Effect of detergents and endogenous lipids on the activity and properties of tyrosinase and its related proteins

ELSEVIER et Biophysics 4cta

Biochimica et Biophysics Acta 1243 (1995) 421-430

Effect of detergents and endogenous lipids on the activity and properties of tyrosinase and its related proteins

Celia Jimhez-Cervantes, Jose Carlos Garcia-Borrh, Jose Antonio Lozano, Francisco Solano *

Department of Biochemistry and Molecular Biology, School of Medicine, University of Murcia, 30100 Murcia, Spain

Received 1 June 1994; accepted 19 October 1994

Abstract

Within mammalian melanocytes, melanin biosynthesis is controlled by three enzymes structurally related: tyrosinase and two tyrosinase related proteins, TRPl and TRP2. These melanosomal enzymes are integral membrane proteins with a carboxyl tail oriented to the cytoplasm, a single membrane-spanning helix and the bulk of the protein located inside the melanosome. Their solubilization is usually carried out by treatment of melanosomal preparations with non-ionic detergents, but, so far, no comparative study of the effect of the detergents employed on the properties of the solubilized proteins has been reported. We have compared the effect of the detergents Brij-35, Nonidet P-40, Tween-20, sodium deoxycholate and Triton X-114 on several properties of the melanogenic enzymes, including the solubilization yield, stability, electrophoretic behaviour and accessibility of epitopes located in the carboxyl tail to specific antibodies. Our data indicate that not only the total amount of enzymes solubilized, but also their relative proportions in the solubilized preparations depend on the detergent used. The non-ionic detergents apparently interact strongly with the melanogenic enzymes, affecting their mobility in SDS-PAGE, and might induce different conformations of the carboxyl tail. Complete replacement of lipids by the detergents results in a decreased stability that can be partially reversed by the addition of endogenous lipids. This treatment also produces a noticeable activation of the ltyrosinase isoenzymes, which is higher for TRPl than for tyrosinase. Taken together, these data show that the transmembrane and carboxyl fragments of the proteins of the tyrosinase family might modulate the stability and activity of the melanogenic enzymes.

Keywords: Tyrosinase; Dopachrome tautomerase; Melanosome; Detergent; Lipid; Melanization; Membrane protein

1. Introduction

Melanin is the main pigment found in skin, hair, eyes and inner ear of mammals. This pigment has a protective function against UV light in the skin, as well as other functions related to tha tissue where it occurs [1,2]. Melanogenesis proceeds in specialized cells called melanocytes. The key enzyme of the pathway is tyrosinase (EC 1.14.18.11, an enzyme catalyzing three different reac-

Abbreviations: CMC, critical micelle concentration; DC, dopachrome or 2-carboxy-2,3-dihydroindole-5,6-quinone; DCT, dopachrome tau- tomerase; DHI, 5,6-dihydroxyindole; DHICA, 5,6-dihydroxyindole-2- carboxylic acid; DQ, dopaquinone or 3,4_quinonephenylalanine; IQ, 5,6- indolequinone; IQCA, 5,6-indcdequinone-2-carboxylic acid; HEMT, high electrophoretic mobility tyrosinase; LEMT, low electrophoretic mobility tyrosinase; MBTH, 3-methyl-2-benzothiazolinone hydrazone; PMSF, phenylmethylsulfonylfluoride; ‘TRP, tyrosinase related protein

* Corresponding author. Fax: + 34 68 830950.

tions: the two initial steps of the pathway, the rate-limiting hydroxylation of L-tyrosine to L-dopa and the oxidation of this o-diphenol to L-dopaquinone, and also the oxidation of 5,6_dihydroxyindole (DHI) to 5,6-indolequinone (IQ) [3]. In addition to this multi-functional enzyme, the regula- tion of the pathway involves other proteins, whose exis- tence was first postulated at the beginning of the 1980s [4-61 and whose enzymatic function was characterized later 17-91. Owing to their structural similarity with tyrosi- nase, these proteins have been named tyrosinase related proteins (TRPs) [lO,ll]. Genetic mapping of mouse and human chromosomes has detected the existence in both species of related genes encoding two members of this family, TRPl and TRP2 [12-151.

Tyrosinase, TRPl and TRP2 are membrane-bound gly- coenzymes, containing copper and maybe other metal ions. They show a high homology in their amino acid sequences [14]. In the three cases, the bulk of the polypeptide chain is located inside the melanosome, with a transmembrane

0304-4165/95/$09.50 0 199.5 Elsevier Science B.V. All rights reserved SSDI 0304-4165(94)00169-3

422 C. Jimknez-Cervantes et al. /Biochimica et Biophysics Acta 1243 (1995) 421-430

hydrophobic fragment spanning the melanosomal mem- brane and a short COOH-terminal peptide directed towards the cytosolic side [lo]. According to this structure, these proteins belong to the type I integral membrane proteins [16]. Because of their extensive homology, the three mem- bers of the family have similar physical properties and their separation by conventional biochemical techniques is very difficult. However, taking advantage of the variability in the amino acid sequences of the carboxyl termini, which is the part where the proteins display the greatest differ- ences, Hearing et al. prepared specific antibodies directed against this region to achieve the immunopurification of these proteins [7,11,17]. On the other hand, some ap- proaches carried out by our group to separate the melanogenic enzymes from the melanosomal membrane by differential solubilization, chromatography and elec- trophoretic techniques, led to the characterization of three proteins, two different tyrosinases, termed HEMT and LEMT on the basis of their high and low electrophoretic mobility [18,19], and DCT [8].

