Tyrosinases are widely distributed in nature. They are copper-containing oxidases belonging to the type 3 copper protein family, together with catechol oxidases and haemocyanins. Tyrosinases are essential enzymes in melanin biosynthesis and therefore responsible for pigmentation of skin and hair in mammals, where two more enzymes, the tyrosinase-related proteins (Tyrps), participate in the pathway. The structure and catalytic mechanism of mammalian tyrosinases have been extensively studied but they are not completely understood because of the lack of information on the tertiary structure. The availability of crystallographic data of one plant catechol oxidase and one bacterial tyrosinase has improved the model of the three-dimensional structure of the active site of the enzyme. Furthermore, sequence comparison of tyrosinase and the Tyrps reveals that the three orthologue proteins share many key structural features, because of their common origin from an ancestral gene, although the specific residues responsible for their different catalytic capabilities have not been identified yet.
This review summarizes our current knowledge of tyrosinase and Tyrps structure and function and describes the catalytic mechanism of tyrosinase and Dct/Tyrp2, which are better characterized.
Tyrosinase (monophenol monooxygenase EC 126.96.36.199) is the key regulatory enzyme involved in the biosynthesis of the melanin pigments. This enzyme is able to catalyse two different reactions: the hydroxylation of monophenols to o-diphenols (monophenolase or cresolase activity, EC 188.8.131.52), and the oxidation of o-diphenols to o-quinones (diphenolase or catechol oxidase activity, EC 184.108.40.206). In plants, we find the highly similar catechol oxidases, which catalyse only the oxidation of o-diphenols. The two enzymes (collectively termed phenoloxidases) and the oxygen carrier proteins haemocyanins belong to the type 3 copper protein family. Phenoloxidases share a very similar active site, almost indistinguishable by sequence and physico-chemical properties other than their differential enzymatic activity. They are responsible for pigmentation of skin and hair, browning of fruit and wound healing in plants and arthropods (Claus and Decker, 2006; García-Borrón and Solano, 2002; van Gelder et al., 1997b; Holm et al., 1996; Marusek et al., 2006; Oetting, 2000). It is interesting to mention that pigmentation also takes place in the nervous system forming neuromelanins, and the role of tyrosinase in that process has been proposed. However, the activity levels seem to be very low, and it is unclear whether the enzyme is beneficial or detrimental to neurons (Greggio et al., 2005).
Tyrosinase is a single chain type I membrane glycoprotein which catalyses the critical rate-limiting hydroxylation of l-tyrosine to l-3,4-dihydroxyphenylalanine (l-dopa), which is rapidly converted into l-dopaquinone. In plants and lower organisms, this is the only step of melanogenesis enzymatically controlled, and the pathway then proceeds spontaneously. In animals, two additional enzymatic proteins displaying significant homology with tyrosinase (Tyrps, tyrosinase-related proteins) and originated by the duplication of the ancestral tyrosinase gene (Jackson, 1994) participate in the catalysis (Figure 1A). They are crucial for the transformation of the unstable quinonic intermediates into the more structured and varied melanin polymer (Ito and Wakamatsu, 2006): l-dopaquinone is a reactive intermediate that, in the absence of thiol compounds, spontaneously undergoes cyclization and further rearrangement yielding l-dopachrome (Prota, 1992). Tyrp2, also called dopachrome tautomerase (Dct, EC 220.127.116.11), catalyses the keto–enol tautomerization of dopachrome into the more stable intermediate 5,6-dihydroxyindole-2-carboxylic acid (DHICA) (Aroca et al., 1990; Palumbo et al., 1991). Alternatively, spontaneous decarboxylation of dopachrome yields 5,6-dihydroxyindole (DHI). This metabolic pathway, with the participation of the melanocyte-specific enzymes, results in the synthesis of eumelanin, a black-brown pigment. In the presence of thiol compounds, yellow-red soluble melanins known collectively as pheomelanins are synthesized (Hennessy et al., 2005; Liu et al., 2005). The types of melanin produced depend not only on the availability of substrates (Ito and Wakamatsu, 2003; Wakamatsu et al., 2006), but also on the relative activities of the three melanogenic enzymes, which leads to a refined control of melanosynthesis in animals compared with lower organisms (Aroca et al., 1992; del Marmol et al., 1993).
