7 March 2013. Humankind’s 200 or so RNA-binding proteins have proved once again to be a plentiful hunting ground for genes linked to neurodegeneration—particularly when they include prion-like sequences. In the March 3 Nature online, researchers led by Paul Taylor of St. Jude Children’s Research Hospital in Memphis, Tennessee, report that mutations in two heterogeneous nuclear ribonucleoproteins—hnRNPA2B1 and hnRNPA1—cause amyotrophic lateral sclerosis (ALS) and multisystem proteinopathy (MSP), also known as inclusion body myopathy-Paget’s disease of bone. These hnRNPs join other RNA-binding proteins that have been linked to neurodegenerative disease, including TDP-43, FUS, TAF15, and EWS (see ARF related news story on Couthouis et al., 2011). They all have prion-like domains where disease mutations occur. Taylor and colleagues found that the prion-like regions in the hnRNPs contain "steric zipper" motifs that encourage protein-protein binding. The disease-causing mutations enhanced protein zipping, promoting pathological aggregation. Other RNA-binding proteins with prion-like domains may also contribute to disease in a similar fashion, the authors propose. Study collaborator James Shorter of the University of Pennsylvania has identified 20 such genes that may be linked to disease.
“These findings further support the important role of RNA binding proteins in ALS, frontotemporal lobar degeneration, and related multisystem proteinopathies,” wrote Christian Haass of Ludwig Maximilians University in Munich, Germany, in an e-mail to Alzforum (see full comment below). Haass, who was not involved in the current work, recently discovered that another hnRNP, A3, occurs in protein deposits in brain samples from people who have ALS or frontotemporal lobar degeneration due to an expansion in the C9ORF72 gene (see ARF related news story on Mori et al., 2013).
Researchers in Taylor’s lab study multisystem proteinopathy, a complex and variable syndrome. Afflicted family members exhibit degeneration in bone, muscle, neurons, or a combination of any of these three. It is unknown why these tissue types are susceptible, Taylor said. The neurodegeneration can take the form of ALS or frontotemporal dementia (Nalbandian et al., 2011; see ARF related news story on Johnson et al., 2010).
Taylor’s group previously discovered that mutations in valosin-containing protein (VCP) cause MSP (Watts et al., 2004). VCP is a kind of chaperone called a segregase, which acts to disassemble many kinds of protein complexes. Among its targets are stress granules—temporary mRNA-laden aggregates that shut down protein translation to help cells conserve energy during stressful events. Suspected pathological stress granules have been associated with TDP-43 (see ARF related news story), FUS (see ARF related news story on Dormann et al., 2010), and tau in Alzheimer’s disease (see ARF related news story and ARF news story on Vanderweyde et al., 2012).
Taylor initially believed that familial multisystem proteinopathy was solely due to VCP mutations, but over time he identified more than a dozen MSP families that possessed normal VCP. Co-first author Hong Joo Kim found the hnRNP mutations when she used exome sequencing to analyze DNA from two of those families. The hnRNPA2B1 and hnRNPA1genes are closely related, and in each case the mutation replaced a conserved aspartate with valine. The researchers, including co-first authors Nam Chul Kim, Yong-Dong Wan, and Jennifer Moore, were intrigued because of the similarities between the hnRNPs and TDP-43. “They are paralogous proteins,” Taylor said. They also work together. TDP-43 binds both hnRNPs, and all three regulate mRNA splicing (Buratti et al., 2005). Further, the researchers thought it important that the aspartate-to-valine substitution occurred in the same carboxyl-terminal, glycine-rich, prion-like region where many TDP-43 ALS mutations turn up.
Taylor contacted collaborators to check if the genetic association held up in other families. Christopher Shaw of King’s College London, U.K., and Rosa Rademakers of the Mayo Clinic in Jacksonville, Florida, each found a person with ALS who harbored an hnRNPA1 mutation. Shaw found a substitution at the same aspartate that Taylor’s group identified, but in this case the mutation was to an asparagine. Rademakers identified a nearby asparagine-to-serine mutation. With four instances of hnRNP variants found in fewer than 600 families, Taylor concluded that, while the mutations are uncommon, they are not extremely rare.
