TPC2 controls pigmentation by regulating melanosomepH and sizeAndrea L. Ambrosioa, Judith A. Boylea, Al E. Aradia, Keith A. Christiana, and Santiago M. Di Pietroa,1

aDepartment of Biochemistry and Molecular Biology, Colorado State University, Fort Collins, CO 80523-1870

Edited by Graça Raposo, Institut Curie-CNRS, Paris, France, and accepted by the Editorial Board April 5, 2016 (received for review January 4, 2016)

Melanin is responsible for pigmentation of skin and hair and issynthesized in a specialized organelle, the melanosome, in melano-cytes. A genome-wide association study revealed that the two poresegment channel 2 (TPCN2) gene is strongly linked to pigmentationvariations. TPCN2 encodes the two-pore channel 2 (TPC2) protein, acation channel. Nevertheless, how TPC2 regulates pigmentation re-mains unknown. Here, we show that TPC2 is expressed in melano-cytes and localizes to the melanosome-limiting membrane and, to alesser extent, to endolysosomal compartments by confocal fluores-cence and immunogold electron microscopy. Immunomagnetic iso-lation of TPC2-containing organelles confirmed its coresidence withmelanosomal markers. TPCN2 knockout by means of clustered reg-ularly interspaced short palindromic repeat/CRISPR-associated 9gene editing elicited a dramatic increase in pigment content inMNT-1 melanocytic cells. This effect was rescued by transient ex-pression of TPC2-GFP. Consistently, siRNA-mediated knockdown ofTPC2 also caused a substantial increase in melanin content in bothMNT-1 cells and primary human melanocytes. Using a newly devel-oped genetically encoded pH sensor targeted to melanosomes, wedetermined that the melanosome lumen in TPC2-KO MNT-1 cellsand primary melanocytes subjected to TPC2 knockdown is less acidicthan in control cells. Fluorescence and electron microscopy analysisrevealed that TPC2-KOMNT-1 cells have significantly larger melano-somes than control cells, but the number of organelles is un-changed. TPC2 likely regulates melanosomes pH and size bymediating Ca2+ release from the organelle, which is decreased inTPC2-KO MNT-1 cells, as determined with the Ca2+ sensor tyrosi-nase-GCaMP6. Thus, our data show that TPC2 regulates pigmen-tation through two fundamental determinants of melanosomefunction: pH and size.

melanosome | pigmentation | two-pore channel 2 | organelle pH |membrane traffic

In humans, melanin is responsible for pigmentation of skin,hair, and eyes and serves to minimize the damage caused by

exposure to UV radiation from sunlight (1, 2). Melanin is synthe-sized in a specialized organelle, the melanosome, which is pro-duced in melanocyte cells in skin and hair follicles and in the eyeretinal and iris pigmented epithelial cells (2–4). The melanosome isa lysosome-related organelle that houses the melanin-synthesizingenzymes: tyrosinase, tyrosinase-related protein-1, and tyrosinase-related protein-2 (2, 3, 5). The color of human skin and hair isdetermined by the amount and chemical composition of the mel-anin produced by melanosomes (2, 6, 7). Importantly, the activityof tyrosinase, the rate-limiting enzyme in melanin synthesis, isgreatly reduced at acidic pH. Melanosomal pH, rather than ty-rosinase expression level, has been shown to regulate tyrosinaseactivity and the amount of melanin produced in melanocytes fromdifferent skin types (2, 7–9). Melanosomes from melanocytes offair-skinned individuals are significantly more acidic and displaylow tyrosinase activity, whereas melanosomes in dark skin mela-nocytes are less acidic or neutral and present higher levels oftyrosinase activity (8, 10). The organelle dimensions are also dif-ferent: Melanosomes from highly pigmented skin are larger thanthose from lightly pigmented skin (11). Thus, revealing how me-lanosome pH and size are controlled is essential to understanding

pigmentation in humans. However, our knowledge of the un-derlying regulatory factors and mechanisms is incomplete (2).Less than 20 of the ∼300 pigmentation genes are known to

directly function in the production of melanin or regulation of itschemical composition (2, 7, 12). Mutation of several of thesegenes, including the three melanogenic enzymes and a few iontransporters, causes oculocutaneous albinism (OCA) in patientsand animal models (2, 5, 12). However, many genes that affectthe color of skin, hair, and eyes encode proteins with unknownfunction in pigmentation (2, 12). TPCN2 is one such gene (13).Two single nucleotide polymorphisms in the TPCN2 gene werestrongly associated with pigmentation variations in a genome-wide association study among 8,460 Icelanders and Dutch indi-viduals (Fig. S1) (13). The TPCN2 gene encodes the two-porechannel 2 (TPC2) protein (13). Indirect clues as to the possibleTPC2 protein function in pigmentation come from studies inother systems. TPC2 is a cation release channel expressed inendosomes and lysosomes in nonspecialized cells and platelet-dense granules, another specialized lysosome-related organelle(14–18). Nevertheless, it has not been experimentally demon-strated that TPC2 regulates pigmentation. It is also unclear howTPC2 would affect pigmentation.Here, we show that TPC2 is expressed in the melanosome-

limiting membrane in human melanocytic MNT-1 cells, a well-characterized system to study melanosome biology (19, 20).Knockout of TPCN2 using the clustered regularly interspacedshort palindromic repeat (CRISPR)/CRISPR-associated 9 (Cas9)gene editing system produced a striking increase in the melanincontent in MNT-1 cells. siRNA-mediated knockdown of TPC2 alsoelicited a significant increase in melanin content in both MNT-1 cellsand primary human melanocytes. Conversely, overexpression ofTPC2 reduced the amount of melanin, suggesting TPC2 is indeedinvolved in the regulation of melanin production by melanosomes.

Significance

Melanin pigments are synthesized in skin and hair cells calledmelanocytes and provide color to skin and hair and protectionagainst UV rays. Inadequate protection poses the risk of accu-mulating genetic mutations in the DNA of skin cells, which canlead to skin cancer. It can also reduce folate levels, which thencauses birth defects. Therefore, understanding pigmentation isimportant for human health. There are several protein com-ponents of the machinery that regulates human pigmentationthat work in unknown ways. Two-pore channel 2 (TPC2) is oneof them. Here we found that TPC2 is located in compartmentsinside melanocytes known as melanosomes, where melanin issynthesized. TPC2 regulates the pH and size of melanosomes,thus controlling the amount of melanin produced.

Author contributions: A.L.A. and S.M.D. designed research; A.L.A., J.A.B., A.E.A., K.A.C.,and S.M.D. performed research; A.L.A. and S.M.D. contributed new reagents/analytictools; A.L.A. and S.M.D. analyzed data; and A.L.A. and S.M.D. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission. G.R. is a guest editor invited by the EditorialBoard.1To whom correspondence should be addressed. Email: santiago.dipietro@colostate.edu.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1600108113/-/DCSupplemental.

5622–5627 | PNAS | May 17, 2016 | vol. 113 | no. 20 www.pnas.org/cgi/doi/10.1073/pnas.1600108113

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 12

, 202

1

We demonstrate that TPC2 regulates melanosomal pH and size inliving cells, providing a direct explanation to the pigmentationphenotype. We also show TPC2 is responsible for Ca2+ release frommelanosomes. Our data show TPC2 works in human pigmen-tation by regulating melanosome function.

ResultsTPC2 Is Present in the Melanosome-Limiting Membrane. To in-vestigate TPC2, we first tagged the endogenous TPCN2 genewith Emerald-GFP (EmGFP) in melanocytic MNT-1 cells, anideal system to study melanosome biology (19–23). To this end, weused the CRISPR/Cas9 gene editing system (24), as depicted in Fig.1A (Right). Briefly, the last exon of the TPCN2 gene was replaced byhomology-directed repair with a version containing the EmGFPcoding sequence right in front of the stop codon. Genotyping withEmGFP-specific primers showed incorporation of the DNA codingfor the fluorescent tag, and analysis with primers annealing outside ofthe substituted region demonstrated correct localization (Fig. S1).Cells expressing TPC2-EmGFP from the endogenous gene werethen transiently transfected with plasmids expressing TPC2-Cherryor the melanosomal markers tyrosinase-iRFP and Cherry-Rab27a(Fig. 1B). Live cell confocal fluorescence microscopy analysis showedendogenous TPC2-EmGFP colocalized with exogenously expressedTPC2-Cherry [Manders’ Overlap Coefficient (MOC) = 0.56 ± 0.05;

nine cells] and, to a lower degree, with tyrosinase-iRFP and Cherry-Rab27a [MOC = 0.39 ± 0.05 (23 cells) and 0.37 ± 0.07 (32 cells),respectively]. Control experiments showed that the empirical maxi-mum MOC detectable for a given protein simultaneously labeledwith two different fluorophores in live cells is 0.71 ± 0.06. Also, byanalyzing the colocalization between TPC2-EmGFP and an exoge-nously expressed peroxisomal marker, mRFP-SKL, in live cells, thebackground MOC was determined to be 0.12 ± 0.03 (Fig. S2). Infixed cells, endogenous TPC2-EmGFP showed a very high level ofcolocalization with the immunostained endogenous melanosomalmarker tyrosinase-related protein 1 (TYRP1; MOC = 0.83 ± 0.01;Fig. 1C). Therefore, TPC2 is expressed in melanocytes and localizesto melanosomes.Thin-section immunogold electron microscopy analysis of

MNT-1 cells labeled with an anti-TPC2 antibody showed the ma-jority of endogenous TPC2 localizes to pigmented, stage III and IVmelanosomes (Fig. 1D) (3, 20, 21). A smaller proportion of the labelwas found associated with electron-lucent compartments likely rep-resenting endosomes (3, 20, 21) (Fig. 1D). Quantification showed68 ± 5% of the label was associated with pigmented melanosomesand 22 ± 4% with endosomes/lysosomes (185 gold particles, eightcells). Labeling of other organelles was very low (nucleus: 5 ± 2%;mitochondria: 3 ± 2%). Similar results were obtained with MNT-1cells transfected with a TPC2-GFP plasmid and labeled with an anti-GFP antibody (Fig. S3; pigmented melanosomes: 57 ± 9% of label;

Fig. 1. TPC2 localizes to melanosomes. (A) Schematicof the CRISPR/Cas 9 design. (B) Confocal fluorescencemicroscopy images of live MNT-1 cells coexpressingendogenous TPC2-EmGFP and exogenous TPC2-Cherry,tyrosinase-iRFP, or Cherry-Rab27a. (Scale bars, 5 μm.)(C) Confocal fluorescence microscopy images of fixed/permeabilized MNT-1 cells expressing endogenousTPC2-EmGFP and immunostained for endogenousTYRP1. (Scale bar, 5 μm.) (D, Upper) Immunogold elec-tron micrograph of an MNT-1 cell labeled with an anti-TPC2 antibody (15,000×). (Scale bar, 500 nm.) (Lower)Higher-magnification view from the region indicated inthe upper panel showing examples of pigmented me-lanosomes (M) and endosomes (E) labeled with theTPC2 antibody. (Scale bar, 250 nm.) (E) Postnuclear su-pernatant extracts of MNT-1 cells expressing TPC2-GFPwere subjected to immunomagnetic isolation withanti-GFP and irrelevant antibodies. Immunoblottingrevealed the presence of tyrosinase, PMEL17, andRab32 in the anti-GFP sample. Blots are representa-tive of three experiments.

Ambrosio et al. PNAS | May 17, 2016 | vol. 113 | no. 20 | 5623

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 12

, 202

1

endosomes/lysosomes: 36 ± 6%; nucleus: 2 ± 2%; mitochondria: 2 ±1%; 982 gold particles, nine cells).To complement the microscopy studies, a biochemical ap-

proach was used to test for TPC2 localization to melanosomes. Apostnuclear supernatant extract of MNT-1 cells transfected with aTPC2-GFP plasmid was submitted to immunomagnetic isolation oforganelles, using an anti-GFP antibody or an irrelevant antibody asa control. Immunoblotting analysis revealed the presence of themelanosome resident proteins tyrosinase, PMEL17, and Rab32 inthe anti-GFP immunoisolated samples, thus confirming the pres-ence of TPC2 in melanosomes (Fig. 1E).

TPC2 Expression Level Determines the Overall Melanin Content.Onceit was established that TPC2 is expressed in melanosomes, weinvestigated the effect of knocking out TPCN2 in MNT-1 cells. Thefirst exon of the TPCN2 gene was excised, using the CRISPR/Cas9system (Fig. 1A, Left), and the deletion was confirmed by genotypingof two independent clones: TPC2-KO1 and TPC2-KO2 (Fig. S1).Real-time PCR of the TPC2-KO clones confirmed negligible TPC2mRNA expression (Fig. 2A). When the TPC2-KO1 and TPC2-KO2cells were pelleted, they appeared strikingly darker than wild-typeMNT-1 cells. Quantification of their melanin content revealed afourfold increase in total melanin compared with wild-type cells (Fig.2B). Importantly, in rescue experiments carried out by transfecting aTPC2-GFP plasmid, the melanin content of the TPC2-KO cells wasrestored to levels similar to wild-type cells, indicating specificity (Fig.2C). As an independent approach, we subjected MNT-1 cells andprimary human melanocytes to TPC2 siRNA and control siRNAknockdown and determined the total melanin content. Following thesame trend as the TPC2-KO cells, the TPC2 siRNA treated MNT-1cells and primary melanocytes more than doubled the melanincontent of the corresponding control siRNA-treated cells (Fig. S4).Conversely, overexpression of TPC2-GFP in control MNT-1 cellshad the opposite effect, significantly reducing the melanin content(Fig. S4). Overall, these results demonstrate an inverse correlationbetween TPC2 expression level and melanin content in MNT-1 cellsand primary human melanocytes.

TPC2 Regulates Melanosome Luminal pH. Tyrosinase is the key en-zyme in melanin synthesis, and its catalytic activity is low at acidicpH and high at neutral pH (2, 7, 8, 10). Therefore, an increase inmelanosome luminal pH in the TPC2-KO cells may be re-sponsible for the increase in melanin content in these cells. Accord-ingly, we investigated the effect of both TPC2 expression level andactivity on melanosome luminal pH. To this end, we engineered agenetically encoded melanosome-localized pH sensor (MELOPS)capable of detecting pH changes in the melanosome lumen of livingcells. MELOPS was created by inserting the coding sequence of thepH sensor protein Nectarine (25) into the first luminal loop of themelanosome protein OCA2, the product of the gene mutated inoculocutaneous albinism type 2 (Fig. S5). This OCA2 loop was pre-viously used to add other tags without affecting the protein traffickingand function (26). Providing OCA2 functions in chloride transport,we used the nonconductive V443I pore mutant to avoid any changeMELOPS expression might cause on melanosomal pH (9). Thefeasibility of using MELOPS to measure melanosomal pH wastested by expressing it in MNT-1 cells and incubating the cells inbuffers with pH values ranging from 4.5 to 6.5 in the presence ofionophores (27). Fluorescence microscopy imaging analysis revealedMELOPS colocalizes with the melanosomal marker tyrosinase-iRFP (MOC = 0.52 ± 0.02), and its fluorescence intensity is pro-portional to the pH of the buffers from pH 5.0 to 6.5 (Fig. S5).Using this calibration curve as a reference, the melanosome luminalpH in control MNT-1 cells was calculated as 5.7 ± 0.6 (44 cells;Fig. S5).Then, we expressed MELOPS in wild-type and TPC2-KO

MNT-1 cells and determined higher fluorescence intensity, andtherefore a higher melanosomal pH in TPC2-KO cells (Fig. 3 Aand B). This effect can be rescued by transient transfection ofTPC2-iRFP in the TPC2-KO cells, corroborating the lack ofTPC2 caused the pH increase in the melanosome lumen (Fig. 3

Fig. 2. TPC2 expression level determines the melanin content in MNT-1cells. (A) RNA was isolated from wild-type MNT-1 cells or two independentclones of CRISPR/Cas9 TPC2-KO homozygous cells and reverse-transcribedinto cDNA. Relative amounts of the TPC2 mRNA were determined in tripli-cate by real-time PCR. (B) Melanin content was determined for the cells in A.(C) Melanin content was determined for mock-transfected wild-type MNT-1cells, CRISPR/Cas9 TPC2-KO cells, and TPC2-GFP transfected TPC2-KO cells 5 dposttransfection (n = 2 for WT+Mock, n = 3 for TPC2-KO+Mock, and n = 9for TPC2-KO+TPC2-GFP cells).

5624 | www.pnas.org/cgi/doi/10.1073/pnas.1600108113 Ambrosio et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 12

, 202

1

A and B). Consistently, primary human melanocytes subjected toTPC2 siRNA knockdown also showed higher melanosomal pHthan control siRNA-treated cells (Fig. S6). Conversely, over-expression of TPC2-GFP in primary human melanocytes causeda lower melanosomal pH, relative to control cells (Fig. S6).TPC2 overexpression in wild-type MNT-1 cells also causes adecrease in melanosomal pH that can be blocked by the TPC2-specific inhibitors Ned19 and tetrandrine (Fig. S7) (28).Furthermore, incubation of wild-type MNT-1 cells with theTPC2-specific inhibitor Ned19 produced a significant increase intheir melanin content relative to MNT-1 cells treated with ve-hicle (Fig. S7). The data all together demonstrate a tight inversecorrelation between TPC2 expression level and melanosomeluminal pH. The fact that Ned19 and tetrandrine inhibit theeffect of TPC2 on melanosomal pH indicates ion flux throughTPC2 is required to exert this function. These results explain, atleast in part, the increase in melanin content of TPC2-deficientMNT-1 cells and primary melanocytes, as well as the decrease inmelanin content of MNT-1 cells overexpressing TPC2.

TPC2 Regulates Melanosome Size.During the course of our studies,we noticed that melanosomes imaged with various markersappeared to be larger in TPC2-KO MNT-1 cells than in wild-typeMNT-1 cells. To better investigate this possibility, we used Rab27a, awell-established marker of pigmented melanosomes (29, 30) that inour hands produces the most crisp confocal fluorescence microscopyimages. Wild-type and TPC2-KOMNT-1 cells were transfected witha plasmid expressing Cherry-Rab27a and subjected to live cell con-focal fluorescence microscopy analysis (Fig. 4A). Quantification of

the images using the Slidebook software automated object analysisfunction showed no increase in the number of melanosomes per cellin TPC2-KO MNT-1 cells compared with wild-type cells (264 ± 89and 285 ± 115, respectively; n = 41 and 37 cells). However, theanalysis revealed melanosomes in TPC2-KO MNT-1 cells were sig-nificantly larger than those in wild-type cells (Fig. 4 A and B).To confirm this result, we analyzed wild-type and CRISPR TPC2-

KO MNT-1 cells by thin-section electron microscopy (Fig. 4 C andD and Fig. S8). Once again, we found that the total amount ofmelanosomes per cell did not differ between wild-type and TPC2-KO cells (66 ± 19 and 63 ± 19 melanosome stages II–IV per cell orthin section, respectively), but TPC2-KO cells showed strikinglylarger melanosomes than wild-type MNT-1 cells (Fig. 4 C and Dand Fig. S8). TPC2-KO MNT-1 cells also revealed more heavilypigmented melanosomes than wild-type cells (Fig. 4C and Fig. S8).These results indicate that TPC2 regulates melanosome size and areconsistent with higher pigment content in TPC2-deficient MNT-1cells compared with control cells.

TPC2 Regulates Ca2+ Release from Melanosomes. TPC2 is a Ca2+channel, and therefore we reasoned that Ca2+ release frommelanosomes into the cytosol could be affected in TPC2-KOcells. To detect Ca2+ released from melanosomes, we attached thegenetically encoded Ca2+ sensor GCaMP6 (31) to the C-terminalcytosolic tail of tyrosinase, which is expressed on the cytosolic sideof the melanosome membrane. A similar approach was previouslyused to measure TPC2-mediated Ca2+ release from other organ-elles such as platelet dense granules and lysosomes (17, 32). In-deed, TPC2-KO cells exhibited a reduced level of GCaMP6fluorescence intensity, indicative of lower local Ca2+ concentration,compared with wild-type cells (Fig. 5 A and B). Similar experimentswere carried out with tyrosinase-GCaMP6-Cherry, a construct inwhich Cherry was added at the C terminus of tyrosinase-GCaMP6to obtain a ratiometric measurement of fluorescence intensity. TheTPC2-KO cells exhibited lower GCaMP6/Cherry fluorescence in-tensity ratio compared with wild-type cells, which is indicative oflower levels of Ca2+ around melanosomes (Fig. S9). Together,these data suggests normal Ca2+ release from melanosomes de-pends on the presence of TPC2 in these organelles.

DiscussionThe complexity of human pigmentation results from hundreds ofproteins functioning in melanocyte development, melanosome bio-genesis, melanin synthesis, and melanosome transfer to neighboringkeratinocytes in skin and hair (1, 2, 7, 12). Even though two TPC2polymorphisms have been associated with pigmentation variations ina genome-wide association study (13), it is unknown where and howTPC2 works to bring about such diversity. Here, we first show thatTPC2 is expressed in melanocytes, and a significant cohort localizesto the melanosome-limiting membrane with a smaller fraction lo-calizing to endosomes and lysosomes, as found in other cell types.TPC2 expressed from the endogenous gene in MNT-1 cells coloc-alizes with melanosome markers such as tyrosinase, Rab27a, andTYRP1 by confocal fluorescence microscopy. Similarly, immuno-gold electron microscopy analysis of MNT-1 cells showed that bothendogenous and exogenous TPC2 localize to pigmented melano-somes. Biochemical experiments confirmed the presence of TPC2 inmelanosomes. Second, the melanin content was substantially in-creased in CRISPR TPC2 knockout and siRNA TPC2-depletedMNT-1 cells, as well as siRNA TPC2-depleted primary humanmelanocytes. Consistently, TPC2 inhibition by Ned19 caused anincrease in melanin content in MNT-1 cells. Conversely, higherTPC2 expression resulted in lower melanin content in MNT-1 cells.These results demonstrate an inverse correlation between the levelof TPC2 expression and melanin content in MNT-1 cells and pri-mary melanocytes. The data show experimentally that TPC2 indeedregulates pigmentation, thus confirming the genome-wide associa-tion study (13), and that TPC2 works in melanocytes by controllingthe amount of melanin produced by melanosomes.How does TPC2 regulate the amount of melanin in melano-

cytes? Melanosome pH is a key determinant of melanin synthesis

Fig. 3. TPC2 regulates melanosome luminal pH. (A) Wild-type or CRISPRTPC2-KO MNT-1 cells expressing MELOPS and tyrosinase-iRFP were analyzedby confocal fluorescence microscopy. In a rescue experiment performed inparallel, TPC2-KO MNT-1 cells expressing MELOPS and TPC2-iRFP were alsoanalyzed. (Scale bar, 5 μm.) (B) The average MELOPS fluorescence intensitywas determined for all treatments (WT+Tyr = 30 cells; TPC2-KO+Tyr = 25cells; TPC2-KO+TPC2 = 21 cells).

Ambrosio et al. PNAS | May 17, 2016 | vol. 113 | no. 20 | 5625

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 12

, 202

1

in melanosomes because of its effect on the catalytic activity ofthe rate-limiting enzyme tyrosinase (2, 7). In addition, sponta-neous late reactions of eumelanin polymerization that take placeafter the Dopachrome stage are also pH-dependent (33). In boththe enzymatic and nonenzymatic reactions, a lower pH decreasesthe amount of melanin produced. Consistently, indirect deter-minations with fixed cells showed melanosomes from light skinmelanocytes are more acidic than those from dark skin mela-nocytes (8, 10). To study melanosome pH, we developedMELOPS, the first genetically encoded melanosome localizedpH sensor that allows live cell determinations. MELOPS shouldbe a useful tool in melanosome research to dissect the effect ofpH by a variety of pigmentation genes. Using MELOPS, weestablished an increase in the melanosome lumen pH in liveTPC2-KO MNT-1 cells compared with wild-type MNT-1 cells.Likewise, primary human melanocytes subjected to TPC2 siRNAknockdown displayed higher melanosome lumen pH, whereasoverexpression of TPC2 elicited lower melanosome lumen pHrelative to control melanocytes. Furthermore, overexpression ofTPC2 in MNT-1 cells produced a decrease in melanosomal pH

that could be blocked using the TPC2-specific inhibitors Ned19or tetrandrine (28). These results both supported the concept ofan anticorrelation between melanosomal pH and TPC2 expres-sion levels and indicated that TPC2 activity regulates melano-somal pH. Therefore, TPC2 controls melanin synthesis inmelanocytes, at least in part, through regulation of melanosomalpH. It is worth noting that TPC2 has also been reported toregulate the pH of lysosomes and platelet-dense granules, andthe gene is present in evolutionarily distant eukaryotes such asplants (14, 17). Thus, melanosomes seem to have incorporatedand adapted an ancient component of the pH regulatory ma-chinery of membrane-bound compartments.A key factor regulating organelle pH is the vacuolar (V)-

ATPase, which is also present in melanosomes (2, 33). Severalother ion channels and transporters involved in human pig-mentation are present in the melanosome-limiting membrane.The recently described OCA2-mediated Cl− channel regulatesmelanin synthesis in part through melanosome pH (7, 9). Themelanosome-localized K+-dependent Na+/Ca2+ exchanger, SLC24A5,regulates pigmentation, potentially through pH changes (2, 34).

Fig. 4. TPC2 regulates melanosome size. (A) Wild-type or CRISPR TPC2-KO MNT-1 cells expressing Cherry-Rab27a were analyzed by confocal fluorescencemicroscopy. Insets are magnifications of the boxed regions. (Scale bar, 5 μm.) (B) The average melanosome size was measured for wild-type and TPC2-KO cells(n = 37 and 41 cells, respectively). (C) Electron micrographs of wild-type (Left) or CRISPR TPC2-KO (Right) MNT-1 cells (18,500×). Roman numerals indicatemelanosome maturation stage. (Scale bar, 500 nm.) (D) The average stage II–IV melanosome size was measured for wild-type and TPC2-KO cells (wild-type:seven cells, 464 melanosomes; TPC2-KO: nine cells, 567 melanosomes).

Fig. 5. TPC2 regulates Ca2+ release from melano-somes. (A) Wild-type or CRISPR TPC2-KO MNT-1 cellsexpressing tyrosinase-GCaMP6 were analyzed byconfocal fluorescence microscopy. (Scale bar, 5 μm.)(B) The tyrosinase-GCaMP6 fluorescence intensitywas determined for wild-type and TPC2-KO cells (n =37 and 41 cells, respectively).

5626 | www.pnas.org/cgi/doi/10.1073/pnas.1600108113 Ambrosio et al.

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 12

, 202

1

The SLC45A2 gene (member of the solute carrier 45A family)is responsible for the OCA4 disease and affects pigmentationthrough melanosome pH, although the identity of the moleculeor molecules transported by this carrier is unknown (35). TPC2thus joins a growing list of melanosome-localized ion channelsand transporters that regulate pigmentation at least in partthrough regulation of organelle pH. Therefore, the melanosomeluminal pH appears to be determined by a complex cross talkbetween channels, transporters, and pumps that control not onlythe amount but also the type of melanin produced by the me-lanosome. Additional work is needed to understand how TPC2may regulate melanosome pH.The second determinant of pigmentation we found to be

regulated by TPC2 is the melanosome size. Melanosomes inMNT-1 cells lacking TPC2 are remarkably larger than in wild-type cells. However, the melanosome count did not differ be-tween wild-type and TPC2-KO cells. This result implies anincrease in melanin content per melanosome in TPC2-KO cells,which was evident in the electron micrographs. The data are alsoconsistent with the fact that melanosomes in dark skin are largerthan those from light skin people (11). Interestingly, TPC2 hasbeen implicated in membrane trafficking events in other celltypes. TPC2 was recently shown to function in endolysosomaltrafficking in the context of Ebola virus infection (28) and inregulation of membrane fusion/fission events between platelet-dense granules (17). A function of TPC2 in regulation of mem-brane trafficking would also be in line with its reported physicalinteraction with the endolysosome tethering and fusion machinery

(18, 36). Such membrane dynamic events are regulated by Ca2+.Our results showing TPC2-KO MNT-1 cells have a reduced[Ca2+] on the cytosolic side of the melanosome membrane sug-gest a decreased Ca2+ release and are consistent with alteredmembrane traffic.In summary, our results show experimentally that TPC2 is a

true regulator of pigmentation. Importantly, TPC2 is expressedin the melanosome membrane, and its activity controls themelanin content of melanosomes, most likely by regulating or-ganelle pH and size.

Materials and MethodsCRISPR Design. Guide RNAs were designed using the CRISPR Design Tool[crispr.mit.edu (37)] to minimize potential off-target effects. Annealing,phosphorylation, and cloning of the guide RNAs (Table S1) into the pX330vector were carried out as described (24). For the TPC2-EmGFP donor plas-mid, the left and right homology arms (855 and 904 bp, respectively) wereamplified from genomic DNA, and mutations were introduced in the PAMsequences to avoid cleavage of the modified alleles. Genotyping was per-formed using the primers listed in Table S2.

Other Procedures. Additional procedures are described in SI Materials andMethods and Table S3.

ACKNOWLEDGMENTS. We thank T. Giddings and C. Ozzello for help withhigh-pressure freezing and electron microscopy, and M. Marks, G. Payne,and P. Kim for their generous gifts of reagents. This work was supported byNIH grant R01HL106186 (to S.M.D.).

1. Sturm RA (2009) Molecular genetics of human pigmentation diversity. Hum MolGenet 18(R1):R9–R17.

2. Kondo T, Hearing VJ (2011) Update on the regulation of mammalian melanocytefunction and skin pigmentation. Expert Rev Dermatol 6(1):97–108.

3. Raposo G, Marks MS (2007) Melanosomes–dark organelles enlighten endosomalmembrane transport. Nat Rev Mol Cell Biol 8(10):786–797.

4. Orlow SJ (1995) Melanosomes are specialized members of the lysosomal lineage oforganelles. J Invest Dermatol 105(1):3–7.

5. Huizing M, Helip-Wooley A, Westbroek W, Gunay-Aygun M, Gahl WA (2008) Disor-ders of lysosome-related organelle biogenesis: Clinical and molecular genetics. AnnuRev Genomics Hum Genet 9:359–386.

6. Yamaguchi Y, Brenner M, Hearing VJ (2007) The regulation of skin pigmentation.J Biol Chem 282(38):27557–27561.

7. Ito S, Wakamatsu K (2011) Human hair melanins: What we have learned and have notlearned from mouse coat color pigmentation. Pigment Cell Melanoma Res 24(1):63–74.

8. Fuller BB, Spaulding DT, Smith DR (2001) Regulation of the catalytic activity of pre-existing tyrosinase in black and Caucasian human melanocyte cell cultures. Exp CellRes 262(2):197–208.

9. Bellono NW, Escobar IE, Lefkovith AJ, Marks MS, Oancea E (2014) An intracellularanion channel critical for pigmentation. eLife 3:e04543.

10. Smith DR, Spaulding DT, Glenn HM, Fuller BB (2004) The relationship between Na(+)/H(+)exchanger expression and tyrosinase activity in human melanocytes. Exp Cell Res298(2):521–534.

11. Sturm RA, Box NF, Ramsay M (1998) Human pigmentation genetics: The difference isonly skin deep. BioEssays 20(9):712–721.

12. Lamoreux ML, Delmas V, Larue L, Bennett DC (2010) The Colors of Mice: A ModelGenetic Network (Wiley-Blackwell, Hoboken), 1st Ed.

13. Sulem P, et al. (2008) Two newly identified genetic determinants of pigmentation inEuropeans. Nat Genet 40(7):835–837.

14. Morgan AJ, Galione A (2014) Two-pore channels (TPCs): Current controversies.BioEssays 36(2):173–183.

15. Zhu MX, et al. (2010) Calcium signaling via two-pore channels: Local or global, that isthe question. Am J Physiol Cell Physiol 298(3):C430–C441.

16. Calcraft PJ, et al. (2009) NAADP mobilizes calcium from acidic organelles throughtwo-pore channels. Nature 459(7246):596–600.

17. Ambrosio AL, Boyle JA, Di Pietro SM (2015) TPC2 mediates new mechanisms ofplatelet dense granule membrane dynamics through regulation of Ca2+ release. MolBiol Cell 26(18):3263–3274.

18. Lin-Moshier Y, et al. (2014) The Two-pore channel (TPC) interactome unmasks iso-form-specific roles for TPCs in endolysosomal morphology and cell pigmentation. ProcNatl Acad Sci USA 111(36):13087–13092.

19. Raposo G, Tenza D, Murphy DM, Berson JF, Marks MS (2001) Distinct protein sortingand localization to premelanosomes, melanosomes, and lysosomes in pigmentedmelanocytic cells. J Cell Biol 152(4):809–824.

20. Hurbain I, et al. (2008) Electron tomography of early melanosomes: Implications formelanogenesis and the generation of fibrillar amyloid sheets. Proc Natl Acad Sci USA105(50):19726–19731.

21. Theos AC, et al. (2005) Functions of adaptor protein (AP)-3 and AP-1 in tyrosinasesorting from endosomes to melanosomes. Mol Biol Cell 16(11):5356–5372.

22. Bultema JJ, Ambrosio AL, Burek CL, Di Pietro SM (2012) BLOC-2, AP-3, and AP-1proteins function in concert with Rab38 and Rab32 proteins to mediate proteintrafficking to lysosome-related organelles. J Biol Chem 287(23):19550–19563.

23. Di Pietro SM, et al. (2006) BLOC-1 interacts with BLOC-2 and the AP-3 complex tofacilitate protein trafficking on endosomes. Mol Biol Cell 17(9):4027–4038.

24. Ran FA, et al. (2013) Genome engineering using the CRISPR-Cas9 system. Nat Protoc8(11):2281–2308.

25. Johnson DE, et al. (2009) Red fluorescent protein pH biosensor to detect concentra-tive nucleoside transport. J Biol Chem 284(31):20499–20511.

26. Sitaram A, et al. (2009) Localization to mature melanosomes by virtue of cytoplasmicdileucine motifs is required for human OCA2 function. Mol Biol Cell 20(5):1464–1477.

27. Lu Y, et al. (2013) Two pore channel 2 (TPC2) inhibits autophagosomal-lysosomalfusion by alkalinizing lysosomal pH. J Biol Chem 288(33):24247–24263.

28. Sakurai Y, et al. (2015) Ebola virus. Two-pore channels control Ebola virus host cellentry and are drug targets for disease treatment. Science 347(6225):995–998.

29. Wu X, et al. (2001) Rab27a enables myosin Va-dependent melanosome capture byrecruiting the myosin to the organelle. J Cell Sci 114(Pt 6):1091–1100.

30. Hume AN, et al. (2001) Rab27a regulates the peripheral distribution of melanosomesin melanocytes. J Cell Biol 152(4):795–808.

31. Chen TW, et al. (2013) Ultrasensitive fluorescent proteins for imaging neuronal ac-tivity. Nature 499(7458):295–300.

32. Medina DL, et al. (2015) Lysosomal calcium signalling regulates autophagy throughcalcineurin and TFEB. Nat Cell Biol 17(3):288–299.

33. Ito S, Suzuki N, Takebayashi S, Commo S, Wakamatsu K (2013) Neutral pH and copperions promote eumelanogenesis after the dopachrome stage. Pigment Cell MelanomaRes 26(6):817–825.

34. Lamason RL, et al. (2005) SLC24A5, a putative cation exchanger, affects pigmentationin zebrafish and humans. Science 310(5755):1782–1786.

35. Bin BH, et al. (2015) Membrane-Associated Transporter Protein (MATP) Reg-ulates Melanosomal pH and Influences Tyrosinase Activity. PLoS One 10(6):e0129273.

36. Grimm C, et al. (2014) High susceptibility to fatty liver disease in two-pore channel2-deficient mice. Nat Commun 5:4699.

37. Hsu PD, et al. (2013) DNA targeting specificity of RNA-guided Cas9 nucleases. NatBiotechnol 31(9):827–832.

38. Bultema JJ, et al. (2014) Myosin vc interacts with Rab32 and Rab38 proteins andworks in the biogenesis and secretion of melanosomes. J Biol Chem 289(48):33513–33528.

39. Shcherbakova DM, Verkhusha VV (2013) Near-infrared fluorescent proteins for mul-ticolor in vivo imaging. Nat Methods 10(8):751–754.

40. Sitaram A, et al. (2012) Differential recognition of a dileucine-based sorting signal byAP-1 and AP-3 reveals a requirement for both BLOC-1 and AP-3 in delivery of OCA2 tomelanosomes. Mol Biol Cell 23(16):3178–3192.

41. Cong L, et al. (2013) Multiplex genome engineering using CRISPR/Cas systems. Science339(6121):819–823.

42. Ambrosio AL, Boyle JA, Di Pietro SM (2012) Mechanism of platelet dense granulebiogenesis: Study of cargo transport and function of Rab32 and Rab38 in amodel system. Blood 120(19):4072–4081.

Ambrosio et al. PNAS | May 17, 2016 | vol. 113 | no. 20 | 5627

CELL

BIOLO

GY

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 12

, 202

1