Learnings from Protein Folding Projected onto Amyloid Misfolding

Viewpoint Cite This: ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

pubs.acs.org/chemneuro

Learnings from Protein Folding Projected onto Amyloid Misfolding Sreeprasad Sreenivasan and Mahesh Narayan*

Downloaded via NOTTINGHAM TRENT UNIV on August 30, 2019 at 11:44:55 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

Department of Chemistry and Biochemistry, The University of Texas at El Paso, El Paso, Texas 79968, United States ABSTRACT: The 1990s saw a revolution in our understanding of the protein folding pathways of both disulfide-bondcontaining proteins and purely conformational folders. High-resolution maps of the folding trajectories, made possible by innovative experimental design, revealed the presence of multiple intermediates, their formation and consumption, and the network of interactions between them that lead to the formation of the folded protein from its unfolded state. The same level of detail has heretofore remained elusive as far as the amyloid aggregation pathways of prion-like proteins are concerned. Nevertheless, a recent development that led to the resolution of intermediates in amyloidogenic trajectories, without resort to their separation, is likely to not only advance our basic understanding of the atomic- and molecular-level interactions guiding amyloid misfolding but also impact interventional efforts in their associated pathologies. KEYWORDS: Protein folding, amyloid proteins, oligomers, protofibrils, conformational folding, oxidative folding, reductive unfolding, atomic force microscopy, vibrational spectroscopy

T

In concert with experiment, molecular dynamics simulations and energy fields also “came-of-age” in that more accurate representations of all-atoms became more tractable as did longer simulation runs. These developments also furthered the study of proteins other than those primarily α-helical in nature. The rigorous application of the experimental and computational innovations described above were key to establishing that the native, or native-like, conformation could often be accessed by multiple routes.2,4 It revealed the presence of onpathway species, kinetically trapped intermediates, and deadend structures. In several cases, the rate-constants leading to, and away from, all species populating the folding trajectory were worked out and the network of inteconnecting pathways driving the transformation of U to N were mapped.2,4 The past few years have seen a redirection of such efforts to understanding the process(es) by which prion-like amyloid proteins aggregate. Amyloid beta (Aβ), α-synuclein, tau, and mutant Huntingtin protein are a few examples of amyloidogenic proteins capable of self-seeding the templated misfolding of their soluble, monomeric (M) counterparts. Their aggregation-prone trajectories typically involve the formation of oligomeric (O) and protofibrillar (Pf) intermediates that precede the formation of mature fibrils (F). The techniques applied to unravel their soluble-monomer to insoluble-fibril transformation include solid-state NMR, IR, and fluorescence studies and molecular dynamics simulations. While significant inroads have been made toward describing interactions driving amyloidogenic pathways, the degree of granularity required for generating accurate road maps of all interconversions constituting the M → O → Pf → F pathway remains unsatisfactory. This is ironic considering that the relatively slow rate (minutes, hours, days) by which amyloid monomers eventually convert into mature fibrils should, in principle and in practice, represent an advantage for dissecting

he 1980s and the 1990s witnessed an exponential growth in our understanding of the mechanisms by which proteins fold.1,2 The efforts of groups led by Baldwin, Scheraga, Creighton, Dobson, Fersht, Wolynes, Englander, and their academic progeny generated detailed blueprints of the protein folding pathways of now classical proteins such as bovine pancreatic trypsin inhibitor, bovine pancreatic ribonuclease A, lysozyme, barstar, etc. (Figure 1). Other proteins like the villinheadpiece lent themselves to computational efforts geared toward describing the interactions that dictate the formation of the native conformation (N) from its unfolded state (U).3 Key to the progress made then were innovative experiments designed to identify the intramolecular interactions that drove the folding process.2,4 Notable examples included the exploitation of disulfide bond formation to sufficiently slow conformational folding which in turn facilitated the stabilization of intermediates that could be studied by crystallography or NMR. The detailed examination of these intermediates, including the free energies required to unfold them, provided clues to the entropic and enthalpic factors that dictated the experimental folding pathways of the aformentioned proteins.2 Other inroads that were crucial, particularly to purely conformational folders, included stopped-flow techniques hyphenated with fluorescence, UV−vis spectrosopy, or H/D NMR. The ability to sample protein folding on the millisecond time scale and “breathing” of proteins on larger time scales provided clues to both initial events that transpire upon introduction of the unfolded polypeptide into a folding milieu and to the nature and “strength” of noncovalent interactions that prevail in the folded protein. By strategically exploiting proline isomerization which provides a kinetic barrier to folding, employing site-directed mutagenesis that can impact folding rates, and by identifying chain-folding initiation sites via distance measurements, it was possible to add additional details to existing folding pathways.4 We eventually learned about the existence of intrinsically disordered proteins through these studies. © XXXX American Chemical Society

Received: August 12, 2019 Accepted: August 13, 2019

A

DOI: 10.1021/acschemneuro.9b00445 ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

Viewpoint

ACS Chemical Neuroscience

Figure 1. Folding pathways of bovine pancreatic trypsin inhibitor (top left; image reproduced with permission from DOI: 10.1073/ pnas.1503909112); bovine pancreatic ribonuclease A (top right); hen-egg white lysozyme (bottom left); and barnase (bottom right; image reproduced with permission from DOI: 10.1016/S0092-8674(02)00620-7).

