Structural Features of the ATP7A and ATP7B Proteins

The sequence of ATP7A predicts a 1500-amino-acid polypeptide and characteristic structural features identify it as a member of the extended family of transmembrane proteins known as P-type ATPases (92). There are at least 80 known P-type ATPases found in all life forms. These enzymes are primarily involved in the ATP hydrolysis-dependent transmembrane movement of various cations, including H+, Na+, K+, Ca2+, and Mg2+ (112). The term "P-type" refers to the existence of a high-energy phosphoryl-enzyme intermediate formed during the reaction cycle of ion translocation. The site of phosphorylation is an invariant aspartic acid in the sequence DKTG. ATP7B is predicted to be 1411 amino acids in length and is very similar to ATP7A with about 57% amino acid identity (94). ATP7A and ATP7B are more similar to the bacterial heavy-metal-transporting P-type ATPases than to the classical mammalian ATPases that transport the alkali metals ions such as Ca2+ and Na+. The cadmium transporter CadA has a very similar predicted structure to that of the Menkes protein (113) and bacterial copper-transporting P-type ATPases, CopA and CopB, have been described (114,115) as well as copper ATPases in yeast (116,117). Hydrophobicity analysis of MNK predicts eight transmembrane domains (92,118). The heavy-metal-transporting ATPases are distinct from the classical P-type ATPases in both membrane topology and primary sequence features. They have been classified as a subclass of the P-type ATPases, called P1-type ATPases or CPx type ATPases (119,120).

Fig. 6. Hypothetical three-dimensional representation of ATP7A/B based on the three-dimensional structure of the calcium ATPases (121). The eight transmembrane domains are shown as cylinders traversing the lipid bilayer and forming a channel that allows copper to traverse the membrane. On the cytoplasmic side of the transmembrane channel, the cysteine residues in the CPC motif are thought to accept copper prior to translocation across the membrane. This copper is shown as coming from the N-terminal copper-binding sites, but the exact details of this transfer is unclear, and copper may also be directly available to the channel if concentrations are high in the cell. The metal-binding sites receive copper from the cytoplasmic chaperone ATOX1. The ATP-binding site in the large cytoplasmic loop is indicated (ATP) and the aspartic acid (D), which is phospho-rylated, is shown as part of a P-pleated-sheet. The enzyme contains an endogenous phosphatase indicated by the motif ITGEA.

Fig. 6. Hypothetical three-dimensional representation of ATP7A/B based on the three-dimensional structure of the calcium ATPases (121). The eight transmembrane domains are shown as cylinders traversing the lipid bilayer and forming a channel that allows copper to traverse the membrane. On the cytoplasmic side of the transmembrane channel, the cysteine residues in the CPC motif are thought to accept copper prior to translocation across the membrane. This copper is shown as coming from the N-terminal copper-binding sites, but the exact details of this transfer is unclear, and copper may also be directly available to the channel if concentrations are high in the cell. The metal-binding sites receive copper from the cytoplasmic chaperone ATOX1. The ATP-binding site in the large cytoplasmic loop is indicated (ATP) and the aspartic acid (D), which is phospho-rylated, is shown as part of a P-pleated-sheet. The enzyme contains an endogenous phosphatase indicated by the motif ITGEA.

A hypothetical three dimensional representation of MNK is shown in Fig. 6, based on the recently described three-dimensional structure of the Ca2+ ATPase (121). The eight transmembrane domains form a channel through which the copper traverses the membrane. As noted, all P-type ATPases have an ATP-binding site and an invariant aspartic acid residue (D in Fig. 6), which is phosphorylated and dephosphorylated during the reaction cycle of cation translocation. The CPC motif characteristic of heavy-metal P-type ATPases is indicated on the cytoplasmic edge of the channel. The cysteines are likely to bind copper during its translocation and are essential for the enzyme activity (122). The proline is found in all P-type ATPases and is important for the conformational changes involved in cation transport. Evidence that ATP7A functions as a P-type ATPase was obtained using a vesicle assay for copper transport (123), which showed that Cu transport was inhibited by orthovanadate, a specific inhibitor of P-type ATPases. The reaction only occurred in the presence of dithiothreitol, supporting the notion that Cu(I) is the ionic form of copper that is translocated. A major feature of the heavy-metal P-type ATPases is the presence of metal-binding sites (MBS) with a canonical sequence MTCXXC in the N-terminal region. ATP7A/B have six of these MBSs and the possible roles that these domains in accepting copper from the chaperone ATOX1 and regulating the trafficking of ATP7/B proteins is discussed next.

