Sustained pigmentation causes DNA damage and invokes translesion polymerase Polκ for repair in melanocytes

Abstract Melanin protects skin cells from ultraviolet radiation-induced DNA damage. However, intermediates of eumelanin are highly reactive quinones that are potentially genotoxic. In this study, we systematically investigate the effect of sustained elevation of melanogenesis and map the consequent cellular repair response of melanocytes. Pigmentation increases γH2AX foci, DNA abasic sites, causes replication stress and invokes translesion polymerase Polκ in primary human melanocytes, as well as mouse melanoma cells. Confirming the causal link, CRISPR-based genetic ablation of tyrosinase results in depigmented cells with low Polκ levels. During pigmentation, Polκ activates replication stress response and keeps a check on uncontrolled proliferation of cells harboring melanin-damaged DNA. The mutational landscape observed in human melanoma could in part explain the error-prone bypass of DNA lesions by Polκ, whose absence would lead to genome instability. Thereby, translesion polymerase Polκ is a critical response of pigmenting melanocytes to combat melanin-induced DNA alterations. Our study illuminates the dark side of melanin and identifies (eu)melanogenesis as a key missing link between tanning response and mutagenesis, mediated via the necessary evil translesion polymerase, Polκ.


INTRODUCTION
Skin pigmentation acts as an important barrier against penetration of harmful ultraviolet (UV) radiations.Melanin polymer with its broad absorption spectrum protects the genome from damage by high energy UV rays ( 1 , 2 ).Hence, skin tanning is an essential physiological response for protection against UV-induced mutagenesis and consequent risk of malignancy.The adapti v e nature of pigmentation is evident in world-wide distribution of basal skin tone as well as the extent of tanning in populations residing across latitudes tha t dif fer in incident UV flux ( 3 ).The presence of melanin laden melanosomes in keratinocytes protects these cells from UV damage ( 4 ).Howe v er, synthesis of melanin within melanocytes poses challenge as the intermediates ar e highly r eacti v e and potentially genotoxic (5)(6)(7)(8)(9); they also have a potential to induce melanoma and facilitate its progression (9)(10)(11).Melanin intermediates are also indicated as immunosuppressants during cancer therapy ( 12 ).Variety of factors contribute to the mutation burden in melanocytes, among these the role of melanogenesis, if any, r emains obscur e.
The nature of DNA damage caused by UV-A, UV-B and UV-C, as well as the cellular response are very wellcharacterized ( 13 , 14 ).UV-signature mutations arise in keratinocytes consequent to unr epair ed DNA photoproducts, such as cyclobutane pyrimidine dimers (CPDs) and pyrimidine 6-4 pyrimidones .These are resolved via Nucleotide Excision Repair (NER) pathway.The autosomal recessi v e syndr ome xer oderma pigmentosum caused by mutations in various genes in this pathway and the translesion polymerase eta (POLH), render individuals sensitive to sunlight ( 15 , 16 ).Tumor suppressor TP53 is the key orchestrator of UV-mediated DNA damage in keratinocytes, that induces POMC gene expr ession, r esulting in ␣-MSH secretion ( 17 ).In neighboring melanocytes, ␣-MSH activates the downstream MC1R signaling.In addition to promoting melanogenic response, this pathway invokes NER which pre-empti v ely pr epar es melanocytes for DNA r epair ( 18 ).
Genetic risk factor of melanoma links to the red hair phenotype of individuals with lighter skin tone due to the presence of pheomelanin.Systematic studies have elucidated the role of oxidati v e damage in UV-independent carcinogenesis, induced by pheomelanin ( 19 ).Epidemiological studies suggest that the highest melanoma risk is associated with intermittent exposure to UV rays ( 20 ).UV-induced free radicals excite electrons in melanin to create a quantum triplet sta te tha t transfer energy to DNA and initiate dark CPD formation.The work shows the transfer of energy to DNA, inducing the formation of CPDs ( 21 ).These lesions are consider ed r esponsible for the characteristic UV-signatur e C > T transitions.Hence, UV directly and in conjunction with melanin is likely to cause DNA lesions and initiate mutations.Additionally, as pheomelanin is capable of mutagenesis independent of UV, it is generally belie v ed that formation of eumelanin is protecti v e and hence its role in causing DNA damage is not systematically investigated.In a recent wor k, inv estiga tors identified tha t melanocytes contain genomic locations that are prone to modifications in se v eral loci implicated in melanoma ( 22 ).We examined the role of melanin and its intermediates generated during sustained melanogenesis to modify DNA and elicit counteracting cellular response.
To investigate this, in the current study we employ a synchronized, temporally resolved pigmenting system using B16 melanoma cells that produce only eumelanin and confirm our results in primary human melanocytes.In a UVindependent manner, the rate of eumelanin production is further enhanced with substrate tyrosine, suppressed with pharmacological inhibition using Phenylthiourea (PTU) as well as CRISPR-based abrogation of melanin synthesis.Sustained eumelanin synthesis increases ␥ H2AX foci and abasic sites formation in B16 melanoma cells.In response, translesion DN A pol ymerase K (Pol ) is induced through ATR-CHK1 signaling.Pol is critical for the induction of cellular replication stress response and in the absence of Pol , sustained melanogenesis results in continued proliferation of melanocytes with DNA damage that could lead to genome instability.Extracellular synthesis of melanin as well as treatment with DHI (dihydroxyindole), one of the intermediates of eumelanogenesis, causes DNA breaks and induces Pol in B16 melanoma cells.Our study establishes that eumelanogenesis through melanin intermediates can produce lesions in DNA.These observations highlight the counteracting role of the translesion polymerase Pol in combating melanin-induced DN A damage.Meta-anal ysis of TCGA-melanoma data shows elevated mutation burden in human melanoma patient samples with high Pol mRNA le v els and a concomitant decrease in surviv al. Thereb y, a complex interplay between melanin-induced DNA damage, err or pr one nature of polymerase Pol and induction of r eplication str ess r esponse sha pes the DN A repair functions of melanocytes.

Cell lines, reagents and media
B16 cells were obtained from ATCC.Primary human melanocytes (Normal Human Epidermal Melanocytes, NHEM) were obtained from Lonza.All cell culture r eagents wer e obtained from GIBCO Life technologies.siR-NAs and shRNA were purchased from Dharmacon (GE Healthcare) and transfections were performed using Dharmafect II for siRNAs and Lipofectamine 2000 for shRNAs.Midiprep plasmid preparation was performed using Qiagen Midi DNA kit.KAPA SYBR FAST qPCR Master Mix was obtained from KAPA Biosystems.Genomic DNA and RNA isolations were performed using Macherey Nagel Nu-cleoSpin ® TriPrep kit (Cat no.: 740966).

Setting of low density based pr ogressiv e pigmentation model of B16 mouse melanoma cells
Cultur es wer e set up as described by ( 23 ).Briefly, depigmented B16 (called as day 0 cells) cells were seeded at a density of 100 cells / cm 2 in DMEM high glucose media (Sigma; D5648) with 10% FBS (Invitrogen, 10270) to make them pigmented by day 6 to day 8.These cells, starting from day 4 until day 8 are the pigmented cells.During the span of pigmentation, the media was not replenished.

Culturing and chemical treatments on melanocytes
B16 wer e cultur ed in DMEM high glucose and 10% FBS, and primary human melanocytes were grown in M254 (Gibco, M-254-500), both at 5% CO 2 concentration.Efforts were diligently made to minimize the influence of ambient light throughout the culture and treatment processes.Special care was taken to ensure minimal exposure during passage and media change, limiting exposure to white light of fewer than 900 lx.This le v el of light intensity is at least three logarithmic orders lower than the energy r equir ed for photosensitiza tion.All trea tments were gi v en at 24 h post seeding.200 M of Phenylthiourea (PTU, Sigma; P7629) was used for depigmentation, and 1 mM of L -tyrosine was used for hyperpigmentation of cells. 2 mM of hydroxyurea (Sigma; H8627) for 3 h was used for inducing replication stress, 5 J / m 2 of UVA and 200 mJ / cm 2 of UVB were used as photooxidati v e dama ging a gents.100 mM H 2 O 2 for 15 min was used as an oxidati v e stress inducer.AZ20 (Sigma; SML1328) was used as an ATR inhibitor at 50 nM concentration.PP242 (Medchem: HY-10474) was used as an mTOR inhibitor at 2 M.

Detection of DNA base modifications by ELISAs
Genomic DNA was isolated using the Macherey Nagel Nucleospin Triprep kit.Abasic sites and 8-o x oGuanine were determined using ELISA kits from cell biolabs (OxiSelect ™ Oxidati v e DNA Damage Quantitation Kit, AP sites, STA-324 and OxiSelect ™ Oxidati v e DNA Damage Quantitation Kit, 8-OHdG, STA-320) using manufacturer's instructions.

NBT assay
Plasmid DNA was incubated either alone (control DNA), with substrate L-DOPA (DNA + DOPA) or with melanin synthesis reaction containing L-DOPA and enzyme tyrosinase (DNA + Mel).These samples, along with the melanin synthesis reaction without DNA (Mel only), or DNA added to the reaction just before column purification [DNA+(Mel)] were column purified after 16 h of incubation, to remove free melanin intermediates, enzyme and melanin from DNA.Hence, only DNA (modified or unmodified as mentioned above) is taken ahead.These samples are then quantified, and equal concentrations are incubated with NBT for 15 min at room temperature.Samples are then transferred in cuvette and absorbance at 570 nm (corrected at 700 nm) is taken for detection of formazan (formed by quinone-DNA adducts).The absorbance is converted to concentration of formazan using the Beerlambert's law, A = ε bC ( A = absorbance, b = path length of cuvette, i.e. 1 cm, C = concentration of formazan in M and ε = molar absorption coefficient of formazan, i.e. 13 000 M −1 cm −1 at 570 nm) ( 24 ).

Detection of reactive oxygen species and apoptotic cells
Relati v e le v els of reacti v e oxygen species were determined using CM-H2DCFDA (2 -7 -dichlor odihydr ofluorescein diacetate) assay kit (Thermo, Cat no.C6827).Detection of apoptotic cells was done using The Alexa Fluor ® 488 annexin V / Dead Cell Apoptosis Kit (Invitrogen, V13241) as per manufacturer's protocol.

Cell cycle analysis of B16 cells
B16 cells were trypsinised and washed once with PBS.Cell pellet of a pproximatel y 3-5 million cells was resuspended in 200-300 l of 70% chilled ethanol and was stored overnight in minus 20 • C. For analysis, cells were spun at 7000 rpm for 10 min at 4 • C. Supernatant ethanol was discarded and washed thrice with 1 PBS to remove traces of ethanol.Finally, cells were resuspended in 100-200 l of (PBS + 0.1% Triton X100) along with RNase A (Sigma; R6513) to a working concentration of 0.5 mg / ml.These were incubated overnight at room temperature.After taking out some cells as unstained control, 2 l of propidium iodide (2 mg / ml) was added to a pproximatel y 100 l of cell suspension and incubated at room temperature for 10 min in the dark.Cells were then directly used for flow cytometric analysis.

Alkaline and neutral comet assay
Corning glass slides wer e pr e-coated with 0.1% LMA and dried at 55 • for 1-2 h on a hot plate.Meanwhile, the cells were trypsinised and 30 000-50 000 cells from each set were taken in a 1.5 ml eppendorf tube and pelleted down by centrifugation.A single wash of 1 × PBS was gi v en for complete removal of media from cell pellet.The cells were uniformly resuspended in 50 l of PBS and 550 l of preheated 0.75% agarose was added onto cells, mixed well and layered on the precoated slides after removing them from heatblock.The slides were now allowed to solidify by keeping them at 4 • C for 20 min.These were then immersed in chilled Alkaline Lysis buffer taken in a container while keeping the buffer on ice for another 20 min.Slides were gently taken out and rinsed thrice in chilled AMQ.While keeping the AMQ on ice, the slides were immersed from sides and taken out without giving any mechanical shock.Finally, the slides were dipped in Electrophoresis buffer (pH-13) and incubated for 40 min on ice.After this, 600 ml of electrophoresis buffer was added to the electrophoresis unit and the slides were immersed in it while aligning them straight in electric field.Electrophoresis was run at 24 V (2 V / cm), 400 mA for 20 min.For neutral comet assay, pr epar ed slides were incubated in chilled Neutral electrophoresis buffer (pH 9) for 1 h, and then electrophoresed in the same buf fer a t 24 V (2 V / cm), 400 mA for 20 min.After completion of run, slides were taken out and rinsed thrice with chilled AMQ and dipped in chilled neutralization buffer (pH 7.4) for 5 min while keeping the buffer on ice.Finally, the slides were taken out and rinsed again in AMQ and kept on hot plate at 55 • C for 3-4 h till complete drying.Slides were stained with 50 M propidium iodide and imaged on Leica microscope.Analysis was performed by KOMET software.
For OGG-1 modified comet assay after the cell lysis, slides were incubated with OGG-1 buffer alone (-OGG-1) or with 0.5 g / ml OGG-1 (+OGG-1) at 37 • C in a humidifying chamber for a period of 45 min.The slides were then processed for alkaline comet assay, as mentioned above.

RNA isolation, cDNA synthesis and real time PCR
RNA isolation was performed using Macherey Nagel triprep kit using manufacturer's instructions.500 ng of RNA was re v erse transcribed using Superscript III cDNA synthesis kit (Life Technologies) according to manufacturer's protocol.Gene expression analysis by quantitati v e real-time PCR was performed on a Roche Light Cycler 480 II real-time cycler using the KAPA SYBR FAST qPCR Master Mix (KAPA Biosystems Cat no.740966) to evaluate transcriptional regulations.Gene specific primers were designed using NCBI Primer Blast tool ( https://www.ncbi.nlm.nih.gov/tools/primer-blast/ ) and obtained from Sigma Aldrich.
RT-PCR primers: Details of all the primers used in the study are gi v en in Supplementary Table S1.

DNA fiber assay
Cells wer e tr eated with 25 mM CldU, and then with 250 mM IdU at 37 • C for 20 min each.Subsequently, the cells were washed twice with PBS, trypsinized and adjusted to a pproximatel y 400 000 cells / ml for further processing.To perform DNA spreading, a 2 l cell suspension was placed on the non-frosting side of a microscopic slide (Corning, Cat no.2948-75x25) and allowed to dry for 5-7 min.Then, 7 l of spreading buffer was added and incubated for 2 min.The slides were then tilted at an angle of 20-40 • , so that the droplet moves to the bottom end of the slide at a constant speed.Slides were then fixed by 1 ml of methanol and incubating for 10 min.After denaturing with 2.5 M HCl for 1 h, the samples were subjected to immunostaining with rat monoclonal anti-BrdU and mouse monoclonal anti-BrdU primary antibodies for CldU and IdU respecti v ely.Slides were imaged using Leica confocal microscopy, and LAS softw are (Leica) w as used to quantify the replication fork length across the samples.

