Buthionine sulfoximine diverts the melanogenesis pathway toward

23 Buthionine sulfoximine (BSO) is a specific inhibitor of γ-glutamylcysteine synthetase, thus 24 blocking the synthesis of glutathione (GSH). It is known that this makes that BSO affects melanin 25 synthesis because of the role of thiols in melanogenesis. However, BSO may also react with the 26 intermediate oxidation products of melanogenesis, a possibility that has not been investigated from 27 the initial steps of the pathway. We created in vitro conditions simulating eumelanogenesis 28 (oxidation of L-DOPA in the absence of GSH) and pheomelanogenesis (oxidation of L-DOPA in 29 the presence of GSH) under presence or absence of BSO. BSO made that eumelanogenesis results 30 in pigments more soluble and less resistant to degradation by hydrogen peroxide than pigments 31 obtained without BSO. A similar but less marked effect was observed for pheomelanogenesis only 32 at subsaturating concentrations of GSH. These results suggest that BSO diverts the melanogenesis 33 pathway toward the production of more soluble and degradable pigments. 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 Buthionine sulfoximine (BSO) is a specific inhibitor of γ-glutamylcysteine synthetase, the enzyme 51 that catalyzes the rate-limiting step in the synthesis of glutathione (GSH), where two of its three 52 constitutive amino acids (glutamate and cysteine) are bonded. BSO thus decreases intracellular 53 GSH levels with no side, toxic effects. The important antioxidant activity of GSH makes that 54 BSO inhibits the growth of different tumour cell lines and increases their sensitivity to 55 antineoplastic drugs. The inhibitory effect of BSO is particularly high against melanoma-derived 56 cell lines, as melanoma may be dependent on the role of GSH and its linked enzymes in melanin 57 synthesis. However, the effect of BSO on the synthesis of different types of melanin has received 58 little attention. This may have important consequences for understanding possible side effects of 59 BSO use. 60 The first step in the melanogenesis pathway, catalyzed by the enzyme tyrosinase, is the 61 oxidation of the amino acid L-tyrosine to L-dopaquinone (see Fig. 1) which undergoes an 62 intramolecular cyclization of the amino group to give L-dopachrome via L-cyclodopa, which in turn 63 suffers a redox exchange with L-dopaquinone that produces L-dopachrome and 3,4-dihydroxy-L64 phenylalanine (DOPA). In this process, the recruiting of L-DOPA to be re-oxidized to L65 dopaquinone is again catalyzed by tyrosinase, so that half of the DOPA oxidized to L-dopaquinone 66 is reduced back to DOPA. L-Dopaquinone is the common precursor of the two connected 67 biosynthetic routes in melanogenesis that lead to the formation of either eumelanin or 68 pheomelanin. In the absence (or below certain concentration) of sulfhydryl groups from thiol 69 compounds such as cysteine or GSH, dopachrome evolves to two dihydroxyindoles, DHICA (5,670 dihydroxyindole-2-carboxylic acid) or DHI (5,6-dihydroxyindole) by tautomerization or 71 decarboxylation, respectively. The resulting DHI/DHICA ratios depend on the level of dopachrome 72 tautomerase activity and/or the presence of some metal ions. DHICA and DHI are further oxidized 73 and polymerized to form eumelanin (Fig. 1). In the presence (or above certain concentration) of 74 sulfhydryl-containing compounds, these conjugate with L-dopaquinone to generate mainly 5-S75 cysteinyldopa (in the presence of cysteine) or 5-S-glutathionyldopa (in the presence of GSH). These 76 and other thiol-DOPA conjugates are further oxidized and polymerized to form pheomelanin 77 (Fig. 1). Given this biochemical process, the reduction of GSH levels exerted by BSO should 78 decrease pheomelanin production and increase eumelanin production. 79 However, the effect of BSO on the synthesis of melanins may not be only mediated by the 80 inhibition of γ-glutamylcysteine synthetase and cysteine production, because the S=NH group of 81 BSO may also behave as the sulfhydryl groups (S-H) of cysteine and GSH, and thus BSO may react 82 directly with some of the intermediate oxidation products of the melanogenesis pathway and that 83 would affect the production of pigments. To investigate the latter possibility, we conducted an in 84 vitro experiment to go deeper into the chemical reactivity properties of BSO. Benathan and 85 coworkers have previously reported increases in the pheomelanin precursor 5-S-cysteinyldopa 86 and decreases in total pigmentation in in vivo cells exposed to BSO. However, the possibility that 87 the S=NH group of BSO reacts with melanogenesis intermediates has never been explored. The 88 expectation should be that BSO reacts with dopaquinone to form BSO-DOPA conjugates and thus 89 diverts the melanogenesis route, decreasing the synthesis of both pheomelanin and eumelanin 90 similarly to other agents containing sulfhydryl groups. 