“It is looking more and more likely that TDP-43 really is the culprit and not a bystander,” said James Shorter, co-principal investigator on the paper along with Aaron Gitler. First author Brian Johnson and colleagues purified soluble recombinant human TDP-43 from bacteria and shook it up. Within minutes, the protein started to glue itself together, as the scientists determined by the turbidity of the solution and the proportion of TDP-43 in the pellet after centrifugation.

Under the electron microscope, the aggregates from the test tubes looked “extremely similar” to those found in samples from people who had ALS or FTLD, Gitler said. The purified protein formed small granular structures as well as filaments, both of which are seen in tissue samples (Lin and Dickson, 2008; Mori et al., 2008).

Other work has shown that the carboxyl-terminal portion of TDP-43 is critical to both its aggregation and toxicity (see ARF related news story on Zhang et al., 2009), and the current study provides more evidence of the same. A carboxyl-terminal fragment (amino acids 188-414) came together in vitro at a similar rate as full-length protein, but a truncate missing the carboxyl end (1-275) failed to coalesce.

Johnson and Gitler previously developed a yeast-based model for TDP-43 pathology; in yeast, the protein forms cytoplasmic inclusions as it does in neurons, and is toxic when expressed at high levels (Johnson et al., 2008). In the current work, they turned again to the single-cell model to analyze the effects of TDP-43 mutations on aggregation. Scientists have discovered more than 25 mutations in TDP-43—all but one in the carboxyl-terminal domain—that are associated with familial ALS (see ARF related news story). Johnson and colleagues expressed seven such mutants in yeast, under an inducible promoter, and used microscopy to count the inclusions that formed. Wild-type YFP-TDP-43 forms more than three aggregates, within six hours, in approximately 4 percent of cells; for six of the mutants, more than 10 percent of cells were affected. One particularly aggregation-prone mutant, Q331K, produced three or more inclusions in more than a quarter of yeast cells. The six mutants that formed more aggregates than the wild-type TDP-43 were also more toxic, limiting yeast cell viability. In addition, the mutant proteins aggregated more quickly in vitro.

These and other data suggest that at least some TDP-43 mutations cause a toxic gain of function. “Presumably the mutations are destabilizing the [protein] structure,” Shorter said. But there was one familial ALS mutant (see Corrado et al., 2009 and Del Bo et al., 2009) that Shorter called “curious.” TDP-43-G294A did not aggregate any more than wild-type in vitro or in the yeast assay, and showed toxicity similar to the wild-type protein, raising questions about its role in the disease. Unlike the other mutations, G294A occurs in a glycine-rich region that Shorter suspects may be involved in protein-protein binding necessary for TDP-43’s normal function. “There could be a loss of function that our yeast model is not able to pick up,” Johnson said, noting that because yeast lack endogenous TDP-43, a loss-of-function mutation is less likely to cause problems.

“There may be multiple mechanisms” for TDP-43 toxicity, agreed Leonard Petrucelli of the Mayo Clinic in Jacksonville, Florida, who was not involved with the current work. The study, he said, “really solidifies our understanding of toxicity and the importance of the C-terminal region.” However, Shorter was careful to note that although there is a correlation between TDP-43 mutations, aggregation, and disease, the experiments do not conclusively show that aggregation directly leads to disease.

Gitler and colleagues are using their yeast system to look for ways to prevent or destroy toxic TDP-43 aggregates. Using a genetic screen, they are searching for genes that suppress or enhance TDP-43 toxicity to learn more about the pathways involved. The yeast or in vitro system could also be useful for screening drugs. In addition, Gitler said, testing several mutations in the relatively quick yeast system could help scientists determine which would be the most promising genetic changes to engineer in mouse models. Based on the current study, Q331K looks like one mutant worthy of further investigation, and G294 might offer an example of a different route to TDP-43 proteinopathy.—Amber Dance.

Reference:
Johnson BS, Snead D, Lee JJ, McCaffery JM, Shorter J, Gitler AD. TDP-43 is intrinsically aggregation-prone and ALS-linked mutations accelerate aggregation and increase toxicity. J Biol Chem. 2009 May 22. Abstract