To understand how the mutations affected the hnRNP proteins, Taylor collaborated with Shorter's group, including co-first authors Emily Scarborough and Zamia Diaz. The researchers purified recombinant versions of the two hnRNPs and found they formed fibrils in vitro. The disease-linked mutants fibrillized faster. Based on computational analysis of the amino acid sequence, the researchers ascribe this aggregation to the steric zipper motif. “The mutations really intensify the strength of that binding,” Shorter said. The resulting complexes, then, act like stuck zippers.
Kim found that in HeLa cervical cancer cultures the mutations affected stress granule assembly, which depends on prion-like protein domains (Kato et al., 2012). Though the wild-type proteins were normally nuclear, they relocated to cytoplasmic stress granules when Kim treated the cells with arsenite. In contrast, the mutant hnRNPs appeared in stress granules even before the treatment. After it, additional mutant hnRNPs joined stress granules, and did so more quickly than the wild-type proteins.
The results led Taylor to propose that MSP is a disease of pathological stress granules held together by stuck zippers. Stress granules are supposed to dissolve when the cells' situation improves. “You normally have an equilibrium,” said Ben Wolozin of Boston University, who was not involved with the Nature study. “If you shift that towards aggregation, then the system begins to develop stress granules that are too stable and too big.” This disequilibrium could underlie the formation of many neurodegeneration-linked aggregates, Wolozin has posited (Wolozin, 2012).
What turns a short-lived stress granule into a disease-causing aggregate? An additional stressor could initiate the transition, Haass proposed, and hnRNP mutations could fit the bill. In that way, he suggested, mutant hnRNPs might seed or even spread aggregation because their zippers are too tight. Alternatively, in people with VCP mutations, the chaperone is unable to unzip the complexes. In both cases, aggregates could cause disease by sequestering RNAs and their binding proteins, Taylor suggested.
The discovery of hnRNPA2B1 and A1 mutations in ALS and MSP fits nicely with a hypothesis that Shorter, along with collaborator Aaron Gitler of Stanford University in Palo Alto, California, has been pursuing. They propose that a broad set of RNA-binding proteins with prion-like domains contribute to neurodegeneration by converting normally folded protein into toxic conformations. Based on computational analysis, the researchers put together a list of proteins with such features that they are checking for prion-like activity and participation in neurodegeneration (King et al., 2012). In fact, hnRNPA2B1 and hnRNPA1 were on Shorter’s list of top 10 gene candidates before he heard from Taylor.
The possibility of prion-like spread could explain the known pathology of ALS and MSP, which tends to start in one part of the body and move to neighboring tissues, noted Amelie Gubitz of the National Institute of Neurological Disorders and Stroke in Bethesda, Maryland, in an e-mail to Alzforum (see full comment below; see also Ravits and La Spada, 2009). This kind of cell-to-cell, templated transmission has already been reported for tau (see ARF related news story on Luk et al., 2012), Aβ (see ARF related news story), and the ALS protein SOD1 (see ARF related news story). While Taylor and colleagues have yet to directly address prion-like transmission, they did show that each mutant protein could seed fibril formation with their wild-type counterparts.
The study authors expect that additional RNA-binding proteins with prion-like segments will join TDP-43 and the hnRNPs as factors in neurodegeneration. Taylor has DNA from other MSP families to sequence, and he anticipates discovering mutations in genes that either bind RNAs, as do the hnRNPs, or disassemble RNA-protein complexes, as does VCP. Shorter and Gitler are exploring the other genes on their list by using yeast as a model system and by staining pathological tissue samples for protein aggregates.—Amber Dance.
Kim HJ, Kim NC, Wang YD, Scarborough EA, Moore J, Diaz Z, MacLea KS, Freibaum B, Li S, Molliex A, Kanagaraj AP, Carter R, Boylan KB, Wojtas AM, Rademakers R, Pinkus JL, Greenberg SA, Trojanowski JQ, Traynor BJ, Smith BN, Topp S, Gkazi AS, Miller J, Shaw CE, Kottlors M, Kirschner J, Pestronk A, Li YR, Ford AF, Gitler AD, Benatar M, King OD, Kimonis VE, Ross ED, Weihl CC, Shorter J, Taylor JP. Mutations in prion-like domains in hnRNPA2B1 and hnRNPA1 cause multisystem proteinopathy and ALS. Nature. 3 Mar 2013. Abstract