A melanosomal two-pore sodium channel regulates pigmentation

Intracellular organelles mediate complex cellular functions that often require ion transport across their membranes. Melanosomes are organelles responsible for the synthesis of the major mammalian pigment melanin. Defects in melanin synthesis result in pigmentation defects, visual deficits, and increased susceptibility to skin and eye cancers. Although genes encoding putative melanosomal ion transporters have been identified as key regulators of melanin synthesis, melanosome ion transport and its contribution to pigmentation remain poorly understood. Here we identify two-pore channel 2 (TPC2) as the first reported melanosomal cation conductance by directly patch-clamping skin and eye melanosomes. TPC2 has been implicated in human pigmentation and melanoma, but the molecular mechanism mediating this function was entirely unknown. We demonstrate that the vesicular signaling lipid phosphatidylinositol bisphosphate PI(3,5)P2 modulates TPC2 activity to control melanosomal membrane potential, pH, and regulate pigmentation.

a dermal melanocyte cell line derived from mice deficient in ocular albinism 1 (Oa1 −/− ) 18 , in which melanosomes are enlarged up to 1.5 μ m diameter 19 . Melanosomes dissected from Oa1 −/− melanocytes were patch-clamped using a NaCl-based pipette solution to mimic luminal conditions and a K + -based bath/cytoplasmic solution containing the impermeant anion gluconate (Gluc − ) (Fig. 1a). Basal whole-melanosome currents were outwardly rectifying with a negative reversal potential (E rev ) characteristic of OCA2-mediated melanosomal Cl − currents 20 (I basal ). We investigated if melanosomal channel activity could be regulated by PI(3,5)P 2 , an organellar phosphatidylinositol bisphosphate that is important in pigmentation 3 and activates the endolyosomal cation channels transient receptor potential mucolipin (TRPML) 4 and two-pore channels (TPCs) 5 . Cytoplasmic treatment with PI(3,5)P 2 activated a large inward whole-melanosome current and shifted the E rev in the positive direction, but did not significantly affect I basal outward current amplitude (Fig. 1b). To isolate the PI(3,5)P 2 -activated current (I PIP2 ), we subtracted I basal from the total current in the presence of PI(3,5)P 2 and found a large inward current with a very positive E rev (I PIP2 ) (Fig. 1b). When we minimized Cl − -dependent I basal by replacing luminal Cl − with the impermeant anion Gluc − , we found that PI(3,5)P 2 activated a current (I PIP2 ) with similar characteristics revealed by I basal subtraction (Fig. 1c). Furthermore, in melanosomes from melanocytes expressing OCA2 siRNA that have reduced I basal , PI(3,5)P 2 activated a similar inward current ( Supplementary Fig. S1a). These results suggest that PI(3,5)P 2 activates a current distinct from and independent of the outward rectifying OCA2-mediated Cl − current. Because inward I PIP2 could result from anions moving into the melanosome but is independent of the major permeant anion Cl − , we conclude that it is mediated by cations moving out of the melanosome. In response to voltage ramps, we recorded an outwardly rectifying current with a negative E rev (black). Application of PI(3,5)P 2 activated an inward current (I PIP2 ) and shifted the E rev in the positive direction (blue). Subtracting the basal current from I PIP2 revealed an inward current with a very positive E rev (red) (representative of 7 melanosomes). (c) Substituting luminal Cl − for Gluc − reduced the outwardly rectifying basal current amplitude and shifted E rev in the positive direction. Application of PI(3,5)P 2 activates an inward current similar to I PIP2 (representative of 5 melanosomes, p < 0.0001 for the difference in current measured at − 120 mV before and after PI(3,5)P 2 ). (d) Melanosomes dissected from freshly isolated bullfrog RPE were used for patch-clamp experiments (scale bar = 10 μ m). (e) In a representative RPE melanosome PI(3,5)P 2 activated an inwardly rectifying current similar to I PIP2 measured in dermal melanosomes. The PI(3,5)P 2 activated component of the current (red) obtained by subtracting the basal current from I PIP2 was inward rectifying, with a positive reversal potential (representative of 3 melanosomes). (f) Substituting luminal Cl − for Gluc − reduced the outwardly rectifying basal current (black) and revealed the inwardly rectifying current activated by PI(3,5) P 2 (blue) (representative of 3 melanosomes, p < 0.0001 for the difference in current measured at − 120 mV before and after PI(3,5)P 2 ). We next tested if the PI(3,5)P 2 -dependent current could also be measured in melanosomes found in the retinal pigment epithelium (RPE) of the eye. By patch-clamping melanosomes from RPE cells of the American bullfrog, Lithobates catesbeianus (Fig. 1d), which are larger than those from mammalian RPE cells, we measured a current with properties similar to dermal melanosomal I PIP2 : PI(3,5)P 2 activated an inward current that did not affect I basal (Fig. 1e). When basal currents were subtracted from these recorded in the presence of PI(3,5)P 2 , an inward current with a very positive E rev was revealed (I PIP2 ) (Fig. 1e). Substituting luminal Cl − with Gluc − also reduced I basal and isolated I PIP2 (Fig. 1f). I PIP2 has properties similar to endolysosomal TPC2. To further characterize the properties of I PIP2 we used a Gluc − -based luminal solution to reduce the contribution of the Cl − -dependent outwardly rectifying current (I basal ). In endolysosomes PI(3,5)P 2 , but not PI(4,5)P 2 , has been shown to activate both TRPML and TPCs 4,5 . To test if phosphoinositide regulation of melanosomal I PIP2 is specific to the organellar species of PIP 2 , we bath-applied the plasma membrane specific PI(4,5)P 2 while in whole-melanosome configuration. Basal currents were unaffected by PI(4,5)P 2 treatment, but, following a wash period, a large inward current was elicited by application of PI(3,5)P 2 to the same melanosome (Fig. 2a). To test if TRPML channels contribute to the PI(3,5)P 2dependent inward current, we used the TRPML-specific agonist ML-SA1 7 . ML-SA1 treatment did not elicit a significant increase in basal currents, while subsequent application of PI(3,5)P 2 evoked large inward currents in the same melanosomes (Fig. 2b). Additionally, acidifying luminal pH from 6.8 to 4.6, a known regulator of TRPML channels, had no effect on I PIP2 (Supplementary Fig. S1c). Treatment with verapamil, however, an antagonist of TPCs 5,9 , nearly abolished I PIP2 in both dermal (Fig. 2c) and RPE melanosomes ( Supplementary Fig. S1b), suggesting that TPCs might contribute to I PIP2 .
in melanosome pH ( Supplementary Fig. S1c), I PIP2 is not likely mediated by TPC1. Another major difference between the two types of channels is their voltage dependence: TPC1 mediates a depolarization-activated noninactivating conductance while TPC2 is voltage independent, similar to leak channels 11 . Because voltage ramps, as used in our study, mask the voltage-dependence of TPC1 activation, we used voltage steps to test for the presence of voltage-activated currents. Voltage steps between − 80 and + 80 mV did not elicit channel activation or measureable tail currents in melanosomes, but PI(3,5)P 2 application activated an inward current (Fig. 2d), suggesting that TPC2 is the most likely melanosomal candidate.