Cell fractionation and western blot analysis
For whole cell lysate B16 cell pellet was resuspended in NP40 Lysis Buffer (Thermo; ALF-J60766-AP) reconstituted with protease and phosphatase inhibitors (750 l of NP40 lysis buffer, 1 × PIC, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 10 mM sodium glycerophosphate, 1 mM PMSF).Cells were incubated in lysis buffer overnight at −80 • C and centrifuged at 13 000 rpm for 20 min a t 4 • C , protein superna tant was collected and estimated using standard BCA protocol (Pierce BCA protein assay kit; Thermo).Cellular fractionation was performed by resuspending the cell pellet in hypotonic Buffer A (10 mM HEPES, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 and 0.5 mM DTT) while keeping the pellet on ice and incubating for 5 min.Cells were further lysed using strokes of a hand-held homogenizer.When > 70% of cellular lysis was achie v ed as visualized under 40 × microscope, suspension was centrifuged at 1000 rpm for 5 min at 4 • C. Supernatant (cytosolic fraction) was separated and nuclear pellet was processed similar to whole cell pellet.30 g of the protein was used for proteins in the size range of 20-150 kDa.Primary antibod y incuba tions were done for overnight in cold room (4 • C).Subsequent to incubation with the primary antibody, the membrane was washed thrice with TBST containing 0.1% tween and were further incubated with corresponding HRP conjugated secondary antibody at 1:5000 dilution in 5% BSA 5% Skim milk or for one hour at room temperature.Membrane was washed in TBST with 0.1% Tween-20, thrice for 15 min each and de v eloped using ECL reagent.
Secondary antibodies conjugated to Alexa fluor for immunofluor escence wer e procur ed from Life technologies (USA).Secondary antibodies conjugated to HRP r equir ed for western blot were obtained from GE Healthcare.

Immunocytochemistry on B16 cells
Cells were seeded on UV treated sterile coverslips (Corning) in six well plates.Cells were fixed with 2% paraformaldehyde in PBS for 10 min at 37 • C. Further, they were permeabilized with 0.1% Triton ® X-100 in PBS for 15 min at room temperature on slow orbital shaking.After three consecuti v e washes with 1 × PBS, cells were blocked with 5% NGS (Jackson's immunor esear ch) for 1 hour at room temperature on slow orbital shaking.Cells were gi v en a single wash of PBS and incubated with ␥ H2AX antibody (CST 9718, 1:200) or anti Pol (ab57070, 1:100 in 1% NGS) for 2 h at room temperature.Post primary antibody incubation cells were washed thrice with PBST (PBS + 0.1% Triton X-100) for 5 min each on slow orbital rotation and then incubated in secondary Alexa Flour 568 anti-mouse antibody (1:500 in 1% NGS) for 1 h at room temperature in dark.Finally, cells were washed thrice with IX PBST and mounted on glass slides with 10 l of antifade DAPI and sealed the coverslips using acetone.Imaging was done using Leica confocal microscopy and quantitated using LAS software (Leica).Average nuclear fluorescence intensity was determined using Leica Application suite AF (LAS AF) software.

Generation of tyrosinase mutant B16 cells
Tyr CRISPR was designed using ECRISP ( http://www.ecrisp.org/E-CRISP/).After overlap PCR of the CRISPR, it was in vitro transcribed using mMessage mMachine T7 ULTRA kit (Thermo; AM1345).For transfections, B16 melanoma cells were seeded at a regular density of 2 × 10 5 cells per well of a six well plate.A complex of CRISPR RNA with Cas 9 protein was pr epar ed and incubated for 20 min at room temperature.500 l of optiMEM (gibco; 31985070) along with 3 l of Lipofectamine 2000 (Invitrogen; 11668019) was added to the above mix and incubated again for 20 min.After incubation, media was removed from culture and gi v en a single wash of 1 × DPBS.CRISPR mix was added to these cells after making up the volume to 1 ml with fresh optiMEM.This was incubated for 6 h after which media was changed to terminate transfections.
These transfected cells were used for setting up multiple LD culture flask in T75.The depigmented colonies arising from a single cell were manually picked up by visual selection under microscope followed by in situ trypsinization on Day 7 of LD culture.These colonies were expanded and analyzed for tyrosinase le v els and activity.Control for this is a normally pigmenting unedited clone from the same experiment.
siRNA transfections in B16 melanoma cells B16 cells were seeded in 6-well pla tes a t a density of 2 × 10 5 cells per well.Transfections were carried out with Dharmafect II reagent as per manufacture's protocol.After 48 h, the cells were trypsinised and pelleted for RNA or protein isolations.For low density cultures (LD), cells were transfected on Day 5 of cycle and terminated on Day 7.During transfections, culture media was stored in sterile falcon or flask.After transfections, cultures were gi v en a single wash of DPBS and replaced back with same media for cells to normally pigment during the course of LD cycle.

shRNA transfections and stable knockdown of pol in B16 melanoma cells
All shRNA's (GIPZ Lentiviral constructs) were purchased from Dharmacon (GE Healthcare).Pol shRNA (RMM4532-EG27015) glycerol stocks were used for generation of stable cell lines for Pol knockdown.B16 cells were transfected with Pol silencing plasmid.24 h post transfections, selection of stable cells was done by treatment with Puromycin.Cells were kept under selection pressure for 3-4 weeks by regular change in media with fresh addition of puromy cin e v ery alternate day.In case of confluency, cultur es wer e passaged, with tr ea tment starting a t 24 h post seeding.This was done until all the cells showed consistent GFP fluorescence.

In vivo model of melanoma induction in mice
All animal procedures were done in accordance with the Indian Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA), (Section 15(1) of the Pre v ention of Cruelty to Animals Act, 1960 ; R egistration no.38 / GO / R e Bi / SL / 99 / CPCSEA).Institutional Animal Ethics Committee approved all experiments and procedures carried out on the animals.For induction of melanoma tumors, Pol knockdown stables or the non-silencing control cells were injected in 5 outbred male / female C57BL / 6 mice each at 4-6 weeks of age (National Institute of Imm unolo gy, Delhi).A single subcutaneous administration of 10 6 cells in 100 l volume of cell suspension in right hand side flank region was gi v en.The time and appearance of the first tumor (latency period) as well as the number of mice with tumors (incidence) was recorded during the study.At different time intervals i.e. on Da y 11, Da y 14 and Da y 17, tumour was excised; length and width were measured using a Vernier Caliper.

Statistical analysis and graphs
GraphPad prism was used to plot the graphs.Statistical analysis to obtain significance in the data was performed using Student's t -test, one-way ANOVA or two-way ANOVA.Adjusted P -value ( P ) > 0.05 is marked as ns, ≤ 0.05 is marked as *, P ≤ 0.005 as **, P ≤ 0.0005 as *** and P ≤ 0.00001 as ****.

Cell-based model to study the effect of pr ogressiv e pigmentation
To systematically investigate the effect of melanogenesis on DNA integrity, we resorted to the use of B16 progressi v e pigmenta tion model tha t permits the kinetic stud y of pigmentation and associated cellular changes in melanocytes ( 23 ).Choice of B16 cells that synthesize purely eumelanin, further enabled segregating the role of pheomelanin ( 25 ).In this model, B16 cells are grown at a high cell density of > 10 000 cells / cm 2 which results in depigmented melanocytes.To induce pigmentation, cells are plated at a low cell density of 100 cells / cm 2 .These cells progressi v ely pigment over a course of 7 days.Thereby, cells on day 0 are depigmented, days 4-5 have intermediate pigmentation and days 6-7 cells are completely pigmented, allowing for progressi v e changes to be studied.To further perturb the pigmenta tion, in this stud y we compare the basal pigmenta tion sta te with hyperpigmenta tion induced by the treatment of 1 mM L -tyrosine (Tyr) that serves as a substrate for tyrosinase enzyme and augments melanogenesis (Supplementary Figure S1A).By the pharmacological inhibition of Tyrosinase with 200 M phenylthiourea (PTU), we achie v e decreased pigmentation in this progressi v e pigmentation model.

