HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 278, Issue 18, 15735-15743, May 2, 2003

This Article Abstract Full Text (PDF) All Versions of this Article: 278/18/15735    most recent M300658200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Olivares, C. Articles by García-Borrón, J. C. Search for Related Content PubMed PubMed Citation Articles by Olivares, C. Articles by García-Borrón, J. C. Social Bookmarking                      What's this?

Conformation-dependent Post-translational Glycosylation of Tyrosinase

REQUIREMENT OF A SPECIFIC INTERACTION INVOLVING THE CuB METAL BINDING SITE^*

Concepcion Olivares, Francisco Solano, and Jose C. García-Borrón^§

From the Department of Biochemistry and Molecular Biology, School of Medicine, University of Murcia, Apto 4021, Campus Espinardo, Murcia 30100, Spain

Received for publication, January 21, 2003, and in revised form, February 19, 2003

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Tyrosinase, the rate-limiting enzyme in mammalian melanogenesis, is a copper-containing transmembrane glycoprotein. Tyrosinase undergoes a complex post-translational processing before reaching the melanosomal membrane. This processing involves N-glycosylation in several sites, including one located in the CuB copper binding site, movement from the endoplasmic reticulum (ER) to the Golgi, copper binding, and sorting to the melanosome. Aberrant processing is causally related to the depigmented phenotype of human melanomas. Moreover, some forms of albinism and several other pigmentary syndromes are considered ER retention diseases or trafficking defects. A critical step in tyrosinase maturation is the acquisition of an ER export-competent conformation recognized positively by the ER quality control system. However, the minimal structural requirements allowing exit from the ER to the Golgi have not yet been identified for tyrosinase or other melanosomal proteins. We addressed this question by analyzing the enzymatic activity and glycosylation pattern of mouse tyrosinase point mutants and chimeric constructs, where selected portions of tyrosinase were replaced by the homologous fragments of the highly similar tyrosinase-related protein 1. We show that a completely inactive tyrosinase point mutant lacking a critical histidine residue involved in copper binding is nevertheless able to exit from the ER and undergo further processing. Moreover, we demonstrate that tyrosinase displays at least two sites whose glycosylation is post-translational and most likely conformation- dependent and that a highly specific interaction involving the CuB site is essential not only for correct glycosylation but also for exit from the ER and enzymatic activity.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Melanogenesis is the biochemical pathway responsible for melanin synthesis. In mammals, three related and highly similar metalloenzymes, tyrosinase (Tyr¹²; monophenol dihydroxyphenylalanine:oxygen oxidoreductase; EC 1.14.18.1), and the tyrosinase-related proteins (Tyrps) 1 and 2 are involved in the catalytic control of the process. Their cDNAs have been cloned and sequenced (1-6), and the genes have been mapped to the mouse albino, brown, and slaty loci, respectively. The amino acid L-tyrosine is the metabolic precursor of the pigment. In the presence of catalytic amounts of L-3,4-dihydroxyphenylalanine (DOPA), Tyr catalyzes L-tyrosine conversion into L-dopaquinone (7). This reactive intermediate undergoes spontaneous cyclization and rearrangement to L-dopachrome in the absence of thiol compounds (8). Tyrp2 (dopachrome ²,⁷-isomerase; EC 5.3.3.12), also called dopachrome tautomerase, catalyzes the tautomerization of dopachrome into the more stable intermediate 5,6-dihydroxyindole-2-carboxylic acid (9, 10). Although the enzymatic function(s) of Tyrp1 is still somewhat controversial, the protein purified from mouse melanocytes has been reported to be a low specific activity Tyr isozyme with both tyrosine hydroxylase and DOPA oxidase activities (11, 12), that catalyzes the oxidation of 5,6-dihydroxyindole-2-carboxylic acid (13, 14) and promotes its incorporation into the eumelanin polymer. Sequence similarity between Tyr and the Tyrps is higher in the metal ion binding sites (Fig. 1), but the proteins also share several conserved N-glycosylation sites (reviewed in Ref. 15).

Tyr and the Tyrps are transmembrane glycoproteins whose final destination is the melanosome, a melanocyte-specific organelle where melanin synthesis is confined. The proteins of the family undergo a complex post-translational processing before reaching the melanosomal membrane in their final catalytically active conformational state. In the mouse, Tyr processing includes N-glycosylation in at least four of the six available glycosylation sites (16) (Fig. 1). Human TYR, displaying seven potential glycosylation sites, can appear as a protein with six or seven glycans, depending on the translation rate (17). During its post-translational processing, the mammalian enzyme also undergoes movement from the ER to the Golgi apparatus, binding of the copper cofactor to two sites designated CuA and CuB (18) and finally sorting to the melanosomes.

The importance of Tyr post-translational processing is highlighted by several observations. Aberrant processing is causally related to the depigmented phenotype of human melanomas (19). On the other hand, oculocutaneous albinism type 1 is an autosomal recessive disease characterized by the absence of pigment in hair, skin, and eyes, with other common features such as severe nystagmus, photophobia, and reduced visual acuity, associated with mutations in the Tyr gene (reviewed in Ref. 20). Some forms of oculocutaneous albinism type 1 are considered ER retention diseases (21-23), and several other pigmentary syndromes are accounted for by processing/trafficking defects (24, 25). In this respect, it has recently been shown that the most common form of albinism worldwide, oculocutaneous albinism type 2, results from mutations in the pink-eyed dilution (p) gene (26) and that the p protein contributes to the correct processing of Tyr and to its traffic to the melanosome (25). Therefore, correct processing and intracellular trafficking of Tyr and also probably of Tyrps is critical to normal pigmentation.

According to current evidence, recently synthesized Tyr is retained by the ER quality control machinery until the proper conformation is acquired. Then glycosylated and correctly folded Tyr is exported to the Golgi, where N-linked oligosaccharide chains are further processed and copper is probably bound. One of the potential glycosylation sites, located in the CuB site, appears particularly relevant. It is conserved in human and mouse Tyr and Tyrps (Fig. 1), and its glycosylation has been reported to correlate with proper cofactor binding and full enzymatic activity (16). Moreover, a natural mutation destroying this glycosylation sequon results in oculocutaneous albinism type 1 (27).

