Advertisement
  • Loading metrics

Electrostatic Repulsion Governs TDP-43 C-terminal Domain Aggregation

Electrostatic Repulsion Governs TDP-43 C-terminal Domain Aggregation

  • Miguel Mompeán, 
  • Avijit Chakrabartty, 
  • Emanuele Buratti, 
  • Douglas V. Laurents
PLOS
x

TDP-43 is a protein with multiple crucial functions in RNA processing and mRNA-protein (mRNP) particle formation, two processes that are strongly implicated in amyotrophic lateral sclerosis (ALS) and frontotemporal lobar dementia (FTLD) [1]. The recently published structural characterization by Lim, Wei, Lu, and Song [2] of the complete (residues 262–414) C-terminal domain (CTD) of TDP-43 and three pathologically relevant variants, and the interaction of this region with membrane-mimetics and DNA oligos, builds on our characterization of a smaller Q/N-rich C-terminal segment (residues 341–367) [3,4]. Together, these studies mark a significant advance in our understanding of this important protein. In particular, the results by Lim, Wei, Lu, and Song [2] show that the CTD exists as an intrinsically disordered conformational ensemble, which can be nudged by certain conditions into forming a “hydrogel” (a putatively functional amyloid) or coerced by mutation into adopting hypothetically pathological amyloid structures. The ability of membrane-mimetics to induce helix formation in a hydrophobic segment of TDP-43's CTD and to increase the aggregation rate suggests a link between membrane interaction and pathological cytotoxicity.

Nevertheless, the interpretation of some of the data could be improved by additional considerations that we would like to address specifically in this comment. In particular, based on the ability of the TDP-43 CTD to interact with membrane-mimetics, the authors claim that membrane disruption could be a disease mechanism. This assertion is plausible in light of results obtained for other aberrant polypeptide aggregates implicated in Alzheimer's disease and type 2 diabetes (as cited by Song and coworkers). However, there are literally thousands of proteins that are known to interact with membranes without disrupting them. For this reason, it is crucial to carry out well-established tests already used for short CTD fragments [5] to determine if aggregates of the complete CTD of TDP-43 can disrupt membranes.

One singularly perplexing proposal of Song and coworkers involves the strong pH dependence of the CTD’s aggregation and amyloid formation. In our opinion, Lim et al. [2] strike a false chord when they propose that

“…in the TDP-43 prion-like domain at pH 4.0, most backbone amyloid protons are involved in forming hydrogen bond networks with side chain oxygen atoms…”

(page 14, second paragraph)

to account for the lack of aggregation and amyloid formation at pH 4.0. They also claim that these interactions would be intramolecular and break down at pH 6.8, leading to faster aggregation. However, this assertion seems to be at odds with the generally very small changes observed in the circular dichroism (CD) spectra and nuclear magnetic resonance (NMR) temperature coefficients over this pH range (see SF1 and SF6, ref. [2]). Furthermore, no mechanism is advanced to explain how these H-bonds weaken in a pH-dependent manner. Finally, this assertion is untenable considering that there are only 60 side chain oxygen atoms in the CTD, a number insufficient to form H-bonds to the majority of the 152 backbone HN groups as claimed.

In fact, regarding this specific issue, we believe that there is a simpler and more logical explanation for their observation.

The solubility of folded proteins has long been known to vary with the pH and is typically lowest at pH values at which the protein carries a small net charge [6]. The solubility increases at those pH values at which the protein molecules present a higher net charge (positive or negative). This idea is supported by results of protein variants carrying multiple charge substitution mutations (Fig 1A) [7]. Moreover, related results have been obtained for Aβ and a series of variants carrying charge substitution mutations [8]; namely, aggregation and amyloid formation depend on the proximity to each polypeptide's isoelectric point (Fig 1B). These results are especially pertinent, since Aβ, like the CTD of TDP-43, is an intrinsically disordered polypeptide with a marked tendency to form amyloid.

thumbnail
Fig 1. Protein aggregation and amyloid formation are favored by a low net charge.