Melanogenesis causes DNA damage
Melanin contains free radicals, and during its biosynthesis se v eral reacti v e quinone intermedia tes are genera ted that can pass the limiting melanosome membrane and result in damage ( 9 , 26 ).By cytometric DCFDA staining of differentially pigmented cells, we observed the highest DCFDA mean fluorescence intensity in tyrosine treated hyperpigmented cells and least in PTU treated cells in which melanin synthesis is minimal (Supplementary Figure S1).We surmised that these free radicals could cause cellular DNA damage.Nuclear ␥ H2AX foci that are associated with DNA damage and strand br eaks, incr eased with tyrosine treatment in pigmented day 6 cells and pretreatment with the melanogenesis inhibitor PTU, reduced the f oci f ormation (Figure 1 A, B).Furthermore, abasic sites detected by direct ELISA in the genomic DNA followed the same pattern as cellular free radical le v els (Figure 1 C).Howe v er, DNA base oxidation as detected by 8hydro xy-2-deo xyguanosine (8-OHdG) le v els were comparable across the three pigmentation groups (Supplementary Figure S1C).Further, we performed quantitati v e realtime PCR on genomic DNA obtained from differentially pigmented cells after subjecting the DNA to 8-o x oguanine DN A gl ycosylase-1 (OGG-1) (Supplementary Figure S1D).This was confirmed using OGG-1 modified comet assay (Supplementary Figure S1E).In these two assays, we did not observe significant changes induced by OGG-1 treatment in the differentially pigmented samples, possibly indica ting tha t the DNA integrity altered by pigmenta tion is not a direct oxidati v e DNA damage.
Surprisingly not only the ␥ H2AX phosphorylation was elevated with tyrosine treatment and decreased with PTU, the total H2AX le v els also followed a similar trend (Supplementary Figure S1F).In alkaline comet assay, the le v el of associated DNA nicks was high in day 7 pigmented cells, tyrosine treated hyperpigmented cells demonstrated maximal DNA nicks.The least tail moment was observed in the PTU treated depigmented cells (Figure 1 D & Supplementary Figure S1G).The mean tail moment progressi v ely increased in control and tyrosine treated cells on day 7, compared to day 5 as the melanin content increased.We then proceeded to assess DNA breaks by neutral comet assay.In the neutral comet assay we observ ed v ery little breaks and the tail moment was close to 1-2 units in both pigmented and unpigmented cells, and did not observe significant differences between these two states (Figure 1 E).
In all, these experiments suggest that ongoing melanogenesis is associated with DNA damage.On a closer look the following points emerge.Melanogenesis causes free radical generation but does not increase 8-OHdG le v els in the cells.Instead, there is a mar ked ele vation in abasic sites, which is clearly dependent on pigmentation.The increased intensity of ␥ H2AX puncta classically associated with double stranded breaks, could also mark single stranded DNA regions as well as other DNA lesions ( 27 , 28 ).Since the neutral comet assay does not show comet formation upon pigmentation, double stranded breaks would not account for the observed DNA breaks.However, the alkaline comet assay increases the tail moment, progressi v ely and predictably in a pigmentation dependent manner (Figure 1 D).It is interesting to note that the abasic sites are prone to strand cleavage and result in enhanced comets as seen in alkaline comet assa y ( 29 ).Theref ore, the elevated tail moment observed in alkaline comet assay could be contributed by the abasic sites (alkali labile) as well as single stranded DNA lesions, if any, generated by melanin-based damage and the relati v e contribution is difficult to assess at this stage.
Finally, to confirm the role of melanin intermediates in causing the DNA damage, we subjected depigmented B16 cells to 1mM Dihydroxy Indole (DHI), one of the intermediates of eumelanin synthesis reaction, as well as exogenously increased melanin synthesis through ex-cello synthesis of melanin (with the substrate L-DOPA and enzyme Tyrosinase) added in the culture medium and performed alkaline comet assay.We observed an increase in the comet tail length with both the trea tments, confirming tha t melanin intermediates can cause DNA damage (Figure 1 F).The DNA damage due to melanin intermediates was known (5)(6)(7)(8)(9).By the systematic investigation of melanogenesis induction and its pharmacological perturbation, we unequivocally demonstrate the DNA damage using this cell-based model system.
Since melanin is a complex polymer, we could not demonstrate direct evidence for covalent conjugation of melanin intermediates to DNA.Howe v er, plasmid DNA incubated with melanin synthesis reaction (which was further purified to remove unbound melanin intermediates) demonstrated quinone-DNA adduct formation using coupled Nitro-blue tetrazolium (NBT) based detection (Supplementary Figure S1H).In support, DNA incubated with melanin intermediates during melanin synthesis showed altered mobility in agarose gel, while the same was not observed when the DNA is incubated with pre-synthesized melanin (Supplementary Figure S1I).

Melanogenesis induced DNA alterations elicit a replication str ess r esponse
Having observed that melanogenesis causes DNA damage, we then proceeded to decipher the nature of response used to combat this unique challenge faced by melanocytes.Towards this we treated either unpigmented day 0 or pigmented day 7 B16 cells with known DNA damaging agents.UVA and UVB to assess photo-oxidati v e damage, hydroxyurea for replication stress and H 2 O 2 for oxidati v e damage were employed.Pigmented and unpigmented cells that were not treated with any DNA damaging agents were included, and the expression of a panel of DNA repair genes were anal yzed by qRT-PCR anal ysis after 24 h trea tment.Signa ture pattern of expression for each of the damage was clearly distinct, and notably the pattern of gene expression induced by pigmentation closely paralleled that of hydroxyurea treatment that results in replication stress (Figure 2 A).
Molecularly, r eplication str ess is slowing or stalling of replication fork progression associated with DNA replication ( 30 ).Although the importance of replication stress response is well-recognized, molecular e v ents associated with this remain enigmatic.During progressi v e pigmentation using our previous microarray data, we observed that an expanded battery of known 'r eplication str ess r esponse genes' was found to be elevated (Supplementary Figure S2A) ( 31 , 32 ).Validation of this set of genes in depigmented (day 0) and pigmented (day 7) B16 cells confirmed that these genes were induced during pigmentation (Figure 2 B).Expression of these genes was further augmented with tyrosine and reduced with PTU treatment confirming elicitation of replication str ess r esponse during pigmentation (Figure 2 C).Protein le v els as well as the phosphorylation status of RPA2 at Ser 33 residue that are hallmarks of repli-cation str ess r esponse also showed a pigmentation dependent induction (Figure 2 D).Based on DNA fiber assay, we further demonstrate decreased fork length and fork speed (Figure 2 E).
When compared to control pigmented cells, a consistent reduction in the number of cells on day 7 upon hyperpigmentation with tyrosine, and an increase upon depigmentation with PTU was observed (Figure 1 C, top).Cell cycle analysis using propidium iodide (PI) staining followed by flow cytometry to quantify cells in different phases of the cell cy cle re v ealed that hyperpigmented cells have reduced S + G2 / M population confirming a reduction in proliferation (Figure 2 F).Tyrosine treatment not only reduced S + G2 / M population, but it also resulted in a sizeable population of sub-G0 cells suggesting apoptosis.Annexin V-PI based apoptosis assay confirmed that higher uncontrolled hyperpigmentation results in an increase in the apoptotic popula tion.W hile this cell dea th was seen to be higher in tyrosine treated cells which are hyperpigmented, control pigmented cells also showed apoptosis compared to depigmented cells (Supplementary Figure S2C, D).Taken together we conclude that melanin causes DNA damage, and a key response is elicitation of replication stress resulting in the arrest of cell cycle progression.

