To the editor

Evidence indicates that Cu/Zn superoxide dismutase 1 (SOD1) mutations linked to amyotrophic lateral sclerosis (ALS) lead to inappropriate release of copper (Cu) redox activity and oxidative damage1,2,3, which can be ameliorated by Cu chelators4,5 and exacerbated by genetic ablation of copper-sequestering metallothioneins6. The recent study by Subramanian and colleagues7 in Nature Neuroscience demonstrates that neurodegeneration in transgenic mice overexpressing mutant SOD1 is not rescued by genetic removal of the copper chaperone of SOD1 (CCS). This study convincingly showed that pathogenesis in the mouse model was not due to the activity of Cu at the active site of SOD1, as CCS is required for efficient loading of Cu into this structure8. The biological activity of the SOD1 enzyme is indeed relatively reduced in CCS−/−/SOD1-mutant crossed mice7.

However, these findings still do not allow us to disqualify Cu from the pathogenesis of this disorder. The view that “ALS is not caused by an altered oxidative activity of SOD1”, proposed in the accompanying News and Views9 in the same issue, is not conclusively supported by the findings, because a pro-oxidative interaction between Cu and SOD1 at sites other than the catalytic Cu binding site has not been excluded.

The pathological, Cu-dependent, pro-oxidant reactions attributed to SOD1 include peroxidation1, peroxynitrite generation2 and inappropriate hydrogen peroxide generation3. The pro-oxidant chemistry of mutant SOD1 has been sometimes incorrectly attributed to “the requirement of copper at the SOD1 catalytic site”9. This is a misconception because pro-oxidant (for example, peroxidative) activity is mediated by other, non-dismutase Cu binding sites on the protein.

There are at least two other documented Cu binding sites on SOD1 that have little or no SOD activity. Cu can incorrectly bind to the zinc binding site10, and Cu can also bind to the surface residues of SOD1 (ref. 11). H46R mutant SOD1, linked to familial ALS, has no SOD activity but binds Cu at surface residue Cys111, a possible transition site for CCS-mediated copper loading11. The Cys111-bound Cu is redox-active and can catalyze deleterious pro-oxidant reactions. To show that CCS ablation abolishes Cu binding to these sites, a decrease in the Cu binding stoichiometry to mutant SOD1 would need to be demonstrated.

The data of Subramanian and colleagues do not establish such a decrease in the CCS−/− mice. The authors have taken the decrease in 64Cu-radiolabeled SOD1 in CCS−/− mice, and accompanying decreased dismutase activity, as proof that Cu stoichiometry is diminished. However, without controls that correlate the dynamic incorporation of 64Cu with the stoichiometric content of Cu bound to SOD1 in the tissue, it is inaccurate to conclude that chemical equilibrium had been reached, and that there was an “80% reduction” in the amount of total Cu bound to SOD1 in the CCS−/− mice7. The marked decrease in dismutase activity that was observed in the CCS−/−/SOD1-mutant mice also does not prove that total Cu binding to SOD1 is decreased, as this activity is only a reflection of Cu binding at the active site; the dismutase activity assay does not reflect Cu bound to other sites on the protein.

What then accounts for the diminished 64Cu incorporation into SOD1 in the CCS−/− mice? It is possible that CCS ablation forces Cu to dwell on the transition loading site of SOD1, which normally would be unoccupied11. This could cause a decrease in dismutase activity, whereas the stoichiometry of Cu binding would be unchanged. The occupied transition site would then act as a chemical log-jam, inhibiting further incorporation of 64Cu. Supporting this possibility, the CCS−/− lesion alone caused a marked (and unexplained) elevation of total tissue Cu (Fig. 3d in ref. 7); this increased Cu could saturate binding sites on SOD1 and inhibit further uptake of 64Cu. Other data in Fig. 3d show that the two strains of mutant SOD1 mice (with CCS+/+) both had ∼3.5 μg/g (∼30%) elevations in brain Cu levels compared to normal mice. This is in agreement with the significant increase in spinal cord Cu levels that we observed in G93A mutant SOD1 transgenic mice (C. Maynard et al., unpublished data). This marked elevation of tissue Cu (∼30 μM) is probably Cu attached to the overexpressed SOD1. The CCS−/− lesion did not decrease total Cu levels in the mutant SOD1 tissue (Fig. 3d). Therefore, the ablation of CCS probably does not decrease Cu binding stoichiometry to mutant SOD1.

In summary, the Cu-mediated oxidative damage by mutant SOD1 is not predicted to be proportional to dismutase activity, but rather, proportional to peroxidative or other pro-oxidant activity, which was not assessed by Subramanian and colleagues. Their findings do not exclude Cu from the pathophysiology of ALS. However, we can now eliminate one of the Cu binding sites on SOD1 from the pathogenesis of ALS.

See Reply to "Is ALS caused by an altered oxidative activity of mutant superoxide dismutase?" by Wong.