View larger version (26K): [in this window] [in a new window]   Fig. 1.   Schematic representation of tyrosinase structure, highlighting the similarity of the metal ion binding sites in mouse Tyr and Tyrps. The dots indicate cysteine residues, and the broken arrows pointing upward indicate the positions of the N-glycosylation sequons. Other relevant structural elements are indicated as appropriate. The metal binding sites are labeled as MeA and MeB, rather than CuA and CuB, because the nature of the metal cofactor is different in Tyr (copper) and Tyrp2/Dct (zinc), whereas the metal cofactor of Tyrp1 has not yet been convincingly characterized. In the amino acid sequence of the metal binding sites, histidine residues thought to be responsible for metal binding are underlined. Within the MeB site, the conserved N-glycosylation sequon is shown in italic and underlined characters.

Therefore, a critical step in pigmentation is the acquisition by tyrosinase of a conformation recognized positively by the ER quality control system. However, the minimal structural requirements allowing for exit from the ER have not yet been identified for Tyr or for other melanosomal proteins. This knowledge would help understand the molecular basis of albinism. Moreover, due to the likely relationships of different post-translational processes such as glycosylation, metal cofactor binding and movement between intracellular compartments, a description of these requirements may also provide a model for other metalloproteins that follow the secretory sorting pathway.

We have addressed this question by analyzing the enzymatic activity and glycosylation pattern of Tyr point mutants and chimeric constructs, where selected portions of the Tyr molecule were replaced by the homologous fragments of Tyrp1. With this approach, we show that although glycosylation of the CuB acceptor sequon is necessary for full enzymatic activity, it is not a "sine qua non" requirement for ER export, copper binding, and complete maturation. Moreover, mutation of a critical histidine residue involved in copper binding and abolishing completely Tyr activity does not block completely Tyr processing to a mature form. Finally, our results prove that a specific interaction involving the CuB site is essential for processing beyond the ER and for the conformation-dependent N-glycan addition to at least two glycosylation sequons.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Reagents-- The radioactive substrate L-[3,5-³H]-tyrosine, specific activity 50 Ci/mmol, was obtained from Amersham Biosciences. The specific PEP1 and PEP7 antisera, recognizing the C-terminal cytosolic extension of mouse Tyrp1 and Tyr, respectively, were a kind gift from Dr. V. Hearing (National Institutes of Health, Bethesda, MD). The goat polyclonal antibody against the carboxyl terminus of calnexin was from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Electrophoresis and Western blot reagents and materials were from Bio-Rad, unless otherwise specified. Reagents and plasticware for cell culture were obtained from Nunc (Roskilde, Denmark) or Invitrogen. Other reagents were from Sigma, Merck, or Prolabo (Barcelona, Spain). Enzymes for DNA cleavage and modification were from Invitrogen or Fermentas (Hanover, MD). Endoglycosidase H (Endo H) and N-glycosidase F (Endo F) were from Roche Molecular Biochemicals.

Preparation of Wild Type, Mutant, and Chimeric Expression Constructs-- All expression constructs were prepared in the pcDNA3 expression vector (Invitrogen) and were based on the mouse Tyr and Tyrp1 clones obtained as described elsewhere (33). Point mutants were created by PCR with the mutagenic primers shown in Table I, as described (33). Concerning the chimeric constructs derived from Tyr and Tyrp1 sequences, their structures are summarized in Fig. 2. Regarding the Tyr(A)-Tyrp1(B) and Tyrp1(A)-Tyr(B) chimerae, we first abolished an existing SphI restriction site at position 361 in the wild type Tyrp1 with a silent single base substitution within the target sequence (primer number 10) and created a new SphI site immediately before the codon corresponding to the first His in CuB (primer number 11). A homologous restriction site is present at this position in wild type Tyr, thus allowing for the fusion of the upstream and downstream portions of the Tyr and Tyrp1 genes with conservation of the protein reading frame. The modified Tyrp1 (coding for a protein carrying one amino acid substitution, L376M) was cloned into pBlueScript KSII and used to generate 5' EcoRI-SphI or 3' SphI-XbaI fragments that were used to replace the homologous fragments of Tyr, cloned into pBlueScript KSII. The full-length chimeric constructs were subsequently subcloned into pcDNA3. The Tyr-MeB chimera was constructed by PCR, using primers 2 and 12 with the H389L Tyr construct as template. This amplicon was then purified and used as a reverse primer, with forward primer 1 and the Tyr(A)-Tyrp1(B) construct as a template, under low astringency conditions compatible with primer extension. All constructs were verified by automated sequencing of both strands, performed at the core facility of the Instituto de Investigaciones Biológicas, Consejo Superior de Investigaciones Científicas (Madrid, Spain).

View larger version (21K): [in this window] [in a new window]   Fig. 2.   Schematic representation of the quimeric constructs analyzed in this study. The arrows pointing downward indicate the location of the glycosylation sequons. The positions of the metal binding sites (labeled CuA and CuB for Tyr-derived sequences or MeA and MeB for Tyrp1-derived fragments) is also shown. The hatched boxes located at the N and C termini in each diagram refer to the signal peptide and transmembrane fragment, respectively. Regions corresponding to the Tyr sequence are shown as an open box, and those pertaining to Tyrp1 are shown as a solid black box.

Transfection of HEK 293T cells was performed with the Superfect reagent (Qiagen, Hilden, Germany), according to the manufacturer's instructions. Cells were harvested 20 h after transfection and processed for enzyme activity determinations or Western blot as described below.

Cell Culture and Preparation of Crude Solubilized Extracts-- B16 mouse melanoma cells were cultured as described previously (28) in 75-cm² flasks and allowed to grow to ~80% confluence. Cells were harvested with trypsin, washed twice with saline phosphate buffer, and solubilized in 10 mM sodium phosphate, pH 6.8, containing 1% Igepal CA-630, 0.1 mM EDTA, and 0.1 mM phenylmethylsulfonyl fluoride, using a ratio of ~1 ml solubilization buffer/10⁷ cells. For expression studies, HEK 293T cells were employed. This is a SV40 T antigen-transformed subline of the permanent line of primary human embryonal kidney cells HEK 293 (ATCC number CRL-1573), widely used for the transient expression of genes cloned into vectors carrying the SV40 origin such as pcDNA3. HEK 293T cells were grown in six-well plates with RPMI 1640, 10% fetal calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin, as described elsewhere (29). Crude extracts were prepared as for B16 cells. The protein content of the extracts was determined by the bicinchoninic acid method, using bovine serum albumin as a standard.

Enzyme Activity Determinations-- The tyrosine hydroxylase activity of tyrosinase was determined by a radiometric method (30). One unit was defined as the amount of enzyme catalyzing the hydroxylation of 1 µmol of L-tyrosine/min, in the presence of a 50 µM concentration of the substrate and 10 µM DOPA as cofactor. DOPA oxidase activity was measured spectrophotometrically in the presence of 4 mM 3-methyl-2-benzothiazolinone, as described by others (31) with a final concentration of 2.0 mM DOPA.

Electrophoretic Procedures-- Analytical SDS-PAGE was performed as described (32) in 12% acrylamide gels. Samples were mixed in a 2:1 ratio with 3× sample buffer (0.18 M Tris-HCl, pH 6.8, 15% glycerol, 0.075% bromphenol blue, 9% SDS, with or without 3 M 2-mercaptoethanol), and electrophoresed at 4 °C. A highly sensitive and specific DOPA oxidase activity stain was carried out by equilibrating gels run under nonreducing conditions at pH 6.0, with 50 mM sodium phosphate buffer, followed by incubation at 37 °C in 2 mM DOPA, 4 mM 3-methyl-2-benzothiazolinone, in 10 mM phosphate buffer, pH 6.8, for 15-30 min (32).

For deglycosylation studies, the extracts were incubated at 37 °C for 4 h in the presence of 5 units of either Endo H or Endo F in 50 mM phosphate buffer, pH 7.0, containing 10 mM EDTA and 0.1% SDS. Samples for Endo F digestion were heated at 95 °C for 5 min prior to incubation at 37 °C and processed as previously described for electrophoresis.

Immunochemical Techniques-- For Western blots, SDS-PAGE gels were run under nonreducing conditions, as described above. Transfer to polyvinylidene difluoride membranes (PolyScreen; PerkinElmer Life Sciences) was done in a semidry unit. Immunodetection of Tyr, Tyrp1, and the mutant proteins was performed with the PEP7 or PEP1 antisera, as appropriate, following previously published procedures (33, 34). Staining and detection were done with the ECL Plus chemiluminescent substrate (Amersham Biosciences). The relative intensities of the specific bands were quantified in a Gel Doc system (Bio-Rad), using the Multi-Analyst software.

For immunoprecipitation experiments, cells were solubilized in 400 mM KCl, 2% Triton X-100, 50 mM Tris-HCl, pH 7.5, buffer with 1× protease inhibitor mixture (50 µg/ml phenylmethylsulfonyl fluoride, 1 µg/ml leupeptin, and 1 µg/ml pepstatin). 200 µl of cell lysate (~2 × 10⁶ cells) were precleared by incubation with 20 µl of prewashed Protein G PLUS-agarose slurry (Santa Cruz Biotechnology) for 1 h on ice. Immunoprecipitation was then performed by incubating 2 µg of calnexin C-20 goat polyclonal IgG (Santa Cruz Biotechnology) with 20 µl of Protein G PLUS-agarose for 1 h at 4 °C with continuous shaking, followed by the addition of the precleared lysate and further incubation from 2 h to overnight at 4 °C. Beads were washed four times with a 10 mM Tris-HCl buffer, pH 7.5, containing 500 mM KCl, 400 mM NaCl, and 0.05% Triton X-100. 20 µl of 1.5× sample buffer for SDS-PAGE with 2-mercaptoethanol were added to the beads and incubated for 10 min at 95 °C. Supernatants were processed for SDS-PAGE and Western blotting with PEP7 as previously described.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Expression and Processing of Wild Type Tyr and Tyrp1 in HEK 293T Cells-- Analysis of the structure-function relationships in the Tyr family proteins by site-directed mutagenesis requires an efficient expression system. To date, a variety of heterologous cells devoid of endogenous melanogenic enzymes have been used for the stable or transient expression of the corresponding genes. Recent studies employed COS7 (23, 34, 35), Chinese hamster ovary (16), or HeLa cells (18, 36). We have found that HEK 293T cells transfected with wild type and mutant Tyr constructs cloned into pcDNA3 transiently express very high levels of active Tyr (33). Therefore, we examined Tyr folding and processing in these cells by means of glycosidase digestion with Endo H and Endo F, followed by Western blot. Endo H efficiently cleaves high mannose glycans, such as the ones found in incompletely processed glycoproteins present in the ER (37). However, upon further processing to complex glycans in the medial Golgi, glycoproteins become resistant to Endo H. Accordingly, sensitivity to Endo H provides a criterion to distinguish early forms of ER-resident, incompletely processed tyrosinase from mature forms of the enzyme (19, 21-23). On the other hand, Endo F removes all forms of glycans, irrespective of their degree of processing and trimming, thus allowing for an estimation of the size of the polypeptide backbone of glycoproteins.

