This study characterizes the dynamic behavior of SOD1 in detail. First, it essentially reproduces previous studies including the ones from the authors' group, as it has been well known that overall structures are similar between wild-type and mutant SOD1 proteins. In addition, significant differences in the dynamic behavior have been observed between Apo and holo forms of SOD1. When the metal ions are removed from the protein, structural disorder increases particularly in the loop regions.

We think that one of the interesting findings in this paper is the increased solvent accessibility of Cys-6 upon metal removal. Cys-6 is one of the four Cys residues (Cys-6, 57, 111, 146) in SOD1 and is buried toward the protein interior in the holo form of SOD1. In an enzymatically active form of SOD1, an intra-molecular disulfide forms between Cys-57 and 146, while Cys-6 and 111 remain reduced. In contrast, pathological inclusions purified from several ALS-model mice contain SOD1 multimers that are cross-linked via non-physiological disulfide bonds (  Read more

This study characterizes the dynamic behavior of SOD1 in detail. First, it essentially reproduces previous studies including the ones from the authors' group, as it has been well known that overall structures are similar between wild-type and mutant SOD1 proteins. In addition, significant differences in the dynamic behavior have been observed between Apo and holo forms of SOD1. When the metal ions are removed from the protein, structural disorder increases particularly in the loop regions.

We think that one of the interesting findings in this paper is the increased solvent accessibility of Cys-6 upon metal removal. Cys-6 is one of the four Cys residues (Cys-6, 57, 111, 146) in SOD1 and is buried toward the protein interior in the holo form of SOD1. In an enzymatically active form of SOD1, an intra-molecular disulfide forms between Cys-57 and 146, while Cys-6 and 111 remain reduced. In contrast, pathological inclusions purified from several ALS-model mice contain SOD1 multimers that are cross-linked via non-physiological disulfide bonds (Furukawa et al., 2006).

It is, however, still controversial which Cys residues are involved in the formation of cross-linked SOD1 multimers under pathological conditions. While we have previously reported that the disulfide formation is not absolutely required for triggering SOD1 aggregation (Furukawa et al., 2008), an important role of Cys-6 and 111 in the formation of disulfide cross-links has been also suggested in the cultured cell model (Niwa et al., 2007). In addition, ALS-causing mutations at position 6 have been reported (i.e., C6G and C6F), implying that the other Cys residues are involved in the formation of disulfide-linked multimers even when Cys-6 is unavailable for disulfide formation. Nonetheless, the increased flexibility and solvent accessibility of Cys-6 upon metal removal will be an important clue to explain a molecular mechanism of the pathological SOD1 oligomer formation.

View all comments by Yoshiaki Furukawa

In a study of epic proportions (X-ray crystallography, calorimetry, NMR, melting, refolding, etc.), Oliveberg and coworkers add to the body of evidence showing that the loss of metals, and loss of zinc in particular, destabilizes SOD1 and promotes aggregation. There is probably now enough evidence to consider these findings irrefutable. They add considerably to the single prior study that demonstrates that the SOD1 electrostatic loop promotes protein aggregation. The authors use the frustrated model (for the non-scientist, this is a theory about how proteins fold, or don't, not Kate Moss at a buffet) to describe how the very residues that stabilize active SOD1 destabilize the demetallated enzyme. Oliveberg takes the loss of metals from wtSOD1, which results in destabilization and aggregation in vitro, as evidence that the more common, sporadic, cases are caused by loss of metals.

So the loss of metals must be a really bad thing for ALS patients, and sporadic ALS patients must have aggregated WT SOD1? Not exactly. The major protein in sporadic patients' inclusions appears to...  Read more

In a study of epic proportions (X-ray crystallography, calorimetry, NMR, melting, refolding, etc.), Oliveberg and coworkers add to the body of evidence showing that the loss of metals, and loss of zinc in particular, destabilizes SOD1 and promotes aggregation. There is probably now enough evidence to consider these findings irrefutable. They add considerably to the single prior study that demonstrates that the SOD1 electrostatic loop promotes protein aggregation. The authors use the frustrated model (for the non-scientist, this is a theory about how proteins fold, or don't, not Kate Moss at a buffet) to describe how the very residues that stabilize active SOD1 destabilize the demetallated enzyme. Oliveberg takes the loss of metals from wtSOD1, which results in destabilization and aggregation in vitro, as evidence that the more common, sporadic, cases are caused by loss of metals.

So the loss of metals must be a really bad thing for ALS patients, and sporadic ALS patients must have aggregated WT SOD1? Not exactly. The major protein in sporadic patients' inclusions appears to be TDP43, and only occasionally SOD1. If you had to choose an ALS mutation, you would do better to take H46R, which doesn't bind Cu but has more than 10 years’ mean survival, than A4V, which binds metals well but has less than a year’s mean survival.

What does this mean for ALS patients? Despite my nitpicking, it is clear that ALS-causing mutations of SOD1 (and probably TDP43) cause the protein to aggregate. Since small aggregates are quite toxic, and large aggregates choke off axons and kill neurons, we ought to be trying to stop aggregation. How many of the more than 50 past ALS clinical trials specifically targeted protein aggregation? My count is zero. How many are planned? My count is zero specific aggregation inhibitors, and one potential aggregation inhibitor. If you believe Oliveberg's data, as I do (we have data from 13 fALS variants that also indicate the electrostatic loop is perturbed), the SOD1 electrostatic loop may be a logical target. If not SOD1, we should target TDP43.

View all comments by Jeffrey Agar