, China) to generate the required P-values. Cartograms were plotted using GraphPad Prism 5 (GraphPad Software, Inc.La Jolla, CA, USA).

SASH1 is induced by the p53/POMC cascade
Our previous study demonstrated that SASH1 protein binds to Gas [8]. Moreover, it has been shown that Gas, a major activator of adenylyl cyclase and cAMP production in melanocytes, is stimulated by a-MSH and melanocortins [17]. a-MSH and other bioactive peptides are identified to be the cleavage products of pro-opiomelanocortin (POMC), a multicomponent precursor for a-MSH (melanotropic hormone), ACTH (adrenocorticotropic hormone) and the opioid peptide b-endorphin. Normal synthesis of a-MSH and ACTH is an important determinant of constitutive pigmentation in human beings and the cutaneous response to UV [3]. P53 directly stimulates the POMC promoter in response to UV and also induces UV-independent pathologic pigmentation [13]. Hence, to determine whether SASH1 is mediated by p53 and POMC, the MC1R ligand, exogenous p53 and exogenous POMC were introduced into HEK-293T cells pre-transfected with exogenous MC1R and NHEMs to investigate the effects of MC1R ligands on SASH1. Endogenous SASH1 was found to be induced by introduction of exogenous p53 and POMC in NHEMs and HEK-293T transfected cells ( Fig. 1A and B).
We also investigated SASH1 responsiveness to POMC inducement. Exogenous POMC was introduced into NHEMs and HEK-293T transfected cells. POMC was shown to increase expression of endogenous SASH1 (Fig. 1C and D). As documented in Figure S1, NHEMs and HEK-293T transfected cells were stimulated by NDP-MSH to secrete more intracellular c AMP as compared to that of DMSO control, which indicated that NHEMs and the transfected HEK-293T cells had MC1R responsiveness.
UV-mediated induction of POMC-MSH signal cascade is directly regulated by p53 because that p53 stimulates the POMC promoter in response to UV [13]. And as SASH1 is identified to be responsive to POMC, the ligand of MC1R signal pathway, we intend to identify whether SASH1 is responsive to p53 activation in vivo. Fresh discarded normal human foreskin specimens were pulled expansion, spread and exposed to gradually increased UV irradiation and stained for the histological analyses of p53, ACTH and SASH1. Immunohistochemical (IHC) analyses revealed p53 is rapidly induced in basal layers at the 0.10 J/cm 2 dose of UV irradiation. The induction of SASH1 and ACTH at 0.50 J/cm 2 dose of UV irradiation in melanocytes is induced by UVB irradiation (Fig. 2). Taken together, these results indicate that SASH1 is induced by p53/POMC signal cascade.

SASH1 binds with MAP2K2, and SASH1 mutations promote binding between SASH1 and MAP2K2
To investigate which pathway that SASH1 is involved in to mediate melanogenesis, pull-down assay, nanoflow LC-MS/MS and bioinformatic analyses were performed to investigate the binding partners that SASH1 is associated with in A375 stable cells.