Western blots of native and glycosidase-treated wild type Tyr expressed in HEK 293T cells showed a major band (apparent M_r 78.4 ± 0.9 kDa, n  5) and a faster migrating minor band (M_r 69.5 ± 0.5 kDa, n  5) (Fig. 3A). Upon treatment with Endo F, a single band of 55.5 ± 1.5 kDa was seen, corresponding to the deglycosylated protein backbone. Therefore, the two bands present in the native extracts corresponded to different glycosylation forms. The expressed Tyr protein was mostly resistant to Endo H, thus showing that it is a mature, post-ER form (Fig. 3A). The enzymatic activity of crude extracts from transfected HEK 293T cells was higher than in B16 mouse melanoma cells (for tyrosine hydroxylase activity, 526 ± 26 microunits/mg protein versus 305 ± 27 microunits/mg in B16 melanoma cells, and for DOPA oxidase activity 51 ± 2 milliunits/mg versus 32 ± 4 milliunits/mg), consistent with an efficient expression, folding, and processing of the protein. In keeping with the behavior of Tyr, wild type Tyrp1 was also processed to an Endo H-resistant protein of 78.1 ± 1.2 kDa (n  3) that could be deglycosylated by Endo F to yield a 57.4 ± 0.9-kDa protein core (Fig. 3B).

View larger version (49K): [in this window] [in a new window]   Fig. 3.   Efficient glycosylation and processing of wild type Tyr (A) and Tyrp1 (B) in HEK 293T cells. Extracts from B16 mouse melanoma cells (8 µg of total protein/lane) or from HEK 293T cells transfected with the wild type Tyr and Tyrp1 genes cloned into pcDNA3 (20 µg of total protein/lane) were electrophoresed, blotted, and probed with PEP7 (for detection of Tyr) or PEP1 (for detection of Tyrp1). For each blot, the migration of molecular molecular weight markers is shown on the left. C, untreated control extracts; eH and eF, extracts digested with Endo H and Endo F, respectively. The protein load was the same for control and glycosidase-treated samples.

Therefore, HEK 293T cells efficiently express and process both Tyr and Tyrp1 and constitute a suitable model to study the folding determinants of the Tyr family proteins.

Chimeric Constructs Reveal Conformation-dependent Glycosylation Sites in Tyr and Tyrp1-- In the course of a study aiming to define structural elements specific for each protein of the Tyr family and to differentiate them from common elements that can fulfill their function within any member of the family, we constructed and analyzed several chimeric proteins. In preliminary experiments, two constructs designated Tyr(A)/Tyrp1(B) and Tyrp1(A)/Tyr1(B) were studied. Tyr(A)/Tyrp1(B) comprises the complete N-terminal portion of Tyr, up to residue 362 (the residue preceding the first His ligand in the CuB site, His³⁶³), and then the complete CuB and C-terminal portion of Tyrp1, including the cytosolic tail, which bears all of their necessary sorting and trafficking signals (reviewed in Ref. 15). Conversely, Tyrp1(A)/Tyr1(B) contains an opposite distribution, with the N-terminal moiety of Tyrp1 followed by the in-frame CuB site and C terminus of Tyr (Fig. 2). The size of the polypeptide backbone of the two chimeric proteins is similar, with 523 amino acids for Tyr(A)/Tyrp1(B) and 547 for Tyrp1(A)/Tyr1(B), as compared with 533 and 537 amino acids for Tyr and Tyrp1, respectively. Thus, should glycosylation and post-translational processing of the constructs and wild type proteins be comparable, the electrophoretic mobility of the chimeric and the parent proteins in reducing SDS gels would also be very similar. However, we detected an abnormal electrophoretic pattern, with a single band of higher mobility than expected (Fig. 4), corresponding to apparent molecular masses of 65.3 ± 1.2 (n = 3) and 64.1 ± 0.8 (n = 3) kDa for the Tyrp1(A)/Tyr1(B) and Tyr(A)/Tyrp1(B) chimerae, respectively. Upon treatment with either Endo F or Endo H, both chimerae yielded a single band of apparent molecular weight of 55.6 and 59.1 kDa, respectively, consistent with their expected size and with the one of the parent proteins. This showed that the higher electrophoretic mobility of the untreated chimeric constructs is accounted for by a severe underglycosylation as compared with Tyr or Tyrp1. Moreover, their sensitivity to Endo H strongly suggested that these chimeric proteins do not fold correctly and, as a result of an aberrant conformation, are retained in the ER, where the N-glycan chains are not processed to complex-type oligosaccharides. Consistent with an improper and incomplete processing, the chimerae were completely devoid of the typical tyrosinase enzymatic activities.

View larger version (47K): [in this window] [in a new window]   Fig. 4.   Glycosylation status of the chimeric constructs Tyr(A)/Tyrp1(B) and Tyrp1(A)/Tyr1(B). Wild type Tyr and Tyrp1 and the chimeric constructs Tyrp1(A)/Tyr1(B) and Tyr(A)/Tyrp1(B) were transiently expressed in HEK 293T cells. Control extracts (lanes labeled C) and extracts treated with Endo H (eH) or Endo F (eF) were analyzed by Western blot, with a protein load of 10 µg/lane. PEP7 was used for the detection of Tyr and Tyrp1(A)/Tyr1(B), since these proteins share the C terminus of Tyr, whereas PEP1 was employed to detect Tyrp1 and Tyr(A)/Tyrp1(B). The migration of molecular weight markers is shown on the left.

Tyr, Tyrp1, and the two chimerae possess the same number of glycosylation sites (Fig. 2). Therefore, the different electrophoretic mobility of the chimeric constructs and the parental proteins must be due to lack of utilization of potential glycosylation site(s) normally occupied in the wild type proteins, as opposed to the withdrawal of N-glycosylation sequons from their primary sequences. Thus, the glycosylation pattern of the chimeric constructs strongly suggested that Tyr and Tyrp1 display conformation-dependent N-glycan acceptor sites. These sites would not reach a glycosylation-competent conformation in the chimeric proteins, due to an aberrant folding.

