Protein Misfolding and Aggregation: Understanding Amyloid


Abstract

The formation of "amyloid-like" fibrils is associated with numerous human pathological diseases. Alzheimers disease and Parkinsons disease, are two of the more recognizable disorders in a group of well over 20 such diseases that have been directly linked to the self assembly of peptides into these ordered amyloid fibrils. (1-6) Although each of these diseases are caused by the assembly of seemingly unrelated peptides/proteins, the aggregates have many common structural and mechanistic details, and as such, a full understanding of these would be largely beneficial, and will be discussed herein.

Introduction

  • Importance of Folding/Misfolding


Their are approximately 100,000 protein sequences in the human genome. In order for a protein to assume proper biologically function, these proteins must fold into their stable native states. Considering that their are over 100 distinct folds available for a peptide backbone, the question becomes how do these proteins reliably assemble. According to the classical thoguht experiment by Levinthal (7), if a polypeptide chain is made of 100 residues, and we assume only two possible conformations for each residue, statistics predicts that there exists about 10^30 possible overal conformations of the polypeptide. A stepwise search for the native state considering all possible conformations, would then take longer than the age of the universe. Since we know that proteins fold reliably and reproducibly in short time scales, there must be some unifying driving forces that enable a more stochastic search for the native state. Over the last few decades, one of the primary goals of biophysicists has been to more fully understand and uncover these driving forces behind protein folding (and conversely misfolding, and aggregation.) Attempts to obtain the rules governing the folding and unfolding of peptides are generally considered as part of the effort to develop a complete and thorough understanding of the role that peptide sequence, solvent environment, hydrogen bonding and various other intermolecular forces play in stabilizing/ destabilizing various secondary and tertiary structures. (3) The intricate, precise, and rapid folding and unfolding of peptides/proteins is necessary for normal cellular acitivity. It follows then, that failure to fold correctly and stabilize, wil lead to the malfunctioning of biological processes and may ultimately lead to disease. (1-8) Hence, interest in these conformational changes arises partially from the role protein misfolding, self -aggregation and subsequent fibrilization plays in many biological patghological diseases. It is now well established that many neurodegenerative diseases, such as Parkinson's and Alzheimers, as well as the systemic and localized amyloidoses are the result of self aggregation and fibril formation of proteins and peptide segments. (1-6) In addition to the negative implications of misfolding and peptide aggregation as it related to disease, there also exists wide ranging positive applications for self assembly in the field of biotechnology. If the structures and mechanistic principals governing the propensity of certains proteins to form self assembled architectures were well understood, one could forsee the design of bio-compatible materials with great potential for use in biologically inspired applications such as drug delivery and tissue engineering. (9,10)
  • Aggregation


The term "protein aggregation" has been given many different meaning in literature. (2, 4,11) As the term has developed, it can loosely be defined as the self assembly or intermolecular association of misfolded proteins or peptide segments that results in insoluble precipitates. Protenacious aggregates can be of two broad types: ordered or disordered. Ordered aggregates typically form highly organized fibrilliar species, wheras disordered aggregates typically form more amorphous conglomerates, an example of which is inclusion bodes. (4) Aggregation in vivo is association with the plaque deposits which decrease the number of native proteins and hence decrease normal activity, create cytotoxic activity, and interupt neural networks.

Considering the above mentioned Levinthal paradox, (7) the question becomes how do these proteins "find" this aggregated state? The current oppinions on this all involve the idea of an energy landscape that is shapped like a funnel, with the global minimum representative of the native state conformaiton. Because on average, native state interactions are more energetically stable that non-native conformation interactions, the protein is able to find this low energy structure rapidly. Misfolded states can then be thought of as representing local minima (or saddle points) in this free energy funnel. In this context, the aggregated state is merely one of these minima that are trapped by high energy transition states due to the large amount of intermolecular forces keeping the aggregate thermodynamically stable. (2,8,12,13)


