It

Fig. 10. apo-CopZ conformers. Bundle of 20 conformers with the lowest residual target function. The orientation of the molecule is the same as in Fig. 11.
Fig. 11. Ribbon diagram of apo-CopZ. apo-CopZ with the lowest residual target function is shown, illustrating the secondary structure elements: aj (14-24), a2 (51-59), P1 (2-7), P2 (28-34), P3 (39-44), and P4 (64-67).
Fig. 12. Copper-binding residues of CopZ. The six best conformers of CopZ are represented by the mean of the backbone coordinates and by a superposition of the two copper-binding cysteine side-chain arrangements in the six conformers.

structural changes of a larger part of the protein, whereas no prolines are present in CopZ. A detailed structural comparison between Atx1 and MerP and mbd4, respectively, can be found in ref. 47.

4.3. Structural Changes of CopZ upon Interaction with Cu(I)

CopZ undergoes significant changes upon interaction with copper(I). Whereas in apo-CopZ, all but a few *H-NMR signals were observable, the signals of residues 11-21 were missing in the NMR spectra of Cu(I)CopZ. Some weak signals were visible, but because of missing NOEs, they could not be assigned unambiguously. Paramagnetic ions [e.g., Cu(II)], could cause a disappearance of the NMR signals in their vicinity, but electron paramagnetic resonance (EPR) measurements showed that no Cu(II) was contained in the sample. Hence, the disappearance of the signals was ascribed to conformational exchange between two or—presumably—more conformations. This coincides with the findings made with Cu(I)Atx1 (47) in the crystal form. Studies of Cu(I)Atx1 in solution revealed a similar behavior only when the protein concentration exceeded 2 mM; otherwise, a well-defined binding site could be observed (48).

A comparison of the backbone 15N, Hw, and Ha chemical shifts revealed that apart from the metal-binding loop and the first helix, no structural changes occurred upon Cu(I) binding (see Fig. 13). This was corroborated by careful examination of the NOESY spectra of Cu(I)CopZ and comparing them to the NOESY spectra of apo-CopZ. Essentially the same NOEs could be found in both spectra and a structure calculation on Cu(I)CopZ yielded an identical structure to that of CopZ—except for the part of the protein, where no signals could be observed. X-ray absorption studies suggested a mixture of 75% diagonally coordinated/25% triagonally coordinated copper for Cu(I)-CopZ, with all ligands being sulfur atoms (49). The origin of the third contributing ligand in addition to Cys-11 and Cys-14 remained unknown. NMR data showed, clearly, that neither of the additional sulfur atoms in CopZ

Fig. 13. Structural differences between apo- and Cu-CopZ. Differences between corresponding backbone chemical shifts in apo-CopZ and Cu(I)-CopZ plotted versus the sequence. Amide protons A8(HN) (a), amide nitrogens A8 15N) (b), a-protons A8(Ha) (c) in the case of glycines the pair of Ha lines with the greatest difference was chosen, where A8=8(apo-CopZ) - 8(Cu(I)-CopZ). (Reprinted with permission from ref. 44.)

Fig. 13. Structural differences between apo- and Cu-CopZ. Differences between corresponding backbone chemical shifts in apo-CopZ and Cu(I)-CopZ plotted versus the sequence. Amide protons A8(HN) (a), amide nitrogens A8 15N) (b), a-protons A8(Ha) (c) in the case of glycines the pair of Ha lines with the greatest difference was chosen, where A8=8(apo-CopZ) - 8(Cu(I)-CopZ). (Reprinted with permission from ref. 44.)

takes part in copper binding. The SHY of Cys-55 can be observed in the NMR spectra of both apo- and Cu(I)CopZ. The Hffi and Cffi of Met-9 does not show a significant difference in chemical shift between apo- and Cu(I)CopZ.

