13 April 2012. What do prion-like proteins do when not converting neighbors to dangerous conformations? One of them helps stabilize functional protein complexes, according to a paper in the April 3 Nature Communications. The authors report that a prion-like domain is key to TDP-43’s cellular location in the nucleus and its function in RNA splicing. They theorize that the domain either brings together normal TDP-43 molecules or binds up abnormal fragments in pathological aggregates. The outcome hinges on the prion sequence itself. When harboring amyotrophic lateral sclerosis-linked mutations, TDP--43’s carboxyl terminus, including the prionesque domain, is released by proteolysis to form the undesirable inclusions, but in the normal form, the same domain promotes associations needed for the protein’s regular function, the authors conclude. TDP-43 is also involved in some forms of frontotemporal lobar dementia (Neumann et al., 2007; Mackenzie et al., 2010).
The results suggest that in TDP-43-linked diseases, neurons suffer when the prion-like domain fails in its regular duties and compromises the protein’s normal function, wrote study first author I--Fan Wang of the Academia Sinica, the national research center in Taipei, Taiwan, in an email to ARF. Che-Kun James Shen, at the same institute, was senior author.
Sequence gazing, Wang and colleagues observed several years ago that the carboxyl end of TDP-43 contains a prion-like stretch similar to that of the yeast protein SupN (Wang et al., 2008). TDP-43’s propensity to aggregate also points to a prionesque phenotype (see ARF related news story on Fuentealba et al., 2010). However, the segment does not fit all aspects of the prion definition, Wang noted; for example, it does not interact with the amyloid-specific dye Congo red. The Taiwan group is one of a handful of labs investigating TDP-43’s prion-like properties (see ARF related news story on Guo et al., 2011; ARF related news story on Fuentealba et al., 2010; ARF related news story on Sun et al., 2011 and Ju et al., 2011).
The current study supports others that found prion-like sequences were part of the problem, but it also emphasizes their natural, healthy role, said James Shorter of the University of Pennsylvania in Philadelphia. In so doing, Shorter said, the work offers a solution to a nagging question: Why would a protein like TDP-43 evolve a prion-like domain if prions cause pathology? According to these data, “the prion-like portion of TDP-43 is like a double-edged sword,” Shorter said. “You need it to achieve TDP-43 function, but at the same time [it] causes problems in protein folding.” Shorter was not involved with the study.
To investigate the similarities between the TDP-43 and yeast prion domains, the researchers swapped TDP-43’s prion-like section for the SupN version and found that anything TDP-43 could do, their TDPSupN construct could do, too. When expressed in 293T human embryonic kidney cultures, each formed two different kinds of nuclear structures visible on fluorescence microscopy: stable granules sized 50-200 nanometers across, and irregular puncta that rapidly formed and dissolved. TDP-43, an RNA-binding protein, is well known to aid excision of exon 9 of the cystic fibrosis transmembrane conductance regulator (CFTR) gene—although the precise mechanism is unknown (Buratti et al., 2001)—and the chimera also completed this task, the team discovered.
TDP-43 appeared to need a prion-like domain to self-assemble into functional complexes. A deletion mutant without that section remained diffuse in the nucleus and associated poorly with full-length TDP-43 in co-immunoprecipitation assays. The deletion construct also failed to properly splice CFTR. Similarly, a mutation in the prionesque section designed to break up the protein’s structure (proline for glutamine 303) downgraded CFTR splicing efficiency by roughly 24 percent. In some transfected cells, this construct moved into the cytoplasm, as occurs in many TDP-43 mutant systems and in ALS. This mutant also partially broke down, forming 24-kDa carboxyl-terminal fragments previously linked to ALS pathology (see ARF related news story on Zhang et al., 2007; ARF related news story on Pesiridis et al., 2011). “There is suspicion that this fragment causes a lot of the problems,” Shorter said. Other ALS-linked mutations in the prion-analog segment (348-cysteine and 361-serine) also partially reduced the exon-skipping skills of full-length TDP-43. However, these mutants assembled normal nuclear complexes, and did not fragment or move into the cytosol. Perhaps, the authors suggested, the mutations only slightly disrupted the prion fold, but not enough to cause full pathology or eliminate splicing activity.
The team also made a structure-breaking glycine-25-proline substitution in their TDPSupN construct (Chang et al., 2008). This version was less able to splice CFTR than the wild-type construct, and was found in both the nucleus and cytoplasm. A carboxyl-terminal fragment broke off at what could be the same cleavage site that releases the 24 kDa fragments from full-length TDP-43. “Impairing prion-like interactions of TDP-43 caused its degradation into a fragment identified in patients with early-stage TDP-43 proteinopathies,” Wang concluded. That is, without the normal interactions mediated by the prion domain, the protein is likely to fall apart—perhaps due to caspase action. Not only does this cause loss of TDP-43’s normal nuclear function, but it also potentially causes a gain of function in the production of toxic fragments that form pathological aggregates. Conversely, in the normal full-length TDP-43, the authors propose that the prion domain is a good player, helping put together the complexes needed to process RNAs. A similar good-prion, bad-prion pathway may be occurring in other ALS-linked proteins that have prion-like domains, such as FUS, Shorter said (see ARF related news story on Couthouis et al., 2011).
Promoting the good prion interaction while combating the bad one, then, should prevent TDP-43 proteinopathy. “Compounds that stabilize functional aggregates of aggregation-prone domains may be investigated as therapies for protein misfolding,” the authors wrote. “This is particularly true for TDP-43 proteinopathies, because cleavage of TDP-43 proteins occurs before inclusion formation during disease progression.” Last November, for example, the European Commission approved tafamidis, a drug that stabilizes tetramers of transthyretin, to treat familial amyloid angiopathy. This disease occurs when transthyretin tetramers dissociate and form insoluble aggregates. Could a similar strategy work for ALS? In their cell cultures, Wang and colleagues tried out a green tea derivative, epigallocatechin gallate (EGCG), which is thought to stabilize oligomers and disassemble amyloids (Roberts et al., 2009; Bieschke et al., 2010). EGCG treatment blocked the degradation of TDP-43, TDPSupN, and the Q303P mutant. Wang has started a company, Garage Brain Science in Taichung, Taiwan, to screen for treatments, including a compound from traditional Chinese medicine.—Amber Dance.
Wang IF, Chang HY, Hou SC, Liou GG, Way TD, James Shen CK. The self-interaction of native TDP-43 C terminus inhibits its degradation and contributes to early proteinopathies. Nat Commun. 2012 Apr 3;3:766. Abstract