In response to melanogenesis, translesion polymerase pol is induced by ATR-CHK1 pathway
Response and repair mechanisms would be operational in the pigmented cell to sense and correct the DNA damage.Cells use specialized polymerases to resume DNA synthesis b y b ypassing it.For achieving this a battery of translesion polymerases are known to play an important role ( 33 ).We ther efor e studied the expr ession changes of all the DNA translesion polymerases during pigmentation in B16 melanoma cells.Profiling of expression changes using the earlier microarray data, in DNA translesion polymerases on different days of progressi v e pigmentation revealed that three of the translesion polymerases P olk , P olh and Rev3l were increased in expression at early and latephase of pigmentation (Supplementary Figure S3A).To investigate whether their regulation is critically dependent on pigmentation, B16 cells were treated with PTU, tyrosine or both and on day 7 were subject to qRT-PCR analysis for these three candidate genes.Regulation of Polk paralleled the observed pigmentation (Figure 3 A).Whereas, the other two translesion polymerases Polh and Rev3l were minimally altered and did not show an expected trend in their regulation with pigmentation and its suppression with PTU (Supplementary Figure S3B, C).Thereby, Pol emerged as a promising candida te tha t could have an important role in DNA damage response during pigmentation.
Immunocytochemical localization of Pol and labelling by ␥ H2AX puncta was sim ultaneousl y carried out in differentially pigmenting B16 cells.We observed high Pol as well as ␥ H2AX labelling in tyrosine treated hyperpigmented cells (Figure 3 B, C).Pre v ention of pigmentation by PTU re v ersed the le v els of both Pol and ␥ H2AX staining, substantiating the induction to be pigmentation dependent.Whole cell lysate confirmed the induction of Pol with progressi v e pigmentation (Supplementary Figure S3D).Subsequent cell fractionation followed by western blot analysis on different days of pigmentation confirmed induction of Pol in the progressi v e pigmentation model (Figure 3 D).From these lysates, a clear induction of nuclear localized phosphorylated ATR and phosphorylated CHK1 during pigmentation further indicated activation of DNA repair response pathway during progressi v e pigmentation.
Having observed an induction of Pol during pigmentation and a concomitant increase in the ATR-CHK1 signaling axis, we hypothesized that the ATR-CHK1 pathway could be responsible for Pol induction.To test this, during pigmenta tion, we trea ted the cells with 50 nM of ATR inhibitor AZ20.Western blot analysis confirmed decrease in p-CHK1 le v els and reduction in Pol (Figure 3 E).Hence we identify Pol induction to be under the control of ATR-CHK1 signaling axis.Thereby, this translesional polymerase is likely to function as a key sensor or a responder in combating the pigmentation induced DNA damage.

Genetic ablation of tyrosinase gene curtails melanogenesis and the consequent DNA damage
It is likely that treatment with chemicals such as tyrosine and PTU could trigger the changes in Pol le v els in cells independent of pigmentation.Hence, we genetically ablated Tyrosinase, the key enzyme involved in melanin synthesis using CRISPR based methodology (Supplementary Figure S4A).The strategy involved transfection of high density unpigmented B16 cells with synthesized single guide RNA (sgRNA) against tyrosinase (Tyr) gene and Cas9 protein complex and plating them at a low density (100 cells / cm 2 ) to initiate pigmentation.Individual depigmented and control pigmented colonies deri v ed from a single cell, were then trypsinized and screened for their ability to pigment, TYR protein le v els and in vitro enzyme acti vity involving L-DOPA zymo gra phy (Supplementary Figure S4B, C).Sequence confirmation identified a single base deletion downstream to the sgRNA sequence, resulting in a frameshift mutation (Tyr fs118 ) encoding a truncated protein (Supplementary Figure S4A).Thereby these mutant cells were compromised in their ability to make melanin and could be compared with B16 WT cells for studying DNA damage under conditions of low-density culturing wherein progressi v e pigmentation is induced.
We first assessed the staining of day 7 B16 WT and Tyr fs118 cells with ␥ H2AX foci and detection of newly synthesized DNA using EdU coupled fluorophore labelling (Figure 4 A).B16 WT pigmented cells have a significant overlap of EdU with ␥ H2AX signals indica ting tha t these ar e r egions of DN A damage w here new DN A synthesis is ongoing.Interestingly, these wild type cells showed heterogeneity in double stained population.On closer analysis, the heavily pigmented cells had se v eral doub le positi v e foci, whereas cells with moderate pigmentation had lower levels of colocalized puncta (Figure 4 B).Analysis of the EdU and ␥ H2AX colocalized spots clearly suggested that the pigmented cells had a higher proportion of DNA breaks, that are in the process of repair wherein fresh DNA synthesis is ongoing.The proportion of these cells was around 15% of the total popula tion, indica ting a possible translesion synthesis in only the minority of heavily pigmented cells.Wher eas B16 Tyr fs118 r emain depigmented and had lower le v el of ␥ H2AX staining concomitant to a high EdU puncta per cell.This suggested higher proliferation, which was confirmed by the growth curve analysis of B16 WT and B16 Tyr fs118 cells (Figure 4 C).DNA breaks determined by comet analysis indicated that the mutant depigmented cells had minimal DNA damage, as inferred from alkaline comet assay (Figure 4 D).Furthermore, western blot analysis showed sustained induction of Pol only in B16 WT pigmented cells but not in the depigmented B16 Tyr fs118 cells, in which the le v els decreased from day 5 to day 7 (Figure 4 E).Thereby the overall DNA damaging effects of melanin and the cellular response of melanocytes by invoking Pol is confirmed using this genetic model.While it is tempting to specula te tha t the dif ference in prolifera tion is due to the lack of sustained Pol induction in Tyr fs118 cells, it cannot be unequivocally concluded.This establishes the causal link between melano genesis, DN A damage and Pol during pigmentation.

Pigmentation induced DNA breaks and elicitation of pol r esponse is r ecapitulated in primary human epidermal melanocytes
Since B16 cells are deri v ed from mouse melanoma, it is likely that the observed response could be restricted to melanoma cells and not a physiological response of melanocytes.Hence, we subjected already pigmented nor- mal human epidermal melanocytes (NHEM) that are deri v ed from healthy skin and r epr esent primary melanocytes, to differential pigmentation with PTU and tyrosine for se v en days (Figure 5 A).Assessment of H2AX phosphorylation by western blot analysis re v ealed a similar pattern of increased phosphorylation and a concomitant increase in total protein le v el with hyperpigmentation (Figure 5 A).This was recapitulated in the foci formation detected by imm unofluorescence anal ysis (Figure 5 B, C).Alkaline comet assay re v ealed that the DNA br eaks wer e highest in hyperpigmented cells and lowest in depigmented cells (Figure 5 D).Hence, these observations indicate that pigmentation is indeed a strong physiological response that causes DNA breaks in non-transformed primary cells.We therefore propose that melanogenesis is potentially genotoxic and causes DNA damage in melanocytes.
Subjecting primary melanocytes with varying le v els of pigmentation to qRT-PCR analysis for POLH and POLK as well as the battery of 'replication stress response genes' identified these DNA repair genes to be differentially expr essed (Figur e 5 E).While POLK showed a correlation with pigmentation response, POLH showed an opposite trend.Validating the B16 based observations, se v eral of the r eplication str ess r esponse genes wer e elevated in primary melanocytes treated with tyrosine, strengthening the possibility of DNA damage associated with a Pol response elicita tion a t the RNA le v el.Pol at the protein le v el was induced upon hyperpigmentation and reduced upon depigmentation in cultured primary human melanocytes (Figure 5 A).Involvement of ATR-CHK1 axis in this induction was confirmed using ATR inhibitor AZ20 (Figure 5 F).As the control primary cells are constituti v ely pigmented, the reduction observed with AZ20 was lower as compared to B16 cells wherein the progressi v e pigmentation could be induced in presence of the inhibitor.
We then assessed the ability of melanin modified DNA to elicit a Pol response in melanocytes that do not make melanin.Transfection of plasmid DNA incubated with melanin synthesis reaction (column purified after 16 h of incubation) in depigmented B16 cells induced Pol protein expr ession (Figur e 5 G).We could observe the induction of Pol κ at the RNA le v el in these cells only when transfected with plasmid DNA incubated with melanin synthesis reaction or extracellular melanogenesis achie v ed with 1mM tyrosine and tyrosinase enzyme in the cultur e medium.Wher eas, mer e incubation of pr e-synthesized melanin with plasmid DNA just befor e transfection, pr esynthesized melanin alone, or melanin synthesis reaction mixture purified through the column did not elicit Pol response in these depigmented cells.Similar to the RNA le v el changes we could confirm melanin-modified plasmid DNA to elevate protein levels of Pol (Figure 5 G, inset).The melanin intermediate DHI, that caused DNA breaks in B16 cells (Figure 1 F), was also able to increase POLK mRNA expr ession.Hence, the r esponse of melanocytes to pigmentation is the induction of Pol through melanin-mediated DNA damage.We then proceeded to investigate the role played by Pol in maintaining genome integrity and cellular homeostasis during sustained pigmentation.

Silencing of pol increases DNA damage caused by melanin
We resorted to silencing Pol in B16 cells stably expressing shRNA targeting P ol (shP ol ) and this was com-pared to non-targeting shRNA against luciferase (shNT).B16 cells were freshly transfected and used as an enriched pool to pre v ent multiple consequences of Pol depletion ( 33 ).Under pigmenting conditions both shNT and shPol showed comparable pigmentation by day 7 (Figure 6 A).We observed that the depigmented day 0 cells had minimal ␥ H2AX puncta, and these were comparable across shNT and shPol cells.Pigmented day 7 shPol cells have more mean fluorescence intensity of the foci indicating higher DNA damage (Figure 6 B).With progressi v e pigmentation the le v el of DNA damage incr eased mor e in the shPol cells compared to shNT cells (Figure 6 C).Similarly, single stranded DNA breaks and alkali sensiti v e (abasic) sites assessed by comet assay showed a progressi v e increase in the mean tail moment with pigmentation.This increase was more in the shPol cells on day 7 when the pigmentation was the highest (Figure 6 D).Sequence independent silencing RN A (siRN A) also confirmed this increase in mean tail moment in pigmented cells (Supplementary Figure S5B).shPol cells resulted in an increase in the number of abasic sites, confirming a role for Pol in resolving these lesions (Supplementary Figure S5C).Abasic sites arise from melanin-modifications (Figure 1 C), and this result confirms a critical role for Pol in its repair perhaps by its TLS activity.These results indicate that Pol is r equir ed to resolve the DNA damage caused by melanin.

Pol is necessary to mount a replication stress response in pigmented melanocytes
So far, we were able to establish that melanin causes DNA damage that consequently delays the progression into cell cycle (Figure 1 C, top & Supplementary Figure S2B).Pigmenting melanocytes elicit induction of Pol as a response through ATR-CHK1 axis (Figure 3 E).Further, silencing studies indicated that Pol is necessary to keep a check on the DNA damage caused during pigmentation.Based on these results we anticipated that silencing of Pol would result in more cell cycle arrest and cell death as a consequence.
Interestingly, the number of cells were significantly higher upon Pol knockdown during progressi v e pigmentation (Figure 6 E).While the growth differences were nonsignificant at earlier days, the differences become prominent on day 7, when the pigmentation was highest.Cell cycle analysis revealed an increase in S + G2 / M population suggesting aberrant progression into cell cycle despite pigmentation-induced DNA damage (Supplementary Figure S5D).Additionally, we verified that pigmented shNT and shPol cells exhibited comparable PI-stained populations, indicating similar cell viability, and the annexin-V labeled population showed similar le v els of apoptosis between them (Supplementary Figure S5E, F).This unex-pected observation prompted us to look at the battery of proteins associated with cell cycle.A panel of orchestrators and effectors were analyzed by western blot on day 7 of pigmentation upon knockdown of Pol .Confirming the silencing, le v els of Pol protein was downregulated (Figure 6 F).PCNA is an effector of Pol and a marker of proliferating cells.Howe v er, upon Pol silencing PCNA le v els remained unaltered, perhaps due to opposing effects of proliferation and the absence of key repair polymerase.Howe v er, both p53 and p21 that function to couple DNA damage to cell cycle arrest were low and CDK2 was elevated, further substantia ting tha t more cells are in the proliferati v e phase of cell cycle.This aberrant proliferation was further validated with injection of shNT and shPol cells into the flank of C57 / BL6 mice to grow autologous pigmented tumors.We consistently observed that shPol cells formed tumors with higher volume, confirming that under this pigmenting condition as well, insufficient induction of Pol results in lack of cell cycle arrest (Figure 6 G, Supplementary Figure S5G,  H).Hence, Pol knockdown does not result in cell cycle arrest or cell death as anticipated, instead the DNA-damaged cells continue to proliferate more.
We then assessed the replication stress response genes and observed a pronounced induction in pigmented day 7 shNT cells as observed earlier for control B16 cells (Figures 6 H,  2B ).Pigmented day 7 shPol cells failed to mount such a response and se v eral of the genes were mildly downregulated compared to shNT day 0 cells.Indica ting tha t Pol is a crucial elicitor of r eplication str ess r esponse when cells are challenged with melanin modified DNA.This was further strengthened by the loss of induction of phosphorylated and total RPA2 le v els in day 7 shPol cells (Figure 6 I).
In the absence of Pol , despite the presence of elevated DN A damage, cells aberrantl y pro gress through the cell cycle and proliferate more.This could potentially render these pigmenting cells vulnerable to genome instability.Substantiating this possibility, a recent systematic investigation of combination of DNA damaging agents and repair gene knockouts carried out in C. elegans highlighted an increased mutation frequency for alkylating agents MMS and DMS upon deletion of the Pol ortholog ( 34 ).In the context of melanocytes, we would expect that Pol would be the primary mitigating response to melanin-induced DNA damage, instead we observe that Pol is the primary mediator of cellular response caused by this damaged DNA.At a molecular le v el the mechanism by which a tr anslesion polymer ase lik e Pol w ould be able to elicit a replication stress response remains enigmatic.It is likely that Pol based flagging of melanin lesions may elicit this response by invoking other repair proteins like CHK1.

Signatures of somatic mutations in human melanoma with varying POLK levels
Having established the role of sustained melanogenesis in causing DNA damage and mapping the Pol response by melanocytes, we decided to explore whether this is relevant in the human melanoma context.We extracted the cutaneous melanoma samples from The Cancer Genome Atlas (TCGA) and segregated the samples based on the mRNA expression of POLK ( 35 ).While 83 melanoma samples did not have detectable expression of POLK mRNA by sequencing, FPKM based binning resulted in 207 patients with high le v els of POLK and 175 patients had lower expression of POLK ( 36 ).Overall number of somatic mutations were highest in high-POLK group and least in samples that did not have detectable POLK (Figure 6 J).A systematic two group comparison of soma tic muta tion burden using Welch Two Sample t-test in POLK expressing (high + low) with POLK non-detected group resulted in a P -value < 0.08.The mean in group 'expressed' is 505 whereas the mean in group 'not-expressed' is 380.Thereby, suggesting that expression of POLK could explain a fraction of mutations in human melanoma, and provide physiological relevance to the observations in cultured cells.The survival plot of patients for whom the POLK mRNA data was available (58 high and 44 low) suggested a trend with higher mortality in the high POLK group compared to low POLK group ( P -value < 0.05), which could be explained by the error prone nature of Pol (Figure 6 K).The status of pigmentation of melanoma (compared to cognate skinderi v ed melanocytes) was not available in the TCGA metadata, to empirically assess this we looked at the expression of se v eral pigmenta tion genes for which the da ta was availab le.We observ ed that most of the pigmentation genes were not significantly different across the three groups, suggesting similar distribution of pigmentation (Supplementary Figure S5G).Subsequently, our focus shifted towards identifying distinct signatures within the array of somatic mutational changes associated with varying le v els of Pol e xpression.In a prior study, we had already compared the somatic mutational spectrum of different cutaneous malignancies with both lesional and non-lesional vitiligo skin samples ( 37 ).In this investigation, we reanalyzed the data, specifically focusing on melanoma samples from TCGA, grouped according to their respective Pol expression levels.Notably, SBS7a and SBS7b were prevalent in both groups of melanoma samples, as well as in various other cutaneous tissues, including healthy skin (Supplementary Figure S5J).These signatures were lik ely link ed to sun exposure, a well-known factor in cutaneous muta tional pa tterns ( 20 , 38 , 39 ).
Moreover, we identified additional signatures enriched in melanoma samples.Among them, SBS6 and SBS21 were enriched e xclusi v el y in low Pol tumors, and interestingl y, they are associated with genome instability.Conversely, SBS31 was found to be enriched solely in high Pol tumors, and it is linked to prior chemotherapy with Platinum drugs, which induce DNA adducts, resembling the proposed mechanism involving melanin in this study.Although it would have been intriguing to explore the correlation of SBS signatures with Pol le v els in the context of patients' prior chemotherapy history, unfortunately, due to a lack of available information, we could not establish conclusi v e e vidence in this regar d.
Thereby, in this study we r einfor ce the corr elation of DNA damage and possible mutations with pigmentation.This provides a compelling reason to further investigate this aspect of eumelanogenesis and its implication in melanoma in future.