Nanoflow LC-MS/mass spectrometry and bioinformatics analyses indicated that SASH1 may be involved in three pathways involved in or mediating melanin biosynthesis: the insulin/IGF pathway, the mitogen-activated protein kinase kinase(MAPK)/MAP kinase cascade and the Ras pathway (Table S1). Obviously, MAP2K2 (P36507), a dualspecificity protein kinase that belongs to the MAP kinase kinase family [18], was shown to be involved in ten pathways including MAPK cascade. And MAP2K2 was showed to have a certain possibility to bind with SASH1 ( Fig. 3A and Table S2). Additionally, SASH1 was showed to activate NF-kB and MAPKs through TLR4 in human endothelial cells to increase production of proinflammatory cytokines [19]. To identify the endogenous protein associations between SASH1 and MAP2K2, immunoprecipitation-Western blot (IP-WB) was performed to identify the associations between SASH1 and MAP2K2. As shown in Figure 3B and C, exogenous SASH1 (GFP-SASH1 or Flag-SASH1) is associated with endogenous MAP2K2 or exogenous MAP2K2 (HA-MAP2K2). And immunoprecipitation with SASH polyclone antibody was performed. As demonstrated in Figure 3D, endogenous SASH1 was showed to immunoprecipitate with endogenous MAP2K2 in NHEMs. We further identify the binding domain of SASH1 to MAP2K2. Deleted SASH1 constructs were created and transfected in HEK-293T cells and immunoprecipitated with GFP antibody. As compared with the binding of SASH1 full-length fragment to MAP2K2, SH3-SAM domain of SASH1 was showed to bind to MAP2K2 (Fig. 3E and F). So, it is believed that SASH1 is associated with MAP2K2 through its SH3-SAM domain corresponding to the Sly-homology domain. We also assessed the effects of SASH1 mutations on MAP2K2. We also further analysed the effects of SASH1 mutations on SASH1-MAP2K2 interaction. IP-WB analyses indicated Flag-SASH1 was immunoprecipitated, and the associated HA-MAP2K2 was detected by Western blot analyses using an anti-HA antibody. (D) Endogenous SASH1 is associated with endogenous MAP2K2 in melanocytes. 90% of cell density of NHEMs after 24-hrs culture was lysed and immunoprecipitated with SASH1 antibody, and the associated immunoprecipitated endogenous MAP2K2 was analysed by Western blot. (E) Schematic view of SASH1 deletion constructs. SASH1 contains two predicted nuclear localization signals (NLS1 and NLS2), the conserved Src-homology 3 domain (SH3) and the two sterile alpha-motifs (SAM1 and SAM2). Amino acid residues of the SASH1 coding sequence are indicated. Four deleted SASH1 fragments were cloned to pEGFP-C3 vector and expressed as GFP-deleted SASH1 fusion proteins. (F) and (G) C-terminal domain (590aa-1247aa) of SASH1 binds to MAP2K2. Four deleted SASH1 fragments including SH3-SAM construct, DC-terminal construct, DN-terminal construct and C-terminal construct and SASH1 full length were cloned into pEGFP-C3 vector and transfected HEK-293T cells. After 36-hrs transfection, cells were lysed and immunoprecipitated with GFP antibody, and the associated endogenous MAP2K2 was detected by Western blot analyses using MAP2K2 antibody. As compared with GFPfull-length SASH1, C-terminal domain (590aa-1247aa) of SASH1 was showed to be associated with endogenous MAP2K2 (indicated by black arrows). (H) SASH1 mutations enhanced the associations between SASH1 and MAP2K2. Wild-type and mutant SASH1-pEGFP-C3 were transfected to HEK-293T cells. At 36 hrs after transfection, transfected cells were lysed and subjected to immunoprecipitation with GFP antibody and the associated MAP2K2 was analysed by Western blot. Left panel: Equal amounts of protein lysates or immunoprecipitates were resolved for Western blotting or immunoprecipitation with anti-GFP or anti-MAP2K2 antibody. Right panel: endogenous MAP2K2 which was immunoprecipitated by GFP-SASH1 was quantified. ※ P < 0.01 compared to WT (wild-type) groups. NHEMs: normal human epithelial melanocytes; SASH1: SAM and SH3 domain containing 1. that mutated SASH1s showed strong binding ability to MAP2K2 (Fig. 3G).

SASH1 mutations regulate ERK1/2/CREB cascade
We further assessed the effects of SASH1 on the phosphorylation of ERK1/2 downstreams of MAP2K2 and phosphorylation levels of CREB. Gradually increasing doses of exogenous SASH1 were firstly introduced to NHEMs and HEK-293T cells. Immunoblot analyses showed that SASH1 increased phosphorylation levels of ERK1/2 and CREB ( Fig. 4A and B).We also assess the effects of SASH1 silencing on phosphorylation levels of ERK1/2 and CREB. Immunoblotting showed that SASH1 knockdown down-regulated the phosphorylation levels of ERK1/2 and CREB (Fig. 4C). Additionally, wild-type and mutant SASH1 were introduced into HEK-293T cells and NHEMs to assess the effects of SASH1 mutations on phospho-ERK1/2 and phospho-CREB. SASH1 mutations increased phosphorylation levels of ERK1/2 and CREB in HEK-293T cells and NHEMs (Fig. 4D and E). Meanwhile, the effects of p53 activation upon UV irradiation and SASH1 mutation on phosphorylation of ERK1/2 and CREB were also assessed in vivo. The fresh human foreskin tissues from a 14-year-old boy exposed at different doses of UVB intensity were used for immunohistochemistry analysis. IHC analyses demonstrated that p53 activation caused by UV irradiation induced increase in phospho-ERK1/2 and phospho-CREB levels at 0.5 J/cm 2 dose (Fig. 5). IHC analyses also showed that Y551D mutation enhanced cytoplasmic and nuclear expression of phosphorylated ERK1/2 and phospho-CREB in the affected epithelial layers as compared to those of normal control (NC) epithelial layers (Fig. 6). Mitf antibody was used to mark melanocytes of affected epithelial layers, and phospho-ERK positive-and phospho-CREB-positive epithelial cells are melanocytes.