Figure 2. Hypothesized evolution of monomeric Aβ (T0) into mature fibrils T6).

The holy grail of amyloid-associated protein misfolding studies is to obtain a high-resolution, and accurate, road map of interactions driving the conversion of M to O, Pf, and F. This would constitute knowledge of the presence of other intermediates that may populate the pathogenic transformation, their structures, the rate constants associated with the interconversions between species, the identification of onpathway conformations, kinetic traps, and dead-end species, the flux of each trajectory leading to interventional targets, etc. Recently, using atomic force microscopy hybridized with vibrational spectroscopy, researchers were able to discriminate between several types of conformers present in a 36 h time point of the soluble → fibril reaction mixture of Aβ without resorting to the separation of constituents therein.5 The

the molecular details of the constitutive driving factors. The inadequacies, nevertheless, could stem from relatively lowresolution data typically obtained from insoluble aggregates, use of isolated methods or techniques that lack relevancy, and ensemble signal-averaging that is typical of many spectroscopic tools that, even collectively, have clouded the optics associated with amyloid-forming studies. Molecular dynamics simulations of the self-templating phenomenon also fall short because of the relatively short simulation times that can be achieved. These limitations result in oversimplified aggregation trajectories, conflicting pathways, or only hypothetical road maps of the process, a scenario further complicated by the sensitivity of intermediates to extrinsic conditions (solvent, pH, salt, temperature, etc.). B

DOI: 10.1021/acschemneuro.9b00445 ACS Chem. Neurosci. XXXX, XXX, XXX−XXX

ACS Chemical Neuroscience

■

landmark foray led to the identification of intermediates leading to both oligomers and protofibrils and to the observation that loosely structured oligomers could form protofibrils via two independent mechanisms, allowing the authors to construct a preliminary road map of the transformation while simultaneously eliminating other hypothetical scenarios. These results are paradigm-shifting in that a heretofore generally postulated unidimensional (linear) pathway is in fact at least two-dimensional in nature. It is to be anticipated that a time-dependent sampling of the aggregation milieu (Figure 2) will help (i) establish the true process by which oligomers, protofibrils, and mature fibrils evolve from the fully soluble monomeric amyloid, (ii) identify other intermediates that may populate the trajectory, (iii) discriminate between on- and off-pathway intermediates and kinetics traps, (iv) provide rate-constants governing the formation and consumption of every species, and (iv) provide structural resolution of the aforementioned species. Importantly, outcomes from such a study are likely to bridge current knowledge gaps in our understanding of the true soluble → toxic species conversion process that has so far created barriers for prophylactic and therapeutic intervention (evident from the limited success achieved using smallmolecules). Considering that toxic oligomers and protofibrils, rather than mature fibrils, are associated with the onset and pathogenesis of their respective neurodegenerative disorders, the need to resolve the atomic and molecular underpinnings leading to the generation of these intermediates is emergent. The future is exciting for amyloid misfolding studies.

■

Viewpoint

REFERENCES

(1) Scheraga, H. A. (2015) My 65 years in protein chemistry. Q. Rev. Biophys. 48, 117−177. (2) Narayan, M., Welker, E., Wedemeyer, W. J., and Scheraga, H. A. (2000) Oxidative Folding of Proteins. Acc. Chem. Res. 33, 805−812. (3) Onuchic, J. N., and Wolynes, P. G. (2004) Theory of protein folding. Curr. Opin. Struct. Biol. 14, 70−5. (4) Baldwin, R. L. (2008) The Search for Folding Intermediates and the Mechanism of Protein Folding. Annu. Rev. Biophys. 37, 1−21. (5) Lipiec, E., Perez-Guaita, D., Kaderli, J., Wood, B. R., and Zenobi, R. (2018) Direct Nanospectroscopic Verification of the Amyloid Aggregation Pathway. Angew. Chem., Int. Ed. 57, 8519−8524.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Mailing address: Dept. of Chemistry and Biochemistry, UT El Paso, 500 W. Univ. Ave., El Paso, TX 79968. Phone: 915-747-6614. Fax: 915-7475748. ORCID

Sreeprasad Sreenivasan: 0000-0002-5728-0512 Mahesh Narayan: 0000-0002-2194-5228 Author Contributions

M.N. wrote the article. S.S. and M.N. contributed to the development of the the work within. Funding

M.N. acknowledges support from NIH 1SC3 GM111200 01A1, the UTEP College of Science (Research Enhancement Award), and the NIH MBRS SCORE Border Biomedical Research Center at The University of Texas at El Paso. This facility is supported by the Grant # 2G12MD007592 and Grant # 5G12MD007592 from the Research Centers in Minority Institutions program of the National Institutes on Minority Health and Health Disparities, a component of the Research Center in Minority Institutions (RCMI) program. Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS M.N. thanks Angela Lopez Velazquez for preparing the figures. M.N. acknowledges support from Mrs. Holly and Dr. Eddie Vazquez (The El Paso Pain Center). C

DOI: 10.1021/acschemneuro.9b00445 ACS Chem. Neurosci. XXXX, XXX, XXX−XXX