The correspondence between the proteins characterized by immunopurification (tyrosinase, TRPl and TRP2) and the enzymes isolated by biochemical techniques has been recently studied. TRP2 has been identified as dopachrome tautomerase, the enzyme which catalyzes the transforma- tion of L-dopachrome (L-DC) into 5,6-dihydroxyindole-2- carboxylic acid (DHICA) [8,11]. However, the assignment of functions to TRPl has been more difficult, since TRPl also displays tyrosinase activity, although with specific activity lower than true tyrosinase [7,17]. This makes difficult the use of differential assays for both enzymes during the purification process. In addition, tyrosinase, TRPl and TRP2 seem to form multienzymic melanogenic complexes in vivo [20]. Nevertheless, the existence of some differential properties, regarding the resistance to proteolysis, deglycosylation, hormonal regulation and cat- alytic parameters, allowed us to propose the correspon- dence of HEMT and LEMT with tyrosinase and TRPl respectively [18,19].

The definitive assignment of HEMT and LEMT has been recently confirmed [21] by their immunochemical recognition with the specific antibodies antiPEP and an- tiPEP1 (directed against tyrosinase and TRPl respectively [7,17]). Such an approach showed some difficulties in the

Table 1 Properties of some detergents used for membrane protein solubilization

Detergent Type mM (1%)

immunorecognition of HEMT and LEMT depending on the nature of the detergent used for enzyme solubilization. Thus, we explored the possible effect of the detergent and melanosomal lipids on the solubilization, isolation, stabil- ity and immunoreactivity properties of each enzyme. Table 1 summarizes the structure and main properties of the detergents used throughout this work [22]. The data ob- tained account for the reasons why HEMT (tyrosinase) and LEMT (TRPl) have different chromatographic properties, hydrophobicities and can be resolved in SDS-PAGE after Brij-35 solubilization, whereas these same proteins cannot be adequately resolved after solubilization with other widely used non-ionic detergents. Moreover, they point out a possible functional role of the carboxyl terminus for the tyrosinase family proteins.

2. Materials and methods

2.1. Reagents

L-[3,5-3H]Tyrosine, specific activity 50 Ci/mmol, was obtained from New England Nuclear (Boston, MA). L-

Tyrosine, bovine serum albumin, hydroxyapatite type I suspension in 1 mM phosphate buffer (pH 6.81, L-dopa, Coomassie Brilliant blue, bromophenol blue, phenyl- methylsulfonyl fluoride, 3,3’-diaminobenzidine, EDTA, MBTH, SDS, Brij-35, sodium deoxycholate and Nonidet P-40 were from Sigma Chemical Co. (St. Louis, MO). Triton X-114 was from Fluka Chemie (Buchs, Switzer- land). Trichloroacetic acid, glycine, Tris, chloroform, ace- tone, methanol, hydrogen peroxide and monosodium and disodium phosphate were from Merck (Darmstadt, Ger- many). Acrylamide, bisacrylamide, TEMED (N,N,N’,N’- tetramethylethylenediamine), ammonium persulfate, 2- mercaptoethanol and Tween-20 were obtained from Bio- Rad (Richmond, CA). Peroxidase-labeled mouse anti-rab- bit IgG was from Chemicon (Temecula, CA). The bicin- choninic acid kit for protein determination was purchased from Pierce Europe (The Netherlands). All other reagents were obtained from Probus (Barcelona, Spain). All reagents were of the highest purity available, and all solutions were prepared in double-distilled water, further purified with a Milli-Q system from Millipore (final resistance around 5 Ma/cm>.

M, CMC (mM) Monomer/micelle

SDS ionic 35 288 3 60 DOC Na+ bile salt 24.1 414.6 1.6 High variability Brij-35 Cl2 E23 8.3 1198 0.09 40 Tween-20 Cl2 Sorb. E20 8.1 1241 0.06 ? Nonidet P-40 Tert- C8-Ph-E9 16.6 602 0.34 - 140 Triton X-114 Tert- C8-Ph-E7,8 18.6 537 - 0.4 - 1.50

Brij-35: Alkyl(12) polyoxyethylene (23) ether; Tween-20: Alkyl(12) polyoxyethylene (20) sorbitan ester; Nonidet P-40 and Triton X-114: Branched-alkyl (8) Phenyl polyoxyethylene ( < 10) ether.

C. Jimtnez-Cervantes et al. /Biochimica et Biophysics Acta 1243 (1995) 421-430 423

2.2. Preparation of solubilized melanosomal extracts from defined as the amount of enzyme that catalyzes the tau- melanoma cells tomerization of 1 pmol of L-DC/min at 37°C.

As a preliminary step in the purification of melanogenic enzymes, the melanosorne-rich fraction of B16 mouse melanomas was obtained by differential centrifugation of tumor homogenates. Freshly excised B16/FlO mouse melanomas were cleaned, washed twice in ice-cold 10 mM phosphate buffer (pH 6.8) 0.25 M sucrose and 0.1 mM EDTA, weighed and homogenized with a Polytron homog- enizer, in buffer supplemented with freshly prepared 0.5 mM PMSF, using a ratio of buffer volume to tumor mass of 2:l (ml/g). All these manipulations were carried out at 4°C. The homogenate was centrifuged at 700 X g, 10 min, at 5°C in a Sorvall SS-34, rotor. The pellet was reextracted with the same volume of buffer and recentrifuged. The melanosomal pellet was obtained by centrifugation of the combined supernatants at 11000 X g for 30 min. Unless otherwise indicated, the pellets were submitted to three consecutive cycles of extraction with the appropriate deter- gent at a final concentration of 1% in 10 mM phosphate buffer (pH 6.8). The solubilization cycles were performed using a volume of detergent solution equal to the starting tumor mass (ml/g). For each cycle, the melanosomal pellet was gently resuspended in the corresponding volume of detergent solution for 30 min with occasional shaking, and then centrifuged at 105 000 X g for 1 h in a Beckman 60 Ti rotor.