Determination of NMR relaxation times finally explained the origin of the third ligand. As can be seen from Fig. 14, the longitudinal relaxation time T1 increased upon copper binding whereas the transverse relaxation time T2 decreased. This is indicative of a decreased molecular tumbling, which in the experimental setup used can only be explained by aggregation—presumably dimerization. These findings were corroborated by light-scattering measurements on a different set of samples. Thus, the third ligand for copper binding comes most likely from a different CopZ molecule. It is, however, possible, that under biological conditions, a small thiol-containing molecule like glutathione or cysteine plays the role as a third ligand.

4.4. Comparison of the CopZ with Other Metal Chaperones

The structures of metal chaperones homologous to CopZ have also been investigated with bound metal ligands. Although the structures and functions of these proteins seem to be very similar, a different behavior upon ligand binding was observed. Table 1 gives an overview over the data reported so far. In general, copper is bound by three sulfur ligands, one of which has a longer distance to the copper ion than the other two. The stoichiometry of copper:protein has been reported to be 1:1 for CopZ (43) and for Hah1 (47), but 0.6-0.8 for Atx1 (50). A well-defined structure of the metal-binding site was obtained only for Hah1 (47) and Atx1 in solution at concentrations below 2 mM (48). CopZ (44) and Atx 1 in the crystal (47) exhibited disordered metal-binding sites.

The structure of Cu(I)Hah1, however, would suggest a stoichiometry copper:protein 1:2 rather than 1:1 found experimentally. CopZ, too, showed aggregation—presumably dimerization—upon interaction with copper. Unlike in X-ray crystallography, the Cu ion cannot be observed directly by

10 20 30 40 50 60 amino acid sequence

Fig. 14. NMR relaxation times for apo- and Cu-CopZ. Relaxation times and steady-state 15N{1H} NOEs measured for the backbone amide nitrogen atoms of apo-CopZ (□) and Cu(I)-CopZ (♦). (A) Tj/T2 recorded at a 15N frequency of 50.7 MHz; (B) longitudinal relaxation time T1; (C) transverse relaxation time T2; (D) 15N{ JH}-NOEs recorded at a 15N frequency of 60.8 MHz. For Cu(I)-CopZ no measurements were obtained for the residues 11-20 (see text). (Reprinted with permission from ref. 44.)

10 20 30 40 50 60 amino acid sequence

Fig. 14. NMR relaxation times for apo- and Cu-CopZ. Relaxation times and steady-state 15N{1H} NOEs measured for the backbone amide nitrogen atoms of apo-CopZ (□) and Cu(I)-CopZ (♦). (A) Tj/T2 recorded at a 15N frequency of 50.7 MHz; (B) longitudinal relaxation time T1; (C) transverse relaxation time T2; (D) 15N{ JH}-NOEs recorded at a 15N frequency of 60.8 MHz. For Cu(I)-CopZ no measurements were obtained for the residues 11-20 (see text). (Reprinted with permission from ref. 44.)

solution NMR. Therefore, nothing definite about the number of copper ions in the CopZ dimers can be said. Cu-Cu scatter peaks in Cu(I)mbd2, however, suggest a copper-binding site with at least two copper ions for this protein domain, which would fit a dimeric state with a 1:1 stoichiometry, but no structural information for mbd2 is available.

Solutions of Cu(I)Atx1 in concentrations higher than 2 mM exhibited a loss of NMR signals from the metal-binding loop and a part of the following helix (47) similar to what was found for Cu(I)CopZ. The formation of precipitate was reported under these conditions, but the aggregation state of the protein remained uncharacterized. In the case of CopZ, the NMR signals of the metal-binding loop and the first two turns of the following helix disappeared; in Cu(I)Atx1 at high concentrations, only the NMR signals of the metal-binding loop and two residues of the first helix turn vanish.

For both the Cu(I)Hahl crystals and the Cu(I)CopZ solution, it has been stated that the observed aggregation/dimerization could be an artifact introduced by the non-physiological sample conditions.

Table 1

Overview of Structural Work Reported on CopZ and Similar Metal Chaperones

Table 1

Overview of Structural Work Reported on CopZ and Similar Metal Chaperones

Protein/

Natural

0 0

Post a comment