RGG/RG domains are common in RNA-binding domains: RNA-recognition motifs (RRM) is the most abundant RNA-binding domain (RBD) and frequently found in the heterogeneous ribonucleoprotein particle (hnRNP) family of proteins (Left). RGG/RG domains also frequently appear in proteins possessing other RBDs and may broaden the specificity of these proteins. (Right) The RGG/RG domains of FET proteins (FUS, EWSR1, and TAF15) are conserved beyond vertebrates.
RGG/RG domains are conserved: RGG/RG domains are enriched in arginines and glycines and yet show diversity in the number and spacing of each RGG/RG. However, for specific proteins this sequence remains conserved suggesting the number and positioning of RGG/RG repeats is not random and has functional significance (Akdogan et al., in submission). Binding of RNA by the RGG/RG domains induces a structural shift in FUS that seeds oligomerization.
Working model for RGG/RG recognition of RNA: RGG/RG domains bind complex RNA structures containing double stranded RNA and do not bind simple double stranded RNA (Akdogan et al., in submission). We hypothesize that two published structures of RGG/RG and RNA complexes demonstrate a general model that nearby structures serve to widen the major grove allowing access for RGG/RG domains (reference shown, left).
FUS spontaneously assembles into liquid-like droplets and then hydrogels: Yes, that is protein. FUS spontaneously assembles into these (Left above), particularly at low temperatures (near freezing). To study these assemblies in vitro, we prefer to allow them to adhere to glass surfaces in order that we can perform binding assays, in which case they have an amorphous shape (second from right). But when unattached from the glass, they can “snap” back into spherical shapes and hence earned the name “droplets”.
Many non-membrane bound organelles compartmentalize cellular biology: (Left) Numerous granular bodies have been described in cells, particularly within the nucleus. One function of FUS in cells is to bind the C-terminal domain (CTD) of RNA Pol II to effect transcription. We can demonstrate this binding using our in vitro assembled droplets. Thus, we hypothesize a novel cellular granule that assembles around active gene promoters. (Right) The LC domain of FUS assembles into protein fibers. RGG/RG domains assemble into fibers as well. Both bind the CTD of RNA Pol II, the significance of which is currently unknown.
Purification of FUS droplets from cells. (A) Cross-linked cell lysates were fractionated by SEC with FUS and RNA Pol II eluting together in complexes between 70-150 nm complexes. For comparison, the crystal structure of the mammalian RNA Pol II holoenzyme is only 25 nm in diameter. (B) Small DNA fragments are protected from degradation by benzonase or sonication and co-eluted with RNA Pol II complexes. Consistent with this observation, next-generation sequencing reveals that these fragments are enriched for genomic regions that are also enriched for RNA Pol II. (C) Droplets were visualized by TEM and diameters confirmed (bar = 100 nm). (Right) In violation of all that is holy in SEC chromatography, we have developed a custom SEC protocol for separating and detecting very large complexes of RNA Pol II and other granular bodies in nuclear and whole cell lysates.
Amyotrophic Lateral Sclerosis (ALS): ALS is a disease of neurons that control muscle movement, called motor neurons. Motor neurons reside both in the cortex and in the spinal cord. Loss of muscle control and muscle wasting (atrophy) is due to motor neurons dying. The disease progresses rapidly with most patients surviving less than 3 years after diagnosis. Age of onset varies considerably. Within Tucson AZ, approximately 15 cases of ALS diagnosed each year. At any time, approximately 25 to 30 patients are currently living with ALS (KS personal communication).
The Normal Function of FUS: (A) FUS is a multi-domain RNA-binding protein with an N-terminal Low Complexity (LC) domain. The most prevalent ALS-causing mutations in FUS are found in the nuclear localization signal (NLS). (B) FUS binds RNA Pol II (RNAP2) and co-localizes with the polymerase to active gene promoters. (C) FUS is one of the most abundant proteins found in the nucleus. It targets transcription by recognizing the earliest transcripts produced by RNA Pol II, and then forms a fibrous scaffold, which recruits more RNA Pol II, other transcription machinery, and regulates phosphorylation of the CTD of RNA Pol II.
Study of FUS function in ALS patient-derived cells: (A) An invaluable resource to ALS research are fibroblast cells donated by ALS patients. (B) Despite mutations to the nuclear localization signal of FUS, in patient samples most of the FUS protein remains in the nucleus. Whereas pathology had been presumed to be associated with toxic cytoplasmic activity of the mislocalized protein, we have explored whether nuclear dysfunction may contribute significantly to pathology.