Normal Processing and Complete Glycosylation of Tyr Is Dependent on a Highly Specific Interaction Involving the CuB Site-- In an attempt to further define the region of the molecule involved in the establishment of a normal glycosylation-competent conformer, we constructed a more refined chimera consisting of the complete Tyr molecule, except for the replacement of the CuB site (residues His³⁶³-His³⁹⁰; see Fig. 1) by the homologous fragment of Tyrp1. This construct was designated Tyr-MeB (Fig. 2), and its enzymatic activity, electrophoretic mobility, and glycosylation pattern were also analyzed after transient expression in HEK 293T cells. Tyr-MeB was devoid of enzymatic activity and failed to undergo normal processing, as shown by sensitivity to Endo H and underglycosylation, with an apparent molecular mass of 68.5 ± 1.6 kDa (n = 3) in control extracts, versus 78.4 ± 0.9 kDa (n = 7) for the wild type Tyr (Fig. 5), despite the identical distribution of N-glycosylation sequons and size of the protein moiety. In agreement with the similarity of the protein backbones, Endo H treatment decreased the molecular weight of the Tyr-MeB construct to 55.9 ± 2.8, a value that compared well with the 55.5 ± 1.5-kDa size observed for wild type Tyr. Moreover, in keeping with the aberrant glycosylation pattern, suggesting ER retention, Tyr-MeB was shown to interact with calnexin to a much higher extent than Tyr, in coimmunoprecipitation experiments performed with an anti-calnexin antibody (Fig. 6). Therefore, replacement of the CuB site sequence had the same effect as the change of the complete C-terminal half of the protein and caused aberrant processing, inability to reach a full N-glycosylation-competent conformation, and ER retention.

View larger version (73K): [in this window] [in a new window]   Fig. 5.   Replacement of the Tyr CuB site by the homologous sequence in Tyrp1 abolishes normal N-glycosylation. Wild type Tyr or the Tyr-MeB construct, identical to Tyr except for the replacement of the CuB site by the homologous fragment of Tyrp1, was transiently expressed in HEK 293T cells. For both proteins, identical amounts of total protein (8 µg/lane) from control extracts (C) or extracts treated with Endo H (eH) were analyzed by Western blot, probed with PEP7. Note the higher electrophoretic mobility of the control, untreated Tyr-MeB construct, indicative of a lower degree of glycosylation, and the complete lack of an Endo H-resistant form for this protein.

View larger version (73K): [in this window] [in a new window]   Fig. 6.   Association of Tyr-MeB, but not Tyr, with calnexin. Extracts from HEK 293T cells transiently expressing wild type Tyr or the Tyr-MeB construct, as indicated at the top of each lane, were immunoprecipitated with a polyclonal antibody against calnexin, followed by Western blotting probed with PEP7. Lanes labeled Controls correspond to the precleared supernatants, before immunoprecipitation (protein load of 10 µg/lane), and those labeled Anti-calnexin correspond to samples submitted to the complete immunoprecipitation procedure. The amount of starting total protein was the same for extracts from cells transfected with either Tyr or Tyr-MeB. The blot shows an experiment representative of three.

Glycosylation of the CuB Sequon Is Not Required for Complete Maturation of Tyr or for Cofactor Binding-- Since the CuB site contains a conserved N-glycosylation site, it was tempting to speculate that underglycosylation of Tyr-MeB in comparison with wild type Tyr could arise directly from lack of N-glycan addition to this sequon. Therefore, we explored the glycosylation status of the CuB sequon in both wild type Tyr and the mutant protein. For this purpose, we constructed two point mutants bearing a conservative N371Q substitution, using the wild type or Tyr-MeB cDNAs as starting material. Since this mutation abolishes the sixth glycosylation signal in both cases, the constructs were designated 6-Tyr and 6-Tyr-MeB. Surprisingly, comparison of the electrophoretic mobility of the expressed proteins proved that the glycosylation sequon located in the CuB site was occupied not only in wild type Tyr but also in Tyr-MeB (Fig. 7).

View larger version (63K): [in this window] [in a new window]   Fig. 7.   The CuB site glycosylation sequon is occupied in wild type Tyr and in the Tyr-MeB chimeric protein. Wild type Tyr and Tyr-MeB, along with the corresponding N371Q point mutants where the glycosylation sequon located in the CuB site was abolished (designated 6-Tyr and 6-Tyr-MeB), were transiently expressed in HEK 293T cells. Equal protein loads of crude extracts (12 µg/lane) were analyzed by Western blot and probed with PEP7. Note the higher electrophoretic mobility of 6-Tyr-MeB as compared with Tyr-MeB, indicative of N-glycan addition to the CuB site N-glycosylation sequon in this latter protein. For comparison, the last lane on the right shows the electrophoretic pattern of wild type Tyr, after complete deglycosylation by Endo F treatment (lane labeled Tyr + eF). The mobility of molecular weight markers is shown on the left.

We also assessed the ability of the CuB glycosylation-deficient 6-Tyr mutant to mature to an Endo H-resistant form and to become enzymatically active. 6-Tyr was processed to an Endo H-resistant protein, and its electrophoretic behavior in the absence of the reducing agent mercaptoethanol was also normal, except for the slightly higher electrophoretic mobility due to the absence of the N-glycan chain (Fig. 8). 6-Tyr retained considerable enzymatic activity, and its kinetic constants were also very similar to those of wild type Tyr (Table II). Taken together, these data prove that the CuB glycosylation sequon is occupied in native wild type Tyr and in the Tyr-MeB construct. They also confirm the importance of CuB site glycosylation for full enzymatic activity reported by others (16). However, they demonstrate that N-glycan addition to this site is not required for further processing, including exit from the ER and binding of the metal cofactor. Therefore, the aberrant processing of the Tyr-MeB and other chimeric constructs cannot be explained in terms of lack of glycosylation of the ³⁷¹NGT³⁷³ sequon.