Figure 1 .
Figure 1.Melanin synthesis causes ␥ H2AX f oci f ormation DNA strand breaks and abasic sites formation.( A ) Immunofluorescence of B16 cells treated with PTU and tyrosine with phosphoryla ted H2AX antibod y.Nuclear DNA stained with DAPI (blue) and ␥ H2AX in (red).Experiment was performed with two biological replicates and a representati v e image is depicted.Scale bar 10 m. ( B ) Quantitation of mean fluorescence intensity per cell of ␥ H2AX from two biological replicates of pigmented day 7 PTU or tyrosine treated cells (shown in A).Data r epr esented as a box plot, horizontal line r epr esents mean and whiskers r epr esent SEM.Ordinary one-way ANOVA was perf ormed f or multiple comparisons.Adjusted P -v alue: * P -v alue < 0.05, *** Pvalue < 0.001, **** P -value < 0.0001.( C ) (Top) Cell pellet of day 7 B16 mouse melanoma cells grown at low density (100 cells / cm 2 ).Cells were left untreated for control treated with tyrosinase inhibitor 200 M phenylthiourea (PTU) or 1mM tyrosinase substrate L-tyrosine (Tyr) for 7 days.Number of cells, mean ± SEM across three biological replicates is depicted below the image of the cell pellet.Numbers r epr esent mean ± SEM cell counts across biological triplicates.(Bottom) Number of abasic sites in the genomic DNA was estimated by an aldehyde specific conjugation of biotin and subsequent detection using streptavidin based detection.Using standards, abasic sites per 10 5 bp is estimated.Bars r epr esent mean ± SEM across duplicate biological experiments, each conducted in triplicates.Ordinary one-w ay ANOVA w as performed for multiple comparisons * P -value < 0.05, ** P -value < 0.01, *** P -value < 0.001.( D ) Single cell electrophoresis followed by comet analysis of B16 cells undergoing varying le v els of pigmentation in the presence of PTU and tyrosine (alkaline comet assay).Experiment was carried out at mid phase (day 5) and late phase (day 7) of pigmentation.Mean tail moment distribution across each population of duplicate biological experiments with atleast 50 comets analyzed is depicted by a violin plot.Two-way ANOVA was performed.Adjusted P values; ns non-significant, * P -value < 0.05, ** P -value < 0.01, *** P -value < 0.001, **** P -value < 0.0001.( E ) Neutral comet assay on B16 unpigmented (day 0) and pigmented (day 7) cells.Mean tail moment distribution across each population of duplicate biological experiments with atleast 50 comets analyzed is depicted by a violin plot.Student's unpaired t-test was performed.P values ns non-significant.( F ) Single cell electrophoresis followed by comet analysis of B16 cells untr eated, tr eated with DMSO for 24 h, melanin synthesis ( ex-cellulo L-tyrosine and tyrosinase added to cell media) for 24 h or cells treated with 1 mM dihydroxyindole (DHI) for 24 h (alkaline comet assay).Mean tail moment distribution across each population of duplicate biological experiments with atleast 50 comets analyzed is depicted by a violin plot.Ordinary one-way ANOVA was performed.Adjusted P values: * P -value < 0.05, **** P -value < 0.00001.

Figur e 2 .
Figur e 2. Melano genesis induces r eplication str ess r esponse.( A ) Fold change in mRNA le v els of a panel of DNA repair genes by qRT-PCR analysis.The heat map r epr esents fold change and compares DNA repair gene signature of depigmented (left), pigmented (middle) B16 cells to various DNA dama ging a gents.Untr eated pigmented cells compar ed to depigmented control is depicted as a heat map (right).( B ) Heat map of expression changes (fold change) in mRNA le v els of a panel of known 'replication stress response' genes by qRT-PCR analysis in B16 melanoma unpigmented day 0 cells compared to pigmented day 7 cells across at least two biological replica tes.( C ) Hea t map of expression changes (fold change) in mRNA le v els of a panel of known DNA replication stress response genes by qRT-PCR analysis in B16 melanoma pigmented day 7 cells treated with PTU or tyrosine across three biological replicates.( D ) Western blot images and quantitation of p-RPA2 and total RPA2 in B16 melanoma unpigmented day 0 cells compared to pigmented day 7 cells.Numbers below r epr esent beta-actin normalized fold changes wrt shNT at day 0. Two biological replicates were performed.( E ) DNA fiber assay of unpigmented day 0 and pigmented day 7 B16 mouse melanoma cells.(Top) Representati v e images of the DNA fiber tracts are shown.(Bottom) Quantification of fork length ( m) and fork speed (in kb per minute) is shown as bar graphs.Bars r epr esent mean ± SEM across triplicate biological experiments.Student's unpaired t -test was performed.* P -value < 0.05.Scale bar 5 m. ( F ) Cell cycle analysis carried out using propidium iodide staining detected by FACS is depicted as stacked bars.Sub G0 population is marked as apoptotic.Experiment was carried out in biological triplicates and mean ± SEM is depicted.Two-way ANOVA was performed.Adjusted P value: * P -value < 0.05 across comparison of G0 / G1 population.

Figur e 3 .
Figur e 3. Melano genesis induces Pol via ATR-CHK1 signaling axis.( A ) mRNA le v els of Polk in B16 cells that are allowed to pigment differentially in the presence of 200 M PTU, 1 mM tyrosine or both.Bars r epr esent per cent mRNA le v els compared to control as mean ± SEM across fiv e biological replicates.Ordinary one-w ay ANOVA w as performed.Adjusted P v alues: * P -v alue < 0.05, **** P -value < 0.00001.( B ) Immunofluorescence of B16 cells treated with 200 M PTU, 1 mM tyrosine or both.Top panel is the bright field (BF) and dark granules are the pigment accumulation.Nuclear DNA stained with DAPI (blue), Pol in red and ␥ H2AX in (green).Scale bar 10 m.Experiments were performed in triplica tes.( C ) Quantita tion of mean nuclear fluorescence intensity of Pol of differentially pigmented day 7 B16 cells treated with PTU, tyrosine or both (shown in B).Bars represent mean ± SEM across two biological replicates.Ordinary one-way ANOVA was performed.Adjusted P values: * P -value < 0.05, ** P -value < 0.005, **** P -value < 0.00001.( D ) Western blot analysis of nuclear and the post-nuclear (cytoplasmic) lysates of B16 cells on day 5, day 6 and day 7 of pigment accumulation.Numbers below the blot correspond to day 5 normalized expression of the indicated protein.Experiment was performed in duplicates.( E ) Western blot analysis of B16 cells treated with DMSO or 50 nM AZ20, a selecti v e inhibitor of ATR kinase.Numbers below the blot correspond to control normalized expression of the indicated protein wrt GAPDH.Experiments were performed in duplicates.