The distinct catalytic functions of the three enzymes is a consequence of the specific binding of different metal cofactors by tyrosinase and the Tyrps: copper in tyrosinase (Lerner et al., 1950), zinc in Tyrp2 (Solano et al., 1996), and still unknown in Tyrp1 (Furumura et al., 1998). However, recent experiments demonstrate the ability of the MeA site from mouse Tyrp1 to bind copper and sustain the typical tyrosinase enzymatic activities (our unpublished data). Thus, it is very likely that the metal cofactor in Tyrp1, as in tyrosinase, is copper.
Tyrosinases are nearly ubiquitously distributed in all domains of life. Tyrosinases and the corresponding genes have been characterized from various sources, including bacteria, fungi, plants and mammals. The former three have been extensively studied not only because of their biological interest and their utility as promising models to get more insights into the structure and enzymatic properties of these enzymes, but also for their potential use in industrial applications. There are excellent reviews summarizing the current knowledge of bacterial, fungal and plant tyrosinases (Claus and Decker, 2006; van Gelder et al., 1997a; Marusek et al., 2006; Selinheimo et al., 2007). Therefore, we will focus here in mammalian tyrosinases and the Tyrps.
In humans, there is a fourth locus within the tyrosinase gene family: the tyrosinase pseudogene (TYRL), which maps to the short arm of chromosome 11 at p11.2→cen and shows 98% sequence identity with TYR but is not expressed in melanocytes (Giebel et al., 1991; Jackson, 1994; Takeda et al., 1991). This pseudogene is also found in the gorilla but not in other higher primates (Oetting et al., 1993), and is thought to be absent in lower vertebrates.
Despite the high heterogeneity concerning the length and overall identity of the published sequences, the structure and the active site area of tyrosinase is highly conserved among different species and shows high homology with Tyrp1 and Tyrp2. The three proteins share numerous structural characteristics and follow quite similar biosynthetic, processing, and trafficking pathways (Hearing and Tsukamoto, 1991). Regarding tyrosinase, there are a few common structural determinants found in all tyrosinases throughout the phylogenetic scale: the copper-binding ligands and other residues establishing important interactions to maintain the globular folding. In animal enzymes, there are some other remarkable structural characteristics, such as the C-terminal motifs for trafficking and targeting, the cysteine clusters and the N-glycosylation sites. A more detailed review on the structural features of mammalian tyrosinase has been published some years ago (García-Borrón and Solano, 2002), and few relevant data have been added ever since.
Basically, the catalytic centre consists of a hydrophobic pocket inside a helix bundle comprising four densely packed helices. The structure is maintained by electrostatic and cation–π interactions among the helical segments. Two histidine-rich regions named CuA and CuB are the peptidic segments involved in binding the two coppers. Both regions contain three perfectly conserved His residues: CuA has the H-x(n)-H-x(8)-H motif and CuB the H-x(3)-H-x(n)-H motif, where n is a variable number of residues. The second His at the CuA motif is the only flexible one, because it is located in a loop structure in the protein, while the other five His are located in the surrounding α-helical fragments (García-Borrón and Solano, 2002; Inoue et al., 2008). Oxygen binds as a side-on (μ–η2:η2) peroxide bridge (Solomon and Lowery, 1993) and, in the resting form of the enzyme, the pair of antiferromagnetically coupled copper ions are penta-coordinated in a distorted square pyramidal geometry.
However, although the six absolutely invariant His are the true ligands for the copper ions, it is important to notice that in all tyrosinases throughout nature there are some other strictly conserved residues involved in the binding and docking of the substrates or in maintaining the structural integrity of the active site. Therefore, it has been proposed a broader consensus for the copper-binding regions which includes these amino acids in the metal binding regions and proposes a new notation to distinguish residues essential for tyrosinase active site folding (García-Borrón and Solano, 2002). Some of these residues are also present in the Tyrps (Figure 1B). According to this conservation, it is possible to correlate the crystallographic data available, as well as some other studies on the functional effects of naturally occurring or artificially created mutations, with the contribution of these residues to the active site structure or a potential role in the catalysis (van Gelder et al., 1997b; Hearing and Jimenez, 1989; Lerch, 1983; Olivares et al., 2002; Spritz et al., 1997). Therefore, some parallelisms in the structure and catalytic mechanism of tyrosinases from bacteria to human can be ventured and will be introduced throughout this review.
