Phenylethylamine Oxidase
Created by Nona Jiang
Phenylethylamine oxidase from Arthrobacter globiformis (AGAO) is a copper-containing amine oxidase (1). Copper amine oxidases catalyze the two-electron oxidative deamination of primary amines to the corresponding aldehyde, ammonia, and hydrogen peroxide. Since these enzymes have many essential functions in the metabolism of biogenic primary amines, the structural characterization of copper amine oxidases is extremely important (2). Copper amine oxidases have also been found in many different organisms and have been purified and characterized in a number of organisms, both prokaryotic and eukaryotic. In Arthrobacter globiformis, the function of phenylethylamine oxidase is to oxidize amine substrates so that they can be used as alternative sources of carbon and nitrogen for the bacteria. Therefore, the main function of phenylethylamine oxidase in Arthrobacter globiformis is to support growth (3).
Studying the structure of phenylethylamine oxidase can give us much insight into its function. Phenylethylamine oxidase exists physiologically as a disulfide-linked mushroom-shaped homodimer with dimensions of 110 x 65 x 45 Å (2). The molecular weight of each subunit of phenylethylamine oxidase is 71677.04 Da and its isoelectric point (pI) is 5.07. Each subunit of the homodimer is composed of three domains: D2, D3, D4, with D4 being the catalytic domain. The domain D4 consists of a large 18-stranded β-sandwich and extends from residue 229-623. Domain D2 is composed of residues 9-91 and domain D3 is composed of residues 103-203. Both domains D2 and D3 can be found on the surface of D4 (2).
The D4 domain constitutes a large part of the dimer interface. Two long β-ribbon arms extend from each D4 domain attaching to the other D4 domain of the other monomer. Arm I (residues 348-369) lies along the bottom of D4 while Arm II (residues 453-474) lies along the top of the D4 domain, on the surface. The functional importance of these two arms is still unknown. Wilce et al. suggests that a potential purpose of these two arms is to regulate access of substrates to the active site while contributing to the stability of the dimer (2). Domain D4 is also known as the catalytic domain as the active site of each subunit is located within D4, between the two arms from the other subunit and close to the dimer interface (3).
Each subunit of the homodimer contains a copper ion and is covalently bound to 2,4,5-trihydroxyphenylalaninequinone (topaquinone, TPQ), an organic redox cofactor (4). The ligands which bind to the copper (II) ion and TPQ itself are all found within adjacent β-sheets in domain D4. TPQ is formed through post-translational modification of a conserved tyrosine residue. In phenylethylamine oxidase, this is Tyr-382. In the active form of the holoenzyme, the copper ion is coordinated with the imidazole groups of three histidine residues, His-431, His-433 and His-592, along with two water molecules in approximately a square pyramidal geometry. The TPQ cofactor is close by to the copper (II) ion but is not coordinated to it. The “equatorial’ ligands consist of the three histidines while the other water is the “axial” ligand. In the inactive form of the enzyme, the copper (II) ion is coordinated by the same three histidine residues along with the phenolic oxygen of the TPQ in an approximately trigonal-pyramidal geometry (2).
Phenylethylamine oxidase in Arthrobacter globiformis catalyzes three reactions: the biogenesis of TPQ, the oxidation of amine substrates to generate reduced TPQ (reductive half-reaction), and the reduction of molecular oxygen by reduced TPQ (oxidative half-reaction). Molecular oxygen is required for both the biogenesis reaction and the oxidative half-reaction (3, 8). There are three main forms of AGAO: the apoenzyme which contains the precursor tyrosine to the TPQ cofactor and no bound copper, the mature native holoenzyme containing both TPQ and copper, and the substrate reduced form of the enzyme with the aminoquinol form of cofactor and copper (2, 3).
The apo-form of the enzyme is the starting structure for AGAO in TPQ biogenesis. In the apo-form, there are four copper ligands, His-431, His-433, His-592, and Tyr-382. Tyr-382 is the residue that will become TPQ after biogenesis (2). These four residues bind to the copper (II) ion in a tetrahedral geometry. In the initial state of the apoenzyme, the copper ion is not yet bound to the protein, but the four copper ligand residues are arranged around an empty metal center, ready to bind. The two water molecules that also bind to the copper ion in the holoenzyme are absent here in the apoenzyme (3).