Amyloid

  • Structure
    beta_filament.jpg
    Figure 1: Amyloid Fibril. Taken and modified from Ref 14

The largest class of protein aggregate related pathologies result from the aggregation of misfolded or intrinsically disordered proteins (IDP's) into insoluble and ordered filaments called "amyloid fibrils". The term amyloid. or amyloid fibril, is widely used to refer to a class protenacious deposits that share certain biophysical and ultrastructural characteristics. Specifically, amyloid fibrils have a "cross Beta spine", in which the individual beta strand peptide backbones are orthogonal to the fibrillar axis, and the sheets are therefore parallel to this axis (6). The antiparallel Beta sheets themseleves self assemble into a bundle of twisted "proto-filaments", groups of which can be pictured as then further intertwining to form the final unbranched amyloid fiber (Figure 1, taken and modified from reference 14). The helical structure is a direct result of the right handed twist of Beta shets (15).

The stability of amyloid-like aggregrates and fibrils is generally related to interstrand H bonding between peptides, formation of interstrand ionic bonds, hydrophobic interactions, and aromatic interactions such as pi-pi stacking. Many mechanistic studies of this phenomena involve using model polypeptide systems that include the primary peptide sequence which is thought to be the "pathological" or "amyloidogenic" sequence involved in each amyloid fibril disease. (16) Some examples of extensively used model systems are (16, 17):
1) yeat prion protein Sup35 "GNNQQNY"
2) Inslet amyloid (Type II diabetes) "NFGAIL"
3) prion protein "AGAAAGA"
3) Amyloid Beta (Alzheimers) "KLVFFAE"
4) Beta2-microglobulin (renal amyloidoses) "DWSFYLLYTEFT"

Despite the sequence diversity of the various known amyloid fibril forming proteins and peptides, and the diverse symptoms of each associated disease, this cross Beta architecture of the resulting fibrils is remarkably simillar in all cases. Morphologically, the final fibers have been shown by various X-ray diffracttion studies to be composed of two to six unbranched protofilaments (continuously stacked and twisted Beta sheets) of 2-5 nm in diamater. This results in fibrils that are typically 6-12nm in diamater. (4,18,20) Optically, amyloid fibers all show a specific green-gold birefringence upon staining with the dye Congo Red. (2, 6,16)

The common secondary, tertiary, and morphological structure of these fibrils is of particular consequence when one considers that the fibrils originate from a variety of different peptides and proteins that are unrelated in residue sequence or peptide length. In addition to the variation in polypeptide sequence of the orginating protein, the native state protein (prior to misfolding or aggregation) may be rich in many other conformations such as alpha helix or beta helix or may be unfolded (11,20)
It is the ubiquitous nature of these fibrils that has led to the hypothesis that the mechanism of aggregation and fibril formation follow certain general rules. The exact pathway to aggregation and fibril formation is still under debate, however, the structural diversity of the amoloidegenic prone proteins coupled with the striking similarity of the final fibril deposit suggests that considerable structural rearrangement must occur in order for fibrilization ot take place. In rigid globular proteins, these types of large conformational can not occur, and so it has been hypothesized that the first step to fibril formaiton is the destabilization of the native structure. (4,11, 21-22, 23-25)

Moreover, recent studies have shown that many short peptides and globular proteins that are unrelated to pathological diseases can also form amyloid fibrils under certain conditions (22). This has lead to the suggestion that amyloid fibril formation is a general feature of all polypeptides. (11,13,26), and that given enough time and the right conditions, all proteins may aggregate. Most recently, it has even been suggested that the native state of proteins is merely a mesostate (8,20), and that the ordered amyloid fibrils represents the global minimum of the gree enrgy landscape of protein folding (12,16).