View larger version (50K): [in this window] [in a new window]   Fig. 8.   A CuB glycosylation-deficient Tyr point mutant is processed to an Endo H-resistant, enzymatically active form. Wild type Tyr and the 6-Tyr mutant, where the N-glycosylation sequon located in the CuB site was abolished by the N371Q substitution, were transiently expressed, and their electrophoretic mobility was probed by Western blot with PEP7. In A, electrophoresis was performed under reducing conditions, for control cell extracts (lanes labeled C) and for equivalent protein amounts (10 µg/lane) of extracts treated with Endo H (eH) or Endo F (eF). The electrophoretic mobility of molecular weight standards is shown on the left. In B, the electrophoretic mobility of the wild type Tyr and 6-Tyr proteins under nonreducing conditions is compared by Western blot. Protein loads were the same as in A. C shows a specific in-gel DOPA oxidase activity stain after nonreducing electrophoretic separation of extracts from cells transfected with wild type enzyme (Tyr), the glycosylation-deficient mutant (6-Tyr), or empty vector (mock), as blank. In this case, the protein load was 20 µg/lane.

View this table: [in this window] [in a new window]   Table II Residual tyrosine hydroxylase activity and kinetic constants of the 6-Tyr mutant lacking the sixth potential N-glycosylation sequon located in the CuB site Values are given as mean ± S.D. for at least three independent experiments.

Acquisition of Endo H Resistance by Transiently Expressed Tyr Mutants Does Not Correlate Strictly with Residual Enzymatic Activity and Is Not Abolished by Mutation of Histidine Residues Involved in Copper Binding-- Since the Tyr-MeB construct is lacking enzymatic activity, the possibility was considered that its deficient processing could be related to inability to bind the metal cofactor. Indeed, cofactor binding promotes a conformational change in most if not all cofactor-dependent enzymes. Therefore, we examined whether improper Tyr-MeB processing and the resulting ER retention could be mimicked by mutation of critical histidine residues involved in copper binding to the CuB site (18, 34). We also wished to determine whether improper processing correlated with the degree of activity loss for several CuB site Tyr point mutants. We used three kinetically well characterized constructs obtained by artificial mutagenesis of selected residues in the CuB site (34). As shown in Fig. 9, the Q378H, H389L, and H390Q mutants were all able to progress, at least partially, to an Endo H mature form, although, in all cases, a sizable fraction of the expressed protein remained Endo H-sensitive. Despite a similar residual activity (around 20% of the tyrosine hydroxylase activity of wild type Tyr), the Q378H mutant yielded a majority band corresponding to the post-ER, Endo H-resistant form, whereas the H389L mutant was more sensitive to Endo H and thus very likely retained in the ER. Therefore, acquisition of Endo H resistance by transiently expressed Tyr mutants, indicative of exit from the ER to the Golgi, does not strictly correlate with residual enzymatic activity. Interestingly, the H390Q protein was also able to undergo substantial maturation to an Endo H-resistant form (Fig. 9). This mutant is absolutely devoid of enzymatic activity, most likely as a result of its inability to bind the metal cofactor due to the absence of one of the His imidazole ligands of the copper atom in the CuB site (34). Accordingly, it can be concluded that processing beyond the ER is not dependent on copper binding.

View larger version (56K): [in this window] [in a new window]   Fig. 9.   CuB active site Tyr mutants are processed to an Endo H-resistant, mature form. HEK 293T cells were transfected with the three constructs shown. Equal amounts of cell extracts (10 µg) were treated with Endo H (lanes labeled eH) or Endo F (eF) and analyzed by Western blot using PEP7 for specific detection. C, control, untreated extracts. The mobility of molecular weight markers is shown on the left. All mutants displayed a significant fraction of Endo H-resistant protein.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Interest in the mechanisms of tyrosinase processing and transport has been recently fostered by the realization that impairment of these processes disrupts melanin synthesis, leading to an amelanotic or albino phenotype. Evidence supporting this view has accumulated rapidly (19, 21-25). Moreover, tyrosinase is considered an excellent model to study the relationship between lectin-like ER chaperones and glycoprotein folding and the role of ER quality control systems in the processing of secretory proteins (reviewed in Ref. 38). However, knowledge of the structural requirements for tyrosinase folding and processing in the ER is still incomplete and mostly limited to the role of N-glycan chains (16, 39, 40) with very little information on the interactions within the protein backbone.

We addressed this question by analyzing the glycosylation pattern and enzymatic activity of Tyr mutants and chimeric proteins based on the Tyr and Tyrp1 sequences. These were transiently expressed in HEK 293T cells, a heterologous cell system that proved to be highly efficient in processing wild type Tyr to an enzymatically active form, with a kinetic behavior comparable with the enzyme in its natural melanocytic environment (34). This system was selected because other heterologous cell types employed in previous studies failed to yield an optimal processing of Tyr. For instance, Tyr is apparently poorly processed by COS7 cells, and high yields of enzymatically active, fully mature protein could only be achieved after cotransfection with the ER-resident, lectin-like chaperone calnexin (41). HeLa cells may also be inefficient in the folding of tyrosinase, resulting in the appearance of temperature-sensitive, incompletely processed forms of the protein and in an exaggeration of the trafficking defects associated with certain mutations such as the temperature-sensitive R402Q (36).

Tyr expressed in HEK 293T cells appeared as a doublet in Western blots probed with PEP7, a specific antibody directed against its C terminus (12). The finding of two Tyr bands is a common feature in most cell types, including heterologous systems (36) and melanocytic cells such as B16 mouse melanoma cells (39), human melanocytes (42), and human melanoma cells (19, 42). The precursor-product relationship of the two bands is suggested by a wealth of experimental data reported by others (42, 43). Accordingly, in our system, both bands correspond to differentially glycosylated isoforms as shown by treatment with Endo F, which yielded a single band of higher electrophoretic mobility than any one of the original forms. Moreover, the higher mobility band appeared Endo H-sensitive and nonreducing SDS-PAGE gels followed by DOPA oxidase activity stain show a single band instead of a doublet (Fig. 8). Taken together, these observations prove that the higher mobility form corresponds to a partially glycosylated, incompletely processed and enzymatically inactive protein that may give rise to the lower mobility band upon further glycosylation. However, our data do not allow us to determine exactly the number of glycosylation signals occupied in each one of the isoforms. Based on studies by others, it can be speculated that the mature, lower mobility band could contain four N-glycan chains. Indeed, using several mutants where specific N-glycosylation sequons were abolished, it was shown that wild type Tyr expressed in Chinese hamster ovary cells bears four glycosylated and two unglycosylated sequons (16). Four occupied sequons were also found in hamster tyrosinase (44), and it has been shown for TYR that, when the rate of protein synthesis is high, a partially processed protein with an unoccupied N-glycosylation site is present (17). However, it is also possible that the different heterologous cellular systems employed for transient expression of tyrosinase might yield a slightly different processing, and therefore, a higher degree of N-glycosylation in our experimental system cannot be ruled out.