Melanogenesis is induced by p53-POMC-SASH1 cascade
We also investigated whether melanogenesis-specific molecules were responsive to activation of p53-POMC-SASH1 cascade. IHC analyses of Mitf, Pmel17 and TYRP1 revealed that the inducement of Mitf started at the 0.05 J/cm 2 dose of UV irradiation in basal layers, reached the maximum at 0.5 J/cm 2 dose and decreased at 1.00 J/cm 2 dose. However, the induction of Pmel17 and TYRP1 started at 0.05 J/cm 2 dose and reached the maximum at 1.00 J/cm 2 (Fig. 7). IHC analyses of SILV and tyrosinase also revealed that SILV was induced at 0.05 J/cm 2 dose of UV irradiation and reached the maximum at 1.0 J/cm 2 dose. However, no obvious inducement of tyrosinase was observed (Fig. S2).

Mutated SASH1s are induced by p53, and hyperpigmentation is induced by SASH1 mutations
Because the ERK1/2 signalling pathway is identified to be responsive to p53-POMC-SASH1 cascade, and the ERK1/2 signalling pathway mediates melanogenesis [10,11], we sought to address whether mutated SASH1s are induced by p53 and mutations in SASH1 promote biosynthesis of melanogenesis-specific molecules. Our previous study [9] showed that SASH1 mutations promote the expression of exogenous p53 and more endogenous p53 was induced by Y551D-SASH1 mutation. Hence, we intend to identify whether p53 could promote mutated SASH1 expression. Our study indicated that mutated SASH1s were more subjected to be induced than wild-type of SASH1 by p53 in the presence of exogenous MC1R in HEK-293T cells (Fig. 8A). So, we deduced that p53 activation plays a pivotal role in the pathogenesis process of this dyschromatosis disorder and activation of p53-POMC-SASH1 cascade triggers the formation of hyperpigmentation phenotype.
Because A375 cells cannot synthesize melanin (data not showed), we used NHEMs to verify the effects of SASH1 on melanin synthesis. In vitro analysis indicated that increased melanin biosynthesis is induced by SASH1 mutations (Fig. 8C). We further tested the enzymatic activity of the rate-limiting melanogenic enzyme tyrosinase. As expected, we found that SASH1 overexpression and mutations in the SASH1 gene induced increased tyrosinase activity in NHEMs (Fig. 8B). To identify the regulation of SASH1/MAP2K2/ERK1/2/CREB cascade to melanogenesis, in NHEMs transiently expressing mutant and wild-type SASH1, endogenous MAP2K2 and endogenous CREB were silenced by their specific siRNA. Deletion of MAP2K2 and CREB could decrease tyrosinase activity and melanin production ( Fig. 8D-I).
Immunohistochemical (IHC) analyses also revealed heterogeneous expression of SASH1, Pmel17, tyrosinase, TYRP1 and melanin in affected epithelial tissues with Y551D mutation. High levels of SASH1, Pmel17, tyrosinase and TYRP1 expression were found in some regions of affected epithelial layers in Y551Daffected individuals (Fig. 9). Excessive melanin secretion was observed in both basal and superbasal layers of affected tissues (Fig. 9).
Taken above, we can conclude that a novel p53/POMC signal cascade transactivates ERK1/2/CREB signalling pathway through a novel SASH1/MAP2K2 crosstalk to mediate melanin biosynthesis (Fig. 10). Mutated SASH1 serves as a molecular motor to activate the p53/ POMC signal cascade, increasing phosphorylated ERK1/2 and CREB to induce pathological hyperpigmentation.