  • Mechanisms and Pathway

Various mechanisms exist for describing amyloid formation. These include subsequent monomer adiition/nucleated polymerization (27,28-30), templated assembly (31,32), and quantitative structure-activity relationship models (23). Although the exact pathway of aggregation and subsequent fibrilization is still being uncovered, all of the proposed mechanisms have key features in common. The starting compound for all mechanisms is the monomeric form of the protein. Recent studies of in vivo globular amyloid proteins have shown that the protein monomer is typically in a non-native conformation, and hence the first step in these cases is a partial un-folding to a de-stabilized state. (21-22) Specifically, Uversky and Fink (20) have shown evidence for the formation of partially mis-folded intermediates (with increased Beta content) that enhance the kinetics of alpha-synuclein fibrilization. Partial unfolding and or misfolding of the native protein prior to fibrilization has also been seen in various other proteins such as the prion protein PrP (23), immunoglobin (24), and lysozyme (25) to name a few.

One common feature of nearly all mechanisms propsed for amyloid aggregation is the existence of a slow lag phase or nucleation phase (Figure 2), indicating nucleation of the monomer. What exactly comprises the nucleus is still debated, but in general, it is hypothesized that it is composed of the misfolded active monomers associated intermolecularly into a Beta sheet rich oligomers. MD simulations have shown that ensembles of soluble Beta Sheet oligomers are formed wih non-naive H-bonding registry, and that the first step is then rearrangemnt into ordered Beta sheets. It has even been suggested that the intermediate oligomers may be the cytotoxic species in many amyloid pathologies. (17,34) This lag phase can take several minutes to days, depending on the system. Nucleation of the monomer is a thermodynamically reversible process in which proteins are added sequentially, forming soluble oligomers, untill a critical nucleus or "seed" is formed. (12,30) This lag phase can be eliminated in in vitro studies by addition of preformed aggregates (i.e. by seeding). The nucleus can then further elongate and self-assemble into the aforementioned protofilaments composed of stacked helical beta sheets. Thid growth phase is typically irreversible, and exponentially faster then the initial nucleation phase as the formation of amyloid aggregates is stable. The stability of thesee aggregates results from the collective nature of the intermolecular forces such as hydrophobic interactions and H-bonding. The protofilaments may form via self templated growth where the ends of existing filaments recruits soluble molecules into aggregates that can themselves multimply through secondary nucleation mechanisms. (32)

nucleation.png
Figure 2: Kinetic Trace for Aggregation. Taken from ref (12) and modified.



  • Biophysical and Spectroscopic Analysis

There exist many biophysical tools used to monitor the formation of amyloid aggregates. These include flourescence (35,36), dynamic light scattering (37), atomic force microscopy (38),Ultra-Violet Circular Dichrosim spectroscopy (39,40), Infared spectroscopy (41,42), and Raman spectroscpy (42). For a full description of all these techniques please refer to Ref (14). Herein, I will breifly review some of the more widely used of these methods.

Atomic Force Microscopy (AFM) allows the direct imaging of peptide fibrils on the surface. Typically, solutions of aggregations at various times are dried onto a mica plate. The use of AFM is particularly beneficial for observing the final supramolecular fibril structure and length scales. (38)
Dynamic light scattering is one of the most widely used methods to directly measure properties affacted by aggregation over time, and hence aggregation kinetics. In this method, the intensity autocorrelation function of scattered light from the sample as it diffuses is monitored. (9) This allows for determination of diffusion coefficients and the hydrodynamic radius of the amyloid aggregates as they form. Using DLS, it has been shown that amyloid fibrils have radii of 18-22 nm. (37)

Flourescene Spectroscopy can be conducted on amyloid fibrils either with an extrinsic or intrinsic flourophore, depending on the specific protein system being analyzed. A classic experiment by Levinne et al showed that Thioflavin T will only associate with final aggregated fibrils of an Alzheimers derived amyloid beta protein. This association resulted in a new excitation at 450nm and emission at 482nm that is distinct from the dye absorption and emission, and hence, monitoring these over time can be an efficient way of obtaining kinetic traces for amyloid fibrilizaiton. (35)

Circular Dichroism Spectroscopy (CD) is sensitive to the chrality of a compound, and hence has been used extensively in the characterization of secondary structure of proteins. Specifically, far UV-CD is widely used to study the time evolution of amyloid aggregation by monitoring the Beta Sheet content over time. Compliation of varioius CD spectra of proteins/peptides with known secondary structures (determined from Xray analysis) have established that individual secondary structure conformations show very distinct CD spectra. (39) In particular, Beta sheets have a characteristic CD absorption spectra with a negative maxima at aproximately 218nm. Since the amyloid fibril is composed of stacked beta-sheets, this provides a suitable tool for qualitative analyzation.(43,37)

Infrared Spectroscopy is also widely used for monitoring the progression of amyloid fibril formation. The amide I normal mode of polypeptides is located at approximately 1620-1640cm-1. This mode is generally described as being composed of carbonyl C=O stretch with some admixture of C-N stretch. Due to vibrational coupling of these local modes along the peptide backbone, the frequency of the amide I is extremely sensitive to secondary structure (i.e. peptide backbone angles phi and psi). The structural sensitivity of the amide I normal mode makes it ideal to study conformational changes such as those that occur upon formation of Beta sheet rich aggregates in amyloid Beta. In particular, Beta sheet aggregates give give distinct amide I bands. Spectroscopically, amide I band intensity decreases and a redshift in frequency is observed upon formation of Beta sheets oligomers (41,42). Decatur et al used isotope edited IR to show that after incubation, Beta strands align such that the hydrophillic residues associate with solvent and the hydrophobis residues pack tightly together forming sites available for H bonding anf formation of Beta sheets. Monitoring the further downshift of the the band upon in register formation of Beta sheets, they showed that in order for the prion protein H1 to form amyloid aggregates it must first form soluble in register beta sheet oligomers via intersheet rearrangement. (42)


Conclusion


Many human pathological diseases are associated with the irreversible formation of protein and peptide aggregates and fibrils. Such diseases vary widely in their symtoms, but share in common that they are the result of amyloid-like aggregation. The structure of all of these aggregates are remarkably simillar considering that the diseases are so varied and that the primary amino acid sequence and native secondary structure of the amyloidogenic specis are unrelated. This has lead many to hypothesize that amyloid formation is a generic feature of all polypeptides, and may represent a global minimum in the energy landscape for protein folding. This being said, unraveling the rules for this ubiquitous aggregation process would be largely beneficial in order to develop ways to prevent it.