91 To investigate the possibility that BSO reacts with intermediates of the melanogenesis 92 pathway, we reproduced under in vitro conditions the initial steps of the melanogenesis pathway 93 oxidizing L-DOPA with tyrosinase in the presence or absence of GSH, and in the presence or 94 absence of BSO. The oxidation of L-DOPA without GSH should thus be similar to 95 eumelanogenesis, while the oxidation of L-DOPA in the presence of GSH should simulate 96 pheomelanogenesis. We used GSH instead of cysteine as a sulfhydryl compound because the 97 former is more abundant in intracellular media than free cysteine, so pheomelanogenesis in the 98 presence of GSH may resemble more closely the in vivo situation. In any case, 5-S99 glutathionyldopa, which is the major thiol-conjugated species formed when L-dopaquinone reacts 100 with GSH, releases 5-S-cysteinyldopa, the main intermediate of the monomeric subunits for 101 pheomelanin, after the action of a dipeptidase and the pathway then progresses as when only 102 cysteine is present. 103 To explore the influence of BSO on melanogenesis, five solutions were prepared in 104 cryogenic vials all containing 10 mM L-DOPA and 90 mM BSO (DL-buthionine-(S,R)105 sulfoximine) in 1 ml of saline phosphate buffer (4mM KHCO3, 2mM CaCl2 2H2O, 20mM NaHCO3, 106 138mM NaCl and 2mM KCl), pH 7.4. One of these solutions did not contain GSH, other two 107 solutions contained GSH at concentrations (0.65 and 3.2 mM) that were lower than the 108 concentration of the substrate L-DOPA, and finally other two solutions contained GSH at 109 concentrations (16.3 and 81.3 mM) that were higher than the concentration of the substrate L110 DOPA but lower than that one of BSO. 36.1 μg mushroom tyrosinase (1715 units/mg) diluted in 20 111 μl of the saline phosphate buffer were added to all tubes. Other five solutions were prepared as 112 explained before, except that these did not contain BSO, thus serving as appropriate controls for 113 standard euor pheomelanogenesis. All solutions were made in duplicates. All products were 114 purchased from Sigma-Aldrich (St. Louis, MO, USA). 115 The chosen concentration of L-DOPA uses to be employed in studies of melanin synthesis 116 in vitro and is around the natural concentrations found in at least some cells. The chosen 117 concentration of BSO was intermediate between the lowest (4.5 mM) and medium (184.4 mM) 118 doses used in a previous in vivo experiment with a bird species (the great tit Parus major), which 119 did not detect significant mortality or effects on body condition. The chosen concentrations of 120 GSH were also around the concentration of total GSH found in the erythrocytes of control birds (3.9 121 mM) in the same study. This makes that the results of this in vitro experiment should be useful to 122 make biologically significant predictions for the influence of BSO on melanogenesis. 123 To characterize the products of the oxidation of L-DOPA under different conditions, after 124 the addition of tyrosinase the solutions were kept in a shaker for 48 h and then centrifuged at 3000 g 125 and 10 oC during 10 min. The absorbance of the resulting supernatant was measured in the range 126 325-800 nm (2 nm intervals) with a Thermo Scientific Genesys 10S Vis spectrophotometer 127 (Thermo Fisher Scientific, Madison, WI, USA). As precipitates were formed in some tubes, we also 128 measured their resistance to chemical degradation. For this, the precipitates were redissolved after 129 previous drying all tubes in a heater at 65 oC for 1.5 h. Then, 250 μl of 10 M NaOH was added to 130 500 μl of supernatant in the tubes where precipitate was not formed and 750 μl of 3.3 M NaOH was 131 added to the tubes with precipitate to keep NaOH concentration 3.3 M in all tubes. The precipitates 132 were then exposed to alkaline degradation in a dry block heater at 90 oC during 24 h until complete 133 re-solution. Then, tubes were centrifuged at 3000 g and 20 oC during 5 min, and the absorbance at 134 400 nm was measured as an estimate of total melanin present. As H2O2 degrades melanin 135 polymers and produces the bleaching of the pigments, 100 μl of 35% H2O2 was then added to all 136 tubes, which were then vortexed two times during 15s and kept for 1.5 h. The absorbance at 400 nm 137 was then measured again in all tubes. By comparing the absorbance at 400 nm of the solutions 138 before and after the addition of H2O2, we could therefore determine the resistance to chemical 139 degradation of the products obtained from the oxidation of L-DOPA in the presence of both GSH 140 and /or BSO at different concentrations as stated above. 141 The shape of the curves with and without BSO was compared by analyzing the mean values 142 of absorbance and the mean values of the slopes of all tangential lines in the curves ((ya-yb/xa-xb); 143 being y absorbance, x wavelength and a and b two adjacent points in the curve). Mean values of 144 absorbance and slopes were compared by one-way ANOVA considering all values of absorbance 145 (N = 238) and slopes (N = 237) in the curves. Mean values in the absorbance decrease produced by 146 H2O2 bleaching were not compared by ANOVA because these means were calculated with only two 147 points (i.e., absorbance at 400 nm of duplicate solutions), so only the mean values are shown. 148 The absorbance spectra of the solutions resulted from the oxidation of L-DOPA under 149 different conditions are shown in Fig. 2. The simulation of eumelanogenesis (i.e., oxidation of L150 DOPA without GSH) showed that, when L-DOPA was oxidized in the absence of BSO, the result 151 was a dark solution with a typical absorbance spectrum of eumelanin with absorbance progressively 152 decreasing with increasing wavelength in the range 300-800 nm. Insoluble black precipitate was 153 formed. When L-DOPA was oxidized in the presence of BSO, the result was a very dark solution 154 with different spectra in which absorbance decreased less steeply and thus absorbance values were 155 greater (Fig. 2). As a consequence, the mean absorbance value and the mean slope value of the 156 curve were higher and lower, respectively, than those of the solution resulted from the oxidation of 157 L-DOPA in the absence of BSO (mean ± SE: absorbance: no BSO = 1.21 ± 0.2, BSO = 1.67 ± 0.02, 158 F1,474 = 180.27, P < 0.0001; slope: no BSO = 2.76 x 10 ± 2.40 x 10, BSO =1.96 x 10 ± 2.40 x 159 10, F1,472 = 5.49, P = 0.019). Precipitate was not observed in the solution with BSO. These results 160 suggest that BSO diverts the eumelanogenesis pathway towards the formation of more soluble 161 products which remain dissolved in the solution. Thus, the solution in the presence of BSO is 162 therefore darker (i.e., higher absorbance and lower slope values) and does not form precipitate (i.e., 163 insoluble eumelanin) than in the absence of BSO, where the extensive polymerization of 164 intermediates gives place to large eumelanin polymers that precipitate leaving a relatively light 165 solution. 166 H2O2 bleached the oxidation products as reflected by a decrease in absorbance at 400 nm in 167 all solutions, but this decrease was higher in the solutions with BSO (mean absorbance at 400 nm: 168 0.54 ± 0.11) than in those without BSO (0.39 ± 0.28). This suggests that the products formed in the 169 presence of BSO may be more easily metabolized and excreted than insoluble eumelanins formed in 170 the absence of BSO. Thus, BSO under these in vitro conditions inhibited the production of standard 171 eumelanin and probably induced the production of other unknown pigments. These pigments may 172 result from the addition of the S=NH group of BSO to L-dopaquinone to form BSO-DOPA 173 conjugates, thus decreasing the formation of L-dopachrome (Fig. 1) and impairing, at least partially, 174 the eumelanogenesis pathway. 175 The simulation of pheomelanogenesis (i.e., oxidation of L-DOPA in the presence of GSH) 176 showed that, at subsaturating concentrations of GSH (0.65 and 3.2 mM), the oxidation of L-DOPA 177 resulted in spectra that were not very different in shape from those generated when L-DOPA was 178 oxidized alone (Fig. 2). Indeed, absorbance values were even lower in the solutions with BSO than 179 in those without BSO (0.65 mM GSH: no BSO = 1.86 ± 0.2, BSO = 1.67 ± 0.02, F1,474 = 69.87, P < 180 0.0001; 3.2 mM GSH: no BSO = 1.81 ± 0.2, BSO = 1.45 ± 0.02, F1,474 = 124.70, P < 0.0001), and 181 the slope values of the curves were also higher in the solutions with BSO (0.65 mM GSH: no BSO 182 = 9.40 x 10 ± 2.27 x 10, BSO = 2.27 x 10 ± 2.27 x 10, F1,472 = 17.16, P < 0.0001; 3.2 mM 183 GSH: no BSO = 1.64 x 10 ± 2.22 x 10, BSO = 2.83 x 10 ± 2.22 x 10, F1,472 = 14.61, P < 184 0.001). At the lowest concentration of GSH (0.65 mM), dark insoluble precipitate, which may be 185 insoluble pheomelanin or eumelanin escaping trapping of dopaquinone by GSH, was formed in the 186 solution without BSO, but not when BSO was present. Under higher GSH concentration (3.2 mM), 187 precipitate was not observed regardless BSO presence. 188 The oxidation products formed in the presence of BSO were always degraded by H2O2 189 easier than when BSO was not present, as reflected by a greater decrease in absorbance of the 190 oxidation products with BSO plus 3.2 mM GSH (mean absorbance at 400 nm: no BSO: 0.10 ± 0.33; 191 BSO: 0.44 ± 0.08). Thus, BSO also inhibits the production of pheomelanin and instead induces the 192 production of other related pigments which may arise from BSO DOPA conjugates avoiding the 193 formation of cysteinyldopas and thus impairing the pheomelanogenesis pathway (Fig. 1). However, 194 this effect is less marked than in eumelanogenesis. 195 Lastly, oxidation of L-DOPA under saturating concentrations of GSH (16.3 and 81.3 mM) 196 resulted in transparent solutions and thus a lack of absorbance, regardless the presence of BSO (Fig. 197 2). This lack of absorbance is expected from the fact that GSH bonds to the copper active site of 198 tyrosinase, thus inhibiting the enzyme. GSH does not react with L-DOPA, but with L199 dopaquinone, so it is necessary that tyrosinase first oxidizes L-DOPA to L-dopaquinone. Under 200 high concentrations of GSH, tyrosinase is almost completely inhibited by GSH. This greatly 201 prevents the formation of 5-S-glutationyldopa and the further generation of pheomelanins. In 202 agreement with that, the solutions were thus transparent because L-DOPA, GSH and BSO (when 203 present) were mostly unaltered without the initial tyrosinase action on L-DOPA. This suggests that 204 the effect of BSO on melanogenesis will depend on the levels of thiol-agents such as L-cysteine and 205 GSH. 206 Our results indicate that BSO reacts directly with some of the intermediate oxidation 207 products of the melanogenesis pathway, diverting the synthesis of eumelanin and pheomelanin 208 toward the production of more soluble pigments. BSO may therefore be a competitive inhibitor of 209 melanogenesis, probably because its S=NH group reacts with dopaquinone to form BSO-DOPA 210 conjugates (Fig. 3) that avoid the formation of dopachrome and cysteinyldopas and thus the 211 synthesis of eumelanin and pheomelanin, respectively. This result is more marked for 212 eumelanogenesis than for pheomelanogenesis, which may be due to a closer similarity between the 213 hypothesized BSO-DOPA conjugates and cysteinyldopas (which are sulfur-containing like BSO) 214 than between those conjugates and dopachrome (which does not contain sulfur), making that the 215 properties of the final products of pheomelanogenesis in the presence of BSO are not significantly 216 different from pheomelanins. However, the possibility that BSO competes with GSH for 217 dopaquinone should not be discarded. This is likely, according to Fig. 3. Future studies should 218 determine the presence and structure of BSO-DOPA conjugates. 219 The impairment of the melanogenesis pathway suggests that BSO has depigmenting 220 properties. Indeed, other depigmenting agents like lipoic acid derivatives exert their inhibitory 221 effects on melanin synthesis by adding their sulfhydryl groups to dopaquinone to form derivatives 222 that avoid dopachrome formation, similarly to the mechanism proposed here for the S=NH group 223 of BSO (Fig. 3). There are at least two other compounds that are known to act as depigmenting 224 agents by acting on dopaquinone during melanin synthesis: ascorbic acid, which reduces 225 dopaquinone and blocks DHICA formation, and cysteamine, which favor pheomelanin synthesis by 226 nucleophilic addition to dopaquinone. Solano et al. have already highlighted the hypopigmenting 227 nature of BSO, but our study shows that this effect is not only mediated by an inhibition of GSH 228 synthesis, but also by direct reactions of BSO with melanogenesis intermediates. This suggests that, 229 if supplemented with GSH to counteract the inhibitory effect of γ-glutamylcysteine synthetase, BSO 230 might have depigmenting effects without affecting the antioxidant capacity, with the corresponding 231 relevance for cosmetic use. Indeed, it has already been shown that the oxidative damage due to 232 BSO-induced depletion of GSH is avoided if BSO is administered with other antioxidants such as 233 ascorbate. BSO is widely used in pharmacology due to its specificity and efficiency in changing 234 the metabolism of sulfur-containing amino acids, so its pharmacological use as a depigmenting 235 agent should be considered. The impairment of conventional pheomelanin synthesis by BSO may 236 be particularly relevant to avoid the phototoxicity of pheomelanin derived from the generation of 237 superoxide anion under exposure to ultraviolet (UV) radiation. Lastly, it may be speculated that 238 the impairment of melanin synthesis by BSO may contribute to avoid the resistance of melanoma to 239 radiotherapy because of the high absorption of energy by eumelanin, which would add to the 240 increase in radiosensitization of melanoma mediated by GSH depletion. This may create synergies 241 with other treatments against melanoma. Future studies should explore these possibilities. 242