Known Structures Of Multicopper Oxidases

Structures of three multicopper oxidases have been determined by X-ray crystallographic analysis: human ceruloplasmin (hCp), Coprinus cinereus (fungal) laccase (Lac), and Cucurbita pepo (zuccini) ascorbate oxidase (AO). hCp is the only ferroxidase whose structure has been determined crystallographically (7,8). Lindley and his coworkers have published a 3.1-A map (PDB accession number, 1KCW). This structure showed that as in Lac (1A65) (9) and AO (1AOZ) (10-12), the type-2 and type-3 copper atoms in hCp are trigonally arranged with an atom-to-atom spacing of approx 3.5 A. This "trinuclear" cluster is, in turn, approx 13 A from the type-1 copper atom. This arrangement of the three copper sites is diagrammed in Fig. 3 using ligand assignments for Fet3p. These assignments are based on the sequence homologies among the multicopper oxidases, as shown in Fig. 4. In all cases, the numbering is based on the encoded protein, not the mature processed one. All of the encoded proteins contain leader signal-recognition sequences that are cleaved during their maturation. As noted earlier, the type-1 Cu(II) is the redox partner of the reducing substrate. The trinuclear cluster, in

Table 1

Spin Hamiltonian and Absorbance Parameters for the Copper Sites in Fet3p

Copper site gM AN (X10-4 cm-1) g± Xmax (nm) e (M-1cm-1)

Type 1 2.19 89 2.05 608 5500

Type 2 2.24 195 2.05 Not resolvable, <100

Type 3 Diamagnetic 330 ~5000

Source: Data from refs. 4-6.

turn, is responsible for the four one-electron transfers to dioxygen that result in the production of two water molecules (13).

The structures of the type-1 sites in hCp, Lac, and AO—and in Fet3p—differ in some details; nonetheless, the electron-transfer reactions at this site in the four proteins are all outer sphere in nature (3). This mechanism is dictated by the fact that type-1 sites, as a class, do not have exchangeable solvent-accessible inner coordination sites to which a reductive ligand could bind. On the other hand, the relationship between the copper atom and the solvent-accessible surface of the protein does vary among these proteins. For example, the type-1 copper in hCp is completely buried (7), whereas this site in Coprinus cinereus Lac is much closer to the surface (9). This difference is apparent in Fig. 5, which displays Connelly surfaces adjacent to the type-1 sites in the two proteins. Connelly surfaces, in effect, show the solvent surface of a macromolecule. The green-hued surface patch in the Lac structure (Fig. 5, left) is the result of the NE of H494; clearly, the edge of this copper ligand is strongly exposed to solvent. In contrast, the corresponding histidine in hCp, H1045, is buried in the protein and completely shielded from solvent (Fig. 5, right). In both cases, the electron transfer from the substrate, irrespective of its nature, is outer sphere; however, given the striking difference in surface accessibility, the precise mechanism (route) for how the electron gets to the Cu(II) may be quite different in the two proteins.

5. ESEEM ANALYSIS OF FET3P COPPER-SITE STRUCTURE: THE TYPE-1 AND TYPE-2 CU(II) SITES

Although the structure of Fet3p is not known, electron spin-echo envelope modulation (ESEEM) data suggest that its type-1 copper has a solvent accessibility that is intermediate between the situation in hCp and Lac (14). ESEEM is a pulsed EPR technique that resolves spin interactions that are weak compared to the interaction of the electron spin magnetization with the instrument magnetic field or with those fields resulting from strongly interacting nuclei. Those are the interactions detected in the cwEPR spectrum as in Fig. 1. Thus, ESEEM that is the result of the weak interaction (approx 1 MHz at a field of 3 kG) of the distal, noncoordinating N in an histidine imidazole found at a Cu(II) site in a protein can be resolved easily (15). The ESEEM spectra of the unpaired electron at the type-

1 Cu(II) (in the T2D mutant) is shown in Fig. 6A (14). The depth and period of modulation of the electron spin magnetization can be correlated to the amount of spin density that is transferred from the Cu2+ to equatorially coordinated histidine imidazoles and the number of such ligands. The modulation depth of the ESEEM spectrum resulting from the type11 Cu(II) is therefore relatively weak because much of the spin density is on the S ligand of Cys484 at this site (1,2). The Fourier transform of the modulation envelope gives the modulation frequencies that generate it. This transform is shown in Fig. 6B and can be best fit with a model that includes two equatorial but inequivalent imidazole ligands at the type-1 Cu2+.

The ESEEM spectrum of the type-2 Cu2+ (in the T1D mutant) is substantially different, both because at this site there is significantly more spin density at the copper (no charge transfer involving

Fig. 3. Spatial relationship among the three copper sites in a multicopper oxidase. The structure (and residue numbering) is a representation of these sites in hCp. The trinuclear cluster is a near-isosceles triangle, approx 3.4 A on a side. The notation Cu(2) and Cu(3) is taken from ref. 10 and corresponds to the notation Type 3'' and

Fig. 3. Spatial relationship among the three copper sites in a multicopper oxidase. The structure (and residue numbering) is a representation of these sites in hCp. The trinuclear cluster is a near-isosceles triangle, approx 3.4 A on a side. The notation Cu(2) and Cu(3) is taken from ref. 10 and corresponds to the notation Type 3'' and

Type 3', respectively, used by Zaitseva et al. (7). Oxygen is thought to bind between the type-2 copper and Cu(

a cysteine sulfur) and because this copper atom is directly coordinated by solvent water. The first of these differences is apparent in the depth of modulation, as seen in Fig. 7A. The second is apparent in the weaker modulation seen at longer times. This is the result of associated with bound water molecules. The Fourier transform of the ESEEM data is given in Fig. 7B and is best fit by assuming one equatorial imidazole and, perhaps, a second imidazole ligand coordinated axially.

The spectra in Figs. 6 and 7 were taken in H2O and therefore do have the potential to show a modulation resulting from :H of any water that might be coordinated to or magnetically "near" the type-2 Cu(II) as well. This modulation was well resolved in the type-2 Cu(II) spectrum as noted but was essentially absent in the type 1-Cu(II) one. In order to better resolve this modulation, however, one takes spectra in 1H2O and 2H2O and calculates the ratio of the experimental modulation patterns: in effect, the pattern in deuterium oxide provides the negative control for water proton modulation (16). Again, the strength and pattern of this ratioed modulation can be interpreted in terms of how water might be coordinated to or how it might magnetically interact with the Cu(II). This is illustrated in the theoretical traces given in Fig. 8A,B for one equatorial water, or for one axial water, or for ambient water only ("outer-sphere" water). These three traces can be compared to the experimental trace for the ratioed envelope for the type-1 site in Fet3p. This experimental trace can be best modeled by a Cu(II) that has a "half-shell" of ambient water only, with no directly coordinated water molecules (Fig. 8D). The fit that is shown is for a model that has this "hemisphere" of water molecules at 3.75 A from the type-1 Cu(II). In comparison to the known structures of Lac and hCp (Fig. 5), this result

Lac 142 Sei"-1 ie-Fel3 79 Ser-Met-hCp IIS Thr-Phe-AO 58 Val-lle-

Lac 187 Trp-Tyr-Fet3 124 Trp-Tvr-hCp 178 Ile-Tvr-AO 101 Phe-Tvr-

Lac 475 Fcü 411 hCp 992 AO 443

Leu-Pro-Gly-Thr-Asp-Leu-Glu-Thr-

His His His His

His-His-His-His

His His His His

Trp-Phe-Ser-Trp-

His His His His

Scr-Ser-Ser-

His His-

His-His

Gly-Mct GÏv-Leu Giv-Ue Gly-Ile

Phc-Ser Thr-Asp lie-Asp - Leu-Gly

0 0

Post a comment