Concerning the different glycosylation pattern of the two Tyr species, the experiments performed with the 6-Tyr mutant show that the higher electrophoretic mobility form should bear at least two additional unglycosylated sequons as compared with the mature form. This is demonstrated by the observation that the mobility of 6-Tyr, lacking one glycosylation site, is intermediate between that of the two glycosylation isoforms observed for the wild type protein (Fig. 7). In any case, our data show that complete Tyr translation yielding a full-length polypeptide backbone recognizable by the C terminus directed PEP7 antiserum can occur without a complete co-translational N-glycosylation of the protein. Therefore, it appears that Tyr contains two types of N-glycan acceptor sites, distinguishable on the basis of their kinetics of glycosylation: sites of rapid, co-translational glycosylation and sites of slow, conformation-dependent glycosylation. Interestingly, complete processing of Tyr is a relatively slow process as compared with other proteins and particularly as compared with the highly similar Tyrp1 (reviewed in Ref. 38). However, the initial glycosylation events must be very rapid, since a native, completely unglycosylated protein is never seen in control samples. This is fully consistent with the co-translational nature of the early glycosylation processes.

On the other hand, chimeric constructs Tyr(A)/Tyrp1(B) and Tyrp1(A)/Tyr(B) were blocked in an underglycosylated and misfolded state as shown by the presence of a single Endo H-sensitive band of lower molecular weight than either Tyr or Tyrp1. Again, the electrophoretic mobility of the chimeric proteins proved that they failed to undergo N-glycosylation in more than one site normally occupied in the parent protein. This inability to progress to a fully mature protein is in contrast with the presence within the chimeric and parent proteins of the same number of N-glycosylation signals, with an identical distribution with respect to the N terminus. Both Tyr and Tyrp1 display six glycosylation sequons. Five of them are located N-terminal to the perfectly conserved sixth site that lies within the CuB metal binding site. The first glycosylation site is also invariant within mouse and human tyrosinase and Tyrps and is located in a region of high sequence similarity in the boundary between the two N-terminal Cys clusters of the epidermal growth factor-like domain (Fig. 1). Accordingly, the chimeric constructs Tyr(A)/Tyrp1(B) and Tyrp1(A)/Tyr1(B) both display six potential N-glycosylation sites (i.e. the same number of sites as the parental proteins). Moreover, as shown in Fig. 2, and due to the strict conservation of the N-glycosylation sequon in the CuB site, all of the potential glycosylation signals in Tyr(A)/Tyrp1(B) are identically located with respect to the N terminus as in Tyr, and the same holds for Tyrp1(A)/Tyr1(B) as compared with Tyrp1.

Since underglycosylation of the chimerae is not related to the withdrawal of target sequons or to a change in their distribution relative to the N terminus, the most likely interpretation is that it must arise from failure of conformation-dependent sites to undergo normal N-glycosylation. Should this be the case, underglycosylation of the chimeric constructs would result from their inability to reach a conformation supporting the recognition of more than one N-glycosylation signal. The chimeric constructs are formed by the N-terminal half of each parental protein, up to the first histidine residue of the CuB site, followed by the in-frame MeB site and C-terminal fragment of the other protein. Thus, the acquisition of the correct glycosylation-competent conformation most likely relies on specific intramolecular interactions between these two parts, that are negated in chimeric constructs.

We attempted to define further the structural elements involved in this interaction, by constructing and analyzing a more selective chimera designated Tyr-MeB. In this construct, exclusively the CuB site of Tyr, comprising amino acids His³⁶³-His³⁹⁰, was replaced by the homologous fragment of Tyrp1. When expressed in HEK 293T cells, the Tyr-MeB construct was enzymatically inactive, underglycosylated in more than one site, to the same extent as the more divergent chimeric constructs and was retained in the ER, as shown by sensitivity to Endo H and strong association with calnexin (Figs. 5 and 6). Therefore, replacement of only the CuB site or of the complete C-terminal part, starting from His³⁶³, yielded proteins with the same behavior. These observations strongly suggest that the structural element located in the C-terminal half of the protein and mainly responsible for the formation of the export-competent conformer is the CuB site. Should this be the case, the most likely scenario would be that the interaction established by this site and responsible for the correct folding would involve the CuA site and the formation of the active site cavity. This interaction must be highly specific, since, within the 28-amino acid stretch replaced in Tyr-MeB as compared with Tyr, 16 positions are invariant, thus leaving a total of only 12 differences in the primary sequence, of which the majority are conservative (Fig. 2). The hypothesis of the early establishment of a CuA-CuB interaction accounts for the intriguing observation that the substrates DOPA and L-tyrosine act to promote folding of human TYR and export of the enzyme from the ER to the Golgi (42). Indeed, this effect implies that the substrates are able to bind to the ER-resident form(s) of the protein. Moreover, the establishment of an active site early during Tyr processing is also consistent with current views of the mechanisms of acquisition of the metal cofactor by metalloenzymes (45). Due to its toxicity, the concentration of free copper is kept exceedingly low in mammalian cells. Metal ions are bound to chaperones and transferred directly to the acceptor sites of metalloproteins in a process requiring that the affinity for the metal cofactor be higher in the acceptor enzyme than in the chaperone (45). Thus, formation of a high affinity metal binding acceptor site must occur before acquisition of the metal cofactor. Although a copper transporter has been localized to the ER membrane (46), and the possibility that copper loading to TYR could occur in this compartment has been mentioned (42), the first DOPA oxidase-positive compartment is the trans-Golgi network (19, 47). Within this compartment, copper is probably delivered by the Menkes and Wilson disease proteins, two P-type ATPases (48), whose mutations cause hypopigmentary disorders. The involvement of the Menkes protein in copper delivery to tyrosinase has been demonstrated (49).