Low affinity C"*1) binding

Fig. 7. Model of the conformational changes in ATP7A/B involved in Cu transport. The P-type ATPases undergo marked conformational changes associated with cation transport. Shown here are two basic forms of the enzyme, with copper bound to the high-affinity cytoplasmic site (CPC) and ATP bound to the enzyme (left diagram). Binding of copper to the metal-binding sites may induce conformational changes that facilitate the transfer of copper to the CPC site. Following phosphorylation of the aspartic acid, the high-affinity site is occluded and the copper is transported to the other side of the membrane, where it is bound to a low-affinity site (shown as methionines and histidines). The aspartic acid is dephosphorylated by the endogenous phosphatase and the enzyme returns to the original state.

Low affinity C"*1) binding

Fig. 7. Model of the conformational changes in ATP7A/B involved in Cu transport. The P-type ATPases undergo marked conformational changes associated with cation transport. Shown here are two basic forms of the enzyme, with copper bound to the high-affinity cytoplasmic site (CPC) and ATP bound to the enzyme (left diagram). Binding of copper to the metal-binding sites may induce conformational changes that facilitate the transfer of copper to the CPC site. Following phosphorylation of the aspartic acid, the high-affinity site is occluded and the copper is transported to the other side of the membrane, where it is bound to a low-affinity site (shown as methionines and histidines). The aspartic acid is dephosphorylated by the endogenous phosphatase and the enzyme returns to the original state.

Much of the understanding of the complex mechanism of ion transport catalyzed by these enzymes has come from the study of the calcium ATPases. This work has now reached the stage of an X-ray crystal structure at 2.6 A resolution (121). Although the topology of the Cpx ATPases and the calcium ATPases differ, in that there are 8 transmembrane channels proposed for the former and 10 have been demonstrated in the latter, the fundamental ATPases core and reaction cycle are conserved. Thus, it is instructive to compare the two, particularly as the three-dimensional structure of the heavy-metal ATPases is not known. It is clear that P-type ATPases undergo considerable conformational changes during the reaction cycle, which can be distinguished by affinity for the metal ion, and various fluorescent changes. We have attempted to represent in Fig. 7the type of changes that may be involved. The binding of the cation to a high-affinity cytoplasmic site (CPC) allows the phosphoryla-tion of the aspartic acid to form the phosphoenzyme, which has a distinct conformation that occludes the high-affinity cytoplasmic binding site and reveals a lower-affinity binding site on the luminal side. In this way, the cation is transferred across the membrane. The internal phosphatase removes the phosphate and the enzyme returns to the intial conformation, where the high-affinity binding site is exposed to the cytoplasm (124). Commonly, only two conformational states are distinguished for P-type ATPases, E1 and E2; in the E1 state, the cation has access to the high-affinity cation-binding sites; for ATP7A, this the CPC in the channel and the corresponding region in the calcium ATPase is VAAIPE-309 forming the type-2 calcium binding site (121). In the E2 state following the substantial conformation change consequent to the phosphorylation, the high-affinity site is occluded, and the metal is bound to the low-affinity sites (125). For ATP7A, this low-affinity site could be the cluster of methionines and histidines on the luminal side (Fig. 7). The simplistic E1/E2 model has been criticized, in that the enzymes will certainly adopt more than just one conformation during the complete reaction cycle (124). A deeper understanding of the conformational changes is necessary when one attempts to understand the various disease phenotypes by consideration of the effects of disease-causing mutations at a molecular level and the mechanism of copper-induced trafficking of the copper ATPases, a mechanism that is a key part of cellular copper homeostasis. This process appears intimately tied to a unique feature of the mammalian Cu-ATPases, the six cytoplasmic copper-binding sites in the N-terminal region of the molecule.