Idealized dependence of the solubility (A) or amyloid formation (B) by proteins with isoelectric points at pH 4 (red), 7 (green), and 10 (blue). Adapted from Fig 15–2 of [6] and Fig 1 of [7] (panel A) and Fig 1 of [8] (panel B). Please see S1 Data, columns A–E and columns G–J for the underlying data of panels A and B, respectively.

https://doi.org/10.1371/journal.pbio.1002447.g001

Bearing these considerations in mind, we have calculated the net charge of the CTD of TDP-43 as a function of pH on the basis of its‎ content of charged termini and residues (1D, 2E, 7H, 1Y, 2K, and 5R) and the reported pKa values for disordered chains (Fig 2A, S1 Data) [9]. The net charge on this CTD construct is high (+12.1) at pH 4, but decreases modestly (to +10.3) by pH 5.0 as the carboxylate groups acquire negative charge, and decreases significantly at pH 6.8 (to +6.0), as the seven imidazolium groups in His264 and the C-terminal hexaHis tag residues lose their positive charges. Strikingly, in fact, this decreased net charge parallels the increased aggregation observed by Lim, Wei, Lu, and Song as the pH increases from 4.0 to 5.0 to 6.8. Therefore, we conclude that the pH dependence of amyloid formation by the CTD is governed by electrostatic repulsion (Fig 2B), in accord with previous results on Aβ and its charge substitution variants (Fig 1 in [8]).

thumbnail
Fig 2. Electrostatic repulsion modulates the aggregation of TDP-43’s C-terminal domain.

A. Calculated total net charge (black diamonds) on the TDP-43 C-terminal domain (CTD) as a function of pH. The contributions of different types of titratable groups are shown as colored symbols (C-terminal carboxylate = red open circles, Asp 406 = small red circles, Glu 271 and 362 = gold open triangles, His 264 and hexaHis tag = green squares, N-terminal amine = open crossed green squares, Tyr 374 = purple circles, Lys 263 and 408 = sky-blue triangles, and Arg 268, 272, 275, 293, and 361 = blue triangles. The red, gold, and blue arrows mark the net charge at pH 4 (+12.1), pH 5 (+10.3), and pH 6.8 (+6.0). The underlying data are shown in S1 Data, columns K–U. B. Model for the electrostatic control of amyloid formation. At pH extremes, polypeptides carry a high positive or negative net charge. The resulting unfavorable charge–charge interactions impede association. At intermediate pH values, the net charge is low and electrostatic repulsion is too weak to prevent oligomerization and amyloid formation, which is favored by strong hydrogen bond networks, hydrophobic interactions, or both.

https://doi.org/10.1371/journal.pbio.1002447.g002

Finally, regarding the pathological TDP-43 CTD variants studied, A315E would decrease the net charge over the pH range studied and Q331K would increase it. It is notable that A315E forms amyloid much faster than Q331K (see Figs 2 and 3 of [2]), as predicted by the charge-repulsion model. Considering that (i) aggressive Aβ mutations like “Arctic” (E22G), “Dutch” (E22Q), “Iowa” (D23N), and “Italian” (E22K) also decrease the net charge at physiological pH [8]; (ii) there is a correlation between higher net charge and slower kinetics of amyloid formation [10]; and (iii) charge–charge interactions modulate aggregation and amyloid formation in alpha synuclein [11], amylin [12], a β2-microglobin-derived peptide [13], huntingtin [14], superoxide dismutase [15], and Tau [16] and affect aggregation kinetics [11] and morphology [17], these results point to the general importance of electrostatic repulsion in governing the amyloid formation rather than H-bond formation as proposed by Lim, Wei, Lu, and Song [2].

Supporting Information

S1 Data. Supporting Data File 1.

This file includes the data used to generate Figs 1 and 2A.

https://doi.org/10.1371/journal.pbio.1002447.s001

(XLSX)

Acknowledgments

We thank Prof. Marta Bruix for critical comments.

Author Contributions

Conceived and designed the experiments: DVL. Wrote the paper: MM AC EB DVL.

References

  1. 1. Buratti E, Baralle FE. TDP-43: Gumming up neurons through protein-protein and protein-RNA interactions. Trends in Biochem Sci. 2012;37: 237–247.
  2. 2. Lim L, Wei Y, Lu Y, Song J. ALS-causing mutations significantly perturb the self-assembly and interaction with nucleic acid of the intrinsically disordered prion-like domain of TDP-43. PLoS Biol. 2016;
  3. 3. Mompeán M, Buratti E, Guarnaccia C, Brito RM, Chakrabartty A, Baralle FE, Laurents DV. Structural characterization of the minimal segment of TDP-43 competent for aggregation. Arch Biochem Biophys. 2014;545: 53–62. pmid:24440310
  4. 4. Mompeán M, Hervás R, Xu Y, Tran TH, Guarnaccia C, Buratti E, et al. Structural evidence of amyloid fibril formation in the putative aggregation domain of TDP-43. J Phys Chem Lett. 2015;6: 2608–2615. pmid:26266742
  5. 5. Liu GC, Chen BP, Ye NT, Wang CH, Chen W, Lee HM, et al. Delineating the membrane-disrupting and seeding properties of the TDP-43 amyloidogenic core. ChemComm. 2013; 49: 11212–11214.
  6. 6. Tanford C. Physical chemistry of macromolecules. 1st ed. New York: J. Wiley and Sons, NY; 1961.
  7. 7. Shaw KL, Grimsley GR, Yakovlev GI, Makarov AA, Pace CN. The effect of net charge on the solubility, activity and stability of Ribonuclease Sa. Prot Sci. 2001;10: 1206–1215.
  8. 8. Guo M, Gorman PM, Rico M, Chakrabartty A, Laurents DV. Charge substitution shows that repulsive electrostatic interactions impede the oligomerization of Alzheimer's amyloid peptides. FEBS Lett. 2005;579: 3574–3578 pmid:15964569
  9. 9. Grimsley GR, Scholtz JM, Pace CN. A summary of the measured pK values of the ionizable groups in folded proteins. Prot Sci. 2009;18: 247–251.
  10. 10. Chiti F, Stefani M, Taddei N, Ramponi G, Dobson CM. Rationalization of the effects of mutations on peptide and protein aggregation rates. Nature 2003;434: 805–808
  11. 11. Buell AK, Galvagnion C, Gaspar R, Sparr E, Vendruscolo M, Knowles TP, et al. Solution conditions determine the relative importance of nucleation and growth processes in α-synuclein aggregation. Proc Natl Acad Sci USA. 2014;111: 7671–7676. pmid:24817693
  12. 12. Patil SM, Alexandrescu AT. Charge-based inhibitors of amylin fibrillization and toxicity J Diabetes Research. 2015;2015:946037.
  13. 13. Valette NM, Radford SE, Harris SA, Warriner SL. Phosphorylation as a tool to modulate aggregation propensity and to predict fibril architecture. ChemBioChem. 2012;13: 271–281 pmid:22174034
  14. 14. Mishra R, Hoop CL, Kodali R, Sahoo B, van der Wel PCA, Wetzel R. Serine phosphorylation suppresses huntingtin amyloid accumulation by altering protein aggregation properties J Mol Biol. 2012;424: 1–14. pmid:22999956
  15. 15. Abdolvahabi A, Shi Y, Rhodes NR, Cook NP, Martí AA, Shaw BF. Arresting amyloid with Coulomb's law: Acetylation of ALS-linked SOD1 by aspirin impedes aggregation. Biophys J. 2015;108: 1199–1212. pmid:25762331
  16. 16. Inoue M, Kaida S, Nakano S, Annoni C, Nakata E, Konno T, Morii T. Phosphorylation regulates fibrillation of an aggregation core peptide in the second repeat of microtubule-bonding domain of human tau. Bioorg & Med Chem. 2014;22: 6471–6480.
  17. 17. Vetri V, Foerà V. The route to protein aggregate superstructures: Particulates and amyloid-like spherulites. FEBS Lett. 2015;589: 2448–2464. pmid:26183565