2.5. Purification of tyrosinase, TRPl and TRP2

2.3. Extraction of total lipids from the melanosomal frac- tion

Total endogenous lipids from the melanosome-rich frac- tion of B16 melanomas were prepared by the chloroform- methanol extraction according to Bligh and Dyer [23]. The amount of total lipids recovered was determined gravimet- rically and a concentrated stock solution (25 mg/ml) in chloroform/methanol 2:l was kept at - 80°C in a nitro- gen atmosphere, until needed.

These enzymes were purified from melanosomal mem- branes solubilized by Brij-35 treatment, as described else- where [8,18,21]. Tyrosinase was purified by bringing the crude melanosomal extract corresponding to the first solu- bilization cycle to 40% saturation with ammonium sulfate and centrifugation at 11000 X g for 30 min. The pellet was discarded and the supematant brought to 75% satura- tion with ammonium sulfate and left overnight with gentle stirring. The resulting pellet was resuspended in a volume of 10 mM phosphate buffer (pH 6.8), 0.1% Brij-35 equal to the starting tumor mass (ml/g). The sample was exten- sively dialyzed and ultracentrifuged at 105 000 X g for 1 h to remove non-soluble material that appeared during the dialysis. The supernatants were incubated 30 min in a Type I hydroxyapatite suspension using equal volumes of extract and settled hydroxyapatite. The slurry was cen- trifuged at 700 X g, 10 min and the pellet was washed twice with one volume of 10 mM phosphate buffer (pH 6.8), 0.1% Brij-35. The combined supematants were con- centrated and submitted to gel-filtration chromatography in a Sephacryl S-300 column (52 X 2.6 cm). The fractions with the highest dopa oxidase specific activity were pooled, concentrated and used as purified tyrosinase. The final preparations were purified around 18-fold, showed specific activities close to 2500 mU dopa oxidase/mg protein, and the yields were around 10%. Since both tyrosinase and TRPl show dopa oxidase activity, the purification factors calculated for these enzymes were not very high because of the separation of both enzymes through the purification process.

2.4. Determination tyros,inase and DCT activities

Tyrosine hydroxylase activity measurements were car- ried out by the method of Pomerantz [24] modified as described elsewhere [25]. Dopa oxidase activity determina- tions were carried out as described by Winder and Harris [26], with minor modifiications. One unit was defined as the amount of enzyme catalyzing the oxidation of 1 pmol dopa/min at 37°C. DCT activity was routinely determined by measuring the decrease in absorbance at 475 mm of a 0.1 mM L-DC solution in 10 mM phosphate buffer (pH 6.0) (E = 3700 M-’ cm- ‘). For the most accurate deter- minations, DCT was assayed by HPLC determination of DHICA formation from L-DC, according to Tsukamoto et al. [ll]. L-DC was prepared in all cases by stoichiometric oxidation of r_-dopa with sodium periodate. One unit was

TRPl was purified from the extracts obtained in the second and third solubilization cycles. This pool, which contains a lower protein content than the first extract [18], was applied to a DEAE-Sephadex column (25 X 3 cm) equilibrated in 10 mM phosphate buffer (pH 6.8), contain- ing 0.1% Brij-35, and allowed to interact with the gel for 30 min. The column was then washed with one bed volume of equilibration buffer, and eluted with a linear gradient of NaCl, from 0 to 0.5 M, in a total volume of 500 ml of the same buffer. Fractions with the highest dopa oxidase activity were pooled and extensively dialyzed against 10 mM phosphate buffer (pH 6.8), 0.1% Brij-35. Then, the solution was concentrated and submitted to gel-filtration chromatography in Sephacryl S-300. The fractions with the highest dopa oxidase specific activity were pooled, concentrated and used as purified TRPl. Final preparations showed a purification factor around 7-fold, showed a specific activity above around 200 mU dopa oxidase/mg protein, and yields around 8%.

The purification of DCT was carried out according to Aroca et al. [8]. The procedure is also initiated with the

424 C. Jim&ez-Ceroantes et al. /Biochimica et Biophysics Acta 1243 (1995) 421-430

solubilization of melanosomes with 1% Brij-35. The solu- bilized preparations were purified by ammonium sulfate precipitation and gel filtration chromatography on Sephacryl S-300. The fractions with the highest DCI and low dopa oxidase activity were pooled, concentrated and used as purified TRP2. The final preparation was purified 55fold, displayed a specific activity of 378 mU/mg pro- tein and a yield of 21%.

600 I _-

??D.0.

??DCT

i CPr0t

400 -I

f 3 !z

2.6. Electrophoretic and Western blotting procedures

Analytical SDS-PAGE was performed according to Laemmli [27] but without 2-mercaptoethanol and heating to preserve enzymatic activities, Electrophoresis was run on 12% polyacrylamide gels at 4°C and at a constant current of 25 mA in a Bio-Rad MiniProtean cell. A highly sensitive stain for dopa oxidase activity was carried out by incubation at 37°C in a sodium phosphate buffered solu- tion of 2 mM L-dopa supplemented with 4 mM MBTH, as described by JimCnez-Cervantes et al. [28]. Densitometric analysis of the gels was performed in an Ultroscan laser densitometer from Pharmacia (Uppsala, Sweden).

0 1 St 2nd 3rd 1st 2nd 3rd

r;, 0 2nd 3rd 1st 2nd 3rd 1st 2nd 3rd

Brij 35 Tween 20 Nonidet P40 Triton X-l 14 DOC

Fig. 1. Dopa oxidase, DCT and total protein solubilized from melanoso- ma1 membranes by 1% detergents in three consecutive extraction cycles, labeled as first, second and third.

For immunoblotting experiments, samples were treated at 80°C for 30 min with sample buffer (0.18 M Tris-HCl (pH 6.81, 15% glycerol, 0.075% bromophenol blue, 9% SDS and 7.5% P-mercaptoethanol), using a 2:l ratio (v/v> of sample to buffer. Western blotting was carried out using nitrocellulose filters from Bio-Rad. After SDS-PAGE of solubilized samples, the gels were equilibrated in transfer buffer (48 mM Tris, 39 mM glycine, 1.3 mM SDS, 20% methanol) for 10 min, and transferred in a Bio-Rad semidry transfer cell. The filters were blocked by gently shaking for 1 h in 2% bovine serum albumin in phosphate saline buffer. The blocked membranes were washed, incubated overnight with 1:2000 dilutions of oPEP1, aPEP or control non-immune rabbit serum, and then with peroxi- dase-labeled mouse antirabbit IgG as described elsewhere [21]. Finally, the filters were washed five times and incu- bated in a solution of 3,3’-diaminobenzidine (0.6 mg/ml) and 0.001% H,O, in the presence of 0.03% CoCl, until the bands were evident.

2.7. Protein determination

Protein concentrations were determined by the bicin- choninic acid test developed by Pierce Co. to avoid the interference of detergents on other currently used assays. Protein content was extrapolated from a standard curve prepared with bovine serum albumin standard solution (1 mg/mO.

3. Results

brane was first analysed. Three solubilization rounds were completed for each detergent at a fixed concentration of 1%. A further fourth round did not extract significant activity in any case (data not shown). A markedly different pattern among the detergents was observed when the amounts of activities extracted in each solubilization round were analysed (Fig. 1). According to the obtained results, two groups of detergents can be defined. The first one, comprising those with molecular mass greater than 1000 Da, and ‘long’ polyoxyethylene chains (Brij-35 and Tween-201, extracted the melanogenic enzymes gradually. The amount of enzymatic activities obtained in the various solubilization rounds were similar, and the third extraction consistently yielded high levels of activity, comparable to those found in the previous rounds. As the amount of protein decreased in the second and third cycles, the specific activities obtained in these rounds were higher than for the first one (results not shown). The second group is comprised of detergents with lower molecular mass, less than 500 Da, having short polyoxyethylene chains or a steroid structure (Nonidet P-40, Triton X-114 and sodium deoxycholate). These detergents were very efficient in the first extraction: around 80% of the total melanogenic activities were solubilized in this round. Moreover, almost all the activities were extracted in the first two cycles. Due to the higher amount of protein solubilized, the specific melanogenic activities were lower in comparison to second and third extracts with Brij-35 and Tween-20. Concerning the total enzymatic activities extracted in the process after the three solubilization rounds, estimated as the sum of the activities obtained in the three extraction cycles, the differences observed were never greater than 25%. Brij-35 was the detergent which ex- tracted the least activity.

The efficiency of several non-ionic detergents to extract The gradual solubilizing action of detergents with a the melanogenic enzymes from the melanosomal mem- long polyoxyethylene chain was apparently related to their

C. Jimknez-Cervantes et al. / Biochimica et Biophysics Acta 1243 (I 995) 421-430 425

??D.O. k?dDCT ??Prot I

1st ma,

1% BRIJ

1 19

2% BRIJ 1% NP40 1% NP40+ 0.1% SDS

Fig. 2. Dopa oxidase, DCI and protein solubilized from melanosomes. Only the results obtained in the first cycle and the total amount (calcu- lated as the sum of the three cyc.les) are shown. Left: comparison behveen 1% and 2% Brij-35. Right: supplementation of 1% Nonidet P-40 with 0.1% SDS.

structure and did not depend on the concentration used. Fig. 2 shows the compar,ative extraction achieved by 2% versus 1% Brij-35 in the first cycle and the total activity after three cycles. The activities obtained were very simi- lar. On the other hand, supplementation of 1% Nonidet P-40 with 0.1% SDS did not increase the total amount of activity extracted (Fig. 2, right). The amount of solubilized protein was increased, thus causing a decrease of the specific activities. In turn, the stability of all the melanogenic enzymes was markedly decreased by 2% Brij-35 or addition of 0.1% SDS in comparison to 1% detergent (results not shown).

The ratio between tyrosine hydroxylase and dopa oxi- dase activity obtained after solubilization also differed depending on the detergent employed. Fig. 3 shows the activities extracted in the first round. Again the behaviour

103 I15

??T.H. ??0.0.

80’ E3Ratio

z r

Sri, 35 Twee” 20 hb”idet P40 Trite” xt 14 ooc

Fig. 3. Tyrosine hydroxylase anli dopa oxidase activities (left) and ratio between both activities (right). These activities were obtained in the first cycle. Data are expressed as percentages with respect to the total activity obtained after the three extraction cycles.

N P40 Brii 35 1st 2nd 1st 2nd

Fig. 4. SDS-PAGE demonstration of the separation of two tyrosinase forms from melanosomal pellets. The gel was stained for the dopa oxidase activity with L-dopa plus MBTH, according to Ref. [28]. A crude melanosomal pellet was solubilized using a detergent/protein ratio of 1:0.75 with Brij-35 and Nonidet P-40. After centrifugation, the pellets were reextracted. The solubilized samples were diluted 1:4 with 10 mM phosphate buffer containing 1% of the appropriate detergent.

of detergents with a long polyoxyethylene chain was dif- ferent from the other group. The activity obtained by the first group was lower, and more importantly, the ratio of percentages (referred to total activity obtained after the three consecutive treatments) of tyrosine hydroxylase and dopa oxidase extracted was also lower than one, whereas the ratios obtained after treatment with Nonidet P-40 or Triton X-114 were higher than one. Thus, these latter detergents extracted preferentially tyrosine hydroxylase ac- tivity, while Brij-35 and Tween-20 extracted preferentially dopa oxidase activity. Since tyrosinase and TRPl display different ratios of tyrosine hydroxylase to dopa oxidase specific activities [l&21], this effect should be related to the different interaction of tyrosinase and TRPl with the melanosomal matrix, or with the detergents used in this study, that would result in a differential solubilization of both enzymes.

The comparison by SDS-PAGE of samples solubilized by Brij-35 and Nonidet P-40 (Fig. 4) showed that the latter detergent solubilized more dopa oxidase in the first extrac- tion, but the resolution of tyrosinase and TRPl was not achieved. However, both enzymes were well resolved in the solubization using Brij-35. Since tyrosinase and TRPl should have very similar molecular mases, as deduced from their sequences and shown by others [7], the rather different electrophoretic mobility observed in the Brij- solubilized samples might be related to the formation of detergent-protein complexes of different stoichiometry for each enzyme, or to the preservation of different conforma- tions in spite of the presence of high concentrations of SDS in the electrophoresis sample buffer. In any case, this observation points to the possibility that the detergent used in the solubilization of the melanosomal membrane might interact strongly with the tyrosinase isoenzymes, so that the excess SDS used in the electrophoresis sample buffer

426 C. Jinknez-Cervantes et al. /Biochimica et Biophysics Acta 1243 (1995) 421-430

might not be able to displace totally the non-ionic deter- gent from the enzymes.

Further evidence for a strong binding of the non-ionic detergents to tyrosinase and TRPl with an effect on the three- dimensional structure of the proteins was obtained by immunoblotting experiments (Fig. 5). Tyrosinase and TRPl solubilized with Nonidet P-40 have been shown to immunoreact strongly with (Y PEP7 (anti-tyrosinase) and aPEP (anti-TRPl). This strong binding has been ex- ploited for the immunoaffinity purification of both proteins [7,17]. However, as can be seen in Fig. 5A, aPEP recognized purified TRPl in Western blots, while aPEP yielded a poor to undetectable staining of purified tyrosi- nase, under identical experimental conditions. Conversely, the reactivity of aPEP against tyrosinase was evident after trichloroacetic precipitation of the purified enzyme, followed by extensive washing of the pellet with acetone

1 A

97.4_

66.2_

55.0_

42.7_

B

97.4 -

66.2_

55.0-

42.7_

2

3

Fig. 5. Immunochemical recognition of purified tyrosinase and TRPl (8 pg per lane). (A) Immunoreactivity of Brij-35 solubilized and purified TRPl probed with aPEP (lane l), and tyrosinase probed with aPEP (lane 2). (B) Immunoreactivity of tyrosinase with aPEP after removal of the Brij-35. Lane 1: tyrosinase (0.15 wg); lane 2: a mixture of tyrosinase and TRPl (0.15 pg each); lane 3: the same mixture of tyrosinase and TRPI probed with nonimmune rabbit serum.

ExtractloniDotergent

Fig. 6. Stability of the dopa oxidase and DCI activities after extraction with different detergents. Results expressed as residual activity after 1 week at 4°C.

and solubilization at 90°C with 3% SDS and 2-mercapto- ethanol (Fig. 5B), a treatment which should remove most if not all the non-ionic detergent present in the original sample. These data suggest that the carboxyl termini of the proteins, which are the epitopes for the antibodies [7,10,17], are accessible in samples obtained with Nonidet P-40, but they are hindered when the solubilization is carried out with Brij-35.

The stability of the enzymatic melanogenic activities solubilized by different detergents was also studied. Fig. 6 shows the percentage of residual dopa oxidase and DCT activities in samples kept at 4°C for 1 week. Dopa oxidase was more stable than DCT in all detergents. Samples obtained with Tween-20 showed maximal stability. The stability of the tyrosine hydroxylase was parallel to dopa oxidase (not shown). The third extraction was less stable than the first and second ones, probably due to the exten- sive delipidation caused after the two solubilization rounds. This suggests that the melanosomal lipids could affect melanogenic activities.

Protein-lipid interactions were studied from a double point of view. First, Fig. 7 shows the immediate effect of re-addition of melanosomal lipids on the dopa oxidase activity of the third extract as well as on purified prepara- tions of tyrosinase and TRPl. The lipids were extracted from a melanosomal pellet, and their re-addition yielded a concentration-dependent stimulation of the dopa oxidase activity, which was maximal at 0.25 mg/ml of lipids. This effect was very pronounced on TRPl, whose activity was activated twofold. The stimulation of the third extraction was never greater than 140%. However, no statistically significant effects were observed when the first extract was tested. Furthermore, the stimulatory effect of the melanosomal lipids was specific for tyrosinase and TRPl, since addition of lipids caused around 20% inhibition on the tautomerase activity of TRP2.