Six copies of the metal-binding-site motif, MTCXXC are found in the N-terminal portion of MNK and have been shown to bind copper in vivo and in vitro (126-128) and a three-dimensional solution structure of MBS4 of ATP7A has been determined (129). Interestingly, only one to three of these MBSs are present in the bacterial and yeast copper-ATPases (115-117,130) and in Cad A. Several studies have investigated the role of these MBSs in copper transport and there is significant disagreement as to whether they are essential for copper transport. Payne and Gitlin showed by progressive mutagenesis of the CXXC motifs of ATP7A to SXXS, a motif that cannot bind copper, that mutation of the first three MBSs abolished copper transport despite the presence of intact MBS4-6, suggesting that the first three MBS had a critical role (131). In contrast, Forbes and Cox, using the same yeast complementation assay, found that only MBS6 of ATP7B was needed for copper-transport activity (132). These conflicting results might indicate that there are fundamental differences between ATP7A and ATP7B in the way in which the six copper-binding sites function (132). Although this seems inherently unlikely in view of the close relationship of the two proteins, there are differences between the two proteins in the N-terminal domain; for example, there is a 78-amino-acid deletion between the first and second metal-binding site in ATP7B relative to ATP7A. The effect of mutation of the copper-binding sites was assessed using an assay for Cu-ATPases in mammalian cell vesicles (133). This work clearly demonstrated that Cu transport into vesicles did not require any of the MBSs; moreover, in whole cells expressing the mutant with all six Cu sites mutated, copper accumulated into vesicles. Previous work had demonstrated that this particular mutant ATP7A had lost its Cu-induced trafficking ability (134). Thus, it appeared that copper accumulation was the result of an active transporter pumping copper into vesicles, but which could not subsequently traffic to the plasma membrane to allow copper release from the cell (133).

At least a partial resolution to the discrepancies noted above has come from studies on the interaction of the copper chaperone, Atx1 in yeast and ATOX1 in mammals, with the N-terminal Cu-binding sites. Interestingly, these chaperones closely resemble a single copper-binding domain of the Cu-ATPases, having the same GMTCXXC motif. Atx1 has been shown to bind copper in a two-or three-coordinate metal cluster, and a direct interaction between Atx1 and the cytosolic domains of the yeast ortholog of ATP7A/B was demonstrated (30). Thus, Atx1 accepts copper from the plasma membrane transporter, Ctr1, and transports the metal through the cytoplasm to CCC2, where the copper is transferred to the Cu site of CCC2 by a series of rapid associative changes involving two-and three-coordinate copper (30). A model has been proposed in which the copper chaperones transfer copper to the N-terminal copper sites and the copper is subsequently donated to the metal at the CPC site in the channel (132). Larin et al. (135) demonstrated that ATOX1 was capable of independent interaction with individual MBSs. MBS1-4 appeared to interact with roughly equal efficiency, but the last two, MBS5 and MBS6, failed to interact. Thus, the results indicate that the fifth and sixth metal-binding sites may well have a different function from the first four. Consistent with this differentiation of function between the metal-binding sites, Forbes et al. found that the first three metal domains (MBS1-3) could not replace MBS5 and MBS6 for copper transport (132). In addition, the most recent studies on the stoichiometry of the copper binding to ATP7A suggest that four Cu(I) ions are bound in an environment shielded from solvent (128). This study, however, did not identify which Cu sites were involved in the binding.

A possible model for integrating these observations is that ATOX1 delivers copper to the first four metal-binding sites, and the metal is subsequently transferred to MBS5 and MBS6, from where it is donated to the CPC in the channel, but direct donation of copper from other MBSs cannot be excluded. Some of the possible paths of copper movement between the MBSs and the CPC are indicated by dotted lines in Fig. 7. These transfer processes are presumably occurring under normal physiological conditions (i.e., with cytoplasmic copper concentrations within the low to normal range). When cyto-plasmic copper concentrations increase as a result of the exposure of the cell to elevated copper concentration, the capacity of the ATOX1 delivery system may be exceeded. Under such conditions, copper may be bound to low-molecular-mass ligands, such as glutathione, and delivered directly to the CPC residues in the channel. This mechanism would explain how an ATP7A molecule with all six copper-binding sites mutated could still pump copper into vesicles in cells exposed to high copper (133).

Whatever the exact role of the various copper-binding sites in copper transport, it is clear that copper binding to this region causes significant conformational changes that are presumably related to the copper-transport cycle (127,128). These changes might be linked, but not necessarily directly, to the conformational changes that accompany the metal transport through the channel. DiDonato et al. showed that a 70-kd fragment containing the six copper-binding domains of ATP7B underwent both secondary- and tertiary-structure changes upon copper binding (127). It has been speculated by a number of authors that the presence of six Cu sites in these proteins may serve as some kind of sensor of the copper concentrations in the cell (32,92). The conformational changes that the N-termi-nal region undergoes upon copper binding has been suggested to be part of the signaling mechanism that stimulates the copper-induced trafficking of ATP7A and ATP7B (127,128,132). There is also evidence that the N-terminal region of ATP7B interacts with part of the cytoplasmic loop containing the ATP-binding site in a copper-dependent manner. This interaction was proposed to be part of the overall conformational changes in the protein that are involved in both copper transport and copper-induced trafficking (136). To explore these ideas further, we now discuss the studies on the cell biology of ATP7A/7B and Cu-induced trafficking of these ATPases.

0 0

Post a comment