Cellular distribution of copper to superoxide dismutase involves scaffolding by membranes

Abstract

Efficient delivery of copper ions to specific intracellular targets requires copper chaperones that acquire metal cargo through unknown mechanisms. Here we demonstrate that the human and yeast copper chaperones (CCS) for superoxide dismutase 1 (SOD1), long thought to exclusively reside in the cytosol and mitochondrial intermembrane space, can engage negatively charged bilayers through a positively charged lipid-binding interface. The significance of this membrane-binding interface is established through SOD1 activity and genetic complementation studies in Saccharomyces cerevisiae, showing that recruitment of CCS to the membrane is required for activation of SOD1. Moreover, we show that a CCS:SOD1 complex binds to bilayers in vitro and that CCS can interact with human high affinity copper transporter 1. Shifting current paradigms, we propose that CCS-dependent copper acquisition and distribution largely occur at membrane interfaces and that this emerging role of the bilayer may reflect a general mechanistic aspect of cellular transition metal ion acquisition.

Keywords: copper homeostasis; copper transfer; copper uptake; membrane scaffold.

Fig. 1.

In vitro characterization of a putative lipid-binding interface in yCCS. (A) Crystal structure of dimeric yCCS (Protein Data Bank code 1QUP) rendered in Chimera with calculated Coulombic surface potential (ε = 20; d = 2.5). Relative orientation with respect to the membrane (black lines) is also shown to indicate the location of the putative binding site. Positive residues hypothesized to play a role in membrane binding are explicitly shown in the enlarged view. (B) Means ± SEM of yCCS association with liposomes of different lipid compositions (n = 11 for the protein-only control; n = 8 for all other data columns); significance was obtained from unpaired Student’s t test . *P < 0.05; **P < 0.005. (C) Mean ± SEM data from liposome floats conducted on wild-type yCCS and membrane-binding mutants of yCCS. Mutants are explicitly identified below their respective columns; combined mutant: yCCS^{K95A,K96A,R101A,K167A}. Statistical significance was determined between wild-type yCCS (n = 8) and membrane-binding mutants (n = 8) using an unpaired Student’s t test. ***P < 0.0005.

Fig. 2.

Effect of CCS lipid-binding deficiencies on ySOD1 in vivo. (A) CCS ablation causes growth defects in S. cerevisiae when grown under aerobic conditions. The top two rows compare growth of wild-type S. cerevisiae in aerobic (Left) and anaerobic (Right) conditions with growth of a CCS-deletion strain under the same conditions. The third row shows the yCCS deletion strain complemented with wild-type yCCS. The bottom panels represent the growth phenotypes of yCCS deletion strain complemented with a membrane-binding mutant (K95A, K96A, R101A). (B) ySOD1 activity after reconstitution of purified, demetallated enzyme in the presence of either the apo-CCS (Left) or copper-loaded (Right) forms of the quintuple mutant of yCCS (P < 0.002). (C) Mean ± SEM data from SOD activity from cytosolic fractions of total yeast cell lysate (n = 8). Anchor: wild-type yCCS fused to a mitochondrial outer membrane anchor sequence from TOM-22. Student’s t test was used to measure significance between wild-type yeast cells (WT) and episomal expression of yCCS wild-type and mutants in RR102 yCCS deletion strain (yCCS KO). *P value <0.05; **P < 0.005; ***P < 0.0005.

Fig. 3.

Copper and SOD1 dependence of CCS membrane binding. (A) Means ± SEM (n = 8) from liposome floats conducted on copper-loaded yCCS (+Cu) and yCCS without copper (−Cu) with 80% POPS liposomes. Binding was found to not be significantly different (unpaired t test, P = 0.3422). (B) Mean values ± SEM from liposome floats of yCCS:ySOD1^H48F complex (n = 7) with comparison with CCS-only control (n = 11; P = 0.0001). (C) Mean values ± SEM from liposome float assays conducted with wild-type yCCS on oxidized yeast extract lipids and nonoxidized yeast extract lipids (t test, P = 0.0045). (D) Displays the mean ± SEM data from liposome floats conducted on ySOD1(H48F). Comparing each condition (n = 4) with the ySOD1^H48F protein-only control (n = 11) by unpaired Student’s t test showed that ySOD1^H48F bound all lipids tested with similar affinity (**P < 0.005).

Fig. 4.

Lipid-binding screen of human CCS- and hCTR1-mediated pull-down of hCCS. (A) Lipid float data for human CCS. All means ± SEM were compared with the protein-only controls, using the Student’s t test to determine significance (n = 8 for all lipid groups and n = 11 for hCCS protein-only control; **P < 0.005; ***P < 0.0005). Lipid mixtures (POPA, POPG, POPS) were mixed with POPC lipids in an 80:20 ratio (wt/wt). (B) hCTR1-mediated pull-down of hCCS measured by SDS/PAGE (Left) and densitometry after Coomassie staining (Right). Significance was determined against the “hCCS-only” control and yielded P < 0.039 (n = 3) for copper-loaded Cu(I)-hCTR1, whereas Apo-hCTR1 without copper (n = 3) did not cause hCCS significant pull-down.

Fig. 5.

Model for membrane-dependent copper distribution to CCS and SOD1. This figure summarizes the putative copper distribution pathway from plasma membrane to SOD1. (A) Copper entry through CTR1. (B) Targeting of apo-CCS to the membrane allows it to screen the membrane surface for the presence of copper-loaded CTR1. (C) Engagement of apo-CCS with CTR1 leads to copper transfer to CCS. (D) Cu-loaded-CCS remains bound by the membrane, where it recruits membrane bound apo-SOD1 for transfer of copper.