Figure 4 .
Figure 4. CRISPR-based genetic ablation of tyrosinase prevents melanogenesis, reduces DNA breaks and curtails Pol induction.( A ) Immunofluorescence images of B16 WT and B16 Tyrosinase mutant (Tyr fs118 ) cells that w ere allow ed to pigment for 7 days.Bright field images indicate the presence of melanin granules.Labelling using 5-ethynyl-2 -deoxyuridine (EdU) (red) and ␥ H2AX antibody (green).Merged images show co-localization of ␥ H2AX with EdU (yellow).A zoomed in view of a single cell is shown as an inset.Scale bars r epr esent 10 m. ( B ) (Top) Cell pellets of B16 WT and Tyrosinase mutant (Tyr fs118 ) cells on days 5, 6 and 7 of pigmentation.(Bottom) Quantitation of average of the number of ␥ H2AX puncta that are single positi v e or double positi v e (colocalizing with EdU puncta) per cell across 50 cells in B16 WT and B16 Tyr fs118 cells is depicted as a violin plot across three biological replicates.The percent population of cells with indicated puncta is mentioned alongside.( C ) Growth curve analysis of B16 WT and Tyr fs118 cells on days 0, 5, 6 and 7 of pigmentation.Each point r epr esents mean ± SEM across biological triplicates.Two-w ay ANOVA w as performed.Adjusted P values: * P -value < 0.05, *** P -value < 0.001.( D ) B16 WT and Tyr fs118 cells on days 5 and 7 of pigmentation were subjected to single cell electrophoresis and alkaline comet assay was performed.Mean tail moment distribution across each population of duplicate biological experiments with at least 50 comets analyzed is depicted by a violin plot.Two-w ay ANOVA w as performed.Adjusted P values ns non-significant: **** P -value < 0.0001.( E ) Western blot analysis of whole-cell lysates of B16 WT and B16 Tyr fs118 cells on day 5 and day 7 stages of pigmentation.Numbers below the blot correspond to day 5 GAPDH normalized expression of the indicated protein.Experiments were performed in biological duplicates with similar results.

Figure 5 .
Figure 5. Normal human epidermal melanocytes (NHEM) respond to pigmentation induced DNA breaks by elevating Pol .( A ) NHEM cells wer e tr eated with 200 M PTU or 1mM tyrosine for 7 days for differential pigmentation.(Top) Cell pellet, (bottom) western blot analysis of cell lysates with POLK, HSC70, phosphorylated H2AX, total H2AX and beta actin antibodies.Numbers below the blot correspond to control normalized expression of the indicated protein.Experiments were performed in biological duplicates.( B ) Immunofluorescence of NHEM treated with PTU or tyrosine with phosphorylated H2AX antibody.Nuclear DNA stained with DAPI (blue) and ␥ H2AX in (red).Experiments were performed with two biological replicates.Scale bar 10 m. ( C ) Quantitation of mean fluorescence intensity per cell of ␥ H2AX from two biological replicates of NHEM treated with PTU or tyrosine (shown in B).Ordinary one-way ANOVA was performed for multiple comparisons.Adjusted P values: * P -value < 0.05, **** P -value < 0.0001.( D ) PTU and tyrosine treated NHEM cells were subjected to single cell electrophoresis and comet analysis (alkaline comet assay).Mean tail moment distribution across each population of duplicate biological experiments with atleast 50 comets analyzed is depicted by a violin plot.Ordinary one-w ay ANOVA w as performed.Adjusted P values: ** P -value < 0.001, **** P -value < 0.00001.( E ) Heat map of expression (fold change) in mRNA le v els of top two translesion polymerases (that were enriched in B16 microarray) (top), and a panel of known DNA replication stress response genes by qRT-PCR analysis in NHEM (Contr ol, PTU or tyr osine trea ted).Da ta r epr esented as mean of triplicate biological e xperiments.( F ) Western b lot analysis of NHEM treated with DMSO or 50 nM AZ20, a selecti v e inhibitor of ATR kinase, for 24 h.Numbers below the blot correspond to control normalized expression of the indicated protein wrt beta-actin.Experiments were performed in biolo gical duplicates.( G ) mRN A le v els of Polk in unpigmented B16 cells mock transfected, or with either control DNA, melanin modified DNA (plasmid DNA was incubated with L-DOPA and tyrosinase and column purified after 24 h) (Mel + DNA), DNA mixed with pre-synthesized melanin and coulmn purified [DNA+(Mel)], in-vitro melanin synthesis ( ex-cellulo L -tyrosine and tyrosinase added to cell media) for 24 h or cells treated with 1 mM DHI (DHI) for 24 h.Bars r epr esent per cent mRNA le v els compared to contr ol acr oss biological triplicates.Ordinary one-w ay ANOVA w as performed.Adjusted P v alues: * P -v alue < 0.05.(Inset) Western blot analysis of B16 cells transfected with onl y DN A (Con DNA) or melanin-modified DNA (Mel + DNA) with Pol antibody normalized to HSC70.Experiments were performed in biological triplicates.

Figure 6 .
Figure 6.Silencing of Pol during pigmentation pre v ents replication stress response despite elevated DNA damage.( A ) Cell pellets of control non-targeting (shNT) and Pol silenced (shPol ) B16 cells on day 0 (left) and day 7 (right) of pigmentation.( B ) Immunofluorescence analysis of day 7 pigmented shNT and shPol cells with phosphorylated H2AX antibody (puncta labelled in green) and the nucleus is counterstained with DAPI (blue).Scale bars represent 10 m. ( C ) Quantitation of mean fluorescence intensity of ␥ H2AX (shown in B) from two biological replicates of shNT and shPol cells across day 0, 5 and 7 of pigmentation induction.Two-way ANOVA was performed for multiple comparisons.Adjusted P values * P -value < 0.05 *** P -value < 0.0005 **** P -value < 0.00001 ns non-significant.( D ) shNT and shPol expressing pigmented B16 cells were subjected to single cell electrophoresis and comet analysis (alkaline comet) on days 0, 5 and 7 of pigmentation induction.Mean tail moment distribution across each population of duplicate biological experiments with at least 50 comets analyzed is depicted by a violin plot.Two-w ay ANOVA w as performed for multiple comparisons.Adjusted P values, ns non-significant, * P -value < 0.05.( E ) Growth curve analysis of shNT and shPol expressing B16 cells on days 0, 5, 6 and 7 of pigmentation.Each point r epr esents mean ± SEM across biological triplicates.Two-way ANOVA was performed.Adjusted P values: * P -value < 0.05, *** P -value < 0.001.( F ) Western blot analysis of DNA repair and cell cycle related proteins in shNT and shPol cells.Numbers below represent tubulin normalized fold changes wrt shNT.Experiments were performed in biological duplicates.( G ) shNT and shPol expressing B16 cells were injected inside the flank of C57 / BL6 mice and allowed to grow as tumors.The volume of the tumor was non-invasi v ely monitored and plotted over time of biological triplicates mean ± SEM.Two-w ay ANOVA w as perf ormed f or multiple comparisons.Adjusted P v alues: * P -v alue < 0.05.( H ) Heat ma p of expression (fold change) in mRN A le v els of a panel of known DNA replication stress response genes by qRT-PCR analysis in shNT (day 0-unpigmented and day 7-pigmented) and shPol (day 0-unpigmented and day 7-pigmented) B16 cells.( I ) Western blot images and analysis of p-RPA2 and total RPA2 in shNT and shPol B16 cells (day 0 unpigmented and day 7 pigmented).Numbers below r epr esent beta-actin normalized fold changes wrt shNT at day 0. Experiments were performed in biolo gical triplicates.( J ) Anal ysis of melanoma samples from TCGA data for mRNA expression of POLK (high, low or not detected) segregated into bar plots and proportion of mutations were plotted on y-axis.( K ) Survival plot of melanoma patients with low or high expression of POLK from TCGA data.Analysis from Human Protein Atlas database.P air ed t -test P value 0.017.