Multiscale simulations reveal TDP-43 molecular-level interactions driving condensation

Abstract

The RNA-binding protein TDP-43 is associated with mRNA processing and transport from the nucleus to the cytoplasm. TDP-43 localizes in the nucleus as well as accumulating in cytoplasmic condensates such as stress granules. Aggregation and formation of amyloid-like fibrils of cytoplasmic TDP-43 are hallmarks of numerous neurodegenerative diseases, most strikingly present in >90% of amyotrophic lateral sclerosis (ALS) patients. If excessive accumulation of cytoplasmic TDP-43 causes, or is caused by, neurodegeneration is presently not known. In this work, we use molecular dynamics simulations at multiple resolutions to explore TDP-43 self- and cross-interaction dynamics. A full-length molecular model of TDP-43, all 414 amino acids, was constructed from select structures of the protein functional domains (N-terminal domain, and two RNA recognition motifs, RRM1 and RRM2) and modeling of disordered connecting loops and the low complexity glycine-rich C-terminus domain. All-atom CHARMM36m simulations of single TDP-43 proteins served as guides to construct a coarse-grained Martini 3 model of TDP-43. The Martini model and a coarser implicit solvent C⍺ model, optimized for disordered proteins, were subsequently used to probe TDP-43 interactions; self-interactions from single-chain full-length TDP-43 simulations, cross-interactions from simulations with two proteins and simulations with assemblies of dozens to hundreds of proteins. Our findings illustrate the utility of different modeling scales for accessing TDP-43 molecular-level interactions and suggest that TDP-43 has numerous interaction preferences or patterns, exhibiting an overall strong, but dynamic, association and driving the formation of biomolecular condensates.

Figure 1

Full-length TDP-43 dynamics and self-interactions. (A) Models of full-length, 414 amino acid long, TDP-43 were constructed at the all-atom (AA) and coarse-grained (CG) scales. Splayed out snapshots are shown for both scales with the main structural domains color coded as shown below. (B) Secondary structure of AA TDP-43 shown for full-length TDP-43 with and without RNA fragment. Eight repeated simulations (from 0.2 to 2 μs) are averaged and average ± SD shown (see Fig. S1 for secondary structure of additional simulations). (C) Radius of gyration (Rg) for TDP-43 from eight 2 μs long AA simulations and eight 10 μs long CG simulations. (D) Average TDP-43 Rg from different simulation conditions. Each is an average ± SE of eight simulations from the 0.4–2 μs for the AA and 2–10 μs for the CG simulations. (E and F) Residue-residue contact maps for the (E) AA and (F) CG resolutions; each is from a single representative simulation averaged over 0.2–2 μs simulation time (white is no data and see Fig. S3 for averages of different times and repeats and Fig. S6 for averages of the different protein variant simulations). To see this figure in color, go online.

Figure 2

TDP-43 cross-interactions. (A) Representative snapshots from one c_full simulation, the backbone of the two TDP-43 proteins is colored differently in cyan and white, and the structure domains are colored according to Fig. 1A with the backbone beads of the two proteins in cyan and white. The left snapshot is from early in the simulation before the two proteins associated. The middle and right snapshots are from close to the end of the simulations and 15 ns apart. (B) Residue-residue contact map between the two c_full proteins, averaged over the eight simulation repeats excluding the first 2 μs of each simulation. (C) Average residue-residue cross-interaction contacts shown for each protein residue, error bars are SE between the eight simulation repeats. (D) Total average contacts between the different protein variants tested and (E) the average contact for the helix region (residues 311–360) only (error bars are SE between the eight simulation repeats). For systems with bound RNA, average contacts are shown including (black) and excluding (gray) RNA. Contact maps and average contacts per residues for the all the variants are shown in Fig. S7. To see this figure in color, go online.

Figure 3

TDP-43 assemblies. (A) Representative snapshots from CG Martini 3 simulations capturing TDP-43 assemblies, shown for a simulation with 24 TDP-43 c_full at time 0 and after 10 μs of simulation where all TDP-43 proteins have assembled. Proteins are colored according to the same scheme as in Fig. 2A, with all protein back bone beads in cyan. (B) Relative distributions of CG Martini 3 particles along the z-dimension of the slab, relative to the center of the box and averaged over the first (dotted light lines) and the last (9–10 μs, solid lines) microsecond of the simulation. (C) Representative snapshot of TDP-43 condensate from a CG slab simulation based on the HPS-Urry model (top). The concentration profile (bottom) of TDP-43 WT versus 6WtoA along the z-dimension of the slab at 310 K. (D) Pairwise intermolecular contact map (top) and per-residue contact probabilities (bottom) in the condensed phase from the WT HPS-Urry CG simulations. To see this figure in color, go online.

Figure 4

A diverse array of interdomain interactions collectively stabilize the TDP-43 condensed phase. The balance between NTD-NTD (via head-to-tail) and CTD-CTD (facilitated by CR-CR and impacted by IDRs) interactions critically determines the formation of soluble, liquid-like condensates. UG-rich RNAs play an important role in maintaining the solubility of TDP-43 by binding to tandem RRMs and CTD (via RGG motifs). Simulations performed in this study suggest that RRMs can also interact with both NTD and CTD, which hints at their additional roles in modulating the LLPS of TDP-43. To see this figure in color, go online.