C. Jimknez-Ceruantes et al. /Biochimica et Biophysics Acta 1243 (1995) 421-430 427

+Tyrosinase

+TRPl

*3rd Ext.

+lst Ext.

-TRPB

-

..~.~~Z~.:~:.-.--.r.~~.~.-.~.-~.~-

E

0.5 1 1.5 2

Lipids (mglml)

Fig. 7. Effect of melanosomal lipids on the DCf activity of TRP2 and on the dopa oxidase activity of purified tyrosinase, purified TRPl, first and third crude extracts. Appropriate amounts of the stock lipid solution were pipetted in a glass test tube and dried under nitrogen. Then, the appropri- ate samples were added, gently shaken for 2 min, and the activity assayed. Control activity (100%) was assayed in the same way but without lipids. The protein contents were 0.12, 0.27, 2.1, 0.52 and 0.96 mg/ml for TRPl, tyrosinase, first extract, third extract and TRP2 respec- tively.

Second, the stabilizing effect of the melanosomal lipids on the melanogenic activities in the third extract solubi- lized by different detergents was also studied. In this regard, all detergents showed a similar pattern, so that only the data for Brij-35 are shown (Fig. 8). The percent of dopa oxidase and DCT ac:tivities remaining in samples left at 4°C for 8 days clearly indicate that the supplementation of the preparations with 0.5 mg/ml of lipids protected the melanogenic enzymes against inactivation.

CONTROL

D.O. Fig. 8. Stabilization of dopa oxidase and DCT activities by addition of 0.25 mg/ml melanosomal lipids to a fraction delipidated (third extract with 1% Brij-35). Results expressed as residual activity after 8 days at 4°C.

4. Discussion

From the results expressed in Fig. 1, it is clear that non-ionic detergents containing a long polyoxyethylene chain (e.g., Brij-35 and Tween-20) achieve a gradual solubilization of proteins from the melanosomal mem- brane. Re-extraction of the residual pellet yields similar levels of dopa oxidase and DCT activities to those ob- tained in the first solubilization cycle. However, non-ionic detergents with a shorter polyoxyethylene chain (e.g., Non- idet P-40, Triton X-114) achieve a more efficient solubi- lization. Most of the activity is extracted in the first treatment of the melanosomal membranes. It could be thought that the gradual solubilizing effect of long-tail detergents was due to the lower molarity of the corre- sponding 1% solutions (see Table l>, but Fig. 2 proves that this was not the case, since 2% Brij-35 did not improve the solubilization got by a 1% solution of the same detergent. Alternatively, the lower solubilizing ability of Brij-35 and Tween-20 could be related to the low CMC of these detergents in comparison to short-tail detergents. The oc- currence of specific interactions of the detergents with the melanogenic proteins might also contribute to the differ- ences observed in the solubilizing efficiency of the deter- gents.

Several observations point to the occurrence of such specific interactions. First, the ratio of tyrosine hydroxyl- ase to dopa oxidase activities solubilized by the two types of detergents were different. Brij-35 and Tween-20 extract preferentially dopa oxidase activity, while short-tail deter- gents preferentially extract tyrosine hydroxylase activity. This suggests a differential solubilization of tyrosinase and TRPl, since we have shown that the two enzymes display different ratios of tyrosine hydroxylase to dopa oxidase activity [18,21]. Second, the electrophoretic resolution of tyrosinase and TRPl solubilized by Nonidet P-40 was very poor, but the resolution of these enzymes in Brij-35 solubi- lized samples was much better. The reasons to account for these differences are not well understood, but they are surely related to the interactions of detergents with tyrosi- nase and TRPl. These interactions should mainly occur through the hydrophobic transmembrane region of the proteins and they seem to be very strong, due to the different electrophoretic behaviour of the solubilized sam- ples even in the presence of 3% SDS. This indicates that SDS is not able to displace the non-ionic detergent. Third, the conformation of the carboxyl terminus of tyrosinase in Brij and Nonidet solution appears to be different, as judged by the lack of reaction of Brij-solubilized tyrosinase with cr PEP7, an antibody recognizing the last 14 amino acids of the protein. This antibody reacts readily with Nonidet P-40 solubilized tyrosinase, or with the Brij-solubilized enzyme after a treatment that removes most of the Brij-35. Al- though the length of the detergent might contribute to the steric hindrance of the epitope, the lack of reactivity of Brij-35 solubilized tyrosinase cannot be totally explained

428 C. Jiminner-Ceruantes et al. /Biochimica et Biophysics Acta I243 (1995) 421-430

in these terms, since aPEP1, directed against the last 14 amino acids of TRPl, readily reacts with the Brij-35 solubilized protein. The different degree of accessibility of the carboxyl terminus of tyrosinase solubilized with Brij-35 and Nonidet P-40 was still evident after SDS-PAGE.

The differences among the melanogenic proteins de- pending on the detergent used for their solubilization accounts for their different chromatographic and elec- trophoretic behaviour, in spite of their sequence amino acids homology. In turn, together with the effect of lipids on the activity and stability of these enzymes, a possible functional role of the transmembrane fragment and car- boxy1 terminal tail of these proteins is suggested. This role might have been overlooked, so far, due to several factors. Thus, genetic studies have revealed that mutations leading to tyrosinase negative albinism cluster in the amino termi- nal side of the protein and in its central core whereas, to date, no mutation leading to enzyme inactivation has been mapped to the carboxyl terminus, beyond the transmem- brane fragment [29].