In addition to the CuA and CuB His-based regions, the structure of all mammalian tyrosinases also comprises the N-terminal signal peptide, which is important for intracellular trafficking and processing, the cysteine-rich domains, the C-terminal hydrophobic transmembrane segment and a short cytoplasmic tail (Figure 1A) (García-Borrón and Solano, 2002; Kwon, 1993; Kwon et al., 1987; Shibahara et al., 1988). The transmembrane and C-terminal domains are necessary for targeting the enzyme to the melanosome (Jimbow et al., 2000a,b). The signals for sorting and targeting of melanosomal membrane proteins have been reviewed by Setaluri (2000).
The cysteine (Cys) residues content in tyrosinase is variable throughout nature: for instance, the Streptomyces enzyme is completely devoid of Cys while mammalian tyrosinases have 17 Cys residues clustered in three regions and 15 of them are perfectly conserved in the Tyrps. Plant catechol oxidases contain only the first cluster. Notably, the second cluster has been denominated the epidermal growth factor (EGF)-like region (Jackson, 1994) as it partially matches the two EGF-like motifs and an EGF-related domain, the laminin-LE. However, although almost all Cys seem to be important for a correct protein folding and copper acquisition (Branza-Nichita et al., 2000; Harrison et al., 2000; High et al., 2000), the function of the three clusters remains mostly unknown, as well as the location of disulfide bridges in mammalian tyrosinases and Tyrps.
Post-translational processing of tyrosinase and its traffic through the secretory pathway are critical for the achievement of an active form of the enzyme. Proper folding of tyrosinase in the endoplasmic reticulum (ER) appears to be crucial for its further transport to the Golgi apparatus. This maturation is assisted by chaperones, calnexin being the most important one (Branza-Nichita et al., 1999, 2000; Halaban et al., 1997; Molinari and Helenius, 2000; Olivares et al., 2003). After further processing of the glycans in the trans-Golgi complex and copper acquisition, tyrosinase is delivered to stage II melanosomes via early endosomal intermediates (Hearing, 2005; Jimbow et al., 2000b; Raposo and Marks, 2007). It is still unclear how or when copper ions (necessary for enzymatic activity) are integrated into apo-tyrosinase. However, it has been demonstrated that the Menkes copper transporter, located in the TGN, is required for copper loading of tyrosinase and, therefore, the activation of the enzyme (Petris et al., 2000), although this event seems to be independent on protein folding (Olivares et al., 2003).
Mutations abolishing tyrosinase activity block the melanogenic pathway, thereby leading to oculocutaneous albinism type 1 (OCA1, OMIM 203100), an autosomal recessive disorder characterized by absence of pigment in hair, skin and eyes (Oetting, 2000). In addition to mutations in hot spots (copper-binding domains), virtually the entire coding sequence of the gene is susceptible to mutations. At present, more than 200 mutations in the tyrosinase gene have been associated with albinism. These include missense, nonsense, frameshift and splicing abnormalities (reviewed by Oetting et al. (2003); Albinism database: http://albinismdb.med.umn.edu/oca1mut.html). Mutant misfolded proteins are retained in the ER and routed for degradation by proteasomes (Halaban, 2002; Halaban et al., 2002; Toyofuku et al., 2001a,b).
The mechanism of catalysis of tyrosinase has been the object of intensive research because of the complexity to account for the peculiarities of the enzyme, such as the existence of two catalytic activities at the same active site and a lag period displayed by the cresolase activity which is dependent on the presence of catechols, the product of this reaction and substrate of the oxidase activity (Riley, 2000). Numerous enzymologists have studied tyrosinase, and it has been included recently as a model in some reviews about biologically inspired oxidation catalysis (Que and Tolman, 2008; Rosenzweig and Sazinsky, 2006).
The most accepted proposed reaction mechanism (Figure 2A) slightly refines former models (Lerch, 1983), and describes a common catalytic site for the two activities with three different forms of the enzyme, called met, oxy and deoxy, according to the absence/presence of oxygen and the oxidation state of the copper ions [Cu(II)/Cu(I)]. The oxy-form of tyrosinase and the appropriate substrates initiate the TH or DO cycles. This mechanism accounts for most kinetic features of the enzyme described so far and contains the evidence for the actual occurrence of physical differences in the catalytic requirements of TH and DO activities previously suggested (Hearing and Ekel, 1976; Tripathi et al., 1992). Some of them are firmly supported by structural evidences from the first crystal structure of tyrosinase published (Streptomyces castaneoglobisporus) (Matoba et al., 2006), which is a good model to interpret the function of some residues in the active site of mammalian tyrosinases because of the high degree of conservation within this region in all known tyrosinases (Figure 1B). Homology modelling of mouse tyrosinase based on this crystal structure has successfully provided a model for the active site which accounts for artificial loss-of-function forms of the protein and naturally occurring albino mutations (Schweikardt et al., 2007).
The high structural homology among the active site of type 3 copper proteins is shown up in the crystal structures of Streptomyces castaneoglobisporus tyrosinase (Matoba et al., 2006), Ipomoea batatas catechol oxidase (Klabunde et al., 1998), and a molluscan hemocyanin (Cuff et al., 1998). They reveal that these proteins have the second domain in common, which folds in a ‘four α-helix bundle’ motif bearing the active site with the two histidine-coordinated copper atoms, and the same relative positions of the six His residues, as we describe for tyrosinase above. In the oxy-form, one dioxygen molecule is coordinated between the coppers, each of which is ligated to the protein matrix by three histidine residues provided by an antiparallel α-helix pair, except for one His at the CuA motif which comes from a flexible loop (Inoue et al., 2008). That His (His54 in the S. castaneoglobisporus enzyme and His202 in the murine enzyme) has been proposed as the residue involved in the proton shift from the ortho position of the substrate tyrosine during the catalysis mediated by the Streptomyces enzyme (Inoue et al., 2008). Some reports on fungal tyrosinases suggest the participation of a cysteine residue as a ligand of CuA (Nakamura et al., 2000), but, if so, this feature is not universal for all tyrosinases. For instance, Streptomyces tyrosinase have no Cys at all, which rules out this possibility. Oxygen binding induces an oxidative change in the valency of the copper ions, which are in the Cu(I) state in the deoxy form, but become Cu(II) upon oxygenation. The Cu2(III,III) state has also been considered (Siegbahn, 2003), but energetically that state seems very unlikely. Furthermore, in Ipomoea batata catechol oxidase, the CuA site is shielded by an aromatic residue (Klabunde et al., 1998); interestingly, all plant and fungal catechol oxidases sequenced have an aromatic residue in the equivalent blocking position, and lack TH activity, while animal tyrosinases have no such bulky residue and they show TH activity. Therefore, it has been suggested that monophenols would dock to CuA but o-diphenols would dock to CuB at the tyrosinase active site (Olivares et al., 2002).
On the other hand, the crystal structures of Streptomyces castaneoglobisporus tyrosinase also leave open the possibility that monophenols bind to CuB (Inoue et al., 2008; Matoba et al., 2006). Other studies support this hypothesis, where both monophenols and o-diphenols coordinate to CuB in an orientation allowing for oxygen transfer to the ortho position in the first case and bridging the two Cu ions in the second one (Tepper et al., 2005). However, these authors used Streptomyces antibioticus tyrosinase for this work and p-nitrophenol as monophenolic substrate, which is a very poor substrate for mammalian tyrosinases. One must not exclude the possibility that p-nitrophenol goes into the tyrosinase active site in a different orientation than l-tyrosine, or that microbial and mammalian tyrosinases roughly share the active site but differ in some features concerning the necessity of a caddie protein in the Streptomyces enzyme and the existence of N-glycosylation in animal tyrosinases that can affect the accessibility, affinity and docking of substrates at the respective active sites during the catalytic cycle. Thus, the proposed role of the second His at the CuA motif could change among different tyrosinases in spite of being the unique flexible residue located in a loop structure of the protein in all tyrosinases (García-Borrón and Solano, 2002; Inoue et al., 2008). To this regard, a tyrosinase with an abnormally high tyrosine hydroxylase/dopa oxidase activities ratio from Ralstonia solanacearum has been characterized (Hernandez-Romero et al., 2006). This higher activity tyrosine hydroxylase versus dopa oxidase was discussed in terms of residues around the active site affecting the accessibility and orientation of the substrate. As far as we know, this type of comparative studies among microbial, plant and mammalian tyrosinases has never been made, and it can be an interesting goal for the future. This could be one of the aspects accounting for the significant differences concerning the turnover number and specificity of enzymes from different sources. Bioinformatics tools for this type of studies are available and have been recently used (Nithitanakool et al., 2009); unfortunately, crystallographic data for a mammalian tyrosinase are still missing.
Therefore, we will outline here the catalytic cycle of tyrosinase, according to our data for the mammalian enzymes. The TH cycle is initiated by the binding of l-tyrosine to the active site (Figure 2A). The structures mentioned before provide information on the possible orientation of monophenols at the active site, although this aspect would not be universal for all tyrosinases. In mammalian tyrosinases, the most probable orientation of l-tyrosine would be caused by the π–π interaction between the aromatic ring in the substrate and the second His in the CuB site, if we extrapolate the three-dimensional structure of the crystallized S. castaneoglobisporus tyrosinase (Matoba et al., 2006) to the mammalian enzymes based on the high degree of conservation within the active site. l-tyrosine is then ortho-hydroxylated and oxidized, and the o-quinone then leaves the reduced bicuprous site, allowing for the entrance of a new oxygen molecule. Note that even in the TH cycle, the releasing substrate is l-dopaquinone, and never l-dopa, as previously suggested (Cooksey et al., 1997). However, the non-enzymatic indirect formation of l-dopa by the chemical reactions accompanying dopaquinone cyclization leads to unusual enzyme kinetics (Land et al., 2003). In the DO cycle, the o-diphenol bound to CuB would also labilise the oxygen, resulting in the oxidation of the organic substrate and release of l-dopaquinone. Biochemical evidences prove that the His residue adjacent to the third copper ligand in the CuB binding site is involved in stereospecific binding of diphenols, but not monophenols (Olivares et al., 2002). The met- bicupric state of the enzyme can bind either l-tyrosine or l-dopa. The latter would dock to the two copper ions by both hydroxyl groups, with higher affinity than when the binding proceeds only through CuB. This aspect of l-dopa acting as a substrate (oxy-Tyr) or as a cofactor (met-Tyr) with different affinities for the enzyme was reported long ago, and it is approximately 100-fold higher as a cofactor than as a diphenolic substrate (Pomerantz and Warner, 1967). This is another characteristic that points at the existence of diverse mechanisms of catalysis between mammalian and non-mammalian tyrosinases. In any case, this orientation accelerates the oxidation of the substrate and the reduced deoxy-Tyr is reoxidized upon oxygen binding. Docking of l-tyrosine to met-Tyr at the catalytic center would lead to a dead-end complex (Figure 2A). The inhibition by l-tyrosine excess reported for tyrosinase (Jara et al., 1988) could be related to the formation of this dead-end complex, and even the possibility of binding two molecules of l-tyrosine to the copper ions has been suggested, but never proven.
Unlike tyrosinase, the mechanism of catalysis of dopachrome tautomerase has been poorly studied. Some reasons account for this scarcity of data, such as the facts that this enzyme was reported much later and it does not exist in lower organisms. Furthermore, its role in controlling melanogenesis is not as crucial as the function of tyrosinase.
The most important factor determining the catalysis at the active site of dopachrome tautomerase is the existence of a couple of zinc ions instead of copper (Solano et al., 1996). Each Zn(II) is expected to be bound to the protein moiety of the enzyme by the three histidine residues conserved in the metal sites MeA and MeB, probably forming a distorted tetrahedron, the characteristic geometry of this ion. In addition, Zn(II) has no redox properties, thus providing the active site of dopachrome tautomerase with the ability to catalyse isomerization reactions, such as a keto–enol tautomerization, opposite to the redox properties of copper ions in tyrosinase that allow for oxidizing reactions. According to these features, l-dopachrome would bind to the Zn(II) ions through the semiquinonic face of the indolic nucleus, displacing a water molecule occupying the fourth position of the distorted tetrahedron and keeping each ion bound to the polypeptidic chain by the three histidine residues (Figure 2B). The formation of the enzyme–substrate complex (dopachrome tautomerase-dopachrome) initiates an electronic rearrangement in the indolic ring leading to the formation of DHICA as the final product of the reaction. As far we know, mouse and human enzymes act on l-dopachrome stereospecifically to release DHICA as the unique product. However, there are a number of reports about other Dct-related enzymes, in mammals (Matsunaga et al., 1999; Odh et al., 1993) and lower organisms, such as cuttlefish (Palumbo et al., 1994) or insects (Sugumaran and Semensi, 1991), which are able to act on d-dopachrome or dopaminochrome to give DHI. The first type of these enzymes is not a tautomerase as it produces a decarboxylation, and the second type might act by a different mechanism, as dopaminochrome does not have stereospecificity. In any case, the stereospecificity of the mouse and human enzymes can be explained by the interaction of the carboxyl group of l-dopachrome with some residue differing the conserved histidines that would anchor the substrate at the active site with the appropriate orientation, similarly to tyrosinase with l-dopa (Olivares et al., 2002). Some attempts to identify this residue have been made (Aroca et al., 1991), but so far they have been unsuccessful.
Conclusion and perspectives
The present review has summarized the main structural and functional features of mammalian tyrosinase and its related proteins. We have compiled the current knowledge on the reaction mechanisms mostly accepted for tyrosinase and Dopachrome tautomerase. The mechanism of catalysis for tyrosinase has been deeply studied because of its complexity and singular properties. The presence of two different catalytic activities, tyrosine hydroxylase and catechol oxidase, with a unique active site, the characteristic lag period shown by the first activity and the different ratio between both activities in tyrosinases from different biological sources should be explained in a common mechanism valid for all of them. The catalytic mechanism of dopachrome tautomerase has been less studied, but it seems simpler to propose because of the nature of the zinc-containing active site and the reaction catalysed. Despite the lack of crystallographic data on mammalian tyrosinases and its related proteins because of its anchoring to the melanosomal membrane, the extrapolation of the recent publication of crystallographic data of a prokaryotic tyrosinase and a plant catechol oxidase have helped to complete the previously existing understanding of the catalytic cycle of tyrosinase, and in some cases to account for naturally occurring albino mutations. However, differences among microbial, plant and animal tyrosinases concerning relative hydroxylase versus oxidase activities are still unsolved, and it is likely related to subtle details in the mechanism of catalysis, such as the accessibility and first docking of the substrates to the dicopper center. Several other questions remain unanswered. The most striking ones refer to the different catalytic activities and substrate specificities of mammalian tyrosinase and the Tyrps and the uncertainties as to the actual role of Tyrp1. Besides some possible stabilization properties of Tyrp1 on tyrosinase, in terms of enzymology, Tyrp1 seems to be a poor tyrosinase with mostly dopa oxidase activity and only some residual hydroxylase activity under special conditions. In agreement with that, although the metal cofactor of Tyrp1 has never been unambiguously characterized, preliminary experiments using quimeric constructs of tyrosinase and Tyrp1 at the metal-binding sites indicate that copper is the metal cofactor not only of tyrosinase but also of Tyrp1. On the other hand, it is becoming increasingly clear that relevant information specifically related to mammalian melanosynthesis is missing, as this pathway is highly regulated in comparison to microbial and plant melanin pigmentation. Moreover, there is good evidence that a melanogenic complex is formed in vivo in mammalian melanosomes and most of our knowledge of these enzymes comes from studies performed with individual proteins. Therefore, although enzymology is not currently a cornerstone field because of the difficulty of taking functional data of the products concentrations and rate of reactions inside melanocytes and melanosomes, the prospects on mammalian tyrosinase and enzymatic control of melanogenesis go through advances in a comprehensive manner of this melanosomal complex.
Work by C. O. is sponsored by grant no. 464/2008 from the Plan de Ciencia y Tecnología, CARM, Spain. We thank Dr. S. Delgado for useful suggestions concerning the mechanism of dopachrome tautomerization and his help with the figures. In memoriam of José Ramón Jara.