Discussion
Dendrite extension and melanin production or secretion are manifestations of melanogenesis that depend on elevated intracellular cAMP levels and are tightly regulated at numerous steps. The cAMP-dependent activation of Ras protein leading to ERK activation independent of protein kinase A (PKA) and Rap1 is an important part of the molecular underpinnings of melanogenesis [20]. Furthermore, recent study had showed that SASH1 is showed to be involved in autosomal dominant lentiginous [21] and autosomal-recessive

2475
SASH1 variants are associated with a new genodermatosis with a pigmentation defects, palmoplantar keratoderma [22].Our work herein suggests that hyperpigmentation phenotype of dyschromatosis universalis hereditaria is attributed to three novel SASH1 point mutations. SASH1 could be a novel exchange factor, which is mediated by p53-POMC-MC1R-cascade, and regulates MAP2K2/ERK1/2/CREB signal pathway to induce melanogenesis process. SASH1 mutations intensify activation of p53-POMC-MC1R-SASH1 cascade and binds with MAP2K2 to mediate ERK1/2 signal pathway, leading to hyperpigmentation of this dyschromatosis disorder.
P53 directly mediates UV induction of POMC/MSH in skin through stimulation of the POMC promoter in response to UV and triggers UV-independent pathologic pigmentation that could mimic the tanning response [13]. Such agonists as a-melanocyte-stimulating hormone (a-MSH) or adrenocorticotropic hormone stimulate MC1R to induce increased melanocyte proliferation and dendricity [19,23], melanosome transfer [24] and an elevated ratio of the darker eumelanin pigment over the red/yellow pheomelanin pigment [25]. Our findings document SASH1 is induced by p53/POMC cascade (Figs 1  and 4). POMC stimulation induces elevated SASH1 expression ( Fig. 1C and D). Hence, SASH1 serves as a novel partner to involve in p53-POMC-MC1R cascade.
The role of G protein subunit-coupled receptors in cell proliferation is rather complex, although the activation of Gs and PKA is generally associated with growth inhibition [26]. This study provides new insight into the GPCR signalling pathway. SASH1 can respond to MC1R stimulation by POMC. PKA-independent and cAMP-dependent pathways have been linked to the regulation of melanogenesis, including the phosphatidylinositol 3-kinase and Ras/Erk pathways [20]. Contrary to many other cell types, B-Raf, mitogen-activated protein kinase/extracellular signal-regulated kinase (MEK) and ERK have been shown to be activated through Ras induction by elevated cAMP in B16 melanoma cells [1,10,20,27]. Ras/ERK activation in these cells has a short-term stimulatory effect, whereby ERK phosphorylates Mitf, leading to increased transcriptional activity of the tyrosinase promoter [10,28,29]. MAP2K2 was initially indentified as an ERK activator from growth factor-treated cells and phosphorylates MAPK1/ERK2 cascade and MAPK2/ ERK3 cascade [18]. A novel c.383C̀->A transversion in exon 3 of MAP2K2 and 19p13.3 microdeletion involving the mitogen-activated protein kinase kinase 2 gene (MAP2K2) were reported to cause the cardio-facio-cutaneous (CFC) syndrome [30,31]. In this study, MAP2K2 is associated with SASH1 to form a novel SASH1/MAP2K2 crosstalk to regulate phosphorylation level of ERK1/2 and CREB, which suggests that SASH1 is involved in MEK/ERK signal cascade (Fig. 3).
Caveolin-1-mediated activation of ERK/CREB signalling may be involved in neuronal transmission pathways implicated in chronic neuropathic pain modulation [32]. Activation of phospho-ERK may partly be accomplished by CREB-dependent gene expression in chronic constrictive injury in rats [33]. Blocking of transient receptor potential canonical (TRPC) channels resulted in suppression of PDGF-induced Pyk2/ERK/CREB activation, but not PI3K/Akt pathway [34]. BMP-6 plays a crucial role in survival of cerebellar granule cells via the noncanonical MEK/ERK/CREB pathway [35]. One of the main targets of the MEK/ERK pathway is CREB. CREB activation has been implicated in synaptic plasticity, learning, and memory and cell survival [36,37]. Our study indicates that the noncanonical MEK/ERK/ CREB pathway is mediated by SASH1 (Fig. 4A and B) and SASH1 is necessary for the regulation of ERK/CREB cascade by p53/POMC/ MC1R cascade (Fig. 4C). Moreover, SASH1 alleles aggravate the activation when overexpressing in HEK-293T cells and NHEMs (Fig. 4D  and E).