Moreover, melanosomal tyrosinase can be conveniently solubilized by mild trypsin digestion to yield a fully active soluble form [30,31]. This treatment is thought to release the bulk of the molecule by cleaving the transmembrane fragment, thus removing the hydrophobic domain and the carboxyl cytosolic tail [32]. It is therefore clear that the carboxyl terminus of tyrosinase, and possibly the other related proteins, is not essential for enzymatic activity. However, our data also show that interaction of this do- main with hydrophobic molecules can modulate the activ- ity of the enzymes. The stimulation produced by lipids on the dopa oxidase activity in contrast to the inhibition produced on DCT supports differences in the interactions of the melanogenic enzymes with hydrophobic molecules and their importance in the modulation of the biological activities.

Concerning the stability of the solubilized melanogenic enzymes, it seems to depend on the interaction of these proteins with the melanosomal lipids, and the likely dis- placement of these lipids by detergents when solubilizing the melanosomal membrane. The amount of detergent needed to obtain a given effect on lipid bilayers depends on several factors, such as the CMC, micelle size, nature of the membrane and the detergent [22], but it is generally accepted that a ratio of detergent to lipid (w/w) within the range 0.1 to 1 causes a selective extraction of peripheral proteins, the membrane bilayer remaining essentially in- tact. When the ratio is around 2, solubilization of the membrane occurs, resulting in the formation of lipid/pro- tein/detergent, protein/detergent and lipid/detergent mi- celles. Finally, ratios around 10 lead to an extensive delipidation around the proteins. Extrapolation to our ex- perimental conditions indicates a detergent:lipid ratio around 2 in the first extract, but around 20 in the third one. Thus, the first extracts should contain mixed micelles, but the preparations coming from the third extraction and also

Fig. 9. Model for the interaction between the hydrophobic transmembrane domain of tyrosinase and TRPl with Nonidet P-40 and Brij-35. The protein model was proposed by Hearing and Tsukamoto [IO], and the length of the polyoxyethylene chains of the detergent has been drawn at the same scale to ilustrate the plausible hindrance of the carboxyl terminus due to Brij-35.

C. Jim&ez-Ceruantes et al. / Biochimica et Biophysics Acta 1243 (1995) 421-430 429

the purified enzymes should be mostly delipidated. Supple- membrane proteins [37]. Thus, melanosomal tyrosinase mentation with melanosomal lipids prevented inactivation family proteins could be a model to approach the study of of the latter fractions. other types of membrane bound proteins.

The finding of different conformations and elec- trophoretic mobilities in the melanogenic enzymes solubi- lized with different detergents accounts for some of the conflicting data about the physicochemical and immuno- chemical properties of the enzymes. Although the molecu- lar basis of the detergent effects is not known, it might be related to the detergent size. The length of the polx- ethylene chains in non-ionic detergents ranges from 13 A for 8 ethylene units (close to Nonidet P-40) to 22 A for chains of 20 units (simiLr to Brij-35) [22]. The estimated lfngths for Nonidet P-40 and Brij-35 are around 17 and 35 A, while the length for a. transmembrane a-helix domain of around 20 amino acids is 30 A. The models drawn in Fig. 9 for the interaction of tyrosinase and TRPl with Nonidet P-40 and Brij-35; are compatible with the experi- mental pattern of these proteins and/or detergents. The length of Brij-35 could reduce its solubilizing efficiency, but increase its action on the melanosomal proteins. In this sense, Brij-35 could shield more efficiently than Nonidet P-40 the hydrophobic regions, and the neighbouring car- boxy1 tails of tyrosinase and TRPl. This steric hindrance could affect: (a> the interaction between tyrosinase and TRPl to form heterodimers in such way that the elec- trophoretic resolution of both proteins would be possible with Brij-35, but it is not with Nonidet P-40; (b) the immunorecognition of the carboxyl tails by the specific antibodies aPEP (anti-TRPl) and (Y PEP7 (anti- tyrosinase), so that the irnmunoreactions in Western blots are stronger in preparations obtained with Nonidet P-40. The observed differences between tyrosinase and TRPl should be related to the divergence in the sequence of the carboxyl tails. First, the membrane-emerging stretch of six consecutive basic amino acids in tyrosinase differs from the alternation of hydroxy and basic amino acids in the equivalent position of TIRPl. Second, mouse TRPl con- tains a putative N-glycosylation NHS sequence very close to the carboxyl end. Alth.ough it is very unlikely that this site will be glycosylated, it was reported that mutant TRPl not fully glycosylated was not appropriately delivered to melanosomes [33]. Related to this, it has been recently reported that a short sequence of the carboxyl terminus is required for the sorting of melanosomal membrane pro- teins [34].

Acknowledgements

We deeply thank Dr. V.J. Hearing for the aPEP1, and aPEP antibodies and for all his suggestions, including permission to use the tyrosinase family model. This work has been supported by Grant number 94/0787 from the Fondo de Investigaciones Sanitarias. C.J.-C. was recipient of a fellowship from the Instituto de Foment0 de la Comunidad Aut6noma de Murcia.

