Soluble Tyrosinase is an Endoplasmic Reticulum (ER)-associated Degradation Substrate Retained in the ER by Calreticulin and BiP/GRP78 and Not Calnexin*

Tyrosinase is a type I membrane protein regulating the pigmentation process in humans. Mutations of the human tyrosinase gene cause the tyrosinase negative type I oculocutaneous albinism (OCAI). Some OCAI mutations were shown to delete the transmembrane domain or to affect its hydrophobic properties, resulting in soluble tyrosinase mutants that are retained in the endoplasmic reticulum (ER). To understand the specific mechanisms involved in the ER retention of soluble tyrosinase, we have constructed a tyrosinase mutant truncated at its C-terminal end and investigated its maturation process. The mutant is retained in the ER, and it is degraded through the proteasomal pathway. We determined that the mannose trimming is required for an efficient degradation process. Moreover, this soluble ER-associated degradation substrate is stopped at the ER quality control checkpoint with no requirements for an ER-Golgi recycling pathway. Co-immmunoprecipitation experiments showed that soluble tyrosinase interacts with calreticulin and BiP/GRP78 (and not calnexin) during its ER transit. Expression of soluble tyrosinase in calreticulin-deficient cells resulted in the export of soluble tyrosinase of the ER, indicating the calreticulin role in ER retention. Taken together, these data show that OCAI soluble tyrosinase is an ER-associated degradation substrate that, unlike other albino tyrosinases, associates with calreticulin and BiP/GRP78. The lack of specificity for calnexin interaction reveals a novel role for calreticulin in OCAI albinism.

Tyrosinase (monophenol, dihydroxy-phenylalanine: oxygen oxidoreductase; EC 1.14.18.1) is the rate-limiting enzyme involved in melanin biosynthesis (1,2). This protein, consisting of a large catalytic lumenal domain that is anchored to the membrane by a C-terminal transmembrane domain and a short cytosolic tail (3), undergoes glycosylation prior to being subjected to the endoplasmic reticulum quality control (4). Although misfolded chains are sorted for the endoplasmic reticulum (ER) 1 -associated degradation (ERAD) pathway, folded tyrosinase is exported out of the ER through the secretory pathway targeting the pigmentation site organelle, the melanosome (5,6).
Mutations in the tyrosinase gene result in the absence of pigmentation and are responsible for oculocutaneous albinism type I (OCAI) in humans (7). OCAI was proposed to be an ER retention disease in which misfolded tyrosinase mutants are retained in the ER by the quality control (8,9). Calnexin and calreticulin, as components of the quality control, were shown to interact transiently with the monoglucosylated N-glycans of the misfolded polypeptides (10,11). These lectin chaperones engage the chains into the de-glucosylation/re-glucosylation cycle catalyzed by glucosidase II and glucosyltransferase, with the latter recognizing only incompletely, folded chains. Although the correctly folded polypeptides leave the cycle and the ER, the incompletely folded ones are re-glucosylated and retrieved by calnexin/calreticulin (12,13). Both lectins bind to membrane and soluble glycoproteins, but calreticulin was proposed to bind preferentially to soluble substrates. In general, the role of calreticulin has been largely overlooked, because until recently, it was assumed that the two lectins can substitute for each other in the ER quality control (14). This cycle is also believed to be effective in the case of some albino tyrosinases (8,9,15). Interestingly, calnexin was shown to associate extensively with these mutants, whereas calreticulin contribution was only rarely reported (8).
Over 100 mutations have been identified in the tyrosinase gene of OCAI patients (16). Although many of them are clustered in the lumenal domain, a number of mutations were found to affect, directly or indirectly, the transmembrane domain of tyrosinase. Thus, a base insertion within the TM domain resulting in a dramatic reduction of its hydrophobicity leads to an albino tyrosinase (3). A deletion of 10 bases occurring at the beginning of the tyrosinase TM domain was recently reported in a tyrosinase-negative OCAI patient (19). It has been proposed that these mutations would interfere with the insertion of tyrosinase into the membrane, hence impairing the whole pigmentation process. Breimer et al. (17) reported a mutational hot spot leading to a soluble tyrosinase in two patients with a temperature-sensitive form of albinism. Because the patients also showed a mutation at codon 422, the temperature sensitive phenotype was associated with the point mutation previously reported to result in a similar phenotype (20). On the other hand, Berson et al. (21) have shown that the C-terminal truncation of both wild type and a temperaturesensitive mutant tyrosinase determined their ER retention.
All of the above OCAI frameshift mutations introducing termination codons, which either delete or decrease the hydrophobicity of the single TM domain, are associated with a complete absence of tyrosinase activity (3,16,17,19). Although these studies indicate the ER retention of different forms of truncated tyrosinases as the cause of albinism, the molecular basis for this ER retention has not been addressed as yet. Herein, the ER maturation pathway of a human soluble tyrosinase construct was investigated. We found that, unlike wild type, the truncated tyrosinase did not associate significantly with calnexin, being retained in the ER by calreticulin and BiP/GRP78. This tyrosinase mutant is an ERAD substrate that does not require ER-Golgi transport and retrieval for degradation. Our findings qualify calreticulin as a major component in the ER retention of albino C-terminally truncated tyrosinases.  (23). The construct pTriEX1.1-E2HCV was a kind gift from Dr. Olivier Argaut (Oxford University, Oxford, UK), and the pTriEX1.1-TYR⌬7 was a gift from Viorica Lastun (Institute of Biochemistry, Bucharest, Romania). Rabbit anti-calreticulin (calregulin C-17) antiserum was purchased from Santa Cruz Biotechnology. Rabbit anti-BiP (anti-Grp78) antiserum was from Stressgen (Victoria, British Columbia, Canada). [ 35 S]Methionine/cysteine (Tran35S-label, specific radioactivity 1100 Ci/mmol) was from MP Biomedicals (Asse-Relegem, Belgium). CHAPS was from Pierce. Lactacystin, N-acetyl-leucyl-leucyl-norleucinal (LLN) and kifunensin were from Calbiochem. Nocodazole, bafilomycin, and all other chemicals were from Sigma.

Reagents
Human Tyrosinase Constructs-Full-length cDNA coding for human tyrosinase in the cloning vector pcTyr was a kind gift from Dr. V. J. Hearing (NCI, National Institute of Health, Bethesda, MD). The full-length human tyrosinase was amplified by PCR from pcTyr using the primers: TYRF1, 5Ј-GCTATACCATGGCCCTCCTGGCT-GTTTTG-3Ј and TYRR1, 5Ј-GGCGCGCCTCGAGTAAATGGCTCT-GATA-3Ј. The resulting PCR product was digested with NcoI and XhoI and was cloned in-frame with the eight histidines of pTriEX-1.1 (Novagen EMD Biosciences) into the corresponding sites of pTriEX-1 to generate pTriEX-1.1-TYR. Tyrosinase lumenal domain cDNA was generated by PCR using pcTyr as a template and the following primers: TYRF1 (shown above) and STYRR1, 5Ј-GTATTCTCGAGC-CGACTCGCTTGTTC-3Ј. The resulting PCR product was cloned in pTriEX-1.1 in the same way as WT tyrosinase to generate pTriEX-1.1-STYR truncated at the 474 codon. Each sequence was confirmed by automated DNA sequencing at the DNA sequencing unit of the Department of Biochemistry, Oxford University, Oxford, UK.
Transfection of Cells and Metabolic Labeling-Constructs were transiently transfected in B16F1, HEK293T, CHO, MEF crt Ϫ/Ϫ , and wild type MEF cells. Semiconfluent cells grown for 1 day in 6-well dishes were used to transiently express tyrosinase cDNAs (1 g of DNA/well) using Lipofectamine Plus (Invitrogen). To ensure similar transfection efficiency for all samples, transfection with pTriEX-1.1-TYR and pTriEX-1.1-STYR was performed at the same time using aliquots of the same cells. Cells were analyzed 24 h after transfection. For metabolic labeling, transfected cells (10 7 cells/ml) were starved in the cysteine/ methionine-free medium for 1 h, pulse labeled with 100 -150 Ci of [ 35 S]methionine/cysteine for 20 min, and chased for the times specified (Figs. 1,(3)(4)(5). In some experiments, lactacystin, LLN, nocodazole (20 M, stock solutions dissolved in Me 2 SO), kifunensin (20 nM, stock solution in Me 2 SO), NB-DNJ (5 mM, stock solution in PBS), and bafilomycin (100 nM, stock solution in Me 2 SO) were added to the medium during starvation and maintained during the pulse and chase period. Immediately after chase, the cells were harvested in cold PBS. The cells were then lysed with CHAPS lysis buffer (50 mM HEPES buffer, pH 7.5, containing 2% CHAPS, 200 mM NaCl, and 0.5% protease inhibitor mixture containing leupeptin, aprotinin, sodium EDTA, bestatin, AEBSF, and E-64) for 1 h on ice. When the samples were to be used for immunoprecipitation of BiP/GRP78, the cells were lysed in the presence of 20 units/ml apyrase (to enzymatically deplete ATP).
Immunoprecipitation and SDS-PAGE-[ 35 S]-labeled cell lysates were centrifuged and supernatants were incubated with T 311 antibodies (1:50) overnight at 4°C, followed by the addition of 20 l of protein A-Sepharose and further incubation for 1 h at 4°C. The slurry was washed three times with 0.5% CHAPS in HEPES buffer. Tyrosinase was eluted by boiling the slurry for 5 min in SDS sample buffer with 5% 2-mercaptoethanol. Chaperone co-immunoprecipitations were performed as described previously (24). Briefly, lysates were immunoprecipitated with anti-calnexin, anti-calreticulin, or anti-BiP/GRP78 antibodies, and the washed slurry was eluted with 1% SDS, diluted ten times with lysis buffer, and re-precipitated with T311 antibodies. The bound proteins were eluted in reducing conditions and resolved by SDS-PAGE. The gels were visualized by autoradiography. Relevant bands were quantified by scanning densitometry.
EndoH Digestions-Immunoprecipitated [ 35 S]-labeled samples were eluted from protein A-Sepharose slurry in EndoH denaturing buffer (0.5% SDS 1% 2-mercaptoethanol) by incubation for 5 min at 100°C. The samples were cooled and mixed with 1 ⁄10 EndoH reaction buffer (0.5 M sodium citrate, pH 5.5). 500 units of EndoH (1 l) was added to one-half of the amount of each sample, whereas the other half contained EndoH buffers alone. The digested samples, together with non-digested controls, were incubated for 18 h at 37°C. In Western blotting experiments, cell lysates containing 20 g of protein were denatured in the EndoH denaturing buffer (0.5% SDS, 1% 2-mercaptoethanol) for 5 min at 100°C, cooled, and mixed with 1 l of EndoH or reaction buffer.
Immunoblotting-Cells were harvested 24 h after transfection and lysed in CHAPS lysis buffer. The proteins from transfected lysed cells were electrophoretically separated in 10% acrylamide gels and transferred to Immobilon membrane (Amersham International, Amersham, UK). Blots were incubated with a 1:200 dilution of anti-tyrosinase antibody (T311) in 1% milk and 0.1% Tween 20 for 2 h at 37°C, followed by sheep anti-mouse IgG-horseradish peroxidase. Immunoreactivity was detected by enhanced chemiluminescent Western blotting (ECL, Amersham Biosciences) according to the manufacturer's instructions.
Immunofluorescence-B16F1 cells were plated on coverslips and transfected with pTriEx-1.1-TYR and pTriEx-1.1-STYR using the method specified above. After 24 h, the cells were rinsed with PBS and fixed and permeabilized with methanol at Ϫ20°C for 5 min. Some cells were incubated with 20 M nocodazole for 3 h before fixation. In that case, the control untreated cells were incubated for the same period with the appropriate dilution of Me 2 SO. After washing eight times in PBS, the cells were incubated with the primary antibodies T311 (1:250), anti-CNX (1:300), and anti-ERGIC-53 (1:100) diluted in PBS for 30 min at room temperature. Following three washes with PBS, they were further incubated with the appropriate Alexa 488 or 594 conjugated secondary antibodies (1:400) in PBS for 30 min at room temperature. Finally, the cells, washed three times with PBS, were mounted in Vectashield mounting medium (Vector Laboratories) and viewed with a Nikon Eclipse E 600 fluorescent microscope. Images were processed using Adobe Photoshop 5.0 software.

Soluble Tyrosinase Is Retained in the ER of Melanocytic and
Non-melanocytic Cells-Constitutively synthesized only by melanocytes and melanoma cells, wild type tyrosinase is localized in specialized organelles named melanosomes. With melanin being synthesized exclusively in melanosomes, sequestration of tyrosinase in any pre-melanosomal compartment leads to an albino phenotype. To study the intracellular localization of a soluble albino tyrosinase, we have constructed a mutant truncated at the 474 codon and expressed it in the mouse melanoma cell line B16F1. Both wild type and truncated tyrosinase were transiently transfected and analyzed by immunofluorescence microscopy (Fig. 1A). It can be seen that wild type tyrosinase displays mainly a distinct punctate cytoplasmic pattern visible also along the dendrites, consistent with an intracellular vesicular distribution specific to a melanosome. Some of the protein displays a perinuclear staining co-localizing with the ER resident protein calnexin. In contrast, soluble tyrosinase is massively co-localized with calnexin, suggesting an ER localization. This has been previously shown for a soluble tyrosinase truncated at the 444 codon and expressed in HeLa cells (21). Although the two tyrosinases were truncated at different codon positions, they displayed similar ER localizations, indicating that truncations occurring ahead of the TM domain result in ER retention.
Next, to confirm the immunofluorescence localization, transiently transfected B16F1 cells were characterized by Western blotting (Fig. 1B). Cell lysates were divided in two, and half of each sample was digested with EndoH and run next to a nondigested control in reducing SDS-PAGE. Digestion with En-doH, which removed high mannose and hybrid N-glycans, resulted in a soluble tyrosinase form running as a single band at ϳ50 -54 kDa, the expected size of the truncated tyrosinase estimated at 54 kDa (Fig. 1B). In contrast, the wild type tyrosinase was expressed as two glycoforms, one EndoH-resistant and one EndoH-sensitive running at ϳ58 -60 kDa, corresponding to the polypeptide (Fig. 1B). Therefore, the ectopical expression of human wild type tyrosinase in B16F1 cells results in a protein able to exit the ER in a proportion of 38% (Fig. 1C) and to be transported through the secretory pathway. In the same system, truncation of the TM domain dramatically decreases the proportion of the ER-exported form, confirming the ER localization observed by immunofluorescence. To address the possibility that this behavior is specific only to melanocytes, we expressed soluble tyrosinase in two non-melanocytic cell lines, CHO and HEK293. As seen in Fig. 1B, we found a similar EndoH pattern as that found in melanocytes. We were not able to detect soluble tyrosinase in the medium of the tested cells referred to above (data not shown). Our results show that the soluble tyrosinase mutant remains in the ER of melanocytic and non-melanocytic cells, consistent with a similar localization of a soluble tyrosinase expressed in HeLa cells (21).
To compare the time course maturation of wild type and soluble tyrosinases, we performed a pulse-chase experiment in B16F1-transfected cells. The cells were pulse-labeled with [ 35 S]methionine/cysteine for 20 min followed by chase for time periods up to 2 h. Immunoprecipitation of the labeled cell lysates with T311 anti-tyrosinase antibodies was followed by SDS-PAGE analysis (Fig. 1C). The mock immunoprecipitation in non-transfected B16F1 cells show the absence of nonspecific bands (Fig. 1C). After a 30-min lag period, a gradual reduction in the amount of the labeled immunoprecipitated soluble and wild type protein was observed during the chase. In cells transfected with ST, digestion with EndoH reduced the pool to a band running at the polypeptide mass (Fig. 1C). As shown by EndoH digestion experiments, WT is synthesized as a high mannose precursor that acquires complex-type glycans in the Golgi in ϳ30 min of chase and is degraded rapidly during 2 h of chase (Fig. 1C). As reported previously (8), a significant population (62%) of the labeled WT remains EndoH-sensitive, indicating a reduced level of productive folding of this protein (Fig.  1D). Although membrane tyrosinase has a half-life of 1.8 h, soluble tyrosinase degrades with a half-life of 1.5 h. Because the half-life of the ER-retained WT (WT-sensitive, Ͼ2 h) is at least twice as high as the one of the secreted population (WTresistant, 1 h), we analyzed the two populations separately. As seen in Fig. 1D, the WT-sensitive population degrades slower than the soluble tyrosinase population.
Although it has been proposed that soluble tyrosinase behaves like the WT-sensitive form of the membrane tyrosinase (21), our degradation kinetics studies indicate an accelerated degradation of the soluble tyrosinase mutant versus the ERretained form of membrane tyrosinase. This prompted us to investigate further the degradation pathways of soluble tyrosinase.
Soluble Tyrosinase Is an ERAD Substrate-Proteins that are not able to fold correctly are stopped by the ER quality control checkpoint in the ER and targeted to degradation. These ERAD substrates are retrotranslocated in the cytoplasm and degraded in proteasomes. Sometimes the misfolded proteins require transport to the early Golgi and retrieval to the ER prior to degradation (25).
To determine whether soluble tyrosinase is an ERAD substrate, we first asked the question as to whether its ER retention was a true ER retention or was the result of exit and retrieval back into the ER. To answer this question, we analyzed the localization of the soluble mutant in the presence and absence of nocodazole, which inhibits the retrograde transport Golgi-ER (26). Immunofluorescence experiments of the transfected cells show only partial co-localization of ST with the ERGIC-53 protein known to recycle between ER and Golgi and mostly localization in the ER/Golgi intermediate compartment (ERGIC) (22,27) (Fig. 2). The distinct localization of the two proteins is even more obvious after the treatment of cells with nocodazole, which induces the distribution of ERGIC in the early Golgi (28). In these conditions, soluble tyrosinase is localized in a different compartment than ERGIC, as seen by the lack of merge between the two proteins. Although ERGIC-53 has a typical peripheral early Golgi pattern, soluble tyrosinase displays an ER distribution, which indicates that ST is not transported beyond the ER, being retained in this compartment without any retrieval steps.
Next, we investigated the effect of inhibitors of the ERAD components on the degradation of the truncated mutant. An ERAD substrate is targeted to proteasomal degradation, and therefore, proteasome inhibitors reduce the degradation rate of these substrates. Associations with lectins recognizing monoglucosylated (CNX, CRT) and Man8 (EDEM) glycans have been reported to be required for the dislocation of the ERAD substrate into the cytoplasm (29). To identify soluble tyrosinase as an ERAD substrate, we have used inhibitors of the proteasomal degradation and of the lectin associations and analyzed the degradation rate. Transfected cells were pulse-chased for 0 and 3 h in the presence of various inhibitors, and the cell lysates were immunoprecipitated with T311 antibody and analyzed by SDS-PAGE. As seen in Fig. 3, the percentage of non-degraded labeled protein increased from 30 to 78% in the presence of the proteasome inhibitor lactacystin and to 95% in the presence of LLN, which is another proteasome inhibitor. Kifunensin, which inhibits the N-glycans trimming by ER mannosidase I (29), showed a substantial inhibitory effect on the degradation of soluble tyrosinase. This inhibitor of the ER N-glycan processing pathway increases the accumulation of soluble tyrosinase at 3 h of chase to 94%. Bafilomycin and nocodazole have very little effect on the degradation rate of soluble tyrosinase, confirming that the mutant is targeted to degradation without any recycling through the ERGIC or Golgi (Fig. 3). Taken together, the data show that soluble tyrosinase is an ERAD substrate degraded mainly through the proteasomal pathway.
Association of Soluble Tyrosinase with the ER Chaperones-The reduced degradation rate observed during the inhibition of the N-glycan processing pathway indicates the contribution of the ER quality control in soluble tyrosinase retention. To address the role of the ER quality control chaperones CNX and CRT in the maturation of ST, we have sequentially immunoprecipitated the metabolically labeled transfected B16F1 cell lysates with antibodies against CNX, CRT, and BiP/GRP78. In these experiments, tyrosinase complexes with chaperones were first immunoprecipitated with chaperone antibodies. Subsequently, the soluble and wild type tyrosinases were recovered from their complexes by immunoprecipitation with anti-tyrosinase antibody. Finally, the unbound protein was precipitated with anti-tyrosinase antibody to allow the determination of the total amount of tyrosinase. Wild type tyrosinase has been used as a positive control known to associate with CNX and CRT (4,6). As can be observed in Fig. 4A, wild type tyrosinase associates with CNX, CRT, and BiP/GRP78 starting from the pulse period. Soluble tyrosinase association with the chaperones CRT and BiP/GRP78 is relatively poor during the pulse and increases from 30 min of chase (Fig. 4A). The poor association of ST with chaperones in the early maturation phase suggests an unproductive folding pathway leading to degradation eventually.
Surprisingly, CNX interaction with soluble tyrosinase is at least 8 -10-fold weaker than its interaction with the membrane tyrosinase (Fig. 4). However, soluble tyrosinase chain associates with CRT and BiP/GRP78, with calreticulin interacting at least 5-fold stronger than calnexin.
Calreticulin Is Required for Soluble Tyrosinase ER Retention-Because soluble tyrosinase associates with CRT rather than CNX, we further addressed the stringency of CRT in ER retention of soluble tyrosinase. We expressed soluble and wild type tyrosinase in a calreticulin-deficient cell line MEF crt Ϫ/Ϫ and in the control MEF crt ϩ/ϩ cells. Western blot analysis of the transfected MEF crt Ϫ/Ϫ cell lysates revealed that a significant population of the soluble tyrosinase pool was secreted from the ER, as shown by its resistance to EndoH treatment (Fig. 5A, res). In contrast, secreted soluble tyrosinase is almost absent in the calreticulin-positive cells (Fig. 5A), as seen in the other cell lines tested by us (Fig. 1B). These data confirm that CRT is one of the chaperones that retain soluble tyrosinase in the ER and also suggest that, in the absence of CRT, soluble tyrosinase association with CNX is still reduced.
To find out whether a similar ER export is shown by membrane tyrosinase, we have expressed it in both MEF cells. As seen in Fig. 5A, more complex forms than those found in the parental cells were seen in MEF crt Ϫ/Ϫ , indicating that membrane tyrosinase is secreted better in the absence of CRT. We should note, however, that the two tyrosinases have similar molecular masses in the calreticulin-deficient cells, which may be explained by an abnormal glycosylation of the membrane tyrosinase in these cells. Although the mechanisms are not completely understood, we propose that CRT retains soluble tyrosinase and, to some extent, wild type tyrosinase in the ER and that, in its absence, the protein is partially secreted.
To further confirm this finding, we have abolished the CRT and CNX interactions in the transfected calreticulin-positive MEF crt ϩ/ϩ cell line and analyzed tyrosinase secretion. Following treatment of these cells transfected with ST and WT with NB-DNJ (an inhibitor of the ER glucosidases) (6), we have obtained heterogeneously glycosylated WT and ST populations. After the EndoH digestion, it was concluded that the tyrosinase populations are comprised of a mixture of high mannose and complex glycoforms (Fig. 5B). This variety of glycoforms could be due to the increased activity of the Golgi endomannosidase from MEF parental cells that process high mannose glucosylated structures to complex glycans. As a control, calreticulindeficient cells transfected with soluble and membrane tyrosinase were run next to the NB-DNJ-treated samples (Fig. 5B). The Golgi tyrosinases (Fig. 5B, res) migrate with similar velocities in NB-DNJ-treated parental MEF cells and in calreticulindeficient MEF cells, confirming the secretion of tyrosinases in the absence of CNX/CRT interactions.
Because both ST and WT were exported out of the ER, we examined the ER retention capabilities of MEF crt Ϫ/Ϫ cells by analyzing the secretion of two control proteins known to be retained in the ER by different mechanisms. Thus, a tyrosinase mutant lacking the last N-glycan (TYR⌬7) retained by CNX (8) and the envelope protein E2 of HCV retained by its TM domain (22) have been transfected in MEF crt Ϫ/Ϫ cells. Both control proteins were EndoH-sensitive at steady state and therefore retained in the ER, as shown by Western blotting analysis (Fig.  5C). This proves that not all ER retention mechanisms are damaged in MEF crt Ϫ/Ϫ cells, supporting the role of CRT in the ER retention of soluble tyrosinase. Taken together, these results show that, in the absence of CRT, soluble tyrosinase escapes from the ER.

DISCUSSION
The major finding of this study is that the truncated tyrosinase is retained in the ER by a different set of chaperones than the previously reported OCAI mutants. We show that the mutant associates with calreticulin and BiP rather than calnexin. This is in startling contrast with the other albino tyrosinases in which ER export is prevented primarily by their prolonged association with calnexin during ER maturation (8,9,15).
That tyrosinase association with calnexin is crucial for its folding and maturation is known from previous reports investigating mouse and human tyrosinase (30,4). We have shown that calnexin interacts with specific glycans, and this interaction is required to prevent the accelerated folding of some regions of the lumenal domain (4). However, a calreticulin role in wild type or albino tyrosinase maturation has been only occasionally investigated (6,8). Thus, the albino tyrosinases T373K and H402A were reported to be retained in the ER by their prolonged binding to both lectin chaperones (8). In our hands, the membrane-anchored chain was assisted by CNX and CRT starting from the early stages until completion of the folding process. In contrast to wild type, soluble tyrosinase chain showed a weak interaction with calnexin and a stronger affinity for CRT and BiP/GRP78. The post-translational associations with CRT and BiP/GRP78 are unable to rescue the chain from misfolding but appear to be responsible for its ER retention.
To confirm the role of CRT in ER retention, we have expressed the truncated tyrosinase in cells deficient in CRT. Indeed, we found that, in these cells as well as in NB-DNJtreated cells, a significant population of soluble tyrosinase was secreted. Membrane tyrosinase was also massively exported in crt Ϫ/Ϫ cells as compared with crt ϩ/ϩ cells, indicating that CRT retains a population of the membrane-anchored chain in the ER. On the other hand, vesicular stomatitis virus G protein, which does not normally bind CRT, was shown to traffic normally in the CRT-deficient cells (14,31). Therefore, these experiments indicate that, in the absence of calreticulin, the soluble tyrosinase chain escapes the ER quality control and leaves the ER. These cells have been recently characterized, and apart from a slight elevation in the BiP/GRP78 level, the other ER folding factors are expressed at normal levels (14); hence, these effects can be attributed to the CRT knock-out. Significantly, CNX is not able to carry out CRT function and retain the chain in the ER. This may be a consequence of an accelerated folding of the soluble tyrosinase resulting in a rapid export out of the ER. Accelerated folding of two viral proteins has been reported in these calreticulin-deficient cells by Molinari et al. (14), confirming a distinct role for CRT versus CNX. Alternatively, CRT may be required in a distal phase during the multiple release-binding steps of the ER cycle, and in its absence, the chain is released from the cycle and secreted.
It is generally accepted that ERAD substrates may select different pathways in eukaryotic cells. These pathways may be different for soluble and membrane proteins, with soluble ones being transported to the Golgi and retrieved to the ER to undergo ERAD (32). They may also vary according to the location of the misfolded domain within the lumenal, TM, or cytosolic region of the protein (33). Our results show that the truncated tyrosinase mutant is targeted to degradation with a more accelerated rate than the ER-retained form of wild type. To analyze the ERAD pathway of soluble tyrosinase, we have investigated its localization in the presence of inhibitors of the retrograde Golgi-ER trafficking. By immunofluorescence mi- FIG. 4. Association of soluble and wild type tyrosinase with calnexin, calreticulin, and BiP. A, to determine the association of ST and WT with chaperones, pulse-chased transfected cells were sequentially immunoprecipitated (IP) with antibodies to CNX, CRT, BiP/GRP78, and tyrosinase. The chaperone-associated tyrosinases were re-immunoprecipitated from the first immunoprecipitation complexes with anti-tyrosinase antibody. To determine the total amount of tyrosinases, the unbound proteins were precipitated from the last supernatant with anti-tyrosinase antibodies. Samples were subjected to SDS-PAGE. B, levels of the chaperone-associated ST and CNXassociated WT (WT-Cnx) were determined by densitometry and expressed as percentage of the total tyrosinase value/ chase point. Values indicate the mean of three independent experiments Ϯ S.D. croscopy, we have shown that soluble tyrosinase is not recycled between ER and Golgi, but simply retained in the ER. Moreover, the degradation of soluble tyrosinase is not stabilized by inhibitors of the retrograde transport, such as bafilomycin or nocodazole, as suggested for some yeast proteins (32,33). This misfolded mutant appears to be unable to gain transport competence and get access to transport vesicles at the ER level, indicating that the ER quality control checkpoint is sufficient for this soluble protein. One possible explanation is that, when the ER quality control is based on calreticulin and not calnexin, its increased efficiency makes any further checkpoint redundant. This may be another hint that calnexin and calreticulin perform different functions in the quality control.
That soluble tyrosinase is an ERAD substrate retrotranslocated from the ER in the cytoplasm and degraded in proteasomes is supported by the considerable inhibition of ERAD observed with proteasome inhibitors. Considering the reported role of the Man8-binding lectin EDEM in ERAD (34), the accumulation of protein in the presence of mannose-trimming inhibitors suggests that EDEM qualifies as a player in this picture. EDEM was proposed to extract the misfolded chains from calnexin (35). However, our data show that ST interacts mainly with BiP/GRP78 and CRT, the latter having a crucial role in the mutant ER retention. Therefore, we propose that there should be a separate link between EDEM and CRT and/or BiP/GRP78. Because no direct interaction of EDEM with CRT has been observed, further investigations are required to understand this pathway in more detail.
Interestingly, tyrosinase displays a secreted form in OCAII albinism, where a mutation in the P gene induces the truncation and secretion of tyrosinase (18,36). Based on the ER location of the P gene, Orlow and collaborators (36) have proposed that tyrosinase folding is impaired by the mutant P protein. Accordingly, this partially misfolded protein would be susceptible to protease cleavage in the ER, or as has been proposed, in a post-ER compartment. As shown above, truncation in the gene results in a soluble tyrosinase retained in the ER and degraded without further transport and recycling from the Golgi. The vast majority of this truncated protein could not be secreted into the medium. Therefore, our data confirm Orlow's hypothesis (36) that, to be secreted, OCAII tyrosinase undergoes cleavage in a post-ER compartment. This also explains why bafilomycin could reverse the P gene defect in OCAII albinism (18). However, bafilomycin had no effect on the soluble tyrosinase mutant. Therefore, although truncated tyrosinases appear to be the cause for some OCAI albino patients and also for most OCAII patients, a potential treatment has to take into account the different intracellular trafficking pathways of the two albino tyrosinases. Thus, bafilomycin or any other inhibitors that dissipate the pH gradient within the secretory pathway cannot be used as drugs in the case of OCAI albino-truncated tyrosinase. In this form of albinism, what is needed are drugs that induce protein folding at the ER level.
Finally, the ER retention of the soluble form of tyrosinase raises interesting questions related to the role of the TM domain in the folding and maturation of this glycoprotein. Clearly, soluble and wild type tyrosinases have different chaperone interaction patterns, with the soluble form associating with calreticulin rather than calnexin. Whether the TM domain may play an active role in the selection of the appropriate chaperone during the early events of the tyrosinase folding is currently under investigation.