Biosynthesis of melanin, the primary pigment synthesized by mammals, is sequestered within melanosomes, which are unique membrane-enclosed structures of melanocytes. Consistent with their singular structure and function, melanosomes contain specific resident proteins expressed only in melanocytes and retinal epithelial Fig. 8 Elevated melanin secretion and increased tyrosinase activity were induced by SASH1 mutations, however, eliminated by silencing MAP2K2 and CREB. (A) Mutated exogenous SASH1 was more induced than wild-type of exogenous SASH1 by p53 in HEK-293T cells. 1 lg of HA-p53 plasmid and 1 lg of GFP-SASH1 (wild-type and mutated) were cotransfected into HEK-293T cells pre-transfected with HA-MC1R plasmid. Cells were lysed and subjected into Western blot analysis along with GAPDH as loading control. Relative protein levels of endogenous SASH1 (bottom panels) were determined by Western blot (top panels) in three independent experiments. ※ P < 0.01 compared to 0 lg groups. (B) and (C) SASH1 mutations increased the activity of tyrosinase (B) and resulted in elevated melanin secretion (C) in NHEMs. NHEMs were infected with wild-type and mutant SASH1 in 10-cm dish and cultured in FBS-free M2 medium. At 48 hrs after infection, cells were washed gently and lysed with IP-WB lysis buffer. The tyrosinase activity analysis and melanin content assay were performed as per the instruction of manufacturer and determined using oneway ANOVA with LSD correction on SPSS version 16.0 and showed as the mean AE S.D. Cartograms were plotted using GraphPad Prism 5. ※ denotes P < 0.01. # denotes P < 0.05. (D) MAP2K2 was silenced by two specific MAP2K2 siRNAs efficiently as indicated by Western blot of NHEMs. (E) MAP2K2 silencing by its specific siRNAs induced decreased tyrosinase activity in NHEMs transiently transfected with wild-type and mutant SASH1 genes as well as empty pEGFP-C3 vector. The tyrosinase activity was determined using one-way ANOVA with LSD correction on SPSS version 16.0 and showed as the mean AE S.D. ※ denotes P < 0.01. # denotes P < 0.05. (F) MAP2K2 silencing by its specific siRNAs decreased melanin secretion in NHEMs transiently transfected with wild-type and mutant SASH1 genes as well as empty pEGFP-C3 vector. Melanin content was determined using one-way ANOVA with LSD correction and showed as the mean AE S.D. (G) Two specific CREB siRNAs could efficiently down-regulate CREB expression which was silenced as indicated by Western blot of NHEMs. (H) CREB silencing induced decrease in tyrosinase activity in NHEMs. (I) CREB deletion by its specific siRNAs caused decreased melanin secretion in NHEMs. Tyrosinase activity and melanin content were determined using one-way ANOVA with LSD correction and showed as the mean AE S.D. ※ denotes P < 0.01. NHEMs: normal human epithelial melanocytes; SASH1: SAM and SH3 domain containing 1. cells. Most of these characterized proteins are integral membrane proteins, such as tyrosinase, tyrosine-related protein 1 (TYRP1) and Pmel17 (gp100). Pmel17 is a matrix protein of premelanosomes, whereas TYRP1 is a component of mature melanosomes [38]. Pmel17 is first enriched in stage I-like structures and accumulates to a higher extent in stage II premelanosomes. In contrast, TYRP1 is almost undetectable in stage I structures, and only a small percentage of labelling is observed in stage II. These published results suggest that Pmel17 is a major biogenetic component of the striations [39]. Our study suggests that SASH1 point mutations result in increased expression levels of Pmel17, TYRP1 and tyrosinase, leading to melanin production and secretion in vivo (Fig. 9). More importantly, our studies demonstrate that p53 regulates the melanogenesis process in a SASH1-dependant manner, which will underscore the p53 role in the formation of hyperpigmentation phenotype in dyschromatosis.   Table S1 Bioinformatic analysis of pathways that SASH1 may involve in Table S2 The peptide sequences of the SASH1 complex identified by SBP-FLAG-SASH1 affinity purification