References

[ll L21

[31 141

[51

[61

[71

[81

[91

[lOI Ml

WI

1131

[141

Hill, Z. (1992) Bioessays, 14, 49- - 56. Prota, G. (1992). in Melanin and Melanogenesis (Jovanovich, H.B., ed.), Academic Press, London. Komer, A. and Pawelek, J. (1982) Science 217, 1163-- 1165. Pawelek, J., Korner, A.M., Bergstrom, A. and Bologna, J. (1980) Nature 286, 617- - 619. Hearing, V.J., Korner, A.M. and Pawelek, J.M. (1982) J. Invest. Dermatol. 79, 16-18. Barber, J.I., Townsend, D., Olds, D.P. and King, R.A. (1984) J. Invest. Dermatol. 83, 145-149. Jimtnez, M., Tsukamoto, K. and Hearing, V.J. (1991) J. Biol. Chem. 266, 1147-1156. Aroca, P., Garcia-Borr6n, J.C., Solano, F. and Lozano, J.A. (1990) Biochim. Biophys. Acta 1035, 266-275. Pawelek, J. (1990) Biochem. Biophys. Res. Commun. 166, 1328- 1333. Hearing, V.J. and Tsukamoto, K. (1991) FASEB J. 5, 2902-2909. Tsukamoto, K., Jackson, I.J., Urabe, K., Montague, P.M. and Hear- ing, V.J. (19921, EMBO J. 11, 519-526. Shibahara, S., Tomita, Y., Sakakura, T., Nager, C., Chaudhuri, B. and Miiller, R.M. (1986) Nucleic Acids Res. 14, 2413-2427. Vijayasaradhi, S., Bouchard, B. and Houghton, A. (1990) J. Exp. Med. 171, 1375-1380. Jackson, I.J., Chambers, D.M., Tsukamoto, K., Copeland, N.G., Gilbert, D.J., Jenkins, N.A. and Hearing, V.J. (1992) EMBO J. 11, 527-535.

These data illustrate the importance of the carboxyl terminus and its interaction with hydrophobic molecules on the physicochemical properties of the tyrosinase family proteins. Moreover, they show the need for a careful choice of the detergent used for solubilization, depending on the type of study to be carried out. Similar results have been recently reported for other membrane proteins, such as cytochrome c oxidase [35] and Ca-ATPase [36]. In turn, differential solubility in non-ionic detergents has been used to distinguish between slowly and rapidly degraded plasma

[151

[161 [171

[I81

1191

DO1

1211

[221

b31

Bouchard, B., Del Marmol, V., Jackson, I.J., Cherif, D. and Du- bertret, L. (1994) Eur. J. Biochem. 219, 127-134. Von Heijne, G. and Gavel, Y. (1988) Eur. J. Biochem. 174,671-678. JimBnez, M., Maloy, W.L. and Hearing, V.J. (1989) J. Biol. Chem. 264,3397-3403. Jimbnez-Cervantes, C., Garcia-Box&, J.C., Valverde, P., Solano, F. and Lozano, J.A. (1993) Eur. J. Biochem. 217, 549-556. Valverde, P., Garcia-Borr6n, J.C., JimBnez-Cervantes, C., Solano, F. and Lozano, J.A. (1993) Eur. J. Biochem. 217, 541-548. Orlow, S.J., Zhou, B., Chakraborty, A.K., Drucker, M., Pifko-Hirst, S. and Pawelek, J. (1994) J. Invest. Dermatol. 103, 196-201. JimCnez-Cervantes, C., Solano, F., Kobayashi, T., Urabe, K., Hear- ing, V.J., Lozano, J.A. and Garcia-Borr6n, J.C. (1994) J. Biol. Chem. 269, 17993-18001. Helenius, H., McCaslin, D.R., Fries, E. and Tanford, C. (1979) in Methods in Enzymology (Fleischer, S. and Packer, L., eds.), Vol. 56, pp. 734-749, Academic Press, NY. Bligh, E.G. and Dyer, W.J. (1959) Can. J. Biochem. Physiol. 37, 911-917.

430 C. Jitntnez-Cervantes et al. /Biochimica et Biophysics Acta 1243 (1995) 421-430

[24] Pomerantz, S. (1964) Biochem. Biophys. Res. Commun. 16, 188- 194.

[25] Jara, J.R., Solano, F. and Lozano, J.A. (1988) Pigment Cell Res. 1, 332-339.

[26] Winder, A.J. and Harris, H. (1991) Eur. J. Biochem. 198, 317-326. [27] Laemmli, U.K. (1970) Nature 277, 680-682. [28] JimCnez-Cervantes, C., Valverde, P., Garcia-Borrbn, J.C., Solano, F.

and Lozano, J.A. (1993) Pigment Cell Res. 6, 394-399. [29] Tripathi, R., Strunk, K.M., Giebel, L.B., Weleber, R.G. and Spritz,

R.A. (1992) Am. J. Med. Gen. 43, 865-871. [30] Nishioka, K. (1978) Eur. J. Biochem. 85, 137-146.

[31] Hearing, V.J., Nicholson, J.M., Montague, P., Ekel, T.M. and Tomecki, K. (1978) B&him. Biophys. Acta 522, 327-339.

[32] Valverde, P., Garcia-Borr6n, J.C., J.H., Solano, F. and Lozano, J.A. (1992) Arch. Biochem. Biophys. 297, 221-227.

[33] Orlow, S.J., Boissy, R.E., Moran, D.J. and Pifko-Hirst, S. (1993) J. Invest. Dermatol. 100, 55-64.

[34] Vijayasaradhi, S. and Xu, Y. (1993) Pigment Cell Res. 6, 276. [35] Frei, T.G. and Murray, J.M. (1994) J. Mol. Biol. 237, 275-297. [36] Cornea, R.L. and Thomas, D.D. (1994) Biochemistry 33,2912-2920. [37] Hare, J.F. and Holocher, A. (19941 J. Biol. Chem. 269, 5981-5988.