Therefore, Tyr most likely acquires a conformation competent for high affinity copper binding before reaching the Golgi apparatus. Moreover, copper binding and maturation appear to be independent events, based on the results obtained for the H390Q Tyr mutant. This mutant, which is enzymatically inactive most likely as a result of impeded or abnormal binding of the copper cofactor (18, 34), is nevertheless partially processed to a mature, Endo H-resistant form. Interestingly, steady state levels of the correctly processed, Endo H-resistant form of H390Q were similar to those of other point mutants in the CuB site retaining considerable enzymatic activity (Fig. 9). This proves that, at least in these cases, there is no strict relationship between the degree of Tyr enzymatic activity impairment and the extent of ER retention.

The Tyr-MeB construct, together with the related point mutants where the CuB site glycosylation sequon was abolished, also enabled us to investigate the relationships between glycosylation of the CuB sequon, enzymatic activity as an index of copper loading, and export from the ER to the Golgi. The ³⁷¹NGT³⁷³ site was glycosylated in the wild type enzyme, as shown by the shift in electrophoretic mobility of the 6-Tyr mutant. Glycosylation of this site in Tyr expressed in Chinese hamster ovary cells has been previously shown by others (16), and this position is also very likely occupied in vivo, within a melanocytic environment (50) (reviewed in Ref. 40). Surprisingly, the N371Q mutation generated within the background of the Tyr-MeB construct also yielded a protein of increased electrophoretic mobility, thus suggesting that the ³⁷¹NGT³⁷³ site is occupied in this chimeric protein. Therefore, this site does not appear to be one of the conformation-dependent sites whose existence was shown by the electrophoretic behavior of the chimeric constructs. Concerning the maturation and enzymatic activity of the CuB site glycosylation-deficient mutant, 6-Tyr and the wild type enzyme were processed to an Endo H-resistant form in a very similar way. In addition, this mutant protein displayed considerable tyrosine hydroxylase and DOPA oxidase residual activities. Of particular relevance is the fact that the V_max values for both the mutant and wild type Tyr were within the same range. This strongly suggests that the degree of copper binding to the active site is similar for both forms, since a diminished cofactor load for 6-Tyr would decrease significantly its V_max. Moreover, the comparable K_m value proves that the presence of the sugar chain in the CuB site of Tyr has no noticeable effect on the affinity for the monophenolic substrate L-tyrosine. Therefore, occupancy of the ³⁷¹NGT³⁷³ site by a glycan chain is not required for acquisition of an ER exit-competent conformation, maturation of the N-glycan chains to complex-type Endo H-resistant oligosaccharides, binding of the metal cofactor, and enzymatic activity. Interestingly, a natural mutation causing the change from Thr to Lys at position 373 in TYR, thus abolishing glycosylation of the CuB site, is associated with oculocutaneous albinism type 1 and aberrant processing of the mutant protein (26, 50). Therefore, the T373K mutation in TYR seems to have a much more dramatic effect than the N371Q change in Tyr, despite their common effect of abolishing one glycosylation site normally occupied in the wild type protein. This higher functional impairment should be related to a specific effect of the amino acid change, instead of to the mere loss of the N-glycosylation site.

In summary, our results and those of others support a model for in vivo folding of Tyr, where the molecule displays at least two N-glycosylation sites whose occupancy is conformation-independent and probably co-translational. One of these sites appears to be the conserved ³⁷¹NGT³⁷³ site located within the CuB region. Interestingly, it has been proposed that calnexin is either divalent or dimeric during its interaction with Tyr (38), and two glycan chains are also needed for stable binding of the chaperone to other proteins such as RNase (51). A chaperone-assisted specific interaction between the CuA and CuB domains will then generate an active site devoid of copper atoms but able to bind substrates. Moreover, the ensuing conformational change will allow for the post-translational N-glycan addition to new conformation-dependent glycosylation sequons. Although post-translational, as opposed to co-translational, N-glycosylation events appear very rare, at least one case involving peptidylglycine -amidating monooxygenase has been rigorously documented (52). The folded tyrosinase protein will leave the ER in an apoenzymatic form and will bind the metal cofactor in the trans-Golgi network. Mutations that would prevent the specific interaction generating the active site will therefore cause aberrant processing and ER retention. Conversely, those compatible with the CuB site-promoted conformational change will allow for transit of at least a fraction of the newly synthesized Tyr molecules from the ER to the Golgi, irrespective of the degree of residual enzymatic activity.

	FOOTNOTES

^* This work was supported by Comisión Interministerial de Ciencia y Tecnología, Spain, Grants PM99-0138 (to J. C. G.-B.) and BIO2001-0140 (to F. S.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Recipient of a fellowship from the Ministerio de Educación y Cultura, Spain.

^§ To whom correspondence should be addressed: Dept. of Biochemistry and Molecular Biology, School of Medicine, University of Murcia, Apto 4021, Murcia 30100, Spain. Tel.: 34-968-364676; Fax: 34-968-830950; E-mail: gborron@um.es.

Published, JBC Papers in Press, February 20, 2003, DOI 10.1074/jbc.M300658200

² According to current conventions, mouse and human tyrosinase are designated Tyr and TYR, respectively. However, we have used the full term "tyrosinase" in sentences referring to a general behavior of the mammalian enzyme, rather to one particular species.

	ABBREVIATIONS

The abbreviations used are: Tyr, tyrosinase; Tyrp, tyrosinase-related protein; DOPA, L-3,4-dihydroxyphenylalanine; Endo F, N-glycosidase F; Endo H, endoglycosidase H; ER, endoplasmic reticulum.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES