This article is the first to demonstrate fibrillation of wild-type SOD1 and familial ALS variants under conditions approaching physiological. Such conditions are required for aggregation results to be disease-relevant, because as Chris Dobson and coworkers have pointed out, most proteins can form fibrils in acidic (non-physiological) conditions. Another major finding here is that a small "seed," or nucleus, of aberrantly modified protein (native disulfide and metals are missing) can cause the aggregation of more "normal" Cu/Zn-SOD1 (native disulfide intact but some metals are missing). Finally, these authors demonstrate that the addition of Cu inhibits fibrillation.

In a nutshell, this article provides the first experimental evidence for the seeding of large quantities of more folded SOD1 by smaller quantities of less folded SOD1. Considering these author's previous results, where "normal" SOD1 wouldn't aggregate even after two years, it appears that aberrantly modified SOD1 is the lynchpin of the aggregate.

For these findings to be relevant to ALS etiology, these SOD1...  Read more

This article is the first to demonstrate fibrillation of wild-type SOD1 and familial ALS variants under conditions approaching physiological. Such conditions are required for aggregation results to be disease-relevant, because as Chris Dobson and coworkers have pointed out, most proteins can form fibrils in acidic (non-physiological) conditions. Another major finding here is that a small "seed," or nucleus, of aberrantly modified protein (native disulfide and metals are missing) can cause the aggregation of more "normal" Cu/Zn-SOD1 (native disulfide intact but some metals are missing). Finally, these authors demonstrate that the addition of Cu inhibits fibrillation.

In a nutshell, this article provides the first experimental evidence for the seeding of large quantities of more folded SOD1 by smaller quantities of less folded SOD1. Considering these author's previous results, where "normal" SOD1 wouldn't aggregate even after two years, it appears that aberrantly modified SOD1 is the lynchpin of the aggregate.

For these findings to be relevant to ALS etiology, these SOD1 aggregates must be toxic. The work of Heather Durham's, David Borchelt's, and numerous other groups demonstrates the propensity to aggregate is the one property shared by all SOD1 mutants tested, and there is evidence that aggregates can be toxic. Conversely, contradictory results convincingly demonstrate that not all aggregates are toxic, and thus the debate as to the nature of aggregate toxicity lives on. One point of common ground may be the following: the histopathological aggregates observed in ALS patients to have grown large enough to essentially "blow up" the only motor axon, thereby causing the motor neuron to die, are toxic by definition.

Our previous Alzforum-featured article demonstrated that the predicted rates of SOD1 mutant aggregation elongation, and SOD1 stability, were synergistic risk factors affecting the survival of SOD1-related ALS patients (Wang et al 2008). The article of Valentine and coworkers demonstrates how nucleation, the exceedingly slow process that precedes elongation (growth), as well as elongation itself, are accelerated by less stable SOD1. Because these results involved modification of wild-type SOD1, they are of particular relevance to sporadic ALS. Notably, the idea that modified SOD1 could be involved in ALS was first put forth by Valentine and coworkers more than 10 years ago.

View all comments by Jeffrey Agar

Despite accumulating evidence supporting the view that misfolding and aggregation of SOD1 is central to the pathogenesis of SOD1-linked ALS, and that perturbation of steps of maturation of SOD1 under non-physiological conditions promotes fibrillation of SOD1, it is not clear whether amyloid formation can be induced in physiological environments. Using a series of biophysical experiments, Joan Valentine provides direct evidence to support convincingly the idea that specific changes in SOD1 under physiological conditions can initiate amyloid nucleation to facilitate subsequent fibrillation of various forms of SOD1, including that of the mature SOD1.

This paper defines the parameters that are necessary for both the initiation and elongation of fibrillation, processes that can be mechanistically separated. These results have important implications for our understanding of SOD1 pathogenesis. They point to future research directions toward clarifying the precise role of oligomeric and/or fibrillar species of SOD1 in the pathogenesis of ALS.

View all comments by Philip Wong

This interesting study describes a mechanism of SOD1 fibrillation that actually reproduces our previous results very well (Furukawa et al., 2008). We have already proposed that complete loss of post-translational modifications leads to fibrillar aggregation of SOD1. A novel proposal in this paper by Joan Valentine’s group is that disulfide-reduced apo-SOD1 can accelerate or trigger fibrillation of the more matured form of SOD1 (e.g., the disulfide-intact form). We have shown that a disulfide-intact SOD1 itself is highly resistant to fibrillation even with vigorous agitation for a week (Furukawa et al., 2008), so it is quite interesting for us to see these new findings that an aggregation-resistant disulfide-intact SOD1 protein can be recruited to the aggregates.

The authors have further shown that a disulfide-reduced apo-SOD1 with an ALS mutation can recruit a wild-type disulfide-intact apo-SOD1 protein into thioflavin T-positive aggregates. As mentioned below, however, we hope that...  Read more

This interesting study describes a mechanism of SOD1 fibrillation that actually reproduces our previous results very well (Furukawa et al., 2008). We have already proposed that complete loss of post-translational modifications leads to fibrillar aggregation of SOD1. A novel proposal in this paper by Joan Valentine’s group is that disulfide-reduced apo-SOD1 can accelerate or trigger fibrillation of the more matured form of SOD1 (e.g., the disulfide-intact form). We have shown that a disulfide-intact SOD1 itself is highly resistant to fibrillation even with vigorous agitation for a week (Furukawa et al., 2008), so it is quite interesting for us to see these new findings that an aggregation-resistant disulfide-intact SOD1 protein can be recruited to the aggregates.

The authors have further shown that a disulfide-reduced apo-SOD1 with an ALS mutation can recruit a wild-type disulfide-intact apo-SOD1 protein into thioflavin T-positive aggregates. As mentioned below, however, we hope that the authors will add more direct evidence to support their hypothesis that a disulfide-intact SOD1 is recruited in the aggregates of disulfide-reduced SOD1. The “recruitment mechanism” implies involvement of wild-type SOD1 in the pathogenesis of familial ALS. It is still controversial whether wild-type SOD1 is involved in the pathological inclusions found in the familial ALS patients, but in Huntington disease, for example, a mutant huntingtin protein with an abnormally expanded polyglutamine tract has been shown to recruit a wild-type huntingtin protein (i.e., non-pathogenic length of polyglutamine) into the aggregates (Busch et al., 2003). A recruitment of wild-type protein into the protein inclusions/aggregates may thus be a more general pathogenic mechanism in neurodegenerative diseases.

We have some technical questions about the present paper. One concerns the experimental conditions when the thiol-disulfide status of SOD1 is analyzed. The authors analyzed the thiol-disulfide status of a SOD1 protein that is collected during the lag phase of an aggregation reaction, and showed significant amounts (approx. 50 percent of total; see Figure 2A) of disulfide-intact SOD1 dimer even in the presence of 5 mM DTT. The analysis of the thiol-disulfide status of SOD1 was, however, performed after separation using size-exclusion chromatography with an elution buffer in the absence of any metal chelators. Given that SOD1 has an unusually high affinity to bind copper and zinc ions, a trace amount of metal ions in a buffer as a contaminant could easily catalyze disulfide formation under aerobic conditions. In our hands, actually, in the presence of 5 mM DTT with 5 mM EDTA, almost all disulfide bonds in SOD1 remain reduced. We would thus recommend to further add EDTA during an aggregation reaction and/or a chromatography step and to test whether disulfide formation really occurs during the lag phase of an aggregation reaction.

It would also be helpful to analyze the thiol-disulfide status of SOD1 in the final aggregates, not in the protein samples during the lag phase. This could be performed by SDS-PAGE after solubilizing the aggregates in 2 percent SDS in the presence of iodoacetamide, or any other thiol-protecting reagent. By this method, disulfide-reduced and disulfide-intact SOD1 can be clearly separated due to their different electrophoretic mobility (e.g., Furukawa et al., 2004), so it is easily confirmed whether the disulfide-intact SOD1 is actually recruited into the final aggregates.

Human SOD1 has four cysteine residues in total, two of which are involved in intramolecular disulfide formation (Cys57 and Cys146). We have previously shown that mutant SOD1 in which all four cysteines are replaced with serine still forms fibrillar aggregates, and suggested that intra- and inter-molecular disulfide formation is not required for SOD1 fibrillation (Furukawa et al., 2008). In that sense, it is surprising for us to see that SOD1 with either a C57S or C146S single mutation did not produce thioflavin T-positive aggregates (Figure 3A). Although the authors concluded that both Cys57 and Cys146 are required to form SOD1 amyloid nuclei, possibly through a disulfide cross-link, we believe it is also possible that C57S- or C146S-mutant SOD1 may “inhibit” fibrillation by forming an aberrant disulfide bond. Actually, the authors performed the aggregation reaction of C57S or C146S SOD1 in the absence of reductants and metal chelators, which facilitates formation of aberrant disulfide cross-links. This could be the reason why C57S or C146S SOD1 did not form fibrillar aggregates under their experimental conditions.