Aberrant FUS function in ALS patients: Despite the majority of FUS being in the nucleus in ALS patients bearing FUS mutations (A), nuclear FUS is trapped in aggregates that cannot migrate on an SDS-PAGE gel and not functional (B). (C) Genes actively transcribed by RNA Pol II are localized within the cell nucleus to discrete foci and this open chromatin stains weaker for DAPI. For ALS patients with FUS mutations, this organization is disrupted and the foci of RNA Pol II becomes fragmented. (D) Our model is that FUS forms droplets, which are scaffolds surrounding expressed genes. We find in ALS patient cells, FUS droplets in the nucleus become aggregate and no longer function to regulate transcription or help organize the transcription machinery.
Affinity-enrichment mass spectrometry reveals interactions shared by FUS and TDP-43: Use of improved mass spectrometry approaches greatly enhances sensitivity. Our preliminary analysis of recent data revealed known interactors, such as NONO, FXR1, and hnRNPU. We also found interactors with known significance for ALS pathology, such as hnRNPA2B1, PPIA, GABBR1, and TUBBA4A.
The role of TDP43 in transcription regulation: (A) Using a next-generation sequencing technique, GRO-seq, we can visualize transcription as it is occurring across the entire genome. Then by removing TDP-43 by siRNA knockdown, we can determine the affect TDP43 has on transcription. (B) Multiple models may explain TDP43 regulation of transcription. TDP43 may be recruited to regulate transcription by binding the transcribed RNA (i) or the DNA (ii). It may bind enhancer RNAs (eRNA) upstream of the gene (iii). TDP43 is a constitutive member of paraspeckles, which raises the possibility of a connect between these bodies and transcription (iv, West et al., J Cell Biol 2016). (C) By replacing TDP43 with those deleting key domains, the mechanism of transcription regulation can be further elucidated.
Multiple genes implicated in ALS are known to be involved in ALS pathology: FUS, VCP, and SETX have all been implicated as directly interacting with DNA repair machinery. SOD1 and p62 have roles in autophagy, which is linked to DNA damage through regulating the production of radical oxidative species (ROS). The role of TDP-43 remains to be determined.
Identification of proteins interacting with FUS and TDP-43 after DNA damage: Heat maps of z-scores indicate change in levels of proteins detected after DNA damage, with green increases and red decreases. Proteins where changes rose to the level of statistic significance indicated by blue bars. Currently, we hypothesize that FUS and TDP-43 cohabit several nuclear bodies, though not interacting directly. After DNA damage, FUS leaves many of its unique complexes and increases its association with complexes that also contain TDP-43. Few interactions for TDP-43 significantly change. Of known interactors of FUS and TDP-43, paraspeckles and the snRNP complex seem unchanged by DNA damage, RNA exosome interactions increase, and hnRNP interactions decrease.
Formation of protein polymers by FUS-Fli1 and EWS-Fli1: (A) The translocation event driving Ewing’s sarcoma fuses the LC domain of FUS or EWSR1 to the DNA-binding domain (DBD) of Fli1, an ETS family transcription factor. (B) In Ewing’s sarcoma, Fli1 fusion proteins bind repetitive GGAA DNA sequences, which raises the hypothesis that these sequences promote oligomerization of the FET-fusion proteins along the DNA. Therefore, like FUS and EWSR1, it may be that the oligomers of FUS-Fli1 and EWS-Fli1 are the active forms of the proteins. Recruitment of alternative co-factors may explain the difference in outcomes to either activate or repress gene expression. (C) We have formed in vitro reconstituted droplets of FET proteins or their LC domains. The LC domain readily incorporates into FUS droplets, suggesting that Fli1 fusion proteins may seed higher order interactions with other FET proteins, including FUS and EWSR1.
FET proteins, and their oncogenic fusions, form phase-separated droplets: The oligomer form of FET proteins are active form to regulate transcription (A). Oligomers of FET proteins coalesce into phase-separated droplets that serve as a molecular scaffold and concentrates binding partners in vitro and in cells (B). The underlying organization of proteins within these droplets remains to be determined. (C) In Ewing’s sarcoma, we hypothesize that the Fli1 fusion proteins reorganize chromatin to form aberrant regions of active gene expression, thereby reprograming mesenchymal stem cells into Ewing’s sarcoma.
Targeting FET protein functions increase Ewing’s sarcoma sensitivity to DNA damage: (A) Our first drug hit RS003 has a synergistic effect with the DNA damage agent SN38 inducing potent cell death at 96 hours for Ewing’s sarcoma but not a non-Ewing’s cell line, U2OS, as measure by an MTT assay. (B) Using cell viability assays, inhibition of PARP with a drug olaparib combined with DNA damage induced by SN-38, which is currently in clinical trials, enhances cell death at 72 hrs but not at 12 hrs. (C) Inhibiting FET family co-factors involved in transcription (MR03) dramatically enhances cell death due to DNA damage for Ewing’s sarcoma cells after only 12 hours of treatment.