The first step in TPQ biogenesis is the binding of copper to apo-AGAO in the vacant metal binding site. After binding, the copper is located in the same position as it would be in the holoenzyme (6). The precursor tyrosine, Tyr-382, at this stage is thought to be protonated. Two of the three histidine residues which bind copper, His-431 and His-433, are in identical positions in both the apoenzyme and holoenzyme forms. However, His-592 can be found in either of two conformations, which suggests that this residue is more flexible than His-431 or His-433. The lengths of each Cu(II)-His(N) bond are all approximately the same, around 2.0 Å. This suggests that while His-592 is more flexible, its bonding interaction with copper is not any weaker (3). The tyrosine precursor, Tyr-382, in the anaeorobic apoenzyme is also thought to be coordinated somehow with the copper (7).
After binding copper, the amine oxidase then binds molecular oxygen. The previously protonated Tyr-382 is then deprotonated and forms a Cu(II)-tyrosinate complex. No other metal atom will significantly support TPQ biogensis, indicating that copper has a crucial role in this process, though its role is still not completely understood. One theory is that the copper plays a key role by forming a transient Cu(I)-tyrosine radical intermediate. In order words, the copper is needed to from a species that reacts with molecular oxygen in order to form a peroxo-bridged adduct. When this adduct breaks down it yields the oxidized quinine form (3,4-dihydroxyphenylalanine, DPQ), and copper (II) hydroxide (3, 5). In order to proceed from the DPQ intermediate, the DPQ ring must rotate to bring the C6 closer to the copper center in order for TPQ biogenesis to continue. When this happens, C3 of DPQ becomes C5in TPW, C6 in DPQ becomes C2 in TPQ, and C6 of DPQ is positioned for attack by the copper bound hydroxide. Since C5 in TPQ corresponds with C3 in DPQ, there appears to be a 180 degree rotation that took place. When the TPQ is first formed, it is in a reduced state. At this stage, there is a hydrogen bond which exists between the oxygen atom at C2 and the backbone carbonyl oxygen of Thr-402, suggesting that the C2 oxygen atom is protonated. This is expected if TPQ is indeed reduced (6).
The final step of this biogenesis reaction is to oxidize TPQ to yield the holoenzyme. This step also requires molecular oxygen, which oxidizes reduced TPQ and yields hydrogen peroxide. Following oxidation of TPQ, the three histidine ligands remain bound to the copper in the same geometry as in DPQ. However, when TPQ is oxidized, the TPQ O4 atom is hydrogen bonded to Tyr-284. This means that the O4 atom is not in the axial copper ligand position as is observed in all of the early reaction intermediates. Another key difference in this oxidized form of the enzyme is that two water molecules are now bound to copper. One of the molecules is bound in an axial position ~1.9 Å from the copper, while the other is bound in an equatorial position ~2.7 Å from the copper (6).
This self-processing of the tyrosine precursor to TPQ in phenylethylamine oxidase demonstrates the exact control of protein motion during catalysis. While recent experiments have provided information about key intermediates in this reaction, several key questions remain. For example, it is not clearly known what drives the 180-degree rotation of DPQ in the active site. Questions of timing and geometry as well as questions on the structural details of this particular rotation still remain (3).
The second reaction that phenylethylamine oxidase catalyzes is the oxidation of amine substrates to generate reduced TPQ and is also known as the reductive half-reaction. The first step of the reductive half-reaction involves the nucleophilic attack of the primary amine substrate at C5 of oxidized TPQ. This results in the formation of a substrate Schiff base. Here, the ligand binding site is shown by ruthenium wire. After the formation of a substrate Schiff base, a conserved aspartic acid, Asp-383, abstracts a protein to form a carbanionic species, which quickly rearranges to form a product Schiff base. The product Schiff base then undergoes hydrolysis to form the product aldehyde. The oxidized TPQ has now effectively been reduced to the aminoquinol state, with the amine nitrogen replacing the oxygen bound to C5 of the cofactor. The end result is that TPQ is reduced by two electrons forming the reduced form of the holoenzyme with the aminoquinol form of cofactor and copper (II). This series of reactions is collectively known as the reductive half-reaction in phenylethylamine oxidase. This reaction mechanism is both well understood and widely accepted (3).
The reduced form of the holoenzyme at the end of the reductive half-reaction directly leads to the oxidative half-reaction. In many ways, this aminoquinol state at the end of the reductive half-reaction can be viewed as the first intermediate of the oxidative half-reaction. During this reaction, molecular oxygen again plays a crucial role in the mechanism. Molecular oxygen first binds to the enzyme and accepts two electrons and two protons from the aminoquinol. This yields hydrogen peroxide and an iminoquinone intermediate. Ammonia is released during the hydrolysis of the iminoquinone and oxidized TPQ is regenerated. This oxidized TPQ is recycled and can participate in another catalytic cycle (3).
However, unlike the reductive half-reaction, the oxidative half-reaction is not well understood and is still being studied. A hotly contested issue is what species the molecular oxygen reacts with. The Cu(II)-aminoquinol state of the enzyme and the Cu(I)-semiquinone species exist in equilibrium in solution. However, how much Cu(I)-semiquinone observed can depend on the source of the enzyme and can sometimes even be undetectable. Therefore, it has been debated whether or not Cu(I)-semiquinone is even in the catalytic pathway. It has been previously proposed that copper (I) from this species would donate the first electron to molecular oxygen during the oxidative half-reaction (9). However, several more recent studies have claimed that this copper does not change oxidation state during the half-reaction (10).
A few structures from the oxidative half-reaction have been solved which give a bit more insight into the binding of molecular oxygen in this half-reaction. Catalytically competent crystals that were anaerobically substrate reduced in the presence of nitric oxide and without nitric oxide were flash-frozen. One structure showed that when the same anaerobic conditions are retained, but nitric oxide is added to mimic dioxygen, the protein will bind the nitric oxide. It is observed that the nitric oxide replaces the axial water, which bridges between the copper and aminoquinol’s protonated O2. It appears that the NO is in its neutral form with an observed angle of 117 degrees to the axial position of copper and therefore is most closely mimicking molecular oxygen. The structure shows that nitric oxide is strongly interacting with the aminoquinol rather than the copper. Overall, this demonstrates that a molecular oxygen mimic can preferentially bind to the aminoquinol in the active site while still being in contact with the copper ion. This structure gives evidence that the first electron transfer to molecular oxygen in this half-reaction is from either copper (I) or aminoquinol (11).
Another intermediate structure revealed that molecular oxygen occupies the same site as nitric oxide, only with a different orientation. In this structure, it was seen that the ends of the dioxygen are 2.8 Å and 3.0 Å from the copper ion. These are fairly large intermolecular distances, and they hint at the peroxide product form of dioxygen (two-electron reduced) and the oxidized iminoquinone state of the cofactor. A simple electrostatic role of copper is also hypothesized from these large distances (11). While the previous structure is consistent with the theory that both electron transfers are principally from the cofactor, they do not rule out that the initial electron transfer that molecular oxygen initially undergoes could be from copper(I). In short, these structures give us clues, but the details of the initial electron transfer event to molecular oxygen in the oxidative half-reaction are still highly controversial and require further research.
The active site of phenylethylamine oxidase in domain D4 requires a conformational change to allow the substrate access. In an experiment conducted by Langley et al., they were able to observe one particular residue, Tyr-296 serving as a “gate” to the active site. The gate to the active site is open when Tyr-296 stays in a hydrophobic depression between residues Leu-137 and Gln-294. In native phenylethaylamine oxidase proteins, the default state is that the gate is closed. The hydroxyl group in Tyr-296 is hydrogen-bonded to one of a chain of four water molecules in the active-site pocket and channel. When the gate is open, an additional water molecule occupies the space, which was previously occupied by the aromatic ring of Tyr-296 (1).
The secondary structure of phenylethylamine oxidase contains beta sheets, alpha helices, and random coils. The majority of the beta sheets are found in the D4 domain with 18 of twisted in a large sandwich. The other two domains, D2 and D3, consist of alpha helices and random coils, with both of them lying on the surface of domain D4.
Phenylethylamine oxidase has an approximate 94% sequence similarity to its homologous protein in Escherichia coli (shown in blue). Both amine oxidases have both primary and tertiary similarities as indicted by the results of DALI (Z=46.0, rmsd=1.7) and protein Blast (E=1e-83) searches (12). The folding patterns in both proteins are also highly similar, with the α/β-roll topology being conserved. The function of amine oxidase in E. coli is virtually identical to that in A. globiformis, though there are subtle changes in structure. While the copper amine oxidase in E. coli (ECAO) is also a mushroom-shaped homodimer consisting of two subunits, the subunits in ECAO have four domains while there are only three domains in AGAO. This extra domain is an amino-terminal domain called D1 and forms the stalk of the protein (8). However, the biological function and significance of this domain is still not known. Another difference is that the positions of the D2 and D3 domains in relation to D4 are slightly different. Other than these structural differences, functionally this protein catalyzes the same reactions in Escherichia coli. The three histidine residues which bind copper in ECAO are His-524, His-526, and His-689. The conserved aspartic acid which abstracts a proton to form a product Schiff base in the reductive half-reaction, is Asp383 as well (8). Like ACAO, ECAO is responsible for oxidizing primary amines in the bacteria in order to utilize these amines for alternative forms of carbon and nitrogen.