In any case, it is necessary as a next step to show evidence that SOD1 can exist as a cross-linked dimer under the experimental conditions used in this study, and that such a cross-linked dimer can act as a nucleus for fibrillation. This can be confirmed by preparing disulfide cross-linked SOD1 dimers in vitro, adding those to the solution containing apo- and disulfide-reduced SOD1, and examining the aggregation kinetics.

View all comments by Yoshiaki Furukawa
View all comments by Nobuyuki Nukina

In this article, Chattopadhyay et al. report results of studies of fibril formation by the amyotrophic lateral sclerosis (ALS)-linked protein superoxide dismutase 1 (SOD1). These studies are important because they show that the phenomena of fibril nucleation and fibril elongation can be distinguished from each other through the use of different forms of the same protein. Distinguishing between these events is important for understanding disease etiology and for the design of therapeutic agents.

SOD1 is an exceptionally stable dimer formed by two monomers, each of which contains a disulfide bond, a Cu(II) ion, and a Zn(II) ion. Chattopadhyay et al. show that small amounts of disulfide-reduced apo-SOD1 are able to initiate fibril formation that then is perpetuated by the highly stable wild-type SOD1. From a phenomenological perspective, this result means that SOD1 amyloid formation may involve both homogeneous and heterogeneous nucleation reactions. The implication is that small amounts of nucleation-capable conformers may be sufficient to initiate SOD1 assembly. This places...  Read more

In this article, Chattopadhyay et al. report results of studies of fibril formation by the amyotrophic lateral sclerosis (ALS)-linked protein superoxide dismutase 1 (SOD1). These studies are important because they show that the phenomena of fibril nucleation and fibril elongation can be distinguished from each other through the use of different forms of the same protein. Distinguishing between these events is important for understanding disease etiology and for the design of therapeutic agents.

SOD1 is an exceptionally stable dimer formed by two monomers, each of which contains a disulfide bond, a Cu(II) ion, and a Zn(II) ion. Chattopadhyay et al. show that small amounts of disulfide-reduced apo-SOD1 are able to initiate fibril formation that then is perpetuated by the highly stable wild-type SOD1. From a phenomenological perspective, this result means that SOD1 amyloid formation may involve both homogeneous and heterogeneous nucleation reactions. The implication is that small amounts of nucleation-capable conformers may be sufficient to initiate SOD1 assembly. This places the entire assembly process in its proper thermodynamic context, namely that it requires only the transitory presence of nucleation-competent conformers to proceed. These conformers may form as a result of mutations that alter the primary structure of SOD1. Such mutations are well known to accelerate SOD1 assembly rates. However, the data of Chattopadhyay et al. also suggest that factors unrelated to gene structure can cause disease by altering the conformer frequency distribution.

In addition to these important general features of the assembly reaction, the authors address the question of how the presence or absence of Cu(II) and Zn(II) affects the process. No fibril formation was observed in the presence of two Cu atoms, either alone or in the presence of bound Zn.

A second question is whether only the cysteines forming the naturally occurring disulfide bond in SOD1, Cys57, and Cys146 are required for fibril nucleation, or whether the remaining cysteines at positions 6 and 111 might also be involved. Through careful substitution studies, the conclusion could be made that the former cysteines are obligatory for fibril nucleation and that the latter cysteines are dispensable.

One hopes, through careful study of homogeneous assembly systems, that a single critical therapeutic target may be identified. The work of Chattopadhyay et al. shows that the SOD1 system is more complex. Nucleation may be mediated by different forms of the same protein, both with respect to primary structure and post-translational modifications. Different protein forms also may participate in fibril elongation processes, albeit with different rate constants.

Where does this leave the effort to understand and cure ALS? It is my opinion that it moves us forward. The rigorous studies of Chattopadhyay et al. add significantly to the SOD1 experimental corpus, which should facilitate a multi-factorial analysis of SOD1 self-assembly. Through this analysis, it may be possible to define the one (or few) factors that are the most critical in controlling assembly and thereby should be targeted in therapeutic approaches.

View all comments by David Teplow