TDP-43 Structure Reveals Two-Faced Amino End

The amino end of TDP-43 has evaded structural characterization because of its penchant for aggregation. Now it has finally revealed its shape. Two shapes, in fact: The domain flips between a well-folded nucleic acid-binding conformation and a loose, aggregation-prone jumble, according to a paper in the Proceedings of the National Academy of Sciences this week. Unexpectedly, the folded conformation looks much like ubiquitin, but this version binds nucleic acids instead of proteins. The authors from the National University of Singapore propose that the folded configuration could work in the nucleus, where TDP-43 regulates gene expression, while the floppy disarray likely promotes cytoplasmic aggregation. Small molecules that lock the domain in the more steadfast form might be therapeutic for TDP-43 proteinopathies, which include amyotrophic lateral sclerosis and frontotemporal dementia, suggested senior author Jianxing Song in an email to Alzforum.

In addition to the amino-terminal domain, TDP-43 contains two RNA-binding motifs and a glycine-rich carboxyl terminus. The majority of disease-linked mutations cluster in the carboxyl end, so scientists have focused their attention on that section, noted Magdalini Polymenidou of the University of Zürich, who was not involved in the study. However, last year researchers at the Mayo Clinic in Jacksonville, Florida, reported that the protein’s amino tip mediates both TDP-43’s function in RNA splicing, and its aggregation (Zhang et al., 2013). Song’s dual structures reveal how the amino end might be involved in these processes. “It really shows that we have overlooked the amino terminus,” Polymenidou said. “We need to study the entire protein.”

Two of a kind.

The structure of the TDP-43 amino end (purple) neatly aligns with that of ubiquitin (yellow). [Image courtesy of Qin et al., PNAS.]

TDP-43, like many proteins involved in neurodegeneration, confounds structural biologists with its aggregation, Song wrote. In TDP-43, the amino terminus has the highest propensity to cluster, he added. Song and study first author Haina Qin came up with a trick to decipher the structures of these typically insoluble proteins by simply dissolving them in purified, salt-free water (Li et al., 2006). Along with co-first authors Liang-Zhong Lim and Yuanyuan Wei, they have already applied this technique to crack the structures of the ALS-related proteins SOD1 and VAPB (Lim et al., 2014; Qin et al, 2013). With the TDP-43 amino domain in pure water, the authors were able to use both nuclear magnetic resonance (NMR) spectroscopy and circular dichroism to solve its structures. They observed the disordered form in equilibrium with the tightly folded one. The more concentrated the protein, the more likely it was to adopt the floppy, aggregation-prone shape.

The stiffer version, surprisingly, folds like ubiquitin (see image above). TDP-43’s amino acid sequence provided no hints of similarity to ubiquitin; it does not even contain the lysines used by ubiquitin to connect with target proteins, noted Philipp Kahle of the University of Tübingen, Germany, who was not part of the study group. Kahle said he was initially skeptical that TDP-43 would form an ubiquitin-like fold, but found the overlay between the TDP-43 and ubiquitin shapes compelling. Other experts contacted by Alzforum also found the structural work to be sound. “The NMR data are convincing,” wrote Elizabeth Meiering of the University of Waterloo, Canada. “I think it is believable,” agreed Gregory Petsko of Weill Cornell Medical College in New York City.

Without those lysines, what does this ubiquitin copycat bind to? Since TDP-43 contains two known RNA-interacting segments, and the amino terminus was previously reported to bind DNA (Chang et al., 2012), Song and colleagues incubated the amino domain with nucleic acids to look for interactions. TDP-43 is known to bind RNAs with alternating uridines and guanines, so they tried those UGUGUG constructs as well as the corresponding single-stranded DNA with thymines and guanines. The amino terminus, when folded, bound both nucleic acids. RNA-binding proteins typically exhibit affinity for DNA as well, Polymenidou noted. Kahle suggested that in the cell, the TDP-43 amino domain would more likely bind RNA than DNA. Thousands of RNAs are known to interact with TDP-43, while only a couple of DNA sequences do so (see Nov 2010 conference story; Mar 2011 news story; Ou et al., 1995; and Lalmansingh et al., 2011).

“To the best of our knowledge, the TDP-43 amino terminus is the first ubiquitin-like fold which can directly bind nucleic acids,” Song wrote. Other proteins might contain similarly structured domains, but they would not be obvious by sequence gazing, he pointed out. Perhaps, Kahle suggested, Song’s pure-water technique will reveal other structural siblings of ubiquitin.

“I find it remarkable that you have a peptide that yields two conformers; one is wonderfully folded and the other is not,” Kahle said. Typically, proteins with alternate conformations toggle between two tightly folded states, though at least one other, a kind of sarcoma homology 3 domain, also has structured and unstructured states (Zhang et al., 1994). In the case of TDP-43, the first 25 residues of the protein are more disordered than those in ubiquitin, Song and colleagues report, which may explain its propensity to hang loose.

One TDP-43—Two Pathways.

Jianxing Song proposes that TDP-43 exists in equilibrium between tightly structured and loose forms. The folded version forms functional oligomers in the nucleus, where it regulates gene expression. The unfolded proteins coalesce to make inclusions in the cytoplasm. [Image courtesy of Qin et al., PNAS.]

Why would the protein contain a segment that makes it prone to unfold and aggregate? “I think it is a compromise of nature,” Kahle suggested. He speculated that the unusually loose folds of TDP-43’s amino end might allow it to be somewhat nonspecific in its binding, so it could find all those different RNA targets. However, he cautioned, this in vitro study cannot confirm that the same folded-to-unfolded switch occurs in cells.

If the two conformations do arise in cells, Song and colleagues suggest a model in which the tightly folded version attends to RNA in the nucleus, while the flexible one aggregates in the cytoplasm. The high concentration of TDP-43 in stress granules might promote the switch to the latter conformer and initiate aggregation, Polymenidou posited.

The results suggest a possible route to treat TDP-43 proteinopathies by preventing unfolding and aggregation. “We are now focused on the discovery and design of molecules to stabilize the folded form of the TDP-43 amino terminus, in an attempt to develop drug leads,” Song wrote. Polymenidou agreed that this therapeutic strategy makes sense.

Copyright © 1996–2018 Biomedical Research Forum, LLC. All Rights Reserved.

tdp-43 topic-preclinical
Share this: