Opposites attract, sometimes so much so that they retreat into their own little bubble. This seems to be the case with tau, a positively charged protein, and RNA, a negatively charged nucleic acid. According to a study published July 6 in PLOS Biology, the pair tightly coupled together in neurons, and coalesced into droplets in a dish. While tau appeared unfettered by the close quarters at first, the protein started showing early signs of fibrillization after prolonged residency in the droplets. The researchers, led by Kenneth Kosik and Songi Han at the University of California, Santa Barbara, proposed that these droplets could play a role in tau aggregation within neurons.

The findings add tau to the growing list of neurodegenerative disease-associated proteins known to undergo liquid-liquid phase separation (LLPS), a phenomenon increasingly implicated in all manner of cellular functions. LLPS occurs when proteins and/or nucleic acids associate closely, changing the viscosity of their local environment (see Li et al., 2012). In so doing, they can form transient “membrane-less organelles,” like pop-up shops, to accomplish specific cellular functions. Stress granules, the nucleolus, and even lipid rafts of signaling receptors are examples. RNA-binding proteins, and other sticky proteins that contain so-called low-complexity domains, are common inhabitants of these liquid organelles. Neurodegenerative bad boys TDP-43, FUS, C9ORF72 dipeptide repeats, hnRNPA1, and hnRNPA2B1 are among those spotted within the droplets, where they are proposed to derail the machinery within them (Oct 2015 webinarOct 2016 news; May 2017 conference news). 

The idea that tau, too, could dabble in droplets stems from observations that it associates with RNA. Though tau is not a bona fide RNA-binding protein or a bearer of low-complexity domains, the protein’s positive charge and intrinsic disordered state make it a prime candidate for an electrostatic liaison with RNA. Previous studies reported that tau’s association with RNA coaxed the protein into fibrils, though with less vigor than the polyanionic aggregation inducer, heparin (see Kamper et al., 1996Wang et al., 2006). 

In the present study, first author Xuemei Zhang and colleagues sought to investigate the nature of tau’s association with RNA. Using a cross-linking technique called PAR-iCLIP, the researchers found that both wild-type and mutant forms of tau associated with RNA in human embryonic kidney (HEK) 293T cells and in neurons derived from induced pluripotent stem cells (iPSCs). A closer inspection of the RNA bound to tau revealed that they were predominantly tiny species.

To the researchers’ surprise, transfer RNA (tRNAs) made up the overwhelming majority of tau’s RNA partners. While tau comingled with many tRNAs, it gave preferential treatment to some. Of the 231 tRNAs tau buddied up with iPSC-derived neurons, it preferred tRNAArg most. The researchers found that at low tau-to-RNA ratios, tau associated with RNA species as a dimer, while ratcheting up the tau concentration generated much larger tau-to-RNA complexes. 

Drip Drop.

Mix tau and RNA together, and droplets form. They merge together (left panel, red circle) and contain tau (right, green). [Courtesy of Zhang et al., PLOS Biology, 2017.]

Mixing tau and RNA together outside of cells created a turbid solution. Under the brightfield microscope, the researchers spied droplets full of tau and RNA. The droplets formed when using full-length tau or Δtau187, and when using tRNA, poly(A)RNA, or poly(U)RNA. Regardless of the type of RNA molecule, droplets existed in a 1:1 charge ratio with their tau associates. They were highly dynamic, readily merged, and the researchers found they could toggle droplet formation up or down by adjusting protein/RNA concentrations, salt concentration, pH, or temperature. Importantly, droplets formed under physiological conditions mimicking those in neurons.

How did the droplet-bound life change tau? Not much at first, the researchers reported. Spin-labeling experiments of tau/RNA droplets revealed that tau did not change its conformation despite its high concentration inside droplets. In contrast, heparin induced dramatic changes corresponding to β-sheet formation. However, Thioflavin T (ThT) fluorescence, an indicator of β-sheets, did gradually rise over 15 hours after the droplets formed. Like other groups studying liquid-liquid phase transition of neurodegenerative disease proteins, Kosik and colleagues suggested that, given enough time and under certain conditions, the droplets could be a prime venue for tau aggregation. In support of this idea, they saw an increase in sarkosyl-insoluble tau when they transfected iPSC-derived neurons with an overload of tRNA.

Kosik told Alzforum that the negatively charged RNA in the droplet likely beckons tau to squeeze more closely together than it would without the attractive charge, and to do so without changing its conformation. A major caveat of the study is that it provides no proof that the tau/RNA droplets exist within cells, Kosik acknowledged. However, he said that if the droplets do form in cells, their integrity could be modulated by cellular stress, a state known to change parameters such as pH and salt concentration. “We think the in vitro tuning conditions we observed for the droplets may have some correlates to stress condition in a living cell,” he said.

Benjamin Wolozin of Boston University praised the study for its rigor, and agreed that cellular stress would influence such droplets inside a cell. For his part, Wolozin has reported that tau facilitates the formation of stress granules, a type of membrane-less organelle that sequesters nonessential RNA transcripts from translation (May 2016 news). He also reported that in response to stress, tau moves from axons into the somatodendritic compartment, where it is likelier to encounter RNA. Perhaps when stress goes on too long, this relationship promotes tau aggregation, he said. This would align with the inklings of aggregation that Kosik and colleagues observed hours after tau/RNA droplets formed.

Wolozin was intrigued by tau’s preference for tRNA, noting that if this association is confirmed, it could explain previous observations that tau inhibits translation (Meier et al., 2016). Perhaps tRNAs deliver tau to the translation machinery, where it gums up the works, he said.

Though still unpublished, findings from other researchers are converging on the idea that tau forms droplets. At a recent meeting in Leuven, Belgium, Susanne Wegmann of Massachusetts General Hospital in Charlestown reported as much. In collaboration with Anthony Hyman at the Max Planck Institute of Molecular Cell Biology and Genetics in Dresden, Germany, Wegmann found that the phosphorylation of tau facilitated droplet formation, and that the droplets could form in neurons (May 2017 conference news). Wegmann speculated that the droplets could provide cells with a ready supply of tau for stabilization of microtubules. This idea is supported by findings the researchers recently posted on bioRχiv (Hernandez-Vega et al., 2017).

Markus Zweckstetter of the Max Planck Institute for Biophysical Chemistry in Gottingen, Germany, will also soon report on liquid phase separation of tau, including differences between tau splice variants in their tendency to form droplets. Zweckstetter told Alzforum that contents of the tau droplets he observed differ from those described in Kosik’s work. Stay tuned for details when Alzforum covers this upcoming paper.—Jessica Shugart 


  1. This work, from the laboratories of Ken Kosik and Songi Han, provides important new insight into the biology of tau protein. The manuscript builds on a growing body of work that is revolutionizing our knowledge of how proteins in the cell interact. For a very long time, we have known that there are many, many proteins that exhibit “sticky,” intrinsically disordered regions. These regions were thought to be a “nuisance,” but our understanding of their role changed dramatically with publication of a manuscript coming from Mike Rosen’s laboratory (Li et al., 2012). This group showed that proteins with multiple intrinsically disordered regions associate in a manner leading to structures analogous to “lipid droplets,” in that they will phase-separate from aqueous solutions. This phase separation turns out to be a major way that the cell organizes its functions, and multiple well-known structures exist based on the biology of liquid-liquid phase separation (LLPS); these structures include the nucleolus and the nuclear pore (Feric et al., 2016; Zhu and Brangwynne, 2015). 

    RNA binding proteins turn out to be one of the major groups of proteins that exhibit the properties of LLPS (Ash et al., 2014; Gitler and Shorter, 2011; Banani et al., 2017; Feric et al., 2016). The phase separation gives rise to RNA granules, which function in the cell to control RNA transport, translation, degradation, and sequestration (Anderson and Kedersha, 2008). Sequestration occurs during stress, when the cell needs to focus RNA translation on protective/reparative transcripts, and sequesters unnecessary transcripts into structures termed stress granules (Panas et al., 2016). Several groups, including my own, extended this idea to the realm of neurodegenerative disease by showing that RNA-binding proteins that are associated with disease (such as TDP-43 and FUS) associate with stress granules, and that disease-linked mutations increase formation of stress granules (Bosco et al., 2010; Colombrita et al., 2009Liu-Yesucevitz et al., 2010). 

    My group demonstrated that tau pathology associates with RNA binding proteins (Vanderweyde et al., 2012). More recently, we demonstrated two surprising results. First, we showed that tau functions during stress to promote stress granule formation; secondly, that the association of tau with stress granules increases tau’s tendency to aggregate (Vanderweyde et al., 2016). In a related paper, Joe Abisambra’s team demonstrated that pathological tau associates with the ribosome and inhibits RNA translation, which is a predicted outcome of the stress granule/translational stress response (Meier et al., 2016). Based on this observation, we also demonstrated that reducing the RNA binding protein TIA1 inhibits tau pathophysiology.

    Enter the present manuscript. It shows that tau undergoes liquid-liquid phase separation in the presence of RNA, and at concentrations as low as 2 μM, which approaches the concentration of tau in the neuron. The group demonstrates the phenomenon of liquid-liquid phase separation using multiple independent approaches. Importantly, they also demonstrate that tRNA has a particularly strong affinity for tau, with a Kd=460 nM, although nonspecific RNA also stimulates tau LLPS. The type of tau structure that forms when associated with RNA depends on the ratio of RNA:tau. At low ratios, the association stimulates formation of oligomers, while at high ratios, the association stimulates larger complexes and more robust LLPS. The LLPS is also sensitive to other conditions, such as salt and temperature.

    This work complements our recent work demonstrating that tau regulates stress granule biology, and provides strong support for a new vision suggesting roles for LLPS, stress granule formation, and the translational stress response in the pathophysiology of tauopathy.


    . Stress granules: the Tao of RNA triage. Trends Biochem Sci. 2008 Mar;33(3):141-50. PubMed.

    . Pathological stress granules in Alzheimer's disease. Brain Res. 2014 Oct 10;1584:52-8. Epub 2014 Aug 7 PubMed.

    . Biomolecular condensates: organizers of cellular biochemistry. Nat Rev Mol Cell Biol. 2017 May;18(5):285-298. Epub 2017 Feb 22 PubMed.

    . Mutant FUS proteins that cause amyotrophic lateral sclerosis incorporate into stress granules. Hum Mol Genet. 2010 Nov 1;19(21):4160-75. Epub 2010 Aug 10 PubMed.

    . TDP-43 is recruited to stress granules in conditions of oxidative insult. J Neurochem. 2009 Nov;111(4):1051-61. Epub 2009 Sep 16 PubMed.

    . Coexisting Liquid Phases Underlie Nucleolar Subcompartments. Cell. 2016 Jun 16;165(7):1686-97. Epub 2016 May 19 PubMed.

    . RNA-binding proteins with prion-like domains in ALS and FTLD-U. Prion. 2011 Jul 1;5(3) PubMed.

    . Phase transitions in the assembly of multivalent signalling proteins. Nature. 2012 Mar 7;483(7389):336-40. PubMed.

    . Tar DNA binding protein-43 (TDP-43) associates with stress granules: analysis of cultured cells and pathological brain tissue. PLoS One. 2010;5(10):e13250. PubMed.

    . Pathological Tau Promotes Neuronal Damage by Impairing Ribosomal Function and Decreasing Protein Synthesis. J Neurosci. 2016 Jan 20;36(3):1001-7. PubMed.

    . Mechanistic insights into mammalian stress granule dynamics. J Cell Biol. 2016 Nov 7;215(3):313-323. PubMed.

    . Interaction of tau with the RNA-Binding Protein TIA1 Regulates tau Pathophysiology and Toxicity. Cell Rep. 2016 May 17;15(7):1455-1466. Epub 2016 May 6 PubMed.

    . Contrasting pathology of the stress granule proteins TIA-1 and G3BP in tauopathies. J Neurosci. 2012 Jun 13;32(24):8270-83. PubMed.

    . Nuclear bodies: the emerging biophysics of nucleoplasmic phases. Curr Opin Cell Biol. 2015 Jun;34:23-30. Epub 2015 May 15 PubMed.

Make a Comment

To make a comment you must login or register.


Webinar Citations

  1. Fluid Business: Could “Liquid” Protein Herald Neurodegeneration?

News Citations

  1. ALS Research ‘Gels’ as Studies Tie Disparate Genetic Factors Together
  2. Protein Liquid-Liquid Phase Transitions: The Science Is About to Gel
  3. Stress Granule Protein Entwines and Misfolds Tau

Paper Citations

  1. . Phase transitions in the assembly of multivalent signalling proteins. Nature. 2012 Mar 7;483(7389):336-40. PubMed.
  2. . RNA stimulates aggregation of microtubule-associated protein tau into Alzheimer-like paired helical filaments. FEBS Lett. 1996 Dec 16;399(3):344-9. PubMed.
  3. . The proline-rich domain and the microtubule binding domain of protein tau acting as RNA binding domains. Protein Pept Lett. 2006;13(7):679-85. PubMed.
  4. . Pathological Tau Promotes Neuronal Damage by Impairing Ribosomal Function and Decreasing Protein Synthesis. J Neurosci. 2016 Jan 20;36(3):1001-7. PubMed.
  5. . Local Nucleation of Microtubule Bundles through Tubulin Concentration into a Condensed Tau Phase. Cell Rep. 2017 Sep 5;20(10):2304-2312. PubMed.

Further Reading


  1. . Mechanistic insights into mammalian stress granule dynamics. J Cell Biol. 2016 Nov 7;215(3):313-323. PubMed.

Primary Papers

  1. . RNA stores tau reversibly in complex coacervates. PLoS Biol. 2017 Jul;15(7):e2002183. Epub 2017 Jul 6 PubMed.