Mutations linked to neurological disease enhance self-association of low-complexity protein sequences

Abstract

Protein domains of low sequence complexity do not fold into stable, three-dimensional structures. Nevertheless, proteins with these sequences assist in many aspects of cell organization, including assembly of nuclear and cytoplasmic structures not surrounded by membranes. The dynamic nature of these cellular assemblies is caused by the ability of low-complexity domains (LCDs) to transiently self-associate through labile, cross-β structures. Mechanistic studies useful for the study of LCD self-association have evolved over the past decade in the form of simple assays of phase separation. Here, we have used such assays to demonstrate that the interactions responsible for LCD self-association can be dictated by labile protein structures poised close to equilibrium between the folded and unfolded states. Furthermore, missense mutations causing Charcot-Marie-Tooth disease, frontotemporal dementia, and Alzheimer's disease manifest their pathophysiology in vitro and in cultured cell systems by enhancing the stability of otherwise labile molecular structures formed upon LCD self-association.

Fig. 1.. Preparation of semi-synthetic derivatives of the TDP-43 LC domain carrying single, methyl-capped peptide backbone nitrogen atoms.

(A) Schematic representation of the LC domain of TDP-43 (residue 262–414) with an ultra-conserved region (residue 316–339) highlighted by darker shading. N^α-methyl amino acid (red) scanning analysis was applied on the ultra-conserved region. (B) Schematic depicting of the strategy to prepare N^α-methyl amino acid (red asterisk)-incorporated TDP-43 LC domains. N^α-methyl amino acids were introduced into synthetic peptides that were conjugated via sequential native chemical ligation reactions to flanking protein fragments produced in bacteria. Ligation products contained cysteine residues in place of alanine residues at both ligation boundaries. Both cysteine residues were chemically desulfurized to alanine such that the only difference between engineered LC domains and the native LC domain was the presence of a single N^α-methyl group at the desired site. (C) Representative result of HPLC trace (left) and high resolution, intact mass spectrometry (right) for confirming ligation production from (B). (D) Light microscopic analyses of phase separation for three representative samples, including the reconstructed, semi-synthetic, native TDP-43 LC domain (WT), a variant carrying a methyl cap on the peptide backbone nitrogen atom of alanine residue 324 (324 meA), and a variant carrying a methyl cap on the peptide backbone nitrogen of methionine residue 337 (337 meM). Each protein was incubated under conditions of neutral pH and physiologically normal monovalent salts allowing for formation of phase separated liquid-like droplets. Droplet formation was assayed in triplicate in normal buffer as well as buffers supplemented by 0.4 M urea, 0.8 M urea and 1.2 M urea. Scale bar = 25 μm. (E) Quantitative analysis of phase separation for all protein samples was monitored by spectroscopic analysis of turbidity.

Fig. 2.. Phase separation capacities of twenty three variants of the TDP-43 LC domain differing according to the presence of a single, methyl-capped peptide backbone nitrogen atom.

(A) Purity and integrity evaluation of all N^α-methyl variants of the TDP-43 LC domain by SDS-PAGE. (B) Normalized turbidity measurement of TDP-43 LC domain phase separation induced in buffer of neutral pH and physiological monovalent salt, or in buffers supplemented by 0.4 M urea, 0.8 M urea and 1.2 M urea. All samples were analyzed in triplicate and further evaluated and photographed by light microscopy (Fig. S1). Turbidity measurements revealed two classes of variants irrespective of the presence or absence of urea. One class composed of nine variants showed, relative to the native TDP-43 LC domain, reduced turbidity and fewer liquid-like droplets under all conditions. The other class, composed of thirteen variants, showed evidence of phase separation indistinguishable from the native protein. (C) Histograms reveal the normalized, average turbidity measurements of twenty four protein samples as evaluated in the absence of urea as well as in the presence of either 0.4 M or 0.8 M urea. Turbidity is graphed on the Y axis of each plot relative to the amino acid sequence of the ultra-conserved region of the TDP-43 LC domain between phenylalanine residue 316 and methionine residue 339 (X axis). Irrespective of assay condition, all nine of variants exhibiting impediments to phase separation clustered between alanine residue 321 and alanine residue 329. (D) Reproductions of published data pinpointing the region of the TDP-43 LC domain footprinted to reveal structure-dependent protection from methionine oxidation (19). (E) Molecular structure of cross-β polymers formed from the TDP-43 LC domain as resolved by cryo-electron microscopy (EM) (27). The region of the protein that was observed to be of the same molecular structure in three independent cryo-EM reconstructions extends from proline residue 320 to methionine residue 337. (F) Top image shows ribbon diagram representations of residues 314–327 from one of the three cryo-EM structures of cross-β polymers made from the TDP-43 LC domain shown in (E) (27). Bottom image shows ribbon diagram representation of residues 314–327 from a different cryo-EM structure of cross-β polymers made from the intact TDP-43 LC domain (32).

Fig. 3.. Preparation and analysis of semi-synthetic derivatives of the TDP-43 LC domain carrying three methyl-capped peptide backbone nitrogen atoms.

(A) Synthetic peptides carrying methyl caps on the peptide backbone nitrogen atoms of residues A321, M323 and A325 (Triple-me (321–323-325)), or residues M323, A325 and Q327 (Triple-me (323–325-327)), were inserted into the full-length LC domain of TDP-43 by chemical ligation as described in Fig. 1. Purity and integrity of the ligation products were evaluated by SDS-PAGE (top panel). Phase separation of the ligation products was quantified by turbidity measurement (bottom panel). The phase separation was induced in buffer of neutral pH and physiological monovalent salt, supplemented with no urea, 0.4 M urea, 0.8 M urea, or 1.2 M urea. (B) Presentation of normalized turbidity measurements in the form of histograms. No evidence of turbidity above background was observed for either sample bearing three methyl-capped peptide backbone nitrogen atoms when assayed in the presence of 0.4 M, 0.8 M or 1.2 M urea. In the absence of supplemented urea, both samples bearing three methyl-capped nitrogen atoms displayed turbidity levels reduced by between 60% and 70% relative to that of the native TDP-43 LC domain. (C) Light microscopic analyses of phase separation for native TDP-43 LC domain (WT), Triple-me (321–323-325) and Triple-me (323–325-327) variants assayed in (A) and (B). Scale bar = 25 μm.

Fig. 4.. Phase separation capacities of twenty six variants of the TDP-43 LC domain differing according to the replacement of a single amino acid residue with glycine.

Single amino acid residues within the ultra-conserved region of the TDP-43 LC domain were changed to glycine by conventional mutagenesis and expression in bacterial cells. Two of the twenty six positions already contained glycine (G335 and G338) in the native sequence and were mutationally changed to serine. Each protein was purified and tested for phase separation in buffer of neutral pH and physiologically normal monovalent salt ions. Light microscopic images were photographed 1 h subsequent to initiation of each reaction. In addition to assays performed in standard buffer, each protein sample was also evaluated for its capacity to phase separate in the presence of 0.4 M urea, 0.8 M urea and 1.2 M urea (Fig. S2). Scale bar = 25 μm.

Fig. 5.. Evidence of the particular importance of proline residue 320 of the TDP-43 LC domain.

(A) Schematic depicting of positions of four proline residues (280, 320, 349 and 363) and ALS-causing M337V variant within the LC domain of TDP-43. Darker shading marked region corresponds to the ultra-conserved region. (B) Protein samples corresponding to the native LC domain of TDP-43 (WT), four proline-to-glycine variants (P280G, P320G, P349G and P363G), and M337V variant were assayed for the formation of liquid-like droplets. Other than the P320G variant, which formed aggregated tangles, all other proline to glycine variants generated spherical droplets indistinguishable from those made from the native TDP-43 LC domain. The M337V variant maintained phase separation capacity but yielded misshaped droplets. Scale bar = 25 μm. (C) U2OS cells were stably transformed with vectors allowing for conditional, doxycycline-mediated expression of FLAG-tagged versions of native, full-length TDP-43 (WT), the ALS-causing M337V variant of TDP-43, or the P320G variant causative of severe precipitation. Single transformant clones were isolated in the absence of doxycycline and screened for the purpose of finding stable clones that expressed equivalent protein levels post-doxycycline induction. Cell lysates were recovered before and 24 hours after doxycycline induction, and western-blotted using a FLAG antibody or antibodies to GADPH as a loading control (left panel). Cell viability was monitored daily post-induction for each clone, revealing minimal effects from doxycycline-induced expression of either the native TDP-43 protein (WT) or the ALS-causing M33V variant, as compared with substantial deficits resulting from expression of the P320G variant (right panel). (D) Constitutive expression vectors encoding fusion proteins linking GFP to the amino terminus of full-length TDP-43 were transiently transfected into U2OS cells. Nucleus-restricted GFP was observed for native TDP-43 linked to GFP, as well as for the proline-to-glycine variants of residues 280, 349 and 363, and the ALS-causing M337V variant. Transient expression of the P320G variant led to predominantly cytoplasmic GFP staining, with residual nuclear staining appearing aberrantly punctate. Scale bar = 20 μm. (E) Semi-synthesis of variants of the TDP-43 LC domain, P320G and P320S, formed tangled precipitates upon tests of phase separation. Methyl capping of the peptide backbone nitrogen atom associated with the glycine residue of the P320G variant, or the peptide backbone nitrogen atom associated with the serine residue of the P320S variant, yielded proteins for which formation of phase separated liquid-like droplets was restored. Scale bar = 25 μm.

Fig. 6.. Effects of methyl-capping peptide backbone nitrogen atoms associated with disease-causing residues within the NFL head domain.

(A) Schematic representation of the strategy to prepare N^α-methyl amino acids (red color, red asterisks) incorporated NFL head domain and NFL full-length protein. Synthetic peptide thioesters corresponding to the amino-terminal twenty five residues of the NFL protein were prepared to contain: (i) the native sequence of NFL; (ii) Charcot-Tooth-Marie (CMT) mutational variants at position 8 or 22; or (iii) CMT variants methyl-capped at the peptide backbone nitrogen in the mutated amino acid. Native chemical ligation was used to produce full-length NFL protein or the isolated head domain bearing different synthetic peptides at the amino-terminus. (B) Full-length NFL proteins carrying P8L, P8Q and P8R variations, P22R, P22S or P22T variations, or N^α-methyl variants of these six residues, were incubated under conditions receptive to the formation of intermediate filaments. Filament assembly was monitored by transmission electron microscopy. Assembly assays for all six CMT variants yielded tangled, amorphous precipitates. Assembly assays for semi-synthetic CMT variants containing a methyl-capped nitrogen (P8meL, P8meQ and P8meR, P22meR, P22meS or P22meT) yielded homogenous intermediate filaments indistinguishable from those made from the native NFL protein (WT). Scale bar = 200 nm. (C) Synthetic peptides used to generate semi-synthetic, full-length NFL proteins were assayed for polymerization as monitored by acquisition of ThT fluorescence. All six peptides bearing a CMT-causing lesion (P8L, P8Q, P8R, P22R, P22S and P22T) yielded ThT curves indicative of time-dependent polymerization. No evidence of polymerization was observed for the parental peptide (WT), nor for any of the six CMT-causing variants (P8meL, P8meQ and P8meR, P22meR, P22meS or P22meT) that were modified to methyl-cap the peptide backbone nitrogen atom of the variant amino acid residue.

Fig. 7.. Effects of eliminating main-chain hydrogen bonding via amide-to-ester substitutions in the peptide backbone, and effects of replacing proline 8 of the NFL head domain with 5,5-dimethyl-L-proline (dmP).

(A) Structures of synthetic peptides with chemical variation at proline residue 8 of the NFL head domain highlighted in red. Synthetic peptides were prepared replacing proline residue 8 of the NFL head domain with leucine, L-leucic acid or dmP. (B) Semi-synthetic, full-length NFL proteins bearing the structures shown in (A) were assayed for assembly of bona-fide intermediate filaments. Reconstruction of the P8L CMT variant yielded tangled, amorphous precipitates. Amide-to-ester backbone substitution at leucine residue 8 (P8esL), thus eliminating a hydrogen bond donor from this site, facilitated assembly of intermediate filaments indistinguishable from those assembled from the native NFL protein. Replacement of proline residue 8 with dmP (P8dmP) allowed assembly of homogeneous intermediate filaments morphologically indistinguishable from those produced by the native protein (WT). Scale bar = 200 nm. (C) Synthetic peptides used to generate full-length NFL proteins were assayed for polymerization as monitored by acquisition of ThT fluorescence. The P8L CMT variant peptide exhibited clear evidence of polymerization not observed for the parental peptide bearing the native sequence of NFL (WT). The WT, P8esL, and P8dmP peptides showed no evidence of polymerization.

Fig. 8.. Restorative effects of eliminating main-chain hydrogen bonding by methyl-capping of peptide backbone nitrogen atoms of the P301S, P301T and P301L mutational variants of tau.

(A) At left, a schematic of the tau (2N4R) protein with the position of the synthetic tau peptide (residues 295–311) and three disease-causing variants (P301S, P301T and P301L) highlighted via a red asterisk. At right, the chemical structures of trans (2S,4R; P-trans) or cis (2S,4S; P-cis) conformer of 4-fluoroproline. (B) tau peptides with native sequence of tau residues 295–311 (WT), disease-causing variants (P301S, P301T and P301L), methyl-capped forms of the three disease-causing variants (P301meS, P301meT and P301meL) and P-trans and P-cis conformer variants were synthesized and assayed for polymerization as monitored by acquisition of ThT fluorescence. Each of the disease-causing variants led to strong increases in ThT fluorescence relative to WT. No such increase was observed if the peptide backbone nitrogen atoms associated with the disease variant residues were methyl-capped. No enhancement of ThT fluorescence was observed for peptides wherein proline residue 301 was replaced by either the P-trans or P-cis conformer of 4-fluoroproline. (C) Each of the peptides bearing a disease-causing substitution (P301S, P301T and P301L) led to the formation of distinct aggregates upon assay in the tau biosensor cell line. Whereas, WT tau peptide, the peptide nitrogen methyl-capped derivatives of disease-causing variants (P301meS, P301meT and P301meL), and P-trans and P-cis conformer variants didn’t induce endogenous tau to form distinct aggregates in the biosensor cells. Scale bar = 50 μm.

Fig. 9.. Restorative effect of eliminating the main-chain hydrogen bond by methyl-capping the peptide backbone nitrogen atom of the P298L mutational variant of hnRNPA2.

(A) At top, a schematic representation of the hnRNP2 LC domain with the position of the disease-causing mutant P298L (red asterisk) and synthetic peptide fragment (dark blue shading) highlighted. Below is an outline of the strategy to prepare semi-synthetic variants of the full-length LC domain of hnRNPA2. (B) The assembled hnRNPA2 LC domains contained the native sequence (WT), P298L variant and P298meL variant were tested for phase separation in an aqueous buffer. WT hnRNPA2 LC domain formed spherical, liquid-like droplets (left). The P298L variant formed distinctly misshapen droplets (middle). P298meL variant formed spherical droplets indistinguishable from those formed by the native protein (right). Scale bar = 25 μm. (C) Synthetic peptides used to assemble semi-synthetic proteins assayed in (B) were incubated in the presence of ThT as an assay of time-dependent polymerization. No evidence of fluorescence increase was observed for WT peptide. A large, time-dependent increase in ThT fluorescence was observed for P298L peptide, and no fluorescence increase was observed for P298meL peptide.