To determine if TPC2 is required for I PIP2 , we generated tpc2-deficient Oa1 −/− mouse melanocytes using the CRISPR-Cas9 system 23 (Supplementary Fig. S3a-d). In melanosomes from Oa1 −/− melanocytes expressing tpc2-targeted CRISPR-Cas9, PI(3,5)P 2 failed to activate an inward current (9 out of 13 melanosomes) (Fig. 4a,b). Expression of GFP-tagged human TPC2 in tpc2-targeted CRISPR-Cas9-expressing mouse melanocytes was sufficient to rescue I PIP2 , suggesting that I PIP2 is mediated by TPC2 (Fig. 4a,b). Because TPC2 has been previously implicated in pigmentation 12 and we now find that it functions in melanosomes, we tested if TPC2 expression in melanocytes contributes to pigmentation. We found that reducing the levels of TPC2 by either tpc2-targeted CRISPR-Cas9 in melan-a or by siRNA in Oa1 −/− melanocytes had markedly increased cellular melanin content, compared with control melanocytes (Fig. 4c and Supplementary Fig. S3e,f). Thus, TPC2 localizes to melanosomes in pigment cells, where it mediates I PIP2 to regulate melanogenesis.
How does TPC2 regulate melanosome function? Because TPC2 forms a Na + leak current, it is likely to contribute to melanosomal membrane potential. To test this, we performed current-clamp recordings of membrane potential in melanosomes from control and tpc2-targeted CRISPR-Cas9-expressing Oa1 −/− melanocytes in the presence of PI(3,5)P 2 . In melanosomes from tpc2-deficient melanocytes that did not exhibit a PI(3,5)P 2 -induced current the resting membrane potential (Ψ m , defined as Ψ m = V cytosol -V lumen ) was greatly reduced compared with melanosomes from control melanocytes (+ 7.9 ± 1.7 mV for tpc2-deficient and + 17.2 ± 0.9 mV for control, Fig. 4d). Thus, TPC2 activity modulates the melanosome's membrane voltage.
TPC2 modulates melanosomal pH to regulate pigmentation. The vacuolar proton pump V-ATPase is modulated by membrane voltage and is present in melanosomes where it regulates luminal pH and melanin production [24][25][26] . Because TPC2 significantly contributes to the membrane potential of melanosomes (p < 0.01 for Ψ m of control vs. tpc2-deficient melanocytes in the presence of PI(3,5)P 2 ), it is conceivable that TPC2-mediated changes in membrane potential modulate V-ATPase activity and melanosomal pH. To test this hypothesis, we sought to measure melanosomal pH from control and tpc2-deficient melanocytes.
Fluorescence-based pH measurements have not been possible in pigmented melanosomes because fluorescent indicator uptake is impaired and melanin interferes with fluorescence emission. To circumvent these difficulties we used B16-F1 mouse melanocytes, which are weakly pigmented in their basal state. To measure pH in the melanosomes of B16-F1 cells we used the albinism-associated V443I-OCA2 mutant as a melanosomal marker ( Supplementary Fig. S4) because it lacks Cl − channel activity, does not significantly increase organellar pH and only modestly induces pigmentation 20,27 . These missing features are present with wild-type OCA2 overexpression, which substantially increases pigmentation in B16-F1 melanocytes (Fig. 5b,c), thereby impairing fluorescence-based pH measurements.
Because our results show that TPC2 acidifies melanosomes and melanogenesis is enhanced at more neutral pH values [24][25][26] , we tested if TPC2 serves as a negative regulator of melanin content by examining its interaction with a positive regulator of melanogenesis, OCA2. Consistent with OCA2's function as a melanosomal anion channel component that raises organellar pH and enhances melanogenesis 20 , overexpression of OCA2 in B16-F1 melanocytes increased pigmentation by ~75%, as illustrated by the presence of numerous darkly pigmented melanosomes (Fig. 5b,c, top two right panels). We tested if TPC2 negatively regulates OCA2-induced pigmentation by coexpressing mCherry-OCA2 and TPC2-GFP, which colocalized (Fig. 5c, bottom right panels). Interestingly, TPC2-GFP expression reduced OCA2-induced pigmentation by ~30%, indicating that it negatively regulates OCA2-induced melanogenesis (Fig. 5b,c, faint appearance of melanosomes in bottom two right panels). In addition, expression of the albinism-associated mutant V443I-OCA2, which has reduced Cl − transport and pH regulation function 20 , also colocalized with TPC2, but resulted in only a slight increase in B16-F1 pigmentation (~27%) that was completely abolished by coexpression with TPC2-GFP (Fig. 5b,d). Thus, the melanosomal cation channel TPC2 acidifies pH to counterbalance the effect of the OCA2-mediated melanosomal anion channel on luminal pH and to serve as a negative regulator of melanogenesis and pigmentation.

Discussion
Our results reveal the first cation conductance in melanosomes. We find that in pigment cells the endolysosomal cation channel TPC2 localizes predominantly to melanosomes, where it mediates a Na + -selective current to modulate melanosomal membrane potential, pH, and pigmentation. Our findings are consistent with the role of TPC2 in pigmentation as a determinant of human hair color 12 , a factor in the development of melanoma 16,17 , and a regulator of pigmentation in Xenopus oocytes 13 .
We find that the important organellar signaling molecule PI(3,5)P 2 regulates the activity of TPC2 in melanosomes. PI(3,5)P 2 is a known modulator of endolysosomal cation channels 4,5 and affects pigmentation 3 , possibly by regulating melanosome biogenesis 28 . Our results show that PI(3,5)P 2 is required for the melanosomal TPC2-mediated Na + current; our experimental conditions do not allow us to evaluate whether in intact cells basal PI(3,5)P 2 is sufficient for TPC2 activity, resulting in a constitutively active Na + current or if an unknown physiological signal modulates PI(3,5)P 2 levels to regulate melanosomal TPC2 activity.
We show that TPC2 functions as a negative regulator of pigmentation by increasing melanosome membrane potential and melanosome acidity, which reduces the activity of the key melanogenic enzyme tyrosinase and subsequent melanogenesis. We hypothesize that TPC2 regulates melanosome pH by providing a cation counterflux to enhance V-ATPase H + transport into the melanosome lumen, consistent with the requirement for an inward cation current in lysosomal acidification 29 . We have recently shown that a melanosomal anion channel mediated by OCA2 reduces organelle acidity to enhance melanogenesis, possibly by decreasing melanosome membrane potential to inhibit V-ATPase activity 20 . We find that the effect of OCA2 on pigmentation is counterbalanced by TPC2, possibly through the regulation of luminal acidity (Fig. 6). It is likely the two conductances regulate one another through voltage-dependent processes and ion homeostasis; it is also possible that the two conductances do not function simultaneously to balance the pH, but rather are differentially regulated by cellular signaling to elicit dynamic changes in melanosomal pH. A more comprehensive model of ionic signaling in melanosomes requires a better understanding of melanosomal membrane potential and luminal ionic concentrations in intact cells; it also entails the molecular identification and characterization of other melanosomal ion channels and transporters. Thus, our studies provide a basis for future work exploring molecular mechanisms underlying ionic signaling in melanosomes and other organelles and could uncover novel therapeutic targets for pigmentation disorders and skin and eye cancers.
Analysis of CRISPR-Cas9 edited cell line. Genomic DNA was extracted using a Blood & Cell Culture Mini Kit (Qiagen) using > 100,000 cells per sample.
Single Clone Sequencing. Genomic DNA was extracted from CRISPR-expressing cells and the CRISPR target sites were amplified with the following primers: The PCR products were run on a 1.5% agarose gel and single bands of the predicted size were extracted using QIAquick Gel Extraction Kit (Qiagen). The resulting DNA products were cloned using the TOPO PCR kit (Invitrogen/Life Technologies) according to manufacturer's protocol. DNA from twenty single cell colonies was extracted and sequenced to screen for gene editing.
Scientific RepoRts | 6:26570 | DOI: 10.1038/srep26570 CRISPR-Cas9 Mutation Detection. PCR product was generated from purified genomic DNA using primers described for single clone sequencing to amplify tpc2 CRISPR sites. Mutation detection was carried out using the Guide-it Mutation Detection Kit (Clontech) according to manufacturer's protocol. In brief, 800-1000 ng of PCR products were re-annealed to enable heteroduplex formation (95 °C for 5 min, 95 °C to 85 °C ramping at 2 °C/s, 85 °C to 25 °C at 0.1 °C/s, and cooling at 4 °C). Re-annealing products were treated with Guide-it Resolvase for 30 minutes at 37 °C and analyzed on 1.5% agarose gels. The fraction of PCR product cleaved (F cut ) = (b + c)/ (a + b + c) where a = undigested product, b and c = digested products.
Mouse tpc2-targeted and control (scrambled) siRNAs were designed with Oligoengine 2.0 software and expressed using the pSUPER RNAi Vector system (Oligoengine). Tpc2-targeted oligos were cloned into pSUPER-GFP-NEO, and expressed in Platinum-E cells to produce retroviral particles. Upon viral transduction of Oa1 −/− mouse melanocytes, stable melanocytes lines expressing tpc2-targetd or control siRNA were created using fluorescence-activated cell sorting (FACS) and subsequently kept under G418 selection (800 μ g/ml). The relative tpc2 mRNA levels in melanocytes expressing tpc2-targeted or control siRNA were calculated by quantitative PCR relative to actin.
Melanosomal electrophysiology. Melanosome dissection. Enlarged dermal melanosomes were individually dissected from Oa1 −/− mouse melanocytes using a borosilicate patch pipette to cut the cell membrane and push out individual organelles. A new pipette was then used for patch clamp experiments. RPE melanosomes from bullfrog RPE cells were released using two patch pipettes, as previously described 20 .
Cation permeability. Relative permeability of I PIP2 was determined by measuring E rev after the substitution of bath/cytoplasmic cations from Na + to K + , Ca 2+ , or NMDG + . Pipette/luminal solution contained 140 mM NaGluc solution and the cytoplasmic/bath solution contained 140 mM X-Gluc, where X = Na + , K + , or NMDG + , or 100 mM CaGluc 2 . Permeability ratios were estimated using the Goldman-Hodgkin-Katz equation: P X / P Na = ([Na + ] Luminal /[X] Cytoplasmic )(exp(E rev F/RT)), P Ca /P Na = (4[Ca 2+ ] Luminal /[Na + ] Cytoplasmic )(exp(E rev F/RT)). Figure 6. Model of ion channel-mediated regulation of melanosome pH and melanogenesis. TPC2 mediates the major melanosomal Na + conductance to increase melanosome membrane potential (Ψ m , Ψ m = V cytosol -V lumen ) and provide a counter cation to enhance vacuolar ATPase (V-ATPase) activity and increase the acidity of the melanosome lumen, thus reducing the activity of the key melanogenic enzyme tyrosinase (TYR). Additionally, the major melanosomal anion channel mediated by OCA2 expression transports Cl − into the cytoplasm to make the melanosome membrane voltage more negative, thereby decreasing V-ATPase-mediated H + transport and luminal acidity, which enhances the activity tyrosinase and melanin production.
Scientific RepoRts | 6:26570 | DOI: 10.1038/srep26570 Immunofluorescence microscopy. Melan-a and B16-F1 melanocytes were seeded on coverslips coated with poly-L-lysine (Sigma) and transfected the following day using 4 μ g DNA and 3.3 μ l of Lipofectamine 2000 for melan-a or 2 μ g of DNA and 6 μ l of Lipofectamine 2000 for B16-F1. Cells were fixed using 4% paraformaldehyde (Sigma) for 10 min at room temperature, washed with PBS and labeled with primary and secondary antibodies diluted in PBS with 0.2% (w/v) saponin, 0.1% (w/v) bovine serum albumin, 0.02% (w/v) sodium azide. Antibodies used were: mouse anti-TYRP1 (TA99/mel-5, 1:50 dilution, BioLegend/Covance Antibody Products), rat anti-LAMP2 (GL2A7, 1:4, deposited by Granger, B.L. to the Developmental Studies Hybridoma Bank). Goat secondary antibodies conjugated to Alexa 568 or 488 were from Invitrogen. Cells were imaged using a laser-scanning confocal microscope (Zeiss LSM 510) with a 63X objective. Final images were generated from single 0.2 μ m stacks and insets magnified using Photoshop software (Adobe). Quantification of colocalization of GFP-tagged proteins and melanin-containing melanosomes or LAMP2 immunostaining was carried out using brightfield and fluorescent confocal images analyzed with ImageJ Fiji software, as previously described 31 . Briefly, after setting an initial threshold the images were made binary for each channel and the area corresponding to fluorescence or melanin-containing melanosomes was quantified. The Image Calculator function was used to multiply the binary images for brightfield melanosomes and fluorescence and the resulting image represents the area of overlap between fluorescently-tagged proteins and melanin-containing melanosomes. The area of overlap of the image resulting from multiplication and the binary image for total fluorescence staining was quantified using the Analyze Particles function to count the area of all structures with an area greater than 5 pixels. The ratio of overlap pixels to total fluorescence pixels gives the % overlap. pH imaging. Control or tpc2-targeted CRISPR-Cas9-expressing B16-F1 melanocytes transfected one day prior to imaging experiments with mCherry-V443I-OCA2 to identify melanosomes were incubated with 1 μ M LysoSensor-ND160 (Invitrogen/Life Technologies) for 5 min. LysoSensor was excited at 405 nm and its emission detected at 417-483 nm (W1) and 490-530 nm (W2) using an Olympus XL fluorescent microscope. The ratio of emissions (W1/W2) in melanosomes expressing mCherry-V443I-OCA2 was assigned to a pH value based on a calibration curve generated with solutions containing 125 mM KCl, 25 mM NaCl, 24 μ M Monensin, and varying concentrations of MES to adjust the pH to 3.5, 4.5, 5, 5.5, 6.5, or 7 (three independent experiments). The fluorescence ratio was linear for pH 4.5-7.0.
Melanin measurements. Melanin from wild type (control), tpc2-targeted CRISPR-Cas9 or siRNA-expressing Oa1 −/− melanocytes was quantified as previously described 32 . The soluble and insoluble fractions of melanocytes were separated after cell lysis with 1% Triton X-100 (Sigma) in PBS pH 7.4. Total protein was measured in the soluble fraction using a BCA Protein Assay Kit (Pierce). The insoluble fraction was dissolved in 1 N NaOH by incubation for 2 h at 85 °C and used to quantify melanin by measuring the optical density of each sample at 405 nm, then fit with a standard curve generated using synthetic melanin (Sigma). Average cellular melanin values were quantified as the ratio between total melanin and total protein (μ g melanin/mg protein) from the same dish. Three independent experiments were performed, each in triplicate.
To increase melanin content in B16-F1 melanocytes, mCherry-tagged wild-type or V443I-OCA2 variants were expressed alone or coexpressed with TPC2-GFP for 24 h prior to fixation and imaging and 48 h prior to melanin analysis. Images from three independent experiments were acquired with a laser-scanning confocal microscope (Zeiss LSM 510) using a 63X objective. The melanin from each B16-F1 condition was quantified as described above. Data analysis. All data are shown as mean ± s.e.m. Data were considered significant if p < 0.05 using paired or unpaired two-tailed Student t-tests or one-way ANOVA.
Ethical Approval. Bullfrog eyes used for RPE cells were obtained as discarded tissue from Dr. Thomas Roberts' laboratory at Brown University under protocol #1303990009, approved by the Institutional Animal Care and Use Committee and used in agreement with the approved guidelines and all the ethics rules and regulations.