References

  1. Bellotti, V., Mangione, P., & Stoppini, M. (1999) Cell. and Mol. Life Sci. 55, 977-991 DOI
  2. Dobson, C., Chiti, F. (2005) Annu. Rev. Biochem. 75, 333-66 DOI
  3. Chiti, F., Dobson, C. (2009) Nature 5. 15-22 DOI
  4. Fink. A. (1998) Folding and Design 3, R9-R23 DOI
  5. Measey, T., Schweitzer-Stenner, R. (2006) J. Am. Chem. Soc., 128, 13324-13325 DOI
  6. Nelson, R., Sawaya, M.R., Balbirnie, M., Madsen, A., Riekel, C., Grothe, R., Eisenberg, D. (2005) Nature 435, 773-778 DOI
  7. Levinthal Paradox Wikipedia
  8. Dobson, C., (2003) Nature 426, 884-890 DOI
  9. Ulijin, R., Bibi, N., Jayawarna, V., Thorton, P. D., Todd, S.J., Gough, J. E. (2007) Materials Today. 10, 40-48 DOI
  10. Kyle, S., Aggeli, A., Ingham, E., McPherson, M. (2009) Trends Biotech. 27, 423-433 DOI
  11. Uversky, V. N., Oldefield, C. J., Dunker, A. K. (2008) Annu. Rev. Biophys. 37, 215-246 DOI
  12. Gazit, E., (2002) Angew. Chem. Int. Ed. 41, 257-259 DOI
  13. Dobson, C., (2003) Nature. 426, 884-890 DOI
  14. Morris, A. M., Watzky, M. A., Finke, R. G., (2009) Biochim. Biophys. Acta. 1794, 375-397 DOI
  15. Sunde, M., Serpell, L.C., Bartlam, M., Fraser, R., Pepys, M. B., Blake, C. (1997) J. Molec. Bio., 273, 729-739 DOI
  16. Gazit, E. (2005) FEBS, 272, 5971-5978 DOI
  17. Santini, S., Mousseau, N., Derreumaux, P. (2004) J. Am. Chem. Soc. 126, 11509-11516 DOI
  18. Kirktadze, M. D., Condron, M. M., Teplow, D. B., (2001) J. Molec. Bio. 312, 1103-1119 DOI
  19. Kirschner, D., Abraham, C., Selkoe, D. J. (1986) PNAS, 83, 503-507
  20. Uversky, V.N., Fink, A., (2004) Biochim. Biophys. Acta. 1698, 131-153 DOI
  21. Wetzel, R. (1996) Cell 86, 699-702 DOI
  22. Sandal, M., Valle, F., Tessari, I., Mammi, S., Bergantino, E., Samori, B. (2008) PLos Biol. 6, 99-108 DOI
  23. Jackson, G. S., Hosszu, L. L. P., Power, A., Hill, A. F., Kenney, J., Collinge, J. (1999) Science 283, 135-1937 DOI
  24. Wetzel, R., (1994) PNAS 91, 5446-5450 Link
  25. Dobson, C., Radford, S., Evans, P. (1992) Nature 358, 302-307 DOI
  26. Chiti, F., Webster, T. P., Taddei, N., Clark, A., Stefani. M. (1999) PNAS, 96, 3590-3594
  27. Roberts, C. J., Andrews, J.M. (2007) J. Phys. Chem. B. 111, 7897-7913 DOI
  28. Hofrichter, J., Ross, P. D., Eaton, W. A. (1974) PNAS 71, 4864-4868 Link
  29. Beaven, G. H., Gratzer, W. B., Davies, H. G. (1969) FEBS. 11, 37-42 DOI
  30. Jarret, J. T., Lansbury,P. T. (1993) Cell 73, 1055-1058
  31. Uratani, Y., Asakura, S., Imahori, K. (1972) J. Mol. Biol. 67, 85-98 DOI
  32. Knowles, T. P. J. (2009) Science 326, 1533-1537 DOI
  33. Chiti, F., Dobson, C. (2004) J. Molec. Biol. 341, 1317-1326 DOI
  34. Lambert, M. P., Barlow, A. K., Chromy, B. A., Edwards, C., Freed, M. (1998) PNAS 95, 6448-6453 Link
  35. Levinne, H. (1993) Proteine Science 2, 404-410 DOI
  36. Munishkina, L. A., Fink, A. L., (2007) Biochim. Biophys. Acta. 1768, 1862-1885 DOI
  37. Modler, A. J., Gast, K., Lutsch, G., Damaschun, G. (2003) J. Mol. Biol. 325, 135-149 DOI
  38. Harper, J.D., Wong, S. S., Lieber, C. M., Lansbury, J. (1997) Chem. Biol. 4, 119-125 Link
  39. Kelly, S., Jess, T. J., Price, N. (2005) Biochim. Biophys. Acta. 1751, 119-139 DOI
  40. Reed, J., Reed, T. A., (1997) Anal. Biochem. 254, 36-40 DOI
  41. Zandomeneghi, G., Krebs, M., McCammon, M. G., Fandrich, M. (2009) Protein Sci. 13, 3314-3321 DOI
  42. Decatur, S. M., Petty, S. A. (2005) PNAS 102, 14272-14277 DOI
  43. Measey, T., Schweitzer-Stenner, R., (2006) J. Am. Chem. Soc., 128, 13324-13325 DOI